Cyclometalated complexes of ruthenium, rhodium and iridium as catalysts for transfer hydrogenation of ketones and imines

Article abstract of DOI:10.1002/adsc.201100152

A library of organometallic compounds derived from primary and secondary amines cyclometalated by ruthenium(II), rhodium(III) and iridi- um(III) was tested in the asymmetric transfer hydrogenation of a number of ketones and imines. All compounds displayed high catalytic activity for the reduction of ketones under mild conditions. The most enantioselective catalysts were based on secondary amines containing two asymmetric carbon atoms bound to the nitrogen atom. For the reduction of aryl alkyl ketones [Ar(C=O)R where R=CH₃ or CH₂R′] the cyclometalated ruthenium and rhodium derivatives of the (2R,5R)-2,5-diphenylpyrrolidine ligand displayed the best results with respect to activity and selectivity (ees up to 97%). However, for the reduction of aryl tert-alkyl ketones [Ar(C=O)R′′ in which R′′ is a tertiary alkyl group] the best catalyst was a ruthenium compound derived from bis[(R)-1-phenylethyl]amine, allowing the reduction of isobutyrophenone and cyclohexyl phenyl ketone which were both reduced with high enantioselectivities (ees up to 98%). This shows that the cyclometalated compounds have a high substrate specificity. In addition, acyclic and cyclic imines were reduced with good selectivities by both rhodium(III) and iridium(III) metalacycles built up with (2R,5R)-2,5-diphenylpyrrolidine. Copyright

Full text of DOI:10.1002/adsc.201100152

FULL PAPERS
DOI: 10.1002/adsc.201100152
Cyclometalated Complexes of Ruthenium, Rhodium and Iridium
as Catalysts for Transfer Hydrogenation of Ketones and Imines
Nicolas Pannetier,a Jean-Baptiste Sortais,a Jean-Thomas Issenhuth,a
Laurent Barloy,a Claude Sirlin,a Alexandre Holuigue,b Laurent Lefort,b
Lavinia Panella,b Johannes G. de Vries,b,* and Michel Pfeffera,*
a
Institut de Chimie, UMR 7177, CNRS/Universitꢀ de Strasbourg, Laboratoire de Synthꢁses Mꢀtallo-Induites, 4, rue Blaise
Pascal, 67000 Strasbourg, France
Fax : (+33)-3-6885-1526; e-mail: pfeffer@unistra.fr
DSM Innovative SynthesisB.V., a unit of DSM Pharma Chemicals, P.O. Box 18,6160 MD Geleen, The Netherlands
b
Fax : (+31)-46-476-7604; e-mail: Hans-JG.Vries-de@dsm.com
Received: March 7, 2011; Revised: May 23, 2011; Published online: October 6, 2011
Abstract: A library of organometallic compounds de- the reduction of aryl tert-alkyl ketones [Ar
rived from primary and secondary amines cyclometa- in which R’’ is a tertiary alkyl group] the best catalyst
lated by ruthenium(II), rhodium(III) and iridi- was a ruthenium compound derived from bis[(R)-1-
um(III) was tested in the asymmetric transfer hydro- phenylethyl]amine, allowing the reduction of isobu-
ACHTUNGTRENN(UNG C=O)R’’
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
genation of a number of ketones and imines. All tyrophenone and cyclohexyl phenyl ketone which
compounds displayed high catalytic activity for the were both reduced with high enantioselectivities (ees
reduction of ketones under mild conditions. The up to 98%). This shows that the cyclometalated com-
most enantioselective catalysts were based on secon- pounds have a high substrate specificity. In addition,
dary amines containing two asymmetric carbon acyclic and cyclic imines were reduced with good se-
atoms bound to the nitrogen atom. For the reduction lectivities by both rhodium
ACHTUGTNREN(UNG III) and iridiumACHTUNGTRNE(NUGN III) met-
of aryl alkyl ketones [Ar(C=O)R where R=CH3 or alacycles built up with (2R,5R)-2,5-diphenylpyrroli-
AHCTUNGTRENNUNG
CH2R’] the cyclometalated ruthenium and rhodium dine.
derivatives of the (2R,5R)-2,5-diphenylpyrrolidine
ligand displayed the best results with respect to ac- Keywords: imines; iridium; ketones; rhodium; ruthe-
tivity and selectivity (ees up to 97%). However, for nium; transfer hydrogenation
Introduction
tallic complexes were built up with chiral bidentate li-
gands such as diamines, amino acids or amino alco-
Amongst the many catalytic applications of Ru(II) in hols.[2,6] The prototype of such complexes was dis-
organic synthesis (hydrogenation, oxidation, metathe- closed in the early 1990s by Noyori et al., who found
sis…),[1] hydrogen transfer has become very popular that a Ru(II) half-sandwich complex ligated by chiral
this last decade. It is indeed a very powerful tool in N-tosyl-1,2-diphenylethylenediamine was a very pow-
asymmetric synthesis, as it allows the formation of erful catalyst precursor for the target reaction.[6b]
new stereocentres from a great variety of prochiral or- Noyori et al. also proposed the concept of metal-
ganic substrates on the way towards useful products ligand bifunctional catalysis after mechanistic stud-
for fine chemical industry.[2,3] This method has many ies.[7] They suggested that the substrate interacts with
advantages compared to alternative competing tech- both a metal hydride unit and an acidic NH function
nologies: (i) the use of safe and low-cost reducing of the ligand via a 6-membered ring transition state.
agents, (ii) an operational simplicity, and (iii) the re-
Organometallic compounds containing a metalacy-
duction of safety constraints associated with the use cle[8] as catalyst precursors in hydrogen transfer are
of molecular hydrogen or stoichiometric amounts of still not very numerous.[9] However, a few groups
metal hydrides.[2,4,5]
have recently described promising catalytic results
[10]
[10a,11]
[12]
[13]
À À
À À
À
À À
A large number of catalyst precursors has already using N C N, P C P,
C P, and C N N
been successfully tested for their ability to catalyse cycloruthenated or cycloiridiated[14] ligands.
the hydrogen transfer reaction, leading to the reduc- Our contribution to this field was triggered by the
tion of ketones and imines. Most of these organome- discovery that cycloruthenated chiral benzylamines
2844
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2011, 353, 2844 – 2852

Cyclometalated Complexes of Ruthenium, Rhodium and Iridium as Catalysts for Transfer Hydrogenation
are also attractive catalyst precursors for the reduc- hydrogen source. We have now extended this study to
tion of acetophenone. They operate under mild condi- a larger set of such compounds based on new chiral
tions (room temperature), are robust (they can be amines. Additionally, we have also prepared some of
used several times provided that more substrate is the related rhodiumACTHNUTRGENN(UG III) and iridiumCAHTUNGTREN(NUGN III) derivatives
added) and allow the achievement of high TONs (up stabilised by a h5-pentamethylcyclopentadienyl (Cp*)
to 30,000 for reactions run at 808C) and high selectiv- ligand. The choice of the cyclometalated ligands used
ities (up to 90%).[15] Mechanistic studies were per- in this library came from a preliminary study whereby
formed on one of these systems: they indicated that a library of 30 in situ-made ruthenium derivatives was
the hydrogen transfer process proceeds via the forma- formed, following the cyclometalation procedure,
tion of a substrate-catalyst complex between the Ru with commercially available optically pure primary
hydrides and the substrate, which is analogous to the and secondary amines. The catalytic properties of
6-membered ring proposed by Noyori et al.[7] In the these non-isolated complexes have been previously
present paper, we present full data concerning the tested in the reduction of acetophenone in parallel
catalytic activities of a library of cyclometalated de- using high-throughput screening.[15]
rivatives [M=Ru(II), Rh
G
G
From this study chiral benzylamine derivatives
a metalacyclic unit, and we highlight their specific be- emerged as the ligand class that induced the highest
haviours.
enantioselectivities. Consequently, we have focused
on this structural motif to prepare the library of com-
pounds of the present study that is depicted in
Figure 1. The catalytic activities of compounds 1, 4
and 6 in the reduction of prochiral ketones to alcohols
have been published.[15] The library was now extended
to the known compounds 2, 5 and 7–13 whose synthe-
Results
Library of Catalyst Precursors
We have shown in preliminary form[15] that a few ses and complete characterisations were described re-
ruthenium complexes obtained via the cyclometala- cently but for which we have not disclosed any cata-
tion reaction on primary or secondary amines were lytic properties as yet.[16,17]
good catalyst precursors for the reduction of aceto-
We synthesised the new compound 3, in order to
phenone in the presence of base and of 2-propanol as check the influence of the steric hindrance at the
Figure 1. Library of cyclometalated compounds.
Adv. Synth. Catal. 2011, 353, 2844 – 2852
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2845

Nicolas Pannetier et al.
FULL PAPERS
Table 1. Asymmetric reduction of acetophenone, A with the
chiral centre on the enantioselectivity of the transfer
hydrogenation. It is worth mentioning that the (R)-a-
isopropylbenzylamine used for the synthesis of 3 was
obtained from the corresponding (S)-alcohol (ee>
98%) that was synthesised by the enantioselective re-
duction of the commercially available isobutyrophe-
none, using compound 6 (see below). The synthesis of
3 was performed following a procedure similar to that
used for 1 and 2.[16a] The spectroscopic data of 3 are
similar to those of 1 and 2, i.e., in solution this com-
pound exists as 2 diastereoisomers, which differ from
one another by the configuration of the ruthenium
atom.
library of catalyst precursors.[a]
Entry
Catalyst
Time
Yield [%][b]
ee [%][b]
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
8
8
9
10
11
12
13
13
40 min
30 min
2 h
10 min
2 h
30 min
2 h
30 min
3 h
2 h
2 h
1.5 h
2 h
1 h
95
96
92
90
93
96
95
92
95
87
90
96
5
39 (S)[c]
51 (S)
60 (S)
57 (S)[c]
54 (S)
76 (S)[c]
77 (S)
88 (S)
90 (S)
90 (S)
34 (S)
89 (S)
9[d]
10[e]
11[d]
12
13
14
15
16[g]
The best Ru-containing catalysts were 6 and 8, i.e.,
those obtained via the cyclometalation of bis[(R)-1-
phenylethyl]amine and (2R,5R)-2,5-diphenylpyrroli-
dine by [(h6-C6H6)RuCl2]2. Consequently, we attempt-
ed the synthesis of the corresponding cyclometalated
[f]
–
24
0
64
89 (S)
–
8 (S)
2 h
30 h
RhACHTUNGTRENNUNG(III) and IrACHTUNGTRENNUNG(III) compounds with these two ligands
[a]
Conditions: ketone (10À1 m) in 10 mL of 2-propanol;
using the Rh and Ir dimers [Cp*MCl2]2 as precursors.
Full details about this chemistry have been published
elsewhere.[17] Contrary to the Ru case,[16a] we found
that the benzylamine ligands were easily dehydrogen-
ated under the cyclometalation conditions once they
[substrate]:ACTHNUGRTENNUG[catalyst]:ACHTUNGTERN[NUGN tBuOK]=100:1:5, 208C.
[b]
Determined by chiral GC. The absolute configuration of
the major enantiomer is given in parentheses.
Published data.[15]
[c]
[d]
[e]
[f]
Ketone (10À1 M) in 100 mL of 2-propanol.
Reaction performed at 08C.
were bound to the RhACHTNUTGRENNUG(III) and IrACHTUTGNREN(NUGN III) centres, leading
to cyclometalated imine derivatives. Thus, whereas
bis[(R)-1-phenylethyl]amine led to intractable mix-
tures of products, (2R,5R)-2,5-diphenylpyrrolidine af-
forded a mixture of cyclometalated amine and imine,
compounds 10–13. Compounds 11 and 13 could be
11 that was used was obtained with the racemic imine
ligand.
Reaction performed at 608C.
[g]
isolated pure, however 10 and 12 were always conta- the iridium-containing catalyst (entry 14) that was sig-
mined by variable amounts of 11 and 13, respectively. nificantly less active than the other compounds.
We independently prepared 11 and 13 directly from
the imine ligand, and found that they were either
Transfer Hydrogenation of Acetophenone
poorly active (11, entry 13) or inactive (13, entry 15)
at room temperature. A mediocre activity was ob-
All organometallic compounds (1–12) derived from served for 13 at 608C in 2-propanol (entry 16), more-
primary and secondary amines (Figure 1) were found over the enantioselectivity was very poor (8%) at this
to be good-to-excellent catalyst precursors for the re- temperature. As pointed out earlier, so far we were
duction of acetophenone (substrate A, Figure 2) unable to obtain pure 10 and 12 as both of these com-
under mild conditions, as far as activity was con- pounds were contaminated by their corresponding cy-
cerned (see Table 1), with the noticeable exception of clometalated imine, i.e., 11 and 13, respectively.
However, these first results show that the activities
and enantioselectivities observed in the transfer hy-
drogenation of acetophenone can reasonably be as-
signed to the cyclometalated pyrrolidine, i.e., 10 and
12 only. It is reasonable to assume that 11 and 13 are
also inactive at room temperature with other more
difficult substrates. Throughout this paper, we shall
thus refer to the mixtures (10, 11) and (12, 13) as
compounds 10 and 12, respectively.
Compounds 2 and 3 differ from 1 by the size of the
substituent at the benzylic position. The substitution
of a methyl by an ethyl and then by an isopropyl
group each time enhanced the enantioselectivity by
ca. 10% with no significant modification of the activi-
Figure 2. Library of ketone substrates.
ty (entries 1–3). However the ees obtained with all
2846
asc.wiley-vch.de
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2011, 353, 2844 – 2852

Cyclometalated Complexes of Ruthenium, Rhodium and Iridium as Catalysts for Transfer Hydrogenation
catalysts, based on primary amines remained modest Substrate Scope of the Best Catalysts
(entries 1–5, maximum=60% for entry 3), though it
was somewhat higher than that of the other com- The catalyst precursors that showed the best perfor-
pounds obtained with primary amines (entries 1–5). mance in the reduction of acetophenone, i.e., 6, 8, 10
The same trend was observed for the rhodium com- and 12, were further tested on a set of prochiral ke-
pound 9 which displayed an even lower selectivity tones (Figure 2). The results under various reaction
than that of 1 (entry 11). It was thus confirmed here conditions (reaction time, temperature and concentra-
that the best catalysts were those obtained with sec- tion) are described in Table 2.
ondary amines, especially the derivatives of the pyrro-
lidine ligand. This is true for both the ruthenium (8) and 10, whereas 6 was less selective and 12 less active.
and the rhodium (10) compounds. The rhodacycle 10 was the most active catalyst of the
Acetylfuran (B) was best reduced by the catalysts 8
Good enantioselectivity was also obtained with cat- library for the reduction of B. Indeed, when the con-
alyst 8 at higher substrate-to-catalyst ratio (entry 9, centration of the ketone was 0.1M and the experi-
S/C=100) but no significant improvement was ob- ment was run at room temperature, the reaction was
served upon lowering the temperature (entry 10).
complete after 15 min and the ee dropped fast from
We also checked that a base must be present. No 82% to 57% after 6 h. This last result is consistent
reaction took place after 48 h in the absence of base. with the high activity of 10 which also catalyses the
There is however no influence of the nature of the reverse reaction, thus leading to an erosion of ee; 10
base (t-BuOK or KOH), nor of its concentration catalysed this reverse reaction at a much higher rate
(varying the base:catalyst ratios from 1.5 to 10) on ac- than the other cyclometalated complexes. Thus, the
tivity or selectivity.
best results were obtained by lowering the tempera-
ture, the reaction time and the concentration of the
substrate (entry 7). With both ruthenacycle 8 and rho-
Table 2. Reduction of the substrates B–G with catalysts 6, 8, 10 and 12.[a]
Entry
Substrate
Catalyst
Temperature
Time [h]
Substrate concentration (M)
Yield [%][b]
ee [%][b]
1
2
3
4
5
6
7
8
B
B
B
B
B
B
B
B
C
C
C
C
C
C
D
D
D
E
E
E
E
F
6
8
8
8
r.t.
r.t.
08C
r.t.
08C
r.t.
08C
r.t.
r.t.
r.t.
08C
r.t.
08C
r.t.
r.t.
r.t.
r.t.
r.t.
r.t.
r.t.
r.t.
r.t.
r.t.
r.t.
r.t.
r.t.
r.t.
r.t.
0.75
1.5
6
7
7
2
0.75
3
0.75
2
4
0.5
3
6
0.5
0.5
7
1.5
1.5
1.5
1.5
1.5
1.5
1.5
3.5
4
10À1
10À1
10À1
10À2
10À2
10À2
10À2
10À1
10À1
10À2
10À2
10À1
10À2
10À1
10À1
10À1
10À1
10À1
10À1
10À1
10À1
10À1
10À1
10À1
10À1
10À1
10À1
10À1
95
82
86
74
97
94
98
37
94
75
88
90
81
91
57
56
39
97
74
88
58
97
95
75
38
32
18
32
86[c]
87[c]
92
89
90
89
92
89
79
8
10
10
12
6
6
6
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
82
87[c]
86[c]
93
8
10
12
8
10
12
6
85
96[c]
91
96
98[c]
48[c]
85
8
10
12
6
81
98[c]
48[c]
75
F
F
F
G
G
G
8
10
12
6
8
10
74
89[c]
45[c]
36
4
4
[a]
Conditions: solvent=2-propanol; [substrate]:
Determined by chiral GC.
ACHUTGTNREN[UNG catalyst]:ACHUTNGTREN[NUGN t-BuOK]=100:1:5.
[b]
[c]
Published data.[15c]
Adv. Synth. Catal. 2011, 353, 2844 – 2852
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2847

Nicolas Pannetier et al.
FULL PAPERS
dacycle 10, running the reaction at 08C led to an in-
crease of the enantioselectivity but also, more surpris-
ingly, to a slight increase in the activity (see entries 4,
5 and 6, 7, respectively).
The chemoselectivity of our catalysts was tested in
the reduction of the 4-vinylacetophenone C. We
found that the ketone function was reduced selective-
ly with all catalysts, whilst the vinylic group remained
intact. The enantioselectivity of 6 and 10 towards the
reduction of C could be significantly improved by
lowering the temperature, and by lowering the con-
centration of the substrate resulting in 87% and 93%
ee, respectively (entry 11 and 13).
Figure 3. Library of imines.
High ees were obtained in the reduction of sub-
strate D using 8, 10 or 12 as catalysts (entries 15–17).
However, the activity of these complexes was medio-
The reaction conditions were optimised using sub-
cre. Only very short reaction times (entries 15 and 16) strates H and I, with catalyst 6. The best source of hy-
could be used for 8 and 10 as longer reaction times in- drogen proved to be a 1:1 mixture of formic acid and
duced significant losses in enantioselectivities with no triethylamine from which traces of water had to be
improvements in the yield of the reaction. The iridi- strictly excluded in order to avoid the hydrolysis of
um derivative 12 was here again rather sluggish the imine function, prior to its reduction. To achieve
(entry 17). Note that 6 was neither active nor selective this, a 5:2 HCOOH/NEt3 mixture was azeotropically
for the reduction of this substrate.
distilled and dry NEt3 was added. A co-solvent had to
An unexpected result was obtained as we found be used in order to solubilise the substrate and di-
that 6 was very efficient for the reduction of the sub- chloromethane was found to be the best choice. The
strates that bear a tertiary carbon atom a to the complexes 1–7 and 9 were tested under the optimised
ketone function in E and F. It was both very active reaction conditions, but none of them gave satisfying-
and selective leading to 98% ees (entries 18 and 22); ly enantioselectivities and they were not used further.
8 was rather inefficient in this instance (<50% ee, en- Indeed the catalysts derived from primary amines led
tries 19 and 23), 10 was a reasonably good catalyst. to amines with ees inferior to 20%, whereas 6 and 7
However, the cyclopropyl ketone G defied this trend; led to mediocre results (ees<40%). Additionally, a
whereas with catalyst 6 a good ee was obtained, its re- fast erosion of ee was observed when using these cata-
duction was slow with 6, 8 and 10.
lysts.
On the other hand, better results were obtained
with catalysts 8, 10 and 12 for the reduction of the
substrates H - J (Table 3).
Reduction of Imines by Transfer Hydrogenation
The iridium compound 12 displayed good activity
Among the methods available to generate optically for all 3 substrates with ees ranging from 57 to 77%.
active amines, asymmetric hydrogen transfer on
imines is of high fundamental and economic impor-
Table 3. Reduction of the imines H–J with catalysts 8, 10
tance. However, it has been much less explored until
and 12.[a]
recently.[18] In the early 1990s, Bꢃckvallꢄs group re-
Entry
Substrate
Catalyst
Yield [%][b]
ee [%][c]
ported the first hydride transfer on imine using
[RuCl2A(PPh3)3], K2CO3 and 2-propanol as the hydride
CHTUNGTRENNUNG
1
2
3
4
5
6
7
8
9
H
H
H
I
I
I
J
J
J
8
52
70
90
11
95
100
21
42
64
60
91
76
71
14
77
60
34
57
source.[19] Noyori then reported the first asymmetric
example of transfer hydrogenation on a few imine
substrates, which occurred with excellent yields and
enantioselectivities.[20] Rhodium complexes related to
Noyoriꢄs catalyst were also successfully used for this
reaction.[21] Some other examples of transfer hydroge-
nation of imines have been reported during the latest
years,[22] including an interesting application in a key
step of a total enantioselective synthesis.[23] We then
decided to test our cyclometalated compounds in the
enantioselective reduction of prochiral imines, and, to
this end, we selected a small set of substrates depicted
in Figure 3.
10
12
8
10
12
8
10
12
[a]
Conditions: imine (0.016M) in 2.5 mL of CH2Cl2,
HCO2H/Et3N (1/1 molar) (0.5 mL), the ratio
[substrate]:CAHTUNGTERN[UNNG catalyst]=100:1, 208C. Time=19 h.
Determined by H NMR.
Determined by chiral HPLC.
[b]
[c]
1
2848
asc.wiley-vch.de
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2011, 353, 2844 – 2852

Cyclometalated Complexes of Ruthenium, Rhodium and Iridium as Catalysts for Transfer Hydrogenation
We have been puzzled by the results observed using
10. For substrate H, it is the most enantioselective
catalyst (ee=91%, entry 3) whereas with substrate I
and the even more related substrate J, it is the worst
one. We have no rational explanation so far for this
result, except it is another illustration that in asym-
metric catalysis, there is no such a thing as a “univer-
sal catalyst”.
Discussion
The cyclometalated amines of the library display
good-to-excellent activities in the transfer hydrogena-
tion of ketones and imines (conversion >90% under
standard conditions for at least one of the substrates
tested). The only ones that showed good-to-excellent
enantioselectivities (ee up to 98%) are those based on
secondary chiral amines. However, no catalyst in our
Figure 4. The likely catalytically active species 8’a and 8’b
derived from 8, and the inactive ones 8’c and 8’d.
library appears to be consistently better than all the are to be re-examined in light of the mechanism of
others for the small selection of prochiral substrates the reaction that was recently studied in detail.[24] It
tested. As usually in enantioselective catalysis, each was shown that the true intermediate is likely to be a
substrate requires a specific catalyst.
ruthenium hydride that interacts with the ketone via
À
The moderate enantioselectivities of cyclometalated both a Ru H and an NH unit according to Noyoriꢄs
primary amines are in line with the mechanistic stud- outer sphere mechanism.[7] Assuming that the stereo-
ies performed recently:[24] the main conclusion that chemistry of the hydrides follows that of the acetoni-
was drawn from these studies was that ruthenacycles trile complexes 6 or 8 that we recently established via
based on chiral primary benzylamines are unable to 1H NMR and X-ray diffraction studies,[16] likely repre-
induce very high enantioselectivities, because of the sentations of the catalysts might well be related to
formation of a diastereomeric mixture of Ru hydride those depicted in Figure 4 (for 8 only). In diastereo-
À
À
intermediates leading to catalysts that display compet- isomers 8’a and 8’b the Ru H and N H vectors are
itive rates but opposite enantioselectivities. cis, therefore they display the correct geometry for
At this stage of our study, we can conclude that 8 the formation of the 6-membered transition state with
and 10 are the best of our catalysts for the reduction the incoming substrate. On the other hand 8’c and 8’d
À
À
of ketones having a methyl or an alkyl group connect- display antiparallel N H and Ru H bonds, and the
ed via a secondary carbon to the C=O unit, and that 6 mechanism above can obviously not take place.
is the best catalyst for the reduction of those ketones
It is interesting to note that the major isomers of 6
that bear an alkyl group connected via a tertiary and 8 (respectively, 69% and 76%), having an aceto-
carbon atom to the carbonyl unit. Although the iridi- nitrile ligand instead of a hydride ligand, present anti-
À
um complex 12 is a mediocre catalyst for the reduc- parallel Ru NCMe and NH bonds. Assuming that re-
tion of ketones, it is rather efficient for the reduction placing acetonitrile with a hydride will not dramatical-
of imines.
ly affect the ratio of isomers, this would mean that
The performances of our best catalysts for the re- the active catalysts are the minor isomers, and hence
duction of ketones (6, 8 and 10) are in line with those the activity of our catalysts could roughly be twice as
of other compounds described recently. For instance, high as that measured.
Noyoriꢄs catalyst[25] was known to be a sluggish and
As the configuration of the alcohol is mainly S, the
moreover non-selective catalyst for the reduction of E most active species should be that having the RN,SM
or F. Morris et al.[26] got significantly improved results configuration as this one should lead specifically to
using a so-called ꢅreverse-tetheredꢄ catalyst derived the S enantiomer of the alcohol.[24] The SN, RM isomer
from Noyoriꢄs catalyst by linking its amino group to leading to the R product is thus much less efficient or
the (h6-arene)Ru unit, obtaining thus excellent ees for much less abundant on the basis of molecular models
the reduction of the substrates E and F (94% and (a strong steric interaction between the non-metalat-
95%, respectively). These latter figures were, howev- ed phenyl unit and the benzene ring is apparent).
er, slightly lower those reached with the catalyst 6
(98% in both cases).
The mechanism of the transfer hydrogenation with
the RhCp* and IrCp* derivatives of the cyclometalat-
The efficiencies of our ruthenium catalysts 6, 7 and ed amines is not known. However, based on literature
8 derived from the metalation of secondary amines data,[27] it is very likely that an outer sphere mecha-
Adv. Synth. Catal. 2011, 353, 2844 – 2852
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2849

Nicolas Pannetier et al.
FULL PAPERS
on a Bruker Daltonics MicroTOF spectrometer. Elemental
analyses were carried out at the Service Central d’Analyse
du CNRS, Vernaison.
nism analogous to the one depicted for the Ru com-
pound is operative as well. No crystal structure of the
pyrrolidine complexes 10 and 12 could be obtained
yet because of the existence of the mixture of com-
pounds observed during the synthesis of these com-
Typical Procedure for the Catalytic Transfer
Hydrogenation of Ketones in Schlenk Tubes
1
plexes. However, the H NMR spectra revealed that
both precatalysts 10 and 12 exist as one diastereomer
only (with a RM, RN configuration). This isomer is the
one having the NCMe and the NH antiparallel to
each other. It is here again rather likely that the
active catalysts are also very minor ones.
The catalyst (10 mmol) was dissolved in 2-propanol (10 mL)
under argon, and the ketone (1 mmol) was added, followed
by t-BuOK (5.6 mg, 50 mmol); the mixture was stirred at
room temperature. The reaction was periodically monitored
by GC on an HP 5890 Series II chromatograph; samples
were filtered over silica gel using Et2O as eluent. The con-
versions and ee values were determined by GC using a
chiral capillary column (Chiraldex b-PM, 50 mꢀ0.25 mm).
Conclusions
Typical Procedure for the Catalytic Transfer
Hydrogenation of Imines
We have synthesised a library of chiral metalacycles
[M=Ru(II), RhACHTUNGTRENNUNG(III) and IrCAHTUNGTRNE(NUGN III)] which were tested as
enantioselective transfer hydrogenation catalysts on a
set of different substrates using first high-throughput
screening as preliminary experiments, and then
Schlenk tube techniques. We observed good activities
for most of the tested substrates and high-to-very-
high enantioselectivities for some of them, the most
selective ones being those compounds derived from
cyclometalation of secondary amines. As often in
asymmetric catalysis, we observed a high specificity of
a given catalyst for a type of substrate with no easy
way to predict a matching catalyst-substrate couple.
The use of metals such as Rh or Ir enabled a diversifi-
cation of the substrate field, with promising results
for imines. This family of cyclometalated catalysts has
the advantage of being easy to synthesise in one step
from chiral commercial ligands. They can be used in
situ and display good robustness, which are two quali-
ties required for routine applications.
To a stirred solution of the imine substrate (1 mmol) and
the catalyst precursor (5 mmol) in CH2Cl2 (2 mL), was added
a freshly prepared solution of HCO2H/Et3N (1/1 molar,
0.5 mL). After stirring under argon at room temperature for
24 h, an aliquot was diluted with Na2CO3 solution and ex-
tracted twice with EtOAc. The extracts were washed with
brine, dried over MgSO4, and evaporated under vacuum. An
aliquot in hexane/2-propanol (90:10) was analysed using a
chiral HPLC column (Chiralcel OD 250ꢆ4.6 mm).
Typical Procedure for High-Throughput Catalytic
Transfer Hydrogenation
The experiment was performed in a liquid handling robot
(Zinner Lizzy), placed in a glove box. Stock solutions of cat-
alysts (1.12 mM), substrates (60 mM), and t-BuOK (5.6 mM)
were prepared in 2-propanol under argon. A few drops of
CH2Cl2 were used to dissolve entirely the catalyst. Aliquots
of 1 mL of each stock solution were mixed in wells placed
on a microplate. 1 mL 2-propanol was added in each well
and the solutions were placed on an orbital shaker and vor-
texed at room temperature for 2 h. Aliquots (0.5 mL) were
put in GC vials and mixed with 0.1 mL of glacial acetic acid
(10% in 2-propanol) to stop the reaction. Samples were di-
luted in EtOAc (1 mL) and submitted to GC analysis. The
conversions and ee values were determined by GC using a
chiral capillary column (Chrompack CP-Sil 5CB, 25 mꢀ
0.25 mm).
Experimental Section
Experiments were carried out under an argon atmosphere
using a vacuum line. Diethyl ether and pentane were dis-
tilled over sodium and benzophenone, dichloromethane and
acetonitrile over calcium hydride, and methanol over mag-
nesium under argon immediately before use. Column chro-
matography was carried out on Merck aluminium oxide 90
standardised. (R)-1-Phenyl-2-methylpropylamine was syn-
thesised following a published procedure for the amination
of alcohols, the optically active alcohol used being that ob-
tained through the reduction of substrate F by catalyst 6,
see entry 22, Table 2.[28] All substrates for the reduction of
ketones and imines were commercially available or prepared
according to published procedures. NMR spectra were ob-
tained at room temperature on a Bruker AC-300 spectrome-
ter and referenced to SiMe4. 1H NMR spectra were recorded
at 300.13 MHz and 13C{1H} NMR spectra (broadband decou-
pled) at 75.48 MHz. The NMR assignments were supported
by 2D spectra. Multiplicity: s=singulet, d=doublet, m=
multiplet, br=broad signal. ES-mass spectra were recorded
Synthesis of {(h6-C6H6)Ru
(NCCH3)}(PF6) (3)
ACHUTGTNRENN[UG C6H4-2-(R)-CHACHTUNGTNER(NUGN iPr)NH2]-
N
ACHTUNGTRENNUNG
To a suspension of [RuACTHNUTRGNEUNG
(h6-C6H6)Cl2]2 (0.298 g, 0.6 mmol),
NaOH (0.048 g, 1.2 mmol) and KPF6 (0.442 g, 2.4 mmol) in
CH3CN (15 mL) was added (R)-1-phenyl-2-methylpropyl-
AHCTUNGTREGaNNUN mine (0.021 g, 0.6 mmol) and the mixture was stirred at
208C under argon during 72 h. The resulting green suspen-
sion was filtered over celite, concentrated under vacuum
and filtered over standardised Al2O3 (12ꢆ3 cm) using
CH3CN as eluent. A light green fraction was collected, con-
centrated under vacuum and vigorously stirred with 20 mL
of hexane during 2 h, in order to extract the residual free
amine. The CH3CN layer was evaporated under vacuum.
The residue was then redissolved in a minimum amount of
2850
asc.wiley-vch.de
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2011, 353, 2844 – 2852

Cyclometalated Complexes of Ruthenium, Rhodium and Iridium as Catalysts for Transfer Hydrogenation
[6] a) V. Rautenstrauch, X. Hoang-Cong, R. Churlaud, K.
Abdur-Rashid, R. H. Morris, Chem. Eur. J. 2003, 9,
4954; b) S. Hashiguchi, A. Fujii, J. Takehara, T. Ikariya,
R. Noyori, J. Am. Chem. Soc. 1995, 117, 7562; c) K. Ev-
eraere, A. Mortreux, M. Bulliard, J. Brussee, A. van der
Gen, G. Nowogrocki, J.-F. Carpentier, Eur. J. Org.
Chem. 2001, 275–291; d) D. Carmona, M. P. Lamata,
L. A. Oro, Eur. J. Inorg. Chem. 2002, 2239.
[7] R. Noyori, M. Yamakawa, S. Hashiguchi, J. Org. Chem.
Figure 5. Numbering scheme for compound 3.
2001, 66, 7931.
[8] A metalacycle is characterised by the presence of a
cyclic unit in which a metal-carbon bond is stabilised
by intramolecular coordination of an heteroatom-metal
bond. Typically the cyclometalation reaction of a ligand
affords metalacycles.
[9] J.-P. Djukic, J.-B. Sortais, L. Barloy, M. Pfeffer, Eur. J.
Inorg. Chem. 2009, 817.
[10] a) P. Dani, T. Karlen, R. A. Gossage, S. Gladiali, G.
van Koten, Angew. Chem. 2000, 112, 759; Angew.
Chem. Int. Ed. 2000, 39, 743; b) M. Albrecht, B. M.
Kocks, A. L. Spek, G. van Koten, J. Organomet. Chem.
2001, 624, 271.
[11] a) D. Amoroso, A. Jabri, G. P. A. Yap, D. G. Gusev,
E. N. Dos Santos, D. E. Fogg, Organometallics 2004, 23,
4047; b) M. Gagliardo, P. A. Chase, S. Brouwer,
G. P. M. van Klink, G. van Koten, Organometallics
2007, 26, 2219.
[12] W. Baratta, P. Da Ros, A. Del Zotto, A. Sechi, E. Zan-
grando, P. Rigo, Angew. Chem. 2004, 116, 3668; Angew.
Chem. Int. Ed. 2004, 43, 3584.
[13] a) W. Baratta, G. Chelucci, S. Gladiali, K. Siega, K. To-
niutti, M. Zanette, E. Zangrando, P. Rigo, Angew.
Chem. 2005, 117, 6370; Angew. Chem. Int. Ed. 2005, 44,
6214; b) W. Baratta, M. Bosco, G. Chelucci, A. DelZot-
to, K. Siega, M. Toniutti, E. Zangrando, P. Rigo, Orga-
nometallics 2006, 25, 4611.
CH2Cl2 and a yellow solid precipitated upon addition of n-
pentane; yield: 150 mg (52%). 1H NMR (see Figure 5 for
numbering system; <300 MHz, CD3CN): major diastereo-
isomer (56%), d=7.79 (d, 1H, H-6, 3JH,H =7.1 Hz), 7.05–
6.85 (m, 2H, H-4+H-5), 6.81 (m, 1H, H-3), 5.57 (s, 6H, h6-
C6H6), 5.18 (br, 1H, NH), 3.76 (m, 1H, H-7), 2.87 (br, 1H,
3
NH), 2.37 (m, 1H, CHMe2), 1.14 (d, 3H, Me, JH,H =7.0 Hz),
0.64 (d, 3H, Me, 3JH,H =6.9 Hz); minor diastereoisomer
(44%), d=7.88 (d, 1H, H-6, 3JH,H =7.4 Hz), 7.05–6.85 (m,
3H, H-3+H-4+H-5), 5.61 (s, 6H, h6-C6H6), 4.40 (br, 1H,
NH), 4.11 (br, 1H, NH), 3.62 (m, 1H, H-7), 1.72 (m, 1H,
3
CHMe2), 0.97 (d, 3H, Me, JH,H =6.8 Hz), 0.87 (d, 3H, Me,
3JH,H =6.8 Hz); 13C{1H} NMR (75 MHz, CD3CN): major
isomer, d=165.7 (C-1), 150.0 (C-2), 140.3 (C-6), 127.3 (C-5),
124.3 (C-4), 124.0 (C-3), 87.8 (h6-C6H6), 70.0 (C-7), 32.7
(CHMe2), 20.0 (Me), 17.6 (Me); minor isomer, d=167.6 (C-
1), 148.2 (C-2), 140.3 (C-6), 127.5 (C-5), 124.5 (C-4), 123.7
(C-3), 87.8 (h6-C6H6), 72.0 (C-7), 31.2 (CHMe2), 20.1 (Me),
14.5 (Me); anal. calcd. for C18H23N2PF6Ru·2.5CH2Cl2
(725.5): C 33.93, H 3.89, N 3.86; found : C 34.20, H 3.82, N
4.18; ES-MS : m/z (%)=369.09 (4) [M]+, 328.10 (100)
[MÀCH3CN]+.
Acknowledgements
[14] C. Azerraf, D. Gelman, Chem. Eur. J. 2008, 14, 10364.
[15] a) J.-B. Sortais, V. Ritleng, A. Voelklin, A. Holuigue,
H. Smail, L. Barloy, C. Sirlin, G. K. M. Verzijl, J. A. F.
Boogers, A. H. M. de Vries, J. G. de Vries, M. Pfeffer,
Org. Lett. 2005, 7, 1247; b) J.-B. Sortais, L. Barloy, C.
Sirlin, A. H. M. de Vries, J. G. de Vries, M. Pfeffer,
Pure Appl. Chem. 2006, 78, 457; c) T. Jerphagnon, R.
Haak, F. Berthiol, A. J. A. Guyet, V. Ritleng, A. Holui-
gue, N. Pannetier, M. Pfeffer, A. Voelklin, L. Lefort, G.
Verzijl, C. Tiarabiono, D. B. Janssen, A. J. Minnaard,
B. L. Feringa, J. G. de Vries, Top. Catal. 2010, 53, 1002.
[16] a) J.-B. Sortais, N. Pannetier, A. Holuigue, L. Barloy, C.
Sirlin, M. Pfeffer, N. Kyritsakas, Organometallics 2007,
26, 1856; b) J.-B. Sortais, N. Pannetier, N. Clꢀment, L.
Barloy, C. Sirlin, M. Pfeffer, N. Kyritsakas, Organome-
tallics 2007, 26, 1868.
[17] L. Barloy, J.-T. Issenhuth, M. Weaver, N. Pannetier, C.
Sirlin, M. Pfeffer Organometallics 2011, 30, 1168.
[18] a) J. S. M. Samec, J.-E. Bꢃckvall, Chem. Eur. J. 2002, 8,
2955; b) J. S. M. Samec, A. H. Ell, J. B. Aaberg, T. Pri-
valov, L. Eriksson, J.-E. Bꢃckvall, J. Am. Chem. Soc.
2006, 128, 14293; c) D. Gnanamgari, A. Moores, E. Ra-
jaseelan, R. H. Crabtree, Organometallics 2007, 26,
1226.
The CNRS (fellowship granted to NP), University of Stras-
bourg, Ministꢁre de la Recherche et de la Technologie (fel-
lowship granted to JBS), ANR (PhotoBioMet, ANR-09-
BLAN-0191-01) fellowship to JTI and DSM are gratefully
acknowledged for partial support of the work.
References
[1] Ruthenium in organic synthesis, (Ed.: S. I. Murahashi),
Wiley-VCH, Weinheim, 2004.
[2] a) S. Gladiali, E. Alberico, Chem. Soc. Rev. 2006, 35,
226; b) A. J. Blacker, in: Handbook of Homogeneous
Hydrogenation , (Eds.: J. G. de Vries, C. J. Elsevier),
Wiley-VCH, Weinheim, 2007, Vol. 3, pp 1215–1244;
c) T. Ikariya, A. J. Blacker, Acc. Chem. Res. 2007, 40,
1300; d) C. Wang, X. Wu, J. Xiao, Chem. Asian J. 2008,
3, 1750.
[3] M. Wills, M. Gamble, M. Palmer, A. Smith, J. Studley,
J. Kenny, J. Mol. Cat. A: Chem. 1999, 146, 139.
[4] J. S. M. Samec, J.-E. Bꢃckvall, P. Andersson, P. Brandt,
Chem. Soc. Rev. 2006, 35, 237.
[5] S. Gladiali, E. Alberico, in: Transition metals for organ-
ic synthesis, 2nd edn., (Eds.: M. Beller, C. Bolm),
Wiley-VCH, Weinheim, 2004, Vol. 2, pp 145–166.
[19] G.-Z. Wang, J.-E. Bꢃckvall, J. Chem. Soc. Chem.
Commun. 1992, 980.
Adv. Synth. Catal. 2011, 353, 2844 – 2852
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2851

Nicolas Pannetier et al.
FULL PAPERS
[20] N. Uematsu, A. Fujii, S. Hashiguchi, T. Ikariya, R.
Noyori, J. Am. Chem. Soc. 1996, 118, 4916.
Ropic, M. Stefinovic, Org. Process Res. Dev. 2006, 10,
457.
[21] a) J. Mao, D. C. Baker, Org. Lett. 1999, 1, 841; b) J. Wu,
F. Wang, Y. Ma, X. Cui, L. Cun, J. Zhu, J. Deng, B. Yu,
Chem. Commun. 2006, 1766; c) L. Li, J. Wu, F. Wang, J.
Liao, H. Zhang, C. Lian, J. Zhu, J. Deng, Green Chem.
2007, 9, 23.
[22] a) D. Gnanamgari, A. Moores, E. Rajaseelan, R. H.
Crabtree, Organometallics 2007, 26, 1226; b) S. Hoff-
mann, A. M. Seayad, B. List, Angew. Chem. 2005, 117,
7590; Angew. Chem. Int. Ed. 2005, 44, 7424; c) G. Ven-
katachalam, R. Ramesh, Inorg. Chem. Commun. 2006,
9, 703; d) S Kuhl, R. Schneider, Y. Fort, Organometal-
lics 2003, 22, 4184–4186; e) D. G. Blackmond, M.
[23] L. F. Tietze, N. Rackelmann, I. Mꢇller, Chem. Eur. J.
2004, 10, 2722.
[24] N. Pannetier, J.-B. Sortais, P. S. Dieng, L. Barloy, C.
Sirlin, M. Pfeffer, Organometallics 2008, 27, 5852.
[25] R. Noyori, S. Hashiguchi, Acc. Chem. Res. 1997, 30, 97.
[26] D. J. Morris, A. M. Hayes, M. Wills, J. Org. Chem.
2006, 71, 7035.
[27] S. Arita, T. Koike, Y. Kayaki, T. Ikariya, Organometal-
lics 2008, 27, 2795.
[28] C. P. Che, K. Prasad, O. Repic Tetrahedron Lett. 1991,
32, 7175.
2852
asc.wiley-vch.de
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2011, 353, 2844 – 2852