Efficient cluster-based catalysts for asymmetric hydrogenation of α-unsaturated carboxylic acids

Article abstract of DOI:10.1002/chem.201200630

The new clusters [H₄Ru₄(CO)₁₀(μ-1,2-P- P)], [H₄Ru₄(CO)₁₀(1,1-P-P)] and [H ₄Ru₄(CO)₁₁(P-P)] (P-P=chiral diphosphine of the ferrocene-based Josiphos or Walphos ligand families) have been synthesised and characterised. The crystal and molecular structures of eleven clusters reveal that the coordination modes of the diphosphine in the [H₄Ru ₄(CO)₁₀(μ-1,2-P-P)] clusters are different for the Josiphos and the Walphos ligands. The Josiphos ligands bridge a metal-metal bond of the ruthenium tetrahedron in the "conventional" manner, that is, with both phosphine moieties coordinated in equatorial positions relative to a triangular face of the tetrahedron, whereas the phosphine moieties of the Walphos ligands coordinate in one axial and one equatorial position. The differences in the ligand size and the coordination mode between the two types of ligands appear to be reflected in a relative propensity for isomerisation; in solution, the [H₄Ru₄(CO)₁₀(1,1-Walphos)] clusters isomerise to the corresponding [H₄Ru₄(CO) ₁₀(μ-1,2-Walphos)] clusters, whereas the Josiphos-containing clusters show no tendency to isomerisation in solution. The clusters have been tested as catalysts for asymmetric hydrogenation of four prochiral α-unsaturated carboxylic acids and the prochiral methyl ester (E)-methyl 2-methylbut-2-enoate. High conversion rates (>94 %) and selectivities of product formation were observed for almost all catalysts/catalyst precursors. The observed enantioselectivities were low or nonexistent for the Josiphos-containing clusters and catalyst (cluster) recovery was low, suggesting that cluster fragmentation takes place. On the other hand, excellent conversion rates (99-100 %), product selectivities (99-100 % in most cases) and good enantioselectivities, reaching 90 % enantiomeric excess (ee) in certain cases, were observed for the Walphos-containing clusters, and the clusters could be recovered in good yield after completed catalysis. Results from high-pressure NMR and IR studies, catalyst poisoning tests and comparison of catalytic properties of two [H₄Ru₄(CO)₁₀(μ-1,2-P-P)] clusters (P-P=Walphos ligands) with the analogous mononuclear catalysts [Ru(P-P)(carboxylato)₂] suggest that these clusters may be the active catalytic species, or direct precursors of an active catalytic cluster species. Copyright

Full text of DOI:10.1002/chem.201200630

FULL PAPER
DOI: 10.1002/chem.201200630
Efficient Cluster-Based Catalysts for Asymmetric Hydrogenation of
a-Unsaturated Carboxylic Acids
Viktor Moberg,[a] Robin Duquesne,[a] Simone Contaldi,[b] Oliver Rçhrs,[a]
Jonny Nachtigall,[a] Llewellyn Damoense,[c] Alan T. Hutton,[e] Michael Green,[c, d]
Magda Monari,[b] Daniela Santelia,[f] Matti Haukka,[g] and Ebbe Nordlander*[a]
In memory of John R. Moss
Abstract:
[H4Ru4(CO)10(m-1,2-P-P)],
(CO)10ACHTUNGTRENNUNG(1,1-P-P)] and [H4Ru4(CO)11ACHTUNGTERN(NUGN P-
P)] (P-P=chiral diphosphine of the fer-
rocene-based Josiphos or Walphos
ligand families) have been synthesised
and characterised. The crystal and mo-
lecular structures of eleven clusters
reveal that the coordination modes of
the diphosphine in the [H4Ru4-
The
new
clusters
[H4Ru4-
between the two types of ligands
appear to be reflected in a relative pro-
pensity for isomerisation; in solution,
the [H4Ru4(CO)10(1,1-Walphos)] clus-
ters isomerise to the corresponding
[H4Ru4(CO)10(m-1,2-Walphos)] clusters,
whereas the Josiphos-containing clus-
ters show no tendency to isomerisation
in solution. The clusters have been
tested as catalysts for asymmetric hy-
drogenation of four prochiral a-unsatu-
rated carboxylic acids and the prochiral
methyl ester (E)-methyl 2-methylbut-2-
enoate. High conversion rates (>94%)
and selectivities of product formation
were observed for almost all catalysts/
catalyst precursors. The observed enan-
tioselectivities were low or nonexistent
for the Josiphos-containing clusters and
catalyst (cluster) recovery was low, sug-
gesting that cluster fragmentation takes
place. On the other hand, excellent
conversion rates (99–100%), product
selectivities (99–100% in most cases)
and good enantioselectivities, reaching
90% enantiomeric excess (ee) in cer-
tain cases, were observed for the Wal-
phos-containing clusters, and the clus-
ters could be recovered in good yield
after completed catalysis. Results from
high-pressure NMR and IR studies,
catalyst poisoning tests and comparison
of catalytic properties of two
[H4Ru4(CO)10(m-1,2-P-P)] clusters (P-
P=Walphos ligands) with the analo-
G
ACHTUNGTRENNUNG(CO)10(m-1,2-P-P)] clusters are differ-
ent for the Josiphos and the Walphos
ligands. The Josiphos ligands bridge
a metal–metal bond of the ruthenium
tetrahedron in the “conventional”
manner, that is, with both phosphine
moieties coordinated in equatorial po-
sitions relative to a triangular face of
the tetrahedron, whereas the phos-
phine moieties of the Walphos ligands
coordinate in one axial and one equa-
torial position. The differences in the
ligand size and the coordination mode
gous mononuclear catalysts [RuACHTUNGTRENNUNG(P-
P)(carboxylato)2] suggest that these
clusters may be the active catalytic spe-
cies, or direct precursors of an active
catalytic cluster species.
Keywords: asymmetric catalysis
·
cluster compounds · hydrogenation ·
ruthenium
[a] Dr. V. Moberg, R. Duquesne, O. Rçhrs, J. Nachtigall,
Dr. E. Nordlander
[d] Dr. M. Green
Present address: School of Chemistry
Newcastle University
Newcastle upon Tyne, NE1 7RU (UK)
Inorganic Chemistry Research Group, Chemical Physics
Center for Chemistry and Chemical Engineering
Lund University
Box 124, 221 00 Lund (Sweden)
Fax : (+46)46-222-0310
[e] Dr. A. T. Hutton
Department of Chemistry
University of Cape Town
Private Bag, Rondebosch 7701 (South Africa)
E-mail: Ebbe.Nordlander@chemphys.lu.se
[b] Dr. S. Contaldi, Prof. M. Monari
Dipartimento di Chimica “G. Ciamician”
Universitꢀ degli Studi di Bologna
Via Selmi 2, 40126 Bologna (Italy)
[f] D. Santelia
Dipartimento di Chimica I.F.M.
Universitꢀ di Torino
Via P. Giuria 7, 10125 Torino (Italy)
[c] L. Damoense, Dr. M. Green
Sasol Technology
[g] Prof. M. Haukka
Department of Chemistry
University of Eastern Finland
Box 111, 80101 Joensuu (Finland)
1 Klasie Havenga Road, Sasolburg
1947 (South Africa)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201200630.
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

Introduction
scale.[1,17,18] Here, we demonstrate the importance of the
chiral ligand for enantioselective cluster-based catalysis. The
use of tetraruthenium cluster catalysts of the type
During the past 40 years there has been a tremendous de-
velopment in homogeneous-catalysed enantioselective hy-
drogenation and a large number of new chiral ligands and
catalysts has been developed for this type of reaction.[1,2]
Asymmetric hydrogenation reactions are predominantly cat-
alysed by either ruthenium- or rhodium-based catalysts.[3]
Enantioselectivities have been found to be dependent on
the substrate used[3] and the most frequently reported results
for asymmetric hydrogenation where high enantioselectivi-
ties are obtained involve substrates such as unsaturated car-
boxylic acids with a-acylamido groups.[1,3,4]
[H4Ru4(CO)10ACHTUNTGRNEUNG(P-P)] containing ferrocene-based ligands
from the chiral Josiphos- and Walphos-diphosphine fami-
lies[19] (Tables 1 and 2) has led to unparalleled catalytic ac-
tivities for cluster-based catalysis, exhibiting both high con-
version rates and enantiomeric excesses exceeding 90% in
the hydrogenation of prochiral a-unsaturated carboxylic
acids. Parts of these results have been published in an earlier
communication.[20]
Table 1. Structures of the Josiphos ligands (1a–1d) used in this investiga-
tion.
There has been a considerable interest in the use of tran-
sition-metal carbonyl clusters as catalysts for various chemi-
cal reactions of industrial relevance[5,6] but very few studies
have been carried out on the potential of clusters to act as
catalysts for asymmetric synthesis. In similarity to pure het-
erogeneous catalysts, transition-metal carbonyl clusters are
often thermally stable and the presence of more than one
metal in a catalyst offers, in principle, the opportunity of
new pathways for the activation of molecules and the regio-
selectivity in reactions due to differentiated metal binding of
a substrate molecule (reactant).[7] It has indeed been found
that metal carbonyl clusters can function as efficient cata-
lysts (or catalyst precursors) for several diverse chemical re-
actions,[5] but cluster-based catalysis has not yet proven to
be commercially viable.
Ligand
R1
R2
Cy[a]
tBu
Cy
1a
1b
1c
1d
Ph
Ph
Cy
Ph
3,5-(CH3)2C6H3
[a] Cy=cyclohexyl.
Table 2. Structures of the Walphos ligands (2a–2g) used in this report.
It has been demonstrated that hydride-containing trinu-
clear and tetranuclear ruthenium carbonyl cluster anions are
effective catalysts/catalyst precursors for hydroformylation,[8]
hydrosilylation[9] and reductive coupling[10] reactions as well
as the water-gas shift reaction.[11] In pioneering systematic
studies, Matteoli and co-workers have shown that clusters of
Ligand
R1
R2
2a
2b
2c
2d
2e
2 f
2g
Ph
Ph
Ph
3,5-(CF3)2C6H3
Ph
Cy
3,5-(CF3)2C6H3
3,5-(CH3)2C6H3
3,5-(CF3)2C6H3
3,5-(CH3)2C6H3
the general formula [H4Ru4(CO)
12À2xACTHUNGTRENUN(NG P-P)x] (x=1,2; P-P=
3,5-(CH3)2-4-(CH3O)C6H2
Ph
Cy
chiral diphosphine) are good catalysts for asymmetric iso-
merisation,[12] hydroformylation[13] and hydrogenation reac-
tions.[14]
3,5-(CH3)2C6H3
The best enantiomeric excesses (ees) hitherto reported for
cluster-based asymmetric catalysis are approximately 45%
in the case of hydrogenation of tiglic acid ((E)-2-methyl-2-
butenoic acid) by the diphosphine-derivatised cluster
Results and Discussion
[H4Ru4(CO)10ACHTUNGTRENNUNG(1,1-bdpp)] (bdpp=(2R/S,4R/S)-2,4-bis(diphe-
nylphosphino)pentane).[15] In contrast to previous attempts
in cluster-based (asymmetric) hydrogenation[14] the catalysis
was found to take place under relatively mild reaction con-
ditions. It was observed that a strong chiral induction was in-
directly shown by the fact that the chirality of the predomi-
nating enantiomer of the product (2-methylbutyric acid) was
dependent on the chirality of the bdpp ligand.[15]
The use of ferrocene-based chiral ligands in homogeneous
asymmetric synthesis has proven to be highly efficient in
conjunction with single-site transition-metal catalysts based
on cobalt, rhodium, iridium and ruthenium.[16] Several mem-
bers of the Solvias ligand families have been successfully
employed with RhI, IrI and PdII complexes to catalyse enan-
tioselective reactions on both laboratory and industrial
The syntheses of the chiral diphosphine derivatives of
[H4Ru4(CO)12] were based on previously published meth-
ods,[15,21] that is, either 1) oxidative decarbonylation (by
using Me3NO) in the presence of the diphosphine in a ben-
zene/methanol solution at ambient temperature or heating
to reflux, or 2) thermal substitution in benzene under elevat-
ed pressures of hydrogen gas (25–30 bar). The latter method
was found to be superior with respect to yield and selectivity
1
in product formation. The clusters were identified by IR, H
and 31P NMR spectroscopy, mass spectrometry and, wher-
ever possible, X-ray crystallography.
Synthesis and characterisation of tetraruthenium clusters
containing the Josiphos ligands 1a–1d: The preparation of
&
2
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Cluster-Based Catalysts for Asymmetric Hydrogenation
FULL PAPER
clusters containing ligands 1a–1d was carried out by oxida-
tive decarbonylation of [H4Ru4(CO)12] by using Me3NO as
the decarbonylation reagent (see the Experimental Section
and the Supporting Information). For each diphosphine, the
[H4Ru4(CO)12] and for the bis-phosphine-derivatised clusters
[H4Ru4(CO)10ACHTUNTRGNNEUG 10ACHUTNGTRENNGUN
(NMDPP)2][21] and [H4Ru4(CO) (PPh3)2].[25]
The Cs symmetry found in the H4Ru4 core seems to be
a common feature shared by [H4Ru4(CO)10(diphosphine)]
clusters and has been reported for both chelating[15] and
bridging[15,21,24] coordination modes of the diphosphine li-
gands. All four hydrides coordinate at least to one of the
ruthenium atoms that are coordinated to the phosphorus
two isomers of [H4Ru4(CO)
phine either chelates one ruthenium atom ([H4Ru4(CO)10-
(1,1-P-P)]) or bridges one Ru–Ru edge ([H4Ru4(CO)10(m-
1,2-P-P)]) were isolated (Table 3), with the latter isomer
10ACHTUNGRTEN(NUNG P-P)] in which the diphos-
ACHTUNGTRENNUNG
À
moieties of the ligand. The Ru Ru bond lengths
are divided into two distinctive classes—the four
Table 3. A summary of cluster compounds and phosphine coordination modes found
in this study. See Tables 1 and 2 for structures of the ligands. Note the difference in
coordination for bridging Josiphos versus Walphos ligands (see text).
À
hydrido-bridged Ru Ru bonds are “long” and the
À
two non-bridged Ru Ru bonds are “short”. All CO
ligands are bound in a terminal monodentate coor-
dination mode and are staggered with respect to
À
the Ru Ru vectors, in contrast to the eclipsed con-
formation of the carbonyls in [H4Re4(CO)12], where
each of the four hydrides bridges a triangular metal
face.[25]
[H4Ru4(CO)112a] (13)
G
G
G
Molecular structure of [H4Ru4(CO)10(m-1,2-1a)] (4):
The molecular structure of cluster 4 is shown in
Figure 1 and relevant bond lengths and angles are
À
presented in Table 4. Ligand 1a bridges one Ru
Ru edge giving rise to a seven-membered dimetalla-
cycle and the H4Ru4 core conforms to a Cs symme-
try as found in other [H4Ru4(CO)10ACHTNUTRGNEUNG(P-P)] clusters.
AHCTUNGTRENNUNG
À
There are two short (Ru Ruaverage =2.784 ꢂ) and
À
four long (Ru Ruaverage =2.973 ꢂ) intermetallic dis-
being the predominant product. The isomers were identified
on the basis of comparison of their IR spectra to those of
tances, where the latter are related to the hydride-bridged
À
Ru Ru vectors (see above). This trend is typical for diphos-
phine-substituted [H4Ru4(CO)12] derivatives, and similar
the corresponding isomers of related [H4Ru4(CO)
10A
clusters[15,21–23] and in the cases of clusters 4 and 7–9, their
structures (in the solid state) were corroborated by X-ray
crystallography. Attempts to affect the product distribution
between the chelating and the bridging isomers by varying
the reaction temperature (0–708C) did not have any notable
effects. However, in the case of ligand 1c it was found that
an autoclave reaction of [H4Ru4(CO)12] with the diphos-
phine at 1258C and 30 bar of H2 gave exclusively the bridg-
ing isomer 8. It has previously been shown that
[H4Ru4(CO)10(m-1,2-P-P)] isomers—for example, P-P=
dppe,[24] bdpp ((2R/S,4R/S)-2,4-bis(diphenylphosphino)pen-
tane),[15] BINAP (2,2’-bis(diphenylphosphino)-1,1’-binaphth-
yl),[22] or DIPAMP (ethane-1,2-diylbis[(2-methoxyphenyl)-
phenylphosphane])[21]—can rearrange to form the corre-
elongation of the Ru Ru bonds bridged by the hydride li-
À
gands was observed in other [H4Ru4(CO)10(diphosphine)]
clusters,
such
as
[H4Ru4(CO)10(m-1,2-bdpp)],[15]
[H4Ru4(CO)10(m-1,2-DIOP)][21] (DIOP=(4R,5R)-2,2-dimeth-
yl-4,5-bis[(diphenylphosphino)methylene]-1,3-dioxolane))
and [H4Ru4(CO)10(m-1,2-dppe)].[24]
sponding [H4Ru4(CO)10ACHTUNRGTNE(NUG 1,1-P-P)] isomers. In contrast to the
corresponding clusters containing the Walphos ligands 2a–
2g (see below), none of clusters 3–10 containing the Josi-
phos ligands 1a–1d showed any propensity towards isomeri-
sation.
Crystal and molecular structures of clusters 4 and 7–9
General description of the molecular structures: The clusters
contain pseudo-tetrahedral H4Ru4 cores. The arrangement
of the four hydrides yields an H4Ru4 cluster core with Cs
symmetry, rather than D2d as observed for the parent cluster
Figure 1. ORTEP drawing of the molecular structure of [H4Ru4(CO)10(m-
1,2-1a)] (4). Thermal ellipsoids are drawn at the 30% probability level.
For the sake of clarity, all hydrogen atoms except the hydrides have been
omitted.
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
3
&
ÞÞ
These are not the final page numbers!

E. Nordlander et al.
Table 4. Selected bond lengths [ꢂ] and angles [8] for clusters 4 and 7–9.
4
7
8
9
À
Ru1 Ru2
2.9656(3)[a]
2.7750(3)
2.9544(3)[a]
2.9812(3)[a]
2.9900(3)[a]
2.7932(3)
2.3418(7)
2.3760(7)
2.9776(3)[a] 2.9768(3)[a]
3.0238(3)[a] 2.7976(3)
3.0940(2)[a] 2.9620(4)[a]
2.9961(3)[a]
2.9580(3)[a]
3.0320(3)[a]
2.7922(3)
À
Ru1 Ru3
À
Ru1 Ru4
Ru2 Ru3
2.7785(3)
2.7823(3)
2.9860(3)[a]
2.9686(3)[a]
À
Ru2 Ru4
2.9288(3)[a]
2.7890(3)
À
À
Ru3 Ru4
2.9289(2)[a] 2.7815(3)
À
Ru1 P1
2.4020(6)
2.3550(6)
3.492
2.3655(7)
2.3594(7)
2.3261(7)
À
Ru2 P2
À
Ru1 P2
2.3148(6)
3.199
P1-Ru1-Ru2
109.52(2)
4.644
113.824(19)
4.617
P···P[b]
P2-Ru2-Ru1
P1-Ru1-Ru2-P2 3.05(3)
112.104(18)
106.706(19)
6.02(3)
Figure 2. ORTEP drawing of the molecular structure of [H4Ru4(CO)10-
P2-Ru1-P1
92.233(18)
87.14(2)
AHCTUNGERTG(NNUN 1,1-1c)] (7). Thermal ellipsoids are drawn at the 30% probability level.
For the sake of clarity, all hydrogen atoms except the hydrides have been
omitted.
À
[a] Ru Ru bond bridged by hydride. [b] Measured with the Mercury soft-
ware (http://www.ccdc.cam.ac.uk/products/mercury/).
The two phosphorus atoms are coordinated in the same
relative positions on a triangular face of the [H4Ru4(CO)12]
framework, as seen for other diphosphine derivatives of
H4Ru4 clusters.[15,21,24] The Ru P bond lengths in 4 are also
À
similar to those observed for H4Ru4-phosphine deriva-
tives,[15,21] but are notably asymmetric, 2.3418(7) (Ru1 P1)
À
À
and 2.3760(7) ꢂ (Ru2 P2). Such a variation is not unprece-
dented in this type of cluster, but it may be related to the
relative p-acceptor ability of each phosphine moiety, which
is affected by the phosphorus substituents. Thus, the aryl
substituents on P1 are expected to make it a slightly better
p acceptor than P2. The P···P separation in 4 (4.644 ꢂ) is in
the range observed for the seven-membered “dimetallacy-
cles” in [H4Ru4(CO)10-m-1,2-DIOP][21] (4.870 ꢂ) and
[H4Ru4(CO)10-m-1,2-bdpp][15] (4.536 ꢂ).
Molecular structure of [H4Ru4(CO)10ACTHNUTRGNEUNG(1,1-1c)] (7): The mo-
Figure 3. ORTEP drawing of the molecular structure of [H4Ru4(CO)10(m-
1,2-1c)] (8). Thermal ellipsoids are drawn at the 30% probability level.
For the sake of clarity, all hydrogen atoms except the hydrides have been
omitted.
lecular structure of cluster 7 is shown in Figure 2 and bond
lengths and angles are summarised in Table 4. The coordi-
nated ligand 1c is chelating the ruthenium atom that is con-
nected to three hydrides and forms a six-membered ring.
The H4Ru4 core conforms to Cs symmetry with four long
À
À
À
À
(Ru Ruaverage =3.006 ꢂ) and two short (Ru Ruaverage
=
(Ru Ruaverage =2.789 and Ru Ruaverage =2.973 ꢂ, respective-
À
À
À
2.780 ꢂ) metal–metal distances. The two Ru P bond lengths
ly). The two Ru P distances differ only very slightly (Ru1
À
À
À
(Ru1 P1=2.402, Ru1 P2=2.355 ꢂ) are asymmetric and
somewhat longer than the corresponding distances in cluster
4. The bite angle (P2-Ru1-P1) generated by the ligand is
92.238, which is similar to that observed for [H4Ru4(CO)10-
P1=2.3655(7) and Ru2 P2=2.3594(7) ꢂ), with the shortest
À
Ru P distance being that of the non-chiral arm of the
ligand, as previously noted for clusters 4 and 7.
(1,1-bdpp)] (ꢀ91.858).[15]
Molecular structure of [H4Ru4(CO)10ACTHNGUTERNNU(G 1,1-1d)] (9): The mo-
lecular structure of cluster 9 is shown in Figure 4 (see
Table 4 for relevant bond lengths and angles). The ligand
1d chelates the ruthenium atom that is coordinated by three
bridging hydrides and forms a six-membered ring. The
Ru4H4 cluster core adapts to Cs symmetry with four long
Molecular structure of [H4Ru4(CO)10(m-1,2-1c)] (8): For
cluster 8, the molecular structure is shown in Figure 3 (bond
lengths and angles are reported in Table 4). The H4Ru4 core
conforms to Cs symmetry. The bridging coordinated diphos-
phine ligand 1c generates a seven-membered ring, similar to
that in cluster 4, with the same relative positions of the two
different phosphine moieties. The lengths of the Ru vectors
follow the same pattern as reported for other [H4R4(CO)10-
À
À
(Ru Ruaverage =2.978 ꢂ) and two short (Ru Ruaverage
=
2.791 ꢂ) metal–metal distances. The bite angle P2-Ru1-P1
(87.148) is notably smaller than that observed for the cluster
7. The two Ru P distances, Ru1 P1 and Ru1 P2, are ap-
proximately equal, 2.326 and 2.315 ꢂ (within the estimated
À
À
À
(P-P)] clusters, that is, two “short” and four “long” distances
&
4
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Cluster-Based Catalysts for Asymmetric Hydrogenation
FULL PAPER
The empirical formulae of clusters 11–21 were confirmed
by mass spectrometry, and their structures in solution and
the solid state were identified by IR and NMR spectroscopy,
as well as X-ray crystallography in certain cases. The IR
spectrum of 12 suggested that the overall structure of the
cluster was not identical to that already established for the
analogous [H4Ru4(CO)10(m-1,2-Josiphos)] clusters 4, 6, 8 and
10 (see above), but it resembled the IR spectrum of clusters
3, 5, 7 and 9, suggesting that the structure of 12 is analogous
to these four clusters with respect to the relative orientation
of the phosphine moieties on the tetraruthenium cluster
framework. The crystal structure of 12 (see below) con-
firmed that this was the case. Comparison of the IR spectra
of 11 and 12 to those of the analogous [H4Ru4(CO)11-
ACHUTNGRENUNG 10ACHTUNGTRENNGUN
(PR3)][21] and [H4Ru4(CO) (1,1-P-P)][15,22] clusters, respec-
tively, indicated their structures and in the case of 13 the
(solid-state) structure could be corroborated by a crystal
structure determination.
Figure 4. ORTEP drawing of the molecular structure of [H4Ru4(CO)10-
ACHTUNGTRENNUNG(1,1-1d)] (9). Thermal ellipsoids are drawn at the 30% probability level.
For the sake of clarity, all hydrogen atoms except the hydrides have been
omitted.
Although the hydrides in the parent cluster
[H4Ru4(CO)12] and most of its (phosphine) derivatives usu-
ally exhibit complete fluxionality at ambient temperature,
four relatively sharp resonances could be detected for the
standard deviation), respectively—significantly shorter than
those found for cluster 7 and [H4Ru4(CO)10ACHTNUGRTENUNG{1,1-(S,S or R,R)-
1
bdpp}][15] (ꢀ2.34 ꢂ). As noted above for clusters 4, 7 and 8,
hydrides in the H NMR spectra of 11 and 12. Variable-tem-
À
the slightly shorter Ru P bond length is located at the non-
perature (VT) 1H NMR spectra recorded for cluster 12
show that the four hydrides are static at 273 K (Figure 5)
and that complete fluxionality is not achieved even at 383 K
(Figure 6). The hydride resonances in cluster 12 could be as-
signed on the basis of its crystal structure, COSY and EXSY
(exchange spectroscopy) spectra (Figures 7 and 8),
chiral arm of the ligand.
Synthesis and characterisation of tetraruthenium clusters
containing the Walphos ligands 2a–2g: The ligand frame-
work of the Walphos ligands 2a–2g differs from that of the
Josiphos ligands 1a–1d by the insertion of a benzyl moiety
into the nonchiral phosphine “arm” of the ligand (see
Table 2). There is thus a six-bond P···P separation in ligands
2a–2g as opposed to a four-bond separation in 1a-1d; the
inherent “bite” of the Walphos ligands is therefore consider-
ably wider than that of the Josiphos ligands and—as will be
shown below—this has repercussions on both the structures
and the dynamics of the Walphos-containing tetraruthenium
clusters 10–21.
1
1
a
31P NMR spectrum, H,31P coupling constants and a H,31P
HetCor spectrum (Figure 9). As may be expected, it was
found that the hydride Hc located in between the two phos-
phines, as demonstrated by its correlation peaks with both
The two synthetic methods used to prepare clusters con-
taining the ligands 1a–1d (see above) were also employed
for the synthesis of clusters with ligands 2a–2g. Although
the yields obtained for each isolated cluster differ, the ob-
served general trend is that the thermal substitution reaction
method is superior to oxidative decarbonylation in terms of
yield. For ligand 2a, the products [H4Ru4(CO)
(11, 7%), [H4Ru4(CO)10(m-1,2-2a)] (12, 43%) and
[H4Ru4(CO)11A(2a)] (13, 44%) were isolated when the oxida-
10ACHTUNGTRNE(NUNG 1,1-2a)]
CHTUNGTRENNUNG
tive decarbonylation method was used. By varying the reac-
tion time and the amount of Me3NO added to the reaction
mixture, relatively high yields (44%) of the cluster “inter-
mediate” 13, in which the diphosphine coordinates as
a monoACHTUNGTRENNUNGdentate ligand through the nonchiral arm, were ob-
tained. On the other hand, cluster 11 could only be isolated
in very small quantities, but the yield was slightly improved
when an excess of the ligand was added to the reaction mix-
ture. Thermal ligand substitution led to the formation of 12
in good yield (73%), as the only isolated product.
Figure 5. VT NMR spectra of cluster [H4Ru4(CO)10(m-1,2-2a)] (12) re-
corded in the temperature range 298–248 K (top to bottom).
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
5
&
ÞÞ
These are not the final page numbers!

E. Nordlander et al.
ligand 2c and the parent cluster [H4Ru4(CO)12] by oxidative
decarbonylation yields two main products, [H4Ru4(CO)10-
AHCTUNGTERNN(GNU 1,1-2c)] (16) (16%) and [H4Ru4(CO)10(m-1,2-2c)] (17)
(56%). The thermal substitution of [H4Ru4(CO)12] with
ligand 2c led exclusively to cluster [H4Ru4(CO)10(m-1,2-2c)]
(17) in 66% yield. The yield for 16 (16%) in the oxidative
Figure 6. VT NMR spectra of cluster [H4Ru4(CO)10(m-1,2-2a)] (12) re-
corded in the temperature range 298–383 K (top to bottom).
phosphorus nuclei in the HetCor spectrum, is the least
prone to exchange.
The diphosphine ligand 2b is structurally very similar to
ligand 2a. It could be coordinated to [H4Ru4(CO)12] most
efficiently
through
thermal
substitution,
yielding
[H4Ru4(CO)10(m-1,2-2b)] (15) as the sole product in 46%
yield. Oxidative decarbonylation gave the two isomers
[H4Ru4(CO)10ACHTUNGTRENNUNG(1,1-2b)] (14) and cluster 15, but in significant-
ly lower yields—14 (2%) and 15 (22%)—compared with
the similar reaction using ligand 2a. The reaction with
Figure 8. Top: EXSY spectrum of cluster [H4Ru4(CO)10(m-1,2-2a)] (12)
recorded at 298 K with a mixing time of 0.5 s. Bottom: EXSY spectrum
of cluster [H4Ru4(CO)10(m-1,2-2a)] (12) recorded at 313 K with a mixing
time of 0.5 s.
Figure 7. COSY spectrum of cluster [H4Ru4(CO)10(m-1,2-2a)] (12) record-
ed at 263 K in CD2Cl2. Note that the peaks at d=À16.4 and À16.6 ppm
do not have any cross couplings.
&
6
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Cluster-Based Catalysts for Asymmetric Hydrogenation
FULL PAPER
a terminal monodentate coordination mode. As observed
À
for clusters 4 and 7–9, the Ru Ru bond lengths may be sep-
arated into two “short” and four “long” metal–metal bonds.
In the molecular structures of clusters 12, 15, 17 and 19–21,
the ligands were found to exclusively coordinate in a bridg-
ing fashion, except for cluster 13 where the diphosphine ex-
hibits
a
monodentate
coordination
mode.
The
[H4Ru4(CO)10(1,2-Walphos)] clusters 12, 15, 17 and 19–21
exhibit a coordination mode for the diphosphine that is dif-
ferent from their Josiphos counterparts 4 and 8.
The “bites” of ligands 2a–2g are so large that it is not
possible for the two phosphorus atoms of the ligand to
reside in the same relative positions on the same triangular
metal face to which both phosphine moieties are coordinat-
ed. In the nine-membered “dimetallacycle” that is formed,
one phosphorus atom coordinates in an equatorial position,
whereas the second phosphorus atom resides in an axial po-
sition of the specific triangular face. This coordination mode
of a diphosphine ligand on the tetrahedral H4Ru4 core has
not been observed previously, with exception for the dimeric
Figure 9. 1H,31P HetCor spectrum of cluster [H4Ru4(CO)10(m-1,2-2a)] (12)
recorded at 223 K in [D3]chloroform solution. The spectrum reveals the
1
following H,31P correlation peaks: the hydrides Ha and Hd correlate with
the phosphorus atom at d=39.1 ppm, the hydride Hb couples with the
phosphorus atom at higher frequency and the hydride Hc correlates with
both phosphorus nuclei.
cluster [{H4Ru4(CO)10ACHTUNTGRENUNG(1,2-DIOP)}2], in which two diphos-
phine ligands form bridges between two tetraruthenium
units so that each H4Ru4 unit is coordinated by two phos-
phorus atoms originating from two different diphosphine li-
gands.[21]
decarbonylation reaction is notably higher than those of
clusters 11 and 14. Ligand substitution with 2d yields one
main product, [H4Ru4(CO)10(m-1,2-2d)] (18), by both the ox-
idative decarbonylation and the thermal substitution meth-
ods. Similarly, reactions with the diphosphine ligands 2e–2g
Molecular structure of [H4Ru4(CO)10(m-1,2-2a)] (12):[20] The
molecular structure of cluster 12 is shown in Figure 11 and
relevant bond lengths and angles are reported in Table 5.
The arrangement of the four hydrides yields a Cs symmetry
of the tetrahedral H4Ru4 cluster core, observed in numerous
examples of diphosphine-derivatised ruthenium clusters of
yield
the
clusters
[H4Ru4(CO)10(m-1,2-2e)]
(19),
[H4Ru4(CO)10(m-1,2-2 f)] (20) and [H4Ru4(CO)10(m-1,2-2g)]
(21). The structures of clusters 15–21 could be identified by
comparison of their IR spectra to those of the analogous
2a-substituted clusters and in the case of clusters 15, 17, 19,
20 and 21 the structures of the clusters were confirmed by
X-ray crystallography (see below). See Table 3 for schematic
structures for all new clusters isolated in this report.
the type [H4Ru4(CO)10(diphosphine)][15,21,24] The Ru Ru
À
bond lengths are divided into two distinctive classes—the
À
À
four hydrido-bridged Ru Ru bonds are “long” (Ru
Crystal and molecular structures of clusters 12, 13, 15, 17
and 19–21
General description of the molecular structures: The clusters
contain pseudo-tetrahedral H4Ru4 cores. All four hydrides
coordinate to at least one of the ruthenium atoms that are
coordinated to the phosphine moieties of the ligand so that
the hydride arrangement yields a Cs symmetry (Figure 10)
of the H4Ru4 cluster core. All carbonyl ligands are bound in
Figure 11. ORTEP
drawing
of
the
molecular
structure
of
[H4Ru4(CO)10(m-1,2-2a)] (12). Thermal ellipsoids are drawn at the 30%
probability level. For the sake of clarity, all hydrogen atoms except the
hydrides have been omitted.
Figure 10. Symmetry of the H4Ru4 cluster framework.
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
7
&
ÞÞ
These are not the final page numbers!

E. Nordlander et al.
Table 5. Selected bond lengths [ꢂ] and angles [8] for clusters 12, 13, 15, 17 and 19–21.
12 13 15 17 19
3.0032(15)[a] 2.9579(6)[a] 3.0417(4)[a] 3.0215(6)[a] 2.9943(5)[a]
“short” and four “long”, where
the four long bonds correspond
to those bridged by hydrides.
20
21
3.0277(7)[a] 2.9732(5)
2.9224(7)[a] 2.9821(5)
À
Ru1 Ru2
Ru1 Ru3
2.7774(14)
2.9960(5)[a] 2.7790(4)
2.7668(6)
2.9894(6)[a] Unlike the related structures
À
2.9205(13)[a] 2.9062(6)[a] 2.9111(3)[a] 2.9239(7)[a] 3.0259(6)[a]
2.9525(15)[a] 2.9883(6)[a] 2.9986(4)[a] 3.0178(7)[a] 2.7761(6)
[H4Ru4(CO)10(m-1,2-bdpp)][15]
À
Ru1 Ru4
2.7635(7)
3.0022(5)
3.0283(6)[a] 2.7832(5)
2.9997(9)[a] 2.7854(5)
À
Ru2 Ru3
and
[H4Ru4(CO)10(m-1,2-
3.0291(15)[a] 2.8936(6)
2.7900(15) 2.7722(6)
2.3249(10) 2.3765(12) 2.369(3)
3.0139(4)[a] 2.9938(7)[a] 2.9215(6)[a]
À
Ru2 Ru4
dppe)],[24] where the longest
À
Ru3 Ru4
2.7785(6)
2.9267(5)
2.7938(4)
2.7787(7)
2.3293(15) 2.3663(13)
2.3898(16) 2.3444(14)
2.7846(6)
À
Ru Ru bond is that bridged by
À
Ru1 P1
2.3641(15) 2.3470(9)
2.3825(14) 2.3570(9)
À
Ru2 P2
2.3598(10)
5.584
2.355(3)
5.557
the diphosphine, the longest
P···P[b]
5.355
5.544
5.544
5.617
À
Ru Ru distance in cluster 15 is
À
Ru2 Ru4 (3.0291(15) ꢂ). The
P1-Ru1-Ru2
P2-Ru2-Ru1
P1-Ru1-Ru2-P2 82.73(3)
116.28(3)
108.23(3)
162.61(3)
120.73(10)
107.15(9)
74.68(13)
121.20
119.28
7.89
N
117.33(2)
105.57(2)
81.86
114.80(4)
106.19(4)
85.48
109.17(3)
116.81(4)
83.03
À
Ru1 P1 bond (2.369(3) ꢂ) is
considerably longer compared
À
to the Ru1 P1 bond found in
cluster 12 (2.3249(10) ꢂ). As in
À
[a] Ru Ru bond bridged by hydride. [b] Measured with the Mercury software (http://www.ccdc.cam.ac.uk/
products/mercury/).
À
Ruaverage =2.99 ꢂ) and the two non-bridged Ru Ru bonds
À
À
are “short” (Ru Ruaverage =2.77 ꢂ). The average Ru Ru
bond lengths in cluster 12 are comparable with other
[H4Ru4(CO)10(m-1,2-diphosphine)] structures reported, such
as
those
of
[H4Ru4(CO)10(m-1,2-bdpp)][15]
and
[H4Ru4(CO)10(m-1,2-dppe)].[24] The Ru1 P1 bond length
À
À
(2.325 ꢂ) is considerably shorter than the Ru2 P2 bond
length (2.360 ꢂ), but not unusual for this type of cluster. Be-
cause of the different coordination mode discussed above,
the P1-Ru1-Ru2-P2 torsion angle (82.738) is exceptionally
large, compared to that observed for clusters 4 (3.058) and 8
(6.028). The P···P separation of 5.584 ꢂ in 12 is also excep-
tionally large when compared to the structures of other
[H4Ru4(CO)10(m-1,2-P-P)] clusters, for example, in
[H4Ru4(CO)10(m-1,2-DIOP)],[21] in which the diphosphine
forms an eight-membered “dimetallacycle”, a P···P separa-
tion of 4.87 ꢂ is found.
Figure 12. ORTEP drawing of the molecular structure of [H4Ru4(CO)11-
AHCTUNGERTG(NNUN 2a)] (13). Thermal ellipsoids are drawn at the 30% probability level.
For the sake of clarity, all hydrogen atoms except the hydrides have been
omitted.
Molecular structure of [H4Ru4(CO)11ACTHNUTRGNEUNG(2a)] (13): The molecu-
lar structure of cluster 13 is shown in Figure 12 and relevant
bond lengths and angles are reported in Table 5. To our
knowledge, this is the first reported molecular structure with
monodentate coordination of a diphosphine to the H4Ru4
À
core. The two short Ru Ru distances have an average value
À
of 2.784 ꢂ, whereas Ru Ruaverage for the four “long” distan-
À
ces is 2.971 ꢂ. The longest Ru Ru bond is found between
Ru1 and Ru4 (3.0022(5) ꢂ), which is notably shorter than
the longest bond in cluster 12, which is 3.0283(6) ꢂ for the
À
À
Ru2 Ru3 bond. The Ru1 P1 distance of 2.3765(12) ꢂ is
somewhat shorter than that found for the same phosphine
moiety in 12 but considerably longer than those found in the
related structures of [H4Ru4(CO)
[H4Ru4A(CO)11A{P
(Me2Ph)}] (2.345 ꢂ)[26] and [H4Ru
(C6F5)3}] (2.362 ꢂ).[27]
(PPh3)] (2.354 ꢂ),[26]
4A(CO)11A{P-
C
E
E
N
CHTUNGTRENNUNG
ACHTUNGTRENNUNG
(m-1,2-2b)] (15):[20] The
Molecular structure of [H4Ru4ACHTNUTRGENNUG(CO)10ACHTUNGTRENNGUN
molecular structure of 15 is shown in Figure 13 and relevant
bond lengths and angles are reported in Table 5. The bond
lengths in 15 are similar to those observed for related clus-
Figure 13. ORTEP
drawing
of
the
molecular
structure
of
[H4Ru4(CO)10(m-1,2-2b)] (15). Thermal ellipsoids are drawn at the 30%
probability level. For the sake of clarity, all hydrogen atoms except the
hydrides have been omitted.
À
ters, and the Ru Ru distances also follow the trend of two
&
8
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Cluster-Based Catalysts for Asymmetric Hydrogenation
FULL PAPER
12, the coordinated ligand in cluster 15 yields a nine-mem-
bered “dimetallacycle”, where one phosphorus atom is coor-
dinated in an equatorial position and the other phosphorus
atom occupies an axial position on the specific triangular
face, with a P1-Ru1-Ru2-P2 torsion angle of 74.688, which is
considerably smaller than that found in cluster 12. The P···P
separation found in 15 (5.557 ꢂ), is slightly shorter than the
corresponding distance in 12 (5.584 ꢂ).
Molecular structure of [H4Ru4(CO)10(m-1,2-2c)] (17): The
molecular structure of 17, which is shown in Figure 14 (rele-
vant bond lengths and angles are reported in Table 5), has
a 2c ligand coordinated in a bridging mode giving rise to
a nine-membered “dimetallacycle”. However, unlike clusters
12 and 15, the phosphine moieties in 17 are coordinated in
the same relative positions to the specific triangular face.
The torsion angle P1-Ru1-Ru2-P2 (7.898) is thus similar to
that found in cluster 8 but deviates markedly from those
found for clusters 12 and 15. It seems that the steric bulk of
the cyclohexyl substituents on P2 are not sufficiently large
(and rigid) to force the phosphorus atoms to occupy equato-
Figure 15. ORTEP
drawing
of
the
molecular
structure
of
[H4Ru4(CO)10(m-1,2-2e)] (19). Thermal ellipsoids are drawn at the 30%
probability level. For the sake of clarity, all hydrogen atoms except the
hydrides have been omitted.
À
rial–axial positions. The Ru P distances of 2.3641(15)
À
À
(Ru1 P1) and 2.3825(14) ꢂ (Ru2 P2) are slightly longer
than those found in 15 but similar to those found in cluster
7. The P···P distance is 5.355 ꢂ, shorter than those of both
12 (5.584 ꢂ) and 15 (5.557 ꢂ).
À
À
Ru2 P2 (2.355(3) ꢂ) found in 15. However, the Ru1 P1
bond length in 19 is shorter than that found in cluster 15,
2.3470(9) and 2.369(3) ꢂ, respectively.
Molecular structure of [H4Ru4(CO)10(m-1,2-2e)] (19): The
molecular structure of cluster 19 is shown in Figure 15 and
relevant bond lengths and angles are summarised in Table 5.
The structure of cluster 19 is very similar to those of 12 and
Molecular structure of [H4Ru4(CO)10(m-1,2-2 f)] (20): The
molecular structure of cluster 20 is shown in Figure 16 and
relevant bond lengths and angles are summarised in Table 5.
The structure is similar to those of clusters 12, 15, and 19.
À
À
15. The longest Ru Ru distance in 19 (3.0417(4) ꢂ) is locat-
As for 19, but unlike clusters 12, 15 and 17, the longest Ru
Ru bond in 20 (3.0215(6) ꢂ) is that bridged by the diphos-
ed between the ruthenium atoms bridged by the diphos-
phine ligand (Ru1 and Ru2). The torsion angle P1-Ru1-
Ru2-P2 measures 81.868. The P···P separation is 5.54 ꢂ. The
À
À
phine (Ru1 Ru2). The Ru1 P1 distance (2.3293(15) ꢂ) is
À
similar to that found in 12. The Ru2 P2 bond
À
Ru2 P2 distance (2.3570(9) ꢂ) is comparable with the
(2.3898(16) ꢂ) is found to be significantly longer than the
À
Ru1 P1 bond, and is similar to that found in cluster 17. The
two phosphorus atoms in the axial–equatorial position on
the specific triangular face are separated by 5.544 ꢂ (within
the estimated standard deviations), which is identical to that
found in cluster 19. The torsion angle P1-Ru1-Ru2-P2 meas-
ures 85.488, which is the largest such angle that has been de-
tected in this study.
Molecular structure of [H4Ru4(CO)10(m-1,2-2g)] (21): The
molecular structure of cluster 21 is shown in Figure 17 and
relevant bond lengths are shown in Table 5. The structure is
À
very similar to those of clusters 12 and 17. The longest Ru
À
Ru distance is the Ru1 Ru4 edge (3.0259(6) ꢂ), which is
bridged by one hydride. The two phosphorus atoms of the
coordinated ligand are separated by 5.62 ꢂ with a torsion
angle of 83.038. This P···P separation is slightly longer than
À
the analogous distance found in 12 (5.58 ꢂ). The Ru P
bonds are slightly asymmetric in lengths (2.3663(13) for
Figure 14. ORTEP
drawing
of
the
molecular
structure
of
À
À
Ru1 P1 and 2.3444(14) ꢂ for Ru2 P2), but do not deviate
from those observed for other clusters reported here.
[H4Ru4(CO)10(m-1,2-2c)] (17). Thermal ellipsoids are drawn at the 30%
probability level. For the sake of clarity, all hydrogen atoms except the
hydrides have been omitted.
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
9
&
ÞÞ
These are not the final page numbers!

E. Nordlander et al.
solution to form 12. Thus, a pure sample of cluster 13 was
dissolved in a dichloromethane/hexane mixture and was left
to recrystallise at room temperature (ꢀ218C). After ap-
proximately two days, red shiny crystals were formed, but
their IR spectrum (in dichloromethane) revealed that cluster
12 had been formed. However, the solids of clusters 11 and
13 could be stored at +48C for months without any ob-
served isomerisation.
The rate of the isomerisation of 11 to 12 was affected
both by temperature and the polarity of the solvent used.
When the ligand substitution reaction was carried out in
a mixture of dichloromethane/acetonitrile, cluster 12 was
the only isolated product. The isomerisation process of 11
into 12 was followed by 1H and 31P NMR spectroscopy
(298 K, CDCl3) (see Figure 18 and the Supporting Informa-
tion) during a period of 30 days. A sample of 11 (10 mg,
6.17 mmol) in CD3Cl (0.5 mL) was prepared in a 5 mm stan-
dard NMR tube. The sample was stored at room tempera-
ture (ꢀ218C) when not being analysed. The proton spectra
(hydride resonances only) (Figure 18) show the transforma-
tion of 11 into 12, which proceeded at a relatively slow rate.
The top spectrum in Figure 18 displays the hydrides of the
freshly made sample of 11, which exhibits relatively sharp
Figure 16. ORTEP
drawing
of
the
molecular
structure
of
[H4Ru4(CO)10(m-1,2-2 f)] (20). Thermal ellipsoids are drawn at the 30%
probability level. For the sake of clarity, all hydrogen atoms except the
hydrides have been omitted.
resonances at d=À15.67(s), À15.97(s), À16.51
ACHTUNGTRENN(UNG brm) and
À16.75 ppm (s). After two days (t=2), a new set of four
new hydride resonances that could be assigned to 12 was ob-
served at d=À16.4, À16.7, À17.6 and À18.5 ppm. In addi-
tion to the appearance of the four hydrides of cluster 12,
Figure 17. ORTEP drawing of molecular structure of [H4Ru4(CO)10(m-
1,2-2g)] (21). Thermal ellipsoids are drawn at the 30% probability level.
For the sake of clarity, all hydrogen atoms except the hydrides have been
omitted.
Interconversion of isomers: As mentioned above, the con-
version of [H4Ru4(CO)10(m-1,2-P-P)] clusters into the analo-
gous, thermodynamically favoured, [H4Ru4(CO) CHTUNGTNER(NUNG 1,1-P-P)]
10A
isomers has been established for
a number of sys-
tems.[15,21,23,24] For the clusters containing ligand 2a the op-
1
Figure 18. Arrayed H NMR spectra illustrating the isomerisation process
of [H4Ru4(CO)10ACTHNUTGRNEG(UN 1,1-2a)] (11) into [H4Ru4(CO)10(m-1,2-2a)] (12) record-
posite isomerisation was detected, that is, in solution
[H4Ru4(CO)
1,2-2a)] (12) without any significant cluster fragmentation.
Furthermore, [H4Ru4(CO)11A(2a)] (13) loses a CO ligand in
10ACHTUNGTRENNUNG(1,1-2a)] (11) isomerises into [H4Ru4(CO)10(m-
ed in CDCl3 at 298 K (t=number of days). The broad signal at d=
À17.15 ppm labelled * is from the four fluxional hydrides of cluster
G
[H4Ru4(CO)11ACTHUNGTERNNU(G 2a)] (13).
&
10
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Cluster-Based Catalysts for Asymmetric Hydrogenation
FULL PAPER
Table 6. Redox potentials (V vs. Ag/Ag+) of the ferrocene reference, the
a broad signal at d=À17.15 ppm became notably strong (t=
5), which may be assigned to the hydride resonances for
cluster 13 (see the Experimental Section). As the isomerisa-
tion proceeds, the resonances from cluster 11 are completely
gone after approximately 30 days. At the end of the experi-
ment, there was still a small amount of 13 present in the so-
lution. The complementary 31P NMR spectra confirm the
1H NMR analyses, that is, transformation of 11 into 12, with
concomitant formation and disappearance of 13, but do also
reveal the appearance of free diphosphine ligand and the
corresponding phosphine oxide during the prolonged reac-
tion time. It appears that decomposition of the cluster
occurs on prolonged standing in solution (30 days), but that
the main species formed is the thermodynamically stable
Walphos ligand 2a and [H4Ru4(CO)10(m-1,2-2a)] (12).[a]
Compound
Oxidation processes
Reduction processes
Epa
Epc
Epc
Epa
ferrocene
2a
+0.15
+0.15
+0.36[b]
+0.45
+0.66
+0.27[c]
+0.40
+0.03
–
+0.28[b]
+0.37[b]
+0.59[b]
+0.21[b,c]
–
12
À1.55
À1.92
–
–
–
À1.01
[a] Measured in CH3CN containing 0.1m [nBu4N]ACHTNUGTRNEUNG[ClO4] at a scan rate of
100 mVsÀ1 (unless otherwise indicated) and referenced to Ag/Ag+ (see
body of table and the Experimental Section for corresponding ferrocene/
ferrocenium values). Epa =anodic peak potential, Epc =cathodic peak po-
tential. [b] Estimated value: peak poorly defined. [c] Scan rate=
À
isomer 12 in which the diphosphine bridges a Ru Ru bond.
1000 mVsÀ1
.
Like cluster 11, [H4Ru4(CO)
10A
reluctance to retain the chelating coordination mode. Even
though this cluster (16) could be synthesised in considerably
higher yield than its analogues 11 and 14, it is also unstable
in solution. Once 16 is exposed to polar solvents (such as di-
chloromethane or chloroform), isomerisation occurs at
a high rate. TLC, IR and NMR analyses reveal that within
minutes a significant amount of cluster [H4Ru4(CO)10(m-1,2-
2c)] (17) has been formed. It is likely that the eight-mem-
bered metallacycle that is formed when a Walphos ligand
coordinates in a chelating mode is more strained than the
nine-membered “dimetallacycle” that is formed in the bridg-
ing mode, and that it is the relief of this strain that is the
driving force for the isomerisation reaction.
+0.28 V (before reaching the potentials of the other oxida-
tive processes mentioned below) resulted in the same irre-
versible profile, showing that the ferrocenyl cation generat-
ed must immediately react to form some non-reducible
product without the participation of the species generated at
more anodic potential. Here three further oxidation waves
were observed (one barely detected, see Figure 19; Ep
values in Table 6), each with an associated but poorly re-
solved reductive counterpart observed upon reversing the
scan; varying the scan rate in steps between 20 and
2000 mVsÀ1 did not enhance the definition of these peaks.
The processes involved at these more positive potentials are
likely to involve oxidation of the two differently substituted
phosphine groups in the ligand, but further interpretation is
not feasible.
Electrochemistry: The ferrocene-based diphosphine ligands
used in this study are not expected to display straightfor-
ward electrochemistry, as the oxidised species will have con-
siderable P-centred radical character that can give rise to
complex reactions on the cyclic voltammetric timescale, in-
cluding decomposition.[28] In order to establish whether the
redox properties of these ligands are affected by their coor-
dination to metals, and whether the metal clusters them-
selves could be reversibly reduced and/or oxidised, a prelimi-
nary electrochemical study of the ferrocenyl-based diphos-
phine ligand 2a and its tetranuclear hydridoruthenium car-
bonyl cluster complex 12 was undertaken by using cyclic
voltammetry at a platinum disk electrode. Experiments
were performed on acetonitrile solutions containing 0.1m
tetra-n-butylammonium perchlorate as background electro-
lyte, by using a standard three-electrode system. The results
are summarised in Table 6.
A somewhat more complicated voltammetric behaviour is
exhibited by the tetraruthenium complex 12, although the
For ligand 2a no reductive electrochemistry was detected
down to the solvent/electrolyte limit of approximately
À2.2 V. However, the anodic electrochemistry (see
Figure 19) displays the expected oxidation of the ligand fer-
rocenyl group at exactly the same potential as recorded for
the reference ferrocene under the same conditions (Epa =
+0.15 V vs. Ag/Ag+, see the Experimental Section), but
with no corresponding reduction peak, even when the scan
rate was increased in steps from the normally used 100 up
to 2000 mVsÀ1. In addition, reversing the anodic scan at
Figure 19. Cyclic voltammogram (oxidation processes) of ligand 2a in
CH3CN at a scan rate of 100 mVsÀ1. The potential scale is referenced to
Ag/Ag+ (see the Experimental Section). The scan started at À0.50 V.
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
11
&
ÞÞ
These are not the final page numbers!

E. Nordlander et al.
general features observed by Hartl, Johnson and co-work-
ers[29] in their detailed study of the a-diimine complexes of
the tetranuclear hydridoruthenium carbonyl clusters
[H4Ru4(CO)10(L)] (L=2,2’-bipyrimidine, 2,3-bis(pyridin-2-
yl)pyrazine or 2,2’-bipyridine) in tetrahydrofuran solution
are retained. Thus, in reductive scans we observe two catho-
dic peaks at rather negative potential (see Figure 20) that,
dence of oxidation of the phosphine groups, in contrast to
the observations made in the study of the free diphosphine
ligand 2a (see above); clearly coordination of the phospho-
rus atoms to the tetraruthenium cluster inhibits such oxida-
tion. However, on increasing the scan rate up to
1000 mVsÀ1, a shoulder on the main anodic peak (which
itself shifts slightly to +0.46 V) becomes resolved at
+0.27 V (see Figure 21).
Figure 20. Cyclic voltammogram (reduction processes) of [H4Ru4(CO)10-
(m-1,2-2a)] (12) in CH3CN at a scan rate of 100 mVsÀ1. The potential
scale is referenced to Ag/Ag+ (see the Experimental Section). The scan
started at À1.50 V and shows that the anodic peak at À1.01 V is only ob-
tained after the reductive scan.
Figure 21. Cyclic voltammogram (oxidation processes) of [H4Ru4(CO)10-
(m-1,2-2a)] (12) in CH3CN at a scan rate of 1000 mVsÀ1. The potential
scale is referenced to Ag/Ag+ (see the Experimental Section). The scan
started at À0.20 V.
by analogy with the corresponding a-diimine complexes,[29]
This can confidently be ascribed to the oxidation of the
ferrocenyl group in the diphosphine ligand, made slightly
more anodic than in the free ligand (by 0.12 V) upon coordi-
nation of the phosphorus atoms to the tetraruthenium clus-
ter. There is evidence for a corresponding cathodic peak at
+0.21 V on the return scan, but this is poorly defined and
not improved by reversing the direction of the potential
scan before the onset of the main oxidation peak.
can be ascribed to a one-electron process, which produces
CÀ
the radical anion 12 (Epc =À1.55 V, irreversible) and
a second one-electron reduction to provide the unstable di-
A
hydrogen to form the dihydrido dianion [H2Ru4(CO)10(1,2-
m-2a)]2À. It is the oxidation of this dihydrido dianion that is
responsible for the anodic peak observed on the return scan
at Epa =À1.01 V, a peak not observed unless the cathodic
peaks referred to above are first traversed (see Figure 20).
In contradistinction to the initial reduction of
[H4Ru4(CO)10(a-diimine)] (in tetrahydrofuran) to produce
the radical anion, which was reported[29] to be electrochemi-
cally and chemically reversible, we find the initial reduction
of the diphosphine complex 12 (in acetonitrile) to be irre-
versible. This was shown by reversing the direction of the re-
ductive scan before onset of the second reduction that forms
the unstable dianion: no oxidative wave was observed over
a wide range of scan rates. The oxidative electrochemistry of
complex 12 displays, at a scan rate of 100 mVsÀ1, a single ir-
reversible anodic peak centred at +0.40 V. Similar behav-
iour was observed in the oxidative behaviour of
[H4Ru4(CO)10(a-diimine)],[29] and this peak can be ascribed
to an anodic process localised on the metal core, which may
Catalytic activity—asymmetric hydrogenation of a-unsatu-
rated carboxylic acids: As mentioned in the introduction, it
is well established that chiral diphosphine ligands play an
important role in asymmetric hydrogenation.[3a,30] In order
to assess the capabilities of the new clusters prepared in this
investigation to function as catalysts for asymmetric hydro-
genation, hydrogenation of the carboxylic acids 24–29
(Scheme 1) in the presence of clusters 4, 6, 8, 10–12, 14, 15
and 17–21 was examined. In the present study, a hydrogen
pressure of 50 bar was used, instead of 130 bar commonly
used in previous cluster-based catalysis tests.[14b] The lower
hydrogen pressure was expected to enhance the chiral in-
duction, as demonstrated by us in a previous study.[15] The
results from the catalytic investigation of clusters 4, 6, 8, 10–
12, 14, 15 and 17–21 are summarised in Table 7. The clusters
that have been investigated in this report are superior to
À
indeed induce cleavage of Ru Ru bonds. There is no evi-
&
12
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Cluster-Based Catalysts for Asymmetric Hydrogenation
FULL PAPER
It has been demonstrated that the carboxylate functionali-
ty plays an important role in the interaction between the
substrate and the catalyst in hydrogenation reactions utilis-
ing mono- and polynuclear ruthenium catalysts.[14a,31] For ex-
ample, Benedetti et al.[14a] observed a significant difference
in optical purity when comparing the two substrates (E)-a-
methyl-cinnamic acid (25) and ethyl (E)-a-methyl-cinna-
mate when they used the cluster [H4Ru4(CO)8{(À)-DIOP}2]
as catalyst. The hydrogenated cinnamic ester obtained was
of considerably lower optical purity than that of the hydro-
genated cinnamic acid.
Scheme 1. Drawings of the substrates used in catalytic experiments. (E)-
2-methyl-2-butenoic acid (tiglic acid, 24), (E)-a-methyl-cinnamic acid
(25), (E)-a-phenyl-cinnamic acid (26), (E)-2-methyl-2-pentenoic acid
(27), (E)-methyl 2-methylbut-2-enoate (28) and (E)-a-acetamidocinnamic
acid (29).
In order to investigate whether the reactivity and the
enantioselectivity in the hydrogenation reactions effected by
the cluster catalysts/catalyst precursors in this study are de-
pendent on substrates having a carboxylate functionality,
a control experiment by using the substrate methyl 2-meth-
ylbut-2-enoate (28, Scheme 1), that is, the methyl ester of
tiglic acid, was performed. The cluster [H4Ru4(CO)10(m-1,2-
2a)] (12) was used as catalyst with a catalyst/substrate ratio
of 1:500. Under standard reaction conditions for catalytic
experiments (see the Experimental Section), the results
from this experiment revealed a rather low conversion of
the substrate (68%). The enantiomeric excess was deter-
mined to be 29%, which is a significant decrease in enantio-
selectivity compared with that obtained when the corre-
sponding carboxylic acid (tiglic acid (24), 82% ee) was used
as substrate under the same reaction conditions (see
Table 7). A possible explanation for this result is that the
carboxylate functionality is an important factor when the
substrate coordinates to the metal centre. A study by Ashby
and Halpern[31] proposes that tiglic acid in the presence of
any previously tested tetrahydrido tetraruthenium clus-
ter[14,15,21] as catalysts/catalyst precursors for asymmetric hy-
drogenation. Clusters 4, 6, 8, 10–12, 14, 15 and 17–21 were
able to efficiently convert tiglic acid (24) (see Table 7) into
2-methylbutanoic acid within 24 h in good yields and in rela-
tively low molar ratios (ncatalyst/nsubstrate =1:500).
Although all catalysts were similar in conversion rates of
the substrate, the catalysts based on Josiphos derivatives
(1a–1d) of [H4Ru4(CO)12] gave no, or very low, enantiose-
lectivity, whereas the clusters derivatised with the Walphos
ligands 2a–2g resulted in good enantioselectivities. Al-
though catalysts 12, 15 and 17–21 were recovered in good
yields (ꢀ70%) after a complete catalytic experiment, only
trace amounts (at best) of catalysts 4 and 7–9 could be re-
covered. Catalyst recycling was investigated by recovering
cluster 18 from the hydrogenation of tiglic acid (24,
Scheme 1) and reusing it in the exact same experiment (cor-
responding to entry 10, Table 7). Identical conversion rates
were observed for the two catalytic runs; the enantioselec-
tivities were not investigated in these specific catalysis ex-
periments.
the mononuclear catalyst [RuII
ACTHUNTRGENNG(U BINAP)ACHTUNGTREN(NUGN OAc)2] (22) coordi-
nates exclusively through the carboxylate group to the
ruthenium as the initial step in the catalytic cycle of a hydro-
genation. It is thus possible that the initial coordination of
the type of substrate used in this study to a potential cluster
catalyst occurs through coordination of the carboxylate
moiety (see below).
In the cluster-based hydrogenation effected by
[H4Ru4(CO)
these clusters were converted to the structural isomers
[H4Ru4(CO)10A{1,1-(R,R or S,S)-bdpp}] (chelating diphos-
10ACHTUNGTRENNUNG
{m-1,2-(R,R or S,S)-bdpp}],[15] it was found that
CHTUNGTRENNUNG
phine) during the catalytic reactions. In agreement with the
opposite trend for the isomerisation reactions mentioned
above, this interconversion was observed during the catalytic
experiments with clusters 11 and 14, that is, the “chelating”
isomers 11 and 14 were converted to the analogous “bridg-
ing” isomers 12 and 15. No remnants of 11 or 14 could be
detected after the catalytic reactions but when cluster 11
was used as a starting material, a second cluster—identified
It is known that mononuclear ruthenium complexes are
viable as catalysts in the asymmetric hydrogenation of vari-
ous types of enamides.[32] In an attempt to hydrogenate (E)-
a-acetamidocinnamic acid (29) a catalytic test was per-
formed by using cluster 12. After a period of 24 h the reac-
tion mixture was analysed but it was not possible to detect
1
any hydrogenated product by H NMR spectroscopic analy-
sis.
as [H4Ru4(CO)8ACHTUNGTRENNUNG(2a)2] on the basis of its mass spectrum—
could be detected in addition to the isomer 12. It thus ap-
pears unlikely that 11 and 14 are active catalysts for the hy-
drogenation reactions.
The catalytic activities of clusters 11, 12, 15 and 17–21
were further examined in the asymmetric hydrogenations of
(E)-a-methyl-cinnamic acid (25), (E)-a-phenyl-cinnamic
acid (26) and (E)-2-methyl-2-pentenoic acid (27)
(Scheme 1). The substrates were selected so that a varying
degree of bulkiness was achieved.
Nature of the active catalyst: Clusters have been used in nu-
merous homogeneous catalysis systems.[5] A recent example
of enantioselective hydrogenation catalysis involving transi-
tion-metal carbonyl clusters is the asymmetric transfer hy-
drogenation of ketones, where Ikariya and co-workers[33] ob-
tained evidence from spectroscopic and reactivity studies
that implicate an active catalyst consisting of a trinuclear
ruthenium cluster with a chiral diiminodiphosphine ligand.
Despite the prevalence of transition-metal carbonyl clusters
in homogenous catalysis, there are no unambiguous proofs
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
13
&
ÞÞ
These are not the final page numbers!

E. Nordlander et al.
Table 7. Summary of the catalytic asymmetric hydrogenation experiments performed in the presence of clus-
ters 4, 6, 8, 10–12, 14, 15 and 17–21 as catalysts.
showed that the cluster cation
[Ru
C6Me6) CTHNUGTRENNUNG
G
(h6-C6H6)(h6-
3ACHUTNGTRENNUG ACHUTNGTRENNUGN
Conversion[a] [%] Selectivity[b] [%] ee [%] Configuration[c]
(m3-O)]+, believed to be
Entry Substrate Catalyst
Solvent
EtOH/Tol[d] 100
1
2
3
4
5
6
7
8
24
24
24
24
24
24
24
24
24
24
24
24
24
25
25
25
25
25
25
25
25
26
26
26
26
26
26
26
26
27
27
27
27
27
27
27
27
4
6
8
97
86
2:9
8:12
no ee
no ee
92
82
63
58
54
77
62
77
50
89
75
69
44
79
65
71
43
83
66
63
57
60
59
72
38
68
68
42
69
62
43
36
24
87
62
80
79
73
60
67
75
47
33
68
30
S
S
–
an active catalyst for benzene
hydrogenation, was actually the
source of colloidal Ru0 particles
that were the origin of the cata-
lytic effect. In order to probe
the nature of the active catalyst
in the hydrogenations effected
by [H4Ru4(CO)10(m-1,2-2a) (12)
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
EtOH/Tol
acetonitrile
100
100
100
100
99
100
100
99
90
10
11
12
14
15
17
18
19
20
21
11
12
15
17
18
19
20
21
11
12
15
17
18
19
20
21
11
12
15
17
18
19
20
21
96
–
100
99
S
S
S
S
R
S
S
S
S
S
S
S
R
S
S
S
S
R
R
R
S
R
R
R
R
S
S
S
R
S
S
S
S
S
S
S
S
R
S
S
S
R
S
S
S
100
100
100
100
99
100
100
64
91
89
82
91
86
70
93
65
90
83
90
89
90
90
77
100
99
99
81
99
100
99
100
100
100
100[f]
100[f]
100[f]
100[f]
100
100
100
100
85
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
87
and
[H4Ru4(CO)10(m-1,2-2b)
100
100
100
100
100
100
100
99
100
100
100
100
100
100
90
(15), which gave good enantio-
selectivity, a number of control
experiments was carried out.
Hydrogenation in the presence
of a chiral diphosphine ligand:
A control experiment was con-
ducted to determine whether
the ligand 2a possesses any cat-
alytic activity in the absence of
ruthenium metal. In a small au-
toclave, ligand 2a and the sub-
strate tiglic acid (24) were
added in a molar ratio of 1:250.
After being heated at 1008C
under 50 bar of H2 for 72 h, it
was found that 23% of 24 was
converted to 2-methylbutanoic
acid with an enantiomeric
excess of 68% (S). This re-
markable result indicates that
2a gives strong chiral induction
and is capable of acting as a cat-
alyst/catalyst precursor itself,
albeit with considerably lower
enantioselectivity than the cor-
responding ruthenium cluster
12. It is possible that the ligand
may be responsible for “back-
ground” activity during the
cluster-catalysed hydrogenation.
However, the rate by which the
free ligand 2a hydrogenates 24
is apparently very low, as indi-
cated by the low conversion
rate, and its contribution to the
48
100
100
100
100
100
100
94
100
100
70
100
100
100
97
24+Hg[e] 12
24+Hg[e] 15
24
25
26
27
24
25
26
27
24
24
12
12
12
12
12
12
12
12
22
23
acetonitrile 100
acetonitrile
acetonitrile
benzene
99
98
99
benzene
59
benzene
84
benzene
58
EtOH/Tol
EtOH/Tol
100
100
79
[a] The amount of substrate consumed in the catalytic experiment, assessed by 1H NMR spectroscopy. [b] The
carboxylate of the substrate was found to react with the solvent (ethanol) to form an ester. [c] Favoured enan-
tiomer. [d] Tol=toluene. [e] Mercury poison test. [f] [EtNH2] was observed as byproduct.
catalytic
hydrogenation
by
of clusters being active catalysts. Besides in situ detection of
an active cluster catalyst, the only proof that can be ob-
tained is probably the opportunity that the enantioselective
catalysis is effected by clusters where the chirality resides in
the cluster framework rather than in a ligand.[34] It is possi-
ble that the active catalysts in cluster-based catalysis are col-
loids or mononuclear species that are formed in situ. For ex-
ample, in a recent study, Sꢃss-Fink and co-workers[35]
using cluster 12 as catalyst must be regarded as limited.
Hydrogenation in the absence of a catalyst/catalyst precur-
sor: To rule out that no metallic residues/contaminations are
responsible for the above-mentioned observation and as
background activity in other experiments, a similar control
experiment was carried out by using standard reaction con-
ditions for the catalytic hydrogenation. In this experiment,
&
14
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Cluster-Based Catalysts for Asymmetric Hydrogenation
FULL PAPER
substrate 24 was dissolved in ethanol/toluene (1:1 v/v) in the
absence of any potential catalyst, and heated under pres-
sures of H2. After 24 h of heating no hydrogenation prod-
ucts were detected. This observation clearly indicates that
the temperature and hydrogen pressure used in our catalytic
experiments are not forcing enough to cause hydrogenation
of 24 and that no apparent metallic contamination that may
cause hydrogenation is present.
ic activity of the two mononuclear complexes 22 and 23 was
investigated by hydrogenation of tiglic acid (24) under the
same reaction conditions used for the cluster-based catalysts.
Both complexes 22 and 23 proved to be efficient catalysts
(see Table 7 entries 48 and 49) with good substrate conver-
sion but formed more byproducts than the cluster-based cat-
alysts. The observed enantioselectivities were significantly
different from those obtained by using the corresponding
cluster-based catalyst. Thus, complex 22 yielded an ee of
68% (R), exceeding the ees reported for [H4Ru4(CO)10{(S)-
BINAP}] under similar conditions1 by more than forty per-
centage units. In contrast, complex 23 generated an enantio-
meric excess of 30% (S), which is more than fifty percent-
age units less than that of complex 12. These results indicate
that mononuclear Noyori-type catalysts are not the catalyti-
cally active species in the cluster-based reactions.
In addition to the above-mentioned studies, a series of
NMR experiments by using a mixture of hydrogen enriched
in the para isomer form (parahydrogen, p-H2) were carried
out in order to utilise the parahydrogen-induced polarisation
(PHIP) effect[40] to detect low concentration species (i.e.,
catalytic cycle intermediates). Several examples of parahy-
drogen addition to transition-metal clusters have been re-
ported previously,[41] including studies of catalytic hydroge-
nation systems.[41e–g] However, signals due to the PHIP effect
could neither be detected for the reaction of
[H4Ru4(CO)10(m-1,2-2a)] (12) with a parahydrogen-enriched
mixture, nor for the hydrogenation of (E)-2-methyl-2-bute-
noic acid (tiglic acid) (24) by using parahydrogen (see the
Supporting Information).
Mercury poisoning test: It is known that forcing conditions,
such as high temperatures and high pressures of hydrogen
gas, increase the probability that a colloidal catalyst will
form. An illustrative example is the well-known Wilkinson
catalyst, [RhICl
ACHTUNGTRENNUNG(PPh3)3], which serves as an excellent homo-
geneous hydrogenation catalyst at ambient temperatures
and hydrogen pressures, but generates a colloidal catalyst
species when heated at 1308C under elevated pressure of H2
(50 psiꢀ3.4 bar).[36] Another example is the Heck reaction,
where the palladium catalyst is readily reduced to Pd0 irre-
spective of the nature of its catalyst precursor, at tempera-
tures of 1208C or higher, leading to the formation of soluble
Pd colloids that mediate the Heck reaction.[37] In order to
assess whether colloidal material is formed during the reac-
tions and may be responsible for the catalytic effects ob-
served for the ruthenium clusters in this study, a mercury
poisoning test[38] was performed. Approximately 2000-fold
excess of metallic mercury was added to the hydrogenation
of tiglic acid (24) by using clusters 12 and 15. The product
analysis revealed that upon addition of Hg0, slightly better
results were obtained than in the reactions without Hg0
added. Not only were the selectivities improved but the
enantiomeric excess was increased by four to five percent-
age units (see Table 7, entry 6 vs. entry 38, and entry 8 vs.
entry 39). These observations suggest that the cluster does
undergo fragmentation and that formation of colloidal mate-
rials occurs during the course of reaction. However, the
effect of the colloidal material seems very limited for the
hydrogenation reaction in general.
Turnover number and turnover frequency: In an attempt to
establish a turnover number (TON) for catalyst 12, a molar
substrate (24)/catalyst/Hg ratio of 10000:1:16000 was used,
1
and the reaction mixture was heated for 96 h. H/31P NMR
and IR spectroscopic analyses showed no other cluster spe-
cies than 12 present in the reaction mixture upon comple-
tion of the experiment. The substrate/product ratios were as-
sessed by 1H NMR spectroscopy and gave the following
result: tiglic acid (0.5%), 2-methylbutanoic acid (87%) and
ethyl 2-methylbutanoate (12.5%); the TON can thus be
considered to be >9900. The enantiomeric excess of the iso-
lated product 2-methylbutanoic acid was determined to be
87% (S). The results from this experiment show that the
enantioselectivity in this catalytic system is retained even
when quite high substrate/catalyst ratios are used (see
Table 7, entry 6).
Our experimental setup has precluded sampling during
the reaction, but in order to establish an understanding of
the catalytic efficiency of complex 12, a set of time-depen-
dent experiments was carried out to determine the turn over
frequency (TOF). A catalyst/substrate ratio of 1:1000 was
used, and reaction times were set to 3, 6, 12 and 24 h. The
Hydrogenation by chiral mononuclear ruthenium com-
plexes: It is reasonable to assume that any colloid-catalysed
hydrogenation would result in no or very low enantioselec-
tive discrimination. However, if cluster fragmentation
occurs, the formation of mononuclear chiral diphosphine
complexes that can lead to effective enantioselective hydro-
genation can not be excluded. In this investigation, the
building blocks for a Noyori-type catalyst with the general
formula [RuII
G
N
systems; however, the formation of such a catalytic species
would require oxidation of the metal. To investigate the pos-
sibility that such a Noyori-type catalyst is formed, and to
compare the enantioselectivities of such catalysts to those
obtained with the present cluster-based catalysts, the ruthe-
nium complexes [RuII{(R)-BINAP}
BINAP=(R)-(+)-2,2’-bis(diphenylphosphino)-1,1’-binaphth-
yl) and [RuII
{(R,R)-2a}(OAc)2] (23) were prepared by using
the synthetic route described by Noyori et al.[39] The catalyt-
ACHTUNGRTEN(NGNU OAc)2] (22) ((R)-
1
Reaction conditions used (see Ref. [14b]): T=1008C, duration=93 h,
solvent=ethanol/toluene (1:1 v/v, 20 mL), pressure H2 =130 bar, molar
ratio substrate/catalyst=2000. Results obtained: conversion=90.5%,
optical purity=28.5 % ee (R).
A
ACHTUNGTRENNUNG
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
15
&
ÞÞ
These are not the final page numbers!

E. Nordlander et al.
results for these reactions are presented in Table 8, and indi-
cate that cluster 12 is quite reactive. The results show that
under the conditions used, the standard reaction time of
Table 8. Results from turn over frequency tests of [H4Ru4(CO)10(m-1,2-
2a)] (12) used in the hydrogenation of tiglic acid (24).
Time [h]
Conversion[a] [%]
Time [h]
Conversion[a] [%]
3
6
14
92
12
24
1
95
99
[a] Conversion of substrate assessed with H NMR spectroscopy.
24 h is unnecessarily long when cluster 12 is used to hydro-
genate tiglic acid, but they do prove that the catalyst/catalyst
precursor is stable for a prolonged time under heat and
pressure. However, the results from this investigation imply
that there is an induction time of approximately 3 h (or
less). This behaviour is usually seen in catalytic systems
where the homogeneous catalyst is transformed into
a heteroACHTUNGTRENNUNGgeneous/colloidal catalyst during the induction time,
after which the reaction starts. An experiment was per-
formed to investigate this issue. In an identical catalytic test,
the pressurised reactor was allowed to “equilibrate” for 3 h
before it was heated to 1008C for another 3 h. The pre-equi-
libration step did improve the conversion by ten percentage
units relative to three hours of reaction without pre-equili-
bration, so that 24% of the substrate was hydrogenated.
In order to investigate this further, a high-pressure NMR
analysis was performed by using a high-pressure sapphire
tube (5.0 mm o.d./3.4 mm i.d.).[42] A saturated solution of
cluster 12 in [D8]toluene (0.4 mL) was prepared and added
to the tube. A reference spectrum was recorded at 298 K
before the tube was pressurised with 20 bar of hydrogen gas.
The pressurised NMR sample was given 48 h to equilibrate
before any measurements were carried out. In addition to
the hydrides typical for cluster 12, the new NMR spectrum
collected in the hydride region showed the presence of
a new signal at d=À17.67 ppm (see Figure 22). Proton spec-
tra were recorded at 60 and 808C. In comparison to the vari-
1
able-temperature H NMR spectroscopy (see Figure 5) per-
formed under normal atmospheric pressure of N2, the pres-
surised sample exhibits three sharp hydride resonances,
whereas the “normal” pressure sample shows two sharp (d=
À16.71 and À17.15 ppm) and two broad signals (d=À16.21
and À18.02 ppm). At 808C the pressurised cluster sample
shows a sharp and a broad signal, and four resonances sig-
nificantly weaker. In contrast, the normal pressure sample
shows one sharp (d=À17.09 ppm) and one broad (d=
À16.86 ppm) resonance together with a resonance of weak
intensity (d=À16.50 ppm). No darkening of the solution or
metallic precipitation was observed, indicating that cluster
fragmentation did not occur during this experiment.
A kinetic study by Doi et al.[43] has shown that the parent
cluster [H4Ru4(CO)12] hydrogenates ethene in heptane at
728C under hydrogen pressures of 0.1–0.4 bar. It has been
demonstrated that evolution of carbon monoxide occurs
when a solution of [H4Ru4(CO)12] in heptane is exposed to
Figure 22. 1H NMR spectra of [H4Ru4(CO)10(m-1,2-2a)] (12) under 20 bar
H2 pressure at a) 298, b) 333 and c) 353 K.
&
16
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Cluster-Based Catalysts for Asymmetric Hydrogenation
FULL PAPER
hydrogen[43,44] and the suppression of catalysis by addition of
CO[43,44] suggests initial decarbonylation of the cluster and
ficiency offer several distinct advantages: they are robust
and are not air sensitive, separation from the products is rel-
atively facile and they permit low catalyst-to-substrate
ratios.
formation of [H4Ru4(CO)
11ACHTUNGERTN(NUNG alkene)] and, subsequently,
[H3Ru4(CO)11A
CHTUNGTRENNUNG
posed that the initial coordination of the alkene occurs in
competition with coordination of H2 to form the hexahydri-
do species [H6Ru4(CO)11] and it may be such a species that
is detected in the above-mentioned high-pressure NMR ex-
periment under H2 pressure, which takes place in the ab-
sence of substrate. We have attempted the catalytic reduc-
tion of tiglic acid (24) by [H4Ru4(CO)10(m-1,2-2a)] (12)
under the conditions used by Doi et al.[43] but no conversion
could be detected after 14 h; again, the hydrogen pressure
(and/or the temperature) appears to be too low to enable
hydrogenation of the substrate. It should be mentioned that
an alternative catalytic pathway[15] involves an initial step
where a metal–metal bond rupture leads to the formation of
a two-electron-deficient 60-electron butterfly cluster fol-
lowed by coordination of the substrate. This alternative re-
action pathway does not require ligand dissociation, but
would probably still be susceptible to competitive reactions,
that is, the association of either the substrate or dihydrogen.
In an attempt to further monitor the reaction system
under catalytic conditions, a high-pressure IR spectroscopy
experiment was conducted. Cluster 12 (20 mg) was dissolved
in an ethanol/toluene mixture (1:1 v/v, 20 mL) and heated to
1008C at a total hydrogen pressure of 50 bar. However, no
change could be observed in the carbonyl region and the ex-
periment was terminated after 5 h of heating. The sample
was left under pressure for another 15 h at ambient temper-
ature, but no change in the spectrum could be seen. In an-
other experiment, cluster 12 and substrate 24 (ncatalyst/nsubstrate
1:100) were dissolved in cyclohexane and heated to 1008C
under a hydrogen pressure of 50 bar. No detection of any
other cluster species than 12 could be observed during the
experiment lasting six hours. This experiment provides fur-
ther evidence of the stability of cluster 12 under catalytic
conditions, suggesting that the cluster may indeed be the
active catalyst, or a direct precursor for an active cluster cat-
alyst.
Experimental Section
General: All reactions were performed under inert atmosphere (nitrogen
or argon) and manipulations of the products were performed in air. The
reactions were performed at room temperature (ꢀ218C) unless stated
otherwise. All solvents used in syntheses and catalysis experiments were
distilled and dried prior to use. The parent cluster [H4Ru4(CO)12] was
prepared according to a literature method.[1,45] 1H and 31P NMR spectra
were recorded on a Varian Unity 300 MHz spectrometer, a Varian Inova
500 MHz spectrometer or
a JEOL EX-400 MHz spectrometer. The
1H NMR spectra were referenced to the relevant resonances of the non-
deuterated solvents (CHCl3, d=7.25 ppm; CH2Cl2, d=5.31 ppm; toluene,
d=2.09 ppm; benzene, d=7.15 ppm). 31P NMR shifts were referenced to
external H3PO4 (85%). Infrared spectra were recorded on a Nicolet
Avatar 360 FTIR spectrometer. Thin-layer chromatography was per-
formed on commercially available 20ꢄ20 cm glass plates, covered with
Merck Kieselgel 60 to 0.25 mm thickness.
A number of prototypical syntheses for the different structural types of
Josiphos- and Walphos-containing clusters (see Table 3) are described
below. Full documentation of synthetic and spectroscopic data for the re-
maining new clusters can be found in the Supporting Information.
Synthesis of [H4Ru4(CO)10ACTHNUTRGNE(UNG 1,1-1a)] (3) and [H4Ru4(CO)10(m-1,2-1a)] (4):
[H4Ru4(CO)12] (23 mg, 31 mmol) and (R)-1-[(S)-2-(diphenylphosphino)-
ferrocenyl]ethyldicyclohexylphosphine (1a) (20 mg, 31 mmol) were dis-
solved in a solvent mixture of dichloromethane (25 mL) and acetonitrile
(10 mL). Under vigorous stirring,
a small excess of Me3NO (4 mg,
36 mmol) dissolved in acetonitrile (15 mL) was added dropwise to the
orange cluster/ligand solution, which instantly started to change colour
towards red. After 3 h the solvents were removed under reduced pres-
sure. The red solid residue obtained was dissolved in a small quantity of
dichloromethane and purified by using preparative TLC (CH2Cl2/n-
hexane 3:2 v/v). Except for traces of starting materials and one stationary
brown band (decomposed material), two bands were isolated from the
TLC plates and extracted with CH2Cl2 and dried under vacuum. The two
solids obtained, orange and red, were identified as [H4Ru4(CO)10ACHTUNGTRENNUNG(1,1–
1a)] (3) (4 mg, 10%); 1H NMR (hydride resonances, 500 MHz, CDCl3):
d=À16.0 (ddd, J=32.7, 15.0, 2.6 Hz), À16.4 (s), À17.7 (t, J=24.4 Hz),
À18.9 ppm (dt, J=8.6, 2.4 Hz); 31P{1H} NMR (202 MHz, CDCl3): d=
63.5, 23.4 ppm; IR (CH2Cl2): n˜CO =2072 (s), 2041 (vs), 2018 (vs), 1997 (s),
1977 (s), 1977 (br), 1965 cmÀ1 (br); MS (FAB): m/z: 1284 [M]+; and
[H4Ru4(CO)10(m-1,2–1a)] (4) (24 mg, 60%) 1H NMR (hydride resonan-
ces, 500 MHz, CDCl3): d=À16.6 (sbr), À17.3 (sbr), À17.7 (sbr),
À18.0 ppm (sbr); 31P{1H} NMR (202 MHz, CDCl3): d=65.8, 32.1 ppm;
IR (CH2Cl2): n˜CO =2065 (s), 2044 (s), 2022 (vs), 2001 (s), 1990 (br), 1979
(br), 1956 cmÀ1 (br); MS (FAB): m/z: 1283 [MÀ1]+.
Conclusion
In summary, we have synthesised and characterised new tet-
raruthenium clusters, derivatised with chiral ferrocenyl-
based diphosphine ligands and used these clusters as cata-
lysts for asymmetric hydrogenation of a-unsaturated carbox-
ylic acids. The catalytic activities for certain clusters report-
ed here outclass all previous attempts to effect the enantio-
selectivity in asymmetric hydrogenation by using
[H4Ru4(CO)10(diphosphine)] clusters as catalysts. Further-
more, the results obtained in this investigation clearly dem-
onstrate the importance of the chiral ligand used. Although
it has not been possible to unambiguously determine the
nature of the active catalyst(s) in these cluster-based sys-
tems, the clusters in this study that exhibit good catalytic ef-
Synthesis of [H4Ru4(CO)
(12)
10ACHTUNGTERNNUG(1,1-2a)] (11) and [H4Ru4(CO)10ACHTUNGTRENNUNG(m-1,2-2a)]
Method A (oxidative decarbonylation/ligand substitution): An excess of
Me3NO (8 mg, 72 mmol) dissolved in methanol (5 mL) was added drop-
wise to a stirred solution of (R)-1-[(R)-2-(2’-diphenylphosphinophenyl)-
ferrocenyl]ethyldiACHTNUTGRNEUNG[bis-(3,5-trifluoromethyl)phenyl]phosphine (2a) (40 mg,
43 mmol) and [H4Ru4(CO)12] (18 mg, 24 mmol) in benzene (20 mL) over
a period of 20 min. During the addition of Me3NO, the colour of the solu-
tion changed gradually from orange to deep red. After 5 h the solvent
was removed under vacuum and the red solid obtained was dissolved in
a small volume of CH2Cl2 (ꢀ2 mL), and purified by using preparative
TLC (eluent: hexane/CH2Cl2 1:1 v/v). Except for traces of unreacted
ligand and a stationary brown band, two new products were observed
and identified as [H4Ru4(CO)
dride resonances, 500 MHz, CDCl3): d=À15.67 (s), À15.97 (s), À16.51
(1,1-2a)] (11) (3 mg, 7%); 1H NMR (hy-
10ACHTUGNTRENNNUG
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
17
&
ÞÞ
These are not the final page numbers!

E. Nordlander et al.
(mbr), À16.75 ppm (s); 31P{1H} NMR (202 MHz, CDCl3): d=43.32 (s),
4.34 ppm (s); IR (CH2Cl2): n˜CO =2078 (s), 2049 (s), 2027 (vs), 2003 (m),
1983 (m, sh), 1965 (v, sh), 1949 cmÀ1 (v sh); MS (FAB): m/z: 1620 [M]+;
scan rate used was 100 mVsÀ1. Under these conditions the ferrocene/fer-
rocenium couple, which was used as a reference, had an E1/2 value of
+0.09 V and DEp =120 mV. All solutions were purged with argon and
voltammograms were recorded under a blanket of argon. The platinum
disk working electrode was polished between runs.
1
and [H4Ru4(CO)10(m-1,2-2a)] (12) (17 mg, 43%); H NMR (hydride reso-
nances, 500 MHz, CDCl3): d=À16.4 (mbr), À16.7 (d, J=7.3 Hz), À17.6
(td, J=11, J=2.4 Hz), À18.5 ppm (brm, Jꢀ30 Hz); 31P{1H} NMR
(202 MHz, CDCl3): d=45.9 (m), 39.1 ppm (m); IR (CH2Cl2): n˜CO =2073
(ms), 2053 (ms), 2028 (vs), 2013 (s), 1998 (w), 1965 cmÀ1 (w); MS (FAB):
m/z: 1620 [M]+. Single crystals suitable for X-ray diffraction analysis
were obtained by re-crystallisation from a dichloromethane/hexane/meth-
anol solution.
Catalysis experiments: In a typical catalytic experiment, a mini bench au-
toclave (Carl Roth, 100 mL) was charged with (cluster) catalyst (10 mg)
and substrate (250 or 1000 molar excess) under inert atmosphere. A de-
gassed toluene/ethanol mixture (5–6 mL, 1:1 v/v) was added. The reaction
vessel was closed and purged with hydrogen gas (4ꢄ20 bar) before final
pressurising to 50 bar. The reaction mixture was continuously stirred with
a magnetic stirrer (ꢀ800 rpm) and heated to 1008C for 24 h. The auto-
clave was allowed to cool to ambient temperature before the hydrogen
gas was cautiously released. The reaction mixture was transferred to
a round bottom flask and concentrated under vacuum. The conversions
for the catalysis runs were calculated on the basis of NMR analyses. To
separate the carboxylic acid from the cluster, the reaction residue was
dissolved in diethyl ether (10 mL) and the carboxylic acid was extracted
with aqueous sodium hydroxide solution (1m, 3ꢄ10 mL) and washed
with diethyl ether (2ꢄ5 mL), leaving the cluster in the organic solvent.
The carboxylate was protonated with sulfuric acid (conc., q.s.), extracted
with diethyl ether (3ꢄ10 mL), washed with water (2ꢄ5 mL) and dried
over magnesium sulfate. The ether was removed under vacuum, yielding
the carboxylic acid quantitatively. The ether phase, from which the car-
boxylic acid was extracted, was concentrated under vacuum to recover
the remaining cluster. In certain cases, where ester formation was ob-
served during the catalytic experiment, the recovered cluster was dis-
solved in a minimum quantity of dichloromethane and the products were
separated by using preparative TLC, eluting with hexane/dichlorome-
thane (1:1 v/v). Usually 60–70% of the cluster was recovered after a cata-
lytic run, and was analysed by IR and NMR spectroscopy.
Synthesis of [H4Ru4(CO)10ACHTUNTRGNEUNG(m-1,2–2a)] (12)
Method B (thermal ligand substitution at high-pressure): In a high-pres-
sure autoclave, [H4Ru4(CO)12] (30 mg, 42 mmol) and 2a (50 mg, 54 mmol)
were suspended in benzene (10 mL). The autoclave was assembled and
pressurised with H2, (25 bar) after being purged with H2 (3ꢄ10 bar). The
autoclave was heated at 1008C for 4 h, after which it was allowed to cool
to room temperature. The autoclave was opened and the solvent of the
reaction mixture was removed by using rotary evaporation. The resultant
red solid was dissolved in a minimum quantity of CH2Cl2 and filtered
through a short plug of Merck Kieselgel 60. The red solid obtained after
concentration under vacuum was identified as [H4Ru4(CO)10(m-1,2-2a)]
(12) (50 mg, 73%).
Synthesis of [H4Ru4(CO)11(2a)] (13)
Method A (oxidative decarbonylation/ligand substitution: A small excess
of Me3NO (5.4 mg, 50 mmol) dissolved in methanol (5 mL) was added
dropwise to a stirred solution of 2a (40 mg, 43 mmol) and [H4Ru4(CO)12]
(18 mg, 24 mmol) in benzene (20 mL) over a period of 20 min. During the
addition of Me3NO the colour of the solution changed gradually from
orange to deep red. After 40 min the solvent was removed under vacuum
and the red solid obtained was dissolved in a small volume of CH2Cl2
The enantiomeric excess of the product was determined by converting
the reduced carboxylic acid with (S)-mandelate and analysing the diaste-
reomeric mixture by NMR spectroscopy, as described by Tyrell et al.[46] It
was found that flash chromatography of the final product(s) was not nec-
essary.
ACHTUNGTRENNUNG
CTHUNGTRENNUNG
(17 mg, 44%); 1H NMR (hydride resonances, 500 MHz, CDCl3): d=
À17.14 ppm (s); 31P{1H} NMR (202 MHz, CDCl3): d=38.90 (s), 4.68 ppm
(s); IR (CH2Cl2): n˜CO =2094 (w), 2086 (m), 2061 (vs), 2053 (vs), 2031
(vs), 2007 (s), 1990 (m), 1963 cmÀ1 (w); MS (FAB): m/z: 1650 [M+2]+;
and 12 (8 mg, 20%). Single crystals suitable for X-ray analysis were
grown from a dichloromethane/hexane/methanol solution.
Catalyst poisoning test by using mercury: The experimental setup fol-
lowed the procedure described above for catalytic experiments, except
that metallic mercury (ca. 2 g) was added to the reaction mixture before
the autoclave was sealed and pressurised. After a complete catalytic run,
the Hg had acquired a faint grey film on some parts of its surface. Prod-
ucts were separated and analysed as mentioned above.
Synthesis of [RuII{(R)-BINAP}(OAc)2] (22) and [RuII
ACHTUNGTRENNGUN ACHTUNTREGG(NNUN 2a)ACHTGUNTREN(NUNG OAc)2] (23):
Compounds 22 and 23 were prepared according to a literature method.[39]
Triethylamine (72 mL, 0.51 mmol) was added to a solution of [RuCl2-
Determination of the TON: In a bench autoclave (Parr instruments,
450 mL) [H4Ru4(CO)10(m-1,2–2a)] (12) (10 mg, 6.18 mmol) and tiglic acid
(6.056 g, 60.6 mmol) were dissolved in an ethanol/toluene mixture
(60 mL, 1:1 v/v) in the presence of metallic mercury (20.03 g, 0.1 mol).
The reaction vessel was sealed and purged several times before being
pressurised with H2 (50 bar). The autoclave was heated at 1008C under
continuous stirring for 96 h. It was then cooled down to room tempera-
ture before carefully opened, and the products were separated and ana-
lysed as described above for the catalytic experiments.
ACHTUNGTRENNUNG(COD)] (COD=cyclooctadiene) (28 mg, 0.1 mmol) and 2a (100 mg,
0.1 mmol) in toluene (5 mL). The mixture was heated to reflux for 12 h
and the colour of the solution changed from yellow-brown to transparent
orange-brown. The reaction mixture was allowed to cool to ambient tem-
perature and the solvent was removed under vacuum. The resultant
brownish solid residue was dissolved in dichloromethane, filtered through
a short plug of celite and concentrated under vacuum. The solid residue
was dissolved in tert-butanol (5 mL) and anhydrous sodium acetate
(40 mg, 0.5 mmol) was added. The mixture was heated to reflux for 12 h.
Once the reaction mixture reached ambient temperature, the solvent was
removed under vacuum, and the solid residue was extracted with diethyl
ether (3ꢄ3 mL) and concentrated under vacuum. The resulting solid was
extracted with ethanol (3ꢄ3 mL) and evaporation of the solvent under
Control experiment with ligand 2a and tiglic acid: In a small autoclave,
2a (10 mg, 10.7 mmol) and tiglic acid (263 mg, 2.7 mmol) were dissolved
in an ethanol/toluene mixture (5 mL, 1:1 v/v). The autoclave was sealed
and purged with H2 (3ꢄ15 bar) before being pressurised with H2 (50 bar)
and heated to 1008C for 78 h. Products were separated and analysed as
described above for the catalytic experiments.
vacuum afforded a yellow-brown solid identified as [RuII
ACTHNUTRGNNEG(U 2a)CAHTUNGTERN(NUGN OAc)2]
(23) (35 mg, 32%); 31P{H} NMR (202 MHz, CDCl3): d=93.2 (sbr),
86.9 ppm (sbr); IR (CH2Cl2): n˜CO =1572 (m), 1463 cmÀ1 (m).
NMR measurements: All NMR solvents (Aldrich) were used as received.
The NMR measurements were made by using 5 mm (o.d.) tubes fitted
with J. Young Teflon valves. Typically approximately 0.5 mL of a saturated
[D2]dichloromethane solution of [H4Ru4(CO)10(m-1,2-2a)] (12) was added
and the tube was sealed under N2 atmosphere. For high-temperature
NMR studies, [D8]toluene was used as solvent. The EXSY experiment[47]
was performed by using standard parameters. The mixing time was se-
lected to be 0.5 s. The 1H,31P HetCor experiment was performed with
a solution of 12 in [D3]chloroform at 223 K.
Electrochemistry: Cyclic voltammetry was performed at ambient temper-
ature by using a Bioanalytical Systems Inc. BAS 100W Electrochemical
Analyser with a one-compartment three-electrode system comprising
a platinum disk working electrode, a platinum wire auxiliary electrode
and a Ag/Ag+ reference electrode (0.01m AgNO3 and 0.1m [nBu4N]-
ACHTUNGTRENNUNG[ClO4] in anhydrous acetonitrile). The reported E values (see Table 6)
are with reference to this electrode. Measurements were made on anhy-
drous acetonitrile solutions that were 2 mm in sample and contained 0.1m
[nBu4N]ACHTUNGTRENNUNG[ClO4] as background electrolyte. Unless otherwise stated, the
High-pressure NMR measurements were carried out in a 5 mm (o.d.)
sapphire tube, which was charged with 0.3 mL of a saturated solution of
&
18
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Cluster-Based Catalysts for Asymmetric Hydrogenation
FULL PAPER
Table 9. Crystallographic data for clusters 4, 7–9, 12, 13, 15, 17 and 19–21.
4
7
8
9
12
13
formula
Mr
C46H48FeO10P2Ru4 C46H60FeO10P2Ru4 C47H62Cl2FeO10P2Ru4 C50H44FeO10P2Ru4 C62H50F12Fe O10P2Ru4 C58H38Cl2F12FeO11P2Ru4
1282.91
1295.01
1379.94
1326.92
1705.09
1731.85
T [K]
l [ꢂ]
100(2)
0.71073
120(2)
0.71073
120(2)
0.71073
120(2)
0.71073
100(2) K
0.71073
120(2)
0.71073
crystal system
space group
a [ꢂ]
b [ꢂ]
c [ꢂ]
orthorhombic
P212121
12.3611(2)
16.7583(2)
23.1317(3)
90
4791.76(11)
4
1.778
monoclinic
P21
monoclinic
P21
orthorhombic
P212121
11.0794(2)
19.1761(3)
23.1694(3)
90
4922.56(13)
4
1.790
monoclinic
C2
orthorhombic
P212121
13.02990(10)
15.0032(2)
32.0496(5)
90
6265.38(14)
4
1.836
10.3579(4)
20.9527(11)
12.0081(10)
110.280(5)8
2444.5(3)
2
10.79530(10)
21.0136(2)
11.66880(10)
99.979(8)8
2607.00(4)
2
44.368(9)
10.745(2)
14.356(3)
103.65(3)8
6651(2)
b [8]
V [ꢂ3]
Z
4
1calcd [gcmÀ3
]
1.759
1.617
1.758
1.621
1.703
1.236
m [mmÀ1
]
1.649
1.609
1.398
crystal size [mm]
q limits [8]
h range
k range
l range
0.32ꢄ0.22ꢄ0.06
2.94 to 27.53
À14 to 16
À21 to 21
À27 to 30
47056
0.28ꢄ0.27ꢄ0.26
2.66 to 27.50
À12 to 13
À27 to 27
À15 to 15
32047
0.61ꢄ0.35ꢄ0.32
2.99 to 27.47
À13 to 14
À27 to 27
À14 to 15
31635
0.57ꢄ0.22ꢄ0.10
3.52 to 27.47
À14 to 12
À24 to 18
À29 to 28
32325
0.26ꢄ0.23ꢄ0.08
2.10 to 27.50
À57 to 57
À13 to 13
À18 to 18
40378
0.22ꢄ0.19ꢄ0.06
2.98 to 26.50
À16 to 15
À18 to 18
À40 to 38
42090
reflns collected
independent reflns
data/restraints/
parameters
10952
10952/0/585
10772
10772/1/585
11524
11524/1/612
10629
10629/0/625
15198
15198/1/793
12963
12963/56/830
goodness-of-fit on F2 1.057
1.022
0.0169
0.0323
1.046
0.0219
0.0567
1.069
0.0218
0.0477
1.015
0.0280
0.0634
1.033
0.0357
0.0745
R1[a]
0.0230
0.0459
wR2[a]
weighting scheme
x=0.0195
y=0.0570
0.543/À0.592
x=0.0120
y=0.1866
0.590/À0.321
x=0.0274
y=2.619
1.142/À1.187
x=0.0230
y=0.7241
0.410/À0.727
x=0.0370
y=5.7466
0.946/À0.770
x=0.0376
y=2.8402
0.980/À0.832
largest diff. peak/
hole [eꢂÀ3
]
Flack parameter
À0.033(13)
À0.002(9)
À0.008(13)
À0.014(12)
À0.003(13)
0.008(18)
15
17
19
20
21
formula
Mr
C52H40FeO10 P2Ru4
1346.91
C52H52FeO10P2 Ru4
1359.01
C58H51FeNO10P2Ru4
1444.07
C57H50Cl2F12FeO10P2Ru4
1715.94
C61H58Cl2Fe O10P2Ru4
1544.04
T [K]
120(2)
120(2)
120(2)
120(2)
120(2)
l [ꢂ]
0.71073
0.71073
0.71073
0.71073
0.71073
crystal system [mm]
orthorhombic
P212121
10.836(2)
14.243(3)
32.129(6)
90
4958.9(17)
4
1.804
monoclinic
P21
orthorhombic
P212121
11.48040(10)
19.9411(4)
24.9251(5)
90
5706.15(17)
4
1.681
orthorhombic
P212121
12.18500(10)
20.0480(4)
25.4160(5)
90
6208.75(18)
4
1.836
monoclinic
P21/n
space group
a [ꢂ]
b [ꢂ]
c [ꢂ]
11.6004(2)
17.8442(5)
12.3817(3)
93.325(2)8
2558.70(10)
2
11.2410(2)
20.8736(5)
26.4175(4)
95.2970(10)8
6172.1(2)
4
b [8]
V [ꢂ3]
Z
1calcd [gcmÀ3
]
1.764
1.550
1.662
1.380
m [mmÀ1
]
1.599
1.396
1.408
crystal size
q limits [8]
h range
k range
l range
0.19ꢄ0.16ꢄ0.09
2.27 to 27.54
À9 to 14
À15 to 15
À41 to 41
27802
0.40ꢄ0.12ꢄ0.05
1.65 to 27.48
À14 to 15
À23 to 22
À15 to 16
25132
0.22ꢄ0.11ꢄ0.10
3.16 to 27.47
À13 to 14
À21 to 25
À32 to 32
46916
0.18ꢄ0.07ꢄ0.06
3.08 to 26.37
À15 to 14
À25 to 24
À31 to 31
63118
0.18ꢄ0.11ꢄ0.08
3.64 to 25.37
À13 to 12
À25 to 25
À31 to 31
76285
reflns collected
independent reflns
data/restraints/
parameters
10239
10239/0/624
10548
10548/1/623
12996
12996/0/707
12642
12642/0/806
11292
11292/0/730
goodness-of-fit on F2
R1[a]
1.123
0.0702
0.1768
1.109
0.0317
0.0688
1.033
0.0292
0.0565
1.041
0.0424
0.0871
1.099
0.0421
0.0915
wR2[a]
weighting scheme
x=0.0488
y=85.0794
1.488/À1.898
x=0.0467
y=0.0000
0.683/À1.402
x=0.0256
y=1.4306
0.869/À0.586
x=0.0358
y=14.3902
1.390/À1.427
x=0.0302
y=21.7512
1.523/À1.335
largest diff. peak/
hole [eꢂÀ3
]
Flack parameter
0.14(6)
À0.02(2)
À0.021(15)
À0.03(2)
–
[a] Iꢁ2s(I), w=1/[s
ACHTUNGTRENNUNG
2(Fo2)+(xP)2+yP], P=(Fo2+2Fc2)/3.
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
19
&
ÞÞ
These are not the final page numbers!

E. Nordlander et al.
12 in [D8]toluene. The NMR tube was pressurised with H2 (20 bar) and
was stored in its safety device at ambient temperature between the NMR
measurements.
[1] H. U. Blaser, F. Spindler, M. Studer, Appl. Catal. A 2001, 221, 119.
[2] a) I. Ojima, N. Clos, C. Bastos, Tetrahedron 1989, 45, 6901; b) W. S.
Knowles, R. Noyori, Acc. Chem. Res. 2007, 40, 1238; c) W. Li, G.
Hou, X. Sun, G. Shang, W. Zhang, X. Zhang, Pure Appl. Chem.
2010, 82, 1429.
[3] a) W. S. Knowles, Angew. Chem. 2002, 114, 2096; Angew. Chem. Int.
Ed. 2002, 41, 1998; b) W. S. Knowles, Acc. Chem. Res. 1983, 16, 106.
[4] a) M. J. Burk, M. F. Gross, J. P. Martinez, J. Am. Chem. Soc. 1995,
117, 9375; b) M. J. Burk, Y. M. Wang, J. R. Lee, J. Am. Chem. Soc.
1996, 118, 5142.
[5] a) A. K. Smith, J. M. Basset, J. Mol. Catal. 1977, 2, 229; b) R. D.
Adams, F. A. Cotton, Catalysis by Di- and Polynuclear Metal Cluster
Complexes, Wiley-VCH, Weinheim 1998; c) W. L. Gladfelter, K. J.
Roesselet in The Chemistry of Metal Cluster Complexes (Eds.: D. F.
Shriver, H. D. Kaesz, R. D. Adams), VCH, Weinheim 1990, pp. 329–
365; d) R. D. Adams, Aldrichimica Acta 2000, 33, 39; e) P. J. Dyson,
Coord. Chem. Rev. 2004, 248, 2443.
Parahydrogen experiments: A 52% para-enriched hydrogen mixture was
obtained by filling a glass spiral full of activated carbon with normal hy-
drogen (pressure=1.2 bar) in a liquid nitrogen bath (77 K) for 30 min.
The parahydrogenation experiments were carried out in NMR tubes with
Young Teflon valves (5 mm o.d.). An aliquot of a solution of 12 (0.5 mL,
1.2 mm) in a [D8]toluene/[D6]ethanol mixture (1:1) was prepared. The so-
lution was deoxygenated, then filled with parahydrogen (4.5 bar) at room
temperature and finally the tube was shaken thoroughly for 30 s, after
which it was placed in the NMR probe. Then single scan 1H or 31P NMR
spectra were acquired.
Several experiments by using [H4Ru4(CO)10(m-1,2-2a)] (12) as a parahy-
drogenation catalyst were carried out in
a mixture of [D8]toluene/
[D6]ethanol (1:1). (E)-2-Methyl-2-butenoic acid (tiglic acid) (24) (31 mg,
0.31 mmol) was added to a deoxygenated solution of 12 (1.2 mm, catalyst/
substrateꢀ1:500). The solution was then filled with parahydrogen
(4.5 bar) at room temperature, shaken for 30 s and then introduced in the
[6] G. Hogarth, S. E. Kabir, E. Nordlander, Dalton Trans. 2010, 39,
6153.
1
NMR magnet for single-scan H NMR measurements.
[7] R. D. Adams, J. Organomet. Chem. 2000, 600, 1.
[8] G. Sꢃss-Fink, G. Herrmann, J. Chem. Soc. Chem. Commun. 1985,
735.
[9] G. Sꢃss-Fink, Angew. Chem. 1982, 94, 72; Angew. Chem. Int. Ed.
Engl. 1982, 21, 73.
High-pressure IR spectroscopy: All high-pressure infrared spectra were
recorded with a Bruker Equinox 55 spectrometer by using the OPUS
software program. The design, construction and mode of operation of the
high-pressure infrared cell have been described earlier.[48] All spectra
were collected by using Blackman-Harris 3-Term apodisation.
[10] G. Sꢃss-Fink, G. Herrmann, Angew. Chem. 1986, 98, 568; Angew.
Chem. Int. Ed. Engl. 1986, 25, 570.
[11] M. W. Payne, D. L. Leussing, S. G. Shore, J. Am. Chem. Soc. 1987,
109, 617.
[12] U. Matteoli, M. Bianchi, P. Frediani, F. Piacenti, C. Botteghi, M.
Marchetti, J. Organomet. Chem. 1984, 263, 243.
[13] M. Bianchi, G. Menchi, P. Frediani, U. Matteoli, F. Piacenti, J. Orga-
nomet. Chem. 1983, 247, 89.
[14] a) C. Botteghi, S. Gladiali, M. Bianchi, U. Matteoli, P. Frediani, P. G.
Vergamini, E. Benedetti, J. Organomet. Chem. 1977, 140, 221; b) U.
Matteoli, V. Beghetto, A. Scrivanti, J. Mol. Catal. A 1996, 109, 45;
c) U. Matteoli, G. Menchi, P. Frediani, M. Bianchi, F. Piacenti, J. Or-
ganomet. Chem. 1985, 285, 281; d) M. Bianchi, U. Matteoli, G.
Menchi, F. Gloria, P. Frediani, F. Piacenti, C. Botteghi, J. Organo-
met. Chem. 1980, 195, 337.
[15] P. Homanen, R. Persson, M. Haukka, T. A. Pakkanen, E. Nordland-
er, Organometallics 2000, 19, 5568.
[16] P. Barbaro, C. Bianchini, G. Giambastiani, S. L. Parisel, Coord.
Chem. Rev. 2004, 248, 2131.
X-ray structure determinations: The crystals of clusters 4, 7–9, 12, 13, 15,
17 and 19–21 were immersed in cryo-oil, mounted in a Nylon loop, and
measured at a temperature of 100–120 K. The X-ray diffraction data
were collected by means of a Nonius Kappa CCD diffractometer by
using MoKa radiation (l=0.710 73 ꢂ). The Denzo-Scalepack[49] or
EvalCCD[50] program packages were used for cell refinements and data
reductions. The structures were solved by direct methods by using
SIR2002,[51] SIR2004[52] or SHELXS-97[53] with the WinGX[54] graphical
user interface. A semi-empirical absorption correction (SADABS[55] or
Xprep in the SHELXTL[56] program package) was applied to all data.
Structural refinements were carried out by using SHELXL-97.[57] For
cluster 13 the fluorine atoms F1 F2, F11 and F12 belonging to two of the
CF3 groups were disordered over two sites. F1 and F2 as well as F11 and
F12 were refined with an effective standard deviation of 0.08 so that
their Uij components approximated to isotropic behaviour. The hydride
hydrogen atoms were either located from the difference Fourier map or
placed on idealised positions by using the XHYDEX[58] program. Other
hydrogen atoms were positioned geometrically and constrained to ride
À
on their parent atoms, with C H=0.95–1.00 ꢂ and
Uiso =1.2–
1.5 Ueq(parent atom). The crystallographic details are summarised in
Table 9. CCDC-876461 (4), 876462 (7), 876463 (8), 876464 (9), 652265
(12), 876465 (13), 876466 (17), 876467 (19), 876468 (20) and 876469 (21)
contain the supplementary crystallographic data for this paper. These
data can be obtained free of charge from The Cambridge Crystallograph-
ic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
[17] C. Godard, A. Ruiz, C. Claver, Helv. Chim. Acta 2006, 89, 1610.
[18] a) H. U. Blaser, F. Spindler, Top. Catal. 1997, 4, 275; b) H. U. Blaser,
B. Pugin, F. Spindler, M. Thommen, Acc. Chem. Res. 2007, 40, 1240.
[19] a) T. Sturm, L. Xiao, W. Weissensteiner, Chimia 2001, 55, 688; b) T.
Sturm, W. Weissensteiner, F. Spindler, Adv. Synth. Catal. 2003, 345,
160.
[20] V. Moberg, M. Haukka, I. O. Koshevoy, R. Ortiz, E. Nordlander,
Organometallics 2007, 26, 4090.
[21] V. Moberg, P. Homanen, S. Selva, R. Persson, M. Haukka, T. A.
Pakkanen, M. Monari, E. Nordlander, Dalton Trans. 2006, 279.
[22] D. Braga, U. Matteoli, P. Sabatino, A. Scrivanti, J. Chem. Soc.
Dalton Trans. 1995, 419.
[23] S. P. Tunik, T. S. Pilyugina, I. O. Koshevoy, S. I. Selivanov, M.
Haukka, T. A. Pakkanen, Organometallics 2004, 23, 568.
[24] M. R. Churchill, R. L. Lashewycz, J. R. Shapley, Inorg. Chem. 1980,
19, 1277.
[25] R. D. Wilson, S. M. Wu, R. A. Love, R. Bau, Inorg. Chem. 1978, 17,
1271.
[26] A. U. Hꢅrkçnen, M. Ahlgrꢆn, T. A. Pakkanen, J. Pursiainen, J. Or-
ganomet. Chem. 1997, 530, 191.
Acknowledgements
This paper is dedicated to the memory of Prof. John R. Moss, an inspiring
colleague, mentor and friend, in recognition of his outstanding contribu-
tions to organometallic chemistry, including pioneering studies in cluster
chemistry. The research has been sponsored by the Swedish Research
Council (VR), the Swedish International Development Agency (SIDA),
the South African National Research Foundation (NRF) and the Cra-
foord Foundation. We thank Prof. Roberto Gobetto for help and advice
in the para-hydrogenation and HetCor NMR experiments. Furthermore,
we thank Ahmed Fawzy Abdel-Magied and Anita Hoang for carrying
out the catalyst recycling study. Finally, we thank Solvias AG, Basel,
Switzerland, for generous gifts of samples of the ligands used in this in-
vestigation.
[27] M. G. Ballinas-Lꢇpez, E. V. Garcꢈa-Bꢉez, M. J. Rosales-Hoz, Poly-
hedron 2003, 22, 3403.
[28] F. Barriꢊre, R. U. Kirss, W. E. Geiger, Organometallics 2005, 24, 48,
and references therein.
&
20
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Cluster-Based Catalysts for Asymmetric Hydrogenation
FULL PAPER
[29] J. Nijhoff, M. J. Bakker, F. Hartl, G. Freeman, S. L. Ingham, B. F. G.
Johnson, J. Chem. Soc. Dalton Trans. 1998, 2625.
[30] W. S. Knowles, Acc. Chem. Res. 1983, 16, 106.
[31] M. T. Ashby, J. Halpern, J. Am. Chem. Soc. 1991, 113, 589.
[32] a) R. Noyori, M. Ohta, Y. Hsiao, M. Kitamura, J. Am. Chem. Soc.
1986, 108, 7117; b) M. Kitamura, Y. Hsiao, R. Noyori, H. Takaya,
Tetrahedron Lett. 1987, 28, 4829; c) M. Kitamura, Y. Hsiao, M.
Ohta, H. Tsukamoto, T. Ohta, H. Takaya, R. Noyori, J. Org. Chem.
1994, 59, 297.
[33] H. Zhang, C.-B. Yang, Y.-Y. Li, Z.-R. Donga, J.-X. Gao, H. Naka-
mura, K. Murata, T. Ikariya, Chem. Commun. 2003, 142.
[34] a) J. Collin, C. Jossart, G. Balavoine, Organometallics 1986, 5, 203;
b) L. Vieille-Petit, G. Sꢃss-Fink, B. Therrien, T. R. Ward, H. Stoeck-
li-Evans, G. Labat, L. Karmazin-Brelot, A. Neels, T. Buergi, R. G.
Finke, C. M. Hagen, Organometallics 2005, 24, 6104.
Angew. Chem. Int. Ed. 2001, 40, 3874; f) D. Blazina, S. B. Duckett,
P. J. Dyson, J. A. B. Lohman, Chem. Eur. J. 2003, 9, 1045; g) D. Blaz-
ina, S. B. Duckett, P. J. Dyson, J. A. B. Lohman, Dalton Trans. 2004,
2108.
[42] a) I. T. Horvꢀth, J. M. Millar, Chem. Rev. 1991, 91, 1339; b) A. Cusa-
nelli, U. Frey, D. T. Richens, A. E. Merbach, J. Am. Chem. Soc.
1996, 118, 5265.
[43] Y. Doi, K. Koshizuka, T. Keii, Inorg. Chem. 1982, 21, 2732.
[44] J. L. Graff, M. S. Wrighton, J. Am. Chem. Soc. 1980, 102, 2123 and
references therein.
[45] M. I. Bruce, M. L. Williams, Inorg. Synth. 1989, 26, 262.
[46] E. Tyrell, M. W. H. Tsang, G. A. Skinner, J. Fawcett, Tetrahedron
1996, 52, 9841.
[47] J. Jeener, B. H. Meier, P. Bachmann, R. R. Ernst, J. Chem. Phys.
1979, 71, 4546.
[35] C. M. Hagen, L. Vieille-Petit, G. Laurenczy, G. Sꢃss-Fink, R. G.
Finke, Organometallics 2005, 24, 1819.
[36] L. N. Lewis, J. Am. Chem. Soc. 1986, 108, 743.
[48] W. Rigby, R. Whyman, K. Wilding, J. Phys. E 1970, 3, 572.
[49] Z. Otwinowski, W. Minor, Methods Enzymol. 1997, 276, 307.
[50] A. J. M. Duisenberg, L. M. J. Kroon-Batenburg, A. M. M. Schreurs,
J. Appl. Crystallogr. 2003, 36, 220.
[37] J. G. De Vries, Dalton Trans. 2006, 421.
[38] a) J. A. Widegren, R. G. Finke, J. Mol. Catal. A 2003, 198, 317;
b) J. A. Widegren, M. A. Bennett, R. G. Finke, J. Am. Chem. Soc.
2003, 125, 10301.
[51] M. C. Burla, M. Camalli, B. Carrozzini, G. L. Cascarano, C. Giaco-
vazzo, G. Polidori, R. Spagna, J. Appl. Crystallogr. 2003, 36, 1103.
[52] M. C. Burla, M. Camalli, B. Carrozzini, G. L. Cascarano, C. Giaco-
vazzo, G. Polidori, R. Spagna, J. Appl. Crystallogr. 2005, 38, 381.
[53] SHELXS-97, Program for Crystal Structure Determination, G. M.
Sheldrick, University of Gçttingen, Gçttingen (Germany), 1997.
[54] L. J. Farrugia, J. Appl. Crystallogr. 1999, 32, 837.
[55] SADABS, Bruker Nonius scaling and absorption correction, v 2.10,
G. M. Sheldrick, Bruker AXS, Inc., Madison, Wisconsin, 2003.
[56] SHELXTL, Bruker Analytical X-ray Systems, G. M. Sheldrick,
Bruker AXS, Inc., Madison, Wisconsin, 2005.
[39] T. Ohta, H. Takaya, R. Noyori, Inorg. Chem. 1988, 27, 566.
[40] a) S. B. Duckett, C. L. Newell, R. Eisenberg, J. Am. Chem. Soc.
1994, 116, 10548; b) T. C. Eisenschmid, R. U. Kirss, P. P. Deutsch,
S. I. Hommeltoft, R. Eisenberg, J. Bargon, R. G. Lawler, A. L.
Balch, J. Am. Chem. Soc. 1987, 109, 8089; c) J. Natterer, J. Bargon,
Prog. Nucl. Magn. Reson. Spectrosc. 1997, 31, 293; d) S. B. Duckett,
C. J. Sleigh, Prog. Nucl. Magn. Reson. Spectrosc. 1999, 34, 71.
[41] a) R. Gobetto, L. Milone, F. Reineri, L. Salassa, A. Viale, E. Rosen-
berg, Organometallics 2002, 21, 1919; b) S. Aime, W. Dastrꢋ, R. Go-
betto, F. Reineri, A. Russo, A. Viale, Organometallics 2001, 20,
2924; c) S. Aime, W. Dastrꢋ, R. Gobetto, A. Russo, A. Viale, D.
Canet, J. Phys. Chem. A 1999, 103, 9702; d) S. Aime, R. Gobetto, D.
Canet, J. Am. Chem. Soc. 1998, 120, 6770; e) D. Blazina, S. B. Duck-
ett, P. J. Dyson, J. A. B. Lohman, Angew. Chem. 2001, 113, 3992;
[57] SHELXL-97, Program for Crystal Structure Refinement, G. M.
Sheldrick, University of Gçttingen, Gçttingen (Germany), 1997.
[58] A. G. J. Orpen, J. Chem. Soc. Dalton Trans. 1980, 2509.
Received: February 25, 2012
Published online: && &&, 0000
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
21
&
ÞÞ
These are not the final page numbers!

E. Nordlander et al.
Asymmetric Catalysis
V. Moberg, R. Duquesne, S. Contaldi,
O. Rçhrs, J. Nachtigall, L. Damoense,
A. T. Hutton, M. Green, M. Monari,
D. Santelia, M. Haukka,
Cluster catalysis: New clusters of the
type [H4Ru4(CO)10(m-1,2-P-P)],
precursors for hydrogenation of a-
unsaturated carboxylic acids (see
scheme). The cluster-based catalytic
systems give excellent conversion rates
and product selectivities, as well as
good enantioselectivities.
[H4Ru4(CO)
10A
[H4Ru4(CO) CHTUNGTRENNUNG
11A(P-P)] (P-P=chiral
E. Nordlander* ............... &&&& — &&&&
diphosphine) have been synthesised
and investigated as catalysts/catalyst
Efficient Cluster-Based Catalysts for
Asymmetric Hydrogenation of a-
Unsaturated Carboxylic Acids
&
22
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!