Water-soluble half-sandwich RuII-arene complexes: Synthesis, structure, electrochemistry, dft studies, and aqueous phase hydroformylation of 1-octene

Article abstract of DOI:10.1002/ejic.201300503

Water-soluble mononuclear ruthenium complexes containing monodentate P-donor ligands [1,3,5-triaza-7-phosphaadamantane (PTA) and P(OMe)₃], together with a mononuclear bidentate salicylaldimine ruthenium-arene complex, were synthesized from a water-soluble dimeric ruthenium(II)-arene precursor. These complexes were used as catalyst precursors in the aqueous biphasic hydroformylation of 1-octene. Hydroformylation reactions with these Ru-arene complexes showed that the complexes are active hydroformylation catalyst precursors in water. The catalyst precursors convert 1-octene into a range of products, which include aldehydes and alcohols, as well as isomerization and hydrogenation products. The less basic the ligand, the more it favoured the production of aldehydes. The relative basicity of the ligands P(OMe) ₃, PTA and the N,O-chelating ligand was elucidated by cyclic voltammetry and supported by DFT calculations. The catalysts show good recyclability without a significant decrease in 1-octene conversion. All of the catalysts could be reused three times. Water-soluble Ru^II-arene complexes were synthesized and characterized by NMR and FTIR spectroscopy, mass spectrometry, elemental analysis and single-crystal X-ray crystallography. The water-soluble Ru complexes catalyze the hydroformylation of 1-octene, which they convert into aldehydes, alcohols, iso-octenes and octane. The catalysts could be recycled four times with sustained conversions and minimal Ru loss.

Full text of DOI:10.1002/ejic.201300503

FULL PAPER
DOI:10.1002/ejic.201300503
Water-Soluble Half-Sandwich RuII–Arene Complexes:
Synthesis, Structure, Electrochemistry, DFT Studies, and
Aqueous Phase Hydroformylation of 1-Octene
Leah C. Matsinha,[a] Peter Malatji,[a] Alan T. Hutton,[a]
Gerhard A. Venter,[a,b] Selwyn F. Mapolie,[c] and
Gregory S. Smith*[a]
Keywords: N,O ligands / Ruthenium / Biphasic catalysis / Hydroformylation
Water-soluble mononuclear ruthenium complexes containing
monodentate P-donor ligands [1,3,5-triaza-7-phosphaadam-
antane (PTA) and P(OMe)3], together with a mononuclear bi-
dentate salicylaldimine ruthenium–arene complex, were syn-
thesized from a water-soluble dimeric ruthenium(II)–arene
precursor. These complexes were used as catalyst precursors
in the aqueous biphasic hydroformylation of 1-octene. Hy-
droformylation reactions with these Ru-arene complexes
showed that the complexes are active hydroformylation cata-
lyst precursors in water. The catalyst precursors convert 1-
octene into a range of products, which include aldehydes and
alcohols, as well as isomerization and hydrogenation prod-
ucts. The less basic the ligand, the more it favoured the pro-
duction of aldehydes. The relative basicity of the ligands
P(OMe)3, PTA and the N,O-chelating ligand was elucidated
by cyclic voltammetry and supported by DFT calculations.
The catalysts show good recyclability without a significant
decrease in 1-octene conversion. All of the catalysts could be
reused three times.
The immobilization of transition metal catalysts in water
has already been successfully applied industrially in the hy-
droformylation reaction, which is one of the oldest and
largest homogeneously catalyzed reactions, first discovered
by Roelen.[3] In this reaction, olefinic substrates add syngas
(H2 and CO) across a double bond in the presence of a
suitable catalyst, normally Rh or Co, to yield aldehydes as
the main products.[4] In 1984, the Ruhrchemie/Rhône Poul-
enc “Oxo” process was developed.[5] This process involves
the hydroformylation of propylene to butyraldehyde or 1-
butene to valeraldehyde by using water-soluble catalysts in
an aqueous biphasic system.[6] This is an example of the
first successful industrial application of aqueous biphasic
catalysis with a water-soluble rhodium(I) catalyst.[7]
Introduction
To provide “greener” methods to effect various catalytic
transformations, researchers have turned their attention to
the use of nontoxic solvents such as water. Aqueous-phase
catalysis combines the virtues of conventional homogen-
eous and heterogeneous systems. The immobilization of or-
ganometallic complexes in the aqueous phase means that a
water-soluble catalyst is “heterogenized” with respect to the
reaction substrates and products found in the organic
phase.[1] This allows easy catalyst–product separation and,
often, full catalyst recovery, which is not only economically
beneficial but also advantageous to the environment and
conforms to green chemistry principles. Water is cheap,
nonflammable, incombustible, nontoxic, odourless and is a
good solvent for many gases.[2] It also acts as a mobile sup-
port for water-soluble complexes and, therefore, immobi-
lizes the catalyst whilst the homogeneous nature of the reac-
tion is maintained.
The replacement of rhodium with less-expensive metals
is an alternative strategy towards the synthesis of new hy-
droformylation catalysts. Half-sandwich ruthenium(II)–ar-
ene complexes catalyze a number of organic reactions.[8–17]
Ruthenium complexes are well known for their ability to
hydrogenate unsaturated C–C bonds[7,18–20] and have also
found use in the synthesis of hydroformylation catalyst pre-
cursors, owing to the variable oxidation states of ruthe-
nium.[21–23] The applications of ruthenium complexes in hy-
droformylation under aqueous biphasic conditions have
found some interest and a few examples of these transfor-
mations have been reported in the literature.[18,24,25] The
first report of this kind, in which an olefin was quantita-
[a] Department of Chemistry, University of Cape Town,
Rondebosch 7701, Cape Town, South Africa
Fax: +27-21-6505195
E-mail: Gregory.Smith@uct.ac.za
Homepage: http://www.gregsmith-research.co.za
[b] Scientific Computing Research Unit and chemistry
Department, University of Cape Town,
Rondebosch 7701, Cape Town, South Africa
[c] Department of Chemistry and Polymer Science, Stellenbosch
University,
Matieland 7602, South Africa
tively converted into an aldehyde, made use of a RuIII
–
Eur. J. Inorg. Chem. 2013, 4318–4328
4318
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
EDTA (EDTA = ethylenediaminetetraacetic acid) complex 25 mgmL–1, respectively). Although 5 is highly soluble in
in an aqueous medium, and the catalyst was recycled sev- water as a chloride salt, this compound was unstable and
eral times without degradation of the metal complex.[18] difficult to isolate. Complex 5 was stabilized as the tet-
[RuCl2(TPPTS)] [TPPTS = P(m-C6H4SO3Na)3] catalyzes raphenylborate salt (which is insoluble in water) by a coun-
the conversion of 1-hexene into n-heptanal and n-heptanol terion exchange performed by metathesis with sodium tet-
in water.[24] The hydroformylation products were formed at raphenylborate (NaBPh4).
high pressure, and good selectivity towards n-heptanal oc-
In general, the aliphatic protons for the 2-hydroxyethoxy
1
curred at 25 °C. [RuCl2(DMSO)] (DMSO = dimethyl sulf- tether appear in the H NMR spectra as two multiplets at
oxide) catalyzed the hydroformylation of 1-hexene, cyclo- δ ≈ 4.38 and 4.03 ppm, which integrate as two protons each.
hexene, 2-methyl-2-pentene and 2,3-dimethyl-2-butene un- The appearance of these signals as multiplets can be attrib-
der aqueous biphasic conditions.[25] For all the substrates, uted to coupling of these protons to the OH proton and
linear aldehydes and isomers were formed.
the aromatic protons, which is evident in the COSY NMR
The increased interest in ruthenium–arene chemistry, spectra of the complexes. For 4 and 5, the 1H NMR spectra
particularly in the field of catalysis, motivated us to develop show the resonance of the imine proton at δ ≈ 7.64 ppm.
new water-soluble ruthenium–arene complexes and to in- The imine signals for L1 and L2 appear more downfield
vestigate their catalytic potential in the aqueous biphasic than those for 4 and 5. This shift and the absence of the OH
1
hydroformylation of 1-octene.
signal observed for the metal-free ligand in the H NMR
spectrum of 4 is evidence of coordination of the ligand to
the Ru centre. The spectra of the complexes also show two
sets of multiplets for the aliphatic protons on each CH2
group of the propyl chain. This splitting of the signals is
induced by the chiral ruthenium centre, which makes these
protons diastereotopic. The 31P{1H} NMR analysis for 2
and 3 reveal one singlet for the phosphorus nuclei. Infrared
spectral analysis of 4 and 5 shows an intense absorption
band for the imine functionality at ca. 1624 cm–1. ESI mass
spectrometry confirmed the integrity of the ruthenium com-
plexes. The ESI spectrum of 5 in the positive ion mode
shows a parent ion at m/z = 423 [M]+.
Results and Discussion
Synthesis and Characterization of the Water-Soluble,
Mononuclear Ruthenium(II)–Arene Catalyst Precursors
The basis for our studies is the known water-soluble
ruthenium(II)–arene precursor 1, which contains a 2-
hydroxyethoxy moiety tethered to the arene and was syn-
thesized by following a method described in the literature
(Scheme 1).[26] Ancillary ligands were chosen for their hy-
drophilicity [1,3,5-triaza-7-phosphaadamantane (PTA)] or
their ability to act as hard and strong donors such as imine
(σ donor) or phenoxide (both σ and π donor) ligands. The
mononuclear ruthenium(II) complexes containing P-donor
ligands were obtained by reaction of the dimeric precursor
Molecular Structures
The molecular structure of 4 was determined by single-
with either P(OMe)3[27] or PTA to give 2 and 3, respectively. crystal X-ray crystallography. Single crystals suitable for X-
The known N,O- and N,N-chelating ligands L1 and L2 were ray diffraction were obtained by slow diffusion of hexane
reacted with the ruthenium dimeric precursor to yield the into a solution of 4 in methanol at room temperature. The
mononuclear Ru complexes 4 and 5, respectively. Only 2, 3 molecular structure of 4 (Figure 1) confirms the expected
and 4 showed good solubility in water (755, 166 and “piano-stool” geometry around the RuII ion; the metal cen-
Scheme 1. Synthesis of 1–5.
Eur. J. Inorg. Chem. 2013, 4318–4328
4319
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
tre is coordinated to the arene moiety, a chloride ligand and parameters of 5 are comparable to those observed in other
a bidentate N,O-chelating ligand (Figure 1). The Ru1A– N,N-ruthenium–arene complexes.[29]
N1A bond length is 2.076(6) Å, and the Ru1A–O1A bond
Table 2. Crystallographic data and structure refinement parameters
for 4 and 5.
length is 2.055(4) Å. The Ru–Cl1A bond length is
2.421(2) Å, which is close to previously reported Ru–Cl
bond lengths of similar complexes.[28] The arene moiety is
planar, and the distances between the C atoms on the
4
5
Molecular formula
C18H22ClNO3Ru
C41H14BClN2O3Ru
732.02
monoclinic
173
planar aromatic ring and the Ru atom are 2.147(7)– Molecular mass [gmol–1] 436.57
Crystal system
Temp. of collection [K]
monoclinic
173
2.266(7) Å.
Space group
a [Å]
b [Å]
c [Å]
P21/c
C2/c
7.7787(4)
21.3660(11)
23.7653(13)
90
18.5053(12)
13.3593(9)
28.786(2)
90
α [°]
β [°]
93.362(2)
90
3943.0(4)
4
105.7750(10)
90
6847.3(8)
8
γ [°]
V [Å3]
Z
λ [Å]
0.71073
1.615
1.083
0.71073
1.420
0.71073
2992
σcalcd. [gcm–3]
μ(Mo-Kα) [mm–1]
F(000)
1944
θ min, max [°]
Index ranges (h, k, l)
1.7, 28.3
–5:10,
–27:28,
–30:31
25590
9759
1.5, 27.1
–13:23,
–17:14,
–36:36
18642
7573
Figure 1. Molecular structure of 4 (hydrogen atoms omitted for
clarity).
Reflections collected
Unique reflections
Reflections with I Ͼ 2σ(I) 6607
4353
7573/365
1.03
Complex 5 was also characterized by single-crystal X-ray
diffraction. Single crystals suitable for X-ray diffraction
Data/parameters refined
9759/453
1.03
Goodness of fit, S
were obtained by slow evaporation of 1:1 acetone/hexane Final R indices [I Ͼ 2σ(I)] 0.0639
0.0788
0.2349
–2.38. 1.14
Final wR2 (all data)
0.1988
solutions at room temperature. The molecular structure of
5 is shown in Figure 2.
Min, max e density [eÅ–3] –2.00, 5.26
The hydroxyethoxy benzene moiety is disordered over
two positions. The structure was modelled in two orienta-
tions (Figure 2). The coordination around the metal centre
is the expected piano-stool geometry, and the iminopyridyl
ligand is bidentate and chelates through its two nitrogen
atoms. The bond angles and lengths for 4 and 5 are summa-
rized in Table 1, and the crystallographic data and structure
refinement parameters are given in Table 2. The geometrical
Complex 5 is insoluble in water and, therefore, was not
used as a catalyst precursor in the aqueous biphasic hydro-
formylation of 1-octene. The water-soluble complexes 2, 3
and 4 were further evaluated by electrochemical measure-
ments and computational studies.
Electrochemistry of 2, 3 and 4
The difference in the σ-donating and π-accepting capaci-
ties of the ligands influences the oxidation potentials of the
RuII centres. These potentials give an indication of the basic
character of the ligands in the complexes, which in turn
have an influence on the chemoselectivity of the complexes
during hydroformylation reactions. To investigate such pos-
sible correlations in the present systems and to provide fur-
ther characterization of the complexes, we performed a pre-
liminary electrochemical study and found that the irrevers-
ible oxidation of RuII to RuIII in 2 occurs at a potential of
+0.84 V vs. the Ag/Ag+ reference electrode; this is very close
to the value (+0.82 V) previously obtained for the same
compound.[27] The voltammogram shows an additional
peak at more positive potential owing to phosphorus oxi-
dation (Figure 3). This was confirmed by performing the
electrochemistry experiment for the free ligand P(OMe)3,
for which the phosphorus oxidation peak was found at
+1.54 V. For 3, an irreversible oxidation peak for RuII/RuIII
Table 1. Selected bond lengths [Å] and angles [°] for 4 and 5.
Atoms
Interatomic distance [Å]
Compound 4
Ru1A–Cl1A
Ru1A–O1A
Ru1A–N1A
Compound 5
Ru(1)–Cl(1)
Ru(1)–N(1)
Ru(1)–N(2)
2.4214(17)
2.055(4)
2.076(6)
2.391(2)
2.099(5)
2.070(7)
Atoms
Angle (o)
Compound 4
Cl1A–Ru1A–O1A
Cl1A–Ru1A–N1A
O1A–Ru1A–N1A
Compound 5
85.94 (14)
86.01(16)
87.53(19)
Cl(1)–Ru(1)–N(1)
Cl(1)–Ru(1)–N(2)
N(1)–Ru(1)–N(2)
83.53(16)
76.20(2)
87.70(19)
Eur. J. Inorg. Chem. 2013, 4318–4328
4320
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Figure 2. Molecular structure of 5 in its two orientations (hydrogen atoms omitted for clarity).
was observed at +0.76 V (Figure 3). This is due to the one- ram may be ascribed to redox processes in the ligands. Mu-
electron oxidation of the Ru centre. The other small peaks rali and co-workers[31] reported that lower RuII/RuIII oxi-
observed in the voltammogram can be ascribed to electro- dation potentials can be attributed to the strong σ-donating
chemical processes in the PTA ligand.
or weak π-accepting properties of chelating ligands.
The RuII centre in 3 is oxidized more easily than that in
2, which indicates that the PTA ligand is a better electron
donor than the P(OMe)3 ligand. A lower RuII/RuIII poten-
tial can be attributed to a stronger σ-donating ability of the
ligand.[30] This implies that there is a higher electron density
on the metal centre in 3 than in 2. The electrochemical pro-
cesses in 4, which contains the N,O-chelating ligand, were
also studied (Figure 3). Here, the oxidation of RuII to RuIII
is again irreversible but occurs at a potential of +0.60 V,
which is significantly lower than the values recorded for 2
Table 3. Summary of Eox values (V vs. Ag/Ag+) for the RuII/RuIII
couple of 2–4.
Eox RuII/RuIII [V] Comment
2
3
4
+0.84
+0.76
+0.60
P(OMe)3: the least σ-donating ability
PTA: moderate σ-donating ability
N,O-chelating ligand: good σ-donat-
ing ability
The metal centre in 4 was easily oxidized compared to
and 3 (Table 3). Other peaks observed in the voltammog- the Ru centres in 2 and 3. This is evidence that the N,O-
Eur. J. Inorg. Chem. 2013, 4318–4328
4321
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Computational Studies on 2, 3 and 4
The structures of the complexes were optimized by den-
sity functional theory (DFT) calculations at the BP86/TZ2P
level (see Computational Details)[32] to quantify the σ-do-
nating and π-accepting ability of the ligands. In addition,
calculations of the ionization energies (IEs) of the neutral
species were used as an indirect indicator of their oxidation
potential, and this information was used to verify the per-
formance of the chosen level of theory. The optimized
structures are shown in Figure 4. Several different confor-
mations were calculated, and the lowest-energy structures
are those that are stabilized by an intramolecular hydrogen
bond between the chloride ligand and the hydroxyl hydro-
gen atom of the hydroxyethoxy benzene ligand. The crystal
structure of 4 contains the conformer stabilized by intermo-
lecular hydrogen bonding to an adjacent ligand, but we
chose these lowest-energy conformers for our comparative
analyses.
The IE can be calculated as either an adiabatic or non-
adiabatic (vertical) process, depending on the pathway fol-
lowed. The latter includes structural relaxation following re-
moval of the electron, whereas the former excludes this
component. The most appropriate choice should be guided
by the timescale of measurement, in accordance with the
Franck–Condon principle.[33] Here, a comparison with oxi-
dation potential is attempted and as this experimental mea-
surement happens over a relatively long timescale, it is ex-
pected to be better approximated by the adiabatic ioniza-
tion energy (AIE). This calculation is accomplished by tak-
ing the energy difference between the relaxed cation and the
neutral complex. For comparison, the vertical ionization
energy (VIE) was also calculated by using both the negative
of the energy (–εHOMO) of the Kohn–Sham highest occu-
pied molecular orbital (KS–HOMO)[34] and the difference
in energy between the cation and neutral complex (both
calculated by using the fixed structure of the neutral sys-
tem).[33] These results and their correlation with the experi-
mental oxidation potentials are shown in Figure 5. The
AIEs for 2–4 are 6.79, 6.64 and 6.36 eV, respectively. Al-
though both VIE and AIE duplicate the measured trend
Figure 3. Voltammograms of 2, 3 and 4 recorded at 100 mVs–1.
chelating ligand is a better σ donor than both PTA and satisfactorily, the best fit (R2 = 0.999) is reproduced, as ex-
P(OMe)3. This ligand displays high basicity, and the effect pected, by using the latter. The comparison with experimen-
of this on the chemoselectivity of the catalyst was investi- tal values is very good and validates the use of BP86/TZ2P
gated in the hydroformylation of 1-octene (see below).
for further analyses.
Figure 4. BP86/TZ2P-optimized structures of 2–4.
Eur. J. Inorg. Chem. 2013, 4318–4328
4322
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Figure 5. The correlation between BP86/TZ2P calculated ioniza-
tion energy (IE) and the oxidation potential. Triangles show values
derived from the negative of the HOMO, diamonds show the verti-
cal IE, and hexagons show the adiabatic IE.
Natural orbitals for chemical valence (NOCV) were used
to quantify the σ-donor and π-accepting ability of the P-
donor and N,O-chelating ligands.[35] The deformation den-
sity Δρ that occurs upon orbital relaxation of the isolated
DFT fragment wavefunctions as they form the complex is
decomposed into complementary pairs of NOCVs:
These pairs define the channels of charge transfer be-
tween fragments. In this representation, the value νk corre-
sponds to the fraction of electron density transferred from
orbital φ–k to orbital φk and can be considered a direct
measure of the strength of the transfer. Subsequent visuali-
zation of these densities allows identification of electron-
transfer channels and can be used to distinguish between
ligandǞmetal donation and metalǞligand back-donation.
Figure 6 shows the leading deformation density contri-
butions and their associated charge transfer. The P(OMe)3
ligand donates 0.99 into a metal orbital with mostly d char-
acter, but there are also contributions from the 3p orbitals
of the chloride ligands (Δρ1 in 2, Figure 6). The back-do-
nation (Δρ2 and Δρ3 in 2, Figure 6) occurs from metal 4d
orbitals into empty orbitals of the P(OMe)3 ligand, which
have predominantly p character localized on the phospho-
rus atom. The contributions into these two orthogonal 3px
and 3py orbitals are 0.33 and 0.31 = 0.64. In 3, the values
for the PTA ligand are 1.04 for σ donation and 0.28 + 0.27
= 0.55 for the π back-donation. The orbital compositions
are similar to those of 2. The ligand in 4, which has two
donor atoms coordinated bidentately, contributes two σ do-
nations along the chelating bonds. The corresponding frac-
tions of charge transferred are 1.00 and 0.56. The contri-
butions Δρ3 and Δρ4 in 4 are 0.32 + 0.25 = 0.57. These
consist of transfer from orbitals with some Ru d character
but are mostly contributions from the Cl px and py orbitals.
The accepting orbitals have p character on the nitrogen and
oxygen ligand atoms.
Figure 6. Contributions from the complementary NOCV pairs to
the deformation density Δρ. The isosurfaces of |Δρ| = 0.002 au are
shown.
Eur. J. Inorg. Chem. 2013, 4318–4328
4323
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
The P(OMe)3 ligand is, thus, the weakest σ donor and of some of the aldehyde products and olefin substrate,
accommodates the least π back-donation, and both the ex- respectively. The isomerization of 1-octene increases and
perimental oxidation potentials and the computed ioniza- better regioselectivity towards the branched aldehydes is
tion energies agree that the metal centre here is least suscep- observed in water. With 3 as the catalyst precursor in the
tible to loss of an electron. This is followed by the stronger biphasic reaction (Table 4), the conversion of 1-octene is
σ donor PTA ligand, which is also a weaker π back-dona- 83%, and the most abundant products formed are internal
tor. Lastly, the N,O-chelating ligand L1 is a much stronger olefins from isomerization of 1-octene. Total aldehydes
σ donor (as expected for a bidentate ligand) and is, thus, formed are only 6% of the total products (Table 4, Entry
best at stabilization of a high oxidation number of the 2). The catalyst favours isomerization of 1-octene and is
metal.
less effective as a hydroformylation catalyst in water. It also
favours hydrogenation of aldehyde products into alcohols,
which indicates the superior hydrogen transfer ability of this
catalyst in water.
Hydroformylation of 1-Octene in Water
With 4 as the catalyst precursor, the conversion of 1-oc-
tene is 88%, which indicates that there is good interaction
between the catalyst and 1-octene in the aqueous phase
(Table 4, Entry 3).
The catalytic activity of the ruthenium complexes 2–4 in
the hydroformylation of 1-octene in water was explored.
This was conducted according to the conditions used for
ruthenium-catalyzed hydroformylation reactions of α-ole-
fins that have been reported in the literature.[21–25,31] The
use of ruthenium–arenes for the hydroformylation reaction
is unprecedented, although a half-sandwich ruthenium(II)–
cyclopentadienyl complex has been used in a similar
study.[23c] The proposed mechanism, which we presume also
occurs in our reactions, involves a ruthenium hydride inter-
mediate followed by the conventional steps for the hydro-
formylation reaction.[23c] Before catalytic testing was at-
tempted, a blank control experiment was conducted for this
reaction. No product was observed in the absence of the
ruthenium catalyst precursor.
However, only 21% aldehydes are formed in the biphasic
reaction. Therefore, catalyst 3 displays better chemoselectiv-
ity towards aldehydes in toluene. Despite this, the regiose-
lectivity of the catalyst towards linear aldehydes is good in
the aqueous biphasic reaction, with 77% nonanal being
formed. Phosphorus ligands (such as those in 2 and 3) are
often involved in hydroformylation catalytic systems, and
their steric and electronic characteristics have been reported
to play a role in the activity of catalysts.[36] Pruett and
Smith reported that these steric and electronic effects in-
fluence the distribution of products in hydroformylation re-
actions.[37] The phosphorus ligands in 2 and 3 induce dif-
ferent activity and selectivity on the catalyst precursors, as
evidenced by the different product distributions. It should
be noted that these reactions were carried out at different
temperatures (optimum conditions for each catalyst precur-
sor): 100 and 125 °C for 2 and 3, respectively.
Aqueous Biphasic Hydroformylation with 2, 3 and 4
The water-soluble Ru-arene catalyst precursors 2, 3 and
4 were dissolved in distilled water (5 mL) and were used in
the aqueous biphasic hydroformylation of 1-octene. Toluene
was used as the organic solvent (5 mL), and the reactions
However, the distribution of the products can give an
were performed for 21 h with a catalyst loading of indication of the different behaviour that the catalyst pre-
1.93ϫ10–5 mol. The reactions were performed under the cursors have because of the different P-donor ligands in
optimum reaction conditions obtained for the three catalyst these complexes. In particular, the ligand basicity has a role
precursors. The results for the aqueous biphasic hydrofor- in determining the product distribution during hydrofor-
mylation of 1-octene with 2 show that the catalyst interacts mylation, that is, the less basic the ligand, the less the donor
with the substrate even though 1-octene has poor water- ability and, hence, the faster hydroformylation rates. This,
solubility (Table 4). The conversion of 1-octene to products therefore, affects the chemoselectivity of the catalyst.[38,30]
is excellent (96%), and there is moderate chemoselectivity From the electrochemistry experiments, P(OMe)3 was found
towards aldehydes of 47% (Table 4, Entry 1). Complex 2 to have lower σ-donor ability than PTA (in 3). Therefore,
favours the production of branched aldehydes. Alcohols are P(OMe)3 is less basic, and the metal catalyst is expected to
formed together with octane, which indicates hydrogenation have lower reducing ability. This should favour the pro-
Table 4. Hydroformylation of 1-octene with 2, 3 and 4 as catalysts precursors.
En-
try
Complex Conv. [%] Aldehydes [%]
Isooctenes [%] Octane [%] Alcohols [%]
n/iso
Total aldehydes n-Aldehydes iso-Aldehydes
1[a]
2[b]
3[c]
2
3
4
96
83
88
47.06
6.08
21.56
29.22
79.96
77.68
70.78
13.22
22.32
34.45
83.74
74.92
16.83
–
3.52
1.52
10.19
–
0.41
6.05
3.48
[a] 50 bar, 125 °C, 1:1 CO/H2, 21 h. [b] 50 bar, 100 °C, 1:1 CO/H2. [c] 50 bar, 100 °C, 1:1 CO/H2; catalyst loading 1.93ϫ10–5 mol, reactions
were performed in 10 mL 1:1 H2O/toluene. GC conversions were obtained by using n-decane as internal standard, average error estimates:
1.46, 1.02 and 0.91% for entries 1, 2 and 3, respectively.
Eur. J. Inorg. Chem. 2013, 4318–4328
4324
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
duction of more aldehyde products. This is in agreement
with the results obtained: more aldehyde products were ob-
tained with 2 as the catalyst precursor.
Recyclability Studies
The catalyst precursors could be reused up to three times.
The hydroformylation reactions were performed under the
optimum conditions established for the three catalysts. Re-
cyclability studies were performed by decanting the organic
layer obtained from the first catalytic run, followed by the
addition of a fresh sample of 1-octene dissolved in toluene
onto the aqueous layer. This was repeated three times by
using the same catalyst-containing aqueous phase, and the
results are shown in Figure 7.
Figure 7. Aqueous biphasic hydroformylation of 1-octene with 2, 3
and 4 in recyclability studies; 1:1 CO/H2, 50 bar, 21 h, 100 (3 and
4) or 125 °C (2), catalyst loading 1.93 x 10–5 mol. Average error
estimates: 1.01 (2), 0.97 (3) and 0.87% (4).
All the catalyst systems displayed good activity and the
catalysts can be used up to four times without a significant
decrease in the conversion of 1-octene. An increase in 1-
octene conversion is observed in the second run when 4 is
employed. This could be because the active species formed
after the first catalytic run becomes more active in the sec-
ond cycle and this subsequently resulted in increased ac-
tivity in the second cycle. In the third cycle, the conversion
drops to about 60%. For 2 and 3, the conversion drops as
the number of runs increases. The decrease in the conver-
sion of 1-octene with each successive run could be caused
by loss of catalyst activity. This could be due to the decom-
position of the catalysts with each successive run. Another
possible reason for the drop in the conversion is loss of the
catalyst owing to leaching into the organic phase.
The regioselectivity of the catalysts was investigated in
all the four catalytic runs. Complex 2 has good selectivity
towards branched aldehydes (Figure 8). As discussed ear-
lier, this could be due to the high isomerization activity of
the catalyst induced by P(OMe)3 before hydroformylation
occurs. The n/iso ratios are very low (between 0.31 and 1.29)
and this complex would be a good catalyst to employ if
branched aldehydes are the products of interest.
Figure 8. Aqueous biphasic hydroformylation of 1-octene with 2, 3
and 4 in recyclability studies; 1:1 CO/H2, 50 bar, 21 h, 100 (3 and
4) or 125 °C (2), catalyst loading 1.9310–5 mol.
Eur. J. Inorg. Chem. 2013, 4318–4328
4325
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
A similar study was performed with 3, and the catalyst the aqueous phase with each successive catalytic cycle in
precursor shows good regioselectivity towards nonanal; less the recyclability studies. This could be one of the reasons
than 25% branched aldehydes were produced in all four for the drop in the conversions of 1-octene in the recycl-
runs (Figure 8). The n/iso ratios for all four catalytic runs ability studies. The amount of metal catalyst lost with 2 as
were between 2.95 and 3.99. There is a slight drop in the the catalyst precursor is 16% after the first catalytic run.
nonanal selectivity of the catalyst with each successive run. Further metal catalyst (5%) was lost in the second run. De-
Complex 4 also displays good selectivity towards normal spite this, the presence of the active species in the aqueous
aldehyde formation under the conditions employed (Fig- phase is evidenced by the activity of the catalyst for two
ure 8). The production of nonanal dropped from the first more catalytic runs with reasonable 1-octene conversions.
run to the third run. The n/iso ratios are moderate and the
The metal loss when 3 was used for the reaction was only
highest ratio of 3.48 was obtained in the first run. The ra- 4% after the second catalytic run. Very little metal was lost
tios drop in the second and third run to 1.63 and 0.73, when this complex was employed. When 4 was used in the
respectively. The turnover frequency of each catalyst reaction, the aqueous phase lost 80% metal in the first cata-
dropped with each subsequent recycle as shown in Figure 9. lytic run. This huge loss could be due to the solubility of
the metal complex in toluene at elevated temperature and
pressure. Despite this, the catalyst shows activity until the
fourth run, which shows that the active species is still pres-
ent in the aqueous phase.
Conclusions
All three water-soluble complexes are active under aque-
ous biphasic hydroformylation conditions. Catalyst precur-
sors 2 and 3 display good regioselectivity towards linear
aldehydes in all runs. Complex 2 has the lowest electron
density at the metal centre owing to the poor σ-donor abil-
ity of the P(OMe)3 ligand; hence, it favours the hydrofor-
mylation reaction, as evidenced by the good chemoselectiv-
ity of the catalyst towards aldehydes both in the organic
and aqueous biphasic experiments. Catalyst 2 gave a good
conversion of 1-octene (96%) in aqueous medium. All cata-
Figure 9. Activity of the catalysts in the aqueous biphasic hydro-
formylation of 1-octene with 2, 3 and 4 in recyclability studies; 1:1
lysts were easy to recover owing to the multiphase nature
of the systems employed and could be reused up to three
times without significant drop in 1-octene conversions. The
facile product separation and recyclability of these homo-
geneous catalysts is a strong drive towards green chemistry
and the application of these compounds as catalysts under
these conditions contributes towards cleaner technology.
CO:H2, 50 bar, 21 h, 100 (3 and 4) or 125 °C (2), catalyst loading
1.93 x 10–5 mol. Average error estimates: 1.01 (2), 0.97 (3) and
0.87% (4).
Of the three catalysts evaluated, catalyst 4 shows the
highest initial activity with a TOF of 242 h–1 compared to
187 and 107 h–1 for 2 and 3, respectively. All three catalysts
show an overall decrease in activity after reuse for a total
of four times. For 2 and 3, we observe a gradual decrease
in activity after each run (Figure 9). Catalyst 4 on the other
hand shows a more dramatic decrease in activity after each
reuse. After the fourth run, the activity is only 24 h–1. This
trend is probably related to the relative stability of the cata-
lysts in solution. It should be noted that catalyst 4 appears
to be the least stable, as indicated by the high degree of
leaching observed for this catalyst relative to the other two
catalysts (2 and 3). The latter two catalysts contain phos-
Experimental Section
General Procedures: All commercially available compounds were
obtained from Sigma–Aldrich and were used without further puri-
fication. Hydrated ruthenium trichloride was obtained from John-
son Matthey/Anglo American Platinum Limited. The dimeric ru-
thenium(II) precursor 1,[26] 2,[27] iminopyridyl ligand L1[39] and sa-
licylaldimine ligand L2[40] were prepared according to previously
reported methods. 1H, 13C and 31P NMR spectra were recorded
phane ligands and show only moderate leaching after each with either a Varian XR300 (1H 300.08 MHz, 13C 75.46 MHz, 31P
121.47 MHz) or
a
Bruker Biospin (1H 400.22 MHz, 13C
run. It would appear that the phosphane ligands have a
stabilising effect on the catalysts. Therefore, metal loss was
studied for each run.
100.60 MHz, 31P 162.00 MHz) spectrometer at ambient tempera-
ture. Analyses for C, H and N were performed with a Thermo
Flash 1112 Series CHNS-O analyzer. Infrared spectra were re-
corded with a Perkin–Elmer Spectrum 100 FTIR spectrometer with
samples in dichloromethane solutions in NaCl solution cells. Mass
spectrometry was performed at the University of Stellenbosch with
a Waters API Quattro Micro Triple Quadrupole electron spray ion-
ization mass spectrometer. Data were recorded in the positive
Leaching Studies
Inductively coupled plasma mass spectrometry (ICP-
MS) was used to analyze the aqueous phase to establish if
there was any loss of the metal catalyst with each successive
run. The data obtained showed loss of metal catalyst from mode.
Eur. J. Inorg. Chem. 2013, 4318–4328
4326
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Cyclic voltammetry was performed at ambient temperature by 49.22, H 5.01, N 3.18. MS (ESI): m/z = 402 [M – Cl]+. S20 °C
=
using a Bioanalytical Systems Inc. BAS 100 W electrochemical ana-
lyzer with a one-compartment three-electrode cell comprising a
platinum disk working electrode, a platinum wire auxiliary elec-
trode and a Ag/Ag+ reference electrode (0.01 m AgNO3 and 0.1 m
[nBu4N][ClO4] in anhydrous acetonitrile). The reported E values
(see text and Figure 3) are referenced to this electrode. Measure-
ments were made with anhydrous acetonitrile solutions with ca.
2 mm sample and 0.1 m [n-Bu4N][ClO4] as background electrolyte.
A scan rate of 100 mVs–1 was used throughout; scans started from
the most negative potential and were initially scanned anodically.
Under these conditions, the half-wave potential of the ferrocene/
ferrocenium couple, which was used as a reference, never varied by
more than 0.03 V from a mean value of E1/2 = +0.18 V vs. the Ag/
Ag+ reference electrode, and had a peak separation of ΔEp = 59–
66 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.
25 mgmL–1 in water.
Ruthenium Complex 5: Ruthenium dimer 1 (0.150 g, 0.242 mmol)
was suspended in methanol (25 mL), and this solution was stirred
at room temperature. Ligand L2 (0.072 g, 0.484 mmol) was dis-
solved in methanol (5 mL), and this solution was added dropwise
to the dimer solution. The mixture was stirred for 1 h at room
temperature. The solution was filtered, and the volume of filtrate
was reduced to ca. 5 mL. Sodium tetraphenylborate (0.210 g,
0.484 mmol) was then added, and the reaction was stirred at 0 °C
for 3 h. The resulting precipitate was collected by using a Hirsch
funnel and dried under vacuum to yield a yellow-brown powder,
yield 0.188 g (51%), m.p. 156–158 °C. 1H NMR (300 MHz, [D6]-
3
DMSO): δ = 9.46 (d, J = 5.2 Hz, 1 H), 8.35 (s, 1 H), 8.14 (m, 1
3
H), 8.04 (d, 3J = 6.6 Hz, 1 H), 7.71 (t, J = 5.1 Hz, 1 H), 7.39–7.32
(m, 8 H), 6.92 (t, 3J = 7.4 Hz, 8 H), 6.78 (t, 3J = 7.2 Hz, 4 H), 6.33
2
(m, 2 H), 5.81–5.33 (m, 2 H), 5.39 (m, 1 H), 4.86 (t, J = 5.4 Hz,
2
1 H), 4.58 and 4.29 (m, 2 H), 4.12 (t, J = 5.8 Hz, 2 H), 3.88 (m,
2 H), 2.09 and 1.96 (m, 2 H), 0.98 (t, 2J = 7.4 Hz, 3 H) ppm.
13C{H} NMR (75 MHz, [D6]DMSO): δ = 167.2, 139.4, 129.1,
127.9, 126.5, 123.9, 125.1, 121.7, 135.7, 93.6, 122.1, 86.8, 74.4, 65.9,
Reaction products from the catalytic studies were analyzed with a
Perkin–Elmer Clarus 580 GC instrument. Inductively coupled
plasma mass spectrometry (ICP-MS) was performed with a Perkin–
Elmer Elan600 quadrupole ICP-MS with a Cetax LSX-200 UV
laser module.
71.7, 67.3, 59.3, 22.7, 12.0 ppm. IR (CH Cl ): ν = 1624 (s, imine,
˜
2
2
C=N) cm–1. C17H22BClN2O2Ru (433.71): calcd. C 66.35, H 5.66,
N 3.78; found C 66.44, H 5.44, N 3.39. MS (ESI): m/z = 423 [M –
BPh4]+. S20 °C: insoluble in water.
Ruthenium Complex 3: To a suspension of ruthenium dimer 1
(0.300 g, 0.484 mmol) in a 1:1 mixture of methanol and dichloro-
methane (18 mL) was added 1,3,5-triaza-7-phosphaadamantane
(PTA, 0.196 g, 1.25 mmol). The mixture was stirred for 30 min at
50 °C and then filtered. The solvent was removed from the filtrate
to yield an orange solid, which was recrystallized from a hot solu-
tion of 1:1 CH2Cl2/MeOH to afford 3 as a red-brown solid, yield
X-ray Crystallography: Single-crystal X-ray diffraction data were
collected with a Bruker KAPPA APEX II DUO diffractometer
with graphite-monochromated Mo-Kα radiation (λ = 0.71073 Å).
Data collection was performed at 173(2) K. The temperature was
controlled by an Oxford Cryostream cooling system (Oxford Cryo-
stat). Cell refinement and data reduction were performed by using
the program SAINT.[41] The data were scaled, and absorption cor-
rection was performed by using SADABS.[42] The structure was
solved by direct methods by using SHELXS-97[42] and refined by
full-matrix least-squares methods based on F2 by using SHELXL-
97[42] and the graphics interface program X-Seed.[43,44] The pro-
grams X-Seed and POV-Ray were both used to prepare molecular
graphic images.
1
0.201 g (89%); m.p. 140 °C (decomp.). H NMR (300 MHz, [D6]-
3
3
DMSO): δ = 5.92 (m, 2 H), 5.43 (d, J = 6.2 Hz, 2 H), 4.94 (t, J
= 5.2 Hz, 1 H), 4.73 (br s, 1 H, OH), 4.12 and 4.33 (m, 12 H), 4.14
(t, 2J = 5.1 Hz, 2 H), 3.99 (m, 2 H) ppm. 13C NMR (75 MHz, [D6]-
DMSO): δ = 140.2, 130.0, 120.8, 114.9, 72.8 (d, JP,C = 6.5 Hz), 72.7
(d, JP,C = 7.1 Hz), 69.9 (d, JP,C = 6.9 Hz), 59.6 (d, JP,C = 17.1 Hz),
52.3 (d, JP,C = 19.2 Hz), 49.1 (d, JP,C = 17.5 Hz), 31.4, 22.4 ppm.
31P NMR (162 MHz, [D6]DMSO): δ = –32.0 ppm. C14H22Cl2N3O2-
PRu (467.30): calcd. C 35.96, H 4.71, N 8.99; found C 35.61, H
4.73, N 8.95. MS (ESI): m/z = 432 [M – Cl]+. S20 °C = 166 mgmL–1
in water.
CCDC-906859 (for 4) and -906861 (for 5) contain the supplemen-
tary crystallographic data for this paper. These data can be ob-
tained free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
Ruthenium Complex 4: To a stirred solution of L1 (0.104 g,
0.635 mmol) in ethanol (50 mL), triethylamine (0.093 mL, Computational Details: The geometries of all complexes were opti-
0.668 mmol) was added dropwise. The yellow solution was stirred
at room temperature for 30 min. Ruthenium dimer 1 (0.197 g,
0.318 mmol) was added to ethanol (10 mL) and this was added
dropwise to the ligand solution. The mixture was stirred for 6 h at
room temperature. The reaction mixture was filtered, and solvent
was removed under reduced pressure. A minimum amount of tolu-
ene was added to dissolve the residue, and this solution placed in
a fridge for 2 d. A crystalline solid formed and was filtered and
dried under vacuum to afford 4 as a brown crystalline solid, yield
0.204 g (71%); m.p. 133–136 °C. 1H NMR (300 MHz, [D6]DMSO):
δ = 7.64 (s, 1 H), 7.18–7.13 (m, 1 H), 6.92–6.87 (m, 1 H), 6.46–6.42
(m, 1 H), 5.94 (t, 3J = 5.6 Hz, 2 H), 5.43 (d, 3J = 5.0 Hz, 2 H),
mized by gradient-corrected DFT calculations by using Becke’s ex-
change functional with Perdew’s correlation functional (BP86)[32]
in conjunction with a Slater-type basis set of triple-zeta quality,
augmented with two sets of polarization functions (TZ2P). An ad-
ditional set of s, p, d, f and g auxiliary basis functions was used for
representation of the molecular density during the calculation of
the Coulomb potential at each self-consistent field (SCF) cycle.
Scalar relativistic effects were incorporated by using the zero-order
regular approximation (ZORA).[45] Vertical ionization energies
were calculated at the BP86/TZ2P level as the difference in energy
between the cation (+1) and the neutral complex at the geometry
of the latter. Adiabatic ionization potentials were determined in a
3
2
2
5.30 (t, J = 6.7 Hz, 1 H), 4.93 (t, J = 5.5 Hz, 1 H), 4.84 (t, J = similar way, but the charged complexes were allowed to relax to
2
5.2 Hz, 2 H), 4.42 (d, J = 5.7 Hz, 2 H), 4.12 and 3.93 (m, 2 H),
2.00 and 1.87 (m, 2 H), 1.34 (t, 2J = 7.3 Hz, 3 H) ppm. 13C{H}
NMR (75 MHz, [D6]DMSO): δ = 167.6, 163.9, 158.6, 135.5, 133.7,
their lowest energy. The former contain only orbital relaxation ef-
fects, whereas the latter incorporate geometrical relaxation as well.
These approaches are often labelled ΔSCF methods.[33] No correc-
129.2, 121.0, 120.2, 118.9, 114.4, 113.0, 69.3, 66.0, 59.4, 24.6, 11.1 tions were attempted for solvent effects. The bonding between se-
ppm. IR (CH2Cl2):
C18H22ClNO3Ru (436.90): calcd. C 49.47, H 5.04, N 3.21; found C
ν
=
1627 (s, imine, C=N) cm–1. lected fragments was decomposed into contributions from the natu-
˜
ral orbitals for chemical valence (NOCV).[35] This procedure de-
Eur. J. Inorg. Chem. 2013, 4318–4328
4327 © 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
[17] P. Singh, A. K. Singh, Organometallics 2010, 29, 6433–6442.
[18] D. U. Parmar, S. D. Bhatt, H. C. Bajaj, R. V. Jasra, J. Mol. Ca-
tal. A 2003, 202, 9–11.
[19] T. Bartik, B. Bartik, B. E. Hanson, T. Glass, W. Bebout, Inorg.
Chem. 1992, 31, 2667–2669.
[20] P. J. Dyson, D. J. Ellis, D. G. Parker, T. Welton, Chem. Com-
mun. 1999, 25–27.
[21] D. U. Parmar, H. C. Bajaj, S. D. Bhatt, R. V. Jasra, Bull. Chem.
Soc. India 2003, 2, 34–39.
[22] G. Süss-Fink, G. F. Schmidt, J. Mol. Catal. 1987, 47, 361–364.
[23] a) R. A. Sanchez-Delgado, J. S. Bradley, G. Wilkinson, J.
Chem. Soc., Dalton Trans. 1976, 399–402; b) E. Cesarotti, A.
Fusi, R. Ugo, G. M. Zanderighi, J. Mol. Catal. 1978, 4, 205–
216; c) K. Takahashi, M. Yamashita, Y. Tanaka, K. Nozaki,
Angew. Chem. 2012, 124, 4459–4463.
composes the deformation density (Δρ) that accompanies the or-
bital interaction between properly constructed fragments into com-
plementary NOCV pairs. Visualization of these densities allows
identification of channels of electron transfer. For 2 and 3, the frag-
ments were the neutral ligands P(OMe)3 or PTA and
Ru(C6H5OC2H4OH)Cl2. For 4, the chosen fragments were the
charged species [L1]+ and [Ru(C6H5OC2H4OH)Cl]– All DFT calcu-
lations, including the NOCV analyses, were performed by using the
ADF 2012.01 program package.[46]
Catalytic Studies: The reactions were performed in a 90 mL stain-
less steel pipe reactor. The reactor was charged with either toluene
(5 mL) or 1:1 toluene/H2O (10 mL), 1-octene (6.40 mg,
5.74ϫ10–5 mol),
internal
standard
n-decane
(0.152 mg,
or 4,
1.00ϫ10–6 mol) and the catalyst precursors (2,
3
[24] M. M. T. Khan, S. B. Halligudi, S. H. R. Abdi, J. Mol. Catal.
1988, 48, 313–317.
1.93ϫ10–5 mol). The reactor was flushed with nitrogen three times
and then twice with syngas (1:1 CO/H2). The reactor was then pres-
surized and heated to the desired pressure and temperature. All
reactions were performed for 21 h, and samples were collected at
the beginning and end of each reaction.
[25] M. Reyes, E. J. Parra, Y. Fonseca, B. Fontal, I. Suarez, F.
Bellandi, R. R. Contreras, I. Romero, L. Prado, P. Cancines,
Av. Quim. 2008, 3, 15–20.
[26] J. Soleimannejad, C. White, Organometallics 2005, 24, 2538–
2541.
[27] B. Lastra-Barreira, J. Diez, P. Crochet, Green Chem. 2009, 11,
1681–1686.
[28] B. Cetinkaya, S. Demir, I. Özdemir, L. Toupet, D. Sémeril, C.
Bruneau, P. H. Dixneuf, New J. Chem. 2001, 25, 519–523.
[29] P. Govender, A. K. Renfrew, C. M. Clavel, P. J. Dyson, B. Ther-
rien, G. S. Smith, Dalton Trans. 2011, 40, 1158–1167.
[30] E. R. Tucci, Ind. Eng. Chem. Prod. Res. Dev. 1970, 9, 516–518.
[31] V. Rajendiran, M. Murali, E. Suresh, M. Palaniandavar, V. S.
Periasamy, M. A. Akbarsha, Dalton Trans. 2008, 2157–2170.
[32] a) A. D. Becke, Phys. Rev. A 1988, 38, 3098–3100; b) J. P. Per-
dew, Phys. Rev. B 1986, 34, 7406–7406; c) J. P. Perdew, Phys.
Rev. B 1986, 33, 8822–8824.
Acknowledgments
The authors gratefully thank the University of Cape Town (UCT),
the National Research Foundation (NRF), South Africa and the
Department of Science and Technology of South Africa (NRF-
DST Centre of Excellence in Catalysis - c*change) for financial
support. A generous donation of ruthenium trichloride from Anglo
American Platinum Limited is gratefully acknowledged.
[33] C. J. Cramer, in: Essentials of Computational Chemistry: Theo-
ries and Models, Wiley, Chichester, UK, 2004.
[1] B. Cornils, W. A. Hermann, in: Aqueous Phase Organometallic
Catalysis, Wiley-VCH, New York, 2004, p. 4–5.
[2] F. Joo, in: Organometallic Catalysis, Kluwer, Dordrecht, The
Netherlands, 2001, p. 8–12.
[3] G. W. Parshall, in: Homogeneous Catalysis, Wiley, New York,
1980, p. 12–24.
[4] G. N. Schauzer, in: Transition Metals in Homogenous Catalysis,
Marcel Dekker, New York, 1971, p. 10–24.
[5] N. Pinault, D. W. Bruce, Coord. Chem. Rev. 2003, 241, 1–25.
[6] M. Beller, J. G. E. Krauter, J. Mol. Catal. A 1999, 143, 31–39.
[7] P. S. Hallman, B. R. McGarvey, G. Wilkinson, J. Chem. Soc.,
Dalton Trans. 1968, 3143–3145.
[34] C.-G. Zhan, J. A. Nichols, D. A. Dixon, J. Phys. Chem. A 2003,
107, 4184–4195.
[35] a) M. P. Mitoraj, A. Michalak, T. Ziegler, J. Chem. Theory
Comput. 2009, 5, 962–975; b) A. Michalak, M. Mitoraj, T. Zi-
egler, J. Phys. Chem. A 2008, 112, 1933–1939; c) M. Mitoraj,
A. Michalak, Organometallics 2007, 26, 6576–6580; d) M. Mi-
toraj, A. Michalak, J. Mol. Model. 2007, 13, 347–355.
[36] C. A. Tolman, Chem. Rev. 1977, 77, 313–315.
[37] R. L. Pruett, J. Smith, J. Org. Chem. 1969, 34, 327–330.
[38] V. K. Srivastava, R. S. Shukla, H. C. Bajaj, R. V. Jasra, Appl.
Catal. A 2005, 282, 31–38.
[8] R. Garcia-Alvarez, J. Diez, P. Crochet, V. Cadierno, Organome-
tallics 2010, 29, 3955–3965.
[39] J. Cloete, S. F. Mapolie, J. Mol. Catal. A 2006, 234, 221–225.
[40] C. Li, G. Zhang, H. Shih, X. Jiang, P. Sun, Y. Pan, C. Cheng,
J. Organomet. Chem. 2009, 694, 2415–2420.
[41] SAINT, v. 7.60a, Bruker AXS Inc., Madison, WI, 2006.
[42] G. M. Sheldrick, SHELXS-97, SHELXL-97 and SADABS, v.
2.05, University of Göttingen, Germany, 1997.
[43] L. J. Barbour, J. Supramol. Chem. 2001, 1, 189–191.
[44] J. L. Atwood, L. J. Barbour, Cryst. Growth Des. 2003, 3, 3–8.
[45] a) E. van Lenthe, A. Ehlers, E.-J. Baerends, J. Chem. Phys.
1999, 110, 8943–8953; b) E. van Lenthe, E. J. Baerends, J. G.
Snijders, J. Chem. Phys. 1994, 101, 9783–9792; c) E. van Len-
the, E. J. Baerends, J. G. Snijders, J. Chem. Phys. 1993, 99,
4597–4610.
[46] ADF, rel. 2012.01, SCM, Theoretical Chemistry, Vrije Uni-
versiteit, Amsterdam, The Netherlands, http://www.scm.com.
Received: April 17, 2013
[9] D. B. Grotjahn, D. A. Lev, J. Am. Chem. Soc. 2004, 126,
12232–12233.
[10] D. B. Grotjahn, E. J. Kragulj, C. D. Zeinalipour-Yazdi, V. Mi-
randa-Soto, D. A. Lev, A. L. Cooksy, J. Am. Chem. Soc. 2008,
130, 10860–10861.
[11] V. Cadierno, P. Crochet, S. E. García-Garrido, J. Gimeno, Dal-
ton Trans. 2004, 3635–3641.
[12] J. W. Faller, P. P. Fontein, Organometallics 2005, 24, 4132–4138.
[13] E. L. Dias, R. H. Grubbs, Organometallics 1998, 17, 2758–
2767.
[14] I. Moldes, E. de la Encarnación, J. Ros, Á. Alvarez-Larena,
J. F. Piniella, J. Organomet. Chem. 1998, 566, 165–174.
[15] J. Soleimannejad, A. Sisson, C. White, Inorg. Chim. Acta 2003,
352, 121–128.
[16] P. Singh, A. K. Singh, Eur. J. Inorg. Chem. 2010, 4187–4195.
Published Online: July 1, 2013
Eur. J. Inorg. Chem. 2013, 4318–4328
4328
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim