Synthesis, spectroscopy, and hydroformylation activity of sterically demanding, phosphite-modified cobalt catalysts

Article abstract of DOI:10.1002/hlca.200590047

The dinuclear complex [Co₂(CO)₆{P(O-2,4-t-Bu ₂C₆H₃)₃}₂] (6) has been synthesised and fully characterised. X-Ray crystal-structure analysis revealed a trans-diaxial geometry, no bridging carbonyls, and Co-Co and Co-P bond lengths of 2.706(5) and 2.134(4) A, respectively. The hydroformylation of pent-1-ene in the presence of 6 was studied at 120-180° at pressures between 20 and 80 bar Syngas. High-pressure (HP) spectroscopy (IR, NMR) was used to detect potential hydride intermediates. HP-IR Studies revealed the formation of [CoH(CO)₃{P(O-2,4-t-Bu₂C₆H₃) ₃}] (2) at ca. 105°, with no significant amount of [CoH(CO) ₄] detectable. The intermediate 2 was synthesised and characterised. The formation of the undesired complex [CoH(CO)₂{P(O-2,4-t-Bu ₂C₆H₃)₃}₂] was completely suppressed due to the large cone angle of the sterically demanding phosphite.

Full text of DOI:10.1002/hlca.200590047

676
Helvetica Chimica Acta ± Vol. 88 (2005)
Synthesis, Spectroscopy, and Hydroformylation Activity of Sterically
Demanding, Phosphite-Modified Cobalt Catalysts
by Reinout Meijbooma)1), Marco Haumanna)2), Andreas Roodt*b), and Llewellyn Damoensea)c)
a) Department of Chemistry, University of Johannesburg, P.O. Box524, Auckland Park 2006, South Africa
b) Department of Chemistry, University of the Free State, P.O. Box339, Bloemfontein 9300, South Africa
(phone: ‡ 27-(0)51-4012547; fax: ‡ 27-(0)51-4307805; e-mail: roodta.sci@mail.uovs.ac.za)
c) Sasol Technology R&D, P.O. Box1, Sasolburg 1947, South Africa
Â
Dedicated to Professor Andre E. Merbach on the occasion of his 65th birthday
The dinuclear complex[Co 2(CO)6{P(O-2,4-t-Bu2C6H3)3}2] (6) has been synthesised and fully characterised.
X-Ray crystal-structure analysis revealed a trans-diaxial geometry, no bridging carbonyls, and CoÀCo and
CoÀP bond lengths of 2.706(5) and 2.134(4) Š, respectively. The hydroformylation of pent-1-ene in the
presence of 6 was studied at 120 ± 1808 at pressures between 20 and 80 bar Syngas. High-pressure (HP)
spectroscopy (IR, NMR) was used to detect potential hydride intermediates. HP-IR Studies revealed the
formation of [CoH(CO)3{P(O-2,4-t-Bu2C6H3)3}] (2) at ca. 1058, with no significant amount of [CoH(CO)4]
detectable. The intermediate 2 was synthesised and characterised. The formation of the undesired complex
[CoH(CO)2{P(O-2,4-t-Bu2C6H3)3}2] was completely suppressed due to the large cone angle of the sterically
demanding phosphite.
Introduction. ± Discovered in 1938 by Otto Roelen at Ruhrchemie [1], hydro-
formylation is, today, the most important application of homogeneous catalysis on an
industrial scale [2], with worldwide annual production capacities exceeding 6 million
tons. In the hydroformylation reaction, also known as oxo synthesis, alkenes react with
carbon monoxide (CO) and hydrogen (H2) in the presence of a transition metal catalyst
to form either linear or branched aldehydes.
The first generation of catalysts was based on unmodified [CoH(CO)4] (1), which
required high reaction pressures to ensure the stability of the active catalyst and to
avoid cobalt plating. In 1966, Shell reported a system where the addition of a tertiary
phosphine stabilised the catalyst to such an extent that reaction pressures below 100 bar
were feasible [3 ± 5]. The larger steric demand of the phosphine ligand, compared to
CO, also led to improved selectivity for the desired linear products [6]. However, the
activity was low compared to that of the unmodified system. More important from an
economic point of view, consecutive hydrogenation of the aldehydes occurred, leading
to alcohols as the primary products. This characteristic of the catalyst system can, thus,
also result in hydrogenation of the alkene feedstock, leading to eventual undesired
alkane formation.
The selectivity can be rationalized by the Heck ± Breslow mechanism, originally
derived for unmodified cobalt catalysis [7]. This mechanism was later generalised and
1
)
)
Present address: University of the Free State, Bloemfontein.
Present address: Institut für Chemische Reaktionstechnik, Universität Erlangen-Nürnberg, D-91058
Erlangen.
2
ꢀ 2005 Verlag Helvetica Chimica Acta AG, Zürich

Helvetica Chimica Acta ± Vol. 88 (2005)
677
Scheme 1. Generalised Heck ± Breslow Mechanism
applied to ligand-modified systems as well (Scheme 1), in which the catalyst precursor
is the monosubstituted hydride, i.e., [CoH(CO)3L] (2).
The unsaturated [CoH(CO)2L] species 3, a p-complex, is formed by the loss of a
CO ligand and rapid addition of an alkene to the intermediary 16-electron species.
Hydrogen transfer to the alkene is influenced by the steric demand of the ligand,
leading either to Markovnikov or, preferentially, anti-Markovnikov addition. In the
case of unmodified catalysis, the symmetry of the species should yield equal amounts of
linear (l) and branched (b) aldehydes (l/b ˆ 1), and addition of CO produces the alkyl
species [(RCH2CH2)Co(CO)3L] (4). Alkyl migration to a coordinated CO ligand
results in the acyl species [{R'C(O)}Co(CO)3L] (5), which is cleaved by H2 to form the
aldehyde, thus regenerating the hydride 2.
Since the original Shell process, based on trialkylphosphines, was introduced, a
variety of phosphine ligands have been studied [8]. Recently, the use of tertiary
phosphines in which the P-atom is incorporated in a limonene bicycle has been
reported [9]. Except for one patent from 1967 [10], no reports on the use of phosphite
ligands in the Co-catalysed hydroformylation have appeared. Since phosphites should
decrease the electron density on the Co centre, relative to phosphines, they are
expected to yield fewer hydrogenation products. In a previous publication, we reported
triphenylphosphite-modified cobalt hydroformylation studied by high-pressure (HP)
NMR, HP-IR, batch autoclave reactions, and independent synthesis of the proposed
hydride intermediates [11]. It was found in that study that a large amount of the
bis(phosphito)cobalt hydride is formed, which is expected to be catalytically relatively
inactive towards hydroformylation, but enhances the isomerisation of the alk-1-enes to
less-reactive internal alkenes. Here, we report our findings on phosphite-modified Co-
catalysed hydroformylation using a ligand with a significantly larger Tolman cone angle
[12]. The increased cone angle of tris(2,4-di-tert-butylphenyl)phosphite (1758) com-
pared to P(OPh)3 (1288) [12] is assumed to prevent the formation of the catalytically
inactive bis(phosphito)cobalt hydride.

678
Helvetica Chimica Acta ± Vol. 88 (2005)
Experimental. General. All manipulations of air- and moisture-sensitive compounds were performed by
means of standard Schlenk techniques under prepurified N2 [13]. Toluene was distilled from Na, and THF from
Na/benzophenone prior to use [14]. All other solvents were used as received. [Co2(CO)8] was purchased from
Strem Chemicals and stored at 48 under N2 in a Schlenk tube wrapped with Al foil. The C9ÀC11 paraffin cut (C8 3,
C9 26, C10 29, C11 33, C12 9%) and pent-1-ene (> 99%), a Fischer± Tropsch derived product, was supplied by
Sasol. Syngas ([CO]/[H2] ˆ 0.5) was purchased from Afrox. All other reagents were purchased from Sigma-
Aldrich and used as received. UV and IR Data: nmax in cmÀ1 (rel. int. %). NMR Data: d in ppm rel. to SiMe4.
Synthesis of [Co2(CO)6{P(O-2,4-t-Bu2C6H3)3}2] (6). The synthesis of compound 6 was similar to that of the
related complex[Co 2(CO)6{P(OPh)3}2] [11][15]. Tris(2,4-di-tert-butylphenyl)phosphite (1.83 g, 2.83 mmol) was
added to a soln. of [Co2(CO)8] (0.422 g, 1.2 mmol) in toluene (10 ml). Gas evolution was observed immediately.
The mixture was stirred and heated to 508. IR Spectroscopy showed mainly dimer formation. The mixture was
stirred overnight. The volatiles were removed in vacuo, and the solid was washed with pentane (2 Â 10 ml). The
straw-coloured solid was dried in vacuo to give 6. Crystals suitable for X-ray diffraction (see below) were
obtained from a sample acquired from an HP-IR run after one year. Yield: 1.76 g (93%). IR (CH2Cl2): 2002
(22), 1984 (100; n(CO)). IR (KBr): 2963 (84, n(CH)), 2871 (21, n(CH)), 2005 (34, n(CO)), 1987 (100, n(CO)),
1979 (92, n(CO)), 1494 (48, das(Me)), 1400 (23), 1363 (31), 1262 (30), 1210 (32, n(PO)), 1186 (49), 1161 (20),
1081 (70), 689 (31), 914 (35), 876 (46), 818 (30, Ar), 517 (29). 1H-NMR (300 MHz, CDCl3): 7.43, 7.03, 6.92
(arom. H); 1.24 (Me), 1.20 (Me). 13C-NMR (75.5 MHz, CDCl3): 148.19; 146.51; 138.51; 124.58; 123.57; 119.40;
34.94; 34.39; 31.32; 30.27. 31P-NMR (121.5 MHz, CDCl3): 156.36.
Synthesis of [CoH(CO)3{P(O-2,4-t-Bu2C6H3)3}] (2). DMF (2.0 ml, 25.8 mmol) was added to a stirred soln.
of [Co2(CO)8] (1.699 g, 4.95 mmol) in toluene (10 ml). A pink precipitate ([Co(DMF)6][Co(CO)4]2) formed
within 1 h, and the supernatant completely decoloured. The mixture was cooled to 08, and conc. HCl was added
(6.0 ml). Immediately, a blue aq. phase and a light-yellow org. phase separated. The mixture was stirred for
30 min, after which the aq. phase was removed by syringe. The org. phase was washed with conc. HCl (3 Â
1.0 ml). IR Spectroscopy showed the presence of [CoH(CO)4]. Then, P(O-2,4-t-Bu2C6H3)3 (2.135 g, 3.3 mmol)
was added and gas evolution was observed. IR Spectroscopy showed the presence of 2 (ca. 1 ml of the soln. was
added to 1 g of ligand and allowed to react; IR analysis showed 2 as the only cobalt carbonyl species). The soln.
was stored in a freezer, and a white solid crystallised. The mother liquor was removed, and the compound was
dried in vacuo. IR (toluene): 2071 (CO, 47), 2020 (CO, 59), 1998 (100, toluene). 1H-NMR (300 MHz, C6D6): 7.89
(m, 1 arom. H); 7.53 (s, 2 arom. H.); 1.71 (s, t-Bu); 1.16 (s, t-Bu); À 11.04 (d, 2J(P,H) ˆ 13.5 Hz). 13C-NMR
(75.4 MHz, C6D6): 148.77; 147.08; 139.27; 125.08; 124.19; 120.14; 35.42; 34.50; 31.40; 30.72. 31P-NMR
(121.4 MHz, C6D6): 149.08.
IR Experiments. All IR spectra were recorded on a Bruker Equinox-55 FT-IR spectrometer and analysed
with the Bruker OPUS-NT software (32 scans, 4 cmÀ1 resolution, Blackman ± Harris 3-term apodisation). Data
for solns. of 2 and 6 were collected using NaCl windows (optical pathlength 0.1 mm). The high-pressure (HP)
experiments were carried out in a 55-ml SS-316 autoclave equipped with a mechanical stirrer (750 r.p.m.),
temperature control, and pressure control (University of Amsterdam) [16]. The soln. was pumped through a
bypass in which ZnS windows were embedded (optical pathlength at 258: 0.3 mm).
The autoclave was flushed with Ar gas prior to use. A soln. of dimer 6 ([Co] ˆ 1500 ppm (mg/kg solvent) in
the final volume) in degassed solvent (10 ml) was transferred under Ar to the autoclave. The appropriate
amount of ligand was added, and the assembled autoclave was purged with Syngas (3Â). The autoclave was
then pressurised at r.t. to 15 bar, resulting in a final pressure of ca. 20 bar at 1408. When the autoclave had
reached the reaction temp., pentene (2 ml) was injected from an attached sample reservoir to give a total
reaction pressure of 50 bar Syngas. Absorbance values at fixed wavelengths were fitted to the appropriate
equations [17] using the SCIENTIST least-squares program [18]. After the reaction, GC samples were taken
from the cooled, depressurised soln. 31P-NMR Spectra of samples taken from the cooled and depressurised
solution showed the presence of only 2, 6, and free phosphite ligand.
NMR Experiments. NMR Spectra were recorded on a Varian Inova spectrometer (1H: 300 MHz, 13C:
75.5 MHz, 31P: 121.5 MHz) at ambient temp. The spectra were referenced rel. to SiMe4 (1H and 13C) or 85%
H3PO4 (31P), using residual solvent signals (d(H) 7.27 or 7.16 for CHCl3 and C6D5H, resp.), the actual solvent
resonances (d(C) 77.0 and 128.0 for CDCl3 and C6D6, resp.), or externally (31P). HP-NMR experiments were
performed in a 10-mm high-pressure ROE cell [19]. For these experiments, the required amount of dimer 6 was
weighed in an Ar-filled sample tube, dissolved in C6D6 (1 ml), and transferred into the NMR tube via syringe.
The appropriate amount of P(O-2,4-t-Bu2C6H3)3 was dissolved in C6D6 (1 ml), and transferred under Ar into the
NMR tube. The cell was purged with Syngas (3Â), pressurised to 40 bar, and left under pressure overnight (ca.
15 h) to allow proper gas dissolution.

Helvetica Chimica Acta ± Vol. 88 (2005)
679
Batch Autoclave Experiments. Gas-uptake experiments were performed with a stainless-steel batch
autoclave (300 ml; Parr model 4560). The autoclave ± equipped with electrical heating, magnetic-drive gas-
dispersion stirrer, and sample dip tube ± was flushed with Ar prior to use. Required amounts of 6 and ligand (L)
were dissolved in degassed paraffin (50 ml). The assembled autoclave was purged with Syngas (3Â) before
being pressurised to 20 bar. When the autoclave setup had reached the desired temp., pent-1-ene was injected
from a 150-ml sample bomb, and the final reaction pressure was adjusted. The gas uptake of a thermostatted, 1-l
gas reservoir was monitored by electronic pressure detection and used to calculate the reaction progress. End-
of-run samples were analysed with an Agilent 6890 GC system equipped with a Hewlett-Packard Pona column
(50 m  0.5 mm). Temp. program: 1008 (hold 10 min), 2508 (28/min), 2508 (5 min hold), 3008 (108/min), 3008
(5 min). N2 Carrier-gas flow was kept constant at 0.7 ml/min.
X-Ray Crystallography3). Crystals of 6 (see Scheme 2) were grown from paraffin as described above. X-
Ray-diffraction data were collected on a Siemens SMART CCD diffractometer using MoKa radiation
(0.71073 Š) and w-scans at 293(2) K. After collection, the first 50 frames were repeated to check for decay,
which was not observed. All reflections were merged and integrated with SAINT [20], and were corrected for
Lorentz, polarization, and absorption effects by means of SADABS [21]. The structures were solved by the
heavy-atom method, and refined through full-matrixleast-squares cycles using the SHELXL97 [22] software
package, with S(/Fo/ À /Fc/)2 being minimised. All non-H-atoms were refined with anisotropic displacement
parameters, while the H-atoms were constrained to parent sites by means of a riding model. The t-Bu groups in
para position, all showing ca. 70% statistical disorder on the Me groups, were restrained by allowing limited
anisotropic movement, using ISOR in SHELXL97. The DIAMOND (Visual Crystal Structure Information
System) software [23] was used for the graphics. Crystal data and details of data collection and refinement are
given in Table 1.
Results and Discussion. ± Synthesis. The dinuclear complex[Co (CO)6{P(O-2,4-t-
2
Bu2C6H3)3}2] (6) was synthesised from dicobalt octacarbonyl and an excess of tris(di-
2,4-tert-butylphenyl)phosphite as ligand (Scheme 2). We recently reported the
molecular structure of the ligand [24], as well as its behaviour in the presence of Rh-
based hydroformylation catalysts. The synthesis of dimer 6 is similar to that of the
previously reported P(OPh)3 complex[Co (CO)6{P(OPh)3}2] [15]. Complex 6 was
2
obtained in excellent yields (90 ± 95%) and characterised by IR, H-, 13C- and 31P-
1
NMR. Crystals of 6 suitable for X-ray diffraction could be obtained from the
decomposition of hydride 2. Compound 6 was used as starting material for the HP-IR,
HP-NMR, and batch autoclave reactions.
The hydrido complex[CoH(CO) 3{P(O-2,4-t-Bu2C6H3)3}] (2) was obtained by
reacting [CoH(CO)4] (1) with 1 equiv. of the ligand (Scheme 2). The product was a
light-yellow, crystalline solid, which was moderately O2- but not moisture-sensitive. It
was thermally sensitive at atmospheric pressure, and dimerised above 08 as a solid, and
above À 208 in solution.
Reaction of 2 with excess phosphite (up to 100 equiv.) did not result in formation of
the bisphosphito complex[CoH(CO) {P(O-2,4-t-Bu2C6H3)3}2], as confirmed by IR
2
spectroscopy (n(CO) in toluene). This is in contrast to the formation of the
bisphosphito hydride when using phosphites with smaller Tolman cone angles [12]
such as P(OPh)3 [11].
Solid-State Structure of Compound 6. A molecular diagram showing the numbering
scheme and thermal ellipsoids for 6 is given in Fig. 1; selected geometric parameters are
3
)
The crystallographic data for 6 have been deposited with the Cambridge Crystallographic Data Centre
(CCDC) as supplementary publication number CCDC-266164. Copies of the data can be obtained, free of
charge, by application to CCDC, 12 Union Road, Cambridge, CB2 1EZ, UK (fax: ‡ 44-1223-336033;
e-mail: deposit@ccdc.cam.ac.uk), or via the internet (http://www.ccdc.cam.ac.uk/products/csd/request).

680
Helvetica Chimica Acta ± Vol. 88 (2005)
Table 1. Crystal Data and Structure Refinement for 6
Empirical formula
Formula weight
Temperature [K]
Crystal size
Wavelength [Š]
Crystal system, space group
a [Š]
C93H150Co2O12P2
1639.93
293(2)
0.30 Â 0.26 Â 0.04 mm
0.71069
Triclinic, P1
Å
11.833(5)
b [Š]
13.147(5)
c [Š]
18.249(5)
a [8]
100.205(5)
b [8]
97.546(5)
g [8]
114.740(5)
2469.7(16)
1, 1.103
0.421
Volume [Š3]
Z, calc. density [Mg mÀ3
Absorption coefficient [mmÀ1
F(000)
]
]
888
q Range [8]
Indexranges
Reflections collected/unique/Rint
Completeness to 2q
5.10 to 25.35
À 14 ꢀ h ꢀ 14, À 12 ꢀ k ꢀ 15, À 21 ꢀ l ꢀ 21
13398/8880/0.0746
98.1%
Max. and min. transmission
Refinement method
Data/restraints/parameters
Goodness-of-fit on F2
0.9838, 0.8872
Full-matrixleast-squares on F2
8880/138/616
0.984
Final R indices [I > 2s(I)]
R1 ˆ 0.0690
wR2 ˆ 0.1523
R1 ˆ 0.1580
R Indices (all data)
wR2 ˆ 0.1912
0.531 and À 0.333
À3
Largest diff. peak and hole [e Š
]
Scheme 2. Synthesis of Dimer 6 and Hydride 2
summarised in Table 2. Fig. 1 clearly shows a dinuclear arrangement formed by two
trigonal [Co(CO)3L] fragments, two phosphite ligands being in trans positions with
respect to the CoÀCo bond (2.706(5) Š). The molecule has an inversion centre, which
bisects the CoÀCo bond, and the phosphite ligands have a measured cone angle of
1768. A trigonal plane is defined by the three CO ligands, and the Co-atom is displaced
by 0.193(3) Š towards the apical phosphite ligand.

Helvetica Chimica Acta ± Vol. 88 (2005)
681
Fig. 1. X-Ray crystal structure of [Co2(CO)6{P(O-2,4-t-Bu2C6H3)3}2] (6) (50% probability). The H-atoms and
the statistical disorder (ca. 70%) at the p-t-Bu groups are omitted for clarity. For the Ph rings, the first digit refers
to the ring number, while the second digit indicates the number of the C-atom in the ring. Primed atoms were
generated via the crystallographic inversion centre.
Table 3 compares the geometric parameters of 6 with those of known, similar
complexes. The distortion induced in these types of compounds is evident, more-bulky
ligands showing more-significant effects. The distortion due to the (sterically demand-
ing) phosphorus donor ligands is manifested in the displacement of the Co-atom from
the trigonal plane formed by the three CO ligands, as well as by the lengthening of the
CoÀCo bond. This is assumed to be due to the steric demand of the P(O-2,4-t-
Bu2C6H3)3 ligands, which introduce intramolecular crowding at the CO ligands, causing
a larger distortion, as manifested by the out-of-plane distance (0.193(3) Š; trigonal
plane formed by the three CO C-atoms) of the individual Co-atoms. This distortion is
further observed in the PÀCoÀCO angles. We are currently further investigating the
influence of phosphorus donor ligands on the displacement of Co-atoms in associated
complexes.
High-Pressure IR Experiments. The formation of the active hydroformylation
species, [CoH(CO)3{P(O-2,4-t-Bu2C6H3)3}] (2), was studied at various ligand/Co ratios
by heating a mixture of P(O-2,4-t-Bu2C6H3)3 and dimer [Co2(CO)6{P(O-2,4-t-
Bu2C6H3)3}2] (6) under Syngas pressures in paraffin. The Co concentration was varied
between 1000 and 2400 ppm, with the [phosphite]/[Co] ratio varying between 4 :1 and
20 :1. The formation of the hydrido species was studied at temperatures and pressures
of up to 1808 and 80 bar. Table 4 summarises these experiments and observations.

682
Helvetica Chimica Acta ± Vol. 88 (2005)
Table 2. Selected Interatomic Bond Distances (Š) and Angles (8) of 6. Symmetry-related atoms are indicated.
CoÀC(2)
CoÀC(1)
CoÀC(3)
CoÀP
1.767(9)
1.769(8)
1.792(9)
2.134(4)
2.706(5)
1.589(6)
1.592(8)
C(2)ÀCoÀC(1)
C(2)ÀCoÀC(3)
C(1)ÀCoÀC(3)
C(2)ÀCoÀP
120.53(32)
118.81(30)
117.14(33)
92.72(18)
96.69(18)
99.48(17)
83.33(18)
82.70(18)
85.21(17)
174.95(4)
104.18(20)
106.14(21)
91.60(19)
113.06(12)
120.00(12)
118.98(13)
178.65(66)
178.55(65)
176.16(58)
124.46(36)
131.86(28)
126.50(31)
CoÀCoa)
PÀO(12)
PÀO(13)
C(1)ÀCoÀP
C(3)ÀCoÀP
C(2)ÀCoÀCoa)
C(1)ÀCoÀCoa)
C(3)ÀCoÀCoa)
PÀCoÀCoa)
O(2)1ÀPÀO(13)
O(12)ÀPÀO(11)
O(13)ÀPÀO(11)
O(12)ÀPÀCo
O(13)ÀPÀCo
O(11)ÀPÀCo
PÀO(11)
1.596(5)
1.146(8)
1.137(9)
1.124(10)
1.399(8)
1.410(6)
1.399(8)
O(1)ÀC(1)
O(2)ÀC(2)
O(3)ÀC(3)
O(11)ÀC(11)
O(12)ÀC(21)
O(13)ÀC(31)
O(1)ÀC(1)ÀCo
O(2)ÀC(2)ÀCo
O(3)ÀC(3)ÀCo
C(11)ÀO(11)ÀP
C(21)ÀO(12)ÀP
C(31)ÀO(13)ÀP
a) À x, 1 À y, À z.
Table 3. Structural Correlations for Complexes of the Type [Co2(CO)6{L}2] (L ˆ PX3, X ˆ alkyl, alkoxy, or
aroyl)
Distance [Š]
CoÀCo
Angle [8]
Ref.
Ligand (L)
CO
CoÀP
±
CoÀ(CO)a) db)
PÀCoÀCO PÀCoÀCo
±
c
2.522(1)
1.82(1)
)
±
1.93(1)d)
1.783(2)
1.776(7)
P(OPh)3
2.6722(4)
P(O-2,4-tBu2C6H4)3 2.706(5)
2.1224(4)
2.134(4)
0.158(1) 95.0
0.193(3) 96.3(2)
177.3(2)
174.95(4)
177.00(6)
180
[11]
e
)
P(OiPr)3
P(nBu)3
PMe3
2.6544(12) 2.1350(12) 1.769(2)
2.665(14)
2.669(1)
0.070
±
92.3
±
93.4
[25]
[26]
[27]
2.178(15)
2.175(1)
1.75(3)
1.772(3)
0.104
180
a) Average of equivalent values. b) Distance of Co from trigonal plane as calculated from data obtained from
CSD. c) Square pyramidal geometry due to bridging CO ligands. d) Bridging CO. e) This work.
The monosubstituted hydride 2 ± having strong IR absorbances at 2071, 2021, and
2000 cmÀ1 ± started to form at 1058, independent of Co concentration or ligand excess.
At this temperature, the strong n(CO) absorbance at 1984 cmÀ1, typical for 6, started to
decrease in intensity, while the second characteristic absorption for the dimer
(2000 cmÀ1) was obscured by the hydride absorption. At 1258, these two bands (1984
and 2000 cmÀ1) had disappeared completely in all experiments, regardless of Co
concentration and ligand excess. Similar behaviour was also observed in HP-NMR
studies in (D6)benzene, where only the modified hydride 2 was observed at temper-
1
atures of up to 1308 (pressure: ca. 50 bar) by H- and 31P-NMR.

Helvetica Chimica Acta ± Vol. 88 (2005)
683
Table 4. Results of High-Pressure IR Studies on the Formation of 2 in Paraffin
b
c
d
[Co]
[L]/[Co]
TMH
)
TDimer
)
TUH
)
Pressure
[bar]
[ppm]
[mm]a)
[8]
[8]
[8]
1000
1400
1400
1800
1800
2400
12.0
17.2
17.2
22.1
22.1
28.8
8 : 1
4 : 1
8 : 1
8 : 1
20 : 1
8 : 1
105
105
105
105
100
105
125
125
125
125
120
125
n.d.e)
135 (25)f)
n.d.
20 ± 30
20 ± 50
20 ± 50
20 ± 80
20 ± 80
20 ± 26
140 (80)f)
n.d.
n.d.
a) Concentration in starting solution at 1 bar and 258. b) Temperature where 2 appeared. c) Temperature where
f
6 disappeared completely. d) Temperature where 1 appeared. e) Not detected. ) Pressure where 1 was formed.
Fig. 2 illustrates the formation of 2 and the disappearance of 6 monitored by IR in
the temperature range between 55 and 1408 at Syngas pressures of 20 ± 26 bar. The Co
concentration was 2400 ppm, with a [P(O-2,4-t-Bu2C6H3)3]/[Co] ratio of 8 :1.
Fig. 2. Temperature-dependent IR spectra showing the formation of modified hydride 2 from dimer 6. [Co] ˆ
2400 ppm, [P(O-2,4-t-Bu2C6H3)3]/[Co] 8 :1, starting pressure ˆ 20 bar Syngas, solvent ˆ paraffin.

684
Helvetica Chimica Acta ± Vol. 88 (2005)
No unmodified hydride, i.e., [CoH(CO)4] (1), could be detected under the above
conditions, even after deconvolution of the spectra. Complex 1 exhibits strong
absorbance at 2030 cmÀ1, which, since being close to the band of 2 at 2021 cmÀ1, can
appear as either a shoulder or a peak at high concentrations. The peak at 2053 cmÀ1,
however, suggests that small amounts of unmodified 1 do, in fact, form under these
conditions.
The formation of complex 2 was studied in detail. Over a period of 40 min, a
solution of 2 was kept at 1058 and at 21 bar pressure. Fig. 3 depicts the decrease of the
dimer absorption at 1984 cmÀ1, and the simultaneous increase of the absorption at
2000 cmÀ1 (for hydride 2). It can be clearly seen that hydride formation at a certain
temperature is time dependent, due to either a nonequilibrated solution or undissolved
dimer particles.
Fig. 3. First-order kinetics for the formation of 2. Lower curve: decrease in the absorbance of n(CO) at
1984 cmÀ1 of 6; upper curve: increase in the absorbance of n(CO) at 2071 cmÀ1 of 2. Conditions: 1058,
[Co] ˆ 1800 ppm, [P(O-2,4-t-Bu2C6H3)3]/[Co] ˆ 8 :1, 21 bar Syngas, solvent ˆ paraffin.
However, when we calculated the rate constants for hydride formation and dimer
decrease, the values were, within experimental error, the same: kDimer ˆ 0.0019 Æ 0.0003
sÀ1, kHydride ˆ 0.0016 Æ 0.0003 sÀ1. This suggests that the way the dimer is split by
dissolved H2 occurs without the formation of intermediates.
The formation of the modified hydride 2 from the unmodified hydride 1 was studied
in order to exclude the presence of 1 as far as possible under conditions with a ligand-
to-metal ratio higher than 4 :1. For these experiments, [Co2(CO)8] at a concentration of
1500 ppm was dissolved in paraffin, and the formation of [CoH(CO)4] (1) was observed
at 1408 and 40 bar Syngas pressure. From the (weak) absorption at 1868 cmÀ1,
indicating bridging CO groups, we concluded that there is an equilibrium between
[Co2(CO)8] and [CoH(CO)4] (1) under these conditions (Scheme 3). The expected IR
absorptions for 1 at 2115, 2053, 2030, and 1998 cmÀ1 were clearly observed.

Helvetica Chimica Acta ± Vol. 88 (2005)
685
Scheme 3. Equilibria between [Co2(CO)8] and Hydrides 1 and 2. Conditions: [Co] ˆ 1500 ppm, pressure ˆ
40 bar Syngas, solvent ˆ paraffin; L ˆ P(O-2,4-t-Bu2C6H3)3.
Fig. 4 further shows the HP-IR spectrum of 1 as a function of ligand concentration.
Upon addition of ligand (>0.5 equiv.) to a solution of 1, the absorption at 1868 cmÀ1
disappeared. The absorptions due to 1 decreased, with a clear increase of the
absorptions indicative of modified hydride 2. After addition of 2.5 equiv. of ligand, the
absorptions due to hydride 1 disappeared almost completely. Only a shoulder remained
for the strongest absorption at 2030 cmÀ1, which could be observed only after Fourier
self-deconvolution. A small amount of unmodified 1 was also indicated by the
absorbance at 2053 cmÀ1. Based on this titration experiment (Fig. 4,b), an equilibrium
constant K was calculated using an equation similar to that reported previously [17].
Thereby, the n(CO) absorbance of 1 at 2053 cmÀ1 was taken, assuming that the CO
pressure was constant, and taking into account that in the first spectrum approximately
30% of the dimer was still present. K was found to be 620 Æ 604) at 1408 and 40 bar
pressure. This confirmed that ca. 1 ± 2% of unmodified 1 would be present under similar
conditions (i.e., [L]/[Co] ˆ 8 :1; [Co] ˆ 1000 ppm; pressure ˆ 40 bar; Tˆ 1408).
To further ensure the absence of [CoH(CO)4] (1) under the reaction conditions,
pressure-variation experiments were performed at 1508. In Fig. 5, the IR spectra of 1,
recorded between 50 and 80 bar Syngas pressure, are shown. No absorbance at
2029 cmÀ1 could be detected, however, above 70 bar, the band at 2021 cmÀ1 broadened.
After cooling the solution to room temperature under 20 bar of Syngas, the three
n(CO) absorptions for 2 were still observed. The ratio between the intensities of these
three bands remained almost constant throughout the experiments. At 1508 and 80 bar,
the relative intensities (%) of the n(CO) absorptions at 2000, 2021, and 2071 cmÀ1 were
100 :56 :39, while, at room temperature, they were 100 :62 :43, respectively. The
relative IR intensities of separately synthesised 2 (in toluene) were 100, 59, and 47% at
1998, 2020, and 2071 cmÀ1, respectively. From these excellent correlations, we
concluded that the species observed at room temperature and 20 bar pressure is the
same as the one observed at 1508 and 80 bar. Furthermore, these results convinced us
that the species observed in the high-pressure studies was indeed, the modified hydride
2.
The stability of [CoH(CO)3{P(O-2,4-t-Bu2C6H3)3}] (2), was accounted for in
various experiments. Table 5 compiles the results of 50-min stability runs. Slight
hydride decomposition was observed at a [L]/[Co] ratio of 8 :1 at a temperature of
1408 (Entries 1 ± 3). Increasing the [L]/[Co] ratio to 12 resulted in a virtually stable
system (Entry 4), even under more-drastic reaction conditions. Thus, if the resting state
of the catalytic cycle (in this case [CoH(CO)3{P(O-2,4-t-Bu2C6H3)3}]) is stabilised by
4
)
When [CO] is constant and L corresponds to P(O-2,4-t-Bu2C6H3), the following equation applies at
K0
‰2Š
equilibrium: K ˆ
ˆ
; see Fig. 4.
‰COŠ ‰1Š‰LŠ

686
Helvetica Chimica Acta ± Vol. 88 (2005)
Fig. 4. a) HP-IR Spectra of the equilibria between [Co2(CO)8], hydride 1, and hydride 2 at 1408 and 40 bar
Syngas pressure in paraffin. [Co] ˆ 1500 ppm in final volume; L ˆ P(O-2,4-t-Bu2C6H3)3. b) Decrease in the IR
2
absorbance at 2053 cmÀ1, fitted to appropriate function ).

Helvetica Chimica Acta ± Vol. 88 (2005)
687
Fig. 5. Stability of Complex 6 in paraffin at 1508 between 50 and 80 bar Syngas pressure, as monitored by HP-IR.
[Co] ˆ 1800 ppm, [P(O-2,4-t-Bu2C6H3)3]/[Co] 8 :1, starting pressure ˆ 20 bar.
Table 5. Stability of Complex 2 in Paraffin, as Monitored by the Relative Change in Selected IR Absorptions. L ˆ
P(O-2,4-t-Bu2C6H3)3.
Entry
[Co] [ppm]
[L]/[Co]
T [8]
Pressure [bar]
Rel. change [%]a)
2000
2021
2071
1
2
3
4
5
1800b)
1800
8 : 1
8 : 1
8 : 1
12 : 1
8 : 1
120
140
140
160
140
20
20
50
80
30
À 2.3
À 8.0
À 6.3
‡ 0.4
À 3.6
À 2.2
À 7.8
À 4.8
À 0.1
À 7.0
À 0.8
À 6.9
À 1.7
À 0.5
À 5.7
1800b)
1800b)
2400
a) IR Signals at 2000, 2021, and 2071 cmÀ1, resp. b) Addition of pent-1-ene (hydroformylation reaction
conditions).
excess ligand, the overall activity of the catalyst decreases. This was, indeed, observed in
batch autoclave experiments (see below).
It was known from previous experiments with less-bulky ligands [11] that the
formation of bis(phosphito)cobalt hydride species occurs at higher ligand-to-metal
ratios. These hydride complexes proved to be less-active hydroformylation catalysts
[11] than the monophosphite complexes. All experiments with P(O-2,4-t-Bu2C6H3)3 as
ligand were conducted with a minimum [phosphite]/[Co] ratio of 4 :1. No bis(phos-
phito)cobalt hydride was observed under these conditions at 1000, 1400, and 1800 ppm

688
Helvetica Chimica Acta ± Vol. 88 (2005)
Co concentration, and increasing this ratio up to 20 :1 ([L] ca. 0.44m at 1 bar, 258) did
not lead to the formation of a bis(phosphito)cobalt hydride species either. This
observation is in agreement with the synthetic study, where the bis(phosphito)cobalt
hydride could not be obtained. The latter should display a different absorbance pattern,
with three n(CO) bands at ca. 2034, 2000, and 1973 cmÀ1, as was observed for P(OPh)3
[11]. Since no such species was observed with P(O-2,4-t-Bu2C6H3)3, the concept of
using very large cone angles to prevent higher substitution on the Co centre proved to
be valid.
To obtain more information on the catalytic activity of modified hydride 2, we
performed a number of hydroformylation runs at Syngas ([CO]/[H2] 1:2) pressures
between 40 and 60 bar, and temperatures between 140 and 1708. The reactions were run
in paraffin (octane) as a solvent. In these experiments, a Co concentration of 1000 ppm
was used to enable correlation with the batch autoclave experiments. After preforming
2 at the desired temperature (starting pressure of ca. 30 bar), pent-1-ene was added,
and the pressure was increased to the predetermined final pressure, which was
maintained for the duration of the hydroformylation experiment.
Repeated scans indicated that hydride 2 was, indeed, an active hydroformylation
catalyst. An increase of the IR absorption at 1734 cmÀ1, indicative of aldehyde
formation, was observed, together with a decrease in the absorptions at 1640 and
1825 cmÀ1, indicative of alk-1-ene, respectively (Fig. 6,a). During the hydroformylation
reaction, the IR spectrum of 2 did not change. No other absorptions in the CO region
(other than aldehyde absorption) were observed, indicating that 2, indeed, corresponds
to the resting state of the catalyst. Fig. 6,a shows an overlay of selected IR spectra of a
typical run (spectra recorded at 1, 2, 3, 5, 7, 10, 15, 20, 25, 30, and 40 min, resp.). Fig. 6,b
shows the increase of the absorption at 1734 cmÀ1, and the decrease of those at 1640 and
1825 cmÀ1 with time.
Assuming the absorption in the mixture to be linear with concentration, the
observed rate constant kobs for pent-1-ene conversion was calculated according to the
first-order rate law (Eqn. 1):
d‰penteneŠ
À
ˆ kobs‰penteneŠ ˆ kobs‰aldehydeŠ
ꢁ1†
dt
The results of these HP-IR hydroformylation experiments are compiled in Table 6. The
absorption at 1825 cmÀ1 is unique for alk-1-enes and, thus, well-suited to calculate kobs
for the consumption of pent-1-ene. The olefin can be hydroformylated to the desired
aldehyde, isomerised to internal pentenes, and hydrogenated to pentane. GC Analysis
of end-of-run samples showed no paraffin formation. Therefore, hydrogenation to
pentane could, under these conditions, be discounted. The kobs value obtained for the
absorption at 1734 cmÀ1 accounts for aldehyde synthesised in the system, which is a
combination of both linear and branched aldehydes. The kobs values correlated to the
resonances at 1640, 1825, and 1734 cmÀ1 were all similar within experimental error;
only the values for 1734 cmÀ1 are, thus, given in Table 6.
The systems studied by HP-IR and by batch autoclave runs were significantly
different (stirring, reactor design, olefin concentration, etc.). Nevertheless, the
calculated kobs values turned out to be comparable, and a similar trend in the values

Helvetica Chimica Acta ± Vol. 88 (2005)
689
Fig. 6. a) Overlayed HP-IR spectra of the hydroformylation of pent-1-ene at 1408 in paraffin. [Co] ˆ 1000 ppm,
[P(O-2,4-t-Bu2C6H3)3]/[Co] 8 :1, pressure ˆ 50 bar Syngas. b) Increase and decrease of the absorptions at 1734,
1640, and 1825 cmÀ1 for the above run.

690
Helvetica Chimica Acta ± Vol. 88 (2005)
Table 6. Rate of Pent-1-ene Hydroformylation in the Presence of Complex 2 (as determined from the change in
IR absorption at 1734 cmÀ1). [Co] ˆ 1000 ppm, solvent ˆ paraffin, L ˆ P(O-2,4-t-Bu2C6H3)3.
Entry
T [8]
Pressure [bar]
[L]/[Co]
k1734 [103 sÀ1
]
1
2
3
4
5
6
140
140
140
150
150a)
160
50
50
50
50
50
50
0 : 1
8 : 1
19 : 1
8 : 1
8 : 1
8 : 1
14 Æ 2
0.73 Æ 4
0.09 Æ 1
0.36 Æ 3
1.76 Æ 2
1.14 Æ 9
a) [Co] ˆ 1500 ppm.
was observed. At high phosphite concentrations (Entry 3, Table 6) a decrease in kobs
was observed. The exact reason is not clear yet, but will be addressed in subsequent
investigations.
Batch Autoclave Experiments. Batch autoclave experiments were carried out using
paraffin as a solvent. A Co concentration of 1000 ppm was used for gas-uptake
measurements. The kinetics were studied at 50 bar Syngas ([CO]/[H2] 1:2 pressure)
between 120 and 1708. The gas uptake of a 1-l vessel was used to calculate the reaction
progress. The gas uptake as a function of time was used to estimate rate constants; in
some experiments, samples were taken at timed intervals over a period of 4 h. GC
Analysis of the samples resulted in time-dependent conversion curves, as depicted in
Fig. 7. From these curves, the observed rate constant for pent-1-ene conversion was
obtained according to Eqn. 1. The results of these batch autoclave hydroformylation
experiments are compiled in Table 7.
Fig. 7. Sampling results for the hydroformylation of pent-1-ene at 1408 (Entry 5 in Table 7). [Co] ˆ 980 ppm,
~
&
[P(O-2,4-t-Bu2C6H3)3]/[Co] 8 :1, pressure ˆ 60 bar Syngas. Symbols: ˆ pent-1-ene, ˆ unbranched aldehyde,
. ˆ branched aldehyde.

Helvetica Chimica Acta ± Vol. 88 (2005)
691
Kinetic data for ꢁunmodifiedꢂ and ꢁP(OPh)3-modifiedꢂ hydroformylation of pent-1-
ene are given in Entries 1 and 2, respectively, of Table 7 [11]. In general, the rate for
phosphite-modified Co-catalysed transformation is very low compared to that with the
unmodified one. None of the experiments in the present study have been optimised,
and, therefore, comparison of different catalysts should only be made under identical
conditions. With the new catalyst, aldehydes were the main product, and only a small
amount of hydrogenation products were observed. In contrast, hydroformylation of
pent-1-ene using P(OPh)3 as a ligand resulted mainly in the formation of internal
alkenes due to isomerisation of pent-1-ene [11]. We proposed that this isomerisation is
catalysed by the bis(phosphito)cobalt hydride species. In the current system, we have
observed neither internal alkenes nor the bis(phosphito)cobalt hydride species by
means of HP-IR studies.
Based on Entries 3 ± 7 (Table 7), an activation energy of 71.8 kJ/mol was calculated
from an Arrhenius plot. This value agrees with literature data of 77 kJ/mol [28] and
95 kJ/mol [29] for hydroformylation with modified homogeneous Co catalysts. The kobs
values determined by gas-uptake measurements were in good agreement with those
calculated from sampling runs (Entries 4 and 5). When decreasing the metal
concentration by 47% from 980 to 520 ppm (Entry 5 vs. 8), the rate constant droped
by 41% to 0.17 Â 10À3 sÀ1, indicating an almost linear relationship between reaction rate
and catalyst concentration.
Table 7. Kinetics of Hydroformylation of Pent-1-ene in the Presence of Catalyst 2 in Paraffin. Batch autoclave
experiments at [Co] ˆ 1000 ppm, L ˆ P(O-2,4-t-Bu2C6H3)3 (see Exper.).
a
Entry
T [8]
Pressure [bar]
[L]/[Co]
kobs ) [103 sÀ1
]
1b)
2b)
3
4
5
6
7
8
9
140
140
120
50
50
50
50
60
50
50
60
50
0 : 1
8 : 1
8 : 1
8 : 1
8 : 1
8 : 1
8 : 1
8 : 1
18 : 1
3.22 Â 103
0.13
0.20
0.29
0.29
1.38
1.90
0.17
1.26
140
140c)
150
170
140d)
170
a) Standard deviations ca. 15%. b) P(OPh)3 as ligand, from [11]. c) 980-ppm Co sampling run using sample dip
tube (Fig. 7). d) [Co] ˆ 520 ppm.
The ligand concentration seems to have a retarding effect on the rate constant
(Entries 7 and 9 in Table 7), as was already observed in the HP-IR studies. This
behaviour is known among metal (e.g., Rh) complexes that can coordinate more than
one ligand; and we recently demonstrated the deactivation of Co catalysts by
triphenylphosphite [11]. Here, the situation seems to be different, since no hydrides
containing more than one phosphite ligand were observed in either the synthesis or HP-
IR studies. From the equilibrium constant, however, it can be calculated that small
amounts of unmodified hydride might be present under these conditions. The
unmodified hydride 1 has a rate constant at least 20 times higher than the modified
hydride 2 (see HP-IR hydroformylation experiments). This means that only 2% of 1

692
Helvetica Chimica Acta ± Vol. 88 (2005)
([Co] ˆ 1500 ppm; [L]/[Co] 8 :1) would lead to an overestimation of kobs by ca. 0.2 Â
10À3 sÀ1 (based on Entry 1 in Table 6).
From the sampling runs using a [L]/[Co] ratio of 8 :1, the l/b (linear vs. branched
aldehyde) ratio was determined to be 2.9 (74% unbranched) in case of 980 ppm Co,
and 2.3 (70% unbranched) in case of 540 ppm Co, respectively. This product
distribution is less favourable (less unbranched aldehyde) than the ones determined
for the smaller P(OPh)3 ligand studied previously. However, our bulky phosphite ligand
resulted in a ten times more-active catalyst system.
Conclusions. ± The complex[Co 2(CO)6{P(O-2,4-t-Bu2C6H3)3}2] (6) was synthesised
in good yield and characterised by IR and NMR spectroscopy and by crystallography.
The proposed hydride intermediate [CoH(CO)3{P(O-2,4-t-Bu2C6H3)3}] (2) was syn-
thesised and characterised by IR and NMR spectroscopy, but, due to its low thermal
stability, no elemental analysis or mass spectrometric data could be obtained.
The dimer 6 was used as a precursor in the hydroformylation of pent-1-ene. High-
pressure (HP) IR studies revealed that, under typical reaction conditions, only hydride
2 was present. The absence of the corresponding bis(phosphito)cobalt hydride was
rationalised by the large cone angle of the phosphite moiety. HP-IR studies revealed
that the unmodified hydride [CoH(CO)4] (1) was present in very low concentrations at
[phosphite]/[Co] ratios higher than 4 :1. Batch autoclave experiments indicated only
modified hydrides, since the activity of 1 is three orders of magnitude higher than that
of 2. The reaction rate was dependent on both the catalyst and the ligand
concentration, and the activation energy for the hydroformylation of pent-1-ene was
calculated to be 71.8 kJ/mol.
Tris(2,4-di-t-butylphenyl)phosphite proved to be large enough in terms of steric
demand (Tolman cone angle) to stabilise the monosubstituted hydride under typical
reaction conditions, but the reaction rate was too low for industrial hydroformylation.
Current research is focused on modifications of steric and electronic properties of
various phosphites to increase the rate of reaction. Further understanding of phosphite-
modified Co-catalysed hydroformylation, as well as more-detailed kinetic investiga-
tions on factors influencing different steps in the catalytic cycle, are necessary.
Financial support from Sasol (including postdoctoral funding for M. H. and R. M.) and the Research Funds
of the University of Johannesburg and the University of the Free State are gratefully acknowledged. Part of this
material is based on work supported by the South African National Research Foundation (SA NRF) (GUN
2053397). Any opinion, findings, and conclusions or recommendations expressed in this material are those of the
authors and do not necessarily reflect the views of the SA NRF. We thank Prof. D. Levendis and Dr. D. Billing
(University of Johannesburg) for the use of their diffractometer, and Mr. A. J. Muller and Mr. L. Kirsten for data
collection and assistance with structure analysis. We also acknowledge Mr. K. Mokheseng (Sasol) for GC
analyses, and Mr. J. A. Vorster (University of Johannesburg) for assistance with HP-NMR experiments. We
thank Prof. M. J. Green, Dr. S. Otto, Dr. C. Grove, and Dr. C. Crause (Sasol) for helpful discussions.
REFERENCES
[1] O. Roelen, Ger. Pat. 949 548, 1938.
[2] C. D. Frohning, C. W. Kohlpaintner, in ꢁApplied Homogeneous Catalysis with Organometallic Com-
poundsꢂ, Eds. B. Cornils, W. A. Herrmann, VCH, Weinheim, 1996, Vol. 1, p. 29; P. W. N. M. van Leeuwen,
ꢁ Homogeneous Catalysis, Understanding the Artꢂ, Kluwer, Dordrecht, 2004.
[3] L. H. Slaugh, R. D. Mullineaux, U.S. Pat. 3.239.569, 1966.

Helvetica Chimica Acta ± Vol. 88 (2005)
[4] L. H. Slaugh, R. D. Mullineaux, U.S. Pat. 3.239.570, 1966.
693
[5] L. H. Slaugh, R. D. Mullineaux, J. Organomet. Chem. 1968, 13, 469.
[6] M. Beller, B. Cornils, C. D. Frohning, C. W. Kohlpainter, J. Mol. Catal., A: Chem. 1995, 104, 17.
[7] R. F. Heck, D. S. Breslow, Chem. Ind. (London) 1960, 467; R. F. Heck, D. S. Breslow, J. Am. Chem. Soc.
1961, 83, 4023.
[8] J. P. Steynberg, K. Govender, P. J. Steynberg, WO Pat., 2002014248, 2002.
[9] C. Crause, L. Bennie, L. Damoense, C. L. Dwyer, C. Grove, N. Grimmer, W. Janse van Rensburg, M. M.
Kirk, K. M. Mokheseng, S. Otto, P. J. Steynberg, J. Chem. Soc., Dalton Trans. 2003, 2036.
[10] J. K. Merzweiler, H. M. Tenney, U.S. Pat. 3351666, 1967.
[11] M. Haumann, R. Meijboom, J. R. Moss, A. Roodt, J. Chem. Soc., Dalton Trans. 2004, 1679.
[12] C. A. Tolman, Chem. Rev. 1977, 77, 313.
[13] D. F. Shriver, M. A. Drezdzon, ꢁThe Manipulation of Air-Sensitive Compoundsꢂ, Wiley-Interscience, New
York, 1986.
[14] D. D. Perrin, W. L. F. Armarego, ꢁPurifiction of Laboratory Chemicalsꢂ, Pergamon Press, Oxford, 1988.
[15] C. Loubscher, S. Lotz, Inorg. Synth. 1992, 29, 178.
[16] W. Rigby, R. Whyman, K. Wilding, J. Phys. E. 1970, 3, 572.
[17] S. Otto, A. Roodt, Inorg. Chem. Commun. 2001, 49, 4.
[18] SCIENTIST for Windows: Program for Least-Squares Parameter Estimation, Version 4.00, Micromath
Inc., Utah, 1990.
Â
[19] D. C. Roe, J. Magn. Res. 1985, 63, 388; I. T. Horvath, J. M. Miller, Chem. Rev. 1991, 91, 1339.
[20] SAINT, Siemens Analytical X-ray Instruments Inc., Madison, Wisconsin, USA, 1995.
[21] SADABS, Siemens Analytical X-ray Instruments Inc., Madison, Wisconsin, USA.
[22] G. M. Sheldrick, SHELXL97, Program for solving crystal structures, University of Göttingen, Göttingen,
Germany, 1997.
[23] K. Brandenburg, M. Berndt, DIAMOND, 1999, Version 2.1c, Crystal Impact GbR, Bonn, Germany.
[24] M. Datt, D. Foster, L. Bennie, C. Steenkamp, J. Huyser, L. Kirsten, G. Steyl, A. Roodt, R. Crous, J. Chem.
Soc., Dalton Trans. 2005, in preparation.
[25] D. Zhao, L. Brammer, Inorg. Chem. 1994, 33, 5897.
[26] T. Bartik, T. Krummeling, B. Happ, A. Sieker, L. Marko, R. Boese, R. Ugo, C. Zucchi, G. Palyi. Catal. Lett.
1993, 19, 383.
[27] J. S. Leigh, K. H. Whitmire, Acta. Crystallogr., Sect. C 1989, 45, 210.
[28] R. V. Gholap, O. M. Kut, J. R. Bourne, Ind. Eng. Chem. Res. 1992, 31, 1597.
[29] R. M. Desphande, R. V. Chaudhari, J. Catal. 1989, 115, 326.
Received November 14, 2004