An NMR study of cobalt-catalyzed hydroformylation using para-hydrogen induced polarisation

Article abstract of DOI:10.1039/b815853c

The syntheses of Co(η³-C₃H₅)(CO) ₂PR₂R′ (R, R′ = Ph, Me; R, R′ = Me, Ph; R = R′ = Ph, Cy, CH₂Ph) and Co(η³-C ₃H₅)(CO)(L) (L = dmpe and dppe) are described,

Full text of DOI:10.1039/b815853c

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
An NMR study of cobalt-catalyzed hydroformylation using para-hydrogen
induced polarisation†
Cyril Godard,a Simon B. Duckett,*a Stacey Polas,b Robert Toozeb and Adrian C. Whitwooda
Received 10th September 2008, Accepted 8th December 2008
First published as an Advance Article on the web 16th February 2009
DOI: 10.1039/b815853c
3
The syntheses of Co(h -C3H5)(CO)2PR2R¢ (R, R¢ = Ph, Me; R, R¢ = Me, Ph; R = R¢ = Ph, Cy, CH2Ph)
3
and Co(h -C3H5)(CO)(L) (L = dmpe and dppe) are described, and X-ray structures for
3
Co(h -C3H5)(CO)(dppe) and the PPh2Me, PCy3 derivatives reported. The relative ability of
3
Co(h -C3H5)(CO)2(PR2R¢) to exchange phosphine for CO follows the trend PMe2Ph < PPh2Me < PCy3
< P(CH2Ph)3 < PPh3. Reactions of the allyl complexes with para-hydrogen (p-H2) lead to the
observation of para-hydrogen induced polarisation (PHIP) in both liberated propene and propane.
Reaction of these complexes with both CO and H2 leads to the detection of linear acyl containing
species Co(COCH2CH2CH3)(CO)3(PR2R¢) and branched acyl complexes Co(COCH(CH3)2)(CO)3-
(PR2R¢) via the PHIP effect. In the case of PPh2Me, additional signals for Co(COCH2CH2CH3)(CO)2-
(PPh2Me)(propene) and Co(COCH(CH3)2)(CO)2(PPh2Me)(propene) are also detected. When the
reactions of H2 and diphenylacetylene are studied with the same precursor, Co(CO)3(PPh2Me)-
(CHPhCH2Ph) is seen. Studies on how the appearance and ratio, of the PHIP enhanced signals vary as
a function of reaction temperature and H2 : CO ratio are reported. These profiles are used to learn
about the mechanism of catalysis and reveal how the rates of key steps leading to linear and branched
hydroformylation products vary with the phosphine. These data also reveal that the PMe2Ph and
PPh2Me based systems yield the highest selectivity for linear hydroformylation products.
Introduction
Cobalt-catalysed alkene hydroformylation represents a reaction
that is of significant industrial importance.1 In this reaction, an
alkene reacts with CO and H2 to form an aldehyde which ideally
should correspond to the linear rather than branched isomer.
Early catalysts were based on Co2(CO)8 and proposed to yield
HCo(CO)4 under the normally employed reaction conditions.2
This “unmodified” system was improved on by replacing a CO for
phosphine (modified) because higher linear to branched product
ratios resulted; their rhodium based counterparts have become
even more important because of their greater selectivity.
Early studies by Heck and Breslow led to an accepted mech-
anism for the hydroformylation reaction that involves a mono-
hydride precursor whichultimatelyyields ametalacylintermediate
that eliminates the aldehyde after reaction with H2. This reaction
sequence is indicated in Scheme 1 and operates at temperatures
close to 150 ◦C with combined CO–H2 pressures of around
200 atmospheres.3 The detection and characterization of the key
species involved in this process has proved to be complex, with
examples of cobalt acyl and alkyl resting states being isolated.4–6
The unmodified species CH3C(O)Co(CO)4, CH3C(O)Co(CO)3
and CH3Co(CO)4 have also been characterised in low temperature
matrices, with the 16-electron species showing little reactivity
Scheme 1 Heck and Breslow hydroformylation mechanism for a modified
catalyst, PR3, showing origin of linear/branched selectivity.
aDepartment of Chemistry, University of York, UK, YO10 5DD
bSasol Technology (UK) Ltd., Purdie Building, North Haugh, St Andrews,
2
towards other ligands resulting in a proposed h -acyl bonding
Scotland, UK 5KY16 9ST
mode.7 Roe has used high pressure NMR methods to demonstrate
that the thermal loss of CO from CH3C(O)Co(CO)4 has a Gibbs
free energy of activation of ca. 90 kJ mol-1 at 363 K and
† Electronic supplementary information (ESI) available: NMR and X-ray
details. CCDC reference numbers 701811–701813. For ESI and crystallo-
graphic data in CIF or other electronic format see DOI: 10.1039/b815853c
2496 | Dalton Trans., 2009, 2496–2509
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
a rate of 11.5 s-1 at 353 K.8 Others have used time-resolved
IR studies to show that the h -acyl containing intermediate
CH3C(O)Co(CO)2(PPh3) re-coordinates CO at a rate of 1.1 ¥
107 dm3 mol-1 s-1 while undergoing competitiv◦e decarbonylation
of the acyl ligand at a rate of 6.2 ¥ 104 s-1 at 25 C.9
Experimental
2
General conditions
All manipulations were carried out under inert atmosphere
conditions, using standard Schlenk techniques (vacuum up to
10-2 mbar, with N2 as an inert atmosphere) or high vacuum tech-
niques (10-4 mbar). Storage and manipulation of samples were car-
ried out using standard glovebox techniques, under an atmosphere
of N2, using an Alvic Scientific Gas Shield glovebox equipped with
a freezer (-32 ◦C), vacuum pump and N2 purge facilities. Sol-
vents were obtained as Analytical Grade from Fisher, and dried
when appropriate by refluxing under nitrogen over sodium wire.
X-Ray data were obtained using a Bruker Smart Apex
Studies on the thermal reactivity of CH3CH2CH2Co(CO)3-
(PBu3) revealed the formation of HCo(CO)3(PBu3), [Co(CO)3-
(PBu3)]2 and CH3CH2CH2C(O)Co(CO)3(PBu3) in addition to H2
◦
and propene at 40 C.10 Similar studies on (CH3)2CHCo(CO)3-
(PBu3) were also described and Rosi et al. concluded that the
isomerisation of these species under standard hydroformylation
conditions does not justify the product mix obtained when
Co2(CO)8–PBu3 is itself examined. The hydroformylation re-
actions selectivity was therefore suggested to result from the
regioselectivity of CoH group transfer to the bound alkene.
˚
diffractometer with Mo-Ka radiation (l = 0.71073 A) using a
SMART CCD camera. Diffractometer control, data collection
and initial unit cell determination was performed using “SMART”
(v5.625 Bruker-AXS).24 Frame integration and unit-cell refine-
ment was carried out with the “SAINT+” software (v6.22, Bruker
AXS). Absorption corrections were applied by SADABS (v2.03,
Sheldrick).
A
more complete picture of the hydroformylation of
propene by HCo(CO)4 has been assembled from density func-
tional theory.11–13 The most stable alkene hydride complex,
=
HCo(CO)3(CH2 CHMe), was found to contain an axial hydride
ligand and equatorial alkene. Subsequent hydride transfer proved
to be reversible, with the regioselectivity being found to be mainly
determined by the stability of the corresponding alkylcobalt
tetracarbonyl product. In the presence of CO, the formation
The syntheses of the cobalt allyl complexes described in this
paper were carried out using modified literature procedures as
detailed below.25,26
2
of the unsaturated, h -stabilised acyls, (CH3)2CHC(O)Co(CO)3
and CH3CH2CH2C(O)Co(CO)3 was indicated. Reactions of these
species with H2 were then considered to proceed via simple and
reversible H2 coordination rather than oxidative addition, which
was shown to correspond to a high-energy product on the way to
aldehyde formation.11,12
Synthesis of Co(g3-C3H5)(CO)2(PPh2Me) 1
Co(CO)4Na was prepared by reduction of Co2(CO)8 (0.6 g,
1.7 mmol) in 100 ml of THF by NaOH powder (ca. 1 g, large excess)
in the glovebox. The solution was stirred for 30 minutes at room
temperature and the colourless solution was then filtered. Addition
of 0.4 ml (27 mmol, large excess) of freshly distilled allybromide
to the eluant was performed at 0 ◦C and the resulting solution
stirred at this temperature for ca. 30 minutes before being allowed
to warm up to room temperature. After 15 minutes, the colour
of the solution had turned bright yellow due to the formation of
The formation of p-allyl cobalt containing complexes has been
shown to be possible via reactions of anions such as Co(CO)4- with
allylbromide,14 or via allyl transfer from chloro-allyl palladium
dimers.15 Furthermore, the reaction of the allyl cobalt complex
3
(h -C3H5)Co(CO)2PPh3 with PPh3 and HSiMe2Ph has been shown
to enable the formation of Co(CO)2(PPh3)2(SiMe2Ph) with the
concomitant formation of propene.16 This corresponds to an
elegant route to the formation of a cobalt–silyl complex. When
such compounds react with H2, however, the formation of para-
magnetic clusters such as [Co(CO)2(PPh3)]3 has been previously
described.17 These clusters are produced by the trimerisation of
CoH(CO)2(PPh3) and result in H2 generation. Such systems have
also been found to be active as arene hydrogenation catalysts.18
3
Co(h -C3H5)(CO)3. A THF solution (20 ml) containing a slight
excess of PPh2Me (700 ml, 3.75 mmol, 1.1 eq.) was then slowly
added via cannula transfer and the reaction mixture stirred at
room temperature for 15–20 minutes. The orange solution was
then filtered, and the solvent removed under vacuum. Extraction
in pentane yielded a dark orange-red solution. Purification was
performed by precipitation in pentane at -80 ◦C, filtration and
the recrystallisation of the precipitate from diethylether. Yield =
420 mg, 34%. nCO in CH2Cl2 (cm-1) 1983, 1923.
3
In this paper we describe studies that employ Co(h -
C3H5)(CO)2PR2R¢ (R, R¢ = Ph, Me; R, R¢ = Me, Ph; R = R¢ = Ph,
Cy, CH2Ph) as precursors to study cobalt based hydroformylation
by NMR spectroscopy. These studies have also involved the
use of para-hydrogen (p-H2).19 p-H2 has been used, because it
leads to the observation of the PHIP (para-hydrogen induced
polarisation)20 effect. This effect has enabled the direct NMR
detection of metal hydride signals for complexes in solution
that would otherwise be invisible.21 PHIP NMR spectroscopy
has also been shown to permit the direct identification of metal
complexes without the need for enhancement of metal hydride
signals, an example being the characterization of intermediates in
palladium-catalyzed hydrogenation through proton resonances of
alkyl ligands.22 This paper builds on a communication where the
3
3
The syntheses of Co(h -C3H5)(CO)2(PPhMe2) 2, Co(h -C3H5)-
3
3
(CO)2(P(CH2Ph)3) 3, Co(h -C3H5)(CO)2(PCy3) 4 and Co(h -
C3H5)(CO)2(PPh3) 5 are described in the ESI.†
Synthesis of Co(g3-C3H5)(CO)2(dppm) (dppm =
diphenyphosphinomethane) 6
3
A THF solution of Co(h -C3H5)(CO)3 (3.4 mmol) was prepared
as described above for 1. A THF solution (20 ml) containing a
slight excess of dppm (1.4 g, 3.75 mmol, 1.1 eq.) was then slowly
added via cannula transfer and the reaction mixture stirred at
room temperature for 15–20 minutes. The orange solution was
then filtered, and the solvent removed under vacuum. Purification
of the product was carried out by recrystallisation from a
dichloromethane solution. Yield = 55%.
3
reactivity of Co(h -C3H5)(CO)2(PCy3) was described.23 We have
employed temperatures up to 120 ◦C but our highest total pressure
is 4 atm.
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 2496–2509 | 2497
©

View Article Online
Synthesis of Co(g3-C3H5)(CO)(dppe) (dppe =
diphenyphosphinoethane) 7
3
A THF solution of Co(h -C3H5)(CO)3 (3.4 mmol) was prepared
as described above for 1. A THF solution (20 ml) containing a
slight excess of dppe (1.5 g, 3.75 mmol, 2.2 eq.) was then slowly
added via cannula transfer and the reaction mixture stirred at room
temperature for 12 hours. The orange solution was then filtered,
and the solvent removed under vacuum. The orange solid was then
re-dissolved in the minimum amount of THF and precipitated by
addition of pentane, filtration and washing of the remaining solid
with pentane. Yield = 30%. nCO in CH2Cl2 (cm-1) 1925.
3
Scheme 2 Synthesis of Co(h -C3H5)(CO)2PR2R¢, where R, R¢ = Ph, Me
(1); R, R¢ = Me, Ph (2); R = R¢= CH2Ph (3), Cy (4), and (5) Ph.
obtained by slow evaporation of a diethylether solution, while
crystals of 4 (PCy3) were obtained from toluene as described
previously.23
para-Hydrogen
ORTEP representations of the structures of 1 and 4 are shown
in Fig. 1. The structures of these species are similar to that of
For the PHIP experiments, hydrogen enriched in the para spin
state was prepared by cooling H2 to 18 K over a paramagnetic
catalyst (Fe2O3 on silica) using the system described previously.27
All the NMR studies were carried out with sample concentrations
of approximately 1 mM and spectra were recorded on a Bruker
3
the previously reported Co(h -C3H5)(CO)2(PPh3) 5 analogue26 and
3
can be described as piano stools with capping h -allyl ligands. The
piano stool includes the Co(CO)2(PR2R) moiety, where in the case
˚
1
of PPh2Me, the Co–P bond length is 2.1715(7) A and the two
DRX-400 spectrometer with H at 400.1, 31P at 161.9 and 13C at
˚ ˚
distinct Co–C(O) bond lengths are 1.772(3) A and 1.757(3) A.
1
100.0 MHz, respectively. H NMR chemical shifts are reported
˚
1
The corresponding Co–P bond length in 4 is 2.2190(4) A, and
in ppm relative to residual H signals in the deuterated solvents
˚
the Co–C(O) distances are almost identical at 1.7812(14) A and
(toluene-d7, d 2.13, and C6D5H, d 7.16), 13C NMR relative
to toluene-d8, d 21.3 and C6D6, d 128.4 and 31P NMR in
ppm downfield of an external 85% solution of phosphoric acid.
Modified COSY, HMQC and EXSY pulse sequences were used as
previously described.28,29
˚
˚
1.7592(14) A. The reported Co–PPh3 distance in 5 is 2.185 A,
and again almost identical to the corresponding distance in 1.
The C–Co–C angle in 1 (R = Ph, Me) was found to be
116.52(13)◦, while the C–Co–P angles were found to be smaller
at 99.65(9)◦ and 96.05(9)◦. The matching angles in 4 (R = Cy) are,
however, 107.73(7)◦, 95.12(5)◦ and 104.39(5)◦ respectively23 while
those for 5 (R = Ph) are reported as being 111.3(2)◦, 93.0(1)◦
and 104.7(1)◦ respectively. This is consistent with the relative
cone angles of the phosphine controlling the resulting geometry.30
In other words, the C–Co–C angle becomes compressed as the
structure changes to reflect the steric effect of the phosphine.
Results and discussion
Syntheses of Co(g3-C3H5)(CO)2PR2R¢ 1–5 (R, R¢ = Ph, Me; R,
R¢ = Me, Ph; R = R¢ = Ph, Cy, CH2Ph)
3
The Co(h -C3H5)(CO)2PR2R¢ complexes 1–5 where R, R¢ = Ph,
˚
Me (1); R, R¢ = Me, Ph (2), and R = R¢ = CH2Ph (3), Cy (4) and Ph
The Co–Callyl distances in 1 were found to be 2.093(3) A,
3
˚
˚
(5) were prepared from in situ generated Co(h -C3H5)(CO)3 which
2.016(3) A and 2.119(3) A, and similar to those found for 5
was obtained from the reaction of Co(CO)4Na with allylbromide.25
Room temperature addition of a small excess of the appropriate
while the corresponding Co–Callyl distances for 4 are 2.0841(15) A,
˚
˚
˚
2.0074(14) A and 2.1197(14) A respectively. There is therefore
no discernable trend in the Co–Callyl distance across the series of
compounds. The two C–C bond lengths within the allyl ligand of
3
phosphine to Co(h -C3H5)(CO)3 led to the desired products,
3
Co(h -C3H5)(CO)2(PR2R¢), in good yield (Scheme 2). Complexes
˚ ˚
1–5 have been characterized by mass spectrometry, and IR and
NMR spectroscopy (Table 1 and experimental). In addition,
crystals of 1 (PPh2Me) that were suitable for X-ray analysis were
1 were determined to be 1.416(4) A and 1.364(4) A, while those
˚
˚
of 5 were reported as 1.399(7) A and 1.385(8) A, and those of 4
˚
˚
are 1.410(2) A and 1.397(2) A. The analogous CO distances for
3
3
3
Fig. 1 ORTEP diagrams of (a) Co(h -C3H5)(CO)2(PPh2Me) 1, (b) Co(h -C3H5)(CO)2(PCy3) 4 and (c) Co(h -C3H5)(CO)(dppe) 7. Ellipsoids drawn at
50% probability level.
2498 | Dalton Trans., 2009, 2496–2509
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Table 1 NMR data for complexes 1–4 and 6–7 in toluene-d8 solution at 295 K
1
1
Species
1H d (multiplicity, intensity)
13C{ H}
31P{ H}
1
4.47 (tt, 1H) Ha, JHH = 6 and 11 Hz
2.44 (dd, 2H) Hb (anti), JHH = 6 Hz, JHP = 4 Hz
1.88 (d, 2H) Hc (syn), JHH = 11 Hz
1.63 (d, 3H) PMePh2, JHP = 8 Hz
7–7.5 (m) PMePh2
205.2 (br) CO
80.0 (s) CH allyl
46.0 (s) CH2 allyl
43.8 (s)
19.4 (d) PMePh2, JPC = 29 Hz
120–135 (m) PMePh2
2
3
4.46 (tt, 1H) Ha, JHH = 6 and 11 Hz
2.35 (dd, 2H) Hb (anti), JHH = 6 Hz, JHP = 4 Hz
1.83 (d, 2H) Hc (syn), JHH = 11 Hz
1.28 (d, 6H) PMe2Ph, JHP = 8 Hz
7–7.8 (m) PMe2Ph
197.6 (br) CO
79.5 (s) CH allyl
45.05 (s) CH2 allyl
19.2 (d) PMe2Ph, JPC = 29 Hz
120–140 (m) PMe2Ph
28.0 (s)
63.9 (s)
69.1 (s)
4.22 (tt, 1H) Ha, JHH = 6 and 10.5 Hz
2.11 (m, 2H) Hb (anti)
1.58 (d, 2H) Hc (syn), JHH = 10 Hz
2.91 (d, 6H) P(CH2Ph)3, JHP = 9 Hz
7–7.2 (m) P(CH2Ph)3
204.5 (d) CO, JPC = 8 Hz
82.0 (s) CH allyl
44.2 (d) CH2 allyl, JPC = 3 Hz
38.6 (d) P(CH2Ph)3, JPC = 17 Hz
125–140 (m) P(CH2Ph)3
4
6
4.78 (tt, 1H) Ha, JHH = 6 and 10.5 Hz
2.61 (dd, 2H) Hb (anti), JHH = 6 Hz, JHP = 3 Hz
1.82 (d, 2H) Hc (syn), JHH = 10.4 Hz
1.0–2.0 (m) PCy3
208 (d) CO, JPC = 8 Hz
79.4 (s) CH allyl
43.4 (s) CH2 allyl
25–40 (m) PCy3
4.31 (m, 1H) Ha
206.7 (br) CO
80.3 (s) CH allyl
46.8 (s) CH2 allyl
53.0 (br d), JPP = 98 Hz
-26.1 (d), JPP = 98 Hz
2.61 (dd, 2H) Hb (anti), JHH = 5 Hz, JHP = 4 Hz
1.87 (d, 2H) Hc (syn), JHH = 10 Hz
3.16 (br) Ph2P–CH2–PPh2
34.1 (dd) Ph2P–CH2–PPh2, JPC = 22 and 32 Hz
6–8 (m) Ph2P–CH2–PPh2
120–140 (m) Ph2P–CH2–PPh2
7
4.32 (m, 1H) Ha
207.0 (br) CO
90.0 (s) CH allyl
47.1 (s) CH2 allyl
58.5 (s)
2.50 (m, 2H) Hb (anti)
1.90 (m, 2H) Hc (syn)
2.61 (d) Ph2P–(CH2)2–PPh2
6–8 (m) Ph2P–CH2–PPh2
27.6 (dd) Ph2P–(CH2)2–PPh2, JPC = 14 and 28 Hz
120–140 (m) Ph2P–(CH2)2–PPh2
◦
3
˚
˚
˚
Table 2 Selected bond lengths (A) and angles ( ) for Co(h -C3H5)-
(CO)2(PPh2Me) 1
1 are 1.146(3) and 1.146(3) while those of 5 are 1.141(4) A and
˚
˚
1.139(6) A and those of 4, are 1.1513(18) A and 1.1522(18) A.
These data suggest that neither the Co–CO nor the allyl
CH2–CH distances change substantially with change in electron
donating power of the phosphine. Key distances and bond angles
for 1 are listed in Table 2 while those for 4 can be found in ref. 23.
Crystal parameters are listed in Table 3.
Assignment
Value
Assignment
Value
Co(1)–P(1)
C(1)–O(1)
C(1)–Co(1)
C(2)–O(2)
C(2)–Co(1)
2.1715(7)
1.146(3)
1.772(3)
1.146(3)
1.757(3)
C(3)–Co(1)
C(4)–Co(1)
C(5)–Co(1)
C(3)–C(4)
C(4)–C(5)
2.093(3)
2.016(3)
2.119(3)
1.416(4)
1.364(4)
Synthesis of Co(g3-C3H5)(CO)2(PPh2CH2PPh2) 6 and
Co(g3-C3H5)(CO)(PPh2CH2CH2PPh2) 7
O(1)–C(1)–Co(1)
O(2)–C(2)–Co(1)
C(4)–C(3)–Co(1)
C(3)–C(4)–Co(1)
C(5)–C(4)–C(3)
C(5)–C(4)–Co(1)
C(4)–C(5)–Co(1)
C(1)–Co(1)–P(1)
C(4)–Co(1)–P(1)
C(3)–Co(1)–P(1)
C(5)–Co(1)–P(1)
177.6(3)
178.4(2)
66.91(16)
72.82(16)
116.5(3)
74.92(18)
66.67(17)
99.65(9)
114.23(9)
154.46(8)
89.16(10)
C(2)–Co(1)–C(1)
C(2)–Co(1)–C(4)
C(1)–Co(1)–C(4)
C(2)–Co(1)–C(3)
C(1)–Co(1)–C(3)
C(4)–Co(1)–C(3)
C(2)–Co(1)–C(5)
C(1)–Co(1)–C(5)
C(4)–Co(1)–C(5)
C(3)–Co(1)–C(5)
C(2)–Co(1)–P(1)
116.52(13)
101.56(13)
125.41(13)
92.32(12)
98.06(12)
40.27(12)
135.39(14)
106.05(15)
38.41(12)
68.26(12)
96.05(9)
The syntheses of the related complexes containing the
chelating phosphines bisdiphenylphosphinomethane, Co(h -
C3H5)(CO)2(dppm) 6, and bisdiphenylphosphinoethane, Co(h -
C3H5)(CO)(dppe) 7 were also completed. The reaction with dppm
proved to readily yield a dicarbonyl complex where the phosphine
ligand is coordinated through only one phosphorus centre. When
a solution of dppe was added slowly to a dilute solution of Co(h -
C3H5)(CO)3, however, the main product proved to be Co(h -
3
3
3
3
C3H5)(CO)(dppe) 7 where the dppe ligand is chelating. These
complexes have also been characterized by mass spectrometry,
IR and NMR spectroscopy (see Table 1).
Crystals of 7 suitable for X-ray diffraction were obtained by slow
evaporation of a toluene solution, and the corresponding ORTEP
diagram is shown in Fig. 1(c). It should be noted, that it has
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 2496–2509 | 2499
©

View Article Online
Table 3 Crystal data and structure refinement information for 1, 7 and [Co(CO)3PPh3]2
Complex
1
7
[Co(CO)3PPh3]2
Empirical formula
Formula weight
Temperature/K
Wavelength/A
Crystal system
C18H18CoO2P
356.22
388(2)
0.71073
Orthorhombic
Pbca
15.2928(17)
13.6440(16)
15.9287(18)
90
90
90
C30H29CoOP2
526.40
115(2)
0.71073
Triclinic
¯
P1
9.4425(7)
9.4559(6)
14.7071(10)
93.077(2)
104.880(2)
101.761(2)
1234.65(15)
C21H15CoO3P
405.23
393(2)
0.71073
Rhombohedral
¯
R3
15.2724(19)
15.2724(19)
13.854(4)
90
90
120
2798.5(9)
˚
Space group
˚
a/A
˚
b/A
˚
c/A
a/◦
b/◦
g /◦
3
˚
Volume/A
3323.6(7)
Z
8
2
6
Density (calculated)/Mg m-3
Absorption coefficient/mm-1
F(000)
1.424
1.132
1472
0.3 ¥ 0.2 ¥ 0.1
2.37 to 30.98
1.416
0.846
548
1.443
1.022
1242
0.19 ¥ 0.18 ¥ 0.02
2.13 to 25.01
Crystal size/mm3
0.21 ¥ 0.11 ¥ 0.05
1.44 to 25.03
-11 ꢀ h ꢀ 5, -11 ꢀ k ꢀ 11,
-16 ꢀ l ꢀ 17
7070
q range for data collection/◦
Index ranges
-22 ꢀ h ꢀ 21, -19 ꢀ k ꢀ 18,
-22 ꢀ l ꢀ 22
-18 ꢀ h ꢀ 18, -18 ꢀ k ꢀ 17,
-16 ꢀ l ꢀ 16
Reflections collected
35594
6523
Independent reflections
Completeness to q = 30.98◦ (%)
Absorption correction
Max. and min. transmission
Refinement method
5049 [Rint = 0.0696]
4325 [Rint = 0.0211]
1091 [Rint = 0.0342]
95.5
99.0
99.2
Multi-scan
Semi-empirical from equivalents
0.960 and 0.822
Full-matrix least-squares on F2
4325/0/307
Semi-empirical from equivalents
1.000 and 0.785
Full-matrix least-squares on F2
1091/0/79
1.000 and 0.824
Full-matrix least-squares on F2
5049/0/200
Data/restraints/parameters
Goodness-of-fit on F2
1.061
1.033
1.110
Final R indices [I > 2s(I)]
R indices (all data)
Largest diff. peak and hole/e A
R1 = 0.0511, wR2 = 0.1133
R1 = 0.0736, wR2 = 0.1227
1.234 and -0.384
R1 = 0.0343, wR2 = 0.0827
R1 = 0.0461, wR2 = 0.0874
0.587 and -0.607
R1 = 0.0242, wR2 = 0.0671
R1 = 0.0261, wR2 = 0.0679
0.339 and -0.293
-3
˚
3
Table 4 Selected bond lengths (A) and angles (◦) for Co(h -
C3H5)(CO)(dppe) 7
3
˚
previously been reported that reactions of Co(h -C3H5)(CO)3 with
3
3
dppe, led to Co(h -C3H5)(CO)(dppe) and [Co(h -C3H5)(CO)2](m-
dppe).31 We have found that increasing the concentrations of
Assignment
Value
Assignment
Value
3
Co(h -C3H5)(CO)3 relative to dppe yields the dimeric species
[Co(h -C3H5)(CO)2(m-dppe)]2 and that Co(h -C3H5)(CO)2(dppe)
can be readily obtained by addition of the dppe to Co(h -
3
1
Co(1)–C(1)
Co(1)–P(1)
Co(1)–P(2)
Co(1)–C(2)
Co(1)–C(3)
Co(1)–C(4)
1.770(3)
2.1640(7)
2.1709(7)
2.084(3)
1.987(3)
2.082(3)
O(1)–C(1)
C(3)–C(2)
C(3)–C(4)
C(5)–P(1)
C(5)–C(6)
P(2)–C(6)
1.150(3)
1.413(4)
1.417(4)
1.869(2)
1.529(3)
1.838(2)
1
C3H5)(CO)4 at 0 ◦C.
The bisphosphine substituted species 7 also exhibits a piano
stool geometry with a capping h -allyl ligand. The piano stool
3
moiety is defined by the metal, the carbonyl ligand and the two
phosphorus centres. The corresponding Co–C and Co–P bond
O(1)–C(1)–Co(1)
P(1)–Co(1)–P(2)
C(1)–Co(1)–C(3)
C(1)–Co(1)–C(4)
C(1)–Co(1)–C(2)
C(1)–Co(1)–P(1)
C(1)–Co(1)–P(2)
C(2)–C(3)–C(4)
C(3)–Co(1)–C(2)
C(3)–Co(1)–P(1)
C(4)–Co(1)–P(1)
178.6(2)
85.06(3)
C(6)–P(2)–Co(1)
C(5)–P(1)–Co(1)
C(2)–Co(1)–P(1)
C(6)–C(5)–P(1)
C(3)–Co(1)–P(2)
C(4)–Co(1)–P(2)
C(2)–Co(1)–P(2)
C(5)–C(6)–P(2)
C(3)–C(4)–Co(1)
C(3)–C(2)–Co(1)
C(4)–Co(1)–C(2)
104.71(8)
110.06(8)
92.05(8)
109.25(16)
114.63(8)
98.66(7)
151.84(8)
107.59(16)
66.07(14)
66.06(14)
69.86(10)
˚
lengths were found to be 1.770(3), 2.1640(7) and 2.1709(7) A
128.64(11)
102.12(11)
104.84(11)
107.89(9)
102.69(8)
114.9(2)
40.53(10)
109.22(8)
148.17(7)
respectively and are almost identical to those seen for 1. The
˚
Co–Callyl bond lengths of 7 are 1.987(3), 1.987(3) and 2.082(3) A
and therefore slightly shorter than those seen for both 1 and 4.
˚
The C–C bonds for the allyl ligand at 1.413(4) and 1.417(4) A are,
however, still similar to those of 1 and 4. Key distances and bond
angles for 7 are listed in Table 4.
Reactivity of 1–7 towards H2
enhanced signals for propene become visible in the associated 1H
NMR spectra.
3
When a toluene-d8 solution of Co(h -C3H5)(CO)2(PPh2Me) 1 is
placed under 3 atm of p-H2 at 295 K and monitored by 1H
NMR spectroscopy, the observation of p-H2 enhanced signals for
propane at d 0.9 (t, JHH = 7 Hz) and d 1.305 (doublet of antiphase
triplets where the separations are 7 and 22 Hz respectively) is
immediately apparent. On warming to 308 K additional p-H2
The observation of p-H2 enhanced signals for free propene
3
and propane requires that Co(h -C3H5)(CO)2(PPh2Me) 1 not only
adds H2, but that hydrogen transfer to the allyl functionality
follows such that two protons of a single p-H2 molecule become
located in the liberated ligand. In the case of the iridium analogue
2500 | Dalton Trans., 2009, 2496–2509
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
3
3
of Co(h -C3H5)(CO)2(PPh2Me), Ir(h -C3H5)(CO)(PPh3)2, the de-
3
tection of the H2 addition product Ir(h -C3H5)(CO)(PPh3)(H)2
was facilitated by p-H2 enhanced NMR methods.32 This sup-
3
ports the suggestion that the short lived intermediate Co(h -
C3H5)(CO)(PPh2Me)(H)2, or its analogous h -H2 form, is pro-
2
duced in this reaction, even though it is not directly observed.
It is also logical to assume that hydride transfer to form
Co(CO)2(PPh2Me)(propene)(H) takes place. This process places
one p-H2 derived proton on the methyl group, while leaving one
behind as a hydride ligand on the metal. In order to account for
the observation of p-H2 enhanced propene and propane signals,
this species must undergo, a further reversible, [1,3]-hydride shift
to exchange a non-enhanced proton of the propene ligand for the
remaining p-H2 derived hydride ligand. This process can occur
via hydride transfer to form a linear or branched alkyl as shown
1
in Scheme 3. The appearance of the H NMR signals for the
liberated propene should therefore reveal information about this
combination of steps.
Fig. 2 (a) 1H NMR spectrum of the vinyl region of p-H2 enhanced
propene at 308 K. (b) NMR simulation for p-H2 based protons at sites
b and c. (c) NMR simulation for p-H2 based protons at sites a and c.
(d) NMR simulation for p-H2 based protons at sites a and b. (e) NMR
simulation for p-H2 based protons at sites c and d. (f) NMR simulation for
p-H2 based protons at sites a and d. (g) NMR simulation for p-H2 based
protons at sites b and d.
is in a methyl group and the other in one of the vinyl sites. The
simulations suggest that in order to reproduce the appearance of
the alpha proton resonance shown in Fig. 2(a), traces b, c, e, f
and g need to be summed together in an approximate ratio of
1.75 : 2 : 1 : 1.75 : 1.
Scheme 3 Hydrogen transfer pathways indicated via the observed po-
3
larisation profiles for a route involving Co(h -C3H5)(CO)PR2R¢(H)2; this
3
would be equally valid for Co(h -C3H5)(CO)PR2R¢(H2).
The proton arrangement necessary to obtain trace e can be
obtained by migration to form a linear alkyl, followed by b-H
transfer back to the metal prior to alkene loss, while traces f and g
can be obtained from a branched alkyl followed by a b-H transfer
back to the metal. The first step in this sequence, the transfer of a
metal hydride to an allyl terminus has been observed previously.34
Traces b and c are obtained by placing a proton from p-H2 on
the central carbon with the second proton being located on one
of the geminal sites. This can be easily achieved if the original
Fig. 2(a) illustrates the vinyl region of such a 1H NMR spectrum,
with each of the three vinyl proton locations showing p-H2
enhanced features. The potential for hydride–hydrogen exchange
means that the two protons that were originally in p-H2 might find
themselves in a variety of positions in the propene product. The
experimental spectrum therefore corresponds to the sum of these
individual arrangements. Fig. 2(b)–(g) illustrate appropriate NMR
spectral simulations33 for the vinyl proton resonances of propene,
where the p-H2 derived protons are present in the indicated sites.
Collectively it can been seen that the spectral trace associated with
Fig. 2(a) corresponds to the summation of simulated traces b, c,
e, f and g. Trace b arises from the situation where the p-H2 based
protons are located cis to one another, trace c where they are trans,
and traces e, f and g are produced when one p-H2 derived proton
3
hydride transfer in Co(h -C3H5)(CO)(PPh2Me)(H)2 proceeds to
the central carbon atom of the allyl, otherwise it requires multiple
rearrangements. The missing permutation, where a and b (the two
geminal sites) are both occupied, requires multiple rearrangements
regardless of the site of the initial proton transfer and is therefore
less probable. The transfer of a hydride to the central carbon of an
allyl ligand has precedent.35
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 2496–2509 | 2501
©

View Article Online
Table 5 Temperatures where similar levels of propene and propane
and that Co(CO)2(PCy3)(propene)(H) is unable to re-form the
initial dihydride allyl complex which would lead, in turn, to
the reformation of 4 after H2 loss. It should be noted that
it has been suggested that such a process is reversible for
1
enhancements become visible in appropriate H NMR spectra that were
3
recorded to monitor the reaction of Co(h -C3H5)(CO)2(PR2R¢) with H2
Propane appearance
temperature/K
Propene appearance
temperature/K
=
Complex (PR2R¢)
Co(P(OMe)3)2(H)(CH2 CHCH2Ph) and that a dihydride form is
accessed.40
1 (PPh2Me)
2 (PMe2Ph)
3 (P(CH2Ph)3)
4 (PCy3)
295
303
318
333
323
313
318
308
308
323
333
328
318
323
Reactivity of complexes 1–5 towards CO
5 (PPh3)
Prior to examining the hydroformylation activity of these sys-
tems, a series of control experiments were undertaken where
a number of toluene-d8 based samples were prepared for each
of the complexes and exposed to an atmosphere of CO. When
6 (dppm)
7 (dppe)
1
Only at temperatures above 308 K does propene liberation
from Co(CO)2(PPh2Me)(propene)(H) become competitive with
propane formation. There is, however, a substantial difference
in appearance of the signals for propene at temperatures above
373 K. This change is consistent with an increase in the degree of
hydride transfer to the central carbon.
The trapping of either of the proposed 16-electron
cobalt alkyl intermediates, Co(CH2CH2CH3)(CO)2(PPh2Me) or
Co(CH2(CH3)2)(CO)2(PPh2Me), that are formed by the second
[1,3]-hydride shift, with a second H2 molecule and subsequent
reductive elimination of the alkane accounts for the enhanced
propane resonances that are seen in these NMR spectra.
the solution of 1 was examined, 31P{ H} NMR spectroscopy
revealed the liberation of PPh2Me, while the associated 1H NMR
3
spectrum confirmed the formation of a new h -allyl complex. This
3
complex was subsequently identified as Co(h -C3H5)(CO)3.25,41
Similar results were obtained for all the phosphines except
PMe2Ph which did not react at room temperature. Interestingly,
1
no evidence for the insertion of CO into an h -allyl ligand was
evident in any of the NMR spectra recorded between 213 K or
383 K. This contrasts with the situation that has been reported
3
for Ir(CO)(PPh3)2(h -C3H5), where the addition of CO leads
3
1
to the detection of Ir(CO)2(PPh3)(h -C3H5), Ir(CO)3(PPh3)(h -
=
=
CH2CH CH2), Ir(CO)2(PPh3)2(COCH2CH CH2) and Ir(CO)3-
42
=
Analogous observations were made when the addition of H2
to complexes 2–7 was studied in a similar way. In fact, the
appearance of the propene 1H NMR signals proved to be similar
regardless of the identity of the phosphine. The only significant
difference observed in the reactivity of complexes 1–7 towards
H2 stems from the temperature needed to see enhanced propene
signals. Table 5 summarizes this information as a function of the
complex. In all cases, the observation of p-H2 enhanced signals
for propane occurred at lower temperatures than those required
for the observation of propene. This confirms that the addition
of H2 to the alkyl complex, formed by hydride migration onto
the bound propene, and the associated reductive elimination of
propane, proceeds more rapidly than the combination of the two
steps, b-H transfer and alkene liberation, which are necessary to
see enhanced signals for free propene. In addition, these data
(PPh3)(COCH2 CHCH2).
These data do, however, confirm that under an atmosphere
of CO, replacement of the phosphine by CO is possible. This
observation indicates that phosphine-modified cobalt catalysts
will produce unmodified species upon reaction with CO. Ligand
exchange between CO and phosphite or phosphines in such allyl
systems has previously been reported.43 In order to compare the
ability of these systems to undergo phosphine replacement by CO
a number of samples were prepared which contained a mixture of
two of the complexes and 3 atm of CO at 295 K. Integration of
the corresponding 31P signals enabled the relative stability of the
complexes to phosphine loss to be assessed. For 1 (PPh2Me) and
4 (PCy3) the corresponding ratio proved to be 4.03 : 1 while for 1
and 3 it increased to 8.33 : 1, and for 1 and 5 it was 10.53 : 1. The
PMe2Ph system, 2, (PMe2Ph) proved to be stable to phosphine
loss at 295 K and 3 atm CO. It can therefore be concluded that the
3
also suggest that Co(h -C3H5)(CO)2(PCy3) 4 undergoes the least
3
effective allyl hydrogenation.
relative ability of Co(h -C3H5)(CO)2(PR2R¢) to lose phosphine
The production of propene and propane must coincide with
the destruction of the starting complex Co(h -C3H5)(CO)2PR2R¢.
Consequently it is not surprising that in the associated NMR
spectra, hydride resonances indicative of the formation of
Co(H)(CO)3(PR2R¢) and Co(H)(CO)2(PR2R¢)2 were also seen at
the end of the experiments.36
follows the trend PMe2Ph < PPh2Me < PCy3 < P(CH2Ph)3 <
PPh3. These changes occur very slowly and were monitored over
one week.
3
Reactivity of complexes 1–7 towards CO and H2
The formation of an allyl hydride complex from a metal
When a toluene-d8 solution of 1 was placed under 3 atm of a 2 : 1
mixture of CO and H2 no reaction was evident at 313 K. Upon
warming to 373 K, the corresponding 1H NMR spectrum showed
much weaker p-H2 enhanced signals for propene and propane than
when such a sample was under 3 atm of p-H2 alone and at 313 K.
In addition, five new p-H2 enhanced signals were also detected in
clear regions of the spectrum at d 3.40, d 3.10, d 1.61, d 1.13,
d 0.86, d 0.80 and d 0.72 (Fig. 3). The corresponding COSY
spectrum connected the d 3.38 signal (t, JHH = 7 Hz) to those at
d 1.61 (sextet, JHH = 7 Hz) and d 1.13 (t, JHH = 7 Hz), confirming
that a linear CH2CH2CH3 group has been formed with protons
2
propene complex is exemplified by the chemistry of W(CO)4(h -
2
2
C3H6)2,37 Fe(CO)4(h -C3H6)38 and Cp*Re(CO)2(h -C3H6).39 In
order to test whether the first hydride migration to the allyl
3
ligand in Co(h -C3H5)(CO)(PR2R¢)(H)2 is reversible, a sample
of 4 was prepared and placed under a 1 : 2 mixture of propene
and p-H2. 1H NMR spectra recorded at 363 K revealed en-
hanced signals for propane, but complex 4 still proved to be
rapidly consumed and the formation of the known complexes
Co(H)(CO)3(PCy3) and Co(H)(CO)2(PCy3)2 indicated. This sug-
gests that the corresponding intermediates are not stabilized
2502 | Dalton Trans., 2009, 2496–2509
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
from p-H2 being placed on all three of the carbon centres. This
spectrum also confirmed that the additional H signal at d 3.40
(septet, JHH = 8 Hz) coupled to a signal at d 1.13 (d, JHH = 8 Hz)
which indicated that a species containing a CH(CH3)2 group was
also formed.
from 13C signals that appeared at d 234.5 and d 239.8 respectively.
Since both the latter resonances exhibited doublet multiplicities
due to 31P–13C couplings (JPC = 51 and 44 Hz respectively), their
origin as metal acyl groups, trans to a single phosphine ligand, was
confirmed. The corresponding complexes are therefore the linear
acyl containing species Co(COCH2CH2CH3)(CO)3(PPh2Me) (1a)
and the branched complex Co(COCH(CH3)2)(CO)3(PPh2Me) (1b)
with structures shown in Scheme 4. In view of the fact that 1a and
1b contain PPh2Me, it is possible to conclude that 1 reacts initially
1
3
with H2 via CO loss and that Co(h -C3H5)(CO)(PR2R¢)(H)2 does
indeed correspond to the initially formed dihydride complex as
suggested earlier. NMR data for these species can be found in
Table 6.
Fig. 3 1H spectrum of a toluene-d8 solution of 1 at 373 K in the presence
of a mixture of H2 and CO.
Scheme
4 Structural representation of Co(COCH2CH2CH3)(CO)3-
(PPh2Me) (1a) and the branched complex Co(COCH(CH3)2)(CO)3-
(PPh2Me) (1b) detected when a toluene solution of 1 is warmed with
CO and p-H2.
When this reaction was repeated with 13CO, both of the terminal
alkyl proton signals showed additional 13C splittings that arose
Table 6 Selected NMR data for the acyl complexes 1a–7a and 1b–7b
Linear acyls (a) Co(COCH2CH2CH3)(CO)3(PR2R¢)
Branched acyls (b) Co(COCH(CH3)2)(CO)3(PR2R¢)
1
Precursor
1H
13COacyl
31P{ H}
1H
13COacyl
a
1
3.10 (t), CH2 , JHH = 7 Hz
1.61 (sextet), CH2b, JHH = 7 Hz
0.87 (t), CH3, JHH = 7 Hz
234.5 (d), JPC = 51 Hz
32.4 (s)
3.40 (septet), CH, JHH = 7 Hz
1.14 (d), JHH = 7 Hz
239.8 (d), JPC = 44 Hz
a
2
3
4
5
6
7
3.06 (t), CH2 , JHH = 7 Hz
—
—
3.34 (septet), CH, JHH = 7 Hz
1.13 (d), JHH = 7 Hz
—
1.60 (sextet), CH2b, JHH = 7 Hz
0.86 (t), CH3, JHH = 7 Hz
a
2.99 (t), CH2 , JHH = 7 Hz
234.8 (d), JPC = 45 Hz
235.6 (d), JPC = 39Hz
234.1 (d), JPC = 57 Hz
234.1 (d), JPC = 41 Hz
234.05 (d), JPC = 35 Hz
50.5 (s)
62.0 (s)
50.1 (s)
—
3.27 (septet), CH, JHH = 7 Hz
1.07 (d), JHH = 7 Hz
239.9 (d), JPC = 43 Hz
240.9 (d), JPC = 40.5 Hz
239.4 (d), JPC = 46 Hz
239.2 (d), JPC = 34 Hz
1.56 (sextet), CH2b, JHH = 7 Hz
0.83 (t), CH3, JHH = 7 Hz
a
3.22 (t), CH2 , JHH = 7 Hz
3.37 (septet), CH, JHH = 7 Hz
1.20 (d), JHH = 7 Hz
1.67 (sextet), CH2b, JHH = 7 Hz
0.92 (t), CH3, JHH = 7 Hz
a
3.18 (t) CH2 , JHH = 7 Hz
3.50 (septet), CH, JHH = 7 Hz
1.17 (d), JHH = 7 Hz
1.64 (sextet), CH2b, JHH = 7 Hz
0.87 (t), CH3, JHH = 7 Hz
a
3.07 (t), CH2 , JHH = 6.5 Hz
3.57 (septet), CH, JHH = 7 Hz
1.08 (d), JHH = 7 Hz
1.58 (sextet), CH2b, JHH = 7 Hz
0.82 (t), CH3, JHH = 7 Hz
a
3.13 (t), CH2 , JHH = 6 Hz
—
3.43 (septet), CH, JHH = 6.5 Hz
1.135 (d), JHH = 7 Hz
239.0 (d)
JPC = 41 Hz
1.61 (sextet), CH2b, JHH = 6.5 Hz
0.85 (t), CH3, JHH = 7 Hz
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 2496–2509 | 2503
©

View Article Online
The original studies by Heck and Breslow led to an accepted
mechanism for hydroformylation involving a mono-hydride pre-
cursor which binds the alkene prior to forming a metal alkyl
which goes on to form a metal acyl complex that eliminates
the aldehyde after H2 addition.3 We have already described
in this manuscript how 1 reacts with p-H2 to form propene
via a metal–propene–hydride. The observation of 1a and 1b
as p-H2 enhanced products can therefore be accounted for
by the trapping of the corresponding 16-electron intermedi-
ates, Co(CO)2(PR2R¢)(C(O)CH2CH2CH3) and Co(CO)2(PR2R¢)
(COCH(CH3)2) with CO rather than H2. This information is
summarised in Scheme 5. The protons that are incorporated in
these products from a single p-H2 molecule are shown in red.
These protons move around the indicated molecular framework
according to the processes illustrated in Scheme 5. We note that
the enhanced signals seen in the NMR spectra for molecules
possessing the linear site B arrangement in product 1a can be
produced directly from the allyl dihydride complex while molecules
corresponding to the arrangement necessary for linear site A in 1a
arise via a re-linearised branched intermediate; this intermediate
reacts competitively with this process to form 1b.
CH(O)CH2CH2CH3 proved to be d 0.7 (t, JHH = 7 Hz), d 1.31
(m), d 1.76 (dt, JHH = 2 and 7 Hz), d 9.29 (t, 2 Hz) while those
for the branched aldehyde CH(O)CH(CH3)2 appeared at 0.77 (d,
JHH = 7 Hz), d 1.62 (m) and d 9.24 (d, 1 Hz).
1
When the H NMR spectra on this sample were recorded at
373 K, additional weak para-hydrogen enhanced signals were seen
at d 4.17 (dd, JHH = 8 and 4 Hz, 3.99 (d, JHH = 6 Hz), 3.92 (dd,
JHH = 7 and 6 Hz), and 3.73 (d, JHH = 6 Hz) (Fig. 4). These peaks
all appear with antiphase character, and occur in a region that is
characteristic of metal–alkene 1H NMR signals.44 When a COSY
spectrum was recorded, the strongest of these peaks, resonating at
d 4.17, proved to be coupled to three other resonances at d 1.85,
1.65 and 1.011. Examination of the multiplicity of these resonances
confirmed that the unique proton yielding the peak at d 4.17 is a to
the methyl of the bound propene which resonates at d 1.011, while
the peaks at d 1.85 and 1.65 correspond to protons that are trans
and cis to this group respectively. This information conclusively
demonstrates that a coupled set of resonances corresponding to a
bound propene ligand has been observed. Additional correlations
1
were visible in such spectra between the H resonance at d 3.92
1
and a new peak at d 1.68, and between the H resonance at d
3.99 and a new peak at d 1.82. The signal to noise ratio for these
signals, however, proved to be insufficient to locate the additional
resonances necessary for the full characterization of the propene
ligands.
Fig. 4 Additional signals observed in the 1H NMR spectrum of 1 in the
presence of H2 and CO at 373 K.
The peaks at d 3.92 (dd, JHH = 7 and 6 Hz) and 3.73 (d, JHH
=
6 Hz) showed a similar dramatic increase in size on warming to
383 K, while the two peaks at d 4.17 and d 3.99 decreased in
intensity by a similar amount. These peaks therefore appear to
belong to four distinct propene complexes. Unfortunately, when
samples containing 13CO labelled materials were employed, none
of these proton signals proved to connect to either a CoCOR
resonance or a CoCO resonance in the corresponding 1H–13C
HMQC experiment. Nonetheless, we suggest that the trapping
of the 16-electron fragments Co(COCH2CH2CH3)(CO)2(PPh2Me)
and Co(COCH(CH3)2)(CO)2(PPh2Me) required to form 1a and
1b, by (PHIP enhanced) propene rather than CO accounts for the
observation of these species. The signals therefore most likely arise
from isomers of Co(COCH2CH2CH3)(CO)2(PPh2Me)(propene)
and Co(COCH(CH3)2)(CO)2(PPh2Me)(propene) such as those
shown in Scheme 6 which might be differentiated by the relative
orientation of the alkene.
Scheme 5 Transformations accounting for the p-H2 enhanced NMR
signals seen at sites where the protons appear in red.
In these 1H NMR spectra, p-H2 enhanced resonances are
also visible at d 0.82 and d 0.72 for the free aldehydes
CH(O)CH(CH3)2 and CH(O)CH2CH2CH3 respectively. The room
temperature proton resonance positions for the linear aldehyde
When the reactions of the corresponding complexes 2–7 with
p-H2 and CO are observed by NMR spectroscopy the analogous
2504 | Dalton Trans., 2009, 2496–2509
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Reactivity of 1 towards H2 and d10-diphenylacetylene
In this experiment, we aimed to form relatively high concen-
trations of p-H2 enhanced cis-stilbene which might in turn
replace the propene ligand in the previously proposed species
Co(CO)(PPh2Me)(H)(propene). This approach has been used
previously to detect a palladium alkyl cation during studies
involving diphenylacetylene and para-hydrogen.22,46 An NMR
sample was therefore prepared that contained a 100-fold excess
of d10-diphenylacetylene relative to 1, and the reaction with p-H2
Scheme 6 Additional species detected during reaction of 1 with p-H2
and CO at 373 K: Co(COCH2CH2CH3)(CO)2(PPh2Me)(propene) and
Co(COCH(CH3)2)(CO)2(PPh2Me)(propene). Only one isomer of each is
illustrated.
∫
examined at 313 K. d10-PhC CPh was indeed hydrogenated under
these reaction conditions, and the formation of para-hydrogen
enhanced cis-stilbene indicated by the observation of an emission
signal at d 6.47. In addition, new para-hydrogen enhanced proton
signals were observed at d 3.51, d 3.33 and d 3.30. These three
resonances proved to be coupled in a COSY experiment, and they
all showed a further31P splitting. A 1H–31P-HMQC data set located
products to 1a and 1b were observed in each case. The NMR
characteristics of these materials are listed in Table 6. Resonances
of the type indicated for bound propene ligands were only
visible in NMR spectra obtained for samples of 1, 2 and 6. It
should further be noted that when samples of 5 were cooled
back to room temperature at the end of the experiments, crys-
tals of [Co(CO)3PPh3]2 consistently precipitated from solution.
The structure of [Co(CO)3PPh3]2 is illustrated in Fig. 5, and
appropriate X-ray data listed in Table 7. Complexes of this
type have been proposed to react with H2 to form the active
precatalyst HCo(CO)3(PR3) necessary for modified cobalt based
hydroformylation catalysis.45
the phosphorus resonance causing these splittings at d 39.1.
13
When mono- C labelled d10-diphenylacetylene, d10-PhC 13CPh
∫
was utilized, the first two proton resonances coupled to a single
13C NMR signal at d 40.02, which appeared as a broad singlet. The
third 1H NMR signal which resonates at d 3.30 coupled to a second
13C resonance, now at d 48.5, which appears as a phosphorus
coupled doublet where JPC = 19 Hz. The formation of an alkyl
group bound directly to cobalt is therefore indicated. When the
same reaction was examined with a cobalt precursor labelled with
13CO, no additional splittings were seen in any of these para-
hydrogen enhanced signals, and the terminal CH2 group failed to
couple to an acyl 13C signal, as would have been expected if this was
a CO insertion product such as those described earlier in this paper.
These results therefore suggest that the large excess of d10-diphenyl-
acetylene facilitates the formation of cis-stilbene which is trapped
by the CoH(CO)2(PPh2Me) intermediate liberated by the forma-
tion of propane. The product that is detected therefore corresponds
to Co(CO)3(PPh2Me)(CHPhCH2Ph) (Scheme 7).
◦
˚
Table 7 Selected bond lengths (A) and angles ( ) for [Co(CO)3PPh3]2
Assignment
Value
Co(1)–Co(1)#1a
Co(1)–P(1)
Co(1–C(1)
2.6506(9)
2.1895(9)
1.7845(16)
1.1467(19)
C(1)–O(1)
C(1)–Co(1)–P(1)
O(1)–C(1)–Co(1)
C(1)–Co(1)–C(1)#2b
94.25(5)
178.83(13)
119.457(12)
a#1 = inversion centred on midpoint of Co–Co bond (11/3 - x, 2/3 - y,
2/3 - z). b#2 = 3-fold rotation about Co–Co axis, (1 - y, x - y, z).
Scheme 7 Additional species detected during the reaction of 1 with p-H2
and d10-diphenylacetylene at 313 K.
Under these conditions we also see a new p-H2 enhanced peak
at d 0.84 which couples to two other protons and appears with a
chemical shift which is very close to that of the methyl resonance of
free propane. In the COSY spectrum this peak proved to couple to
additional resonances at d 1.59 and d 3.09. Of these two signals,
the former resonance is para-hydrogen enhanced, and the peak
multiplicities of the PHIP enhanced signals indicate the presence
of couplings to 2 protons and 5 protons respectively. Since the d
3.09 resonance does not appear in the one dimensional 1H NMR
spectrum it can be concluded that the associated protons located
in this site do not arise from para-hydrogen. These data confirm
however the formation of a linear alkyl group, with the lower
Fig. 5 ORTEP diagram of [Co(CO)3PPh3]2. Ellipsoids drawn at 50%
probability level. The asymmetric unit corresponded to 1/6th of the
¯
molecule as labelled. The R3 space-group defines the rest of the molecule
(inversion centre located at the midpoint of the Co–Co bond and 3-fold
rotation about the Co–Co axis).
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 2496–2509 | 2505
©

View Article Online
temperatures used here facilitating its detection. The chemical
shown for the linear acyl containing species 4a [( ) and ( )] refer to
signals obtained from the p-H2 enhanced proton NMR resonances
at the COCH2 site and the terminal methyl site respectively; the
resonance for the connecting CH2 site appeared in a congested
region of the spectrum and could only be located by COSY
spectroscopy. The terminal methyl resonance seen for the branched
acyl containing species 4b was used to obtain the data represented
by the ( ) label. para-Hydrogen enhanced signals were also used
to probe the branched to linear ( ) free aldehyde composition at
very early stages in the reaction. In order to establish that the p-H2
based signal strength trends are meaningful, we also measured
the linear/branched aldehyde ratio at the end of the experiment
when the associated signals arise from “normal” magnetisation.
These data are indicated with the ( ) label. It can be seen that the
ratio of free linear to branched aldehyde follows the same trend
as that obtained from p-H2 enhanced signals that are seen at early
stages in the reaction. This validates the approach and suggests
meaningful trends might be observed when the intensities of p-H2
enhanced signals are compared.
shift of the CH2 moiety at d 3.09 however suggests the trapping
∫
of Co(CO)2(PPh2Me)(COCH2CH2CH3) by PhC CPh, the ligand
in solution that has the highest concentration, and hence the
∫
detection of Co(CO)2(PPh2Me)(COCH2CH2CH3)(PhC CPh).
In view of the high cost of the 13C labelled d10-diphenylacetylene,
when this experiment was repeated with the labelled precursor,
only a 10-fold excess of the substrate relative to metal was used.
1
The H NMR resonance at d 3.09 now proved to be enhanced
in accordance with more extensive hydrogen scrambling. This
suggests that when a large excess of substrate is employed, the
ability of the system to undergo the necessary proton exchange to
see enhancement is substantially reduced.
When
a sample of 1 was shaken with CO, H2 and
d10-diphenylacetylene, these extra p-H2 enhanced species are
produced in much smaller amounts than when p-H2 and
d10-diphenylacetylene alone are employed. The reaction with
CO therefore now competes effectively with hydrogenation and
reduces the amounts of these species. As a consequence, we were
unable to obtain more definitive data and thereby confirm their
identity.
The relative signal strengths for the COCH2 group ( ) and the
terminal methyl group ( ) of 4a become equal at temperatures
greater than 377 K. This suggests that the p-H2 label is now
randomly distributed within the corresponding acyl group and
hence indicates that the proton exchange pathways shown in
Scheme 5 should be facile.
Catalytic studies on the reactivity of complexes 1–8 towards
CO and H2
In order to compare the reactivity of these systems under
hydroformylation conditions, we have undertaken a number of
studies using constant concentrations of the catalyst precursor
and a CO : H2 ratio of 2 : 1.
Roe has used high pressure NMR methods to demonstrate
that the thermal loss of CO from CH3C(O)Co(CO)4 occurs
at a rate of 11.5 s-1 at 353 K.8 This rate would be sufficient
to see magnetisation transfer from such a species into the
free aldehyde by NMR spectroscopy. A series of 1D EXSY
experiments were therefore undertaken to test for magnetization
transfer from 4a and 4b into the corresponding aldehydes. None
was observed. This result indicates that the electron donating
properties of the phosphine dramatically stabilise 4a with respect
to CO loss when compared to the unmodified form. Further-
more, it confirms that the enhanced NMR signals seen for the
free aldehydes arise from the trapping of the corresponding
16-electron intermediates Co(COCH2CH2CH3)(CO)2(PCy3) and
Co(COCH(CH3)2)(CO)2(PCy3) with H2 (which leads to rapid
aldehyde elimination) rather than CO (which leads to 4a or 4b).
The variation in the observed aldehyde product ratios illustrated
in Fig. 6 therefore reflects higher turnovers for the linear product
at higher reaction temperatures. Logically, this effect might
correspond to a depletion in the amount of 4a that is detected
which in turn will reduce the observed ratio of 4a : 4b. This effect
is supported by the data presented in both Fig. 6 and 7 where
the final aldehyde ratios are plotted as a function of temperature
with the PCy3 data appearing as a and showing an increase with
temperature as predicted.
1
For the PCy3 precursor, 4, the relative size of the observed H
NMR signals due to the aldehydes, propene and propane, relative
to those of the solvent, consistently increased with an increase in
temperature. There was, however, an increased preference for the
formation of the linear aldehyde at higher temperatures. Fig. 6
summarizes these variations. The two distinct points of reference
Fig. 7 also demonstrates that the PMe2Ph and PPh2Me ligand
systems show a similar increase in selectivity for the linear product
with increase in reaction temperature. In contrast, however, the
PPh3, dppe and dppm systems show a fall in selectivity over
the same temperature range. This situation is complicated for the
activating P(Ch2Ph)3 phosphine where the selectivity initially falls
and then increases with temperature. This observation actually
reflects the fact that the P(Ch2Ph)3 system produces secondary
oligomerization products at higher temperatures and consequently
the full product distribution is not accounted for in this plot.47
Fig. 6 Plot of the p-H2 based NMR signal intensities for species 4a,
4b and the corresponding aldehydes as a function of temperature for
samples of 1 reacting in toluene with 3 atm of a 2 : 1 mixture of CO and
H2 normalised relative to free branched aldehyde. (Note, linear aldehyde
(n) ( ) and branched aldehyde ( ) data obtained from 1H signals measured
using normal methods when the thermally polarized states are visible).
2506 | Dalton Trans., 2009, 2496–2509
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Fig. 7 Final free linear/branched aldehyde product ratios observed as a
function of temperature and phosphine for an initial 2 : 1 ratio of CO to
H2. Phosphines identified in key.
Fig. 9 Ratio of the normalised 1H signal intensities for linear site B
to linear site A in complexes 1a–7a as a function of temperature and
phosphine for a CO : H2 ratio of 2 : 1.
Nonetheless, it is clear thatthe phosphine plays aroleincontrolling
this reaction selectively.
clearly seen in the case of the dppm system where the ratio starts
off at 1.7 : 1 and falls towards 1 : 1 with increase in temperature.
Consequently, there is therefore an increase in hydride transfer into
the branched site with increase in temperature which corresponds
with the fall in reaction selectivity seen for this phosphine. For the
PCy3 system, however, the initial ratio is less than one, and tends
towards this value as the temperature is raised. This suggests that
the propensity for hydride transfer into the terminal site increases
with temperature and matches the trend for increased selectivity
for linear aldehyde with increase in temperature.
Fig. 8 shows how the ratio of the PHIP enhanced signals
for the linear to branched intermediates (a) and (b) vary with
temperature. Overall, the highest selectivity for the linear product
Co(COCH2CH2CH3)(CO)2(PR3) is found with PCy3 as might be
expected for the largest phosphine. These data also reveal that the
linear CO addition products, Co(COCH2CH2CH3)(CO)2(PR2R¢),
are not always favoured.
For the other phosphines, there is a slight increase in this ratio
with temperature. This suggests that there is essentially a random
distribution of the para-hydrogen label to start and hence the
ratio of the corresponding linear and branched intermediates is
initially under thermodynamic control. The reactions observed
selectivity should therefore reflect directly the ratio of these species
in solution rather than the rate constants for hydride ligand
transfer to propene as suggested for the dppm system. When the
ratio increases with temperature, it would therefore appear that
CO insertion now competes more effectively with hydrogen label
scrambling.
In order to explore how the relative ratios of H2 and CO affect
these reactions, we recorded a number of experiments at 373 K for
the different precursors where the total pressure was maintained
at 3 atm, but the CO to H2 ratio was changed from 4 : 1, through
1 : 1 and 2 : 1 to 3 : 1. Fig. 10 shows the results of this process
and reveals that the reaction final product distribution shows a
Fig. 8 Ratios of the observed PHIP enhanced 1H NMR signals for
linear cobalt product (a) to branched cobalt product (b) as a function
of temperature for an initial 2 : 1 ratio of CO to H2.
In addition, these data reveal for the PMe2Ph and PPh2Me
systems that the ratio of the corresponding linear and branched
intermediates hardly changes over the temperature range 368–
393 K. In contrast, there is a much higher change in the free
aldehyde product ratio. These data must therefore mask a further
trend. This process is revealed in Fig. 9, which shows how the ratio
of the normalised signal strengths for the intensities of site B (the
CH3 group) and site A (the COCH2 group as indicated in Scheme
5) of the linear species Co(COCH2CH2CH3)(CO)2(PR2R¢) vary
with phosphine. This plot therefore provides an insight into both
the selectivity of the hydride transfer step and its reversibility. For
example if a random arrangement of the p-H2 label were to result
the ratio would be of 1 : 1. As the temperature increases, it might
be expected that the reversible proton transfers of Scheme 5 get
faster and hence the observed ratio tends to 1 : 1. This trend is most
Fig. 10 Effect of the CO : H2 ratio on the free aldehyde product
distribution as a function of phosphine for a series of reactions that were
completed at 373 K.
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 2496–2509 | 2507
©

View Article Online
complex dependence on the CO to H2 ratio. The 1 : 2 ratio shows
the highest selectivity for the formation of linear aldehyde in the
case of both the PCy3 and PPh3 based systems. In contrast, both
the PMe2Ph and PPh2Me systems show a mutually similar trend
where the selectivity rises with CO composition. The dppe and
dppm systems also result in higher selectivity for linear aldehyde
at higher CO ratios, but their reactions are far more sensitive to this
change than either the PMe2Ph or PPh2Me systems. We cannot
however exclude potential variations in the contributions from
chelated and non-chelated species.
about how the relative rates of key steps leading to linear and
branched hydroformylation products compete for the different
phosphines. While the PPh2Me system yields one of the largest
differences in the ratio of the signal strengths seen for the
aldehyde products, the corresponding difference in signal ratios
for the intermediates Co(COCH2CH2CH3)(CO)3(PPh2Me) and
Co(COCH(CH3)2)(CO)3(PPh2Me) is close to one. This suggests
that there is less of the Co(COCH2CH2CH3)(CO)3(PPh2Me) form
present because more of it is converted into the aldehyde via H2
addition to Co(COCH2CH2CH3)(CO)2(PPh2Me) rather than CO
coordination. Furthermore, the ratio of the signal intensities of
the linear site B to that of A for PPh2Me is always greater than one
and increases with increasing temperature. These data therefore
reveal that the degree of alkyl/alkenehydride interconversion falls
with increase in temperatures and hence confirm that productive
aldehyde formation competes with alkyl group isomerisation.
Consequently it can be seen that at low temperatures, the
reaction selectivity is best viewed on the basis of the initial
kinetic preference for hydride insertion to form a branched
intermediate over that which leads to the linear analogue for
dppm. At higher temperatures, there is a rapid equilibration
between these intermediates as evidenced by the fact that the para-
hydrogen label is randomly arranged between sites A and B. This
suggests that their ratio reaches a level which is controlled by their
thermodynamic stability. The ensuing reactions selectivity must
then reflect the relative rates of H2 addition to the corresponding
16-electron intermediates.
Conclusions
3
The syntheses of Co(h -C3H5)(CO)2PR2R¢ (R, R¢ = Ph, Me; R,
3
R¢ = Me, Ph; R = R¢ = Ph, Cy, CH2Ph) and Co(h -C3H5)(CO)(L)
(L = dmpe and dppe) are described, and X-ray structures
3
for Co(h -C3H5)(CO)(dppe) and the PPh2Me, PCy3 derivatives
3
reported. The relative ability of Co(h -C3H5)(CO)2(PR2R¢) to
exchange phosphine for CO follows the trend PMe2Ph < PPh2Me
< PCy3 < P(CH2Ph)3 < PPh3. According to Tolman,30 the basicity
of these ligands follows the order PCy3 > PMe2Ph > P(CH2Ph)3
> PPh2Me > PPh3 while their steric effect increases according to
PMe2Ph < PPh2Me < PPh3 < P(CH2Ph)3 < PCy3. Consequently
we can deduce that the ability of 1–6 to undergo phosphine
exchange is highly influenced by the steric effect of the phosphine.
Reactions of these allyl complexes with para-hydrogen (p-H2)
lead to the observation of para-hydrogen induced polarisation in
both liberated propene and propane. The temperature at which
these products are first seen varies with the phosphine according
to the listing PPh2Me < PMe2Ph < dppm < dppe ~ P(CH2Ph)3
< PPh3 < PCy3. These reactions proceed as a consequence of CO
loss. Consequently, the better the electron donating properties of
the phosphine, the stronger the CoCO bond; this accounts for the
higher activation temperatures of the PCy3 system, but given the
fact PPh3 containing system also requires a high temperature steric
effects must also play their role on this reaction.
The reaction of the allyl complexes with both CO and
H2 leads to the detection of linear acyl containing species
Co(COCH2CH2CH3)(CO)3(PR2R¢) and branched acyl com-
plexes Co(COCH(CH3)2)(CO)3(PR2R¢) via the PHIP effect. In
the case of PPh2Me, additional signals that have been at-
tributed to Co(COCH2CH2CH3)(CO)2(PPh2Me)(propene) and
Co(COCH(CH3)2)(CO)2(PPh2Me)(propene). Similar signals were
also seen during studies on the PMe2Ph and dppm systems.
This information suggests that when the phosphine is elec-
tron rich, and relatively small, the liberated propene can bind
sufficiently well as to allow the detection of these species.
Furthermore, when the reactions of H2 and diphenylacetylene
are studied with the PPh2Me based precursor, the detection of
Co(CO)3(PPh2Me)(CHPhCH2Ph) rather than its acyl form is
achieved. This must reflect the stronger M–CHPh- bond strength
when compared to M–CH2- which acts to reduce the rate of acyl
intermediate formation.
Acknowledgements
SBD and CG are grateful to Sasol Technology UK Ltd. for
support. The EPSRC is acknowledged for providing funding for
the diffractometer.
References
1 (a) For example see: G. W. Parshall and S. D. Ittel, Homogeneous
Catalysis, Wiley, New York, USA, 1992; (b) B. Breit, Acc. Chem. Res.,
2003, 36, 264.
2 B. Cornils and W. A. Hermann, Applied homogeneous catalysis with
organometallic compounds: a comprehensive handbook in 2 volumes,
VCH, Weinheim, New York, USA, 1996.
3 R. F. Heck and D. S. Breslow, J. Am. Chem. Soc., 1962, 84, 2499.
4 W. Hieber and E. Lindner, Chem. Ber., 1960, 95, 2042.
5 R. F. Heck and D. S. Breslow, J. Am. Chem. Soc., 1962, 84, 2499.
6 V. Galamb, G. Palyi, R. Boese and G. Schmid, Organometallics, 1987,
6, 861.
7 R. L. Sweany, Organometallics, 1989, 8, 175.
8 D. Roe, Organometallics, 1987, 6, 942.
9 S. M. Assick, J. G. Rabor, S. Elbers, J. Marhenke, S. Bernhard, J. R.
Schoonover and P. C. Ford, Inorg. Chem., 2000, 39, 3098.
10 L. Rosi, A. Salvini, M. Bianchi, P. Frediani and F. Piacenti, J.
Organomet. Chem., 1997, 535, 143.
11 C.-H. Hua, Y.-W. Li, M. Beller and H. Jiao, Organometallics, 2003, 22,
4665.
12 L. Versluis and T. Ziegler, Organometallics, 1990, 9, 2985.
13 M. Torrent, M. Sola and G. Frenking, Chem. Rev., 2000, 100, 439.
14 R. F. Heck and D. S. Breslow, J. Am. Chem. Soc., 1961, 83, 1097.
15 R. F. Heck, J. Am. Chem. Soc., 1968, 90, 317.
16 U. Schubert and J. Muller, J. Organomet. Chem., 1988, 340, 101.
17 G. F. Pregaglia, A. Andreetta, G. F. Ferrari and G. Montrasi, J.
Organomet. Chem., 1971, 33, 73.
18 E. L. Muetterties and J. R. Bleeke, Acc. Chem. Res., 1979, 12, 324.
19 C. R. Bowers and D. P. Weitekamp, Phys. Rev. Lett., 1986, 57, 2645.
20 (a) C. R. Bowers and D. P. Weitekamp, J. Am. Chem. Soc., 1987,
109, 5541; (b) T. C. Eisenschmid, R. U. Kirss, P. A. Deutsch, S. I.
Studies on how the appearance, and ratio, of the PHIP enhanced
signals seen in these reaction vary as a function of reaction
temperature and H2 to CO ratio have also been described. These
data revealed that the PMe2Ph and PPh2Me based systems yield
the highest selectivity for the linear hydroformylation products.
The PHIP enhanced signal profiles also provide information
2508 | Dalton Trans., 2009, 2496–2509
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Hommeltoft, R. Eisenberg, J. Bargon, R. G. Lawler and A. L. Balch,
J. Am. Chem. Soc., 1987, 109, 8089.
21 (a) J. Natterer and J. Bargon, Prog. Nucl. Magn. Reson. Spectrosc.,
1997, 31, 293; (b) S. B. Duckett and C. J. Sleigh, Prog. Nucl. Magn.
Reson. Spectrosc., 1999, 34, 71; (c) S. B. Duckett and D. Blazina, Eur.
J. Inorg. Chem., 2003, 2901; (d) D. Blazina, S. B. Duckett, J. P. Dunne
and C. Godard, Dalton Trans., 2004, 2601.
22 J. Lo´pez-Serrano, S. B. Duckett, S. Aiken, K. Q. Almeida Len˜ero, E.
Drent, J. P. Dunne, D. Konya and A. C. Whitwood, J. Am. Chem. Soc.,
2007, 129, 6513.
23 C. Godard, S. B. Duckett, S. Aiken, S. Polas, R. Tooze and A.
Whitwood, J. Am. Chem. Soc., 2005, 127, 4994.
32 C. Godard, S. B. Duckett, C. Henry, S. Polas, R. Toose and A. C.
Whitwood, Chem. Commun., 2004, 1826.
33 P. H. M. Budzelaar, g-NMR; version 5, Cherwell Scientific Publishing
Limited, Oxford, UK, 1995–97.
34 (a) D. L. Cedeno and E. Weitz, Organometallics, 2003, 22, 2652; (b) T.
H. Tuip and J. A. Ibers, J. Am. Chem. Soc., 1979, 79, 4202.
35 P. W. Jennings and L. L. Johnson, Chem. Rev., 1994, 94, 2241; M. D.
Fryzuk, X. Gao and S. J. Rettig, J. Am. Chem. Soc., 1995, 117, 3106.
36 C. D. Wood and P. E. Garrou, Organometallics, 1984, 3, 170.
37 T. Szymanska-Buzar and K. Kern, J. Organomet. Chem., 1999, 592,
212.
38 T. M. Barnhart, J. De Fellippis and R. J. McMahon, Angew. Chem.,
24 G. M. Sheldrick, SHELXS-97, Program for solution of crystal struc-
tures, University of Go¨ttingen, Germany, 1997.
25 R. S. Dickson, P. Yin, M. Ke, J. Johnson and G. B. Deacon, Polyhedron,
1996, 15, 2237.
Int. Ed. Engl., 1993, 32, 1073.
39 J.-M. Zhuang and D. Sutton, Organometallics, 1991, 10, 1516.
40 J. R. Bleeke and E. L. Mutterties, J. Am. Chem. Soc., 1981, 103,
556.
26 P. V. Rinze and U. Muller, Chem. Ber., 1979, 112, 1973.
27 D. Blazina, S. B. Duckett, T. K. Halstead, C. M. Kozak, R. J. K. Taylor,
M. S. Anwar, J. A. Jones and H. A. Carteret, Magn. Reson. Chem., 2005,
43, 200.
28 (a) B. A. Messerle, C. J. Sleigh, M. G. Partridge and S. B. Duckett, J.
Chem. Soc., Dalton Trans., 1999, 1429; (b) P. Hubler and J. Bargon,
Angew. Chem., Int. Ed., 2000, 39, 3701.
29 (a) S. A. Colebrooke, S. B. Duckett and J. A. B. Lohman, Chem.
Commun., 2000, 685; (b) S. A. Colebrooke, S. B. Duckett and J. A.
B. Lohman, Chem.–Eur. J., 2004, 10, 2459.
30 C. A. Tolman, J. Am. Chem. Soc., 1970, 92, 2953.
31 C. Loubser, H. M. Roos and S. Lotz, J. Organomet. Chem., 1991, 402,
393.
41 W. R. McClellan, H. H. Hoehn, H. N. Cripps, E. L. Muetterties and B.
W. H ow k , J. Am. Chem. Soc., 1961, 83, 1601.
42 C. K. Brown, W. Mowat, G. Yagupsky and G. Wilkinson, J. Chem. Soc.
A, 1971, 850.
43 (a) R. F. Heck, J. Am. Chem. Soc., 1965, 87, 2572; (b) R. F. Heck, J.
Am. Chem. Soc., 1963, 85, 655.
44 R. Rehderd, Coord. Chem. Rev., 1991, 110, 161.
45 K. W. Kramarz, R. J. Klinger, D. E. Fremgen and J. W. Rathke, Catal.
Today, 1999, 49, 339.
46 J. P. Dunne, S. Aiken, S. B. Duckett, D. Konya, K. Q. Almedia Lenero
and E. Drent, J. Am. Chem. Soc., 2004, 126, 16708.
47 G. Margheri, L. Rosi, A. Salvini, M. Bianchi and P. Frediani, J. Mol.
Catal. A: Chem., 1998, 132, 189.
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 2496–2509 | 2509
©