Wide bite angle amine, arsine and phosphine ligands in rhodium-and platinum/tin-catalysed hydroformylation 1

Article abstract of DOI:10.1039/b002126l

New wide bite angle amine, arsine and mixed phosphineamine and phosphinearsine ligands based on xanthene back-bones were synthesized. The co-ordination chemistry and the catalytic performance of these ligands were compared to those of the parent phosphine

Full text of DOI:10.1039/b002126l

View Article Online / Journal Homepage / Table of Contents for this issue
Wide bite angle amine, arsine and phosphine ligands in rhodium-
and platinum/tin-catalysed hydroformylation†
Lars A. van der Veen, Peter K. Keeven, Paul C. J. Kamer and Piet W. N. M. van Leeuwen*
Institute of Molecular Chemistry, University of Amsterdam, Nieuwe Achtergracht 166,
1018 WV Amsterdam, The Netherlands. E-mail: pwnm@anorg.chem.uva.nl
Received 16th March 2000, Accepted 16th May 2000
Published on the Web 14th June 2000
New wide bite angle amine, arsine and mixed phosphineamine and phosphinearsine ligands based on xanthene back-
bones were synthesized. The co-ordination chemistry and the catalytic performance of these ligands were compared
to those of the parent phosphine ligands. The amine based xanthene ligands do not form rhodium–hydride complexes
and therefore give very poor rhodium hydroformylation catalysts. The catalytic performance of the xantarsine and
the mixed xantphosarsine ligands is comparable with that of the xantphos ligands and they form similar (ligand)-
Rh(CO)2H and [(ligand)Rh(CO)2]2 complexes. In the platinum/tin-catalysed hydroformylation the xantarsine and the
mixed xantphosarsine ligands proved to be superior to the xantphos ligands. The remarkably high selectivity and
activity that is displayed by the mixed xantphosarsine ligand is explained by its wide natural bite angle and the
formation of cis co-ordinated platinum complexes.
Pt(H)(PH3)2(SnCl3)(C2H4) complex shows that the major role
Introduction
Ϫ
of the SnCl3 ligand is to stabilise the five-co-ordinated inter-
Hydroformylation of alkenes is one of the world’s largest
mediates of this complex and to weaken the platinum–hydride
homogeneously catalysed reactions in industry, producing more
bond trans to it, thus favouring alkene insertion.29
than six million tons of aldehydes and alcohols annually.1–3 The
Both in the rhodium- and the platinum/tin-catalysed hydro-
commercial hydroformylation processes are run exclusively on
formylation phosphines have proven to be superior to other
cobalt or rhodium complexes as catalysts. Platinum complexes
donor ligands, such as nitrogen- or arsine-containing com-
can also give active hydroformylation catalysts, but are mainly
pounds.15,20,30,31 Owing to their extensive use in catalysis, the
of academic interest. Extensive research has been devoted to
improve the selectivity of these catalyst systems toward the
formation of the industrially more important linear aldehydes.
influence of phosphine structure on catalytic performance has
been studied in great detail.2 Tolman introduced the concept
of the cone angle θ and the electronic parameter χ to classify
High selectivities in the hydroformylation of terminal alkenes
phosphorus ligands with respect to their steric bulk and phos-
have been reported for both diphosphine and diphosphite
phine basicity.32,33 Casey and Whiteker developed the concept
modified rhodium catalysts.4–10 Novel rhodium catalysts for the
selective hydroformylation of internal alkenes to linear alde-
of the natural bite angle as an additional characteristic of
diphosphine ligands.34 The pronounced influence of the natural
hydes, which is of great interest in both industry and synthetic
bite angle of diphosphine ligands on the activity and selectivity
organic chemistry, were developed recently.11,12
Both terminal and internal alkenes can be hydroformylated
of several catalytic reactions has been reviewed recently.35,36
In a number of these reactions, including the rhodium- and
selectively also by employing platinum–diphosphine complexes
platinum/tin-catalysed hydroformylation, diphosphines having
activated by tin chloride as the co-catalyst.13–19 Tin free catalyst
wide natural bite angles proved to be favourable to the catalytic
systems, however, have been reported as well.20–23 Despite the
performance. Unfortunately, ligands having relatively wide
very high linear over branched (l:b) aldehyde ratios induced
natural bite angles are scarce. A family of diphosphine ligands
by the platinum/tin–diphosphine catalysts, these systems have
having wide natural bite angles was developed in our group
mainly been applied to asymmetric hydroformylation so far.
by computer-assisted design.6,7,12 These ligands, based on the
Major drawbacks of these catalyst systems are extensive iso-
readily available heterocyclic xanthene-type backbones, give
merisation and hydrogenation of the substrate alkenes.3,24 In the
asymmetric platinum/tin-catalysed hydroformylation very high
very efficient catalysts for several reactions.6,7,12,37–39
Based on these results we wondered whether wide natural
enantioselectivity has been obtained for a variety of prochiral
bite angles would also improve the catalytic performance of
alkenes applying diphosphine ligands derived from 4-hydroxy-
ligands having donor atoms other than phosphorus. Since the
xanthene backbone is an excellent scaffold for the construction
-proline.25
The role of the tin chloride in the platinum/tin-catalysed
of ligands with wide natural bite angles, we set out to synthesize
hydroformylation is not yet fully understood, since it can act as
the (mixed) Group 15 derivatives of the xantphos ligands. Here
a Lewis acid, as a ligand directly bonded to platinum, and as
we report the synthesis of the amine and arsine analogues of
the xantphos ligands 1a and 2a (Table 1), and their performance
in the rhodium-, and platinum/tin-catalysed hydroformylation
reaction. The amine based xanthene ligands give very poor
Ϫ
SnCl3 counter ion.26,27 According to the ionic mechanism
for platinum/tin–diphosphine catalysed hydroformylation
Ϫ
postulated by Tóth and co-workers, the effect of the SnCl3
ligand is an increase of the cationic character of platinum.28
rhodium hydroformylation catalysts, since they do not co-
A recent theoretical study on the alkene insertion in the
ordinate to rhodium at all. The catalytic performance of the
xantarsine and the mixed xantphosarsine ligands, on the other
hand is comparable with that of the xantphos ligands and they
form the same (ligand)Rh(CO)2H and [(ligand)Rh(CO)2]2
complexes. In the platinum/tin-catalysed hydroformylation the
† Electronic supplementary information (ESI) available: details of
syntheses and characterisations. See http://www.rsc.org/suppdata/dt/b0/
b002126l/
DOI: 10.1039/b002126l
J. Chem. Soc., Dalton Trans., 2000, 2105–2112
This journal is © The Royal Society of Chemistry 2000
2105

View Article Online
phase transfer catalyst (Scheme 1).41 Boiling 2,7-di-t-butyl-
4,5-diiodo-9,9-dimethylxanthene and diphenylamine or
5
phenoxazine, in the presence of copper powder, potassium
carbonate and a catalytic amount of 18-crown-6 under reflux
overnight in o-dichlorobenzene gave compounds 1b, 2b, and 6
in 58, 52, and 58% yield, respectively. Compound 5 was
obtained in high yield via dilithiation of 4,5-dibromo-2,7-
di-t-butyl-9,9-dimethylxanthene 4 and subsequent reaction
of the formed dilithio compound with 1,2-diiodoethane.42
Monolithiation of xantamine 6 followed by reaction with
chlorodiphenylphosphine gave xantphosamine ligand 1d in
62% yield.
The syntheses of the xantarsine and xantphosarsine ligands
1c, 2c and 1e were performed via a procedure similar to the
syntheses of the xantphos ligands 1a and 2a (Scheme 1).43
Reaction of dilithiated 4 with chlorodiphenylarsine or 10-
chlorophenoxarsine gives ligands 1c and 2c in 83 and 64%
yield, respectively. Chlorodiphenylarsine and 10-chlorophen-
oxarsine were prepared according to literature procedures.44,45
Xantphosarsine ligand 1e was obtained in 49% yield by mono-
lithiation of xantphos 7 and consequent reaction of the
lithiated compound with chlorodiphenylarsine. Compound 7
was synthesized by monolithiation of 4, followed by reaction
with chorodiphenylphosphine. The yield obtained is only 48%,
which is relatively high keeping in mind that the statistical yield
for this reaction is 50%.
xantarsine and the mixed xantphosarsine ligands are even
superior to the xantphos ligands.40
Results
Ligand synthesis
Rhodium complexes
The xantamine ligands 1b, 2b, and 6 were prepared by a modifi-
cation of the Ullmann reaction employing 18-crown-6 as a
We used IR and NMR spectroscopy to study the rhodium
complexes that are formed upon mixing the ligands and
Rh(CO)2(acac) under an atmosphere of CO/H2 (1:1). The high
pressure (HP) IR spectra obtained for the xantamine ligands
1b and 2b did not show the formation of the corresponding
(ligand)Rh(CO)2H complexes, the catalyst’s resting state in
rhodium–diphosphine catalysed hydroformylation.46
Table 1 Natural bite angles and flexibility ranges for the xantamine-,
xantarsine-, and xantphos ligands
Natural bite
Flexibility
Ligand
E, EЈ
angle (βn)a/Њ
rangea/Њ
Instead,
the Rh(CO)2(acac) was slowly converted into poorly soluble
rhodium–carbonyl clusters and the xantamine ligands 1b and
2b do not seem to co-ordinate to rhodium at all. The three
absorption bands in the carbonyl region of the IR spectrum
(2073, 2043, and 1820 cmϪ1) observed for the reaction using
ligand 1b can fully be ascribed to the hexanuclear rhodium–
carbonyl cluster Rh6(CO)16.47,48 For ligand 2b five carbonyl
absorptions (2074, 2068, 2043, 1886, and 1820 cmϪ1) were
displayed in the IR spectrum, indicating that both the hexa-
nuclear as well as the tetranuclear rhodium–carbonyl cluster,
Rh4(CO)12, were formed. The latter is most likely an intermedi-
ate in the formation of the larger rhodium–carbonyl cluster.49
Compared to ligands 1b and 2b, Rh(CO)2(acac) was converted
much faster in the HP IR experiment using the xantphosamine
ligand 1d. The HP IR spectrum for ligand 1d again showed the
1a
2a
1b
2b
1c
2c
1d
1e
3
P, P
P, P
110.1
121.4
98–134
107–138
N, N
N, N
As, As
As, As
N, P
As, P
P, P
112.9
123.0
98–132
107–142
111.4
102.0
97–130
92–120
a The natural bite angle (βn) and the flexibility range were calculated as
by Casey and Whiteker.34 βn is defined as the preferred chelation angle
determined only by ligand backbone constraints and not by metal
valence angles. The flexibility range is defined as the accessible range of
bite angles within 12.6 kJ molϪ1 excess strain energy from the calculated
natural bite angle.
Scheme 1 Ligand synthesis. (i) n-BuLi/ICH2CH2I/THF/Ϫ20 ЊC; (ii) R2NH/Cu/K2CO3/18-crown-6/C6H4Cl2/220 ЊC; (iii) n-BuLi/R2AsCl or Ph2PCl/
THF/Ϫ60 ЊC.
2106
J. Chem. Soc., Dalton Trans., 2000, 2105–2112

View Article Online
Table 2 31P-{1H} NMR Data of the (ligand)Pt(SnCl3)Cl complexesa
Ligand
% cisb
1J(Pt,PtransCl)/Hz
1J(Pt,PtransSn)/Hz
1J(Pt,PtransP)/Hz
2J(P,P)/Hz
1
67
74
100
3706
3697
3840
3046
2660
3117
1954
—
—
14.3
—
13.0
1e
3c
a In CD2Cl2 at Ϫ73 ЊC. b Percentage of the complex in which the ligand is co-ordinated cis to platinum. c T = Ϫ50 ЊC.
Fig. 2 cis and trans complexes (ligand)Pt(SnCl3)(Cl).
Fig. 1 Structures of ee, ea, and dimeric complexes.
Platinum complexes
formation of the hexanuclear rhodium–carbonyl cluster along
with the formation of another rhodium–carbonyl complex.
The structure of this complex is not clear, however the absence
of a rhodium–hydride signal in the 1H NMR spectrum of
the complex mixture excludes the desired (ligand)Rh(CO)2H
complex. The 31P-{1H} NMR spectrum of the complex mixture
shows one broad doublet at δ 41.6, with a relatively large
rhodium–phosphorus coupling constant (1J(Rh,P)) of 184 Hz.
The composition of the complex mixture was not affected by
the addition of 1-octene.
The (ligand)Pt(SnCl3)Cl complexes of ligands 1a, 1d, and 1e
were investigated using 31P-{1H} NMR spectroscopy to
elucidate the co-ordination mode of the ligands. For com-
parison also the platinum/tin complex of homoxantphos7
3
was measured. The platinum–phosphorus coupling constants
(1J(Pt,P)) of the square-planar (ligand)Pt(SnCl3)Cl complexes
are tools to differentiate between cis and trans co-ordination
of phosphine ligands (see Fig. 2). The platinum–phosphorus
coupling constants of the trans platinum() complexes are
usually considerably lower than those of the cis com-
plexes.25,26,52,53 The (ligand)Pt(SnCl3)Cl complexes were pre-
pared in situ in CD2Cl2 from PtCl2(cod), ligand and anhydrous
tin() chloride. In general broad signals were observed in the
1H and 31P-{1H} NMR spectra at room temperature, indicating
the occurrence of dynamic processes on the NMR timescale.
These dynamic processes were frozen at low temperature.
The platinum–phosphorus coupling constants of the com-
plexes of ligands 1a, 1e, and 3 measured at Ϫ73 or Ϫ50 ЊC are
summarised in Table 2.
Xantarsine ligand 2c reacted very slowly with Rh(CO)2-
(acac). In the HP IR spectrum formation of the hexanuclear
rhodium–carbonyl cluster and of another complex with
carbonyl absorptions at 2060, 2045, 2027, and 1795 cmϪ1 was
observed. The latter complex is most likely a rhodium–carbonyl
cluster of the type Rhx(CO)y(2c)n.47,48 The HP IR spectrum of
the complexes obtained with xantarsine ligand 1c clearly
showed the four carbonyl absorptions that can be assigned to
two isomers of the rhodium–hydride complex Rh(CO)2(1c)H
(2036, 1998, 1971, and 1953 cmϪ1), in which the ligand is
co-ordinated in the equatorial–equatorial (ee) and in the
equatorial–apical (ea) fashion, respectively (Fig. 1).46 Pre-
dominantly, however, formation of the rhodium–dimer
complex [Rh(CO)2(1c)]2 (see Fig. 1) was observed (2012, 1981,
1768, 1756 cmϪ1).43,50,51 The formation of the rhodium–hydride
complex was confirmed by 1H NMR spectroscopy. In the
1H NMR spectrum the rhodium–hydride signal is observed at
δ Ϫ9.68 as a doublet with a rhodium–proton coupling constant
(1J(Rh,H)) of 11.1 Hz.‡ Also for the xantphosarsine ligand
1e both the rhodium–hydride complex Rh(CO)2(1e)H (2037,
1998, 1978, 1953 cmϪ1, mixture of ee and ea isomers) and
the rhodium–dimer complex [Rh(CO)2(1e)]2 (1974, 1767, 1754
cmϪ1) were formed according to the IR spectrum, but with this
ligand the dimer complex is formed in a considerably smaller
amount. In the 1H NMR spectrum of Rh(CO)2(1e)H the
rhodium–hydride signal appears as a double doublet at δ Ϫ8.65
with a rhodium–proton coupling constant of 8.4 Hz and a
relatively large phosphorus–proton coupling constant (2J(P,H))
of 78 Hz. The 31P-{1H} NMR spectrum of Rh(CO)2(1e)H
exhibits a doublet at δ 26.9 with a rhodium–phosphorus
coupling constant (1J(Rh,P)) of 112 Hz. Both the intensities of
the carbonyl vibrations in the IR spectrum and the hetero-
nuclear coupling constants in the NMR spectra of Rh(CO)2-
(1e)H indicate that the ea complex isomer is favoured over the
ee isomer.46
At low temperature three signals were observed in the 31P-
{1H} NMR spectrum for the (ligand)Pt(SnCl3)(Cl) complex
containing xantphos 1a. Based on the mutual phosphorus–
phosphorus coupling 2J(P,P), the doublets at δ 16.4 and 4.3 are
assigned to the cis isomer.26 From the chemical shifts and the
platinum–phosphorus coupling constants it can be concluded
that the doublet at δ 16.4 belongs to the phosphine positioned
trans to tin, and that at δ 4.3 to the phosphine positioned trans
to chlorine. The singlet at δ 15.2 is assigned to the trans com-
plex. According to the relative intensities of the signals in the
NMR spectrum, the cis:trans isomer ratio of the complex
Pt(SnCl3)(Cl)(1) is 2:1. The complex of xantphosarsine 1e
exhibited two signals in the 31P-{1H} NMR spectrum with
relative intensities of 3:1. The relatively large platinum–
phosphorus coupling constant and the relatively small tin–
phosphorus coupling constant (2Javg(Sn,P))54 of 212 Hz of the
major signal at δ 2.1 are indicative of a phosphine positioned
trans to chlorine and cis to tin and, hence, is assigned to
a cis complex. The platinum–phosphorus coupling constant
of the minor signal at δ Ϫ2.6 is in between those of cis and
trans complexes. Considering the absence of a small cis tin–
phosphorus coupling constant, however, it is assigned to the
other cis complex isomer in which arsenic is positioned trans
to chlorine. The 3:1 ratio of the cis complexes is in agreement
with the relative trans influences of the phosphorus and arsenic.
The cis complex in which phosphorus is positioned trans
to chlorine is favoured, since phosphorus has a larger trans
directing effect than arsenic.55 For homoxantphos 3 it is clear
that only the cis complex isomer is formed. In the 31P-{1H}
NMR spectrum of the complex Pt(SnCl3)(Cl)(3) only two
doublets with relatively large platinum–phosphorus coupling
constants are observed.
‡ The Rh(CO)2(1c)H complex consists of a mixture of ee and ea iso-
mers that is in dynamic equilibrium on the NMR timescale and, there-
fore, only one rhodium–hydride signal is observed in the 1H NMR
spectrum. This dynamic behaviour is consistent with that of xanthene-
type (diphosphine)Rh(CO)2H complexes, and cannot be frozen out
even at 200 K.7,43,46
J. Chem. Soc., Dalton Trans., 2000, 2105–2112
2107

View Article Online
Table 3 Results of the rhodium-catalysed hydroformylation of 1-octene at 80 ЊCa
%
l:b
% Linear
aldehydeb
%
Ligand
Conversionb
ratiob
Isomerisationb
tofb,c
—
1ad
2ae
1b
2b
1c
96
22
30
55
62
90
55
96
17
30
1.9
50
67
3.0
3.0
2.9
3.5
2.7
8.9
8.5
21
94
89
21
21
65
52
52
87
88
68
3.9
10
72
71
12
33
28
3.2
1.4
180
240
1600
90
94
500
200
450
84
2c
1d
1e
3f
37
a Conditions: P(CO/H2) (1:1) = 20 bar, ligand:Rh = 5:1, substrate:Rh = 637:1, [Rh] = 1.00 mM, number of experiments = 3. Hydrogenation was
observed in none of the experiments. b Percent conversion, l:b ratio, percent linear aldehyde, percent isomerisation to 2-octene, and turnover
frequency were determined after 1 h by GC analysis. c Averaged turnover frequency = (mol of aldehyde) (mol of Rh)Ϫ1 hϪ1 d The reaction was
.
analysed after 0.5 h. e The reaction was analysed after 6 min. f The reaction was analysed after 4.5 h.
As was the case for rhodium, xantphosamine 1d forms other
platinum/tin complexes than the arsine and phosphine ligands.
According to the 31P-{1H} NMR spectrum, two platinum
complexes are formed in a 9:2 ratio using ligand 1d. The
major complex displays two signals at δ 22.9 and Ϫ0.05 with
platinum–phosphorus coupling constants of 3081 and 3601 Hz,
respectively. This strongly indicates that in this complex two
xantphosamine ligands are co-ordinated to platinum only via
the phosphine moieties, which are probably in mutually cis
positions. The formation of (ligand)2MCl2 complexes starting
from MCl2(cod) and ligand has been observed before for other
mixed phosphorus/nitrogen ligands.56,57 For the minor complex
only one signal is observed, which has a large platinum–
phosphorus coupling constant of 5721 Hz. The structure of
this complex is unclear, but the large platinum–phosphorus
coupling constant could be indicative of the chelate complex
Pt(SnCl3)Cl(1d) with the phosphine and the nitrogen group in
mutual trans positions.
For ligands 1a and 3 the platinum/tin–diphosphine com-
plexes present during hydroformylation conditions were also
studied using HP IR spectroscopy. The catalyst precursor com-
plexes (ligand)Pt(SnCl3)(CO)H were prepared in situ by stirring
the solutions of the (ligand)Pt(SnCl3)Cl complexes in CH2Cl2
at 60 ЊC under an atmosphere of 40 bar of CO/H2 (1:1).
Formation of the (ligand)Pt(SnCl3)(CO)H complexes was
evidenced by the observation of a single platinum–carbonyl
vibration in the carbonyl region of the IR spectrum. The IR
carbonyl bands for xantphos ligands 1a and 3 are at 2049 and
2041 cmϪ1, respectively. The platinum–hydride vibration was
not observed for either complex. The (ligand)Pt(SnCl3)(CO)H
complex proved to be the catalyst resting state for ligand 3,
since the addition of 1-octene to the reaction mixture did not
lead to the formation of new complexes. The occurrence of the
hydroformylation reaction was clearly evidenced by the decay
of the absorption of 1-octene at 1639 cmϪ1, and the rise of
the strong acyl absorption of 1-nonanal at 1724 cmϪ1. Upon the
addition of 1-octene to the platinum–hydride complex of ligand
1a the platinum–carbonyl absorption band of this complex
disappeared completely and two new absorptions at 1828 and
1683 cmϪ1 emerged. The band at 1683 cmϪ1 can be ascribed to
the platinum–acyl complex, trans-Pt(SnCl3)(COC8H17)(1a).26,27
The vibration at 1828 cmϪ1 is characteristic of the presence
of bridging carbon monoxide ligands and thus suggests the
formation of oligomeric platinum–acyl complexes [Pt(SnCl3)-
(COC8H17)(CO)n(1a)]m in which xantphos 1a and or the
carbonyl ligand bridge two platinum centres. The formation of
oligomeric platinum–acyl complexes has been reported before
by Scrivanti et al.26
Hydroformylation of 1-octene
(a) Rhodium. The rhodium-catalysed hydroformylation of
1-octene was performed at 80 ЊC and 20 bar of 1:1 CO/H2
using a 1.0 mM solution of rhodium diphosphine catalyst
prepared from Rh(CO)2(acac) and five equivalents of ligand.
The production of octene isomers, nonanal, and 2-methyl-
octanal was monitored by gas chromatography. The results of
the experiments with ligands 1–3 are shown in Table 3.
For the xantamine ligands 1b and 2b almost the same selec-
tivity for linear aldehyde formation and the same extent of
isomerisation are observed as for hydroformylation experi-
ments without added ligand. The application of ligands 1b and
2b only leads to a slight increase of the l:b ratio and a lower
hydroformylation rate. The xantphosamine ligand 1d clearly
forms an active hydroformylation catalyst. The high activity
displayed by this ligand is, however, accompanied by a con-
siderable amount of isomerisation and a low l:b ratio, resulting
in a low selectivity for linear aldehyde formation of 52%. These
results are comparable to those reported for other phosphorus/
nitrogen ligands.58–60
The results obtained with the xantarsine ligands 1c and
2c and the xantphosarsine ligand 1e show that substituting
phosphine for arsine gives less selective but equally active
catalysts. While substitution of one diphenylphosphine group
for one diphenylarsine group decreases the hydroformylation
rate by a factor of three, substitution of both diphenylphos-
phine groups for diphenylarsine groups doubles the activity. For
the xantphos ligands the substitution of the diphenylphosphine
moieties for phosphacyclic groups results in enhanced catalytic
activity.43 The phosphacyclic xantphos 2a is more than six times
as active as xantphos 1a. The opposite effect is observed for
the xantarsine ligands. The arsacyclic ligand 2c proves to be
only half as active as the non-cyclic xantarsine 1c. The mixed
xantphosarsine ligand 1e proves to be the most selective arsine
modified ligand. Although ligand 1e is less selective and active
than xantphos ligands 1a and 2a, it is equally selective and more
than twice as active as xantphos ligand 3.
(b) Platinum/tin. The platinum/tin-catalysed hydroformyl-
ation of 1-octene was performed at 60 ЊC and 40 bar of 1:1 CO/
H2 using a 2.5 mM solution of platinum/tin–diphosphine
catalyst prepared from PtCl2(cod), two equivalents of ligand,
and two equivalents of SnCl2. The production of octene iso-
mers, nonanal, and 2-methyloctanal was monitored by gas
chromatography. The results of the experiments with ligands
1a, 1d, 1e and 3 are shown in Table 4. Averaged turnover
frequencies were determined at 20–30% conversion. A relatively
2108
J. Chem. Soc., Dalton Trans., 2000, 2105–2112

View Article Online
small natural bite angle appeared to be beneficial for the
reaction rate. Ligand 3, having a natural bite angle of 102Њ, is
40 times as active as ligand 1a, which has a natural bite angle
of 110Њ. The increase in activity, however, is accompanied by a
considerable increase of the isomerisation rate. It is striking
that the selectivities for linear aldehyde obtained for ligands 1a
and 3 are virtually identical to those obtained in the rhodium-
catalysed hydroformylation.
in Table 5. Xantphos 3 shows a very good activity in the hydro-
formylation of methyl trans-3-pentenoate. The high activity,
however, is accompanied by a substantial amount of hydro-
genation. The xantarsine ligand 1c and xantphosarsine ligand
1e displayed less hydrogenation, but also lower hydroformyl-
ation activity compared to xantphos 3.
Discussion
Remarkably, the substitution of one phosphine group of
ligand 1a for arsine increases the hydroformylation activity
twentyfold, without losing the excellent selectivity for linear
aldehyde formation. To the best of our knowledge the high
activity and selectivity displayed by the mixed xantphosarsine
ligand 1e under these mild conditions is unprecedented.
Usually, the hydrogenation and isomerisation rates increase
along with the hydroformylation activity. Substituting both
phosphine groups for arsine results in a twelvefold increase
of the hydroformylation rate compared to xantphos 1a.
Xantarsine ligand 1c is still a very active ligand, but not as
selective as xantphosarsine ligand 1e. The xantamine ligand 1b
and the xantphosamine ligand 1d gave very poor results in the
platinum/tin-catalysed hydroformylation. The strong electron
donating capacity of the diphenylamine group evidently
renders the platinum/tin complexes unsuitable for efficient
hydroformylation.
The high activities displayed by ligands 1c, 1e, and 3 in the
platinum/tin-catalysed hydroformylation of 1-octene prompted
us to test these ligands also in the platinum tin-catalysed hydro-
formylation of methyl trans-3-pentenoate. Meessen et al. have
recently reported that in the hydroformylation of methyl
trans-3-pentenoate the xantphos ligands display high regio-
selectivities toward linear aldehyde.19 Selective linear hydro-
formylation of unsaturated carboxylic acid esters, such as
methyl 3-pentenoate, is of great industrial interest, since it
can provide new synthetic routes to polyamides and polyesters.
The platinum/tin-catalysed hydroformylation of methyl trans-3-
pentenoate was performed similarly to the hydroformylation
of 1-octene only at a temperature of 80 ЊC and a pressure of
15 bar of 1:1 CO/H2. The results of the experiments are shown
The results of the rhodium-catalysed hydroformylation with the
(mixed) xanthamine, and xantarsine ligands are consistent with
earlier work. The superiority of phosphines among the Group
15 ligand series was demonstrated before by Carlock in the
rhodium-catalysed hydroformylation of 1-dodecene.61 The fact
that the xantamine and xantphosamine ligands 1b, 2b, and 1d
do not form selective hydroformylation catalysts is in agreement
with the HP IR experiments. According to the IR spectra
ligands 2b and 2c do not co-ordinate to rhodium at all
and rhodium–carbonyl clusters are formed. The decreased
hydroformylation rates found for these ligands can be explained
by the fact that the ligands slow down the reaction of
the rhodium–carbonyl clusters with hydrogen to form the
Rh(CO)4H complex, which is the active catalyst in unmodified
rhodium-catalysed hydroformylation. The IR and NMR data
obtained for the rhodium complex containing ligand 1d are
inconclusive. The structure of the catalytic complex formed for
ligand 1d is unknown and no conclusions can be drawn
whether, besides the phosphine, also the amine function of
the ligand is co-ordinated to rhodium. Probably a dimeric or
a cluster rhodium–carbonyl complex containing one or more
ligands is formed.
Contrary to the xantamine ligands, the xantarsine and xant-
phosarsine ligands 1c and 1e give efficient rhodium hydro-
formylation catalysts. The activity of 1c and the selectivity of
1e resemble those of the xantphos ligands. This shows that
diarsine or phosphine–arsine ligands can perform as well as
diphosphine ligands. For these xantarsine ligands the same
rhodium–hydride and the same dimeric rhodium complexes are
formed as for the xantphos ligands. The poor results obtained
for ligand 2c are explained by the fact that this ligand, like
1d, gives undefined rhodium–carbonyl clusters instead of the
rhodium–hydride complex.
Table 4 Results of the platinum/tin-catalysed hydroformylation of
1-octene at 60 ЊCa
The HP IR and NMR spectra of the (ligand)Rh(CO)2H
complex containing ligand 1e indicate that, compared to
xantphos 1a, ligand 1e shows an increased preference for ea co-
ordination. This is best explained by the electronic dissimilarity
of arsine and phosphorus. Based on the relatively small
rhodium–phosphorus and large phosphorus–proton coupling
constants, it can be concluded that in the ea isomer of
Rh(CO)2(1e)H the phosphine group is co-ordinated pre-
dominantly at the apical site of the trigonal bipyramidal com-
plex, while the arsine group is located mainly in the equatorial
plane. These results are consistent with the work of Casey et al.
on electronically dissymmetric diphosphines, which give
enhanced ea co-ordination with the stronger donor ligand on
the apical position trans to the hydride ligand.62 The relatively
large rhodium–hydride coupling constant observed in the
1H NMR spectrum of Rh(CO)2(1c)H could be an indication
l:b
% Linear
aldehydeb
%
Ligand
ratiob
Isomerisationb
tofb,c
1a
1b
1c
1d
1e
3
230
1.9
>250
3.9
200
>250
95
8.5
92
23
96
88
4.5
87
8.0
71
3.1
12
18
<1
210
6.5
350
720
a Conditions: P(CO/H2) (1:1) = 40 bar, ligand:SnCl2 :Pt = 2:2:1,
substrate:Pt = 255:1, [Pt] = 2.50 mM, number of experiments = 2.
Hydrogenation was observed in none of the experiments. b l:b ratio,
percent linear aldehyde, percent isomerisation to 2-octene, and turn-
over frequency were determined at 20% conversion. c Turnover
frequency = (mol of aldehyde) (mol of Pt)Ϫ1 hϪ1
.
Table 5 Results of the platinum/tin-catalysed hydroformylation of methyl trans-3-pentenoate at 80 ЊCa
Ligand
t/h
% Conversionb
% Hydrogenationb
% 2-MPb
% Aldehydeb
l:b ratiob,c
tofb,d
1c
1e
3
16
16
16
39
36
95
4.0
3.4
16
27
29
19
69
68
65
2.9
3.7
3.4
4.5
3.9
10
a Conditions: P(CO/H2) (1:1) = 15 bar, ligand:SnCl2 :Pt = 2:2:1, substrate:Pt = 326:1, [Pt] = 2.50 mM, number of experiments = 2. b Percent
conversion, -hydrogenation, -methyl 2-pentenoate (MP), -linear aldehyde, and l:b ratio were determined by GC analysis. c l:b ratio includes all
branched aldehydes. d Averaged turnover frequency = (mol of aldehyde) (mol of Pt)Ϫ1 hϪ1
.
J. Chem. Soc., Dalton Trans., 2000, 2105–2112
2109

View Article Online
that also xantarsine 1c exhibits an increased preference for
ea co-ordination. This cannot be confirmed by HP IR spectro-
scopy since the carbonyl absorptions of the Rh(CO)2(1c)H
complex are partly hidden under the intense absorption bands
of the [Rh(CO)2(1c)]2 dimer.
cis complex has to occur. This clearly explains why for tri-
phenylphosphine and for ligands with wide natural bite angles
hydrogenolysis is found to be the rate determining step in
the catalytic cycle.16,27 The increased preference for trans co-
ordination with increasing natural bite angle is demonstrated
by the co-ordination behaviour of xantphos ligands 1a and 3
in the (ligand)Pt(SnCl3)Cl complexes. Ligand 3 gives the cis
complex exclusively according to the 31P-{1H} NMR spectrum,
while 1a gives the trans complex in 33% yield.
The very good catalytic performances of xantarsine ligand
1c and xantphosarsine ligand 1e in the platinum/tin-catalysed
hydroformylation are remarkable. Arsine ligands have been
employed in platinum/tin-catalysed hydroformylation before,
but like in the rhodium-catalysed hydroformylation, without
exception they performed worse than phosphine ligands.15,20,31
Especially the high hydroformylation activity and excellent
selectivity for linear aldehyde formation displayed by ligand 1e
are striking. The xantphosarsine modified platinum/tin catalyst
performs even better than the rhodium–xantphos catalysts
in the hydroformylation of 1-octene. The low activity of xant-
phosamine ligand 1d in the platinum/tin-catalysed hydro-
formylation is not surprising. As was demonstrated by Dekker
et al. for other phosphine–amine ligands, the barrier for
migratory insertion reactions in square-planar platinum()
complexes is higher for this type of ligand than for symmetric
diphosphine or diamine ligands.57 Alternatively, it is likely that
ligand 1d does not form chelate complexes at all and during
catalysis two ligands are co-ordinated to platinum only via the
phosphine moieties. This is evidenced by the fact that according
to NMR spectroscopy the complex Pt(SnCl3)Cl(1d)2 is formed
to the greater extent when ligand 1d is treated with PtCl2(cod)
and anhydrous tin chloride. The symmetric xantamine 1b
performs even worse than ligand 1d. This can be explained by
the strong electron donating capacity of the diphenylamine
groups which have an unfavourable effect on the hydroformyl-
ation cycle.
In the platinum/tin-catalysed hydroformylation homoxant-
phos 3, having the smaller natural bite angle, showed a higher
activity than the xantphos ligand 1a. These results are con-
sistent with the previous work of Kawabata and Hayashi and
co-workers.13,16,63 It can be deduced from their work that in
the platinum/tin–diphosphine catalysed hydroformylation the
activity increases with increasing natural bite angles up to a
value of approximately 102Њ (DIOP, 4,5-bis(diphenylphos-
phinomethyl)-2,2-dimethyl-1,3-dioxolane),5 but for wider bite
angles the activity drops abruptly. This suggests that the rate-
determining step shifts from one point in the hydroformylation
cycle to another. For diphosphines having narrow natural bite
angles alkene co-ordination or carbon monoxide insertion
could be rate limiting, while hydrogen addition could be
rate determining for diphosphines with wide natural bite
angles. This is confirmed by the hydroformylation experiments
with xantphos ligands 1a and 3 that were monitored by IR
spectroscopy. The observation of the platinum–carbonyl
complex as the catalyst resting state for ligand 3 suggests a
rate-limiting step early in the catalytic cycle, while the
platinum–acyl complexes observed for ligand 1a indicate that
the reaction with hydrogen is rate determining.
The trend of increasing activity with increasing natural bite
angle that is observed for the ligands possessing narrow natural
bite angles16 can be explained by both the step of alkene inser-
tion and the step of carbon monoxide insertion. A theoretical
study of Rocha and De Almeida on the complex Pt(H)(PH3)2-
(SnCl3)(C2H4) shows that upon alkene insertion the P–Pt–P bite
angle increases from 106.8 to a maximum of 146.2Њ.29 Based on
these results it can be concluded that widening of the natural
bite angle enhances the alkene insertion reaction and, therefore,
the hydroformylation rate. The influence of the natural bite
angle on the rate of carbon monoxide insertion has been
demonstrated both experimentally and theoretically.55,65,66
Besides an effect on the activity, an effect of the natural bite
angle on the selectivity for linear aldehyde formation is also
observed. This bite angle effect is comparable to that observed
in the rhodium–diphosphine catalysed hydroformylation, and
can be explained in a similar way.7,46 Widening of the bite angle
will lead to an increase of the steric congestion around the
platinum centre, and, hence, to more selective formation of
the sterically less hindered linear platinum alkyl species.
The major difference, however, between the platinum/tin- and
rhodium-catalysed hydroformylation is the fate of the branched
metal–alkyl species that is also formed in the hydroformylation
cycle. The branched rhodium–alkyl species gives both branched
aldehyde formation and isomerisation to 2-octene, while for the
branched platinum–alkyl species isomerisation to 2-octene
occurs almost exclusively.
The high activity of the xantphosarsine ligand 1e in the
platinum/tin-catalysed hydroformylation is thus explained
by the fact that, despite a wide natural bite angle, this ligand
gives cis co-ordination in the platinum/tin complexes, as was
observed by 31P-{1H} NMR. The enhanced preference for cis
co-ordination of ligand 1d compared to 1a can be imposed by
the electronic dissymmetry of the ligand. On the other hand,
the high activity of xantarsine 1c suggests that this symmetric
ligand also is a better cis co-ordinating ligand than 1a, which
might be the result of the smaller trans directing influence of
arsine compared to phosphine.55 The wide natural bite angle
of ligand 1d, which is comparable to the one of xantphos 1a,
explains the high selectivity for linear aldehyde formation
observed for this ligand.
Conclusion
Wide bite angle xantarsine and mixed xantphosarsine ligands
give active and selective rhodium catalysts for the hydro-
formylation of 1-octene. The xanthene based amine ligands
give very poor hydroformylation catalysts. Under hydro-
formylation conditions the xantamine ligands do not co-
ordinate to rhodium at all, while the xantphosamine ligand
behaves most likely as a monophosphine ligand instead of
a chelating phosphorus/nitrogen ligand.
In the platinum/tin-catalysed hydroformylation the xant-
amine and xantphosamine ligands do not give active catalysts
either. The xantarsine and xantphosarsine ligands on the other
hand give very efficient catalysts for the selective hydroformyl-
ation of 1-octene. Especially the high activity and selectivity
that is observed for the xantphosarsine ligand is remarkable.
To the best of our knowledge these are the first examples of
arsine based ligands that are superior to phosphine ligands in
the platinum/tin-catalysed hydroformylation. The high activity
The shift of the rate-determining step is probably explained
by the different reactivity of the cis and trans isomers of
the platinum() complexes. In general trans co-ordination of
phosphines in the square-planar platinum/tin–acyl complexes
is thermodynamically favoured over cis co-ordination.26,27,64
However, for the occurrence of migratory insertion reactions in
the platinum/tin–hydride and –alkyl complexes and for hydro-
genolysis of the platinum/tin–acyl complex to form the alde-
hyde, cis co-ordination of the phosphines is a prerequisite.28,29,65
Diphosphines having narrow natural bite angles can only form
the cis complexes due to bite angle constraints and give, there-
fore, active hydroformylation catalysts. Diphosphines with wide
natural bite angles and monophosphines will give the more
stable trans platinum/tin–acyl complexes. The hydrogenolysis
rate of these complexes is slowed down considerably, since first
an energetically unfavourable isomerisation of the trans to the
2110
J. Chem. Soc., Dalton Trans., 2000, 2105–2112

View Article Online
2
and selectivity of the xantphosarsine ligand is explained by the
fact that it has approximately the same natural bite angle as
xantphos, but has an enhanced preference for the formation of
cis co-ordinated platinum/tin complexes.
14.3, PtransSn), 15.2 (1J(Pt,P) = 1954 (satellites), Javg(Sn,P) =
1
234 (satellites), PtransP) and 4.3 (d, J(Pt,P) = 3706 (satellites),
2Javg(Sn,P) = 117 (satellites), 2J(P,P) = 14.3 Hz, PtransCl).
Pt(CO)(SnCl3)Cl(1a) 11b. HP IR (CH2Cl2, carbonyl region,
cmϪ1): 2049 (PtCO).
Experimental
Computational details
Hydroformylation experiments
(a) Rhodium catalysed. Hydroformylation reactions were
carried out in an autoclave, equipped with a glass inner
beaker, a substrate inlet vessel, a liquid sampling valve, and a
magnetic stirring rod. The temperature was controlled by an
electronic heating mantle. In a typical experiment the 50
µmol of ligand were placed in the autoclave and the system
was evacuated and heated to 50 ЊC. After 0.5 h the autoclave
was filled with CO/H2 (1:1) and a solution of 10 µmol
Rh(CO)2(acac) in 8.5 ml of toluene. The autoclave was pres-
surised to 16 bar, heated to 80 ЊC and stirred for 1.5 h to
form the active catalyst. Then 1.0 ml of substrate (filtered
over neutral activated alumina to remove peroxide impurities)
and 0.5 ml of the internal standard n-decane were placed in
the substrate vessel, purged with 10 bar CO/H2 (1:1), and
pressed into the autoclave with 20 bar CO/H2 (1:1). After 1 h
samples of the reaction mixture were taken, quenched with
tri-n-butyl phosphite, and analysed by temperature-controlled
gas chromatography.
The molecular mechanics calculations were performed using
the CAChe WorkSystem version 4.0,67 on an Apple Power
MacIntosh 950, equipped with two CAChe CXP coprocessors.
Calculations were carried out similarly to the method described
by Casey and Whiteker,34 using a Rh–P bond length of 2.315 Å
and a Rh–As bond length of 2.415 Å.68 Minimisations were
done via the block-diagonal Newton–Raphson method,
allowing the structures to converge fully with a termination
criterion of a rms factor of 0.00042 kJ molϪ1 ÅϪ1 or less.
Preparations
Ligand 1b. This compound was prepared via a modified
version of the Ullmann reaction.41 A slurry of 1.50 g of com-
pound 5 (2.66 mmol), 1.35 g of diphenylamine (7.98 mmol),
2.5 g of copper (40 mmol), 11.0 g of potassium carbonate
(80 mmol), and 0.10 g of 18-crown-6 (0.27 mmol) was boiled
under reflux overnight in 20 ml of 1,2-dichlorobenzene. The
reaction mixture was filtered and the residual solids were
washed with dichloromethane. The filtrate was evaporated to
dryness and the residue washed with hexanes and crystallised
from ethanol–THF. Yield: 1.01 g white crystals of 1b (58%). mp
(b) Platinum/tin-catalysed. The platinum/tin-catalysed hydro-
formylation experiments were performed similarly to the
procedure described for rhodium. The catalyst complex (ligand)-
Pt(SnCl3)Cl was preformed outside the autoclave by adding
9.5 mg of SnCl2 (50 µmol) to a solution of 9.4 mg PtCl2(cod)
(25 µmol) and two equivalents of ligands in 8.5 ml of CH2Cl2.
The reactions were stopped by cooling on ice.
1
4
259–260 ЊC. H NMR (CDCl3): δ 7.31 (d, J(H,H) = 2.4, 2H;
H1,8), 7.00 (m, 10H; CH), 6.80 (t, 3J(H,H) = 7.3, 4H; CH), 6.73
(d, J(H,H) = 7.5 Hz, 8H; CH), 1.70 (s, 6H; CH3) and 1.25
3
(s, 18H; t-butyl). Calc. for C47H48N2O: C, 85.93; H, 7.37;
N, 4.26. Found: C, 86.18; H, 7.43; N, 4.06%.
Ligand 2b and compound 6 were prepared similarly to 1b.
Detailed descriptions of the syntheses and characterisations are
included in the supplementary material.
HP FT IR experiments
(a) Rhodium complexes. In a typical experiment the HP IR
autoclave was filled with two to five equivalents of ligand, 4 mg
of Rh(CO)2(acac), and 15 ml of cyclohexane. The autoclave
was purged three times with 15 bar CO/H2 (1:1), pressurised to
approximately 18 bar, and heated to 80 ЊC. Catalyst formation
was monitored in time by FT IR and was usually completed
within one h.
Ligand 1c. At Ϫ60 ЊC 2.8 ml of n-butyllithium (2.5 M in
hexanes, 6.9 mmol) were added dropwise to a stirred solution
of 1.50 g of compound 4 (3.12 mmol) in 50 ml of THF. The
resulting beige suspension was stirred for 1 h. Next a solution
of 1.82 g of chlorodiphenylarsine (6.9 mmol) in 20 ml of THF
was added and the reaction mixture slowly warmed to room
temperature overnight. It was diluted with 25 ml of ethyl
acetate and hydrolysed with 25 ml of a one to one mixture
of brine and dilute hydrochloric acid. The water layer was
removed and the organic layer dried over MgSO4. The solvents
were removed in vacuo and the residual off-white powder
washed with hexanes and crystallised from ethanol–THF.
Yield: 2.01 g white crystals of 1c (83%). mp 207–208 ЊC.
(b) Platinum/tin complexes. In a typical experiment 14 mg of
anhydrous SnCl2 (75 µmol) were added to a stirred solution
of 14 mg PtCl2(cod) (38 µmol) and two equivalents of ligand
in 15 ml of CH2Cl2. After 0.5 h the reaction was transferred
to the HP IR autoclave. The autoclave was purged three times
with 15 bar CO/H2 (1:1), warmed to 60 ЊC and pressurised with
CO/H2 (1:1) to 40 bar. After 1 h 1.0 ml of 1-octene was added
and the reaction monitored in time by FT IR.
4
1H NMR (CDCl3): δ 7.41 (d, J(H,H) = 2.3, 2H; H1,8), 7.27
(m, 20H; CH), 6.69 (d, 4J(H,H) = 2.3 Hz, 2H; C3,6), 1.72 (s, 6H;
CH3) and 1.15 (s, 18H; t-butyl). Calc. for C47H48As2O: C, 72.49;
H, 6.22. Found: C, 72.65; H, 6.25%. Ligands 1d, 1e, and 2c
and compound 7 were prepared similarly to 1c. Detailed
descriptions of the syntheses and characterisations are included
in the supplementary material.
Acknowledgements
Financial support from the Technology Foundation (STW) of
the Netherlands Organization for Scientific research (NWO) is
gratefully acknowledged.
References
Pt(SnCl3)Cl(1a) 11a. To a stirred solution of 3.7 mg of
PtCl2(cod) (10 µmol) and 7.6 mg of compound 1a (11 µmol)
in 0.75 ml of CD2Cl2 were added 2.1 mg of anhydrous SnCl2
(11 µmol). After 1 h the reaction mixture was transferred to a
1 P. W. N. M. van Leeuwen and G. van Koten, in Homogeneous
Catalysis with Transition Metal Complexes, eds. P. W. N. M. van
Leeuwen and G. van Koten, Elsevier, Amsterdam, 1993, pp. 201 and
222.
2 M. Beller, B. Cornils, C. D. Frohning and C. W. Kohlpaintner,
J. Mol. Catal. A: Chem., 1995, 104, 17.
3 C. D. Frohning and C. W. Kohlpaintner, in Applied Homogeneous
Catalysis with Organometallic Compounds:
Handbook in Two Volumes, eds. B. Cornils and W. A. Herrmann,
VCH, Weinheim, 1996, vol. 1, pp. 27–104.
1
0.5 cm NMR tube and analysed. H NMR (25 ЊC, CD2Cl2):
δ 7.54 (bm, 6H; CH), 7.37 (bm, 6H; CH), 7.25 (bm, 12H; CH),
1.84 (s, 3H; CH3), 1.81 (s, 3H; CH3), 1.26 (s, 9H; t-butyl) and
1.22 (s, 9H; t-butyl). 31P-{1H} NMR (25 ЊC, CD2Cl2): δ 16.4
(1J(Pt,P) = 2172 Hz, satellites). 31P-{1H} NMR (Ϫ73 ЊC,
a Comprehensive
CD2Cl2):
δ
18.1 (d, 1J(Pt,P) = 3046 (satellites), 2J(P,P) =
J. Chem. Soc., Dalton Trans., 2000, 2105–2112
2111

View Article Online
4 T. J. Devon, G. W. Phillips, T. A. Puckette, J. L. Stavinoha and
J. J. Vanderbilt (to Eastman Kodak), U.S. Pat., 4,694,109, 1987
(Chem. Abstr., 1988, 108, 7890).
5 C. P. Casey, G. T. Whiteker, M. G. Melville, L. M. Petrovich,
J. A. Gavney, Jr. and D. R. Powell, J. Am. Chem. Soc., 1992, 114,
5535.
6 M. Kranenburg, Y. E. M. van der Burgt, P. C. J. Kamer, P. W. N. M.
van Leeuwen, K. Goubitz and J. Fraanje, Organometallics, 1995, 14,
3081.
7 L. A. van der Veen, P. H. Keeven, G. C. Schoemaker, J. N. H. Reek,
P. C. J. Kamer, P. W. N. M. van Leeuwen, M. Lutz and A. L. Spek,
Organometallics, 2000, 19, 872.
35 P. C. J. Kamer, J. N. H. Reek and P. W. N. M. van Leeuwen,
Chemtech, 1998, 28, 27.
36 P. Dierkes and P. W. N. M. van Leeuwen, J. Chem. Soc., Dalton
Trans., 1999, 1519.
37 M. Kranenburg, P. C. J. Kamer, D. Vogt and P. W. N. M. van
Leeuwen, J. Chem. Soc., Chem. Commun., 1995, 2177.
38 W. Goertz, P. C. J. Kamer, P. W. N. M. van Leeuwen and D. Vogt,
Chem. Commun., 1997, 1521.
39 M. Kranenburg, P. C. J. Kamer and P. W. N. M. van Leeuwen,
Eur. J. Inorg. Chem., 1998, 25.
40 L. A. van der Veen, P. K. Keeven, P. C. J. Kamer and P. W. N. M. van
Leeuwen, Chem. Commun., 2000, 333.
8 E. Billig, A. G. Abatjoglou and D. R. Bryant (to Union Carbide),
Eur. Pat., 213,639, 1987 (Chem. Abstr., 1987, 107, 7392r).
9 A. van Rooy, P. C. J. Kamer, P. W. N. M. van Leeuwen, K. Goubitz,
J. Fraanje, N. Veldman and A. L. Spek, Organometallics, 1996,
15, 835.
10 M. E. Broussard, B. Juma, S. G. Train, W.-J. Peng, S. A. Laneman
and G. Stanley, Science, 1993, 260, 1784.
11 P. M. Burke, J. M. Garner, W. Tam, K. A. Kreutzer, A. J. J. M.
Teunissen, C. S. Snijder and C. B. Hansen (to DSM/Du Pont),
WO Pat., 97/33854, 1997 (Chem. Abstr., 1997, 127, 294939r).
12 L. A. van der Veen, P. C. J. Kamer and P. W. N. M. van Leeuwen,
Angew. Chem., Int. Ed., 1999, 38, 336.
41 S. Gauthier and J. M. J. Frechet, Synthesis, 1987, 383.
42 K. McWilliams and J. W. Kelly, J. Org. Chem., 1996, 61, 7408.
43 L. A. van der Veen, P. C. J. Kamer and P. W. N. M. van Leeuwen,
Organometallics, 1999, 18, 4765.
44 N. K. Bliznyuk, G. S. Levskaya and E. N. Matyukhina, J. Gen.
Chem. USSR, 1965, 35, 1253.
45 W. L. Lewis, C. D. Lowry and F. H. Bergheim, J. Am. Chem. Soc.,
1921, 43, 893.
46 L. A. van der Veen, M. D. K. Boele, F. R. Bregman, P. C. J. Kamer,
P. W. N. M. van Leeuwen, K. Goubitz, J. Fraanje, H. Schenk and
C. Bo, J. Am. Chem. Soc., 1998, 120, 11616.
47 R. Whyman, Chem. Commun., 1970, 230.
13 Y. Kawabata, T. Hayashi and I. Ogata, J. Chem. Soc., Chem.
Commun., 1979, 462.
14 C.-Y. Hsu and M. Orchin, J. Am. Chem. Soc., 1975, 97, 3553.
15 I. Schwager and J. F. Knifton, J. Catal., 1976, 45, 256.
16 T. Hayashi, Y. Kawabata, T. Isoyama and I. Ogata, Bull. Chem. Soc.
Jpn., 1981, 54, 3438.
17 F. Ancillotti, M. Lami and M. Marchionna, J. Mol. Catal., 1991,
66, 37.
18 F. Ancillotti, M. Lami and M. Marchionna, J. Mol. Catal., 1990,
63, 15.
19 P. Meessen, D. Vogt and W. J. Keim, J. Organomet. Chem., 1998,
551, 165.
20 S. C. Tang and L. Kim, J. Mol. Catal., 1982, 14, 231.
21 P. W. N. M. van Leeuwen, C. F. Roobeek, R. L. Wife and J. H. G.
Frijns, J. Chem. Soc., Chem. Commun., 1986, 31.
22 P. W. N. M. van Leeuwen and C. F. Roobeek, New J. Chem., 1990,
14, 487.
48 B. L. Booth, M. J. Else, R. Fields and R. N. Haszeldine,
J. Organomet. Chem., 1971, 27, 119.
49 P. Chini and S. Martinengo, Inorg. Chim. Acta, 1969, 3, 315.
50 B. R. James, D. Mahajan, S. J. Rettig and G. M. Williams,
Organometallics, 1983, 2, 1452.
51 A. Castellanos-Páez, S. Castillón, C. Claver, P. W. N. M. van
Leeuwen and W. G. J. de Lange, Organometallics, 1998, 17, 2543.
52 E. Paumard, A. Mortreux and F. Petit, J. Chem. Soc., Chem.
Commun., 1989, 1380.
53 C. J. Cobley and P. G. Pringle, Inorg. Chim. Acta, 1997, 265, 107.
54 The averaged value of 2J(117Sn,P) and 2J(119Sn,P).
55 G. K. Anderson and G. J. Lumetta, Organometallics, 1985, 4,
1542.
56 G. K. Anderson and R. Kumar, Inorg. Chem., 1984, 23, 4064.
57 G. P. C. M. Dekker, A. Buijs, C. J. Elsevier, K. Vrieze, P. W. N. M.
van Leeuwen, W. J. J. Smeets, A. L. Spek, Y. F. Wang and C. H.
Stam, Organometallics, 1992, 11, 1937.
23 C. Botteghi, S. Paganelli, U. Matteoli, A. Scrivanti, R. Ciociaro and
L. M. Venanzi, Helv. Chim. Acta, 1990, 73, 284.
24 F. Agbossou, J.-F. Carpentier and A. Mortreux, Chem. Rev., 1995,
95, 2485.
25 J. K. Stille, H. Su, P. Brechot, G. Parrinello and L. S. Hegedus,
Organometallics, 1991, 10, 1183.
26 A. Scrivanti, C. Botteghi, L. Toniolo and A. Berton, J. Organomet.
Chem., 1988, 344, 261.
27 M. Gomez, G. Muller, D. Sainz, J. Sales and X. Solans,
Organometallics, 1991, 10, 4036.
58 S. Gladiali, L. Pinna, C. G. Arena, E. Rotondo and F. Faraone,
J. Mol. Catal., 1991, 66, 183.
59 C. Abu-Gnim and I. Amer, J. Chem. Soc., Chem. Commun., 1994,
115.
60 C. Basoli, C. Botteghi, M. A. Cabras, G. Chelucci and M.
Marchetti, J. Organomet. Chem., 1995, 488, C20.
61 J. T. Carlock, Tetrahedron, 1984, 40, 185.
62 C. P. Casey, E. L. Paulsen, E. W. Beuttenmueller, B. R. Proft,
B. A. Matter and D. R. Powell, J. Am. Chem. Soc., 1999, 121, 63.
63 Actually, their reports are the first examples in which the catalytic
performance can be related to the natural bite angle of diphosphine
ligands.36
28 I. Tóth, T. Kegi, C. J. Elsevier and L. Kollar, Inorg. Chem., 1994,
33, 5708.
29 W. R. Rocha and W. B. De Almeida, Organometallics, 1998, 17,
1961.
30 B. Cornils, in Hydroformylation. Oxo Synthesis, Roelen Reaction,
ed. J. Falbe, Springer-Verlag, Berlin, 1980, pp. 1–225.
31 H. C. Clark and J. A. Davies, J. Organomet. Chem., 1981, 213, 503.
32 C. A. Tolman, Chem. Rev., 1977, 77, 313.
64 G. K. Anderson and R. J. Cross, Acc. Chem. Res., 1984, 17, 67.
65 N. Koga and K. Morokuma, Chem. Rev., 1991, 91, 823.
66 G. P. C. M. Dekker, C. J. Elsevier, K. Vrieze and P. W. N. M. van
Leeuwen, Organometallics, 1992, 11, 1598.
67 CAChe WorkSystem, Version 4.0, CAChe Scientific Inc.,
Beaverton, OR, 1997.
33 C. A. Tolman, J. Am. Chem. Soc., 1970, 92, 2953.
34 C. P. Casey and G. T. Whiteker, Isr. J. Chem., 1990, 30, 299.
68 According to the Cambridge Crystallographic Database, a dis-
tance of 2.415 Å is typical for Rh–As bond lengths.
2112
J. Chem. Soc., Dalton Trans., 2000, 2105–2112