Fluorous derivatives of [Rh(COD)(dppe)]BX4 (X=F, Ph): Synthesis, physical studies and application in catalytic hydrogenation of 1-alkenes and 4-alkynes

Article abstract of DOI:10.1016/S0040-4020(02)00217-X

Fluorous derivatives of 1,2-bis(diphenylphosphino)ethane (dppe) (1), containing a para-(1H,1H,2H,2H-perfluoroalkyl)silyl function, were used in the synthesis of fluorous derivatives of [Rh(COD)(dppe)]BF₄. The single crystal X-ray crystallographic structure of [Rh(COD)(1a)]BPh₄ (7) was determined (1a=[CH₂P(C₆H₄-SiMe₂CH ₂CH₂C₆F₁₃-p)₂] ₂). Large differences in catalytic activity and selectivity (higher hydrogenation activity and lower isomerization activity) relative to [Rh(COD)(dppe)]BF₄ were observed in the hydrogenation of 1-octene using non-fluorous, silyl-substituted [Rh(COD)(2)]BF₄ (5; 2=[CH₂P(C₆H₄-SiMe₃-p) ₂]₂ or [Rh(COD)(3)]BF₄ (6; 3=[CH₂P(C₆H₄-SiMe₂C₈H ₁₇-p)₂]₂) and even more so with fluorous [Rh(COD)(1a)]BF₄ (4a). For 4a and 6, the presence of aggregates was demonstrated by dynamic light scattering (DLS) and cryogenic transmission electron microscopy (cryoTEM), which is most probably responsible for these differences. Similar differences between (fluorous) silylated catalysts and non-substituted [Rh(COD)(dppe)]BF₄ were observed in the semi-hydrogenation of 4-octyne. Recycling of highly fluorous catalyst [Rh(COD)(1c)]BF₄ (4c; 1c=[CH₂P(C₆H₄-Si(CH₂CH ₂C₆F₁₃)₃-p)₂] ₂) was investigated in two different fluorous biphasic solvent systems. The catalyst could be recycled with 97.5% (for PFMCH/acetone, 1:1 v/v) and >99.92% (for FC-75/hexanes, 1:3 v/v) retention of rhodium, respectively. The leaching of phosphorus was comparable to the leaching of rhodium (in PFMCH/acetone), showing that dissociation and leaching of free ligand does not take place for these rhodium diphosphine catalysts.

Full text of DOI:10.1016/S0040-4020(02)00217-X

TETRAHEDRON
Pergamon
Tetrahedron 58 02002) 3911±3922
Fluorous derivatives of [RhꢀCOD)ꢀdppe)]BX4 ꢀXˆF, Ph):
synthesis, physical studies and application in catalytic
hydrogenation of 1-alkenes and 4-alkynes
Elwin de Wolf,a Anthony L. Spek,b Bonny W. M. Kuipers,c Albert P. Philipse,c
Johannes D. Meeldijk,d Paul H. H. Bomans,e Peter M. Frederik,e Berth-Jan Deelmanf,p
and Gerard van Kotena
aDepartment of Metal-Mediated Synthesis, Debye Institute, Utrecht University, Padualaan 8, CH-3584 Utrecht, The Netherlands
bDepartment of Crystal- and Structural Chemistry, Bijvoet Centre for Biomolecular Research, Utrecht University, Padualaan 8,
CH-3584 Utrecht, The Netherlands
cVan't Hoff Laboratory for of Physical and Colloid Chemistry, Debye Institute, Utrecht University, Padualaan 8, CH-3584 Utrecht,
The Netherlands
dDebye Institute, Utrecht University, Padualaan 8, CH-3584 Utrecht, The Netherlands
eEM-unit, Universiteit Maastricht, P.O. Box 616, MD-6200 Maastricht, The Netherlands
fATOFINA Vlissingen B.V., P.O. Box 70, AB-4380 Vlissingen, The Netherlands
Received 31 October 2001; revised 21 February 2002; accepted 22 February 2002
AbstractÐFluorous derivatives of 1,2-bis0diphenylphosphino)ethane 0dppe) 01), containing a para-01H,1H,2H,2H-per¯uoroalkyl)silyl
function, were used in the synthesis of ¯uorous derivatives of [Rh0COD)0dppe)]BF4. The single crystal X-ray crystallographic structure
of [Rh0COD)01a)]BPh4 07) was determined 01aˆ[CH2P0C6H4±SiMe2CH2CH2C6F13-p)2]2). Large differences in catalytic activity and selec-
tivity 0higher hydrogenation activity and lower isomerization activity) relative to [Rh0COD)0dppe)]BF4 were observed in the hydrogenation
of 1-octene using non-¯uorous, silyl-substituted [Rh0COD)02)]BF4 05; 2ˆ[CH2P0C6H4±SiMe3-p)2]2 or [Rh0COD)03)]BF4 06;
3ˆ[CH2P0C6H4±SiMe2C8H17-p)2]2) and even more so with ¯uorous [Rh0COD)01a)]BF4 04a). For 4a and 6, the presence of aggregates
was demonstrated by dynamic light scattering 0DLS) and cryogenic transmission electron microscopy 0cryoTEM), which is most probably
responsible for these differences. Similar differences between 0¯uorous) silylated catalysts and non-substituted [Rh0COD)0dppe)]BF4 were
observed in the semi-hydrogenation of 4-octyne. Recycling of highly ¯uorous catalyst [Rh0COD)01c)]BF4 04c; 1cˆ[CH2P0C6H4-
Si0CH2CH2C6F13)3-p)2]2) was investigated in two different ¯uorous biphasic solvent systems. The catalyst could be recycled with 97.5%
0for PFMCH/acetone, 1:1 v/v) and .99.92% 0for FC-75/hexanes, 1:3 v/v) retention of rhodium, respectively. The leaching of phosphorus
was comparable to the leaching of rhodium 0in PFMCH/acetone), showing that dissociation and leaching offree ligand does not take place for
these rhodium diphosphine catalysts. q 2002 Elsevier Science Ltd. All rights reserved.
1. Introduction
0MeO)2PCH2CH2P0OMe)2 as ligand, i.e. K[Ru{0F15C7)-
C0O)CHC0O)0C7F15)}3] and in situ generated
[Rh0COD){CH2P0OCH2CH2C6F13)2}2]OTf, respectively.5
Possibly, even higher ¯uorous character is necessary to
dissolve these ionic complexes in ¯uorous solvents, since
in ¯uorous solvents with low dielectric constant, strong
solvent±solute interactions needed to separate ion pairs
are lacking. For neutral ¯uorous bidentate phosphine
complexes it was previously noted that at least eight
per¯uoroalkyl tails are necessary to achieve good solubility
in ¯uorous solvents.3
Diphosphine transition metal complexes are widely used as
catalyst in several types of homogeneous catalysis.1 Since
these complexes often contain expensive metals and 0chiral)
ligands, recycling of these complexes by ¯uorous phase
separation techniques is of major interest.2 Only a few tran-
sition metal complexes containing ¯uorous bidentate phos-
phine ligands are known in the literature,3 all of which
display suboptimal catalytic activity.4 Also, to the best of
our knowledge, only two ¯uorous phase-soluble ionic
complexes have been reported. These employ ¯uorous
derivatives of the 2,4-pentanedionate anion or
The use of the ±SiMe32b0CH2CH2±)b spacer 0bˆ1±3)6
enabled us to synthesize ¯uorous derivatives of 1,2-bis-
0diphenylphosphino)ethane 01a±c) containing up to 12
tails per molecule.7 Here, we report the use of diphosphines
1±3 in the synthesis of rhodium complexes, along with a
Keywords: ¯uorous biphasic catalysis; catalyst recycling; phosphine;
hydrogenation; rhodium.
Corresponding author; e-mail: b.j.deelman@chem.uu.nl
p
0040±4020/02/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020002)00217-X

3912
E. de Wolf et al. / Tetrahedron 58 :2002) 3911±3922
Table 1. NMR data of [Rh0COD)0dppe)]BF4 and its silyl-substituted 0non)¯uorous derivatives
Compound
31P NMR d 0ppm)a
1JRhP 0Hz)
103Rh NMR d 0ppm)a
[Rh0COD)0dppe)]BF4
5
6
4a
4bb
4cc
7
56.5
56.4
56.4
56.5
56.8
57.5
56.1
148.4
148.4
148.3
148.3
149.3
148.3
148.4
2513
2512
±
2509
±
±
±
a
In CDCl3, unless noted otherwise.
In C6D6/CF3C6H5, 1:1 0v/v).
In C6D6/FC-72, 1:1 0v/v).
b
c
single crystal X-ray crystallographic study of [Rh0COD)-
01a)]BPh4. These novel complexes were applied as catalyst
precursors in the hydrogenation of 1-alkenes and alkynes.
Moreover, their recycling using ¯uorous biphasic separation
was investigated. The large differences in catalytic per-
formance observed between the ¯uorous complexes and
their non-¯uorous counterparts stimulated us to study the
steric and electronic effects imposed by the per¯uoroalkyl
substitution as well as the aggregation behavior of these
complexes in polar solvents. To investigate the shape and
size of any aggregates, both dynamic light scattering 0DLS)
and cryogenic transmission electronmicroscopy 0cryoTEM)
were employed.
P distance etc.). Therefore, the similarity of the 103Rh NMR
chemical shifts of [Rh0COD)0dppe)]BF4, 4a and 5 indicates
that both the para-trimethylsilyl and a para-01H,1H,2H,2H-
per¯uoroalkyl)dimethylsilyl substituents have no signi®cant
consequences for the electronic environment and/or coordi-
nation geometry of the respective rhodium centers.
Interestingly, single crystals of [Rh0COD)01a)]BPh4 07)
suitable for X-ray structural analysis were obtained.
Complex 7 is, to the best of our knowledge, the ®rst ¯uorous
diphosphine±metal complex of which the molecular struc-
ture has been studied by single crystal X-ray crystallography
and is one of the few crystal structures of per¯uoroalkylated
phosphine complexes that have been reported.11 An ORTEP
representation showing the molecular geometry of 7 in the
solid state is presented in Fig. 1. Selected bond lengths and
angles as well as torsion angles are given in Table 2. Unfor-
tunately, this structure cannot be directly compared with its
non-¯uorous analogue [Rh0COD)0dppe)]X 0Xˆa weakly
coordinating anion), since no examples of these compounds
have been studied crystallographically. Therefore, com-
parison is made with rhodium±CODcomplexes containing
other ethanediyl bridged diphosphine ligands.
Ê
Rh±P bond lengths 02.27309) and 2.27208) A) are similar to
reported values for [Rh0COD)0L2)]1 0average Rh±P
2. Results and discussion
Ê
distanceˆ2.27 A for L2ˆCHIRAPHOS, DiPAMP or
DuPHOS).12 The P±Rh±P bite angle 082.403)8) of 7 is at
the low end of the range observed for the above mentioned
The ¯uorous complexes [Rh0COD)01a±c)]BF4 04a±c),
trimethylsilyl-substituted [Rh0COD)02)]BF4 05), dimethyl-
0octylsilyl) substituted [Rh0COD)03)]BF4 06) and
[Rh0COD)01a)]BPh4 07) were synthesized from
[Rh0COD)2]BF4 or [Rh0acac)0COD)], respectively, follow-
ing slightly modi®ed literature procedures.8 All compounds
1
were identi®ed by both H and 31P NMR spectroscopy and
elemental analyses and in some cases also by 19F NMR
spectroscopy 04a,b, 5). H NMR spectra of 4b and c at
1
room temperature in C6D6/FC-72 01:1 v/v) 0FC-72 is
mixture of per¯uorohexanes) as solvent system displayed
broad peaks with unresolved multiplicities.9
1
The 31P NMR chemical shifts and the JRhP coupling
constants of 4, 5, 6 and 7 were found to be in close agree-
ment with data found for non-substituted [Rh0COD)-
0dppe)]BF4 0Table 1). Also, the 103Rh NMR chemical shifts
of [Rh0COD)0dppe)]BF4, 4a and 5 were found to be very
similar. It is known that the 103Rh NMR chemical shift is a
very sensitive probe for both electronic and steric in¯u-
ences_.10 For example, shifts of 20±100 ppm are observed
upon small geometrical variations 0P±Rh±P bite angle, Rh±
Figure 1. ORTEP drawing of 7 at 30% probability level. Hydrogen atoms
and the BPh42 residue are omitted for clarity.

E. de Wolf et al. / Tetrahedron 58 :2002) 3911±3922
Table 2. Selected bond lengths 0A), angles 08) and torsion angles 08) with estimated standard uncertainties in parentheses for [Rh0COD)01a)]BPh4 07)
3913
Ê
Rh01)±P01)
Rh01)±P02)
2.27309)
2.27208)
2.2203)
2.2603)
2.2803)
2.2003)
2.1402)
2.1302)
C069)±C070)
C073)±C074)
C059)±C060)
C060)±C061)
C061)±C062)
C062)±C063)
C063)±C064)
1.3504)
1.4104)
1.4703)
1.6004)
1.4804)
1.5404)
1.5004)
Rh01)±C069)
Rh01)±C070)
Rh01)±C073)
Rh01)±C074)
Rh01)±Cg069±70)a
Rh01)±Cg073±74)a
P01)±Rh01)±P02)
P02)±Rh01)±Cg073±74)p
82.403)
169.105)
P01)±Rh01)±Cg069±70)p
173.205)
86.208)
Cg069±70)±Rh01)±Cg073±74)p
P01)±C065)±C066)±P02)
P01)±Rh01)±P02)±C066)
Rh01)±P02)±C066)±C065)
C011)±C012)±C013)±C014)
C012)±C013)±C014)±C015)
50.8018)
4.809)
235.3019)
164.5019)
C013)±C014)±C015)±C016)
C059)±C060)±C061)±C062)
C060)±C061)±C062)±C063)
C061)±C062)±C063)±C064)
15902)
5803)
166.2018)
5603)
16202)
a
Cg0x±y) denotes the center of gravity between C0x) and C0y).
ethanediyl bridged diphosphine complexes 083±878),12 but
is similar to the natural bite angle calculated for dppe
081.78).13 Possibly, the difference in P±Rh±P bite-angle
between CHIRAPHOS/DiPAMP and 1a is due to the
sterically more demanding environments in the bridge
0CHIRAPHOS) or on the aryl ring 0DiPAMP). The ®ve-
membered Rh±P±C±C±P chelate ring is close to an
envelope conformation with C065) lying out of the Rh±P±
are not coordinated perpendicular to the P±Rh±P plane.
This is also obvious from the angle between the planes
de®ned by Cg069±70)±Rh01)±Cg073±74) and P±Rh±P
013.108)8), which makes the environment around the
Rh-atom distorted square planar. This also results in forma-
tion of two longer 0Rh01)±C070), ±C073): 2.2603) and
Ê
2.2803) A) and two shorter 0Rh01)±C069), ±C074): 2.2203)
Ê
and 2.2003) A) Rh±C bonds, which can be explained by
Ê
C±C±P least squares plane by 0.3203) A. It has been noted
different steric hindrance exerted by the two aryl sub-
stituents of each phosphorus atom on the CODligand.
12
before that for [Rh0hfacac)0L2)] 0hfacacˆ[CF3C0O)CH-
C0O)CF3]2; L2ˆethanediyl-bridged diphosphine) no
systematic preference for either a twist or an envelope
conformation was found. Possibly, steric interactions and
packing effects determine the ultimate conformation of the
chelate ring.14
The four per¯uoroalkyl tails can be divided in two sets. The
three tails radiating away from Si01), Si02) and Si03) show
exclusive trans conformations along the carbon-skeleton
with dihedral angles ranging from 159 to 1708, resulting
in a helical 0pig-tail) structure. The fourth tail radiating
from Si04), however, adopts a conformation with two
dihedral angles 0C060)±C061) and C062)±C063) are 5803)
and 5603)8), which are typical for a gauche conformation.
The remaining dihedral angles are all trans conformations.
The C±C bond lengths in the CODring are normal, except
Ê
for C073)±C074) 01.4104) A) which is due to the rather poor
quality of the diffraction data. As reported for [Rh0COD)-
0CHIRAPHOS)]1, the CvC double bonds of the COD-ring
Figure 2. Packing diagram of 7 0F blue; P yellow; Rh green; B red; all other atoms are omitted for clarity).

3914
E. de Wolf et al. / Tetrahedron 58 :2002) 3911±3922
Scheme 1.
Indeed, the preferred conformation along the carbon chain
in a per¯uoroalkyl group in the solid state is trans.15
However, different dihedral angles close to the alternative
gauche conformations have also been observed in various
crystal structures of transition metal complexes with
¯uorous ligands, the conformational changes most probably
being induced by crystal packing effects.11 The irregular
C±C interatomic distances within the per¯uoroalkyl groups
are artifacts of the quality of the dataset.
The packing diagram of 7, shown in Fig. 2, clearly shows
the presence of ¯uorous and non-¯uorous regions in the
Figure 3. Conversion vs time plots of the hydrogenation of 1-octene using 0a) [Rh0COD)0dppe)]BF4, 0b) 5 or 0c) 4a. 0Legend: Bˆ1-octene; Xˆ2-octenes;
Oˆ3- and 4-octenes; Vˆn-octane; points are measured results from GC, solid line was ®tted to the datapoints).

E. de Wolf et al. / Tetrahedron 58 :2002) 3911±3922
Table 3. Observed quantities and kinetic modeling of the hydrogenation of 1-octene
3915
[Rh0COD)0dppe)]BF4
5
6
4a
para-Substituent
TOF50%
±H
902)
±SiMe3
2504)
0.8801)
±SiMe20C8H17)
2504)
0.8701)
±SiMe2CH2CH2Rf RfˆC6F13
5905)
0.9401)
a
cis/trans ratiob
0.2901)
8302)% 010 min)
Isomerizationc
5902)% 065 min)
5602)% 060 min)
3102)% 045 min)
k1obs 0min21
k2obs 0min21
)
)
1.005)
702)£101
201)
0.1007)
301)
1.307)
0.703)
201)
0.1007)
1.501)
0.2001)
0.501)
0.2008)
0.201)
0.1006)
302)£101
0.1008)
1.702)
0.1007)
0.401)
0.2007)
0.1006)
0.1007)
302)£101
0.602)
1.202)
0.201)
0.604)
0.201)
k22obs 0min21
)
k3obs 0min21
k4obs 0min21
)
)
k24obs 0min21
)
0.00108)
0.301)
krestobs 0min21
K 0M21
)
)
802)£101
Conditions: Tˆ408C, p0H2)ˆ1 bar, stirring speedˆ1100 rpm, catalyst/1-octeneˆ1:83, [Rh]ˆ6 mM. For the de®nition of the parameters, see Scheme 1. All
experiments were performed in duplo. Deviations are given in brackets.
Turn-over frequency in mole n-octane/mole of Rh/hour determined when 50% of the maximum amount of n-octane was formed.
a
b
Initial cis/trans-ratio of 2-octenes.
Maximum amount of cis- and trans-2-octene; time given in brackets.
c
crystal lattice. The per¯uoroalkyl parts of the molecules are
assembled in separate layers. These layers are alternated
with non-¯uorous layers which contain the organic and
charged parts of the molecules, i.e. this structure organizes
the borate, rhodium and phosphorus containing parts of the
molecules in single layers. This layered structure is clearly
related to the amphiphilic character of 7 and represents a
nice demonstration of the structural consequences of
¯uorous behavior in the solid state.
¯uorous 4a, are dif®cult to explain in terms of electronic
or steric effects. This is con®rmed by the very similar 103Rh
NMR data of [Rh0COD)0dppe)]BF4, 4a and 5 and the
analysis of the X-ray crystallographic structure of 7 0same
¯uorous cation as 4a).
Stimulated by the crystallographic structure of 7, which is
clearly indicative of the tendency of this compound to
aggregate, we wondered whether such `¯uorous self-
assembly' could also be operative in solution. To obtain
insight in the differences observed between [Rh0COD)-
0dppe)]BF4, [Rh0COD)03)]BF4 06) and [Rh0COD)01a)]BF4
04a) in solution, 5 mM solutions in acetone were studied by
dynamic light scattering 0DLS). Since the results of direct
deposit TEM 05±10 nm spheres) did not match the results
obtained by DLS, the solutions were studied by cryo
transmission electron microscopy 0cryoTEM), where the
sample is vitri®ed and kept in liquid nitrogen 0Table 4 and
Figs. 4 and 5). For para-dimethyloctylsilyl derivative 6, the
presence of aggregates was indeed established by both DLS
and cryoTEM 0Table 4). Similar and even larger structures
were observed for 4a. The sizes of the aggregates are within
the typical range for vesicle-like structures.16 Furthermore,
the low transmission intensity in the center of the aggregates
0Figs. 4 and 5) indicates that also the inside of the
aggregates contains rhodium species, excluding a vacuole-
like structure with a single bilayer and a solvent-containing
interior. Surprisingly, small aggregates 030±50 nm) were
even found for non-substituted [Rh0COD)0dppe)]BF4
using DLS. However, this could not be con®rmed by
cryoTEM.
The hydrogenation of 1-octene in acetone as solvent
served as a model reaction to study the consequences of
silyl-functionalization on the catalytic reactivity by
comparing the respective precatalysts [Rh0COD)-
0dppe)]BF4, 4a, 5 and 6 0Scheme 1). The resulting conver-
sion vs time plots 0Fig. 3) were ®tted to a simpli®ed kinetic
model 0see Section 4). The kinetic, catalytic and thermo-
dynamic data thus obtained are represented in Table 3.
From inspection of these data it is clear that trimethylsilyl-
05) and dimethyloctylsilyl substitution 06) result in similar
rate constants for hydrogenation but reduced isomerization
obs
obs
obs
obs
rate constants 0k1 and k3 relative to k2 and k4
)
compared to parent [Rh0COD)0dppe)]BF4. It can be
concluded that 5 and 6 relative to [Rh0COD)0dppe)]BF4
are more effective in hydrogenation because less alkene is
®rst isomerized to internal alkene isomers that subsequently
are more dif®cult to hydrogenate. Interestingly, both the cis-
2-octene/trans-2-octene ratio after 15 min and the alkene
association constant are increased for 5 and 6.
Surprisingly, the effects observed upon trimethylsilyl- or
dimethyloctylsilyl-substitution are even more pronounced
upon 1H,1H,2H,2H-per¯uoroalkyldimethylsilyl-function-
alisation 04a), i.e. compared to 5 and 6, the hydrogenation
activity increased and the isomerization activity decreased.
This is again accompanied by a higher initial cis/trans ratio
for the 2-octene isomers formed, but also by a signi®cantly
higher alkene association constant K.
Table 4. Sizes of aggregates in 5 mM acetone solutions of derivatives of
[Rh0COD)0dppe)]BF4
Compound
cmc 0mM)a
Radius 0nm)
CryoTEM
DLS
[Rh0COD)0dppe)]BF4
4a
6
±
0.7
1.2
30±50
100±400
110±130
Not observed
90±360
110±300
The large differences observed between the non-¯uorous
parent compound [Rh0COD)0dppe)]BF4 and non-¯uorous
silyl-substituted 5 and 6, and to an even higher extent
a
Critical micelle concentration determined from conductivity measure-
ments.

3916
E. de Wolf et al. / Tetrahedron 58 :2002) 3911±3922
Figure 5. CryoTEM micrographs of 5 mM acetone solutions of 6.
Figure 4. CryoTEM micrographs of 5 mM acetone solutions of 4a.
the square root of the concentration 0Fig. 6). For both
compounds, the plot consists of two linear parts intersecting
at a concentration which, following the interpretation of
Furthermore, also some smaller, less well identi®able struc-
tures were observed for 4a that were either free ¯oating in
solution or attached to the larger spheres. These structures
0see Fig. 4, boxed area), are most likely bilayered segments,
their cross diameter 0ca. 4 nm) being in agreement with
approximately twice the length of a 1H,1H,2H,2H-per-
¯uorooctyl-moiety. Related bilayered, vesicle-like struc-
tures of ¯uorinated compounds in organic solvents were
observed before for per¯uoroalkyl diester derivatives of
HOOC±CH0NHR)±0CH2)2COOH 0Rˆalkyl or oligo glycol
moiety) in cyclohexane.17 It has been noted previously that
the higher rigidity of per¯uoroalkyl groups relative to
normal alkyl groups, results in a smaller loss of entropy
upon aggregation of the molecules containing these alkyl
groups.18 Consequently, compounds containing per¯uoro-
alkyl groups have a higher tendency to aggregate than the
corresponding non-¯uorinated species. The irregular
surface of the spheres in Fig. 4 might be due to an equi-
librium between the ¯uorous rhodium species in the
aggregate and freely dissolved in solution.
conductivity data in Ref. 19, can be identi®ed as 0cmc)1/2
.
The linear decrease in the molar conductivity L with c1/2
observed above and below the cmc for both 4a and 6, is
consistent with the behavior of strong electrolytes 0the inter-
actions between ions at higher concentrations retard their
mobility20). The slope of the plot and consequently the
extrapolated y intercept 0L0), is in both cases lower at
concentrations above the cmc, indicating that the aggre-
gated rhodium complex has a lower L0. The molar conduc-
tivities above the cmcs are still considerable, indicating that
the aggregates carry signi®cant net charge. Interestingly, the
cmc of ¯uorous 4a is lower compared to the cmc of non-
¯uorous 6 0Table 4, Fig. 5). This is in agreement with the
¯uorous compound's higher tendency to aggregate.
The differences observed between the three types of
catalysts 0[Rh0COD)0dppe)]BF4, non-¯uorous silyl-
substituted 5 and 6, and ¯uorous 4a) can be explained by
the differences in aggregation behavior as described above.
For aqueous systems, there is ample evidence in the
literature to conclude that aggregation of homogeneous
Critical micelle concentrations 0cmc) of 4a and 6 in acetone
could be determined from conductivity measurements
0Table 4), plotted as the equivalent conductance against

E. de Wolf et al. / Tetrahedron 58 :2002) 3911±3922
3917
Figure 6. Conductivity measurements of acetone solutions of 4a 0V) and 6 0B) for the determination of their critical micelle concentration.
catalysts can have bene®cial effects on catalytic activity
and/or selectivity.21,22 Especially noteworthy are positive
effects on enantioselectivity.21,22 For aqueous emulsions
containing micelle-forming surfactants, the solubility of
apolar gasses 0like CH4, O2, Ar, C3H8) can be 1.2 to 20
times higher compared to that in pure water.23 This
increased solubility of gaseous reactants could explain the
positive effect of these surfactants on the catalytic activity in
a number of cases.
result in the more pronounced difference in catalytic
performance relative to reference compound [Rh0COD)-
0dppe)]BF4.
Except for the positive effects on the hydrogenation rate
constants and the higher K, alkene isomerization 0k2),
which most likely involves allylic C±H activation of the
alkene,25 appears to be disfavoured in the 0¯uorous)
aggregates. In general, the nature of the effect responsible
for different 0enantio)selectivities in micellular environ-
ments remains unclear and it is also not immediately
obvious why isomerization should be more dif®cult in an
aggregate.21,22
In addition, the more lipophilic inside of an aggregate
0lower dielectric constant) can cause a higher concentration
of lipophilic substrate molecules as well as a lowered
solvent concentration in the interior or vicinity of the aggre-
gate compared to that of the bulk of the surrounding 0more
polar) solvent. Both effects can lead to an apparent higher
association constant for a particular catalyst±substrate
adduct. In our speci®c case, the rhodium complex is less
well stabilized by solvent. Consequently, stabilization by
association of substrate becomes more important, which
leads to an apparently higher association constant K for
the bulk of the reaction mixture.16 Such modulation of
substrate±catalyst interaction appears to be a quite general
phenomenon in metalloaggregates.16 However, it should be
noted that, since the association constant K is an average for
all alkenes present in the bulk of the reaction mixture and
trans-alkenes compared to cis- and primary alkenes have in
general a lower tendency to complex to metals,24 the higher
association constant found for 4a, 5 and 6 relative to
[Rh0COD)0dppe)]BF4 could be partly due to the lower
proportion of internal trans-alkenes formed.
The difference in initial cis/trans-2-octene ratio observed
between the 0¯uorous) silyl-substituted catalysts
00.87±0.94) and reference catalyst [Rh0COD)0dppe)]BF4
00.29) could be explained by the higher isomerization
activity of the latter. A thermodynamic 1:3 cis/trans-2-
octene mixture was observed for [Rh0COD)0dppe)]BF4,
while the kinetically 1:1 mixture, as predicted by the
p-allyl mechanism,25 was detected for the silyl-substituted
catalysts with lower isomerization activity.
To further explore the impact of aggregation on the catalytic
selectivity, the semi-hydrogenation of 4-octyne was studied
0Scheme 2). For all catalysts tested, 4-octyne was converted
with moderate selectivity to cis-4-octene, which was subse-
quently hydrogenated to n-octane or ®rst isomerized to other
internal alkenes and then hydrogenated. Similar behavior
had been observed previously for the hydrogenation of
2-hexyne by [Rh0nbd)0dppe)]BF4.26 Activities and
maximum selectivities towards cis-4-octene for different
catalytic systems are given in Table 5. Much like the
observations described above for the hydrogenation of
1-octene, the hydrogenation activity increased upon
trimethylsilyl-functionalization and increased even further
upon attachment of dimethyl01H,1H,2H,2H-per¯uoro-
octyl)silyl-groups. The higher hydrogenation activity may
explain the relatively higher proportion of overhydrogena-
tion to n-octane and a consequently lower selectivity
When translated to the case of an aggregated catalyst in a
polar but organic acetone solution, it can be argued that the
higher catalytic activity observed for non-¯uorous 5 and 6
and ¯uorous 4a is caused by the combined effects of a
locally increased dihydrogen concentration and an aggre-
gate-induced higher af®nity of the catalyst for the alkene
substrate. The latter effect appears to be the dominant one
0Table 3). The larger size and more apolar environment of
the ¯uorous aggregates of 4a 0relative to 5 and 6), possibly

3918
E. de Wolf et al. / Tetrahedron 58 :2002) 3911±3922
Scheme 2.
towards cis-4-octene for the functionalized catalysts. As
observed for the hydrogenation of 1-octene, the ¯uorous
catalysts again produced lower amounts of isomerization
products. A further increase of the number of per¯uoroalkyl
groups to 8 04b) resulted in a lower hydrogenation activity
but similar maximum selectivity compared to 4a.
the conversion was found. Leaching of both rhodium and
phosphorus was higher for the system containing acetone in
contrast to the very low leaching values obtained when
hexane was used as organic layer. For PFMCH/acetone
the amount of Rh and P found in the organic layer corre-
sponded to a 1:2 molar ratio, indicating that leaching of the
intact catalyst and not of free ligand occurred. This shows
that the use of bidentate diphosphines solves the problem of
ligand-leaching, which was encountered previously for
monodentate phosphines in Wilkinson-type catalysts.27
For FC-75/hexane, a small excess of phosphorus was
found in the organic layer, most likely indicating that
some catalyst had decomposed as no rhodium was found.
The amount of leaching of the ¯uorous solvent into the
organic layer, which is a known issue in ¯uorous biphasic
catalysis, was ca. 20% per cycle for FC-75/hexane but
was not observed for PFMCH/acetone due to their low
miscibility.
Different solvent systems were needed to perform hydro-
genations using the highly ¯uorous complex 4c. Since it is
known that the solvent in¯uences the catalytic activity, no
direct conclusions regarding the effect of per¯uoroalkyl-
functionalization by comparison with the activities and
selectivities of 4a,b, 5 and [Rh0cod)0dppe)]BF4 can be
drawn. Three types of solvent system were considered: a
biphasic combination of a non-coordinating ¯uorous solvent
0PFMCH) and a coordinating organic solvent 0acetone), a
biphasic combination of a potentially coordinating, ¯uorous
solvent 0FC-75) and a non-coordinating, organic solvent
0hexanes) and ®nally a monophasic combination also
consisting of hexanes and FC-75 but in a 1:3 volumetric
ratio. Using acetone as the organic phase, the hydrogenation
activity of 4c was, despite the biphasic nature of the solvent
system, comparable to the activity of 4b in pure acetone.
With apolar hexane as the organic solvent, the observed
activity was lower than that observed using acetone.
Furthermore, the monophasic version of this reaction
medium resulted in higher activity than the biphasic system,
possibly due to the absence of mass transfer limitations. It
should also be noted that under the nearly homogeneous
conditions, 4c displayed comparable activity to non-
¯uorous [Rh0COD)0dppe)]BF4 and a similar maximum
selectivity compared to 4a.
3. Conclusions
Fluorous diphosphines 1a±c and silyl-substituted, non-
¯uorous diphosphines 2 and 3 were used for the synthesis
of ¯uorous and non-¯uorous rhodium complexes. The latter
were shown to be active in the hydrogenation of 1-octene
and 4-octyne.
Important differences were observed in the catalytic per-
formance of the three types of catalyst, i.e. ¯uorous catalysts
4a±c, non-¯uorous silyl substituted catalyst 5±6 and
[Rh0COD)0dppe)]BF4. Formation of aggregates, con®rmed
by DLS and cryoTEM for [Rh0COD)01a)]BF4 04a) and
[Rh0COD)03)]BF4 06) are most likely responsible for the
lower isomerization activities and higher alkene-
associations constants observed for silyl-substituted, non-
¯uorous catalysts 5 and 6. The similar but more pronounced
effects observed for ¯uorous catalysts 4 in acetone, possibly
due to the larger size of the aggregates and/or a higher
degree of ordering, can be explained in this way as well.
Recycling experiments using 4c were performed in
PFMCH/acetone 01:1) or FC-75/hexane 01:3) as solvent
systems 0Table 6). The higher activity of 4c in PFMCH/
acetone is re¯ected in the higher conversions obtained
after 2.5 h. In both systems, the catalyst can be ef®ciently
recycled, since no signi®cant drops in conversions are
observed and in PFMCH/acetone even a small increase of
Table 5. Results of hydrogenation of 4-octyne 0conditions: Tˆ408C, p0H2)ˆ1 bar, stirring speedˆ1100 rpm, catalyst/4-octyneˆ1:83, catalyst loading: 6 mmol
per mL of reaction volume) using parent compound [Rh0COD)0dppe)]BF4, 4a±c or 5
Catalyst precursor
Solvent systema
TOFb
Max. selectivityc
Max. isomersd
[Rh0COD)0dppe)]BF4
5
Me2CO
Me2CO
Me2CO
Me2CO
PFMCH/Me2CO 01:1)
FC-75/n-hexane 01:1)
FC-75/n-hexane 01:3)
1
1
1
1
2
2
1f
32
43
77
36
31
12
23
83 0185 min)
74 0105 min)
70 075 min)
67 0115 min)
61 0185 min)
70 0310 min)
68 0205 min)
31 0240 min)
29 0180 min)
18 0115 min)
16 0165 min)
15 0295 min)
17 0340 min)e
17 0295 min)
4a
4b
4c
4c
4c
a
Solvents and number of liquid phases are given. Volume ratios are given in brackets
TOFˆturn over frequency in mole of substrate/mol of Rh/hour at 50% conversion.
Towards cis-4-octene.
Maximum amount of isomerization products of cis-4-octene, time given in brackets.
The reaction was not complete after 340 min.
Most of the catalyst is emulsi®ed in the otherwise homogeneous reaction mixture.
b
c
d
e
f

E. de Wolf et al. / Tetrahedron 58 :2002) 3911±3922
3919
Table 6. Recycling and reuse of 4c in 4-octyne hydrogenation for two
¯uorous 0biphasic) solvent systems
the diffusion coef®cient was estimated from 5t 0tˆhalf
time). The hydrodynamic radius, Rh, was calculated
using the Stokes±Einstein relation, D0 ˆ kT=6phRh; in
which kT is the thermal energy and h the solvent
viscosity.
Cycle
Conversion after 2.5 h 0%)
PFMCH/acetone 01:1)
FC-75/hexane 01:3)
1
2
3
89
93
59
63
57
4.3. CryoTEM measurements
94
2.5% 036 ppm)
2.2% 019 ppm)
Rh-leachinga
P-leachinga
,0.08% 0,1 ppm)b
0.4% 03 ppm)
Specimen preparation and vitri®cation was done with
a
Vitrobot 0see http://www.maastrichtinstruments.nl/
projects/vr/index.html). Quantifoil grids 0200 mesh copper
grids R 2/2, Quantifoil Micro Tools GmbH, Jena, Germany)
were used for sample preparation. In the chamber of the
Vitrobot, 3 mL of the sample was applied to the grid, the
grid was blotted twice 0each blotting action lasted 2 s) and
shot in liquid nitrogen for vitri®cation. Liquid ethane, an
ef®cient coolant for aqueous solutions/suspensions, was
found incompatible with the solvent used 040% acetone);
apparently the vitri®ed material dissolved in liquid ethane.
Liquid nitrogen, with a much slower heat exchange than
liquid ethane, proved to be inert and despite poor heat
exchange the solvent matrix was vitri®ed under the experi-
mental conditions. After preparation of a vitri®ed thin ®lm
on a quantifoil grid, the grid was mounted in a cryo-holder
0Gatan 626) and transferred to a Philips CM 12 transmission
microscope for observation at 21708C and 120 kV. Low
dose conditions were used to take micrographs to ensure
that the area of interest had minimal beam exposure before
micrographs were taken.
For conditions, see Table 5. Phase separation was performed at 08C.
Determined by ICP-AAS analysis of the organic layer of the ®rst cycle.
Below the detection limit 01 ppm).
a
b
It may be clear that formation of aggregates of `light
¯uorous'28 catalyst molecules in polar organic solvents
opens up a new and interesting ®eld for application of
¯uorous catalysts in homogeneous catalysis.
Furthermore, we have demonstrated that highly ¯uorous
rhodium catalyst 4c can be ef®ciently recycled in two
different ¯uorous biphasic systems, resulting in 2.5% 0for
PFMCH/acetone) and ,0.08% 0for FC-75/hexane) leaching
of rhodium. It is important to note that the use of
diphosphines solves the problem of dissociation and leach-
ing of free ligand, which was observed for catalytic systems
containing monophosphines.
4. Experimental
4.1. General
4.4. X-Ray structure determination of 7
X-ray data were collected on a Nonius TurboCAD4 diffract-
Ê
ometer 0Rotating anode, Mo Ka, lˆ0.71073 A, Tˆ150 K,
All experiments were performed in a dry dinitrogen
atmosphere using standard Schlenk techniques. Solvents
were stored over sodium benzophenone ketyl or CaH2 and
were distilled before use. Fluorinated solvents 0c-CF3C6F11
0Lancaster), FC-72 and CF3C6H5 0Acros)) were degassed
and stored under dinitrogen atmosphere. Derivatives of
umaxˆ21.58) for
a
yellowish plate shaped crystal
00.5£0.5£0.05 mm3) glued with per¯uor oil on top of a
Lindemann glass capillary. The crystal was found to diffract
poorly at higher diffraction angles and with structured
re¯ection pro®les, resulting in data of limited quality,
though suf®cient for the purpose of this study. The structure
was solved with Patterson techniques and completed with
difference Fourier syntheses. Correction for absorption was
deemed unnecessary. Hydrogen atoms were included at
calculated positions. Full matrix least squares re®nement
on F2 0SHELXL-9731) converged at Rˆ0.15 0wR2ˆ0.37,
Sˆ1.05, 676 parameters, 12410 re¯ections, max and min
dppe 01±2),7 [Rh0COD)0dppe)]BF4,8 [Rh0acac)0COD)],29
30
[Rh0COD)2]BF4
and 1,2-bis[bis0p-bromophenyl)phos-
phino]ethane7 were prepared according to previously
reported procedures. Other chemicals were obtained from
commercial suppliers 0Acros, Aldrich, Lancaster) and used
as delivered. Microanalyses were carried out by H. Kolbe,
Mikroanalytisches Laboratorium, MuÈlheim an der Ruhr. 19F
NMR spectra were externally referenced against CFCl3
0dˆ0 ppm) and 31P NMR spectra were referenced against
H3PO4 0dˆ0 ppm).
Ê 23
residual density 0.88 and 20.66 e A ). Rhodium and
silicon were re®ned with anisotropic parameters, all other
non-hydrogen atoms with isotropic displacement
parameters. Geometrical calculations and the ORTEP
illustration were done with PLATON.32 Pertinent data:
Ê
4.2. DLS measurements
monoclinic, space group P2 =c, Zˆ4, aˆ14.8704013) A,
1
Ê
Ê
bˆ41.04404) A,
cˆ22.69102) A,
bˆ128.52307)8
Solutions were ®ltered through a P4 ®lter 0pore sizeˆ10±
16 mm). Dynamic light scattering 0DLS) results were
obtained using an argon ion laser 0Spectra Physics Series
2000) operating at 514.5 nm. Auto correlation functions
were measured with a Malvern Multibit K7025 128
points correlator. Diffusion coef®cients D0 were obtained
from a second order cumulant ®t using auto correlation
functions obtained from angles between 35 and 1208. If a
single exponential ®t of the autocorrelation function
could not be obtained, possibly due to polydispersity,
C74H72F52P2RhSi4.C24H20B,
MWˆ2545.74,
Dxˆ
1.5606 g cm23, m0Mo Ka) 0.37 mm21. Full details have
been deposited with the Cambridge Crystallographic Data
Centre 0CCDC 172832).
4.4.1. 1,2-Bis[bis{4-ꢀꢀdimethyloctyl)silyl)phenyl}phos-
phino]ethane ꢀ3). A solution of 0.74 g 1,2-bis[bis0p-bromo-
phenyl)phosphino]ethane 01.04 mmol) in THF 020 mL) was
cooled to 2908C and 5.5 mL of a solution of t-BuLi 01.5 M;
8.3 mmol) was added. After stirring for 30 min while keep-

3920
E. de Wolf et al. / Tetrahedron 58 :2002) 3911±3922
ing the temperature lower than 2608C, 0.86 g 04.15 mmol)
of C8H17SiMe2Cl was added. After heating to room
temperature, all volatiles were evaporated. The residue
was dissolved in CH2Cl2/degassed water. The organic
layer was separated, dried on MgSO4 and evaporated to
dryness, yielding 1.04 g 093%) of a clear light brown oil.
Anal. calcd for C66H112P2Si4: C 73.41, H 10.45; Found: C
4.4.6. [RhꢀCOD)ꢀ3)]BF4 ꢀ6). Compound 3 00.14 g;
0.13 mmol) dissolved in CH2Cl2 020 mL) was added to a
solution of [Rh0COD)2]BF4 053 mg; 0.13 mmol) in
CH2Cl2 010 mL). After stirring for 30 min, all volatiles
were removed in vacuo and the product, 0.17 g 095%) of a
red oil, was not puri®ed any further. Anal. calcd for
C74H124BF4P2RhSi4: C 64.51, H 9.07; Found: C 61.59, H
8.14; 1H NMR 0200 MHz, CDCl3): d 0.31 0s, 24H), 0.86 0m,
20H), 1.26 0m, 48H), 2.2±2.6 0m, 12H), 4.95 0br, 4H), 7.60
0m, 16H).
1
69.67, H 10.93; H NMR 0200 MHz, CDCl3): d 0.28 0s,
3
24H), 0.77 0m, 8H), 0.92 0t, JHHˆ3.4 Hz, 12H), 1.29 0m,
48H), 2.16 0ps t, Jˆ3.5 Hz, 4H), 7.35±7.54 0m, 16H); 31P
NMR 0CDCl3, 81.0 MHz): d 212.0.
4.4.7. [RhꢀCOD)ꢀ1a)]BPh4 ꢀ7). Compound 1a 00.82 g;
0.41 mmol) dissolved in THF 015 mL) was added to a solu-
tion containing [Rh0acac)0COD)] 0128 mg; 0.41 mmol) and
NaBPh4 00.28 g; 0.82 mmol). After stirring for 10 min and
evaporating all volatiles in vacuo, CH2Cl2 010 mL) was
added. After ®ltration and drying, 0.76 g 073%) of an orange
solid was obtained. Crystals suitable for X-ray structure
analysis were grown from CH2Cl2/hexane. Anal. calcd for
C98H92BF52P2RhSi4: C 46.43, H 3.66, P 2.44; Found: C
4.4.2. [RhꢀCOD)ꢀ1a)]BF4 ꢀ4a). Compound 1a 00.76 g;
0.38 mmol) dissolved in CH2Cl2 030 mL) was added to a
solution of [Rh0COD)2]BF4 0153 mg; 0.38 mmol) in
CH2Cl2 010 mL). After stirring for 10 min and evaporating
all solvent in vacuo, the residue was washed with benzene.
Drying of the precipitate yielded 0.62 g 070%) of an orange
solid. Anal. calcd for C74H72BF56P2RhSi4: C 38.42, H 3.14,
P 2.68; Found: C 38.49, H 3.20, P 2.48; 1H NMR 0200 MHz,
CDCl3): d 0.39 0s, 24H), 1.03 0m, 8H), 1.91±2.42 0m, 20H),
4.94 0m, 4H), 7.64 0m, 16H); 19F NMR 0282.4 MHz,
CDCl3): d 2154.0 0m, 4F), 2126.6 0m, 8F), 2123.6 0m,
8F), 2123.4 0m, 8F), 2122.5 0m, 8F), 2116.3 0m, 8F),
281.3 0m, 12F).
1
46.05, H 3.70, P 2.62; H NMR 0200 MHz, CDCl3): d
0.38 0s, 24H), 1.02 0m, 8H), 2.10 0m, 20H), 4.93 0m, 4H),
6.76 0m, 4H), 6.86 0m, 8H), 7.30 0m, 8H), 7.44 0m, 8H), 7.60
0m, 8H); 31P NMR 0CDCl3, 81.0 MHz):
01JRhPˆ148.4 Hz).
d
56.1
4.5. Catalytic reactions and recycling
4.4.3. [RhꢀCOD)ꢀ1b)]BF4 ꢀ4b). Compound 1b 03.24 g;
0.97 mmol) dissolved in CF3C6H5 015 mL) was added to a
solution of [Rh0COD)2]BF4 0393 mg; 0.97 mmol) in CH2Cl2
010 mL). After evaporation of all volatiles, the crude oil was
dissolved in a biphasic system consisting of CH2Cl2 010 mL)
and PFMCH 025 mL). Phase separation and drying of the
¯uorous layer gave 3.36 g of an orange to red oil 095%).
Anal calcd for C102H76BF108P2RhSi4: C 33.64, H 2.10, P
1.70, F 56.34; Found: C 33.46, H 1.96, P 1.63, F 56.28;
1H NMR 0200 MHz, C6D6/FC-72, 1:1 v/v): d 0.12 0br,
12H), 0.85 0br, 16H), 1.5±2.6 0br, 28H), 4.86 0br, 4H),
7.57 0br, 16H); 19F NMR 0282.4 MHz, CDCl3/CF3C6H5,
1:1 v/v): d 2152.2 0m, 4F), 2126.9 0m, 16F), 2123.8 0m,
16F), 2123.5 0m, 16F), 2122.5 0m, 16F), 2116.4 0m, 16F),
281.8 0m, 24F).
Three hydrogenation reactions were performed simul-
taneously in a Micronic 12 Place Heated Carousel Reaction
Station. In a typical experiment, 36 mmol of catalyst was
dissolved in 6 mL of the appropriate solvent or solvent
mixture. The inert nitrogen atmosphere was replaced by
hydrogen and the mixture was stirred at 1100 rpm at 408C
for 1 h.33 Subsequently, substrate 00.44 mL) and n-decane
00.40 mL, internal standard) were added. Samples taken at
regular time intervals were analyzed by GC. In recycling
experiments, the reaction mixture was cooled to 08C after
2.5 h while stirring was stopped and the hydrogen
atmosphere was maintained. The organic layer was removed
and new substrate and n-decane were added. Subsequently,
the reaction mixture was heated to 408C and a new cycle
was started.
4.4.4. [RhꢀCOD)ꢀ1c)]BF4 ꢀ4c). Compound 1c 01.57 g;
0.336 mmol) dissolved in PFMCH 010 mL) was added to
a solution of [Rh0COD)2]BF4 0136 mg; 0.336 mmol) in
CH2Cl2 015 mL), while stirring vigorously 01500 rpm).
Phase separation and evaporation to dryness yielded
1.66 g of an orange to red oil 099%). Anal. calcd for
C130H80BF160P2RhSi4: C 31.42, H 1.62, P 1.25; Found: C
31.32, H 1.56, P 1.31; 1H NMR 0200 MHz, C6D6/FC-72, 1:1
v/v): d 1.08 0br, 24H), 1.96 0br, 36H), 4.79 0br, 4H), 7.57
0br, 16H).
4.6. Kineticmodel for [RhꢀL )]1-catalyzed
2
hydrogenation ꢀL2ˆchelating ligand) and kinetic ®tting
The rate equation for each step x in Scheme 1 0xˆ1, 2, 22,
3, 4, 24, rest) can be de®ned as:34
kxobsK‰RhŠ ‰alkeneŠ
tot
rx ˆ
1 1 K‰alkeneŠ
which corresponds to the mechanistic pathways depicted in
Scheme 3. [Rh]tot is the total concentration of all rhodium
species in solution, [alkene] is the concentration of the
respective alkene involved in each step x and kx is the
observed rate constant for step x.
4.4.5. [RhꢀCOD)ꢀ2)]BF4 ꢀ5). Compound
2 00.23 g;
0.32 mmol) dissolved in CH2Cl2 020 mL) was added to a
solution of [Rh0COD)2]BF4 0131 mg; 0.32 mmol) in
CH2Cl2 05 mL). Removal of all volatiles, washing with
pentane and drying in vacuo yielded 0.31 g of an orange
solid 096%). Anal. calcd for C46H68BF4P2RhSi4: C 56.09,
H 6.96; Found: C 53.73, H 6.68; 1H NMR 0200 MHz,
CDCl3): d 0.32 0s, 24H), 2.31 0m, 12H), 4.96 0br, 4H),
7.61 0m, 16H); 19F NMR 0282.4 MHz, CDCl3): d 2154.0
0m, 4F).
obs
The 0constant) hydrogen pressure 0total pressure kept at
obs
1 bar) is included in kx of the hydrogenation steps 0xˆ1,
3 or rest). In this model, it is assumed for simplicity that all
alkenes have an average association constant K. Further-
more, no distinction between cis- and trans-alkenes was

E. de Wolf et al. / Tetrahedron 58 :2002) 3911±3922
3921
Scheme 3.
1998, 37, 1174. 0g) Cornils, B. Angew. Chem., Int. Ed. Engl.
1997, 36, 2057.
made. For [RH0COD)0dppe)]BF4, Winstein±Holness
kinetics35 were applied to the 3- and 4-octene hydrogenation
steps by de®ning one overall rate constant 0kobsrest). In the
case of catalysts 4a, 5 and 6, no signi®cant amount of 4-
octenes was observed.
3. 0a) Hope, E. G.; Kemmitt, R. D. W.; Stuart, A. M. J. Chem.
Soc., Dalton Trans. 1998, 3765. 0b) Kainz, S.; Koch, D.;
Baumann, W.; Leitner, W. Angew. Chem., Int. Ed. Engl.
1997, 36, 1628.
4. Hope, E. G.; Kemmitt, R. D. W.; Paige, D. R.; Stuart, A. M.
J. Fluorine Chem. 1999, 99, 197.
Conversion vs time plots were ®tted, using the program
Micromath Scientist for Windows, version 2.0, to the
model as shown in Scheme 1. Fitting was proceeded until
the squared difference between measured and calculated
concentrations was lower than 0.005.
5. 0a) Klement, I.; LuÈtjens, H.; Knochel, P. Angew. Chem., Int.
Ed. Engl. 1997, 36, 1454. 0b) Haar, C. M.; Huang, J.; Nolan,
S.; Petersen, J. L. Organometallics 1998, 17, 5018.
6. 0a) Kleijn, H.; Jastrzebski, J. T. B. H.; Gossage, R. A.;
Kooijman, H.; Spek, A. L.; van Koten, G. Tetrahedron
1998, 54, 1145. 0b) Richter, B.; de Wolf, E.; van Koten, G.;
Deelman, B.-J. J. Org. Chem. 2000, 65, 3885.
Acknowledgements
7. de Wolf, E.; Richter, B.; Deelman, B.-J.; van Koten, G. J. Org.
Chem. 2000, 65, 5424.
ATOFINA Vlissingen B. V., NWO0ALS) and the Dutch
Ministry of Economic Affairs 0SENTER/BTS) are grate-
fully thanked for the ®nancial support. Ing. J. M. Ernsting
0University of Amsterdam) is thanked for the 103Rh NMR
measurements.
8. Schrock, R. R.; Osborn, J. A. J. Am. Chem. Soc. 1971, 93,
2397.
9. No signi®cant changes in line broadening were observed upon
heating to 508C.
10. Leitner, W.; BuÈhl, M.; Fornika, R.; Six, C.; Baumann, W.;
Dinjus, E.; Kessler, M.; KruÈger, C.; Ru®nska, A.
Organometallics 1999, 18, 1196.
References
11. 0a) Bhattacharyya, P.; Croxtall, B.; Fawcett, J.; Fawcett, J.;
Gudmunsen, D.; Hope, E. G.; Kemmitt, R. D. W.; Paige, D. R.;
Russell, D. R.; Stuart, A. M.; Wood, D. R. W. J. Fluorine
Chem. 2000, 101, 247. 0b) Fawcett, J.; Hope, E. G.; Kemmitt,
R. D. W.; Paige, D. R.; Russell, D. R.; Stuart, A. M. J. Chem.
Soc., Dalton Trans. 1998, 3751. 0c) Guillevic, M. A.; Arif,
1. See for example: Brunner, H. In Applied Homogeneous
Catalysis with Organometallic Compounds, Cornils, B.,
Herrmann, W. A., Eds.; VCH: Weinheim, 1996; p. 201
Chapter 2.2.
2. 0a) de Wolf, E.; van Koten, G.; Deelman, B.-J. Chem. Soc.
Rev. 1999, 28, 37. 0b) Hope, E. G.; Stuart, A. M. J. Fluorine
Chem. 1999, 100, 75. 0c) Cavazzini, M.; Montanari, F.; Pozzi,
G.; Quici, S. J. Fluorine Chem. 1999, 94, 183. 0d) Rutherford,
Â
A. M.; Horvath, I. T.; Gladysz, J. A. Angew. Chem., Int. Ed.
Engl. 1997, 36, 1612. 0d) Fawcett, J.; Hope, E. G.; Kemmitt,
R. D. W.; Paige, D. R.; Russell, D. R.; Stuart, A. M.; Cole-
Hamilton, D. J.; Payne, M. J. Chem. Commun. 1997, 1127.
Â
D.; Juliette, J. J. J.; Rocaboy, C.; Horvath, I. T.; Gladysz, J. A.
Â
Catal. Today 1998, 42, 381. 0e) Horvath, I. T. Acc. Chem. Res.
1998, 31, 641. 0f) Curran, D. P. Angew. Chem., Int. Ed. Engl.
Â
0e) Guillevic, M. A.; Rocaboy, C.; Arif, A. M.; Horvath, I. T.;
Gladysz, J. A. Organometallics 1998, 17, 707.

3922
E. de Wolf et al. / Tetrahedron 58 :2002) 3911±3922
12. Burk, M. J.; Feaster, J. E.; Nugent, W. A.; Harlow, R. L. J. Am.
Chem. Soc. 1993, 115, 10125. 0b) Giovanetti, J. S.; Kelly,
C. M.; Landis, C. R. J. Am. Chem. Soc. 1993, 115, 4040.
0c) Ball, R. G.; Payne, N. C. Inorg. Chem. 1977, 16, 1187.
13. Casey, C. P.; Whiteker, G. T. Isr. J. Chem. 1990, 30, 299.
Grassert, I.; Paetzold, E.; Meisel, R.; Drexler, K.; Fuhrmann,
H. Coord. Chem. Rev. 1999, 185±186, 585. 0b) Oehme, G. In
Aqueous-Phase Organometallic Catalysis, Concepts and
Applications, Cornils, B., Herrmann, W. A., Eds.; Wiley-
VCH: Weinheim, 1999; p. 193.
23. 0a) Hai, M.; Han, B.; Yan, H.; Han, Q. Phys. Chem. Chem.
Phys. 1999, 1, 445. 0b) Radwan, N.; King, Jr., A. D. J. Colloid
Interf. Sci. 1997, 194, 120. 0c) Roy, S.; Mehra, A.; Bhowmick,
D. J. Colloid Interf. Sci. 1997, 196, 53.
È
14. Angermund, K.; Baumann, W.; Dinjus, E.; Fornika, R.; Gorls,
H.; Kessler, M.; KruÈger, C.; Leitner, W.; Lutz, F. Chem. Eur.
J. 1997, 3, 755.
15. Bunn, C. W.; Howells, E. R. Nature 1954, 174, 549.
16. Scrimin, P. In Supramolecular Control of Structure and
Reactivity, Hamilton, A. D., Ed.; Perspectives in Supra-
molecular Chemistry; Wiley: Chichester, 1996; Vol. 3, p. 101.
17. Ishikawa, Y.; Kuwahara, H.; Kunitake, T. J. Am. Chem. Soc.
1989, 111, 8530.
24. Graff, J.; Sanner, R. D.; Wrighton, M. S. Organometallics
1982, 1, 837.
25. 0a) Buriak, J. M.; Klein, J. C.; Herrington, D. G.; Osborn, J. A.
Chem. Eur. J. 2000, 6, 139. 0b) Budzelaar, P. H. M.; Moonen,
N. N. P.; de Gelder, R.; Smits, J. M. M.; Gal, A. W. Eur. J.
Inorg. Chem. 2000, 753.
18. 0a) Wang, K.; Karlsson, G.; Almgren, M.; Asakawa, T.
J. Phys. Chem. 1999, 103, 9237. 0b) Blokzijl, W.; Engberts,
J. B. F. N. Angew. Chem., Int. Ed. Engl. 1993, 32, 1545.
19. Guo, W.; Brown, T. A.; Fung, B. M. J. Phys. Chem. 1991, 95,
1829.
20. The molar conductivity L is de®ned as LˆL02Sc1/2 where S
is linearly dependant on the molar conductivity at in®nite
dilution 0L0). Fernandez-Prini, R. In Physical Chemistry of
Organic Solvent Systems, Covington, A. K., Dickinson, T.,
Eds.; Plenum: London, 1973 Chapter 5.
26. Schrock, R. R.; Osborn, J. A. J. Am. Chem. Soc. 1976, 98,
2143.
27. Richter, B.; Spek, A. L.; van Koten, G.; Deelman, B.-J. J. Am.
Chem. Soc. 2000, 122, 3945.
28. Zhang, Q.; Luo, Z.; Curran, D. P. J. Org. Chem. 2000, 65,
8866.
29. Cramer, R. J. Am. Chem. Soc. 1964, 86, 217.
30. Schenck, T. G.; Downes, J. M.; Milne, C. R. C.; Mackenzie,
P. B.; Boucher, H.; Whelan, J.; Bosnich, B. Inorg. Chem.
1985, 24, 2334.
21. 0a) Grassert, I.; Oehme, G. J. Organomet. Chem. 2001, 621,
158. 0b) Grassert, I.; Schmidt, U.; Ziegler, S.; Fischer, C.;
Oehme, G. Tetrahedron: Asymmetry 1998, 9, 4193.
31. Sheldrick, G. M. SHELXL: Program for Crystal Structure
È È
Re®nement; University of Gottingen: Gottingen, Germany,
1979.
È
0c) Selke, R.; Holtz, J.; Riepe, A.; Borner, A. Chem. Eur. J.
32. Spek, A. L. PLATON: A Multipurpose Crystallographic Tool;
Utrecht University, Utrecht, The Netherlands, 2001.
33. Shorter pre-hydrogenation times resulted in induction times
for hydrogenation of 1-octene. Pre-hydrogenation for longer
than 1 h did not change the activity of the catalyst any further.
34. 0a) Landis, C. R.; Halpern, J. J. Organomet. Chem. 1983, 250,
485±490. 0b) Reiss, J.; Het¯ejs, J. React. Kinet. Catal. Lett.
1986, 31, 309±313.
1998, 4, 769. 0d) Yonehara, K.; Ohe, K.; Uemara, S. J. Org.
Chem. 1999, 64, 9381. 0e) Goedheijt, M. S.; Hanson, B. E.;
Reek, J. N. H.; Kamer, P. C. J.; van Leeuwen, P. W. N. M.
J. Am. Chem. Soc. 2000, 122, 1650. 0f) Hanson, B. E.; Ding,
H.; Kohlpainter, C. W. Catal. Today 1998, 42, 421. 0g) Ding,
H.; Hanson, B. E.; Bartik, T.; Bartik, B. Organometallics
1994, 13, 3761.
22. For reviews on micellar catalysis, see: 0a) Oehme, G.;
35. Winstein, S.; Holness, N. J. J. Am. Chem. Soc. 1955, 77, 5562.