Fluorous versions of Wilkinson's catalyst. Activity in fluorous hydrogenation of 1-alkenes and recycling by fluorous biphasic separation

Article abstract of DOI:10.1021/ja993562v

Two novel fluorous tris(arylphosphine)rhodium(I) chloride complexes RhCl[P{C₆H₄-p-SiMe₂(CH₂)₂C(n)F(2n+1)}₃]₃ (3a: n = 6, 3b: n = 8) being fluorous analogues of Wilkinson's catalyst show high activity in hydrogenation of 1-alkenes under single phase fluorous conditions and can be effectively recycled using the new concept of fluorous biphasic separation. For comparison the p-(trimethylsilyl)-substituted derivative RhCl[P(C₆H₄-p- SiMe₃)₃]₃ (3c) has also been synthesized and the X-ray structure of dimeric [Rh(μ-Cl){P(C₆H₄p-SiMe₃)₃}₂]₂ (4c) was determined. A comparison of the catalytic activity of the fluorous catalysts and the nonfluorous derivatives under identical conditions revealed a decreasing activity in the order 3c > 3a > RhCl-(PPh₃)₃ > 3b. The recycling efficiency of the new catalysts (> 98% for 3b) is much better than expected on the basis of the fluorous phase affinity of the free phosphine ligands themselves and is in fact the result of the much higher fluorous phase affinity of the phosphine-containing rhodium species that are present during and after catalysis. Hence it is demonstrated that assembling multiple fluorous ligands in a specific configuration at one metal center can drastically improve the fluorous phase affinity of the resulting complexes.

Full text of DOI:10.1021/ja993562v

J. Am. Chem. Soc. 2000, 122, 3945-3951
3945
Fluorous Versions of Wilkinson’s Catalyst. Activity in Fluorous
Hydrogenation of 1-Alkenes and Recycling by Fluorous Biphasic
Separation
Bodo Richter,† Anthony L. Spek,‡ Gerard van Koten,† and Berth-Jan Deelman*,§
Contribution from the Debye Institute, Department of Metal-Mediated Synthesis, Utrecht UniVersity,
Padualaan 8, 3584 CH Utrecht, The Netherlands, BijVoet Center for Biomolecular and Structural
Chemistry, Utrecht UniVersity, Padualaan 8, 3584 CH Utrecht, The Netherlands, and Elf Atochem
Vlissingen B.V., P.O. Box 70, 4380 AB Vlissingen, The Netherlands
ReceiVed October 4, 1999
Abstract: Two novel fluorous tris(arylphosphine)rhodium(I) chloride complexes RhCl[P{C6H4-p-SiMe2-
(CH2)2CnF2n+1}3]3 (3a: n ) 6, 3b: n ) 8) being fluorous analogues of Wilkinson’s catalyst show high activity
in hydrogenation of 1-alkenes under single phase fluorous conditions and can be effectively recycled using
the new concept of fluorous biphasic separation. For comparison the p-(trimethylsilyl)-substituted derivative
RhCl[P(C6H4-p-SiMe3)3]3 (3c) has also been synthesized and the X-ray structure of dimeric [Rh(µ-Cl){P(C6H4-
p-SiMe3)3}2]2 (4c) was determined. A comparison of the catalytic activity of the fluorous catalysts and the
nonfluorous derivatives under identical conditions revealed a decreasing activity in the order 3c > 3a > RhCl-
(PPh3)3 > 3b. The recycling efficiency of the new catalysts (> 98% for 3b) is much better than expected on
the basis of the fluorous phase affinity of the free phosphine ligands themselves and is in fact the result of the
much higher fluorous phase affinity of the phosphine-containing rhodium species that are present during and
after catalysis. Hence it is demonstrated that assembling multiple fluorous ligands in a specific configuration
at one metal center can drastically improve the fluorous phase affinity of the resulting complexes.
Introduction
ability of fluorinated solvents to form mono- or biphasic systems
with conventional organic solvents, above and below the
consolute temperature, respectively.4 The major advantage is
that the catalytic reaction can now be performed above the
consolute temperature under truly homogeneous conditions
whereas biphasic catalyst separation can be accomplished by
cooling.
A number of catalytic reactions employing FBS-based
separation techniques have been reported: i.e., hydroformyl-
ation,3b-d,5 hydroboration6 and epoxidation of alkenes,7 oxidation
of hydrocarbons,8 and cross-coupling and allylic substitution
reactions.9 However, reported turnover frequencies (TOF) and
Catalyst recycling and reuse can be of crucial importance in
homogeneous catalysis if toxic and/or expensive metal catalysts
are involved.1 To achieve efficient catalyst recovery, while
maintaining the high activity and selectivity of nonimmobilized
systems, various concepts for catalyst immobilization in the
presence of a mobile substrate phase have been developed.2 In
liquid biphasic catalysis the catalyst is immobilized in one of
the two phases of a biphasic system consisting of either water
and an organic solvent2a,b or an ionic liquid and an organic
solvent.2c Furthermore membrane techniques based on den-
drimer- or other macromolecule-attached metal catalysts have
been developed.2d However, a major disadvantage of these
techniques is the reduced activity due to mass-transport limita-
tion as a result of phase boundaries present between reactants
and catalysts.
(3) (a) Vogt, M. Ph.D. Thesis, Rheinisch-Westfa¨lische Technische
Hochschule, Aachen, Germany, 1991. (b) Horva´th, I. T.; Ra´bai, J. Science
1994, 266, 72-75. (c) Horva´th, I.; Ra´bai, J. U.S. Patent 5463082, 1995.
(d) Horva´th, I. Acc. Chem. Res. 1998, 31, 641-650. (e) Keim, W.; Vogt,
M.; Wasserscheid, P.; Driessen-Ho¨lscher, B. J. Mol. Catal. A: Chem. 1999,
139, 171-175.
(4) De Wolf, E.; Deelman, B.-J.; Van Koten, G. Chem. Soc. ReV. 1999,
28, 37-41 and references therein.
(5) (a) Kainz, S.; Koch, D.; Baumann, W.; Leitner, W. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 1628-1630. (b) Horva´th, I. T.; Kiss, G.; Cook, R.
A.; Bond, J. E.; Stevens, P. A.; Ra´bai, J.; Mozeleski, E. J. J. Am. Chem.
Soc. 1998, 120, 3133-3142.
(6) (a) Juliette, J. J. J.; Horva´th, I. T.; Gladysz, J. A. Angew. Chem., Int.
Ed. Engl. 1997, 36, 1610-1612. (b) Juliette, J. J. J.; Rutherford, D.; Horva´th,
I. T.; Gladysz, J. A. J. Am. Chem. Soc. 1999, 121, 2696-2704.
(7) (a) Pozzi, G.; Banfi, S.; Manfredi, A.; Montanari, F.; Quici, S.
Tetrahedron 1996, 52, 11879-11888. (b) Pozzi, G.; Cinato, F.; Montanari,
F.; Quici, S. Chem. Commun. 1998, 877.
(8) (a) Klement, I.; Knochel, P. Synlett 1995, 1113-1115. (b) Vincent,
J.-M.; Rabion, A.; Yachandra, V. K.; Fish, R. H. Angew. Chem., Int. Ed.
1997, 36, 2346-2349. (c) Klement, I.; Lu¨tjens, H.; Knochel, P. Angew.
Chem., Int. Ed. Engl. 1997, 36, 1454-1456. (d) Pozzi, G.; Montanari, F.;
Quici, S. Chem. Commun. 1997, 69. (e) Guillevic, M.-A.; Horva´th, I. T.;
Gladysz, J. A. Angew. Chem., Int. Ed. Engl. 1997, 36, 1612-1614.
In Fluorous Biphasic Systems (FBS), first applied by Horva´th
and Vogt3 as an elegant alternative to aqueous biphasic catalyst
separation, this problem is not present because of the unique
† Debye Institute.
‡ Bijvoet Center for Biomolecular and Structural Chemistry.
§ Elf Atochem Vlissingen B.V., currently based at the Debye Institute.
E-mail: b.j.deelman@chem.uu.nl. Fax: +31 30 2523615.
(1) Cornils, B.; Herrmann, W. A. In Applied Homogeneous Catalysis
with Organometallic Compounds; Cornils, B., Herrmann, W. A., Eds.;
Weinheim: New York, 1996; Chapter 4.1, pp 1167-1197.
(2) (a) Cornils, B.; Herrmann, W. A. In Applied Homogeneous Catalysis
with Organometallic Compounds; Cornils, B., Herrmann, W. A., Eds.;
Weinheim: New York, 1996; Chapter 3.1, pp 577-600. (b) Chauvin, Y.;
Olivier-Bourbigou, H. CHEMTECH 1995, 25, 26. (c) Chauvin, Y.;
Mussmann, L.; Olivier, H. Angew. Chem., Int. Ed. Engl. 1995, 34, 2698-
2700. (d) Knapen, J. W. J.; van der Made, A. W.; de Wilde, J. C.; van
Leeuwen, P. W. N. M.; Wijkens, P.; Grove, D. M.; van Koten, G. Nature
1994, 372, 659-663.
10.1021/ja993562v CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/06/2000

3946 J. Am. Chem. Soc., Vol. 122, No. 16, 2000
Richter et al.
Scheme 1
Figure 1. Fluorous triphenylphosphine derivative with silicon as the
branching point and spacer group.
turnover numbers (TON) of fluorous rhodium(I)-catalyzed
hydrogenation of terminal and cyclic alkenes which employ
fluorous alkylphosphines (TON ) 98-120 in 16 h)10a or aryl
phosphites (TON ) 16-1073 in 24 h)10b are low and fall far
below the activity of Wilkinson’s catalyst RhCl(PPh3)3 (TOF
is ca. 150-600 h-1 for 1-alkenes).11,12,16b One possible explana-
tion could be that the donor properties of the phosphine ligand
are influenced in a negative sense by the electron-withdrawing
effect of the perfluoro tail. Also in olefin hydroformylation the
activity of the Rh(I)/P[CH2CH2(CF2)5CF3]3 system was found
to be 1 order of magnitude lower than that of Rh(I)/PPh3.5b This
latter observation is not unexpected because of the demonstrated
improvement in the activity of aryl- versus alkylphosphine Rh
complexes.11 Stimulated by these considerations we became
interested in using previously developed fluorous triarylphos-
phines13 P[C6H4-p-SiMe2(CH2)2CnF2n+1]3}3 (Figure 1) to obtain
analogues of Wilkinson’s catalyst12 and to investigate their
potential advantages in the context of homogeneous hydrogena-
tion of 1-alkenes. The weakly electron donating silicon atom,
we reasoned, could compensate for the strongly electron
withdrawing effect of the fluorotail(s) on the aryl ring preventing
a negative influence on the donor properties of the phosphorus
atom. Furthermore it has the added advantage that it can serve
as a branching point for connecting up to three fluoro tails per
aryl ring and provides a means to enhance the fluorous character
of the phosphine.13
fluorous Rh(I)-complexes 3a,b and the nonfluorous derivative
3c were synthesized from [(COD)RhCl]2 (COD ) 1,5-cyclooc-
tadiene, Scheme 1). Another efficient method to obtain fluorous
3a,b is the successive substitution of the PPh3 ligands of RhCl-
(PPh3)3 by their fluorous derivatives (Scheme 1).
The reaction of [(COD)RhCl]2 with 6 equiv of 1a-c in
benzene was monitored by 31P NMR and found to lead initially
to complexes (COD)RhCl(PR3) (2a-c) (δP 30.6, d). A mono-
nuclear structure of compounds 2 is proposed on the basis of
1
the JRh,P coupling constant (e.g., 151 Hz for 2a), which is
characteristic for monomeric complexes of this type rather than
1
dimeric alternatives; i.e., a JRh,P coupling constant of 188 Hz
Results and Discussion
was reported for dimeric [(COE)RhClL]2 (COE ) cyclooctene)
whereas for monomeric (COD)RhClL (L ) P(H)[CH(SiMe3)2]2)
a 1JRh,P of 151 Hz was found.14a Complexes 2 were obtained as
dark red viscous oils (2a,b) contaminated with 1 and 3 or as
yellow solid (2c). Complex 2c, being less soluble in n-pentane
than 3c, was easily purified by recrystallization. If allowed to
react further for 12-15 h complexes 3a,b were obtained as dark
red viscous oils, which separated from the aromatic solvent and
contained 10 to 30% of uncoordinated phosphine. However,
decantation of the upper organic phase and washing of the
remaining oil with toluene or benzene yielded analytically pure
complexes 3a and 3b in reasonable to high yields (44 and 87%,
respectively). These novel complexes were characterized by 1H,
31P, and 13C NMR (2a) spectroscopy.
From the chemical shift and the JRh,P and JP,P values for 3 in
comparison with RhCl(PPh3)3 and RhCl[P(C6H4-p-Me)3]3 (Table
1) it is clear that similar square-planar structures are retained
in solution. In particular from the very similar JRh,P coupling
constants it can be concluded that the character of the Rh-P
bond is not influenced significantly by the presence of the
perfluoro tails, thereby reflecting comparable electronic sur-
roundings for the rhodium centers in 3 and nonfluorous
derivatives 3c, RhCl(PPh3)3, and RhCl[P(C6H4-p-Me)3]3. Hence
similar catalytic activity would be expected. Furthermore, the
formation of complexes 3 via substitution of the COD ligand
in 2 indicates that phosphines 1 belong to the group of sterically
less demanding phosphines such as PPh3, PPhMe2, and P(nBu)3
since more bulky phosphines are known to prohibit the
formation of compounds with cis-positioned ligands.14
Synthesis. Using previously developed fluorous and non-
fluorous silyl-substituted triarylphosphines 112,13 (Figure 1),
(9) (a) Betzemeier, B.; Knochel, P. Angew. Chem., Int. Ed. Engl. 1997,
36, 2623-2624. (b) Caroll, M. A.; Holmes, A. B. Chem. Commun. 1998,
1395-1396. (c) Kling, R.; Sinou, D.; Pozzi, G.; Chopin, A.; Quignard, F.;
Busch, S.; Kainz, S.; Koch, D.; Leitner, W. Tetrahedron Lett. 1998, 39,
9439-9442.
(10) (a) Rutherford, D.; Juliette, J. J. J.; Rocaboy, C.; Horva´th, I. T.;
Gladysz, J. A. Catal. Today 1998, 42, 381-388. (b) Haar, C. M.; Huang,
J.; Nolan, S. P.; Petersen, J. L. Organometallics 1998, 17, 5018-5024.
(11) (a) Chaloner, P. A.; Esteruelas, M. A.; Joo´, F.; Oro, L. A.
Homogeneous Hydrogenation; Kluwer: Boston, 1994; pp 8-9. (b) Mon-
telatici, A.; van der Ent, A.; Osborn, J. A.; Wilkinson, G. J. Chem. Soc.
(A) 1968, 1054-1058. (c) Jardine, F. H. In Prog. Inorg. Chem.1981, 28,
117-131.
(12) (a) Richter, B.; Deelman, B.-J.; van Koten, G. European Patent Appl.
98203308.6. (b) Richter, B.; Deelman, B.-J.; van Koten, G. J. Mol. Catal.
(A), Chem. 1999, 145, 317. (c) Deelman, B.-J.; Richter, B.; Van Koten, G.
CERC3 Young Chemists Workshop “Homogeneous Catalysis”; University
of Rostock, April 21-24, 1999; Book of abstracts. (d) Deelman, B.-J.;
Richter, B.; Van Koten, G. 218th American Chemical Society National
Meeting; New Orleans, August 22-26, 1999, New Orleans, Louisiana;
American Chemical Society: Washington, DC, 1999; Book of abstracts.
(13) Richter, B.; de Wolf, E.; Deelman, B.-J.; van Koten, G. J. Org.
Chem. Accepted for publication.
(14) (a) Murray, B. D.; Hope, H.; Hvoslef, J.; Power, P. P. Organome-
tallics 1984, 3, 657-663. (b) Denise, B.; Pannetier, G. J. Organomet. Chem.
1978, 148, 155.
(15) (a) Tolman, C. A.; Meakin, P. Z.; Lindner, D. L.; Jesson, J. P. J.
Am. Chem. Soc. 1974, 96, 2762-2774. (b) Halpern, J.; Okamoto, T.;
Zakhariev A. J. Mol. Catal. 1976, 2, 65-69.
(16) (a) Young, J. F.; Osborn, J. A.; Jardine, F. H.; Wilkinson, G. Chem.
Commun. 1965, 131-132. (b) Osborne, J. A.; Jardine, F. H.; Young, J. F.;
Wilkinson, G. J. Chem. Soc. (A) 1966, 1711-1732. (c) Osborn, J. A.;
Wilkinson, G. Inorg. Synth. 1967, 10, 67-71. (d) Jardine, F. H.; Osborn,
J. A.; Wilkinson, G. J. Chem. Soc. (A) 1967, 1574-1578.

Fluorous Versions of Wilkinson’s Catalyst
J. Am. Chem. Soc., Vol. 122, No. 16, 2000 3947
Table 1. 31P NMR Data and Assignment for the Fluorous Wilkinson-Type and Related Complexesa
trans-P-Rh-X (X ) Cl, O, H)
trans-P-Rh-P
complex
δP (multiplicity)
1JRh,P (Hz)
2JP,P (Hz)
δP (multiplicity)
1JRh,P (Hz)
2JP,P (Hz)
(COD)RhCl[P{C6H4-p-SiMe2Rf6}3] (2a)b
(COD)RhCl[P{C6H4-p-SiMe2Rf8}3] (2b)b
(COD)RhCl[P{C6H4-p-SiMe3}3] (2c)b
30.6 (d)
30.7 (d)
30.7 (d)
48.1 (dt)
48.0 (dt)
48.0 (dt)
47.4 (dt)
46.2 (dt)
51.1 (dt)
52.8 (d)
50.0 (d)
51.4 (d)
51.8 (d)
49.5 (d)
28.2 (dt)
26.3 (dt)
28.0 (dt)
28.1 (dt)
20.7 (dt)
19.9 (dt)
19.7 (dt)
19.3 (dt)
151
154
151
190
190
192
192
189
188
195
198
196
195
196
151
151
152
152
90
b
RhCl(PPh3)3
37.9
37.8
37.8
37.6
38
31.2 (dd)
31.4 (dd)
31.4 (dd)
31.0 (dd)
30.2 (dd)
36.2 (dd)
145
143
145
143
143
142
37.8
37.6
37.4
37.5
38
RhCl[P(C6H4-p-SiMe2-Rf6)3]3 (3a)c
RhCl[P(C6H4-p-SiMe2-Rf8)3]3 (3b)d
RhCl[P(C6H4-p-SiMe3)3]3 (3c)b
e
RhCl[P(C6H4-p-Me)3]3
f
RhCl[P(C6H4-m-Rf6)3]3
38
38
b
[RhCl(PPh3)2]2
[RhCl{P(C6H4-p-SiMe2-Rf6)3}2]2 (4a)c
[RhCl{P(C6H4-p-SiMe2-Rf8)3}2]2 (4b)b
[RhCl{P(C6H4-p-SiMe3)3}2]2 (4c)b
e
[RhCl{P(C6H4-p-Me)3}2]2
i
RhCl(O2)[P(C6H4-m-SO3Na)3]3
24
17.5 (dd)
12.0 (dd)
13.7 (dd)
14.2 (dd)
40.3 (dd)
39.9 (dd)
41.1 (dd)
39.9 (dd)
98
99.6
99.6
99.7
114
116
114
115
24
RhCl(O2) [P(C6H4-p-SiMe2-Rf6)3]3 (5a)b
RhCl(O2)[P(C6H4-p-SiMe2-Rf8)3]3 (5b)b
RhCl(O2)[P(C6H4-p-SiMe3)3]3 (5c)b
25.8
25.5
25.5
17.5
g
18.3
g
25.5
25.5
25.5
17.5
19.0
18.2
14.7
j
cis-RhCl(H)2(PPh3)3
cis-RhCl(H)2[P(C6H4-p-SiMe2-Rf6)3]3 (6a)b
cis-RhCl(H)2[P(C6H4-p-SiMe2-Rf8)3]3 (6b)h
cis-RhCl(H)2[P(C6H4-p-SiMe3)3]3 (6c)b
89.6
88.4
91.5
a Rf6 ) CH2CH2(CF2)5CF3; Rf8 ) CH2CH2(CF2)7CF3. b C6D6. c FC-72/C6D6, 1:1 (v/v). d CF3C6H5/C6D6, 1:1 (v/v). e C6H5CH3, see ref 15. f THF-
d8, see ref 5a. g not resolved. h n-C6D14/FC-72 1:1 (v/v). i H2O, see ref 17c. j At -25 °C, CH2Cl2, see ref 15a.
Scheme 2
Scheme 3
An attempt to convert a mixture of 1a, 2a, and 3a completely
to 3a by hydrogenation of the COD ligand under a dihydrogen
atmosphere (1 bar) failed and led to the formation of a cis
dihydride species 6a whereas 2a remained unchanged. The pale
yellow cis dihydride complexes 6 can be reconverted into
square-planar 3 by dihydrogen elimination under reduced
The presence of oxygen causes the formation of the η2-O2-
complexes 5a-c (31P NMR, Table 1), which decompose over
a period of days affording dimeric complexes 4 and fluorous
triarylphosphine oxide (OdPR3 in Scheme 3) which is in line
with earlier observations for RhCl(PPh3)3O2.17e If excess of free
phosphine 1 is added either initially or after initial reaction with
oxygen (2 equiv per Rh), rhodium-catalyzed oxidation of
phosphine takes place17e with compounds 1, 3, and OdPR3 as
the sole species present in the final product mixture (Scheme
3). The 31P NMR spectroscopic data of oxo complexes 5 are in
close agreement with data reported for RhCl(O2)[P(C6H4-m-
SO3Na)3] (Table 1).17c
The results of an X-ray structural analysis of 4c indicate that
its molecular structure (Figure 2) is analogous to the dimeric
structure displayed by [Rh(µ-Cl)(PPh3)2]2.18a The Rh-Cl and
Rh-P bond lengths in 4c (2.408(5) Å (Rh-Cl), 2.394(5) Å
(Rh-Cl(a)), 2.220(5) Å (Rh-P(1)), and 2.211(5) Å (Rh-P(2))
are within experimental error comparable to the corresponding
distances in [RhCl(PPh3)2]2 (2.394(2) Å (Rh-Cl), 2.424(2) Å
(Rh-Cl′), 2.200(2) Å (Rh-P(1)) and 2.213(2) Å (Rh-P(2)),
respectively. However, in contrast to [RhCl(PPh3)2]2, 4c has a
1
pressure or by the addition of an excess of 1-hexene. H and
31P NMR data indicate that complexes 6 under a dihydrogen
atmosphere exhibit octahedral structures with one hydride trans
2
to phosphorus (6c: doublet, δH ) -9.1, JH,P ) 141 Hz) and
the other located trans to the chloride ligand (6c: singlet, δH )
-16.5). Related compounds are well-known for PPh315 and also
for fluorous alkylphosphine P(CH2CH2C6F13)3.10a
The substitution of PPh3 in RhCl(PPh3)3 by its fluorous and
nonfluorous phosphines 1 was demonstrated for 3b and 3c. In
the case of 3b the complex precipitated as a red oil from
benzene. In contrast, treatment of RhCl(PPh3)3 with 1c (3 equiv
and excess) led to the formation of a complicated mixture of
several mixed phosphine complexes. These results indicate that
there is a significant shift of the equilibrium for 1b toward
complete PPh3 substitution, which is most probably driven by
the decreased solubility of the fluorous complex 3b in benzene.
The equilibrium of the square-planar complex RhCl(PPh3)3
with the dimeric species [RhCl(PPh3)2]2 and free ligand in
solution is well-known (Scheme 2).11b,16 In the case of RhCl-
(PPh3)3 itself, a dimer-to-monomer ratio amounting to 0.2 was
found in CF3C6H5/C6D6 as solvent (1:1 (v/v), 25 °C). Complexes
3a and 3b show a similar behavior. However, in the presence
of an excess of 10 mol % of free ligand (relative to 3a,b) the
dimer-to-monomer ratio was found to be <0.01.
(17) (a) Baird, M. C.; Lawson, D. N.; Mague, J. T.; Osborn, J. A.;
Wilkinson, G. Chem. Commun. 1966, 129. (b) Bennett, M. J.; Donaldson,
P. B. Inorg. Chem. 1977, 16, 1581-1585. (c) Larpent, C.; Dabard, R.; Patin,
H. New J. Chem. 1988, 12, 907. (d) Suzuki, H.; Matsuura, S.; Moro-Oka,
Y.; Ikawa, T. J. Organomet. Chem. 1985, 286, 247-258. (e) Atlay, M. T.;
Carlton, L.; Read, G. J. Mol. Catal. 1983, 19, 57.
The oxygen sensitivity of RhCl(PPh3)3 and its derivatives has
long been noted17 and was also observed for complexes 3a-c.

3948 J. Am. Chem. Soc., Vol. 122, No. 16, 2000
Richter et al.
Table 2. Catalytic Hydrogenationa of 1-Octene Using 3a,b as
Pre-catalysts
leaching (%)b
(cumulative) Rh
conv t1/2 TOF50%
complex cycle (%) (min) (h-1
TON
)
P
3ac
3bd
1
1
2
3
4
5
6
7
8
9f
95
92
96
99
92
56
69
58
47
43
298
531
337
688
1048
1383
1747
2100
2444
2808
3117
0.27 8.10
0.11 2.34
0.05 2.06
0.10 3.24
0.11 3.20
0.18 3.52
0.08 2.15
0.05 2.43
0.33e 4.70e
0.08 3.41
177
212
261
277
304
397
507
617
142
99.8 40
96.8 30
94
99.7 20
85 75
24
a Conditions: solvent ) c-CF3C6F11, T ) 80 °C, p(H2) ) 1 bar.
b At T ) 0 °C, estimated error of ICP-AAS analysis (0.013%.
c Containing 10% of 1a, [1-octene]/[Rh] ) 521, [Rh]0 ) 0.0060 mol/
L. d [1-octene]/[Rh] ) 365, [Rh] ) 0.0087 mol/L in the initial cycle.
e High value due to experimental error during phase separation. f Fresh
c-CF3C6F11 was added as compensation for losses of fluorous solvent
to reestablish the starting conditions (see text).
Figure 2. Displacement ellipsoid illustration of 4c at the 30%
probability level. Hydrogen atoms and disordered solvent were omitted.
Selected bonding distances (Å) and angles (deg): Rh‚‚‚Rh(A) )
3.450(5), Rh-Cl ) 2.408(5), Rh-Cl(a) ) 2.394(5), Rh-P(1) ) 2.220-
(5), Rh-P(2) ) 2.211(5), P(1)-C(11) ) 1.820(14), P(1)-C(21) )
1.837(12), P(1)-C(31) ) 1.848(11), P(2)-C(41) ) 1.851(11), P(2)-
C(51) ) 1.817(12), P(2)-C(61) ) 1.823(14), Cl-Rh-P(1) ) 87.97(14),
Cl-Rh-P(2) ) 172.10(14), Cl-Rh-Cl(a) ) 80.78(14), P(1)-Rh-
P(2) ) 97.83(15), P(1)-Rh-Cl(a) ) 168.63(15), Rh-Cl-Rh(a) )
91.86(15), C(11)-P(1)-C(21) ) 104.7(6), C(11)-P(1)-C(31) )
100.3(6), C(21)-P(1)-C(31) ) 103.0(5), C(41)-P(2)-C(51) )
105.4(5), C(41)-P(2)-C(61) ) 102.4(5), C(51)-P(2)-C(61) )
101.5(6).
folded Rh(µ-Cl)2Rh moiety (fold angle (λ) ) 141.3(2)°), which
is the largest value for nonplanar structures of this type (λ is
usually <134°).18 As a consequence the intramolecular Rh-
Rh distance of 3.449(5) Å is around 0.2 to 0.4 Å longer
compared to other folded structures.18 Recent discussions
propose steric (cone angle and interligand repulsion) rather than
electronic features of the phosphine as the origin of a more or
less folded moiety.18c Extended Hu¨ckel calculations using
[RhCl(PH3)2]2 and [RhCl(dphm)]2 (dphm ) H2PCH2PH2) as
model compounds resulted in planar structures being the most
stable ones but also revealed an energy cost of only 10 kJ mol-1
to fold these structures to λ ) 140°. Hence, packing effects
and/or subtle intramolecular steric interactions will most prob-
ably be sufficient to determine the actual folding angle in the
solid state.
Catalytic Hydrogenation. Rhodium complexes 3a,b obtained
from [(COD)RhCl]2 and fluorous phosphines 1a,b were found
to be active catalysts for hydrogenation of 1-alkenes under
single-phase fluorous conditions at 80 °C (Table 2). The
hydrogenation products were readily isolated from the catalyst
layer by cooling of the reaction mixture to 0 °C followed by
phase separation of the resulting biphasic system. In this way
hydrogenation of 1-octene afforded n-octane in >95% yield
(GC, GC-MS). Less than 4.3% internal olefins resulting from
isomerization were present. In the case of 3b we were able to
recycle the fluorous catalyst layer multiple times while still
maintaining high turnover numbers and high conversions per
cycle (Table 2 and Figure 3).
Figure 3. Plot of the turnover number (TON) and turnover frequency
(TOF) for nine cycles. Conditions: catalyst ) RhCl[P(C6H4-p-SiMe2-
(CH2)2(CF2)7CF3)3]3; solvent ) CF3C6F11; T ) 80 °C; p(H2) ) 1 bar;
[1-octene]/[Rh] ) 365; [Rh]0 ) 0.0087 mol/L. Fresh c-CF3C6F11 was
added in the 9th cycle as compensation for losses of fluorous solvent
to reestablish the starting conditions (see text).
It was found for 3b that the activity increased with the number
of cycles (up to TOF ∼ 600 h-1 in the 8th cycle). This is most
probably caused by the combined effects of a non-zero-order
rate dependence in rhodium and the (observed) loss of fluorous
solvent due to evaporation during phase separation under H2
-flow and nonzero miscibility of c-C6F11CF3 in the product layer
even at 0 °C (ca. 5 vol %). Consequently, an average loss of
fluorous solvent of ca. 12% per cycle (ca. 0.25 mL) took place.
However, restoring the amount of c-C6F11CF3 in the 9th cycle
to its initial volume showed that the catalyst activity had dropped
only by 19% (entry 9, Table 2).
As anticipated from the limited drop in activity upon recycling
of the catalyst layer resulting from 3b, rhodium leaching into
the organic phase (as determined by ICP-AAS) was indeed low.
On average 0.12% of Rh per cycle was lost corresponding to a
rhodium concentration of 3 ppm (by weight) in the product
phase. Over 9 cycles only 1% of rhodium was lost despite the
rather simple phase separation method used. Not surprisingly,
leaching was higher for the pre-catalyst with a lower fluorine
content, e.g., for 3a, 0.3% (6 ppm) of rhodium was present in
the organic phase after phase separation.
(18) (a) Curtis, D. M.; Butler, W. M.; Greene, J. Inorg. Chem. 1978, 17,
2928-2931. (b) Kubas, G. J.; Ryan, R. R. Cryst. Struct. Commun. 1977, 6,
295-300. (c) Wang, K.; Goldman, M. E.; Emge, T. J.; Goldman, A. S. J.
Organomet. Chem. 1996, 518, 55-68. (d) Hofmann, P.; Meier, C.; Hiller,
W.; Heckel, M.; Riede, J.; Schmidt, M. U. J. Organomet. Chem. 1995,
490, 51-70.
The amount of leached rhodium enabled us to calculate the
average partition coefficients (P) of the rhodium species present

Fluorous Versions of Wilkinson’s Catalyst
J. Am. Chem. Soc., Vol. 122, No. 16, 2000 3949
Table 3. Average Partition Coefficients (P)a of Rhodium Species
Present in the Fluorous Biphasic Product Mixture Resulting from
Hydrogenation of 1-Octene
mol‚L-1‚h-1) were measured for 3c at relatively high rhodium
and olefin concentrations (entry 2, Table 4). Under these
conditions a zero-order dependence in [1-octene] was found up
to ca. 80% conversion resulting in a linear conversion versus
time plot. In all other experiments ([Rh] ) 4.0-8.0 mM,
[1-octene]0 ) 1.46 M) rates of dihydrogen uptake were lower
(<11 mol‚L-1‚h-1) and conversion versus time plots cor-
responded to a rate law -d[1-octene]/dt ) kobs[1-octene] for
>98% conversion, which allows the catalyst activity to be
P
F-content of
pre-catalyst
phosphine (wt %)
T ) 25 °C
T ) 0 °C
3a
3b
48.7
53.2
76
n.d.
293
887
a Calculated from the amount of Rh found in the organic product
phase of the first cycle by ICP-AAS analysis.
evaluated in terms of the observed first-order rate constant kobs
.
Because the rates of dihydrogen uptake for these experiments
were well below the maximum value obtained in entry 2,
diffusion limitation of dihydrogen can be excluded in these
cases. In addition to kobs, turnover frequencies (TOF in
mol‚mol-1‚h-1) were derived from the tangent of a 4th order
polynomal fit to conversion versus time plots at 25% conversion
and served as a measure of catalyst activity independent of any
assumed kinetic relationship with substrate.
during phase separation (Table 3). The values found clearly
demonstrate a significant positive effect of longer fluorochains
and lower temperature. The partitioning of rhodium found for
3b at 0 °C is higher than that reported for RhCl[P(CH2)2(CF2)n-
CF3)3]3 (696 and 811 for n ) 5 and 7, respectively, in c-C6F11-
CF3/toluene at 27 °C).6b Although a direct comparison is not
possible because of the different solvent system used, this result
is still surprisingly good for arylphosphine rhodium complexes
taking into account the significant lower partition coefficients
of the fluorous arylphosphines (1a, P ) 0.26; 1b, P ) 2.2, in
c-C6F11CF3/toluene at 0 °C)12 as compared to the fluorous
alkylphosphines P[(CH2)2(CF2)nCF3)3]3 (n ) 5, P ) 82 and n
) 7, P ) 332 in c-C6F11CF3/toluene at 27 °C).6a This result
emphasizes that the fluorophase preference of the rhodium
complexes cannot solely be deduced from the partition coef-
ficient of the phosphine ligand itself, although an increase of
fluorphase affinity of the phosphine does correspond to a higher
fluorphase affinity of the derived Rh complex. This latter
observation also becomes apparent if the improvement in
fluorphase partitioning of 1b relative to 1a is compared with
the higher fluorphase partition coefficient of 3b relative to 3a.
Clearly the relative orientation of the phosphine ligands in the
derived rhodium complexes is important. A fluorophilic van
der Waals surface, made up of a relatively high proportion of
CF2 and CF3 units that shield the fluorophobic metal environ-
ment of the catalyst, may well be a prerequisite for high
fluorphase affinity and hence good recycling efficiency in
fluorous biphasic separation.
It should be noted that efficient retention of uncoordinated
fluorous phosphine in the fluorphase is as important for recycling
the intact catalyst system as is retention of rhodium. In fact,
significant amounts of fluorous phosphine were present in the
alkane product phases (3a (3b): 130 (64) ppm, corresponding
to 8.1 (3.0)% of total phosphorus present in the precatalyst)
indicating that leaching of fluorous ligand is more significant
than that of rhodium itself and to a large extent responsible for
the low recycling efficiency of the 3a-derived catalyst solution.
A monomer-dimer equilibrium involving dissociation of phos-
phine as in Scheme 2 or more likely one involving dissociation
of phosphine from dihydrides 6 which are the predominant
species present under dihydrogen atmosphere16 may well be
responsible for this behavior (eq 1).
From Table 4 it is clear that 3c displays by far the highest
activity being >1.5 times more active than Wilkinson’s catalyst.
This result clearly shows the beneficial influence of the p-silyl
substitution on catalytic activity. However, the most important
observation is the fact that 3a (entry 3) is comparable in activity
to RhCl(PPh3)3 (entry 4) despite its fluoro-tail functionalization
and the presence of small quantities of free phosphine (<10%),
which is known to inhibit catalytic activity.11b,c The somewhat
lower activity of 3b (entry 5) suggests a small negative
electronic influence on catalytic activity caused by the longer
fluorotail. However, solvation effects cannot be excluded either.
The overall activities of 3a and 3b in perfluoromethylcyclo-
hexane and solvent R,R,R-trifluorotoluene also compare favor-
ably with turnover frequencies for 1-alkene hydrogenation by
the Wilkinson catalyst in conventional solvents (TOF ) 150-
600 h-1 for hydrogenation of 1-heptene in benzene).16b
The overall first-order kinetics observed are noteworthy. More
complicated relationships between substrate concentration and
rate of hydrogenation are found in the literature (e.g. 1/rate )
C(1/[olefin]) + C′ with C and C′ being constants),16b which is
due to parasitic formation of inactive olefin complexes RhClP2-
(olefin) (P ) phosphine), especially at higher olefin concentra-
tions. Our data appear to be in line with the conventional
hydrogenation mechanism proposed for Wilkinson’s catalyst,
involving rate-limiting elimination of alkane from a dihydrido
olefin species which itself is formed in a rapid preequilibrium
with dihydride species and olefin15b but without kinetically
important formation of RhClP2(olefin) complexes. Although we
did not study this further, the insignificance of such species
under our conditions might well be the result of the different
solvent properties of the partially fluorinated solvent R,R,R-
trifluorotoluene.
Conclusions
The catalytic results presented above indicate that as far as
hydrogenation of 1-alkenes is concerned we have indeed
designed and prepared a fluorous alternative to Wilkinson’s
catalyst with the added advantage of facile catalyst separation
and recycling. Recycling efficiencies are >98% (as defined by
phosphine leaching and activity) for the catalysts with the longer
fluortails (3b). This recycling efficiency is much better than
expected on the basis of the fluorous phase affinity of the free
phosphine ligands alone and is in fact the result of the much
higher fluorous phase affinity of the phosphine-containing
rhodium species that are present during and after catalysis. It is
proposed that this unprecedented effect is a reflection of the
(1)
To allow comparison of the activities of fluorous derivatives
3a and 3b with those of nonfluorous 3c and conventional
Wilkinson catalyst RhCl(PPh3)3, 1-octene hydrogenation experi-
ments under single-phase conditions in the hybrid solvent R,R,R-
trifluorotoluene at atmospheric dihydrogen pressures were
carried out. The results obtained with the different pre-catalysts
are listed in Table 4. Highest rates of dihydrogen uptake (14.2

3950 J. Am. Chem. Soc., Vol. 122, No. 16, 2000
Richter et al.
Table 4. Comparison of the Different Pre-catalysts RhCl[P(C6H4-p-R)3]3 in the Hydrogenation of 1-Octenea
c
d
[Rh]
(mM)
[1-octene]
(M)
conversion
(%)
kobs
(h-1
TOF(25%)
entry
R
TONb
)
(h-1
)
1
2
3
4
5
Me3Si (3c)
4.0
8.0
4.0
4.0
4.0
1.46
2.92
1.46
1.46
1.46
92
99
99
98
99
336
361
363
358
355
7.5(2)
e
4.2(1)
4.0(1)
3.2(1)
1610
1910
1110
960
Me3Si (3c)
Rf6(CH2)2SiMe2 (3a)f
H
Rf8(CH2)2SiMe2 (3b)g
870
a Conditions: T ) 80 °C, p(H2) ) 1 bar, solvent ) R,R,R-trifluorotoluene, stirring speed ) 900 rpm. b TON ) turnover number (mole of
olefin/mole of Rh). c Obtained by fitting the data to Xt ) [1 - exp(-kobst)] (Xt ) conversion). d TOF ) turnover frequency (mole of olefin/mole of
Rh/hour) at 25% conversion. e Zero order in olefin up to 80% conversion, kobs ) 14.2(1) mol‚L-1‚h-1
.
f Rf6: -(CF2)5CF3. g Rf8: -(CF2)7CF3.
1
106.3 (δ-C, COD), 71.0 (d, JC,Rh ) 13.1 Hz, R-C, COD), 33.7 (â-C,
particular spatial orientation of multiple fluorous ligands as-
sembled on one metal center resulting in a highly fluorous van
der Waals surface. Catalytic activities as measured in C6H5CF3
are remarkably close to Wilkinson’s catalyst stressing the
successful application of the -SiCH2CH2- spacer in fluoro-
tail functionalization of homogeneous catalysts. Since is was
found that catalyst recovery critically depends on the amount
of fluorous character of the ligand, we now seek to prepare
ligands that give rise to increased fluorous van der Waals
surfaces and to employ them in catalysis.13
2
COD), 29.5 (γ-C, COD), 26.5 (t, JC,F ) 23.0 Hz, CH2CF2), 5.51 (s,
1JC,Si ) 50.2 Hz, CH2Si), -3.69 (s, JC,Si ) 54.8 Hz, Me). 3a:
1
C
144H126F117ClSi9P3Rh: calcd C 37.9, H 2.78, F 48.71, Si 5.54, P 2.04;
found C 37.76, H 2.85, F 48.48, Si 5.61, P 2.11. 1H NMR (300 MHz,
FC-72/C6D6, 1:1 (v/v)): δ 7.61 (m, 18 H, o-H, Si-C6H4), 6.97 (m, 18
H, m-H, Si-C6H4), 1.87 (m, 18 H, CH2CF2), 0.80 (m, 18 H, CH2Si),
0.04 (m, 54 H, Me).
Preparation of (COD)RhCl[P{C6H4-p-SiMe2(CH2)2(CF2)7CF3)}3]
(2b) and RhCl[P{C6H4-p-SiMe2(CH2)2(CF2)7CF3}3]3 (3b): (a) Method
1. 1b (2.801 g, 1.578 mmol) dissolved in 60 mL of benzene was treated
with 0.130 g (0.263 mmol) of [(COD)RhCl]2 at 25 °C. After being
stirred for 15 h, when a waxy, yellow precipitate was observed, the
reaction mixture was treated with c-C6F11CF3 (10 mL). The color of
the fluorous phase turned to dark red, while the organic upper phase
remained yellow. After phase separation the organic phase was
evaporated to dryness affording a red viscous oil, which consisted of
1b and 2b (1:1.6 molar ratio). The fluorous phase was washed two
times with 10 mL of benzene and all volatiles were removed in vacuo
(0.1 bar). Further drying of the remaining dark red oil in high vacuum
(10-6 bar) for 12 h yielded 2.497 g (0.457 mmol, 86.9% based on Rh)
of a highly viscous dark red oil being pure 3b.
(b) Method 2. 1b (0.914 g, 0.515 mmol) and 0.158 g (0.171 mmol)
of RhCl(PPh3)3 were dissolved in 20 mL of benzene at 25 °C. Instantly
a dark red oil precipitated. c-CF3C6F11 (2.5 mL) was added when the
formation of a dark red bottom layer and an orange upper layer was
observed. The upper layer was decanted and all volatiles of the lower
layer were removed in vacuo. The remaining red oil was further dried
in high vacuum (10-6 bar, 12 h). A highly viscous dark red oil (0.690
g) containing 0.119 mmol of 3b (69.6% based on Rh) and 0.15 equiv
Experimental Section
General Comments and Materials. All reactions and workups were
conducted under nitrogen atmosphere unless noted otherwise. Solvents
were employed as follows: c-C6F11CF3, CF3C6H5 (Acros, Alfa)
degassed and stored over molecular sieves; C6F6 (Acros), C6D6, CDCl3,
n-C6D14 (Cambridge Isotope Laboratories, Aldrich) degassed and stored
over molecular sieves; and 1-octene degassed and distilled over sodium
sand. Hydrogen gas (Hoekloos, 5.0) was used as received. Elemental
analyses and ICP-AAS measurements were carried out by H. Kolbe,
Microanalytisches Laboratorium, Mu¨hlheim an der Ruhr. NMR spectra
were obtained from Varian INOVA 300 and MERCURY 200 spec-
1
trometers. H, 13C, 29Si NMR spectra are referenced relative to TMS,
31P NMR relative to 85% H3PO4, and 19F NMR relative to CFCl3
(external standard). The 19F-decoupler frequency in 13C{19F}NMR
experiments was set to δ -121 to decouple from either the 19F-nuclei
or the CF2-moieties. GC analyses were performed on UNICAM PU610
apparatus with a 30 m J&W Scientific AT-SILAR capillary column
and a flame ionization detector. Product yields were determined by
peak area analysis using n-decane as the internal standard. [(COD)-
RhCl]2 and RhCl(PPh3)3 were prepared as described in the literature.19
Preparation of (COD)RhClP[C6H4-p-SiMe2(CH2)2(CF2)5CF3]3
(2a) and RhCl[P{C6H4-p-SiMe2(CH2)2(CF2)5CF3}3]3 (3a). 1a (3.210
g, 2.176 mmol) dissolved in 20 mL of n-hexane and 5 mL of toluene
was treated with 0.179 g (0.363 mmol) of [(COD)RhCl]2 at 25 °C.
After the solution was stirred for 15 h the volume was reduced by
two-thirds and toluene was added (20 mL). Stirring was continued for
15 min affording a biphasic system. The upper phase was carefully
decanted and evaporated to dryness affording a red viscous oil, which
consisted of 1a, 2a, and 3a (3.5:1:1.5 molar ratio). The fluorous phase
consisting of a dark red oil was washed two times with toluene (10
mL) and predried in vacuo (0.1 bar). Further drying in high vacuum
(10-6 bar) for 12 h afforded 1.50 g of a highly viscous dark red oil
containing 3a (0.318 mmol, 43.8% based on Rh) and 0.1 equiv of 1a.
2a: 1H NMR (200 MHz, C6D6): δ 7.92 (m, 6 H, o-H, Si-C6H4), 7.22
(m, 6 H, m-H, Si-C6H4), 5.94 (s, 2 H, COD), 3.28 (s, 2 H, COD), 2.18
(m, 2 H, COD), 1.67 (dm, 4 H, COD), 1.85 (m, 6 H, CH2CF2), 0.78
(m, 6 H, CH2Si), -0.02 (s, 36 H, Me). 13C{1H} -NMR (75.5 MHz,
C6D6): δ 140.5 (s, ipso-C, Si-C6H4), 135.2 (d, JP,C ) 10 Hz, C6H4),
1
of 1b was obtained. 2b H NMR (300 MHz, C6D6): δ 7.95 (m, 6 H,
o-H, Si-C6H4), 7.24 (m, 6 H, m-H, Si-C6H4), 5.96 (s, 2 H, COD), 3.29
(s, 2 H, COD), 2.20 (m, 4 H, COD), 1.68 (dm, 4 H, COD), 1.88 (m,
6 H, CH2CF2), 0.81 (m, 6 H, CH2Si), 0.01 (s, 18 H, Me). 3b:
C
162H126F153ClSi9P3Rh: calcd C 35.61, H 2.32, F 53.20, Si 4.63, P 1.70;
found C 35.74, H 2.37, F 53.03, Si 4.66, P 1.65. 1H NMR (300 MHz,
CF3C6F11/C6D14, 1:1 (v/v)): δ 7.43 (m, 18 H, o-H, Si-C6H4), 7.00 (m,
18 H, m-H, Si-C6H4), 1.89 (m, 18 H, CH2CF2), 0.85 (m, 18 H, CH2Si),
0.15 (m, 54 H, Me). 13C{1H} NMR (75.5 MHz, CF3C6F11/C6D14, 1:1
(v/v)): δ 138.7, 138.4, 137.0, 135.3, 134.7, 132.4, 128.7, 126-102
2
(m, CF3, CF2), 26.4 (t, JC,F ) 23.8 Hz, CH2CF2), 5.51 (s, CH2Si),
-4.35 (s, JC,Si ) 52.6 Hz, Me).
1
Preparation of (COD)RhCl[P(C6H4-p-SiMe3)3] (2c) and RhCl-
[P(C6H4-p-SiMe3)3]3 (3c). 1c (1.436 g, 2.999 mmol) and 0.177 g (0.359
mmol) of [(COD)RhCl]2 were dissolved in 20 mL of n-hexane and 5
mL of toluene. After the mixture was stirred for 15 h all volatiles were
removed in vacuo. The orange residue was dissolved in 10 mL of
n-hexane, while the mixture was warmed up to ∼40 °C. The solution
was stored at -10 °C for 12 h. The precipitate mainly consisting of 2c
was filtered off and the solution was stored for 3 days at -10 °C. Again,
the precipitate was filtered off and all volatiles of the solution were
removed in vacuo. 3c (0.657 g, 0.417 mmol, 58.1% based on Rh) was
1
2
133.7 (d, JP,C ) 10 Hz, C6H4), 119.3 (tt, JC,F ) 253 Hz, JC,F ) 32.9
1
2
Hz, R-CF2), 118.1 (qt, JC,F ) 290 Hz, JC,F ) 32.9 Hz, CF3), 112.2
1
obtained as an orange solid. 2c. H NMR (300 MHz, C6D6): δ 7.99
1
2
1
(tquin, JC,F ) 268 Hz, JC,F ) 31.4 Hz, â-CF2), 112.1 (tquin, JC,F
)
)
(m, 6 H, o-H, Si-C6H4), 7.35 (m, 6 H, m-H, Si-C6H4), 5.96 (s, 2 H,
COD), 3.38 (s, 2 H, COD), 2.19 (m, 4 H, COD), 1.66 (dm, 4 H, COD),
0.13 (s, 27 H, Me). 3c: C81H117ClSi9P3Rh: calcd C 61.78, H 7.49, Si
16.05, P 5.90; found C 61.65, H 7.56, Si 15.83, P 5.94. 1H NMR (200
MHz, C6D6): δ 7.9 (m, 18 H, o-H, Si-C6H4), 7.2 (m, 18 H, m-H, Si-
271 Hz, 2JC,F ) 31.4 Hz, γ-CF2), 111.3 (tquin, 1JC,F ) 273 Hz, 2JC,F
1
2
31.7 Hz, δ-CF2), 109.4 (tqt, JC,F ) 260 Hz, JC,F ) 30.5 Hz, ꢀ-CF2),
(19) Osborn, J. A.; Wilkinson, G. Inorg. Synth. 1990, 28, 77-79 and
88-90.

Fluorous Versions of Wilkinson’s Catalyst
J. Am. Chem. Soc., Vol. 122, No. 16, 2000 3951
C6H4), 0.18 (m, 81 H, Me). 13C{1H} NMR (75.5 MHz, C6D6): δ 141.3,
minutes a color change from red to pale yellow occurred. Analysis by
1H and 31P NMR spectroscopy indicated the complete conversion of
complexes 3 to dihydrides 6. 6c: 1H NMR (200 MHz, C6D6): δ 7.8-
7.2 (m, 36 H, C6H4), 0.18 (m, 81 H, Me), -9.1 (br, d, 1 H, 2JH,P ) 141
Hz, Rh-H), -16.5 (br, s, 1 H, Rh-H).
Catalytic Hydrogenation Reactions. The catalytic experiments were
carried out in a 30 mL Schlenk flask, under dihydrogen atmosphere (1
bar). The catalyst was dissolved in either a hydrogen-saturated toluene
or perfluoro solvent and the mixture was stirred with a magnetic stirring
bar (900 rpm). The olefinic substrate was added after initial dihydrogen
uptake had ceased. The dihydrogen uptake was monitored using two
mineral oil filled gas burets. During the hydrogenation reactions one
buret was opened to the reaction vessel while the other was recharged
with dihydrogen.
141.0, 137.1 (m), 135.5 (m), 134.7 (d), 133.8 (d), 132.7 (m), 132.1
1
(d), (C6H4), -0.75 (s, JC,Si ) 52.5 Hz, Me).
Preparation of [Rh(µ-Cl){P(C6H4-p-SiMe3)3}2]2 (4c). Single crys-
tals of 4c were obtained from a solution of 3c in a biphasic system
consisting of n-octane and c-C6F11CF3 (1:1, v/v, Tc ) 42 °C). The
molecular structure of 4c was determined by a single-crystal X-ray
structure determination study.20
Formation of RhCl(O2)[P{C6H4-p-SiMe2R}3]3 (R ) (CH2)2-
(CF2)5CF3 (5a), (CH2)2(CF2)7CF3 (5b), Me (5c)). NMR tube samples
of complexes 3 dissolved in C6D6/FC-72 (1:1 (v/v) (3a,b)) and C6D6
(3c), respectively, were exposed to the air resulting in formation of
complexes 5. Complexes 3 were reformed in 3-4 days upon addition
of an excess of 1.
Formation of RhCl(H)2[P{C6H4-p-SiMe2R}3]3 (R ) (CH2)2-
(CF2)5CF3 (6a), (CH2)2(CF2)7CF3 (6b), Me (6c)). NMR tube samples
of complexes 3 in C6D6/FC-72 (1:1 (v/v) (3a,b)) and C6D6 (3c),
respectively, were frozen (-78 °C) and the dinitrogen atmosphere was
replaced by dihydrogen gas. Upon warming to room temperature within
(a) Recycling Experiments (3a and 3b). Only catalysts prepared
through method 1 were used. Catalytic reactions were carried out under
single-phase fluorous conditions in c-CF3C6F11 (2 mL) at 80 °C. After
>90% conversion (monitored by the H2 uptake) the homogeneous
reaction mixture was cooled to 0 °C and the upper organic layer was
siphoned off. Between cycles the fluorous layer was kept under H2
atmosphere. By warming the solution up to 80 °C and adding a fresh
portion of 1-octene (12.74 mmol) a new cycle was started. The organic
phases resulting from the first cycle were analyzed by GC and ICP-
AAS.
(20) X-ray data for 4c (molecular formula: C108H156Cl2P4Rh2Si12
+
solvent) were collected on an Enraf Nonius CAD-4T/Rotating Anode
diffractometer (Mo KR, λ ) 0.71073 Å, T ) 200 K) for an orange needle
shaped crystal mounted with the inert oil technique at the end of an open
Lindemann glass capillary. Crystal quality was limited as indicated by
relatively broad, structured reflection profiles and very limited diffracted
intensity beyond θ ) 20°. Crystals are monoclinic, space group C2/c with
cell dimensions a ) 39.60(2) Å, b ) 13.39(2) Å, c ) 31.00(2) Å, â )
127.72(8)°, V ) 13002(26) Å3, Z ) 4. A total of 11060 reflections up to
θ ) 20° were scanned and averaged into a unique set of 6046 reflections
(of which 3806 had I > 2σ(I)). The structure was solved by standard
Patterson techniques using the program DIRDIF-9621 and refined on F2
with SHELXL97.22 The crystal structure was found to contain four sites
filled with disordered solvent of unreliable composition. The contribution
of the scattering power of the disordered solvent in that area to the structure
factors was taken into account using the BYPASS procedure23 as imple-
mented as the SQUEEZE option in the program PLATON.24 A total density
equivalent with the scattering of 50 electrons per solvent accessible void
of 317 Å3 was found. Hydrogen atoms were taken into account at calculated
positions and refined riding on their carrier atoms. Convergence was reached
at R ) 0.0745, wR2 ) 0.187, S ) 1.01, 1/w ) σ2(Fobs2) + (0.0808P)2. No
residual electron density was found in a final difference map outside the
range -0.84 < ∆F < 0.68 electron/Å3. The trimethylsilyl moieties show
relatively high librational motion as illustrated in the ORTEP Figure 2.
Geometrical calculations and the ORTEP illustration were done with
PLATON24.
(b) Experiments in CF3C6H5. Catalytic reactions were carried out
under single-phase conditions in CF3C6H5 at 80 °C under atmospheric
hydrogen pressure. After >90% conversion (monitored by the H2
uptake) volatiles were distilled off and analyzed by GC. Activities
(TOF) were determined for 25% conversion.
Acknowledgment. This work was supported by Elf Atochem
Vlissingen B.V., U-Cat (Utrecht) and The Netherlands Founda-
tion for Scientific Research (NWO/CW, to A.L.S.). The authors,
but especially B.R., would like to dedicate this work to Prof.
K.-H. Thiele, on the occasion of his 70th birthday on the 9th of
March 2000.
Supporting Information Available: Crystal data and details
of the structure determination, final coordinates and equivalent
isotropic displacement parameters of the non-hydrogen atoms,
hydrogen atom positions and isotropic displacement parameters,
(an)isotropic displacement parameters, bonds distances, and
bond angles for 4c (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
(21) Beurskens, P. T.; Admiraal, G.; Beurskens, G.; Bosman, W. P.;
Garcia-Granda, S.; Gould, R. O.; Smits, J. M. M.; Smykalla, C. The
DIRDIF96 program system; Technical Report of the Crystallography
Laboratory; University of Nijmegen, The Netherlands, 1996.
(22) Sheldrick, G. M. SHELXL97; Program for crystal structure refine-
ment; University of Go¨ttingen: Go¨ttingen, Germany, 1997.
(23) Van der Sluis, P.; Spek, A. L. Acta Crystallogr. 1990, A46, 194.
(24) Spek, A. L. Acta Crystallogr. 1990, A46, C-34.
JA993562V