Molecular engineering in homogeneous catalysis: One-phase catalysis coupled with biphase catalyst separation. The fluorous-soluble HRh(CO){P[CH2CH2(CF2)5CF3]3}3 hydroformylation system

Article abstract of DOI:10.1021/ja9738337

The hydroformylation of decene-1 was studied in the presence of the fluorous-soluble P[CH₂-CH₂(CF₂)₅CF₃]₃ modified rhodium catalyst at 100 °C and 1.1 MPa of CO/H₂ (1:1) in a 50/50 vol % toluene/C₆F₁₁CF₃ solvent mixture, which forms a homogeneous liquid phase at and above 100 °C. P[CH₂CH₂(CF₂)₅CF₃]₃ was selected on the basis of a semiempirical calculation of the electronic properties of P[(CH₂)(x)(CF₂)(y)CF₃]₃ (x = 0, y = 2, 4 and x = 0-5, y = 2) and prepared by the reaction of PH₃ with CH₂=CH(CF₂)₅CF₃. The solution structure of HRh(CO){P[CH₂CH₂(CF₂)₅CF₃]₃}₃ (1) in C₆F₁₁CF₃ is similar to that of HRh(CO)(PPh₃)₃ (2) in toluene and HRh(CO){P(m-C₆H₄SO₃Na)₃}₃ (3) in water. High-pressure NMR of 1 under 2.1-8.3 MPa of CO/H₂ (1:1) revealed that 1 is in equilibrium with HRh(CO)₂{P[CH₂CH₂(CF₂)₅CF₃]₃}₂ (4). Kinetic studies show that the reaction is first order in both rhodium and decene-1. While the reaction is inhibited by P[CH₂-CH₂(CF₂)₅CF₃]₃, the normal/iso (n/i) ratio of the aldehyde increases with increasing phosphine concentration. The catalytic activity of the Rh/P[CH₂CH₂(CF₂)₅CF₃]₃ catalyst is similar to that of the nonfluorous analogue Rh/P[(CH₂)₇CH₃]₃ catalyst and is an order of magnitude lower than that of the Rh/PPh₃ catalyst. Surprisingly, the n/i product selectivity of Rh/P[CH₂CH₂(CF₂)₅CF₃]₃ is closer to the selectivity of the Rh/PPh₃ catalyst than that of the Rh/P[(CH₂)₇CH₃]₃ catalyst. The fluorous biphase catalyst recovery concept was tested in a semicontinuous hydroformylation of decene-1 with the Rh/P[CH₂CH₂(CF₂)₅CF₃]₃ catalyst. During 9 consecutive reaction/separation cycles, a total turnover of more than 35 000 was achieved with a loss of 1.18 ppm of Rh/mol of undecanals. The fluorous-soluble Rh/P[CH₂CH₂(CF₂)₅CF₃]₃ catalyst was also tested for the continuous hydroformylation of ethylene using the high-boiling fluorous solvent FC-70, which allows continuous removal of propanal at the reaction temperature of 110 °C. The long-term stability of the Rh/P[CH₂CH₂(CF₂)₅CF₃]₃ catalyst is better than that of the Rh/PPh₃ catalyst. Thus, the Rh/P[CH₂CH₂(CF₂)₅CF₃]₃ catalyst is the first catalyst system which can be used for the hydroformylation of both low and high molecular weight olefins and provides facile catalyst separation for both low and high molecular weight aldehydes.

Full text of DOI:10.1021/ja9738337

J. Am. Chem. Soc. 1998, 120, 3133-3143
3133
Molecular Engineering in Homogeneous Catalysis: One-Phase
Catalysis Coupled with Biphase Catalyst Separation. The
Fluorous-Soluble HRh(CO){P[CH2CH2(CF2)5CF3]3}3
Hydroformylation System
Istva´n T. Horva´th,*,† Ga´bor Kiss,*,† Raymond A. Cook,† Jeffrey E. Bond,†
Paul A. Stevens,*,† Jo´zsef Ra´bai,†,‡ and Edmund J. Mozeleski§
Contribution from the Corporate Research Laboratories, Exxon Research and Engineering Company,
Annandale, New Jersey 08801, and Exxon Chemical Company, Annandale, New Jersey 08801
ReceiVed NoVember 7, 1997
Abstract: The hydroformylation of decene-1 was studied in the presence of the fluorous-soluble P[CH2-
CH2(CF2)5CF3]3 modified rhodium catalyst at 100 °C and 1.1 MPa of CO/H2 (1:1) in a 50/50 vol % toluene/
C6F11CF3 solvent mixture, which forms a homogeneous liquid phase at and above 100 °C. P[CH2CH2(CF2)5-
CF3]3 was selected on the basis of a semiempirical calculation of the electronic properties of P[(CH2)x(CF2)yCF3]3
(x ) 0, y ) 2, 4 and x ) 0-5, y ) 2) and prepared by the reaction of PH3 with CH2dCH(CF2)5CF3. The
solution structure of HRh(CO){P[CH2CH2(CF2)5CF3]3}3 (1) in C6F11CF3 is similar to that of HRh(CO)(PPh3)3
(2) in toluene and HRh(CO){P(m-C6H4SO3Na)3}3 (3) in water. High-pressure NMR of 1 under 2.1-8.3 MPa
of CO/H2 (1:1) revealed that 1 is in equilibrium with HRh(CO)2{P[CH2CH2(CF2)5CF3]3}2 (4). Kinetic studies
show that the reaction is first order in both rhodium and decene-1. While the reaction is inhibited by P[CH2-
CH2(CF2)5CF3]3, the normal/iso (n/i) ratio of the aldehyde increases with increasing phosphine concentration.
The catalytic activity of the Rh/P[CH2CH2(CF2)5CF3]3 catalyst is similar to that of the nonfluorous analogue
Rh/P[(CH2)7CH3]3 catalyst and is an order of magnitude lower than that of the Rh/PPh3 catalyst. Surprisingly,
the n/i product selectivity of Rh/P[CH2CH2(CF2)5CF3]3 is closer to the selectivity of the Rh/PPh3 catalyst than
that of the Rh/P[(CH2)7CH3]3 catalyst. The fluorous biphase catalyst recovery concept was tested in a
semicontinuous hydroformylation of decene-1 with the Rh/P[CH2CH2(CF2)5CF3]3 catalyst. During 9 consecutive
reaction/separation cycles, a total turnover of more than 35 000 was achieved with a loss of 1.18 ppm of
Rh/mol of undecanals. The fluorous-soluble Rh/P[CH2CH2(CF2)5CF3]3 catalyst was also tested for the
continuous hydroformylation of ethylene using the high-boiling fluorous solvent FC-70, which allows continuous
removal of propanal at the reaction temperature of 110 °C. The long-term stability of the Rh/P[CH2CH2(CF2)5-
CF3]3 catalyst is better than that of the Rh/PPh3 catalyst. Thus, the Rh/P[CH2CH2(CF2)5CF3]3 catalyst is the
first catalyst system which can be used for the hydroformylation of both low and high molecular weight olefins
and provides facile catalyst separation for both low and high molecular weight aldehydes.
Introduction
catalyst can be controlled by the careful selection of metals and
ligands.1 Some of the recent results include fine-tuning catalytic
activity2 and/or selectivity,3 novel catalyst designs for facile
product separation,4 and control of runaway reactions.5 Despite
recent advances,1c the lack of efficient catalyst recycling
represents a major obstacle for the industrial applications of
Biological systems frequently combine molecular and engi-
neering principles to achieve total control of a single chemical
reaction or a multistep chemical process. On the contrary, many
industrial processes have been developed by performing exten-
sive chemical engineering on chemistries which have intrinsic
limitations on maximum performance. Furthermore, changing
environmental concerns and regulations are placing additional
requirements on process performance, which cannot always be
achieved by reengineering existing processes. Since many of
these limitations have a molecular origin, a new interdisciplinary
field, frequently called molecular engineering, is emerging with
the ultimate goal to solve chemical process problems through
detailed structural modification at the molecular level.
(1) (a) Parshall, G. W.; Ittel, S. D. Homogeneous Catalysis; Wiley: New
York, 1992. (b) Cornils, B.; Herrmann, W. A. Applied Homogeneous
Catalysis with Organometallic Compounds; VCH: Weinheim, 1996. (c)
Herrmann, W. A.; Cornils, B. Angew. Chem., Int. Ed. Engl. 1997, 36, 1049.
(2) (a) Drent, E.; Arnoldy, P.; Budzelaar, P. H. M. J. Organomet. Chem.
1994, 475, 57. (b) Zhang, X.-X.; Wayland, B. B. J. Am. Chem. Soc. 1994,
116, 7894.
(3) (a) Coates, G. W.; Waymouth, R. M. Science 1995, 267, 217. (b)
Johnson, L. K.; Killian, C. M.; Brookhart, M. J. Am. Chem. Soc. 1995,
117, 6414.
(4) (a) 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. (b) Tsang, S. C.; Davis, J. J.; Green, M. L. H.; Hill, H. A.
O.; Leung, Y. C.; Sadler, P. J. Chem. Commun. 1995, 1803. (c) Bergbreiter,
D. E.; Caraway, J. W. J. Am. Chem. Soc. 1996, 118, 6092. (d) Neumann,
R.; Cohen, M. Angew. Chem., Int. Ed. Engl. 1997, 36, 1738.
(5) Bergbreiter, D. E.; Zhang, L.; Mariagnam, V. M. J. Am. Chem. Soc.
1993, 115, 9295.
Homogeneous organometallic catalysis is particularly well
suited for molecular engineering, as the performance of the
† Exxon Research and Engineering Company.
‡ Permanent address: Department of Organic Chemistry, Eo¨tvo¨s Uni-
versity, Budapest, Hungary.
§ Exxon Chemical Company.
S0002-7863(97)03833-X CCC: $15.00 © 1998 American Chemical Society
Published on Web 03/13/1998

3134 J. Am. Chem. Soc., Vol. 120, No. 13, 1998
HorVa´th et al.
many homogeneous organometallic catalysts. For example, the
hydroformylation of olefins is an important industrial process
for the production of aldehydes from olefins, CO, and H2 in
the presence of homogeneous transition-metal catalysts.6 One
of the most challenging problems associated with the triphen-
ylphosphine-modified rhodium catalyst system (Rh/PPh3)7 has
Scheme 1
O
O
RCH CH2 + CO + H2
RCH2CH2C
normal
+ R(CH3)CHC
iso
H
H
fluorous biphase concept was first demonstrated for the hydro-
formylation of decene-1 using the tris(1H,1H,2H,2H-perfluo-
rooctyl)phosphine-modified rhodium catalyst (Rh/P[CH2-
CH2(CF2)5CF3]3).12 It was designed to achieve several process
requirements with a single catalyst, including (1) the possibility
to use the same catalyst for the hydroformylation of lower and
higher molecular weight olefins, (2) facile and effective separa-
tion of the aldehydes from the catalyst, (3) appropriate coordina-
tion power of the ligand to keep the active rhodium species
stable and prevent rhodium from leaching out of the fluorous
phase, (4) comparable activity and selectivity with the com-
mercially used Rh/PPh3 catalyst system, and (5) the advantages
of single-phase catalysis with biphase product separation by
running the reaction at higher temperatures, where the system
forms a single phase, and separating the products from the
fluorous catalyst at lower temperatures. We now report our
detailed study on the fluorous-soluble hydroformylation catalyst,
HRh(CO){P[CH2CH2(CF2)5CF3]3}3 (1).
been the separation of higher aldehydes (Cn > 8) from the
catalyst without deactivating the catalyst.8 The use of an
aqueous biphase system, in which the water phase contains the
P(m-C6H4SO3Na)3-modified rhodium catalyst,9 offers an easy
separation of the organic products, and it has been used
commercially for the hydroformylation of propylene.10 How-
ever, since the catalytic reaction occurs in the aqueous phase,
the potential application of the aqueous biphase system is limited
by the solubility of the olefins (Cn > 7) in the catalyst containing
the aqueous phase.11
We have recently developed a novel concept for performing
stoichiometric and catalytic reactions using fluorous biphase
systems (FBSs).12 The FBS is based on the limited miscibility
of perfluorinated alkanes, perfluorinated dialkyl ethers, and
perfluorinated trialkylamines with nonfluorinated compounds.13
The fluorous biphase systems consist of a fluorous phase
containing a dissolved reagent12,14 or catalyst12,15 and a product
phase with limited solubility in the fluorous phase which could
be any common organic or nonorganic solvent (Scheme 1). The
Results and Discussions
(6) Falbe, J. New Synthesis with Carbon Monoxide; Springer-Verlag:
Berlin, 1980.
Selection of the Fluorous-Soluble Phosphine. Although
perfluoroalkylated triphenylphosphines were already known,16
we were concerned that the aromatic C-H and C-F bonds do
offer dipole-dipole interactions,17 making aryl- and perfluo-
roaryl-containing ligands soluble in common organic solvents
and thus less desirable for fluorous biphase systems. According
to the available liquid-liquid miscibility data,13 it was clear
that the best fluorous solubility could be achieved by attaching
three perfluorinated alkyl groups (fluorous ponytails) to the
phosphorus atom.12 However, the attachment of the highly
electron-withdrawing fluorous ponytails directly to the phos-
phorus atom could significantly lower the coordinating power
(7) (a) O’Connor, C.; Yagupsky, G.; Evans, D.; Wilkinson, G. Chem.
Commun. 1968, 420. (b) Evans, D.; Yagupsky, G.; Wilkinson, G. J. Chem.
Soc. A 1968, 2660. (c) Evans, D.; Osborn, J. A.; Wilkinson, G. J. Chem.
Soc. A 1968, 3133. (d) Brown, C. K.; Wilkinson, G. J. Chem. Soc. A 1970,
2753. (e) Pruett, R. L.; Smith, J. A. J. Org. Chem. 1969, 34, 327. (f) Pruett,
R. L. AdV. Organomet. Chem. 1979, 17, 1. (g) Gregorio, G.; Montrasi, G.;
Tampierei, M.; Cavaliezi d’Oro, P.; Pagani, G.; Andreetta, A. Chim. Ind.
(Milan) 1980, 62, 389. (h) Cavaliezi d’Oro, P.; Raimondi, L.; Pagani, G.;
Montrasi, G.; Gregorio, G.; Andreetta, A. Chim. Ind. (Milan) 1980, 62,
572. (i) Oswald, A. A.; Merola, J. S.; Mozeleski, E. J.; Kastrup, R. J.; Reisch,
R. V. In AdVances in Chemistry Series; Quin, L. D., Verkade, J. G., Eds.;
American Chemical Society: Washington, DC, 1981; Vol. 171, p 503. (j)
Kastrup, R. J.; Merola, J. S.; Oswald, A. A. In AdVances in Chemistry Series;
Alyea, E. C., Meek, D. W., Eds.; American Chemical Society: Washington,
DC, 1982; Vol. 196, p 43. (k) Oswald, A. A.; Hendriksen, D. E.; Kastrup,
R. J.; Merola, J. S. Prepr. Pap., Am. Chem. Soc., DiV. Pet. Chem. 1982,
27, 292. (l) Brown, J. M.; Canning, L. R.; Kent, A. G.; Sidebottom, P. J.
Chem. Commun. 1982, 721. (m) Brown, J. M.; Kent, A. G. Chem. Commun.
1982, 723. (n) Brown, J. M.; Kent, A. G. J. Chem. Soc., Perkin Trans. 2
1987, 1597. (o) Moser, W. R.; Papile, C. J.; Brannon, D. A.; Duwell, R.
A.; Weininger, S. J. J. Mol. Catal. 1987, 41, 271.
(14) (a) Bergbreiter, D. E. Chemtracts Org. Chem. 1995, 8, 108. (b)
Curran, D. P. Chemtracts Org. Chem. 1996, 9, 75. (c) Curran, D. P.; Hadida,
S. J. Am. Chem. Soc. 1996, 118, 2531. (d) DiMagno, S. G.; Dussault, P.
H.; Schultz, J. A. J. Am. Chem. Soc. 1996, 118, 5312. (e) Curran, D. P.;
Hoshino, M. J. Org. Chem. 1996, 61, 6480. (f) Larhed, M.; Hoshino, M.;
Hadida, S.; Curran, D. P.; Hallberg, A. J. Org. Chem. 1997, 62, 5583. (g)
Studer, A.; Hadida, S.; Ferritto, R.; Kim, S. Y.; Wipf, P.; Curran, D. P.
Science 1997, 275, 823. (h) Horner, J. H.; Martinez, F. N.; Newcomb, M.;
Hadida, S.; Curran, D. P. Tetrahedron Lett. 1997, 38, 2783. (i) Studer, A.;
Jeger, P.; Wipf, P.; Curran, D. P. J. Org. Chem. 1997, 62, 2917. (j) Studer,
A.; Curran, D. P. Tetrahedron 1997, 53, 6681. (k) Bergbreiter, D. E.;
Franchina, J. G. Chem. Commun. 1997, 1531.
(8) (a) Jamerson, J. D.; Pruett, R. L.; Billig, E.; Fiato, R. A. J. Organomet.
Chem. 1980, 193, C43. (b) Morrell, D. G.; Sherman, P. D. U.S. Patent
4260 828, 1981. (c) Bryant, D. R.; Billig, E. U.S. Patent 4277 627, 1981.
(d) Abatjoglou, A. G.; Billig, E.; Bryant, D. R. Organometallics 1984, 3,
923. (e) Oswald, A. A.; Jermasen, T. G.; Westner, A. A.; Huang I. U.S.
Patent 4668 809, 1987. (f) Deshpande, R. M.; Divekar, S. S.; Gholap, R.
V.; Chaudhari, R. V. J. Mol. Catal. 1991, 67, 333.
(9) (a) Kuntz, E. G. CHEMTECH 1987, 17, 570. (b) Herrmann, W. A.;
Kohlpaintner, W. Angew. Chem., Int. Ed. Engl. 1993, 32, 1524.
(10) (a) Wiebus, E.; Cornils, B. Chem.-Ing.-Tech. 1994, 66, 916. (b)
Cornils, B.; Wiebus, E. CHEMTECH 1995, 25, 33.
(11) Horva´th, I. T. Catal. Lett. 1990, 6, 43.
(12) (a) Horva´th, I. T.; Ra´bai, J. Science 1994, 266, 72. (b) U.S. Patent
5463 082, 1995.
(13) (a) Grosse, A. V.; Cady, G. H. Ind. Eng. Chem. 1947, 39, 367. (b)
Scott, R. L. J. Am. Chem. Soc. 1948, 70, 4090. (c) Hildebrand, J. H.;
Cochran, D. R. F. J. Am. Chem. Soc. 1949, 71, 22. (d) Zadymova, N. M.;
Markina, Z. N.; Yevseyeva, N. K.; Schukin, E. D. Kolloidn. Zh. 1988, 50,
825. (e) Dorset, D. L. Macromolecules 1990, 23, 894. (f) Kloubek, J Colloids
Surf. 1991, 55, 191.
(15) (a) Pozzi, G.; Montanari, F.; Quic, S. Chem. Commun. 1997, 69.
(b) Klement, I.; Lutjens, H.; Knochel, P. Angew. Chem., Int. Ed. Engl. 1997,
36, 1454. (c) Juliette, J. J. J.; Horva´th, I. T.; Gladysz, J. A. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 1610. (d) Vincent, J.-M.; Rabion, A.; Yachandra,
V. K.; Fish, R. H. Angew. Chem., Int. Ed. Engl. 1997, 36, 2346. (e)
Betzemeier, B.; Knochel, P. Angew. Chem., Int. Ed. Engl. 1997, 36, 2623.
(16) (a) Tamborski, C. U.S. Patent 3499 041, 1970. (b) Tamborski, C.;
Snyder, C. E. U.S. Patent 4011 267, 1977. (c) Tamborski, C.; Snyder, C.
E.; Christian, J. B. U.S. Patent 4454 349, 1984. (d) Gavezotti, P.; Strepparola,
E. U.S. Patent 4681 693, 1987.
(17) Filler, R. In Fluorine Containing Molecules; Liebman, J. F.,
Greenberg, A., Dolbier, W. R., Jr., Eds.; VCH: Weinheim, 1988; Chapter
2.

Molecular Engineering in Homogeneous Catalysis
J. Am. Chem. Soc., Vol. 120, No. 13, 1998 3135
Table 1. Electronic and Physical Properties of
P[(CH2)x(CF2)yCF3]3, P[(CH2)3CH3], and PPh3
perfluoroalkyl moiety. The effect of perfluoroethyl ponytails,
which are sufficient models for longer perfluoroalkyl groups,
is small for two (x ) 2) and essentially negligible for three (x
) 3) methylene units. However, the differences between the
electronic properties of P[(CH2)xCF2CF3]3 with more than three
methylene groups (x > 3) and P[(CH2)3CH3]3 are small but
finite. These results suggest that the insertion of two methylene
groups (x ) 2) is enough to lower the electron-withdrawing
effect of the fluorous ponytails. To obtain high fluorous
solubility, the attachment of perfluorohexyl groups seems
appropriate, as the three perfluorohexyl groups provide a
fluorous domain of about 75% of the total volume of P[CH2-
CH2(CF2)5CF3]3.
P lone-pair P Mulliken protonation P-H, HPL,
level, eV population (q) energy, eV
species
Å
deg
P[CF2CF3]3
-11.7
-11.7
-10.6
-9.9
-9.5
-9.3
-9.2
-8.7
-9.0
0.83
0.83
0.62
0.48
0.40
0.38
0.36
0.33
0.67
-6.5
-6.4
-7.7
-8.3
-8.6
-8.8
-8.9
-9.3
-9.9
1.189 85.9
1.192 85.4
1.205 86.3
1.218 92.3
1.225 91.8
1.226 92.0
1.228 91.8
1.230 91.7
1.226 61.6
P[(CF2)3CF3]3
P[CH2CF2CF3]3
P[(CH2)2CF2CF3]3
P[(CH2)3CF2CF3]3
P[(CH2)4CF2CF3]3
P[(CH2)5CF2CF3]3
P[(CH2)3CH3]3
PPh3
Synthesis and Characterization of P[CH2CH2(CF2)5CF3]3.
Since the reported synthesis for P[CH2CH2(CF2)5CF3]3 gave
rather low yields in our hands,21 we developed a new procedure
based on the well-known reaction of PH3 with olefins.22 The
reaction of PH3 (25 mmol) with CH2dCH(CF2)5CF3 (100 mmol)
was performed in the presence of azobis(isobutyronitrile)
(AIBN) at 100 °C for 2 h. 31P NMR of the reaction mixture
showed the formation of 20% P[CH2CH2(CF2)5CF3]3, 4%
HP[CH2CH2(CF2)5CF3]2, and 2% H2P[CH2CH2(CF2)5CF3]. Fur-
ther addition of AIBN and heating the solution to 80 °C for 8
h resulted in the disappearance of the mono- and dialkylphos-
phines. After purification, we obtained 26% P[CH2CH2(CF2)5-
CF3]3. It has recently been shown that the yield could be
increased to 53% by performing the addition of PH3 to
CH2dCH(CF2)5CF3 at lower temperatures (80-85 °C) for 24
h.23 NMR and MS data are in agreement with the formula of
P[CH2CH2(CF2)5CF3]3.
of the phosphine. We hypothesized that the insertion of
“insulating” methylene groups [-(CH2)-] before the fluorous
ponytails could circumvent this problem. The selection of the
number of -(CH2)- units between the phosphorus atom and
the perfluoroalkyl moiety was based on semiempirical calcula-
tions of the electronic properties of P[(CH2)x(CF2)yCF3]3 (x )
0, y ) 2 or 4 and x ) 0-5, y ) 2), P[(CH2)3CH3], and PPh3.
The calculations were performed using the UniChem version
of MNDO93 and employed the PM3 parameter set.18 Full
geometry optimizations were performed. Since the purpose of
the calculations was to determine the number of -(CH2)-
spacer units necessary to minimize the effect of the strongly
electron-withdrawing perfluoroalkyl moieties, the relative trends
in the results and not the absolute numbers are of interest.
Although similar calculations using different parameter sets or
more advanced techniques might give different absolute num-
bers, it is felt that the established accuracy of the method used
is sufficient to determine the relative trends.19
Phosphines (PR3) are primarily electron-donating ligands with
donation occurring through the lone pair of the phosphorus atom.
Besides direct electronic interactions, steric factors also play
an important role in the coordination chemistry of phosphines.20
In a molecular orbital picture, the ability of species to donate
charge to a given acceptor can be characterized by the energy
gap between the donor and the acceptor orbitals and their spatial
overlap. Thus, a first approximation to the effect of the
perfluoroalkyl units is the position of the P lone-pair orbital. In
addition, the amount of charge on the P atom should also
correlate with the size of the lone pair and, hence, its ability to
donate electrons. The P lone-pair eigenvalues and the P
Mulliken charge for P[(CH2)x(CF2)yCF3]3 (x ) 0, y ) 2 or 4
and x ) 0-5, y ) 2), P[(CH2)3CH3], and PPh3 are listed in
Table 1. In addition, the protonation energies (∆E) of the PR3
species for the reaction of H+ + PR3 f HPR3+ were calculated
and are included in Table 1. Zero-point-energy corrections
would need to be included to obtain the corresponding ∆H’s.
Negative values imply that the protonated ligand is more stable.
It is unrealistic to compare these results directly with the
measured values because solvent effects are not included. For
comparison, the value calculated for H2O is -5.8 eV. In an
attempt to characterize the potential for steric interactions, the
minimum angles subtended by the proton, phosphorus, and R
are also reported in Table 1. The results clearly show that the
electronic properties of P[(CH2)x(CF2)yCF3]3 (x ) 0, y ) 2 or
4 and x ) 0-5, y ) 2) can be tuned by varying the number of
methylene groups between the phosphorus atom and the
Characterization of HRh(CO){P[CH2CH2(CF2)5CF3]3}3
(1). The fluorous-soluble 1 was prepared by reacting 0.1 mmol
of Rh(CO)2(acac) with 0.3 mmol of P[CH2CH2(CF2)5CF3]3 in
C6F11CF3 (perfluoromethylcyclohexane) and treating the fluo-
rous phase with 2.74 MPa of CO/H2 (1:1) at 80 °C for 2 h. The
1H NMR in C6F11CF3 shows a quartet at -11.82 ppm (JH-P
)
14 Hz), which becomes a doublet of quartets by preparing the
sample with 99% enriched 13CO (JH-C ) 35 Hz). The 31P NMR
spectrum shows a doublet of doublets at 19.76 ppm (JP-Rh
)
149.3, JP-H ) 14 Hz) for the phosphine ligand. When the same
solution is prepared with 13CO, the 31P{1H} NMR spectrum
shows a doublet of doublets (JP-Rh ) 149.3, JP-C ) 10 Hz).
Finally, the 13C{1H}-NMR shows a doublet of quartets at 205.4
ppm (JRh-C ) 50, JP-C ) 10 Hz). The IR spectrum of 1 in
C6F11CF3 shows two bands at 2004 and 1976 cm-1, which are
attributed to the rhodium hydride and the terminal carbonyl
ligand vibrations, respectively. This was confirmed by preparing
a
13CO-labeled 1, which resulted in a shift of the rhodium
terminal carbonyl ligand vibration from 1976 to 1930 cm-1
.
Thus, the overall structure of 1 is very similar to those of HRh-
(CO)(PPh3)3 (2)7a-d and HRh(CO){P(m-C6H4SO3Na)3}3 (3).24
Mechanistic studies on phosphine-modified hydroformylation
catalysts (Rh/PR3) have clearly established that the coordina-
tively unsaturated species {HRh(CO)(PR3)2} and {HRh(CO)2-
(PR3)} are key intermediates. Their reaction with olefins leads
to competing catalytic cycles involving one or two phosphine
ligands on the rhodium (Scheme 2).7 It is believed that the
active rhodium species with two phosphine ligands or one
(21) Benefice-Malouet, S.; Blancou, H.; Commeyras, A. J. Fluorine
Chem. 1985, 30, 171. In this, the 31P NMR chemical shift for P[CH2-
CH2(CF2)5CF3]3 was erroneously reported at +26.4 ppm.21
(22) Rauhut, M. M.; Currier, H. A.; Semsel, A. M.; Wystrach, V. P. J.
Org. Chem. 1961, 26, 5138.
(18) UniChem is a computational chemistry software package developed
by Cray Research, Inc. Included within the package is the MNDO93
program developed by W. Thiel and co-workers.
(19) James, J. P. S. In ReViews in Computational Chemisty; Lipkowitz,
K. B., Boyd, D. B., Eds.; VCH: New York, 1990; Vol. 1, p 45.
(20) Tolman, C. A. Chem. ReV. 1977, 77, 313.
(23) Juliette, J. J. J.; Gladysz, J. A. Personal communication, 1997.
(24) Horva´th, I. T.; Kastrup, R. V.; Oswald, A. A.; Mozeleski, E. J. Catal.
Lett. 1989, 2, 85.

3136 J. Am. Chem. Soc., Vol. 120, No. 13, 1998
HorVa´th et al.
Hydroformylation of Decene-1. Our initial study on the
hydroformylation of decene-1 with the Rh/P[CH2CH2(CF2)5-
CF3]3 catalyst proved the feasibility of the fluorous biphase
concept.12 We report here a detailed kinetic study in which
the catalytic performance of the Rh/P[CH2CH2(CF2)5CF3]3
catalyst is compared with that of the commercially used Rh/
PPh3 and the nonfluorous analogue Rh/P[(CH2)7CH3]3 catalysts.
The possible kinetic and selectivity effects of the fluorous
solvent C6F11CF3 have also been tested by using the Rh/PPh3
catalyst, which is soluble in a 50/50 vol % toluene/C6F11CF3
solvent mixture at and above 100 °C.
The hydroformylation of decene-1 with all three catalysts was
carried out in batch autoclaves at 100 °C and 1.1 MPa of CO/
H2 (1:1) in a 50/50 vol % toluene/C6F11CF3 solvent mixture or
in toluene. The concentration of Rh ranged from 0.02 to 2.2
mmol/L, and in all experiments, an excess of phosphine was
used (P/Rh > 3). It is important to note that at the reaction
conditions, the 50/50 vol % toluene/C6F11CF3 solvent mixture
formed a single homogeneous phase in the presence of all three
catalysts at the ligand concentrations studied.
The reaction is first order in the concentration of rhodium
for all three catalysts. The turnover frequency (TOF: mole of
aldehyde/mole of Rh/second) vs conversion plot for the Rh/
P[CH2CH2(CF2)5CF3]3 catalyst (Figure 2) is linear, indicating
first-order reaction in decene-1. The nonzero intercept reflects
a less than 100% hydroformylation selectivity which will be
addressed later. As expected, for a first-order reaction, the ln-
(1 - conversion/100) vs time plots are linear to more than two
half-lives (Figure 3). The slopes of these plots yield reaction
rates (integral method) which are essentially identical with the
TOF/[decene-1] values obtained from the differential method
described in the Experimental Section. The reaction is also first
order in the concentration of the substrate with the nonfluorous
analogue Rh/P[(CH2)7CH3]3 at all conditions studied. However,
the kinetic response of the Rh/PPh3 catalyst is more complicated.
While at high concentrations of PPh3 (ca. 0.2 mol/L) only first-
order kinetics can be observed, at low phosphine concentrations
(ca. 0.02 mol/L), a saturation in decene-1 occurs (see the
Appendix for data and detailed discussion).
Figure 1. High-pressure 31P{1H} NMR spectra of HRh(CO){P[CH2-
CH2(CF2)5CF3]3}3 (1) in the presence of 3 equiv of P[CH2CH2(CF2)5-
CF3]3 under (A) N2, (B) 2.07 MPa of CO/H2 (1:1), and (C) 8.27 MPa
of CO/H2 (1:1) in C6F11CF3.
Scheme 2
The kinetic response of the Rh/P[CH2CH2(CF2)5CF3]3 catalyst
to changes in decene-1 concentration is more similar to that of
the Rh/P[(CH2)7CH3]3 catalyst than that of the Rh/PPh3 catalyst.
These results are in good agreement with our modeling
prediction on the electronic nature of the fluorous-soluble P[CH2-
CH2(CF2)5CF3]3. The two CH2 groups can indeed shield the
electron-withdrawing effect of the fluorous ponytails, and thus,
P[CH2CH2(CF2)5CF3]3 is more basic than PPh3. In fact, recent
solution calorimetric investigation of CO substitution in [Rh-
(CO)2Cl]2 with P[CH2CH2(CF2)5CF3]3 indicates that its elec-
tronic parameters are similar to those of PMe2Ph.25
While the hydroformylation of decene-1 is inhibited by
P[CH2CH2(CF2)5CF3]3, the n/i ratio of the aldehyde increases
with increasing phosphine concentrations (Figure 4 and Table
2). GC analysis of the aldehyde products shows the formation
of undecanal and 1-methyldecanal only up to 98% conversion
of decene-1. The highest measured n/i ratio is 7.84 at 0.3 mol/L
P[CH2CH2(CF2)5CF3]3 concentration. The olefin isomerization
selectivity is about 10%, and it does not seem to be affected by
the concentration of the ligand. Since the hydroformylation rate
of internal olefins is much slower than that of a terminal double
bond,7 the formed internal olefins remain largely unconverted
in the reactor. The dominating isomer olefin product is decene-
2. The overall isomer distribution of the product decene mixture
phosphine ligand gives higher and lower normal/iso (n/i) ratios,
respectively. These two species could form via ligand dis-
sociation from HRh(CO)x(PR3)4-x (PR3 dissociation, x ) 1 or
2; CO dissociation, x ) 2 or 3). While HRh(CO)(PR3)3 and
HRh(CO)2(PR3)2 have been shown to be the dominant species
in the equilibrium under CO/H2 (1:1), HRh(CO)3PR3 has not
been observed.7
The reaction of 0.1 mmol of 1 with carbon monoxide in the
presence of 0.3 mmol of P[CH2CH2(CF2)5CF3]3 under different
CO/H2 (1:1) pressures was studied by high-pressure 31P{1H}
NMR (Figure 1). When a solution of 1 in C6F11CF3 was treated
with 2.07 MPa of CO/H2 (1:1), a new doublet appeared at 24.8
ppm (JP-Rh ) 148.5 Hz), the intensity of the resonance at 19.76
ppm for 1 decreased, and the intensity of the resonance at -25.2
ppm for the free ligand increased according to Scheme 2,
indicating the formation of HRh(CO)2{P[CH2CH2(CF2)5CF3]3}2
(4). By increasing the CO/H2 (1:1) pressure to 8.27 MPa, the
concentration of 1 decreased further and the concentrations of
2 and the free ligand increased proportionally. Finally, when
the CO/H2 (1:1) pressure was decreased to 0.1 MPa and the
solution was purged with N2, 1 re-formed quantitatively.
Detailed high-pressure NMR studies on the equilibria shown
in Scheme 2 involving HRh(CO)x(PR3)4-x (x ) 1-2; R )
(CH2)7CH3, CH2CH2(CF2)5CF3, Ph, and m-C6H4SO3Na) are in
progress and will be published elsewhere.
(25) Li, C.; Nolan, S. P.; Horva´th, I. T. Organometallics Submitted.

Molecular Engineering in Homogeneous Catalysis
J. Am. Chem. Soc., Vol. 120, No. 13, 1998 3137
Figure 2. Rate of hydroformylation of decene-1 with Rh/P[CH2CH2(CF2)5CF3]3 catalyst in a 50/50 vol % toluene/C6F11CF3 solvent mixture at 100
°C and 1.1 MPa of CO/H2 (1:1) as a function of conversion. [Rh] ) 0.814 mmol/L, [P[CH2CH2(CF2)5CF3]3] ) 22.1 mmol/L, and [decene-1]t)0
1.012 mol/L.
)
Figure 3. ln(1 - conversion/100) vs time plot rate for the hydroformylation of decene-1 with Rh/P[CH2CH2(CF2)5CF3]3 catalyst in a 50/50 vol %
toluene/C6F11CF3 solvent mixture at 100 °C and 1.1 MPa of CO/H2 (1:1). [Rh] ) 0.814 mmol/L, [P[CH2CH2(CF2)5CF3]3] ) 22.1 mmol/L, and
[decene-1]t)0 ) 1.012 mol/L.
conforms with a stepwise double-bond migration through
repeated olefin addition and â-elimination cycles. Finally, less
than 1% of decene-1 is hydrogenated to decane.
Based on the most coherent mechanism of the reaction,26 it
should be clear that the distribution of Rh intermediates with
different CO/phosphine ligand ratios depends on the relative
concentrations of CO and phosphine. At constant partial
pressure of CO (as in this study), the determining factor thus
becomes the concentration of the phosphine ligand. Our data
are in full agreement with this prediction (Table 2). Although
it is customary in the literature to report kinetic and selectivity
data as a function of P/Rh ratio,1b,6,27 we did not find any
correlation between those values when excess phosphine (P/
Rh > 3) is present. Rather, we found that the TOF/[decene-1]
and product selectivities are independent of the P/Rh ratio and
solely depend on the concentration of the phosphine, as long
as the CO/H2 ratio, total pressure, solvent, and temperature are
(27) (a) Oliver, K. L.; Booth, F. B. Hydrocarbon Process. 1970, 49 (4),
112. (b) Deshpande, R. M.; Bhanage, B. M.; Divekar, S. S.; Chaudhari, R.
V. J. Mol. Catal. 1993, 78, L37. (c) Claridge, J. B.; Douthwaite, R. E.;
Green, M. L. H.; Lago, R. M.; Tsang, S. C.; York, A. P. E. J. Mol. Catal.
1994, 89, 113. (d) Bartik, T.; Ding, H.; Bartik, B.; Hanson, B. E. J. Mol.
Catal. A: Chem. 1995, 98, 117. (e) Kurkin, V. I.; Sakulin, V. V.; Kobyakov,
A. K.; Slivinskii, Ye. V.; Basner, M. Ye.; Aronovich, R. A.; Korneyeva,
G. A.; Loktev, S. M. Petrol. Chem. 1994, 34 (5), 393.
(26) Tolman, C. A.; Faller, J. W. Mechanistic Studies of Catalytic
Reactions Using Spectroscopic and Kinetic Techniques. In Homogeneous
Catalysis with Metal Phosphine Complexes; Pignolet, L. H., Ed.; Plenum:
New York, 1983.

3138 J. Am. Chem. Soc., Vol. 120, No. 13, 1998
HorVa´th et al.
Figure 4. Effect of P[CH2CH2(CF2)5CF3]3 concentration on the rate and n/i ratio in the hydroformylation of decene-1 with Rh/P[CH2CH2(CF2)5-
CF3]3 catalyst in a 50/50 vol % toluene/C6F11CF3 solvent mixture at 100 °C and 1.1 MPa of CO/H2 (1:1).
Table 2. Effect of the Phosphine Concentration on the Activity
and n/i Selectivity of the Rh/P[CH2CH2(CF2)5CF3]3 Catalyst in the
Hydroformylation of Decene-1 in 50/50 vol % Toluene/C6F11CF3
Solvent Mixture at 100 °C and 1.1 MPa of CO/H2 (1:1)
Table 4. Effect of Solvent on the Rate and Selectivity of the
Rh/PPh3 Catalyst in the Hydroformylation of Decene-1 at 100 °C
and 1.1 MPa of CO/H2 (1:1)a
selectivity, %
TOF/[decene-1],
[P[CH2CH2(CF2)5CF3]3],
mmol/L
TOF/[decene-1],
(L/mol)/s
solvent
(L/mol)/s
n/i aldehyde isomrzn decane
P/Rh
n/i
toluene
toluene/C6F11CF3 ) 1
6.3 ( 0.3
4.1 ( 0.8
3.2
3.6
91.8
91.2
8.2
8.8
0.0
0.0
22.1
22.5
22.3
41.3
82.1
152.2
304.0
27
39
28
76
79
103
102
0.60
0.49
0.51
0.40
0.21
0.08
0.04
3.25
3.25
3.20
4.35
5.04
6.32
7.84
a [PPh3] ) 23.2 ( 0.3 mmol/L.
approximately 30% lower in the toluene/C6F11CF3 solvent
mixture than in toluene, the n/i selectivity is somewhat lower
in toluene than in toluene/C6F11CF3. These effects might be
due to the differences in gas solubilities. The aldehyde and
olefin isomerization selectivities are virtually the same in both
solvents. The results for the toluene/C6F11CF3 solvent systems
are consistent with that of a single-phase homogeneous hydro-
formylation reaction.
Table 3. Effect of the Ligands on the Activity and Selectivity of
the Phosphine-Modified Rh Catalyst in the Hydroformylation of
Decene-1 in 50/50 vol % Toluene/C6F11CF3 Solvent Mixture at 100
°C and 1.1 MPa of CO/H2 (1:1)
selectivity, %
TOF/[decene-1]
liganda
(L/mol)/s
n/i aldehyde isomrzn decane
The fluorous biphase catalyst-recovery concept was tested
in a semicontinuous decene-1 hydroformylation experiment.
After each batch reaction, the product was cooled and the formed
product and fluorous phases were allowed to separate. The
supernatant product phase was siphoned off, and the fluorous
catalyst solution left in the autoclave was used in the next run.
In 9 consecutive runs, a total turnover number (TON: mole of
aldehyde/mole of Rh) of more than 35 000 was measured with
a 4.2% Rh loss (Table 5). This loss represents 1.18 ppm of Rh
loss/mol of aldehyde. The measured Rh loss is quite constant
from run to run (Figure 5) and is due to the low but finite
solubility of the Rh/P[CH2CH2(CF2)5CF3]3 catalyst in the
product phase. Although no mass balance was established for
the fluorous phosphine ligand, the activity and selectivity
changes during the experiment indicate some ligand leaching
as well. Thus, the rate of reaction and the olefin isomerization
selectivity increase while the product aldehyde n/i ratio decreases
in consecutive runs as the concentration of the ligand in the
catalyst solution is decreased. The olefin hydrogenation
selectivity is low and essentially constant through the nine
P[CH2CH2(CF2)5CF3]3 0.53 ( 0.05 3.2
91.1
91.8
96.4
8.9
8.2
3.6
0.4
0.0
0.0
PPh3
6.27 ( 0.30 3.2
0.22 ( 0.03 2.3
P[(CH2)7CH3]3
a [L] ) 23.0 ( 1.5 mmol/L.
the same. Similar results were obtained for the Rh/PPh3 catalyst
as well (see Appendix).
The catalytic activity of the Rh/P[CH2CH2(CF2)5CF3]3 catalyst
is similar to that of the nonfluorous analogue Rh/P[(CH2)7CH3]3
and is an order of magnitude lower than that of the Rh/PPh3
catalyst (Table 3). Unlike the activity, the n/i product selectivity
of Rh/P[CH2CH2(CF2)5CF3]3 is closer to the selectivity of the
Rh/PPh3 catalyst than that of the Rh/P[(CH2)7CH3]3 catalyst
(Table 3). The hydrogenation of decene-1 is negligible, with
all three catalysts in agreement with earlier low-temperature and
low-pressure data.7d
The introduction of a fluorous solvent results in minor
changes in the activity and selectivity of the Rh/PPh3 catalyst
in the hydroformylation of decene-1 at 100 °C and 1.1 MPa of
CO/H2 (1:1) (Table 4). While the activity of the catalyst is

Molecular Engineering in Homogeneous Catalysis
J. Am. Chem. Soc., Vol. 120, No. 13, 1998 3139
Table 5. Rhodium Content, Activity, and Selectivity Data for the
Semicontinuous Hydroformylation of Decene-1 with the Rh/
P[CH2CH2(CF2)5CF3]3 Catalyst in 50/50 vol % Toluene/C6F11CF3
Solvent Mixture at 100 °C and 1.1 MPa of CO/H2 (1:1)a
The stability of the catalyst was tested in a continuous stirred
tank reactor (CSTR) with ethylene at 110 °C for 2 months. As
it is depicted by Figure 6, no measurable deactivation occurred
during that period despite the low ethylene concentration and
relatively high temperature, both of which are known to
destabilize the industrial Rh/PPh3 catalyst.8b,d,f In fact, the
activity maintenance of the Rh/P[CH2CH2(CF2)5CF3]3 catalyst
was better than the one reported for Rh/PPh3,8c which later lost
0.2% activity per day at the same temperature but a higher
ethylene partial pressure (46 kPa). These results again manifest
the similarity of P[CH2CH2(CF2)5CF3]3 as a modifying ligand
to trialkylphosphines rather than to triarylphosphines. Replace-
ment of the aryl groups of the phosphine ligand with alkyl
groups has been reported8e to increase the stability of the
phosphine-modified rhodium catalysts.
selectivity, %
Rh in the
cycle reactor, µmol
TOF/[decene-1],
(L/mol)/s
n/i aldehyde isomrzn decane
1
2
3
4
5
6
7
8
9
15.626
15.544
15.455
15.375
15.304
15.226
15.139
15.060
14.976
0.279
0.315
0.379
0.356
0.384
0.419
0.496
0.484
0.488
4.47
4.09
3.94
3.76
3.64
3.59
3.51
3.57
3.51
91.5
89.4
89.3
89.5
89.1
89.0
88.9
88.3
88.3
7.83
9.86
10.1
9.93
10.30
10.33
10.57
11.01
11.03
0.7
0.7
0.6
0.6
0.6
0.6
0.5
0.6
0.6
a [P[CH2CH2(CF2)5CF3]3] in cycle 1: 42.7 mmol/L.
cycles. These trends can be recognized in the previously
discussed (vide supra) batch kinetic experiments when the ligand
concentration is the independent variable. Although the fluorous
components are slowly removed with the organic product, it
has to be emphasized that this semicontinuous experiment serves
only as a demonstration of the FBS concept, and neither the
process nor the catalytic system is optimized. The use of heavier
fluorous solvents and longer fluorous ponytails on the phosphine
should further decrease the rhodium loss, and with simple
engineering solutions (e.g., distillation of the product and catalyst
recycle), virtually all Rh and ligand losses can be eliminated.
Continuous Hydroformylation of Ethylene. The perfor-
mance of the Rh/P[CH2CH2(CF2)5CF3]3 catalyst was also tested
in the hydroformylation of ethylene (Table 6) using FC-70
(isomers of perfluorotripentylamine) as a solvent. Just like in
the hydroformylation of decene-1, the reaction was found to
be first-order for both ethylene and Rh. Since the solubility of
ethylene has not been measured in the reaction mixture, the
rate of reaction is referenced to the partial pressure of ethylene
in the gas phase. The results show that while the rate of reaction
is the same within experimental error at CO partial pressures
of 460 and 210 kPa, the selectivity of ethane is doubled at the
lower CO partial pressure. These observations are in agreement
with some earlier reports. Thus, zero kinetic order in CO has
been reported7h for the hydroformylation of propylene. Also,
the strong suppression of olefin hydrogenation by CO has been
reported by Wilkinson et al.7c The negative first order for CO
in the hydrogenation of the substrate, on the other hand, is
predicted by the mechanism26 since the 16-electron alkyl
intermediate (RRh(CO)3-xLx, x ) 1 or 2) can be trapped by
CO and follow the hydroformylation route rather than olefin
hydrogenation.
Conclusions
We have established that the Rh/P[CH2CH2(CF2)5CF3]3
catalyst system can be used for the hydroformylation of ethylene
and decene-1 and the product propanal and undecanals can be
separated from the catalyst by distillation or phase separation,
respectively. The solution structure of HRh(CO){P[CH2-
CH2(CF2)5CF3]3}3 (1) is similar to that of HRh(CO)(PPh3)3 (2)
and HRh(CO){P(m-C6H4SO3Na)3}3 (3). High-pressure NMR
revealed that 1 is in equilibrium with HRh(CO)2{P[CH2-
CH2(CF2)5CF3]3}2 (4) under CO/H2 (1:1). The long-term
stability of the Rh/P[CH2CH2(CF2)5CF3]3 catalyst during the
continuous hydroformylation of ethylene is better than that of
the Rh/PPh3 catalyst. Kinetic studies on the hydroformylation
of ethylene and decene-1 show that the reaction is first order in
both rhodium and olefin. While the hydroformylation of
decene-1 is inhibited by P[CH2CH2(CF2)5CF3]3, the n/i ratio of
the undecanals increases with increasing phosphine concentra-
tion. The catalytic activity of the Rh/P[CH2CH2(CF2)5CF3]3
catalyst is similar to that of the nonfluorous analogue Rh/
P[(CH2)7CH3]3 catalyst and is an order of magnitude lower than
that of the Rh/PPh3 catalyst. In contrast, the n/i product
selectivity of Rh/P[CH2CH2(CF2)5CF3]3 is closer to the selectiv-
ity of the Rh/PPh3 catalyst than that of the Rh/P[(CH2)7CH3]3
catalyst. The fluorous biphase catalyst recovery concept was
tested in a semicontinuous hydroformylation of decene-1 with
the Rh/P[CH2CH2(CF2)5CF3]3 catalyst. During 9 consecutive
reaction/separation cycles, a total turnover of more than 35 000
was achieved with a loss of 1.18 ppm of Rh/mol of undecanals.
Rh/P[CH2CH2(CF2)5CF3]3 is the first catalyst system which
can be used for the hydroformylation of both low and high
molecular weight olefins and provides facile catalyst separation
for both low and high molecular weight aldehydes. Finally,
the most intriguing aspect of this catalyst system is that it
combines the advantages of one-phase catalysis with biphase
product separation by running the reaction at higher temperatures
and separating the products at lower temperatures.
The activation energy of the reaction was measured at two
CO/H2 ratios, and the average value of 20.0 kcal/mol (83.7 kJ/
mol) is essentially the same as reported for the hydroformylation
of propylene.7h Oswald and co-workers reported 22 kcal/mol
for butene-1.7j Chaudhari et al. have reported in a series of
papers the activation energies for the hydroformylation of
hexene-1,28a decene-1,28c and dodecene-128e to be 28, 11.8, and
13.7 kcal/mol, respectively. These latter values, however, were
obtained from an engineering fit of experimental data; thus, the
activation energies can be model dependent, which may explain
their wide variance and some of those unusually low values.
Experimental Section
Solvents (toluene, anhydrous, Aldrich; perfluoromethylcy-
clohexane or C6F11CF3, Fluorochem; FC-70, isomers of per-
fluorotripentylamine, 3M; and Freon-225, isomers of dichloro-
pentafluoropropane, Asahi Glass Co. Ltd.; and decene-1, 99.0%,
Chemsampco) were thoroughly deaerated but otherwise were
used as received. All gases (ethylene, UHP grade, and CO/H2
(1:1), UHP grade, both from MG Industries), PPh3 (M&T),
P[(CH2)7CH3]3, hexadecane (anhydrous, 99+%, Aldrich), and
Rh(acac)(CO)2 (99%, Strem) were used as received. Tris-
(1H,1H,2H,2H-perfluorooctyl)phosphine, P[CH2CH2(CF2)5-
(28) (a) Deshpande, R. M.; Chaudhari, R. V. Ind. Eng. Chem. Res. 1988,
27, 1996. (b) Deshpande, R. M.; Chaudhari, R. V. J. Catal. 1989, 115,
326. (c) Divekar, S. S.; Deshpande, R. M.; Chaudhari, R. V. Catal. Lett.
1993, 21, 191. (d) Deshpande, R. M.; Purwanto, Delmas, H.; Chaudhari,
R. V. Ind. Eng. Chem. Res. 1996, 35, 3927. (e) Bhanage, B. M.; Divekar,
S. S.; Deshpande, R. M.; Chaudhari, R. V. J. Mol. Catal. A: Chem. 1997,
115, 247.

3140 J. Am. Chem. Soc., Vol. 120, No. 13, 1998
HorVa´th et al.
Figure 5. Rate of hydroformylation of decene-1 and rhodium content of the reactor as a function of reaction cycle in the semicontinuous
hydroformylation of decene-1 with Rh/P[CH2CH2(CF2)5CF3]3 catalyst in a 50/50 vol % toluene/C6F11CF3 solvent mixture at 100 °C and 1.1 MPa
of CO/H2 (1:1). [Rh] ) 0.814 mmol/L, [P[CH2CH2(CF2)5CF3]3] ) 42.7 mmol/L, and [decene-1]t)0 ) 1.012 mol/L.
Table 6. Continuous Hydroformylation of Ethylene with the
Rh/P[CH2CH2(CF2)5CF3]3 Catalyst in FC-70
Addition of more AIBN (0.25 g) and heating the solution at 80
°C for 8 h resulted in the disappearance of the mono- and
dialkylphosphines. The reaction mixture was diluted with C6F11-
CF3 (25 mL) and washed with toluene (4 × 15 mL). Distillation
under vacuum (155 °C at 0.3 mmHg) yielded 26% tris-
(1H,1H,2H,2H-perfluorooctyl)phosphine. 31P{1H} NMR (C6F11-
CF3): -25.2 ppm. 1H NMR (neat): 1.87 (br, 6H, C6F13CH2)
and 1.39 (br, 6H, CH2P) ppm. 19F NMR (neat): -80.8 (t, J )
10.2 Hz, CF3), -114.3 (CH2CF2), -120.9 (CF2), -121.9 (CF2),
-122.4 (CF2), and -125.5 (CF2) ppm. MS: M+ ) 1072
(calcd: 1072).
TOF/pC H
,
2
4
selectivity, %
mmol/L mmol/L kPa kP2a °C Rh/kPa C2H4/s propanal ethane
[Rh],
[L],a pCO
,
pH
,
T,
mmol oxo/mol
2.171
2.171
2.171
2.171
3.205
3.205
3.205
3.205
43.4
43.4
43.4
43.4
44.1
44.1
44.1
44.1
458 417 79.9
457 416 90.0
462 415 100.0
216 437 101.8
210 431 80.0
206 430 89.9
210 432 100.0
210 435 110.0
0.612
1.583
3.292
3.260
0.645
1.600
3.060
4.651
99.3
99.3
99.3
98.5
98.8
98.7
98.5
98.2
0.7
0.7
0.7
1.5
1.2
1.3
1.5
1.8
a L ) P[CH2CH2(CF2)5CF3]3.
In Situ Preparation of HRh(CO){P[CH2CH2(CF2)5CF3]3}3
(1). A 10-mL vial was charged with 0.026 g (0.1 mmol) of
solid Rh(acac)(CO)2 and 0.322 g (0.3 mmol) of P[CH2-
CH2(CF2)5CF3]3 in 3 mL of C6F11CF3 under N2 in a glovebox.
After the mixture was stirred for 30 min, the light-yellow
solution was placed in a sapphire NMR tube. The tube was
placed in a safety shield and charged with 2.74 MPa of CO/H2
(1:1). The tube was transferred to a heater and kept at 80 °C
for 2 h. The tube was placed into the safety shield and
transferred to a hood, and the pressure was slowly released.
The solution was transferred into a 10-mm glass NMR tube
and purged with N2. Multinuclear NMR and IR measurements
showed the quantitative formation of 1.
CF3]3, was prepared, purified, and analyzed as described later.
All organic compounds were checked for purity by GC (Perkin-
Elmer Autosystem, Hewlett-Packard Pona column). NMR
spectra were recorded on a Varian 400-MHz spectrometer. High-
pressure NMR experiments were performed using single-crystal
sapphire NMR tubes; their design has been described else-
where.29 IR spectra were recorded on a Mattson Galaxy 5020
spectrometer. Rhodium analyses were performed on an induc-
tively coupled plasma-atomic emission spectrometer (ICP/
AES), Jarrell-Ash Model 1100, with argon gas flow and torch
alignment. The ICP/AES instrument was calibrated with 10
ppm rhodium solution.
Synthesis of Tris(1H,1H,2H,2H-perfluorooctyl)phosphine
A 100-mL glass-lined autoclave was charged under N2 with 35
g (100 mmol) of 1H,1H,2H,2H-perfluoro-1-octene, 0.6 g of
azobis(isobutyronitrile) (AIBN), and 0.85 g (25 mmol) of PH3
at room temperature. The mixture was stirred and heated to
100 °C and kept at that temperature for 2 h. After the reactor
was cooled to room temperature, the unreacted PH3 was vented
to a scrubber containing an aqueous solution of 37% formaline
and 0.05% RhCl3. GC and 31P NMR (in CF2ClCCl2F) showed
the formation of H2PCH2CH2(CF2)5CF3 (2%, -139.3 ppm, t,
JP-H ) 189 Hz), HP[CH2CH2(CF2)5CF3]2 (4%, -67.1 ppm, d,
JP-H ) 194 Hz), and P[CH2CH2(CF2)5CF3]3 (20%, -24.9 ppm).
High-Pressure NMR of 1 under CO/H2 (1:1). A sapphire
NMR tube was charged with a solution of 0.335 g (0.1 mmol)
of 1 and 0.335 g (0.3 mmol) of P[CH2CH2(CF2)5CF3]3 in 3 mL
of C6F11CF3 under N2 in a glovebox. The tube was placed in
a safety shield and pressurized with CO/H2 (1:1) to the required
pressure. After the tube was transferred in the safety shield to
the top of the magnet, the tube was lowered into the probe of
the magnet. After the measurement was complete, the tube was
lifted into the safety shield and transferred to a hood, and the
pressure was slowly released. The solution was transferred to
a 10-mm glass NMR tube and purged with N2. 31P{1H} NMR
measurement showed that only 1 was present.
Hydroformylation of Decene-1. For kinetic measurements,
(29) Horva´th, I. T.; Ponce, E. C. ReV. Sci. Instrum. 1991, 62, 1104.
Rh stock solutions were prepared by using toluene for PPh3 or

Molecular Engineering in Homogeneous Catalysis
J. Am. Chem. Soc., Vol. 120, No. 13, 1998 3141
Figure 6. Continuous hydroformylation of ethylene with Rh/P[CH2CH2(CF2)5CF3]3 catalyst in FC-70 at 110 °C and 1.04 MPa of total pressure
using 745.7 cm3/min feed rate. [Rh] ) 3.2 mmol/L and [P[CH2CH2(CF2)5CF3]3] ) 44.1 mmol/L. The steady-state composition in the reactor is CO
(206 kPa, 19.6 vol %), H2 (426 kPa, 40.6 vol %), CH2dCH2 (4 kPa, 0.4 vol %), CH3CH3 (<1 kPa, <0.1 vol %), CH3CH2CHO (32 kPa, 3.0 vol
%), and CH4 (376 kPa, 35.8 vol %).
P[(CH2)7CH3]3 and C6F11CF3 or FC-70 for P[CH2CH2(CF2)5-
CF3]3. Stock solutions were made by weighing ((0.1 mg) the
solvent, Rh(acac)(CO)2, and the phosphine (P/Rh ratio of ca.
10) into a sealed bottle. The final catalyst solutions with the
required composition were prepared by weighing the Rh stock
solution, additional phosphine, and solvent(s).
Kinetic experiments were performed in a 150- or a 300-mL
Autoclave Engineers’ autoclave. The configurations of both
units were identical. The reactors were connected to their
volume-calibrated gas feed vessels (PVT tank) through a high-
precision pressure-regulator valve. The reactors and feed vessels
were equipped with certified Omega digital thermocouples and
pressure gauges. The temperatures of the reactors were
maintained within (0.5 °C.
The Rh charge at each condition was adjusted to ensure that
the initial gas-consumption rate was below the mass-transfer
limit, established by variable stirrer-speed experiments. The
rate and selectivity values obtained in the two autoclaves at the
same conditions were statistically indistinguishable. The re-
producibility was checked by carrying out at least two runs at
each condition.
(1:1). The pressure in the reactor was maintained constant by
making up the consumed CO/H2 (1:1) from the PVT tank. The
conversion was followed by recording the pressure drop in the
PVT tank. The reaction was stopped by rapid cooling of the
reactor after 80-95% conversion.
After the reaction mixture was cooled to room temperature,
the product was removed under N2. In the case of 50/50 vol %
toluene/C6F11CF3 biphase mixtures, Freon-225 was added to the
stirred reactor before removing the products. The addition of
Freon-225 increased the mutual solubilities and created a single
liquid phase. Aliquot samples of the product were prepared
under nitrogen for GC analysis by adding a known amount of
n-hexadecane as an internal standard. Mass balances were
established for all organic components and were typically 100
( 5%, except for the phosphines, for which the mass balances
were typically 100 ( 20%.
The turnover frequency (TOF) used in the kinetic analysis is
defined as the rate of olefin conversion by 1 mol of rhodium
and is expressed as the moles of olefin converted to aldehydes
per mole of rhodium per second:
TOF ) {(dnolefin/dt)t}/nRh
In a typical experiment, carried out in the 300-mL autoclave,
80 mL of catalyst solution was prepared in a Vacuum
Atmospheres glovebox from the rhodium stock solution, the
corresponding phosphine, and the solvent(s). The solution was
charged into the reactor under N2. The reactor was then flushed
and pressurized with CO/H2 (1:1) to ca. 500 kPa. Twenty
milliliters of olefin was charged in the glovebox into a 75-mL
injection bomb, which was then mounted into the CO/H2 (1:1)
feed line of the reactor. The catalyst was performed under CO/
H2 (1:1) at the reaction temperature. High-pressure 31P{1H}
NMR experiments were carried out under identical conditions
to verify the formation of the equilibrium mixture of the rhodium
hydrides with a general formula of HRh(CO)x(PR3)4-x (x ) 1
and 2).
The rate of olefin conversion is calculated from the PVT pressure
vs time data by approximating (dnolefin/dt)t with the finite
difference of (∆nolefin/∆t)t and utilizing the fact that ∆nolefin
(constant)∆pPVT
)
:
(dnolefin/dt)t = (∆nolefin/∆t)t ) (constant)(∆pPVT/∆t)t
i
i
i
where
constant ) (total moles of oxo product)/
(total PVT pressure drop)
The reaction was initiated by injecting the olefin into the
reactor and bringing the pressure up to 1.1 MPa with CO/H2
For the semicontinuous hydroformylation of decene-1, the
autoclave was fitted with a dip leg that reached ca. 5 mm above

3142 J. Am. Chem. Soc., Vol. 120, No. 13, 1998
HorVa´th et al.
Figure 7. Rate of hydroformylation of decene-1 with Rh/PPh3 catalyst at low PPh3 concentrations in toluene at 100 °C and 1.1 MPa of CO/H2
(1:1) as a function of conversion. 4: [Rh] ) 0.036 mmol/L, [PPh3] ) 23.0 mmol/L, and [decene-1]t)0 ) 1.04 mol/L. O: [Rh] ) 0.025 mmol/L,
[PPh3] ) 23.1 mmol/L, and [decene-1]t)0 )1.08 mol/L.
Table 7. Effect of PPh3 Concentration on the Activity and n/i
the heavier fluorous phase. Individual runs were carried out as
Selectivity of the Rh/PPh3 Catalyst in the Hydroformylation of
Decene-1 in Toluene at 100 °C and 1.1 MPa of CO/H2 (1:1)
described earlier for the kinetic experiments. At the end of each
run, the reactor was cooled to 15 °C; then the stirrer was stopped,
and the supernatant product phase was removed from the reactor.
The weight and the GC analysis of the product provided a mass
balance for all organic components. For Rh balance, the
concentration of Rh in each product and in the final catalyst
solution was determined by ICP-AES. For each consecutive
run, the catalyst solution was rapidly heated to the reaction
temperature and a fresh charge of decene-1 was injected to
initiate the reaction.
[PPh3], mmol/L
P/Rh
TOF/[decene-1], (L/mol)/s
n/i
228.6
228.8
23.7
23.1
23.0
602
993
223
639
917
0.78
0.73
6.25
6.58
5.98
4.8
4.7
3.2
3.2
3.2
Appendix
Hydroformylation of Ethylene. The continuous-flow hy-
droformylation of ethylene was carried out in a 500-mL
Autoclave Engineers’ zipperclave. The catalyst solution was
prepared by reacting 0.202 g (0.78 mmol) of Rh(acac)(CO)2
with 11.57 g (10.80 mmol) of P[CH2CH2(CF2)5CF3]3 in 465 g
of FC-70 under N2 in a glovebox and was charged into the
reactor as described for the hydroformylation of decene-1. The
gaseous feed components, H2, CO, C2H4, and CH4 internal
standard, were individually fed via Brooks mass-flow control-
lers.
Hydroformylation of Decene-1 with the Rh/PPh3 Catalyst. The
kinetic response of the Rh/PPh3 catalyst with respect to the concentration
of the decene-1 is not uniform. While at high concentrations of PPh3
(ca. 0.2 mol/L) only first-order kinetics can be observed, at low
phosphine concentrations (ca. 0.02 mol/L), a saturation kinetics occurs
(Figure 7). Mass-transfer limitation can be excluded since the measured
maximum rate of gas consumption is well below the experimentally
determined mass-transfer limit and the maximum TOF is independent
of the concentration of Rh (Figure 7).
Saturation kinetics for decene-128c and dodecene-128e using the Rh/
PPh3 catalyst has been previously observed. Similar kinetic response
has been reported7d for the hydroformylation of hexene-1 in benzene
at 25 °C with HRh(CO)(PPh3)3 (2), with no excess PPh3 present. In
their study, the maximum TOF does depend on the concentration of 2.
However, at their conditions (no excess ligand and low [Rh]), the kinetic
order for Rh is less than one, while with excess ligand it is one for a
variety of substrates and ligands.7i,h,28 Data obtained in the present study
are also consistent with a first kinetic order for Rh.
Mass balances for each component were obtained indepen-
dently of each other and typically were 100 ( 10% for CO and
H2 and 100 ( 5% for ethylene. At the end of the experiment,
the formation of 4% OdP[CH2CH2(CF2)5CF3]3 was observed
by 31P{1H} NMR.
Rhodium Analysis. Aliquots of the product or the catalyst
phase were removed. The residue was ashed with 1 mL of
concentrated H2SO4 and muffled at 530 °C for at least 8 h. The
remaining residue was fused with 0.5 g of K2S2O7. The fusate
was solubilized in a 10% H2SO4 solution. The rhodium analysis
was perfomed using ICP-AES instrumentation with matrix-
matched standards (3.3% RSD).
The existence of decene-1 saturation suggests a rate-limiting step at
the saturation limit before which there is a quasi-equilibrium involving
the olefin. In the most coherent mechanism published to date,26 the
earliest such possible equilibrium step is the formation of HRh-
(C7H15CHdCH2)(CO)x(PPh3)3-x (x ) 1 or 2) but can also involve the
alkyl [C9H19Rh(CO)x(L)4-x and C7H15CH(CH3)Rh(CO)x(L)4-x] or the
acyl [C9H19CORh(CO)x(L)4-x and C7H15CH(CH3)CORh(CO)x(L)4-x] (x
) 1, 2, or 3) intermediates as well. The corresponding rate-limiting
steps are olefin and CO insertion or H2 activation, respectively. Tolman
Acknowledgment. We are indebted to R. L. Espino, P. J.
Guzi, A. Kaldor, M. G. Matturro, S. C. Mraw, and W. Weissman
for their support and encouragement.

Molecular Engineering in Homogeneous Catalysis
J. Am. Chem. Soc., Vol. 120, No. 13, 1998 3143
and Faller have proposed26 that at the saturation limit, the rate-limiting
step is H2 activation, while at low olefin concentrations the limiting
rate is olefin addition to HRh(CO)x(PPh3)3-x (x ) 1 or 2). The buildup
of a non-acyl intermediate has been reported from in situ CIR-FTIR
studies of reacting phosphine-modified Rh catalyst systems.7o The
authors have assigned the observed IR spectrum to an 18-electron alkyl
complex. They also proposed26 that the shift from first to zero order
is due to a shift in the rate-limiting step from CO insertion in the alkyl
intermediate to CO dissociation from HRh(CO)2(L)2, which later is the
dominating species under typical hydroformylation conditions with Rh/
PPh3. NMR kinetic data,7l,n however, make this later scenario less
likely, since it has been found that the dissociation of CO and phosphine
is rapid on the time scale of catalytic hydroformylation.
For comparison, the effect of PPh3 concentration on the activity and
selectivity of the Rh/PPh3 catalyst in the hydroformylation of decene-1
in toluene was also established under similar conditions (Table 7). We
found that the TOF/[decene-1] and product selectivities are independent
of the P/Rh ratio and solely depend on the concentration of the
phosphine, as long as the CO/H2 ratio, total pressure, solvent, and
temperature are the same.
JA9738337