Effect of Ligand Modification on Rhodium-Catalyzed Homogeneous Hydroformylation in Supercritical Carbon Dioxide

Article abstract of DOI:10.1021/om990560a

Several fluoroalkyl- and fluoroalkoxy-substituted tertiary arylphosphines were synthesized and investigated in the homogeneous catalytic hydroformylation of 1-octene using HRh-(CO)L₃ (L = tertiary arylphosphine). The activity of the rhodium complex (formed in situ from Rh(CO)₂(acac) and L) increased with decreasing basicity of the phosphine according to the series [3,5-(CF₃)₂C₆H₃]₃P > [4-CF₃C₆H₄]₃P ≈ [3-CF₃C₆H₄]₃P > [4-CF₃C₆H₄]₃P > [4-F(CF₂)₄(CH₂)₃C₆H ₄]₃P. The very weakly basic phosphine (C₆F₅)₃P did not complex with Rh(CO)₂(acac), most likely due to a combination of electronic and steric factors. Steric effects did not play a role in either the activity or selectivity of the rhodium catalysts that were formed under hydroformylation conditions.

Full text of DOI:10.1021/om990560a

Organometallics 2000, 19, 81-86
81
Effect of Liga n d Mod ifica tion on Rh od iu m -Ca ta lyzed
Hom ogen eou s Hyd r ofor m yla tion in Su p er cr itica l Ca r bon
Dioxid e
Daniel R. Palo and Can Erkey*
Environmental Engineering Program, Department of Chemical Engineering, University of
Connecticut, 191 Auditorium Road, Storrs, Connecticut 06269
Received J uly 19, 1999
Several fluoroalkyl- and fluoroalkoxy-substituted tertiary arylphosphines were synthesized
and investigated in the homogeneous catalytic hydroformylation of 1-octene using HRh-
(CO)L3 (L ) tertiary arylphosphine). The activity of the rhodium complex (formed in situ
from Rh(CO)2(acac) and L) increased with decreasing basicity of the phosphine according to
the series [3,5-(CF3)2C6H3]3P > [4-CF3C6H4]3P ≈ [3-CF3C6H4]3P > [4-CF3OC6H4]3P >
[4-F(CF2)4(CH2)3C6H4]3P. The very weakly basic phosphine (C6F5)3P did not complex with
Rh(CO)2(acac), most likely due to a combination of electronic and steric factors. Steric effects
did not play a role in either the activity or selectivity of the rhodium catalysts that were
formed under hydroformylation conditions.
In tr od u ction
The continued quantification of electronic effects by
such measurements as basicity,17,18 FTIR,19,20 NMR,16,18,21
ionization potential,22 rate of complexation,23 enthalpy
of protonation,24 electrochemistry,25 enthalpy of reac-
tion,26-31 and QALE32-38 has supplemented Tolman’s
original ø-value concept. Steric and electronic effects are
Tertiary phosphines are by far the most extensively
used ligands in homogeneous catalysis.1,2 Phosphine-
modified systems offer several benefits over unmodified
systems, including increased catalyst stability, improved
reaction rates and selectivities, and enhanced partition-
ing in two-phase systems. Consequently, hundreds of
reports have been published investigating phosphine
synthesis, stability, and characterization.
(15) Moore, S. J .; Marzilli, L. G. Inorg. Chem. 1998, 37, 5329.
(16) Leitner, W.; Buhl, M.; Fornika, R.; Six, C.; Baumann, W.;
Dinjus, E.; Kessler, M.; Kruger, C.; Rufinska, A. Organometallics 1999,
18, 1196.
(17) Allman, T.; Goel, R. G. Can. J . Chem. 1982, 60, 716.
(18) Grim, S. O.; Yankowsky, A. W. J . Org. Chem. 1977, 42, 1236.
(19) Li, C.; Nolan, S. P.; Horvath, I. T. Organometallics 1998, 17,
452.
(20) Howell, J . A. S.; Lovatt, J . D.; McArdle, P.; Cunningham, D.;
Maimone, E.; Gottlieb, H. E.; Goldschmidt, Z. Inorg. Chem. Commun.
1998, 1, 118.
(21) Allen, D. W.; Taylor, B. F. J . Chem. Soc., Dalton Trans. 1982,
51.
(22) Weiner, M. A.; Lattman, M.; Grim, S. O. J . Org. Chem. 1975,
40, 1292.
(23) Amatore, C.; Carre, E.; J utand, A.; M’Barki, M. A. Organome-
tallics 1995, 14, 1818.
(24) Sowa, J . R.; Agelici, R. J . Inorg. Chem. 1991, 30, 3534.
(25) Zizelman, P. M.; Amatore, C.; Kochi, J . K. J . Am. Chem. Soc.
1984, 106, 3771.
(26) Serron, S. A.; Luo, L.; Li, C.; Cucullu, M. E.; Stevens, E. D.;
Nolan, S. P. Organometallics 1995, 14, 5290.
(27) Serron, S. A.; Nolan, S. P. Organometallics 1995, 14, 4611.
(28) Serron, S.; Nolan, S. P.; Moloy, K. G. Organometallics 1996,
15, 4301.
(29) Serron, S.; Huang, J .; Nolan, S. P. Organometallics 1998, 17,
534.
Early research published by Tolman on the steric (θ-
value) and electronic (ø-value) properties of phosphine
ligands3-6 inspired numerous subsequent studies aimed
at defining and understanding these effects in transition
metal systems. Some investigators have published ad-
ditional θ-measurements, while others have sought to
modify the cone angle concept, introducing the ideas of
ligand intermeshing, variable cone angle, cone angle
radial profiles, and the accessible molecular surface
(AMS) model.7-16
* Corresponding author. E-mail: cerkey@engr.uconn.edu. Tel: (860)
486-4601. Fax: (860) 486-2959.
(1) Cornils, B.; Herrmann, W. A. Applied Homogeneous Catalysis
with Organometallic Compounds; VCH: Weinheim, Germany, 1996;
Vol. 1.
(2) Parshall, G. W.; Ittel, S. D. Homogeneous Catalysis, 2nd ed.;
Wiley: New York, 1992.
(3) Tolman, C. A. J . Am. Chem. Soc. 1970, 92, 2956.
(4) Tolman, C. A. J . Am. Chem. Soc. 1970, 92, 2953.
(5) Tolman, C. A.; Reutter, D. W.; Seidel, W. C. J . Organomet. Chem.
1976, 117, C30.
(30) Huang, J .; Serron, S.; Nolan, S. P. Organometallics 1998, 17,
4004.
(31) Haar, C. M.; Nolan, S. P.; Marshall, W. J .; Moloy, K. G.; Prock,
A.; Giering, W. P. Organometallics 1999, 18, 474.
(32) QALE stands for quantitative analysis of ligand effects.
(33) Rahman, M. M.; Liu, H. Y.; Prock, A.; Giering, W. P. Organo-
metallics 1987, 6, 650.
(6) Tolman, C. A. Chem. Rev. 1977, 77, 313.
(7) Angermund, K.; Baumann, W.; Dinjus, E.; Fornika, R.; Gorls,
H.; Kessler, M.; Kruger, C.; Leitner, W.; Lutz, F. Chem.sEur. J . 1997,
3, 755-764.
(8) Trogler, W. C.; Marzilli, L. G. J . Am. Chem. Soc. 1974, 96, 7589.
(9) Clark, H. C. Isr. J . Chem. 1977, 15, 210.
(34) Rahman, M. M.; Liu, H.-Y.; Eriks, K.; Prock, A.; Giering, W. P.
Organometallics 1989, 8, 1.
(10) DeSanto, J . T.; Mosbo, J . A.; Storhoff, B. N.; Bock, P. L.; Bloss,
R. E. Inorg. Chem. 1980, 19, 3086.
(11) Luo, L.; Nolan, S. P. Organometallics 1994, 13, 4781.
(12) Smith, J . M.; Taverner, B. C.; Coville, N. J . J . Organomet.
Chem. 1997, 530, 131.
(13) J oerg, S.; Drago, R. S.; Sales, J . Organometallics 1998, 17, 589.
(14) Pruchnik, F. P.; Smolenski, P.; Wajda-Hermanowicz, K. J .
Organomet. Chem. 1998, 570, 63.
(35) Woska, D. C.; Wilson, M.; Bartholomew, J .; Eriks, K.; Prock,
A.; Giering, W. P. Organometallics 1992, 11, 3343.
(36) Woska, D. C.; Bartholomew, J .; Greene, J . E.; Eriks, K.; Prock,
A.; Giering, W. P. Organometallics 1993, 12, 304.
(37) Wilson, M. R.; Liu, H.; Prock, A.; Giering, W. P. Organometallics
1993, 12, 2044.
(38) Wilson, M. R.; Woska, D. C.; Prock, A.; Giering, W. P. Orga-
nometallics 1993, 12, 1742.
10.1021/om990560a CCC: $19.00 © 2000 American Chemical Society
Publication on Web 12/10/1999

82 Organometallics, Vol. 19, No. 1, 2000
Palo and Erkey
Sch em e 1. Hyd r ofor m yla tion of Olefin s
Sch em e 2. Syn th esis of F lu or oa lk yl-Su bstitu ted
Ar ylp h osp h in e 8 w ith Meth ylen e “Sp a cer s”
not easily separated, though some successful attempts
have been made to quantify each with respect to
phosphine behavior.13,15,16,20,25,30,33,39 In addition, several
studies have been published in which the electronic
effect is further separated into its σ- and π-compo-
nents.13,20,33,34,39,40
These studies have resulted in a plethora of tabulated
information (pKa, ø, θ, ν(CO), δ(1H), δ(31P), J (M-P)) that
can now be utilized in catalyst design. Such fundamen-
tal information is quite useful in the modification of
phosphine ligands for use in nontraditional applications
such as aqueous, fluorous,41-49 or supercritical fluid
systems.50,51 Quite a few of the studies in such nontra-
ditional applications have focused on hydroformylation,
which is the oldest and most widely used homogeneous
catalytic reaction involving olefins. This reaction in-
volves the formation of branched or linear aldehydes by
the addition of H2 and CO to a double bond according
to Scheme 1.
The linear aldehydes are the preferred products, and
the selectivity in such reactions is usually expressed in
terms of n:iso ratio, which is the ratio of the linear
aldehyde to the branched aldehyde. Rhone-Poulenc/
Ruhrchemie successfully demonstrated and commercial-
ized an aqueous-phase process for lower olefin hydro-
formylation,1,52 where meta-SO3Na substitution of the
aryl rings on the phosphine resulted in excellent catalyst
solubility in water. More recently, Horva´th and Ra´bai
reported the successful modification of transition-metal/
phosphine systems for hydroformylation in fluorous
biphase systems (FBS)41,42 using tri(perfluoroalkyl)-
phosphines. Using a technique similar to that in the
FBS application, Leitner and co-workers performed
ligand modification leading to the demonstration of the
first transition-metal/phosphine complex that is highly
soluble in supercritical carbon dioxide (scCO2).53,54
We recently reported on the catalysts RhCl(CO)(PR3)2
(1) and HRh(CO)(PR3)3 (2), where R ) p-CF3C6H4.55-57
Both catalysts were found to be moderately soluble in
scCO2 and active in the hydroformylation of various
unsaturated substrates at temperatures between 35 and
70 °C. Species 1 showed a significant induction period
due to the formation of HCl when 1 is converted to the
active catalyst species HRh(CO)(PR3)2 or HRh(CO)2-
(PR3) under reaction conditions. In our recent report on
the hydroformylation of 1-octene in scCO2,58 the kinetic
behavior of 2 in scCO2 was found to be significantly
different from the behavior of HRh(CO)(PPh3)3 in
organic solvents. Primary differences included ∼0.5
order dependence of the rate on hydrogen concentration
and lack of substrate inhibition. Several different factors
may be contributing to this behavior, including scCO2
solvent effects, high reactant gas concentrations, and
the nature of the modified phosphine. Therefore, com-
parison of the behavior of various substituted arylphos-
phines in scCO2 is one step toward determining the
reasons for the altered kinetics.
This paper describes the synthesis and characteriza-
tion of several fluoroalkyl- or fluoroalkoxy-substituted
arylphosphines and demonstrates their use in the
rhodium-catalyzed homogeneous hydroformylation of
1-octene in scCO2. Complexes analogous to 2 were
formed in situ under hydroformylation conditions, and
the effects of fluoroalkyl- and fluoroalkoxy-substitution
on activity and selectivity were investigated and quanti-
fied.
(39) Drago, R. S.; J oerg, S. J . Am. Chem. Soc. 1996, 118, 2654.
(40) Golovin, M. N.; Rahman, M.; Belmonte, J . E.; Giering, W. P.
Organometallics 1985, 4, 1981.
(41) Horvath, I. T.; Kiss, G.; Cook, R. A.; Bond, J . E.; Stevens, P.
A.; Rabai, J .; Mozeleski, E. J . J . Am. Chem. Soc. 1998, 120, 3133.
(42) Horvath, I. T.; Rabai, J . Science 1994, 266, 72.
(43) Guillevic, M. A.; Arif, A. M.; Horvath, I. T.; Gladysz, J . A.
Angew. Chem., Int. Ed. Engl. 1997, 36, 1612.
(44) J uliette, J . J . J .; Horvath, I. T.; Gladysz, J . A. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 1610.
(45) Bhatacharyya, P.; Gudmunsen, D.; Hope, E. G.; Kemmitt, R.
D.; Paige, D. R.; Stuart, A. M. J . Chem. Soc., Perkin Trans. 1 1997,
3609.
(46) Fawcett, J .; Hope, E. G.; Kemmitt, R. D. W.; Paige, D. R.;
Russell, D. R.; Stuart, A. M.; Cole-Hamilton, D. J .; Payne, M. J . Chem.
Commun. 1997, 1127.
(47) Fawcett, J .; Hope, E. G.; Kemmitt, R. D. W.; Paige, D. R.;
Russell, D. R.; Stuart, A. M. J . Chem. Soc., Dalton Trans. 1998, 3751.
(48) Sinou, D.; Pozzi, G.; Hope, E. G.; Stuart, A. M. Tetrahedron
Lett. 1999, 40, 849.
(49) Kling, R.; Sinou, D.; Pozzi, G.; Choplin, A.; Quignard, F.; Busch,
S.; Kainz, S.; Koch, D.; Leitner, W. Tetrahedron Lett. 1998, 39, 9439.
(50) Rathke, J . W.; Klinger, R. J . Cobalt Carbonyl Catalyzed Olefin
Hydroformylation in Supercritical Carbon Dioxide, U.S. Patent, 1993.
(51) J essop, P. G.; Ikariya, T.; Noyori, R. Chem. Rev. 1999, 99, 475.
(52) Cornils, B.; Kuntz, E. J . Organomet. Chem. 1995, 502, 177.
(53) Kainz, S.; Koch, D.; Baumann, W.; Leitner, W. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 1628.
Resu lts a n d Discu ssion
All the phosphines were synthesized from the corre-
sponding perfluoroalkyl- or perfluoroalkoxy-substituted
aryl bromides, which were all commercially available
except 4-bromo-(1H,1H,2H,2H,3H,3H-perfluoroheptyl)-
benzene (8c). Here, we describe an alternative method
for preparation of [4-F(CF2)4(CH2)3C6H4]3P (8) (Scheme
2.)
Comparison was made between the reaction rate for
2 to the rate for the species formed in situ from the
phosphine ligand and Rh(CO)2(acac). As can be seen
from Figure 1, no significant difference exists between
the two systems in either reaction rate or selectivity.
The result of this comparison was taken as confirmation
(55) Palo, D. R.; Erkey, C. Ind. Eng. Chem. Res. 1998, 37, 4203.
(56) Palo, D. R.; Erkey, C. Ind. Eng. Chem. Res. 1999, 38, 2163.
(57) Palo, D. R.; Erkey, C. Rhodium Catalyzed Homogeneous Hy-
droformylation of Unsaturated Compounds in Supercritical Carbon
Dioxide with [HRh(CO)(P(p-CF3C6H4)3)3]. In Reaction Engineering for
Pollution Prevention; Hesketh, R., Abraham, M., Eds.; Elsevier Science
(accepted for publication).
(54) Koch, D.; Leitner, W. J . Am. Chem. Soc. 1998, 120, 13398-
13404.
(58) Palo, D. R.; Erkey, C. Ind. Eng. Chem. Res. 1999, 38, 3786.

Rh-Catalyzed Homogeneous Hydroformylation
Organometallics, Vol. 19, No. 1, 2000 83
the significant effects phosphine ligands have on the
nature of the catalytic species. It is well-known that
fluorine has a very strong electron-withdrawing effect
on the phosphine, causing the phosphorus lone pair to
become less basic, and leading to a significant decrease
in the electron density at the metal center.17,20,24,34 The
trend of activity observed among the five phosphines of
Figure 2 indicates that less basic phosphines lead to
more active hydroformylation catalysts.
The catalytic behavior of the unmodified Rh(CO)2-
(acac) catalyst differs fundamentally from that observed
for the phosphine-modified catalysts, as seen by com-
paring Figure 2 and Figure 3. Not only does Rh(CO)2-
(acac) show a significant induction period (see Table 1),
but the rate of reaction does not decrease with increased
conversion as seen for the phosphine-modified systems.
This seems to indicate a zero-order dependence of the
rate on substrate concentration. Behavior of the system
modified with (C6F5)3P (7) is identical to that of Rh(CO)2-
(acac) alone (see Figure 3, Table 1), indicating that 7
does not significantly displace either CO or acac to form
a rhodium-phosphine complex. The lack of interaction
between 7 and Rh(CO)2(acac) is probably due primarily
to the extremely low basicity of 7, but no doubt partially
due to its steric bulk (θ7 ) 184°).3 Not surprisingly, an
attempt to produce HRh(CO)[(C6F5)3P]3 from 7, Rh-
(CO)2(acac), H2, and CO in scCO2 also failed.
F igu r e 1. Comparison of catalytic behavior of HRh(CO)-
[P(p-(CF3C6H4)3]3 with the species formed in situ from Rh-
(CO)2(acac) and P(p-(CF3C6H4)3: P ) 273 atm, T ) 50 °C,
[1-octene]0 ) 0.95 M, [H2]0 ) [CO]0 ) 1.1 M, [Rh] ) 1.27
mM.
Syntheses of the various catalysts in scCO2 were
performed for the purpose of measuring the NMR and
FTIR values listed in Table 2. These NMR measure-
ments on the catalysts shed further light on the
fundamental difference between the substituted phos-
phines and triphenylphosphine (PPh3). The most rec-
ognizable feature of the 1H NMR spectrum of the
complex HRh(CO)L3 is a quartet at δ ≈ -9.8, corre-
sponding to the rhodium hydride. The quartet arises
from coupling of the hydride with the three equivalent
phosphorus nuclei of the complex. For PPh3, this quartet
is only observed at low temperatures, where the phos-
phine exchange rate is reduced enough to allow resolu-
tion of the quartet.59 For phosphines (4-CF3C6H4)3P (3),
(3-CF3C6H4)3P (4), 5, and (4-CF3OC6H4)3P (6), the
quartet is well resolved at 30 °C, indicating a much
slower exchange rate for the fluoroalkyl-substituted
phosphines than for the more basic PPh3. However, the
quartet is not resolved for the complex formed from
[4-F(CF2)4(CH2)3C6H4]3P (8), indicating a similarity to
PPh3.
It can be seen from Figure 4 that the rate of hydro-
formylation correlates very well with the value of ν(CO)
for the various catalysts, showing a linear dependence.
The IR stretching frequency of metal carbonyls is a
sensitive indicator of electron density at the metal
center, yielding a relative measure of the amount of
π-back-bonding from an occupied metal d-orbital to the
empty π*-orbital of the carbonyl.3,19,20,29,40 The trend in
ν(CO) corresponds well with the amount and proximity
of the electron-withdrawing fluoroalkyl or fluoroalkoxy
groups of the phosphines, with 5 having the greatest
effect and 8 having the least effect relative to the value
for PPh3. The phosphines 3 and 4 show identical
electronic behavior and almost identical activity/selec-
F igu r e 2. Hydroformylation of 1-octene using HRh(CO)-
L3: P ) 273 atm, T ) 50 °C, [1-octene]0 ) 0.95 M, [H2]0 )
[CO]0 ) 1.1 M, [Rh] ) 1.27 mM, [P]/[Rh] ) 3.0.
that the active catalytic species is the same for the in
situ preparations as for the preformed catalysts, HRh-
(CO)L3. Thus, in situ preparations of the active catalyst
species allow direct comparison of the various catalysts
without the need to separately synthesize and purify
each complex.
Figure 2 shows the concentration versus time data
for the hydroformylation of 1-octene using Rh(CO)2-
(acac)/phosphine. All reactions were conducted at iden-
tical conditions, as described in the figure caption. Each
reaction proceeded smoothly, producing both normal
and branched C9 aldehydes. When [3,5-(CF3)2C6H3]3P
(5) was the phosphine, approximately 1% conversion to
C8 species resulted, but experiments employing the
remaining phosphines of Figure 2 showed no detectable
isomerization or hydrogenation.
Table 1 lists the initial rates and selectivities observed
for the various ligands under standard reaction condi-
tions. Initial rates were calculated from the linear
portion of each rate curve by estimating the slope in
mol dm-3 s-1. The selectivities listed are those measured
at the end of each experiment (between 90% and 98%
conversion). Among the various catalyst systems, the
initial rates of reaction differ more than 5-fold between
the slowest and the fastest. This fact alone speaks of
(59) Ahmad, N.; Levison, J . J .; Robinson, S. D.; Uttley, M. F. Inorg.
Synth. 1990, 28, 81.

84 Organometallics, Vol. 19, No. 1, 2000
Palo and Erkey
Ta ble 1. Hyd r ofor m yla tion of 1-Octen e in scCO2 Usin g Rh (CO)2(a ca c) a n d Va r iou s F lu or in a ted P h osp h in e
Liga n d s
106 × initial rate,a
induction period,
min
selectivity
(n:iso ratio)
ν(CO),b
ligand
mol dm-3 s-1
cm-1
θ, deg
[3,5-(CF3)2C6H3]3P 5
(4-CF3C6H4)3P 3
(3-CF3C6H4)3P 4
no ligand
(C6F5)3P 7
(4-CF3OC6H4)3P 6
[4-F(CF2)4(CH2)3C6H4]3P 8
263
188
185
166
161
94.8
51.2
0
0
0
10
10
0
3.02
3.29
3.13
1.54
1.55
2.99
3.09
1963
1950
1950
2068, 2000
160c
145d
153e
184f
145d
145d
1934
1921
0
a
Standard conditions: T ) 50 °C, P ) 273 atm, V ) 54 mL, [substrate]0 ) 1.0 M, [Rh] ) 1.2 mM, [P]/[Rh] ) 3.0, [CO]0 ) [H2]0 ) 1.1
M. For HRh(CO)L3. c From ref 20. Assumed to be the same as PPh3. e Estimated as the average between phosphines 3 and 5. f From
b
d
ref 3.
an effect is not observed in this study, where the
selectivities for the sterically identical but electronically
different phosphines 3, 6, and 8 are around 3. Hydro-
formylation studies conducted in organic solvents usu-
ally show that the n:iso ratio decreases with increasing
steric bulkiness of the phosphines.1 This is usually
attributed to the formation of the HRh(CO)2(R3P) spe-
cies, which favors the formation of branched products.
Surprisingly, such an effect is also not observed in this
study, where steric differences among the phosphines
have very little if any effect on the n:iso ratio. Phos-
phines 3 and 4, which are electronically identical but
sterically different (θ3 ) 145°, θ4 ) 153° 61), exhibit quite
similar activities and selectivities. Furthermore, phos-
phine 5 (θ5 ) 160°) shows no decrease in selectivity,
despite a significant increase in cone angle over phos-
phine 3.
F igu r e 3. Hydroformylation of 1-octene using Rh(CO)2-
(acac) alone or modified by P(C6F5)3 (3 equiv): P ) 273 atm,
T ) 50 °C, [1-octene]0 ) 0.95 M, [H2]0 ) [CO]0 ) 1.1 M,
[Rh] ) 1.27 mM.
Another result of this study was the observation that
Rh(CO)2(acac) and Rh(CO)2(hfacac) exhibit significantly
different behavior in scCO2. The induction period for Rh-
(CO)2(acac), as shown in Table 1, is only 10 min, while
the induction period for Rh(CO)2(hfacac) inferred from
Leitner’s data is somewhere between 30 and 60 min.54
The trifluoromethyl groups of the hfacac moiety seem
to inhibit the formation of the active rhodium species
under hydroformylation conditions.
tivity behavior, while 5, with a much higher ν(CO)
value, also has a significantly higher activity. These
results are in agreement with an earlier report that the
electronic effect of trifluoromethyl substitution is inde-
pendent of ortho, meta, or para placement and that the
effects of multiple trifluoromethyl substitutions are
cumulative.20 The ability of oxygen and methylene
groups to insulate against the electron-withdrawing
effects of the fluoroalkyl moieties can be seen for
phosphines 6 and 8, where significant decreases in ν-
(CO) are accompanied by 50% and 70% decreases in
activity, respectively. These results confirm the observa-
tion by Leitner that methylene spacers effectively
insulate the phosphorus lone pair from the electron-
withdrawing effect of the fluoroalkyl chain53 and also
show that oxygen spacers exhibit a similar though less
profound effect. While these spacers are effective insula-
tors, they actually decrease the catalytic activity relative
to “noninsulated” phosphines. These results on activity
are also in agreement with a study conducted in organic
solvents using modified phosphine ligands of the form
(4-XC6H4)3P, where the rate of hydroformylation of
1-hexene was found to increase with decreasing basicity
of the phosphine according to the series [4-CF3C6H4]3P
> [4-ClC6H4]3P > [4-FC6H4]3P > [4-HC6H4]3P > [4-OCH3-
C6H4]3P > [4-N(CH3)2C6H4]3P.60 In the same study, it
was also observed that the n:iso ratio increased with
decreasing basicity of the phosphines. Surprisingly, such
Con clu sion
We have shown that differences in the extent and
location of fluoroalkyl or fluoroalkoxy substitution on
tertiary phosphines have a significant effect on the
activities of rhodium-based hydroformylation catalysts
made therewith. It was shown that the electron-
withdrawing effect of fluoroalkyl moieties actually
increases the activity of rhodium hydroformylation
catalysts of the form HRh(CO)L3, where L is the
substituted arylphosphine ligand. Good correlation was
observed between activity and carbonyl stretching
frequency for the rhodium complexes formed from
substituted arylphosphines.
Exp er im en ta l Section
Gen er a l Meth od s. All phosphine and catalyst syntheses
were carried out under a nitrogen atmosphere. Phosphines 3
and 7 were purchased from Strem Chemicals and used as
received. Fluoroalkyl- or fluoroalkoxy-substituted bromoben-
zenes, allylbenzene, perfluorobutyl iodide, butyllithium, bro-
(60) Moser, W. R.; Papile, C. J .; Brannon, D. A.; Duwell, R. A.;
S.Weininger J . Mol. Catal. 1987, 41, 271.
(61) Estimated as the average between phosphines 3 and 5.

Rh-Catalyzed Homogeneous Hydroformylation
Organometallics, Vol. 19, No. 1, 2000 85
Ta ble 2. P r op er ties of Va r iou s F lu or in a ted P h osp h in e Liga n d s a n d Th eir Com p lexes w ith Rh od iu m a s
HRh (CO)L3
complexed as HRh(CO)L3
free δ(31P),
δ(31P),
ppm
J (Rh-P),
δ(1H),
ppm
3J (H-P),
ν(Rh-H),
ν(CO),
ligand
ppm
Hz
Hz
cm-1
cm-1
(C6H5)3P
(4-CF3C6H4)3P 3
(3-CF3C6H4)3P 4
[3,5-(CF3)2C6H3]3P 5
(4-CF3OC6H4)3P 6
(C6F5)3P 7e
-6.0a
-6.3
-5.4
-4.4
-8.9
-75
39.8b
40.2
42.9
44.0
37.2
155b
156
156
157
155
-9.7c
-9.9
14d
14
14
15
14
2040d
2038
2025
2033
2016
1923d
1950
1950
1963
1934
-9.7
-10.0
-9.9
[4-F(CF2)4(CH2)3C6H4]3P 8
-8.1
37.4
153
-9.7
f
1969
1921
a
b
d
From ref 6. From ref 64. c From ref 59. From ref 65. e HRh(CO)L3 did not form. f Unresolved at 30 °C.
Hz), 7.42 (1H, t, J ) 7 Hz) ppm. 31P{1H} NMR (CDCl3): δ -5.4
(s) ppm. Anal. Calcd for C21H12F9P: C, 54.09; H, 2.59. Found:
C, 53.66; H, 2.43.
Syn th esis of 6. Phosphine 6 was produced according to the
same procedure used for 4, starting with 4-bromotrifluo-
romethoxybenzene. Addition of a small amount of iodine was
necessary to initiate the Grignard reaction. Fractional distil-
lation at reduced pressure yielded a clear, slightly yellow,
viscous liquid. Yield: 12%. 1H NMR (CDCl3): δ 7.31-7.27 (2H,
m), 7.20 (2H, d, J ) 8 Hz) ppm. 31P{1H} NMR (CDCl3): δ -8.9
(s) ppm.
Syn th esis of 8. Using a modification of the procedure by
Ishihara et al.,63 tetrakis(triphenylphosphine)palladium (2.35
g, 2.03 mmol) was added to a solution of allylbenzene (6.76 g,
57.2 mmol) and n-perfluorobutyl iodide (19.7 g, 56.9 mmol) in
dry hexanes (40 mL) at 0 °C. The solution was slowly heated
to 25 °C and maintained at this temperature for 72 h, after
which the mixture was filtered and concentrated, yielding a
light-sensitive orange/brown oil. Distillation yielded 17.9 g
(68%) of 2-iodo-1H,1H,2H,3H,3H-perfluroheptylbenzene (8a ).
1H NMR (400 MHz, CDCl3): δ 4.5 (1H, m, C4F9CH2CHI-
CH2C6H5) ppm.
F igu r e 4. Correlation between rate of 1-octene hydro-
formylation in scCO2 and ν(CO) for HRh(CO)L3.
mine, and Brockmann alumina were purchased from Aldrich
and used as received. Magnesium turnings were obtained from
Acros Chemicals and used as received.
Compound 8a (17.9 g, 38.5 mmol) was dissolved in diethyl
ether in an addition funnel and added dropwise through a
condenser into a round-bottomed flask containing LiAlH4 (2.85
g, 75.2 mmol) dissolved in diethyl ether (20 mL), at 0 °C. The
solution was slowly heated to 25 °C and stirred for 2 h, after
which distilled water (3 mL), 1 M NaOH (3 mL), and distilled
water (6 mL) were added. The two-phase solution was filtered,
and the ether phase was removed and dried over MgSO4.
Reduction under vacuum gave a crude orange oil, which was
vacuum distilled to yield 1H,1H,2H,2H,3H,3H-perflurohep-
Hyd r ofor m yla tion Exp er im en ts. Hydroformylation reac-
tions were conducted batchwise in a 54 cm3 custom-built, high-
pressure stainless steel reactor fitted with sapphire windows.
The experimental setup is described in detail elsewhere.57
1-Octene was freshly distilled from sodium under nitrogen
before each experiment. For each hydroformylation experi-
ment, Rh(CO)2(acac) (6.5 × 10-5 mol, vacuum sealed in a glass
ampule), phosphine (2.0 × 10-5 mol), and 1-octene (0.054 mol)
were charged to the reactor under nitrogen. The reactor was
then sealed, heated to reaction temperature (50 °C), and
charged with H2/CO (0.055 mol each). Finally, CO2 was
charged to the reactor, bringing the total pressure to 273 atm.
The presence of a single fluid phase was confirmed visually,
with the catalyst being completely dissolved in the reaction
mixture. Periodic samples taken during each reaction were
analyzed by gas chromatography.
When working with high-pressure equipment, appropriate
safety devices should be used, including but not limited to
pressure relief mechanisms and explosion barriers.
P h osp h in e Syn th esis. Standard Grignard procedures
followed by phosphination were utilized to produce phosphines
4 and 6 from the corresponding commercially available fluo-
roalkyl- or fluoroalkoxy-substituted bromobenzene. Lithiation
was employed in the synthesis of phosphine 5 from the
commercially available starting material.
1
tylbenzene (8b) (5.6 g, 30%) as a clear oil. H NMR (CDCl3):
δ 7.4-7.1 (5H, m), 2.7 (2H, t, J ) 8 Hz), 2.1-1.8 (4H, m) ppm.
Anal. Calcd for C13H11F9: C, 46.17; H, 3.28. Found: C, 46.06;
H, 3.39.
Compound 8b (3.17 g, 9.4 mmol) dispersed on Brockmann
alumina (10.2 g) and bromine (1.9 g, 11.9 mmol) dispersed on
Brockmann alumina (10.5 g) were combined and shaken for 1
h. The product was extracted with diethyl ether and fraction-
ally distilled at reduced pressure to yield 4-bromo-
(1H,1H,2H,2H,3H,3H-perfluoroheptyl)benzene (8c) (3.41 g,
1
87%). H NMR (CDCl3): δ 7.4 (2H, d, J ) 6 Hz), 7.0 (2H, d, J
) 8 Hz), 2.67 (2H, t, J ) 8 Hz), 2.1-1.8 (4H, m) ppm. Anal.
Calcd for C13H10F9Br: C, 37.43; H, 2.42. Found: C, 38.33; H,
2.69.
The phosphine, 8, was produced from 8c (3.0 g, 7.2 mmol)
and PCl3 (1.1 g, 8.0 mmol) using standard lithiation tech-
niques. Yield: 350 mg, 14%. 31P NMR (400 MHz, CDCl3): δ
-8.1 (s) ppm. Calcd for C39H30F27P: C, 44.93; H, 2.90. Found:
C, 44.76; H, 3.08.
Syn th esis of 4. Phosphine 4 was synthesized from 3-bro-
mobenzotrifluoride and PCl3 according to the procedure given
by Miller et al. for producing the para-substituted phosphine.62
Fractional distillation at reduced pressure yielded 4 as a
colorless, viscous liquid. Yield: 47%. 1H NMR (CDCl3): δ 7.65
(1H, d, J ) 8 Hz), 7.59 (1H, d, J ) 8 Hz), 7.50 (1H, t, J ) 8
(63) Ishihara, T.; Kuroboshi, M.; Okada, Y. Chem. Lett. 1986, 1895.
(64) Wu, G.; Washliyen, R. E.; Curtis, R. D. Can. J . Chem. 1992,
70, 863.
(62) Miller, G. R.; Yankowsky, A. W.; Grim, S. O. J . Chem. Phys.
1969, 51, 3185.
(65) Evans, D.; Yagusky, G.; Wilkinson, G. J . J . Chem. Soc., A 1968,
2660.

86 Organometallics, Vol. 19, No. 1, 2000
Palo and Erkey
Syn th esis of HRh (CO)L3. To determine FTIR and NMR
values for the various catalysts in this study, each catalyst
was synthesized in small quantities in pure scCO2 in the high-
pressure reactor described above for the hydroformylation
experiments. For each synthesis, Rh(CO)2(acac) (∼20 mg)
along with the respective phosphine (3 equiv) and a magnetic
stir bar were charged to the reactor, which was then sealed
and purged with a stream of 1:1 H2/CO for several minutes to
remove oxygen. The reactor was then heated to 35-40 °C,
charged with H2/CO (2 equiv), pressurized with CO2 (205 atm),
and stirred for ∼15 min. It was confirmed visually that the
phosphine and Rh(CO)2(acac) were completely dissolved (no
solid in reactor, and yellow scCO2 solution). Cooling and
venting yielded a crude solid product containing the respective
catalyst species, HRh(CO)L3, which were analyzed by 1H NMR,
31P NMR, and FTIR spectroscopy (see Table 2).
Ack n ow led gm en t is made to the donors of the
Petroleum Research Fund, administered by the ACS,
for partial support of this research (ACS-PRF 32299-
AC1). We would like to acknowledge the contribution
of J ennifer A. Hay in helping develop the synthetic route
for phosphine 8 and T. Davis in helping to synthesize 5
and the corresponding rhodium complex. We are also
indebted to Gary Lavigne for his assistance with FTIR
measurements and to Dr. Thomas Leipert for his
assistance with NMR spectroscopy.
OM990560A