phase are greatly reduced and vary little as the temperature is
increased, suggesting that the stability of the catalyst and ligand
is greatly increased in the absence of toluene. In the absence of
toluene the polar aldehyde phase separates as it forms, greatly
reducing the possibility for destructive reactions between the
phosphite and aldehyde. Increasing the concentration of the
catalyst and the ligand (experiment 9) increases the rate
somewhat, but also increases the l+b ratio, mainly again because
more isomerised alkene is produced at the expense of branched
aldehyde, although the best selectivity to linear aldehyde
(75.6%) using this ligand is obtained with the higher catalyst
and ligand loading at 60 °C (experiment 9). The leaching, in
terms of percentage lost to the organic phase is also lowest at
2.6% (20.6 mg Rh (mol of aldehyde)21) and 3.9% (28.6 mg P
(mol of aldehyde)21). We have noted similar high l+b ratios
when using triaryl-phosphites for reactions carried out in
supercritical fluids7 or ionic liquid–supercritical fluid biphasic
systems.8
Given the problems of catalyst–ligand instability experienced
with the fluorous derivatised phosphite, we investigated the use
of the analogous phosphine, P(4-C6H4C6F13)3,6 which gives
high rates for hydroformylation in scCO2,9 again using only the
fluorous solvent, although in this case perfluoromethylcyclo-
hexane (experiments 10–12). This ligand proved to be much
more stable than the phosphite and gave good rates and linear
selectivities with little evidence for decomposition up to 80 °C,
although increased Rh leaching occurred when reactions were
carried out at 90 °C (experiment 12). The best results were
obtained with a rhodium loading of 2 mmol dm23 and a P+Rh
ratio of 10+1 (experiment 10). Even at this ratio, the phosphine
concentration is only 20 mmol dm23, compared with the
loading of 152 mmol dm23 used by Horváth and co-workers to
obtain a similar l+b ratio with their phosphine,4 and a loading of
200–300 mmol dm23 used in commercial systems employing
PPh3.1 The results obtained under these conditions, presented
diagrammatically in Fig. 1, show that a linear selectivity over
80% can be obtained at a high rate and with excellent retention
of rhodium (99.95%, loss of 0.08 mg Rh (mol aldehyde)21) and
phosphorus (96.7%, loss of 15.9 mg (mol aldehyde)21) in the
fluorous phase. These figures are compared with results
obtained by Horváth and co-workers4 and with results from
publications and patents that are closely related to the
commercial systems1,10,11 in Fig. 2. This comparison shows that
at similar rhodium concentrations and pressures to those used
commercially, but at much lower temperatures and phosphine
loadings (important because of the cost of the fluorous
phosphine), experiment 10 gives much higher rates (expected
because of the lower phosphine loading) and comparable
selectivities (l+b and linear aldehyde), especially when it is
remembered that the commercial system uses propene, which
cannot isomerise and in experiment 10, 3.7% of the selectivity
is lost to alkene isomerisation. It seems that the major advantage
in our system over the commercial systems is that high linear
selectivities can be obtained at much lower phosphine loading
and hence high rates do not need to be compromised.
Fig. 2 Comparison between results obtained in this work (experiment 10)
for hydroformylation of oct-1-ene with those reported in ref. 4 (dec-1-ene)
and those from patents and publications that are closely related to the
commercial processes for propene run by Union Carbide (UCC)1,2,10 and
BASF.1,2,11
they acknowledged that it was significant; we have estimated it
at 3% per reaction (i.e. comparable to the losses we observe)
based on the change in l+b ratio and rate that they observed.
We conclude that triaryl-phosphines or phosphites bearing
fluorous ponytails can give superior performance in terms of
reaction rate, linear selectivity, and retention into the fluorous
phase at lower ligand loadings than trialkylphosphine ana-
logues. Extra advantages in terms of all of these parameters are
obtained by working purely in the fluorous phase, which offers
the added advantage that the alkene is completely miscible with
the solvent, whilst the aldehyde phase separates. In addition,
there are considerable advantages to the processing of the
product since it does not have to be separated from an organic
solvent such as toluene.
We thank the Royal Society (EGH and AMS) and the EPSRC
for a Postdoctoral Fellowship (DJA) and for funding to use the
Catalyst Evaluation and Optimisation Service (CATS). We are
indebted to Ed McCurdy, Agilent Technologies UK Ltd. for
carrying out the ICPMS analyses.
Notes and references
‡ Low ligand concentrations are essential for the successful use of
expensive fluorinated ligands.
§ Since the separation is always carried out at room temperature, a stable
catalyst at a given concentration of ligand and metal should give the same
amount of leaching regardless of the reaction temperature.
1 C. D. Frohling and C. W. Kohlpaintner, in Applied Homogeneous
Catalysis with Organometallic Compounds, ed. B. Cornils and W. A.
Herrmann, VCH, Weinheim, 1996.
2 P. W. N. M. Van Leeuwen and C. Claver, Rhodium catalysed
hydroformylation, Kluwer, Dordrecht, 2000.
3 I. T. Horváth and J. Rabái, Science, 1994, 266, 72.
4 I. T. Horváth, G. Kiss, R. A. Cook, J. E. Bond, P. A. Stevens, J. Rabái
and E. J. Mozeleski, J. Am. Chem. Soc., 1998, 120, 3133.
5 B. Cornils, in Applied Homogenous Catalysis with Organometallic
Compounds, ed. B. Cornils and W. A. Herrmann,VCH, Weinheim,
1996.
Horváth and co-workers used a lower catalyst loading but
higher [phosphine], so it is not surprising that their rate was
lower. Their selectivity was also lower but the leaching of
rhodium per mol of aldehyde produced is similar in the two
systems. Although, they did not measure phosphorus leaching
6 P. Bhattacharyya, D. Gudmunsen, E. G. Hope, R. D. W. Kemmitt, D. R.
Paige and A. M. Stuart, J. Chem. Soc., Perkin Trans. 1, 1997, 3609.
7 M. F. Sellin and D. J. Cole-Hamilton, J. Chem. Soc., Dalton Trans.,
2000, 11, 1681.
8 M. F. Sellin, P. B. Webb and D. J. Cole-Hamilton, Chem. Commun.,
2001, 781.
9 A. M. B. Osuna, W. P. Chen, E. G. Hope, R. D. W. Kemmitt, D. R.
Paige, A. M. Stuart, J. L. Xiao and L. J. Xu, J. Chem. Soc., Dalton
Trans., 2000, 4052.
10 E. A. V. Brewester and R. L. Pruett, US Pat., 1970, 4,247,486, Chem.
Abstr., 1976, 84, 24271.
11 P. Zehner, H. Hoffmann, W. Richter, D. Stuetzer, M. Strohmeyer, W.
Helmut and E. Weippert, Eur. Pat., 1987, 254,180.
Fig. 1 Diagrammatic representation of experiment 10; the catalytic
hydroformylation of oct-1-ene in perfluorocyclohexane catalysed by a
rhodium complex of P(4-C6H4C6F13)3.
CHEM. COMMUN., 2002, 722–723
723