Rhodium(I) complexes of unsymmetrical diphosphines: Efficient and stable methanol carbonylation catalysts

Article abstract of DOI:10.1039/b002802i

Rhodium complexes of unsymmetrical diphosphines of the type Ph₂PCH₂CH₂PAr₂ are catalysts for the carbonylation of methanol; several features of the catalysis are reminiscent of iridium carbonylation catalysts.

Full text of DOI:10.1039/b002802i

Rhodium( ) complexes of unsymmetrical diphosphines: efficient and stable
I
methanol carbonylation catalysts†
Charles-Antoine Carraz,a Evert J. Ditzel,b A. Guy Orpen,a Dianne D. Ellis,a Paul G. Pringle*a and
Glenn J. Sunleyb
a School of Chemistry, University of Bristol, Cantocks Close, Bristol, UK BS8 1TS.
E-mail: paul.pringle@bristol.ac.uk
b BP Chemicals, Hull Research and Technology Centre, Saltend, Hull, UK HU12 8DS
Received (in Cambridge) 7th April 2000, Accepted 17th May 2000
Rhodium complexes of unsymmetrical diphosphines of the
type Ph2PCH2CH2PAr2 are catalysts for the carbonylation
of methanol; several features of the catalysis are reminiscent
of iridium carbonylation catalysts.
lower than the commercial [RhI2(CO)2]2 catalyst. The follow-
ing observations suggest that the catalyst is a diphosphine–
Methanol carbonylation (eqn. (1)) is one of the most successful
industrial applications of homogeneous catalysis and is cur-
rently carried out on a scale of several million tonnes per
annum.1 The rhodium–iodide catalysed process gives acetic
acid in better than 99% selectivity and the mechanism has been
thoroughly reviewed by Maitlis et al.2 The new Cativa process
uses an iridium–iodide catalyst and offers many advantages
over the rhodium system.3 The conditions shown in eqn. (1)
limit the use of phosphine modified catalysts because of the
potential problem of quaternisation of the ligand.1 Thus, while
the rhodium complexes of P,O-, P,N- and P,S-donor ligands
have all been reported4–8 to be methanol carbonylation
catalysts, either the conditions used were not the industrial
conditions of eqn. (1) and therefore the process would not be
viable,4–6 or the catalysts were found to be unstable.7,8 Here we
report that rhodium complexes of unsymmetrical ethylene
diphosphine ligands are more efficient catalysts than their
symmetrical dppe analogues for methanol carbonylation and
longer-lived than any previously reported ligand-modified
catalysts under the harsh conditions of eqn. (1).9
(3)
rhodium complex throughout and not [RhI2(CO)2]2. IR spectra
obtained in situ during catalysis with ligand 1e showed the
absence of the intense n(CO) bands for [RhI2(CO)2]2 at 2059
and 1988 cm21. At the end of the catalysis, when the solution
cooled, a homogeneous orange solution or red crystalline
precipitate was present and 31P NMR and IR spectra showed the
presence of a mixture of diphosphine rhodium(III) carbonyl
complexes (1J(RhP) ca. 100 Hz, n(CO) ca. 2090 cm21). The
product fac-[RhI3(CO)(1b)] 7 was isolated from the reaction
mixture using the catalyst derived from 1b and the crystal
structure of its pentane solvate was determined (see Fig. 1). The
octahedral geometry at the rhodium(III) centre is somewhat
distorted (iodine atoms I(1) and I(2) lie 0.140 and 0.184 Å
respectively from the RhP2 plane). The carbonyl ligand lies cis
to the diphosphine phosphorus atoms, which show identical
Rh–P bond lengths, any difference in the values being masked
by the disorder.‡ The rate of catalysis is constant throughout a
run and, after all the methanol was consumed, a second aliquot
of methanol was injected and the rate was the same as in the first
run. This final observation not only confirms the integrity of the
(1)
The diphosphines 1a–f, some of which are known,10,11 were
made according to eqn. (2) and fully characterised. The
(2)
Table 1 Methanol carbonylation data†
catalysts were prepared by addition of diphosphines 1a–f or 2a–
d to [Rh2(m-Cl)2(CO)4] in methanol [eqn. (3)]. Under these
conditions the complexes 3a–f and 4a–f are formed in
approximately 1 1 ratio in each case; these isomeric mixtures
have been isolated and fully characterised. Similarly the iodo
analogues 5e and 6e have been prepared and characterised as a
1 1 mixture.
The carbonylation catalysis was carried out under the
conditions given in eqn. (1) and the results are presented in
Table 1.† In each case the conversion of methanol was greater
than 98% and the selectivity for acetic acid was greater than
99%; the rates for the diphos systems (entries 1–11) were all
Entry
Ligand
Ratea
CH3CH2CO2Hb
1
2
3
4
5
6
7
8
9
1a
1b
1c
1d
1e
1f
1b + Ruc
2a
2b
2c
2d
—
2.0
5.0
5.6
8.5
7.6
2.9
13.7
1.9
2.0
2.3
1.7
18.5
6
45
50
79
54
31
188
4
13
20
12
276
10
11
12d
a At 10% conversion, in mol l21 h21 with estimated errors of 5–10%.
b Concentration in ppm. c [RuI2(CO)4] (0.576 g, 1.23 mmol) added to the
catalyst mixture. d Catalyst is [RhI2(CO)2]2.
† Electronic supplementary information (ESI) available: typical experi-
mental procedure for catalysis. See http://www.rsc.org/suppdata/cc/b0/
b002802i/
DOI: 10.1039/b002802i
Chem. Commun., 2000, 1277–1278
This journal is © The Royal Society of Chemistry 2000
1277

catalyst but also shows its longevity is greater than any previous
rhodium–phosphine catalyst under these conditions7 since it
shows that all the diphosphine catalysts undergo over 500
turnovers without noticable diminution of activity.
In the following respects, the rhodium–diphosphine catalysts
resemble the iridium Cativa catalysts. The main inefficiencies
in traditional rhodium-catalysed methanol carbonylation are the
water gas shift reaction and the formation of by-products such
as MeCHO, EtI and CH3CH2CO2H; this problem is much
reduced with the iridium catalysts.3 The amount of propionic
acid reported in Table 1 for the diphosphine catalysts (entries
1–10) is significantly less than with [RhI2(CO)2]2 as catalyst
under these conditions (entry 12). 31P NMR studies in CH2Cl2
show that oxidative addition of MeI to 5e/6e is very rapid. The
greater nucleophilicity of [RhI(CO)(diphosphine)] complexes12
than [RhI2(CO)2]2 may partly explain the similarities between
the rhodium–diphosphine and the iridium catalysts.
Since the iridium catalysts are promoted by iodide-abstract-
ing ruthenium complexes,3 we investigated whether
[RuI2(CO)4] would also promote the rhodium catalyst from
diphosphine 1b; by comparing entries 7 and 2 in Table 1, it is
clear that the addition of the Ru complex has more than doubled
the rate.
From the data in Table 1, it can be deduced that the influence
of the phosphorus substituents is complicated. The rate data are
plotted in Fig. 2 as a function of the Hammett constants for the
aryl substituents. The plot shows that increasing the electron-
withdrawing power of the substituents on the aryl rings in the
unsymmetrical diphosphines generally increases the catalyst
activity up to a point, beyond which the rate decreases. The
significance of the maximum in the curve might be interpreted
in terms of a balance of s-donor and p-acceptor qualities being
required to optimise the rate. However entries 4 and 9 in Table
1 are with ligands 1d and 2b which would be expected to have
similar overall donor/acceptor properties by virtue of the same
number of meta-fluoro substituents and yet they show very
different catalytic performance. In fact, all of the symmetrical
diphosphines 2a–d yield catalysts of similarly low activity
(entries 8–11). Thus the asymmetry of the diphosphine is
apparently crucial. Casey et al.10 have shown that un-
symmetrical diphosphines are superior to the symmetrical
analogues for hydroformylation catalysis and associated this
with a preference of the better s-donor for the axial site in the
trigonal bipyramidal intermediates. It is notable that P,O-, P,N-
and P,S-donor ligands used previously4–8 for methanol carbo-
nylation are all unsymmetrical with one strong and one medium
or weak donor. For the best one (Ph2PCH2CH2P(S)Ph2), Baker
et al.7 showed that only one isomer (with the S-donor trans to
CO) is formed in the reaction of [Rh2I2(CO)4] with the ligand.
By contrast, we find no such diastereoselectivity in the reaction
of diphosphines 1a–f with [Rh2X2(CO)4] (X = Cl or I). In the
presence of CO, 31P NMR spectroscopy shows that the
diastereoisomers 3/4 and 5/6 interconvert rapidly (eqn. (3) and
thus the ca. 1 1 mixtures observed represent the thermody-
Fig. 2 Plot of the rate of methanol carbonylation (from Table 1) as a function
of the Hammett substituent constant, s for the Ar substituents in the ligands
Ph2PCH2CH2PAr2. The error bars represent a 7.5% error in the rate
measurement.
namic proportions. Hence there is little difference in the
stability of the [RhI(CO)(diphos)] precursors under ambient
conditions in CH2Cl2 but it is possible that under the radically
different conditions of the catalysis, one of the isomers is
preferred or one is significantly more reactive.
In conclusion we have established that unsymmetrical
diphosphine–rhodium complexes are very active and selective
catalysts for methanol carbonylation under industrially sig-
nificant conditions and these catalysts have several features in
common with the iridium Cativa catalysts.
We thank BP Chemicals and the EPSRC for financial
support.
Notes and references
‡ Crystal structure analysis of [Rh(CO)I3[Ph2P(C2H4)P(3-C6H4F)2]·
0.5C5H12, 7·0.5C5H12. Crystal data: C29.5H22F2I3OP2Rh, M = 976.02,
orthorhombic, space group Pbca (no. 61), a = 18.850(3), b = 15.338(4), c
= 20.992(4)Å, T = 173 K, U = 6069(6) Å3, Z = 8, m = 3.755 mm21, 6885
unique data, R1 = 0.038. The fluorine atoms are disordered occupying one
meta site on each of the four aryl rings equally as a consequence of the
enantiomers of 7 crystallising at the same site in the unit cell. CCDC
182/1645. See http://www.rsc.org/suppdata/cc/b0/b002802i/ for crystallo-
graphic files in .cif format.
1 (a) M. Gauss, A. Seidel, P. Torrence and P. Heymans, in Applied
Homogeneous Catalysis with Organometallic Compounds, ed. B.
Cornils and W. A. Herrmann, VCH, New York, 1996; (b) M. J. Howard,
M. D. Jones, M. S. Roberts and S. A. Taylor, Catal. Today, 1993, 18,
325.
2 P. M. Maitlis, A. Haynes, G. J. Sunley and M. J. Howard, J. Chem. Soc.,
Dalton Trans., 1996, 2187.
3 (a) M. J. Howard, G. J. Sunley, A. D. Poole, R. J. Watt and B. K.
Sharma, Stud. Surf. Sci. Catal., 1999, 121, 61; (b) T. Ghaffar, H. Adams,
P. M. Maitlis, G. J. Sunley, M. J. Baker and A. Haynes, Chem.
Commun., 1998, 1023; (c) see also ref 3 in J. Yang, A. Haynes and
P. M. Maitlis, Chem. Commun., 1999, 179.
4 R. W. Wegman, A. G. Abatjoglou and A. M. Harrison, J. Chem. Soc.,
Chem. Commun., 1987, 1891.
5 A. Bader and E. Lindner, Coord. Chem. Rev., 1991, 108, 27.
6 M. S. Balakrishna, R. Klein, S. Uhlenbrock, A. A. Pinkerton and R. G.
Cavell, Inorg. Chem., 1993, 32, 5676.
7 M. J. Baker, M. G. Giles, A. G. Orpen, J. Taylor and R. J. Watt, J. Chem.
Soc., Chem. Commun., 1995, 197.
8 J. R. Dilworth, J. R. Miller, N. Wheatley, M. J. Baker and G. Sunley,
J. Chem. Soc., Chem. Commun., 1995, 1579.
9 M. J. Baker, E. Ditzel, G. Sunley, C. A. Carraz and P. G. Pringle, Br.
Pat., 1999, 9907447.8.
10 C. P. Casey, E. L. Paulsen, E. W. Beuttenmueller, B. R. Proft, B. A.
Matter and D. R. Powell, J. Am. Chem. Soc., 1999, 121, 63 and
references therein.
Fig. 1 The molecular structure of 7 showing one of the two orientations of
the meta-C6H4F groups. Important molecular dimensions include: bond
lengths (Å) Rh(1)–C(3) 1.885(6), Rh(1)–P(1) 2.335(2), Rh(1)–P(2)
2.3370(15), Rh(1)–I(1) 2.7337(7), Rh(1)–I(2) 2.7296(7), Rh(1)–I(3)
2.6869(6); bond angle (°) P(1)–Rh(1)–P(2) 86.03(6).
11 H. Brunner and A. Stumpf, J. Organomet. Chem., 1993, 459, 139; P. N.
Kapoor, D. D. Pathak, G. Gaur and M. Kutty, J. Organomet. Chem.,
1984, 276, 167.
11 L. Gonsalvi, H. Adams, G. J. Sunley, E. Ditzel and A. Haynes, J. Am.
Chem. Soc., 1999, 121, 11 233.
1278
Chem. Commun., 2000, 1277–1278