Activity of SPANphos rhodium dimers in methanol carbonylation

Article abstract of DOI:10.1002/anie.200500226

(Chemical Equation Presented) Contrary to earlier speculation, mononuclear rhodium species that contain trans diphosphines, such as SPANphos, are inactive towards the oxidative addition of methyl iodide, whereas the corresponding dinuclear species are ver

Full text of DOI:10.1002/anie.200500226

Angewandte
Chemie
Homogeneous Catalysis
Activity of SPANphos Rhodium Dimers in
Methanol Carbonylation**
Zoraida Freixa,* Paul C. J. Kamer, Martin Lutz,
Anthony L. Spek, and Piet W. N. M. van Leeuwen
Oxidative addition to organometallic compounds is one of the
key elementary steps in many catalytic processes. For this
reason, the reaction has been extensively studied, and in the
[*] Dr. Z. Freixa
Institute of Chemical Research of Catalonia (ICIQ)
Avda. Països Catalans s/n., Tarragona (Spain)
Fax: (+34)97792-0221
E-mail: zfreixa@iciq.es
Dr. P. C. J. Kamer, Prof. P. W. N. M. van Leeuwen
Van’t Hoff Institute for Molecular Sciences
University of Amsterdam
Nieuwe Achtergracht 166, Amsterdam (The Netherlands)
Fax: (+31)20525-5406
Dr. M. Lutz, Prof. Dr. A. L. Spek
Bijvoet Center for Biomolecular Research
Department of Crystal and Structural Chemistry
Utrecht University
Padualaan 8, 3584 CH Utrecht (The Netherlands)
[**] Dr. Gisela Colet and Dr. Lluís Solà from the Research Support Area
of ICIQ are kindly acknowledged for the help on the use of the React
IR 4000 spectrometer.
Angew. Chem. Int. Ed. 2005, 44, 4385 –4388
DOI: 10.1002/anie.200500226
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4385

Communications
last few years many publications have appeared concerning its
implication to ligand-promoted methanol carbonylation.[1] It
is commonly accepted that the rate-limiting step of this
reaction is precisely the oxidative addition of MeIto a square
planar d8 complex. This reaction is thought to proceed by an
SN2 mechanism. Accordingly, many groups have been work-
ing on the acceleration of the overall catalytic reaction by
using ligands that increase the electronic density on the metal
center as well as stabilizing the ligand through chelation to
make it resistant to the harsh conditions of the process.[2]
Recent breakthroughs on this topic concern the use of
benzoic acid-based flexible trans-chelating diphosphines and
the use of bis(imino)carbazolide ligands recently reported by
Haynes and co-workers, which, to our knowledge, led to the
fastest systems published until now.[3,4]
Herein we report on SPANphos mononuclear rhodium
complexes, which, even though they resemble the commonly
accepted catalyst precursor for methanol carbonylation with
phosphine ligands, do not undergo oxidative addition with
methyl iodide. As an alternative, we present unprecedented
reactivity of SPANphos dinuclear complexes and its implica-
tion to the reaction mechanism of the carbonylation of
methanol.
Recently, our group reported a trans-spanning diphos-
phine ligand (SPANphos, Scheme 1)[5] for which crystal
structures of its mono- and dinuclear rhodium chloro carbonyl
Figure 1. Upper: Displacement ellipsoid plot (50% probability level) of
[RhCl(SPAN-PPh2)(CO)] (1). The Cl and CO groups are substitutionally
disordered; only the major configuration is shown. Hydrogen atoms
are omitted for clarity. Lower: Displacement ellipsoid plot (50% proba-
bility level) of [Rh2(m-Cl)2(SPAN-PPh2)(CO)2] (2). Hydrogen atoms and
disordered CH2Cl2 solvent molecules have been omitted for clarity.
Scheme 1. The SPAN-diphosphine family. Cy=cyclohexyl
complexes have been characterized by X-ray diffraction
(Figure 1).[6] These systems are closely related to the those
published by Süss-Fink and co-workers and the dinuclear
SPANphos rhodium compound resembles the bis-acyl com-
plexes isolated at the end of the reaction by that group.[3]
These binuclear complexes, which are formed by the elimi-
nation of phosphine, were considered to be inactive but act as
a reservoir to the mononuclear, putative catalyst.[3]
Inspired by these results, we considered using complexes
of SPANphos ligands, which is a truly trans-chelating diphos-
phine family that cannot cis coordinate, as a catalyst for
methanol carbonylation. The results obtained showed that
SPAN-diphosphines are very active catalyst for this reaction
(Table 1). The TOF obtained compare well with those
reported by Süss-Fink and co-workers for other trans-chelat-
ing diphosphine ligands under the same catalytic conditions.[3]
In their report, the catalytic cycle is proposed to proceed
through a mononuclear species that contains diphosphine
ligands in a trans disposition.[3]
Table 1: Methanol carbonylation with Rh catalyst and SPAN-diphosphine
ligands.[a]
Entry
Ligand[b]
Conv. [%]
t [h]
TON[c]
TOF[d]
1[e]
2[e]
no ligand
SPAN-PPh2
SPAN-PCy2
SPAN-Bz
SPAN-POP
no ligand
SPAN-PPh2
SPAN-PCy2
SPAN-Bz
20
93
86
84
0.25
1
190
902
831
817
766
0
1093
880
880
725
2764[g]
762
902
1247
1225
2299
0
728
586
586
3[e]
0.67
0.67
0.3
1.5
1.5
1.5
1.5
1.5
1.5
4[e]
5[e]
79
6[f]
no conv.
11.3
9.1
9.1
7.5
7[f]
8[f]
9[f]
10[f]
11[f,g]
SPAN-POP
SPAN-POP
483
14.3
1842[g]
[a] Conditions: T=1508C, P(at 228C)=22 bar of CO, 700 rpm, [{Rh(m-
Cl)(CO)2}2] 28 mmol, ligand/Rh=1. [b] PPh2, P Cy, Bz, and POP refer to
2
the substitution of the phosphorous atom (see Scheme 1). [c] TON=
turnover number (in mol conversion per mol Rh). [d] TOF=turnover
frequency (in mol conversion per mol Rh per h). [e] MeOH/MeI/H2O/
Rh=967:100:682:1. [f] MeOH/MeI/H2O/Rh=9670:1000:6820:1. [g] Di
meric [Rh2(m-Cl)2(SPAN-POP)(CO)2], ligand/Rh=1:2, TON in mol
conversion per mol Rh2, TOF in mol conversion per mol Rh2 per h.
When a MeOH/rhodium ratio of about 10000:1 was used
(Table 1 entries 6–11), high activities were still observed.
These results can be considered as an indication that the
4386
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2005, 44, 4385 –4388

Angewandte
Chemie
observed activity cannot be just attributed to a medium effect
caused by phosphine quaternization taking place during the
reaction.
Surprisingly, the study of the rate of the oxidative addition
of MeIto the mononuclear [ClRh(SPAN-PPh 2)(CO)] (1)
rendered no rhodium(iii) methyl or acetyl compounds, not
even after 24 h in pure MeI. Only [RhI(SPAN-PPh2)(CO)],
which arises from halide metathesis, was isolated at the end of
the reaction (Scheme 2, reaction a). A previously reported
Figure 2. IR decay curves/surface for oxidative addition of MeI to 2 at
283 K. [MeI]=1.6m (1 spectrum per minute).
co-workers for dinuclear Rh complexes with diamido bridging
ligands.[9] Simultaneous NMR spectroscopy experiments
performed on reaction mixtures and GC–MS analysis of
reaction products are consistent with this hypothesis.
Kinetic studies by using in situ FTIR have been performed
on the reaction of MeIwith the dinuclear compound 2, which
showed that the reaction is first order with respect to both
MeIand 2. From the pseudo first-order constants obtained
over a range of MeIconcentrations and temperatures, the
Scheme 2. a) Reaction of 1 with MeI. b) Reaction of 2 with MeI and
the proposed structure for the resulting acyl compound; configuration
at the Rh center is unclear.
activation
parameters
(DH° = 32.15 kJmolÀ1,
DS° =
À163.73 JmolÀ1 KÀ1) were calculated. These values are in
concordance with those previously reported for this reaction
with mononuclear complexes.[1c,4,8a,8b]
example of halide exchange instead of oxidative addition was
found for a cis coordinated mixed P,S-bidentate ferrocenyl
ligand.[7] A crystal structure examination of the corresponding
[RhCl(CO)(L)] complex showed that one face of the square
planar complex is blocked by the ferrocenyl moiety of the
ligand backbone, thus inhibiting MeIaddition. [7] Both of these
cases can be considered as an extreme situation of the
extensively studied steric effects on the rate of MeIoxidative
addition.[1c,8]
We conclude that trans-spanning diphosphine ligands
form mononuclear rhodium compounds that are inactive for
the oxidative addition of methyl iodide. The lack of reactivity
as established for SPAN-diphosphines should be attributed to
one face of the square-planar compound being completely
blocked by the backbone of the ligand. Consequently, the
activity observed with mononuclear systems as precursors for
MeOH carbonylation is most likely to arise from dinuclear
species formed by partial dissociation of the ligand. A
complete catalytic cycle that involves dinuclear species
should be considered. These dimetal compounds showed an
unprecedented activity for MeIoxidative addition, and also
for MeOH carbonylation, and constitute one of the fastest
systems reported until now.
When dimeric [Rh2(m-Cl)2(SPAN-POP)(CO)2] was tested
as a catalyst for methanol carbonylation (Table 1, entry 11)
the results obtained showed that this catalyst is nearly four
times more active than the corresponding mononuclear
compound (entry 10). These results, together with the fact
that when using other trans-spanning diphosphines dinuclear
structures were isolated at the end of the reaction,[3] brought
us to consider a different approach: the activity observed in
the case of trans-spanning diphosphine ligands should per-
haps be attributed to a dinuclear species rather than to the
mononuclear species, [RhX(trans-P-P)(CO)] (X = Cl, I). To
prove this hypothesis, we studied the rate of MeIoxidative
addition to the dinuclear compound 2 (Scheme 2, reaction b).
The results obtained showed that the oxidative addition to
this dimetal species is only ten times slower compared to the
most active catalyst reported until now (Figure 2),[4] but it
constitutes the most active phosphine-based catalyst. In the
IR spectrum of the reaction, the fact that the band at
1981 cmÀ1 decays but does not disappear at the end of the
reaction suggests that the migration, and probably also the
oxidative addition, occurs only at one of the two rhodium
centers. The addition of MeIat only one of the two metallic
centers of a rhodium dimer to render a RhI–RhIII monoacyl
species is not unprecedented. Similar reactivity, at much
lower reaction rates, has already been reported by Oro and
Experimental Section
SPAN-diphosphines were synthesized according to
a modified
reported procedure published for the diphenyl phosphine deriva-
tive.[5]
Compounds 1 and 2 were synthesized in Schlenk tubes by mixing
the appropriate quantities of SPAN-PPh2 and [Rh2(m-Cl)2(CO)4],
dissolving the mixtures in degassed CH2Cl2, passing the resulting
solutions through celite, and removing the solvent by evaporation.
Yields were quantitative. Suitable crystals for X-ray diffraction were
grown by slow diffusion of pentane into CH2Cl2 solutions of the
rhodium complexes.
1: 1H NMR (400 MHz, 298 K, CD2Cl2, TMS): d = 1.22 (s, 3H,
CH3), 1.24 (s, 3H, CH3), 1.31 (s, 6H, CH3), 1.77 (d, J = 13.8 Hz, 1H,
CH2), 1.86 (d, J = 13.8 Hz, 1H, CH2), 1.96 (s, 3H, CH3), 2.00 (s, 3H,
CH3), 2.45 (d, J = 13.8 Hz, 1H, CH2), 2.49 (d, J = 13.8 Hz, 1H, CH2),
6.23 (dd, J = 9.6, J = 1.6 Hz, 1H, CHarom), 6.40 (dd, J = 9.6, J = 1.6 Hz,
1H, CHarom), 7.01 (d, J = 1.6 Hz, 1H, CHarom), 7.02 (d, J = 1.6 Hz, 1H,
CHarom), 7.15–7.45 (m, 16H, CHarom), 7.72 (dd, J = 9.6, J = 8.4 Hz, 1H,
CHarom), 7.84 ppm (dd, J = 9.6, J = 8.4 Hz, 1H, CHarom); 13C NMR
Angew. Chem. Int. Ed. 2005, 44, 4385 –4388
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4387

Communications
(100 MHz, 298 K, TMS): d = 21.4 (CH3), 21.5 (CH3), 29.3 (CH3), 29.5
[4] J. A. Gaunt, V. C. Gibson, A. Haynes, S. K. Spitzmesser, A. J. P.
White, D. J. Williams, Organometallics 2004, 23, 1015 – 1023.
[5] Z. Freixa, M. S. Beentjes, G. D. Batema, C. B. Dieleman, G. P. F.
van Strijdonck, J. N. H. Reek, P. C. J. Kamer, J. Fraanje, K.
Goubitz, P. W. N. M. van Leeuwen, Angew. Chem. 2003, 115,
1322 – 1325; Angew. Chem. Int. Ed. 2003, 42, 1284 – 1287.
[6] a) Crystal structure determinations: X-ray intensities were mea-
sured on a Nonius KappaCCD diffractometer with a rotating
anode (graphite monochromator, l = 0.71073 Š) at 150 K. The
structures were solved with direct methods (SHELXS-97)[6b] and
refined with SHELXL-97[6b] against F2 of all reflections. Structure
calculations and checking were performed with the PLATON
program.[6c] Compound 1: C48H46ClO3P2Rh, Mr = 871.15, yellow
needles, 0.18 ” 0.12 ” 0.06 mm3, orthorhombic, Pna21 (no. 33), a =
17.271(2), b = 13.1760(12), c = 17.5576(18) Š, V= 3995.4(7) Š3.
Z = 4, 1calcd = 1.448 gcmÀ3. A total of 50605 reflections were
measured of which 7749 were unique. The crystal appeared to be
cracked into two fragments with a rotation of 48 approximately
about the a axis. The structure was thus refined as a general twin
of two domains with a twin fraction of 0.284(2). The CO and Cl
ligands were substitutionally disordered to each other in a ratio of
0.62:0.38, respectively; 524 parameters, 4 restraints; R values: R1
(I > 2s(I)): 0.0566, wR2 (all refl.) = 0.1096; Flack x parameter
À0.06(4); S = 1.080. Compound 2: C49H46Cl2O4P2Rh2·2CH2Cl2,
Mr = 1207.37, orange plates, 0.35 ” 0.33 ” 0.12 mm3, monoclinic,
P21/c (no. 14), a = 19.7446(1), b = 14.4777(1), c = 19.2880(1) Š,
b = 110.1247(2)8, V= 5176.97(5) Š3. Z = 4, 1calcd = 1.549 gcmÀ3. A
total of 79995 reflections were measured of which 11842 were
unique. The CH2Cl2 solvent molecules were disordered;
620 parameters, 69 restraints; R values: R1 (I > 2s(I)): 0.0279,
wR2 (all refl.) = 0.0747; S = 1.043. CCDC-260236 (1) and -260237
(2) contain the supplementary crystallographic data for this
paper. These data can be obtained free of charge from the
Cambridge Crystallographic Data Centre via www.ccdc.cam.
ac.uk/data_request/cif; b) G. M. Sheldrick, SHELXS-97 and
SHELXL-97, programs for crystal structure solution and refine-
ment, University of Gꢀttingen, Germany, 1997; c) A. L. Spek, J.
Appl. Crystallogr. 2003, 36, 7 – 13.
(CH3), 30.6 (CH3), 30.8 (CH3), 47.9 (CH2), 48.1 (CH2), 86.3 (C), 86.5
(C), 117.2 (C), 127–138 ppm (Carom); 31P NMR (162 MHz, 298 K): d =
22.29 (dd, J = 365.4, J = 132.0 Hz), 26.50 ppm (dd, J = 365.4, J =
132.0 Hz); MS [MÀCl].
2: 1H NMR(400 MHz, 298 K, CD2Cl2, TMS): d = 0.73 (s, 3H,
CH3), 1.18 (s, 3H, CH3), 1.70 (d, J = 14.4 Hz, 1H, CH2), 2.00 (s, 3H,
CH3), 2.10 (d, J = 14.4 Hz, 1H, CH2), 6.28 (dd, J = 12.0, J = 1.6 Hz,
1H, CHarom), 7.05 (d, J = 1.5 Hz, 1H, CHarom), 7.15–7.35 (m, 8H,
CHarom), 7.73 ppm (dd, J = 11.6, J = 7.6 Hz, 1H, CHarom); 13C NMR
(100 MHz, 298 K, TMS): d = 21.1 (CH3), 29.5 (CH3), 31.0 (d, CH3, J =
20.6 Hz), 50.462 (CH2), 102.5 (C), 117.2 (C), 128–137 ppm (Carom);
31P NMR (162 MHz, 298 K): d = 45.88 ppm (d, J = 178.16 Hz); MS
[MÀCl], [MÀ2Cl].
Catalytic experiments: In a typical experiment 29.3 mL of a
mixture MeOH/H2O was placed in a 100-mL Hastelloy autoclave.
The autoclave was then pressurized to 10 bar of carbon monoxide and
heated to 1508C with vigorous stirring (700 rpm). When the reaction
temperature was reached and had stabilized,
a solution of
[Rh2Cl2(CO)4] and the appropriate quantity of the corresponding
ligand in MeIwere added through a liquid injection port and the
autoclave was pressurized to the desired reaction pressure. Pressure
was maintained constant during the reaction by feeding gas from a 1-
L autoclave pressurized with over 60 bar of carbon monoxide. After
the required reaction time, the autoclave was cooled to room
temperature. The solution was analyzed by NMR spectroscopy to
determine the selectivity. The conversion was calculated from gas
consumption of the reservoir autoclave.
Kinetic experiments: Kinetic measurements for the reaction of 1
or 2 with MeIin CH 2Cl2 were performed in situ by using a React IR
4000 from Mettler Toledo equipped with a diamond probe. A solution
that contained the corresponding rhodium compound in distilled
CH2Cl2 was placed in the sample vessel under an argon atmosphere.
The temperature was fixed by using a thermostat jacket. When the
desired temperature was achieved and remained constant, freshly
distilled MeIwas added by using a syringe. The RI spectra were
scanned in the region 400–2200 cmÀ1 and saved at regular time
intervals (usually 10 s) under computer control.
[7] R. Broussier, M. Laly, P. Perron, B. Gautheron, I. E. Nifantꢁev,
J. A. K. Howard, L. G. Kuzꢁmina, P. Kalck, J. Organomet. Chem.
1999, 587, 104 – 112.
Received: January 20, 2005
Published online: June 8, 2005
[8] a) L. Gonsalvi, J. A. Gaunt, H. Adams, A. Castro, G. J. Sunley, A.
Haynes, Organometallics 2003, 22, 1047 – 1054; b) L. Gonsalvi, H.
Adams, G. J. Sunley, E. Ditzel, A. Haynes, J. Am. Chem. Soc.
2002, 124, 13597 – 13612; c) L. Cavallo, M. Solꢂ, J. Am. Chem.
Soc. 2001, 123, 12294 – 12302; d) L. Gonsalvi, H. Adams, G. J.
Sunley, E. Ditzel, A. Haynes, J. Am. Chem. Soc. 1999, 121, 11233 –
11234.
Keywords: carbonylation · homogeneous catalysis ·
phosphine ligands · rhodium
.
[1] a) E. Daura-Oller, P. J. M. Poblet, C. Bo, Dalton Trans. 2003, 92 –
98; b) M. Sharma, N. Kumari, P. Das, P. Chutia, D. K. Dutta, J.
Mol. Catal. A 2002, 188, 25 – 35; c) H. C. Martin, N. H. James, J.
Aitken, J. A. Gaunt, H. Adams, A. Haynes, Organometallics 2003,
22, 4451 – 4458; d) V. Chauby, J.-C. Daran, C. S.-L. Berre, F.
Malbosc, P. Kalck, O. D. Gonzalez, C. E. Haslam, A. Haynes,
Inorg. Chem. 2002, 41, 3280 – 3290; e) M. Bassetti, A. Capone, L.
Mastrofrancesco, M. Salamote, Organometallics 2003, 22, 2535 –
2538.
[9] M. V. Jimꢃnez, E. Sola, M. A. Egea, A. Huet, A. C. Francisco, F. J.
Lahoz, L. Oro, Inorg. Chem. 2000, 39, 4868 – 4878.
[2] a) C.-A. Carraz, E. J. Ditzel, A. G. Orpen, D. D. Ellis, P. G.
Pringle, G. J. Sunley, Chem. Commun. 2000, 1277 – 1278; b) D. K.
Dutta, J. D. Woollins, A. M. Z. Slawin, D. Konwar, P. Das, M.
Sharma, P. Bhattacharyya, S. M. Aucota, Dalton Trans. 2003,
2674 – 2679; c) J. Rankin, A. D. Poole, A. C. Benyei, D. J. Cole-
Hamilton, Chem. Commun. 1997, 1835 – 1836; d) J. Rankin, A. C.
Benyei, A. D. Poole, D. J. Cole-Hamilton, J. Chem. Soc. Dalton
Trans. 1999, 3771 – 3782.
[3] a) C. M. Thomas, R. Mafua, B. Therrien, E. Rusanov, H. Stoeckli-
Evans, G. Süss-Fink, Chem. Eur. J. 2002, 8, 3343 – 3352; b) C. M.
Thomas, G. Süss-Fink, Coord. Chem. Rev. 2003, 243, 125 – 142.
4388
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2005, 44, 4385 –4388