Steric and electronic effects on the reactivity of Rh and Ir complexes containing P-S, P-P, and P-O ligands. Implications for the effects of chelate ligands in catalysis

Published on Web 10/16/2002

Steric and Electronic Effects on the Reactivity of Rh and Ir

Complexes Containing P-S, P-P, and P-O Ligands.

Implications for the Effects of Chelate Ligands in Catalysis

Luca Gonsalvi,^†,‡Harry Adams,^‡Glenn J. Sunley,^§Evert Ditzel,^§and

Anthony Haynes*^,‡

Contribution from the Department of Chemistry, UniVersity of Sheffield, Sheffield S3 7HF, U.K.,

and Hull Research and Technology Centre, BP Chemicals Ltd., Saltend, Hull HU12 8DS, U.K.

Received November 27, 2001

Abstract: Kinetic studies of the reactions of [M(CO)(L-L)I] [M ) Rh, Ir; L-L ) Ph₂PCH₂P(S)Ph₂(dppms),

Ph₂PCH₂CH₂PPh₂(dppe), and Ph₂PCH₂P(O)Ph₂(dppmo)] with methyl iodide have been undertaken. All

the chelate ligands promote oxidative addition of methyl iodide to the square planar M(I) centers, by factors

of between 30 and 50 compared to the respective [M(CO)₂I₂]^-complexes, due to their good donor properties.

Migratory CO insertion in [Rh(CO)(L-L)I₂Me] leads to acetyl complexes [Rh(L-L)I₂(COMe)] for which X-ray

crystal structures were obtained for L-L ) dppms (3a) and dppe (3b). Against the expectations of simple

bonding arguments, methyl migration is faster by a factor of ca. 1500 for [Rh(CO)(dppms)I₂Me] (2a) than

for [Rh(CO)(dppe)I₂Me] (2b). For M ) Ir, alkyl iodide oxidative addition gives stable alkyl complexes [Ir-

(CO)(L-L)I₂R]. Migratory insertion (induced at high temperature by CO pressure) was faster for [Ir(CO)-

(dppms)I₂Me] (5a) than for its dppe analogue (5b). Reaction of methyl triflate with [Ir(CO)(dppms)I] (4a)

yielded the dimer [{Ir(CO)(dppms)(µ-I)Me}₂]²⁺(7), which was characterized crystallographically along with

5a and [Ir(CO)(dppms)I₂Et] (6). Analysis of the X-ray crystal structures showed that the dppms ligand adopts

a conformation which creates a sterically crowded pocket around the alkyl ligands of 5a, 6, and 7. It is

proposed that this steric strain can be relieved by migratory insertion, to give a five-coordinate acetyl product

in which the sterically crowded quadrants flank a vacant coordination site, exemplified by the crystal structure

of 3a. Conformational analysis indicates similarity between M(dppms) and M₂(µ-dppm) chelate structures,

which have less flexibility than M(dppe) systems and therefore generate greater steric strain with the “axial”

ligands in octahedral complexes. Ab initio calculations suggest an additional electronic contribution to the

migratory insertion barrier, whereby a sulfur atom trans to CO stabilizes the transition state compared to

systems with phosphorus trans to CO. The results represent a rare example of the quantification of ligand

effects on individual steps from catalytic cycles, and are discussed in the context of catalytic methanol

carbonylation. Implications for other catalytic reactions utilizing chelating diphosphines (e.g., CO/alkene

copolymerization and alkene hydroformylation) are considered.

Introduction

important, although the precise way in which the effect of the

bite angle is transmitted to the active site of the metal complex

Ligand steric and electronic effects play key roles in

determining organometallic reactivity trends and catalytic

properties. For monodentate ligands, a number of quantitative

parameters (e.g., Tolman’s cone angle,¹Giering’s QALE

method,²and Drago’s ECW model³) are well-established.

Several important homogeneous catalysts now utilize bidentate

ligands for which the stereoelectronic properties are less well

understood. The ligand bite angle^4,5has been shown to be

is still a matter of some debate, with electronic and steric effects

both contributing.⁶Attempts to quantify the steric requirements

of diphosphine ligands have recently been made by Koide et

al., who introduced the pocket angle concept,⁷and Angermund

et al., who used an accessible molecular surface model.⁸Despite

the large body of work in this area, the quantification of ligand

effects on individual steps from catalytic cycles is quite rare.

In this paper we report kinetic, crystallographic, and theoretical

data which illuminate the steric and electronic factors influencing

key steps in ligand-promoted catalytic methanol carbonylation.

* Address correspondence to this author. E-mail: a.haynes@sheffield.ac.uk.

^†Current address: Institute for the Chemistry of Organometallic

Compounds, National Research Council (ICCOM-CNR), Via J Nardi 39-

41, 50132 Firenze, Italy.

(5) Dierkes, P.; van Leeuwen, P. W. N. M. J. Chem. Soc., Dalton Trans. 1999,

1519. van Leeuwen, P. W. N. M.; Kamer, P. C. J.; Reek, J. N. H.; Dierkes,

P. Chem. ReV. 2000, 100, 2741.

(6) Casey, C. P.; Paulsen, E. L.; Beuttenmueller, E. W.; Proft, B. R.; Petrovich,

L. M.; Matter, B. A.; Powell, D. R. J. Am. Chem. Soc. 1997, 119, 11817.

(7) Koide, Y.; Bott, S. G.; Barron, A. R. Organometallics 1996, 15, 2213.

(8) Angermund, K.; Baumann, W.; Dinjus, E.; Fornika, R.; Go¨rls, H.; Kessler,

M.; Kru¨ger, C.; Leitner, W.; Lutz, F. Chem.sEur. J. 1997, 3, 755.

^‡University of Sheffield.

^§BP Chemicals Ltd.

(1) Tolman, C. A. Chem. ReV. 1977, 77, 313.

(2) Wilson, M. R.; Liu, H.; Prock, A.; Giering, W. P. Organometallics 1993,

12, 2044.

(3) Joerg, S.; Drago, R. S.; Sales, S. Organometallics 1998, 17, 589.

(4) Casey, C. P.; Whiteker, G. T. Isr. J. Chem. 1990, 30, 299.

9

J. AM. CHEM. SOC. 2002, 124, 13597-13612

13597

A R T I C L E S

Gonsalvi et al.

Scheme 1. Cycle for Methanol Carbonylation Catalyzed by 1a

promoted system, where despite the much lower CO pressure,

the sole observed species was [Rh(CO)₂(dppeo)I] containing a

monodentate P-coordinated dppeo ligand.¹²

Our investigation has been directed at quantifying the kinetics

of oxidative addition and migratory insertion in the (dppms)Rh

system, and comparing the rates with corresponding reactions

of Rh complexes containing dppmo and Ph₂P(CH₂)₂PPh₂(dppe),

each of which also forms five-membered chelate rings. Bidentate

phosphines Ph₂P(CH₂)_nPPh₂(n ) 1-4) are known to promote

the reductive carbonylation (using a CO/H₂mixture) of methanol

to acetaldehyde and derivatives.¹⁴A range of symmetrical and

unsymmetrical derivatives of dppe, containing fluorinated

substituents on the phenyl rings, have recently been tested for

acetic acid production, but all were found to be less active than

the conventional [Rh(CO)₂I₂]^--catalyzed system.¹⁵

In addition to our kinetic studies on the rhodium systems,

we have made comparative measurements on some analogous

iridium complexes, where oxidative addition generally gives

stable alkyl complexes. A crystallographic study of [Ir(CO)-

(dppms)I₂Me] (5a), an iridium model for the reactive intermedi-

ate 2a, provides an insight into the steric effects on reactivity

in these systems. Other X-ray crystal structures are presented

which support the importance of steric effects in accelerating

migratory insertion, while ab initio quantum mechanical calcula-

tions suggest that electronic effects also contribute, particularly

for complexes with S trans to CO. A preliminary account of

some of the experimental results has been published.¹⁶

Although our study is primarily directed at complexes and

reactions involved in catalytic methanol carbonylation, our

observations are more generally applicable to ligand effects on

organometallic and catalytic reactivity. This is exemplified by

consideration of ligand effects in catalytic CO/alkene copolym-

erization and hydroformylation reactions.

The findings are related to the behavior of other catalytic

systems which use bidentate phosphine ligands.

Our studies were inspired by the findings of Baker et al.,^9,10

who reported an 8-fold enhancement in the rate of rhodium/

iodide-catalyzed methanol carbonylation (at 185 °C and 70 bar

of CO) on addition of the mixed P,S donor ligand Ph₂PCH₂P-

(S)Ph₂(dppms). Mixed P,S¹¹and P,O¹²ligands have been

employed previously for the Rh/iodide-catalyzed carbonylation

of methanol to acetic acid. Whereas Ph₂PCH₂CH₂P(O)Ph₂

(dppeo) was reported¹²to be an effective promoter for methanol

carbonylation under relatively mild conditions (80 °C, ∼3 bar

of CO), it gave only a marginal rate improvement under the

more commercially relevant conditions employed by Baker et

al. Likewise, Ph₂PCH₂P(O)Ph₂(dppmo) and Ph₂PN(Ph)P(S)-

Ph₂did not prove to be good promoters at 185 °C.

Results and Discussion

Preliminary mechanistic studies¹⁰identified the active species

in the dppms-promoted system as the Rh(I) chelate complex

[Rh(CO)(dppms)I] (1a), which was the only species observed

by in situ high-pressure IR spectroscopy under catalytic condi-

tions. The promotion is not therefore due to iodide salts formed

by quaternization of the dppms ligand. The proposed catalytic

cycle (Scheme 1) closely resembles the well-known Monsanto

system¹³based on the anionic complex [Rh(CO)₂I₂]^-. A first-

order dependence of catalytic rate on [MeI] was consistent with

rate-determining oxidative addition of MeI to 1a to give the

methyl intermediate [Rh(CO)(dppms)I₂Me] (2a). The stoichio-

metric reaction of 1a with MeI, in the absence of CO, yielded

the stable acetyl complex [Rh(dppms)I₂(COMe)] (3a), resulting

from facile migratory insertion in 2a. Coordination of CO and

reductive elimination of acetyl iodide completes the organo-

metallic cycle. All the observed catalytic intermediates retain

the Rh(dppms) chelate structure, in contrast with the dppeo-

Synthesis of Rh(I) and Ir(I) Complexes [M(CO)(L-L)I].

The square planar rhodium(I) complexes [Rh(CO)(dppms)I]

(1a), [Rh(CO)(dppe)I] (1b), and [Rh(CO)(dppmo)I] (1c) were

synthesized by reaction of the dimeric precursor [Rh(CO)₂I]₂

with the appropriate bidentate ligand (eq 1). Spectroscopic

¹/₂[Rh(CO)₂I]₂+ L-L f [Rh(CO)(L-L)I] + CO (1)

parameters for the dppe complex 1b were in accordance with

the literature data reported by Moloy and Wegman.¹⁴For the

dppms and dppmo complexes, there exists the possibility of two

geometrical isomers with CO coordinated either cis or trans to

the coordinated S or O heteroatom of the bidentate ligand. The

³¹P NMR data indicate the presence of a single isomer in each

case, and on the basis of similar spectroscopic data for the

known chloride analogues [Rh(CO)(dppms)Cl]¹⁰and [Rh(CO)-

(dppmo)Cl],¹²for which X-ray crystal structures have been

determined, we assign the stereochemistry with CO trans to S

or O. This geometry is reflected in the infrared spectra, which

exhibit ν(CO) bands at relatively low frequency for 1a (1987

cm^-1) and 1c (1983 cm^-1) compared to 1b (2011 cm^-1). The

(9) Baker, M. J.; Dilworth, J. R.; Sunley, J. G.; Wheatley, N. European Patent

Application 632,006, 1995.

(10) Baker, M. J.; Giles, M. F.; Orpen, A. G.; Taylor, M. J.; Watt, R. J. J.

Chem. Soc., Chem. Commun. 1995, 197.

(11) Cavell, R. G. PCT Patent Application WO 92/04118, 1992. Dilworth, J.

R.; Miller, J. R.; Wheatley, N.; Baker, M. J.; Sunley, J. G. J. Chem. Soc.,

Chem. Commun. 1995, 1579.

(12) Wegman, R. W.; Abatjoglou, A. G.; Harrison, A. M. J. Chem. Soc., Chem.

Commun. 1987, 1891.

(14) Moloy, K. G.; Wegman, R. W. Organometallics 1989, 8, 2883.

(15) Carraz, C. A.; Ditzel, E.; Orpen, A. G.; Ellis, D. D.; Pringle, P. G.; Sunley,

G. J. Chem. Commun. 2000, 1277.

(16) Gonsalvi, L.; Adams, H.; Sunley, G. J.; Ditzel, E.; Haynes, A. J. Am. Chem.

Soc. 1999, 121, 11233.

(13) Dekleva, T. W.; Forster, D. AdV. Catal. 1986, 34, 81. Forster, D. AdV.

Organomet. Chem. 1979, 17, 255. Maitlis, P. M.; Haynes, A.; Sunley, G.

J.; Howard, M. J. J. Chem. Soc., Dalton Trans. 1996, 2187.

9

13598 J. AM. CHEM. SOC. VOL. 124, NO. 45, 2002

Effects of Chelate Ligands in Catalysis

A R T I C L E S

Table 1. Selected Bond Lengths (Å) and Angles (deg) for

Rhodium Acetyl Complexes 3a and 3b

3a

3b

Rh-C(26)

Rh-S

1.951(6)

2.357(2)

2.256(2)

2.6497(19)

2.7024(19)

1.178(7)

1.491(9)

2.027(3)

Rh-C(27)

Rh-P(1)

Rh-P(2)

Rh-I(1)

2.013(7)

2.267(2)

2.284(2)

2.7035(12)

2.7029(13)

1.178(9)

Rh-P(2)

Rh-I(1)

Rh-I(2)

O-C(26)

C(26)-C(27)

P(1)-S

Rh-I(2)

O-C(27)

C(27)-C(28)

1.490(11)

C(26)-Rh-S

C(26)-Rh-P(2)

C(26)-Rh-I(1)

C(26)-Rh-I(2)

P(2)-Rh-S

93.62(19)

92.80(19)

92.23(18)

104.84(19)

90.47(9)

91.74(9)

173.64(4)

161.77(4)

83.96(8)

92.12(8)

120.1(6)

123.8(5)

116.1(5)

C(27)-Rh-P(1)

C(27)-Rh-P(2)

C(27)-Rh-I(1)

C(27)-Rh-I(2)

P(1)-Rh-P(2)

P(2)-Rh-I(1)

P(1)-Rh-I(1)

P(2)-Rh-I(2)

P(1)-Rh-I(2)

I(2)-Rh-I(1)

91.2(2)

92.5(2)

103.4(2)

99.1(2)

84.74(7)

90.73(6)

164.90(5)

167.28(5)

89.80(6)

91.62(4)

123.4(7)

124.5(6)

112.1(5)

P(2)-Rh-I(1)

S-Rh-I(1)

Figure 1. X-ray structure of [Rh(dppms)I₂(COMe)] (3a). Hydrogen atoms

on the dppms ligand are omitted for clarity. Selected geometric data are

given in Table 1.

P(2)-Rh-I(2)

S-Rh-I(2)

I(1)-Rh-I(2)

O-C(26)-C(27)

O-C(26)-Rh

C(27)-C(26)-Rh

O-C(27)-C(28)

O-C(27)-Rh

C(28)-C(27)-Rh

shift in ν(CO) (and the preference for the observed geometry)

is attributed to the ability of the coordinated S or O atoms to

act as π (as well as σ) donors in 1a and 1c as opposed to the

moderate π-acceptor PPh₂moiety trans to CO in 1b.

Scheme 2. Reaction of 1b with Methyl Iodide and Isomers of 2b¹⁴

The iridium(I) complexes [Ir(CO)(dppms)I] (4a) and [Ir(CO)-

(dppe)I] (4b) were synthesized in the manner of Fisher and

Eisenberg¹⁷by reaction of the anionic iridium precursor Bu₄N-

[Ir(CO)₂I₂] with the appropriate bidentate ligand (eq 2).

Bu₄N[Ir(CO)₂I₂] + L-L f

[Ir(CO)(L-L)I] + CO + Bu₄NI (2)

Spectroscopic data for 4b were in accordance with literature

values,¹⁷and data for the new dppms complex 4a were

consistent with the geometry displayed by the Rh analogue 1a

described above. The ν(CO) bands (1972 cm^-1for 4a and 1994

cm^-1for 4b) showed shifts of ca. 15 cm^-1to low frequency

compared to those of the Rh complexes, reflecting the enhanced

back-donating ability of the third-row metal. Attempts to

synthesize the complex [Ir(CO)(dppmo)I] by a method similar

to that used for 4a and 4b were unsuccessful.

Reactions of [M(CO)(L-L)I] with Methyl Iodide. The

reaction of 1a with excess methyl iodide leads smoothly to the

acetyl product 3a, which was fully characterized. When the

reaction was performed at high [MeI], a weak ν(CO) band at

2062 cm^-1was observed in the IR spectrum in addition to the

bands of 1a and 3a, consistent with the presence of a small

amount of the intermediate methyl complex 2a. The behavior

of this intermediate is dealt with in more detail when the kinetics

of this reactions are considered (vide infra). An X-ray crystal

structure of 3a (Figure 1) revealed a distorted square pyramidal

geometry with an apical acetyl group. The methyl group of the

acetyl ligand approaches an eclipsed conformation with respect

to one of the basal iodides (dihedral angle C(27)-C(26)-

Rh(1)-I(2) ) 19°), which leads to I(2) being displaced from

the ideal basal plane of the square pyramid (angle C(26)-Rh-

I(2) ) 105° compared with C(26)-Rh-I(1) ) 92°). A very

similar conformation of acetyl and iodide ligands is found in

the structure of [Rh(dppp)I₂(COMe)] (dppp ) Ph₂P(CH₂)₃-

PPh₂).^18,19

The reaction of the Rh(dppe) complex 1b with methyl iodide

has been reported previously, leading to [Rh(dppe)I₂(COMe)]

(3b) via the relatively long-lived intermediate [Rh(CO)(dppe)-

I₂Me)] (2b), for which isomers with CO cis or trans to methyl

1

were identified in solution by H and ³¹P NMR spectroscopy

(Scheme 2).¹⁴Since the original report did not state the ratio of

isomers observed, we have reproduced the ³¹P NMR experiment

(in CD₂Cl₂containing 3.2 M CH₃I at ambient temperature) and

found that the ratio cis-2b:trans-2b remains at ca. 2:3 for most

of the duration of the experiment.²⁰In the first spectrum recorded

after addition of the methyl iodide, however, a slightly higher

proportion of cis-2b was apparent (ratio cis-2b:trans-2b ) ca.

4:5), suggesting that cis-2b is the kinetic product and that

equilibrium between the two isomers is attained quite rapidly.

A recent theoretical investigation predicted that the two isomers

are virtually isoenergetic,²¹in agreement with our experimental

observations. The mechanism for isomerization likely involves

a five-coordinate intermediate formed by loss of a CO or iodide

ligand or by dechelation of one arm of the dppe ligand. Our

data do not distinguish these possibilities.

An X-ray crystal structure of the stable acetyl product 3b

(Figure 2) revealed a distorted square pyramidal geometry with

the acetyl ligand in the apical position, similar to structures

(19) Søtofte, I.; Hjortkjaer, J. Acta Chem. Scand. 1994, 48, 872.

(20) ³¹P NMR resonances were observed for cis-2b at δ 36.2 (dd, J_RhP) 103

Hz) and 61.1 (dd, J_RhP) 115 Hz, J_PP) 14 Hz) and trans-2b at δ 56.2 (d,

J_RhP) 120 Hz), in agreeement with the data reported in ref 14.

(21) Cavallo, L.; Sola, M. J. Am. Chem. Soc. 2001, 123, 12294.

(17) Fisher, E.; Eisenberg, R. Inorg. Chem. 1984, 23, 3216.

(18) Moloy, K. G.; Petersen, J. L. Organometallics 1995, 14, 2931.

9

J. AM. CHEM. SOC. VOL. 124, NO. 45, 2002 13599

A R T I C L E S

Gonsalvi et al.

Figure 2. X-ray structure of [Rh(dppe)I₂(COMe)] (3b). Hydrogen atoms

on the dppe ligand are omitted for clarity. Selected geometric data are given

in Table 1.

Figure 3. X-ray structure of [Ir(CO)(dppms)I₂Me] (5a). Selected hydrogen

atoms on the dppms ligand are omitted for clarity. Selected geometric data

are given in Table 2.

reported previously for other [Rh(L-L)I₂(COMe)] complexes

(L-L ) Ph₂PCH₂PPh₂, dppm,²²and dppp^18,19). In 3b, the plane

of the acetyl ligand approximately bisects the I-Rh-I angle

and the two C_acetyl-Rh-I bond angles (103° and 99°) show

less asymmetry than the corresponding angles in 3a.

Table 2. Selected Bond Lengths (Å) and Angles (deg) for Iridium

Alkyl Complexes 5a, 6, and 7

5a

6

7

Ir-C(2)

Ir-C(1)

Ir-S

Ir-P(2)

Ir-I(1)

Ir-I(2)

P(1)-S

O-C(1)

2.143(5)

1.867(4)

2.175(8)

1.826(9)

2.416(2)

2.284(2)

2.7173(6)

2.8086(6)

2.009(3)

1.167(11)

2.130(5)

1.878(5)

The reaction of the Rh(dppmo) complex 1c with methyl

iodide (1.2-3.2 M, large excess) in CH₂Cl₂gave an equilibrium

mixture of methyl and acetyl products indicated in the IR

spectrum by a terminal ν(CO) band at 2057 cm^-1for 2c and a

series of bands in the acetyl region at 1704, 1673, and 1651

cm^-1, presumably due to isomers or conformers of 3c. No pure

product was isolated from this mixture. The ³¹P NMR spectrum

of a solution of 1c dissolved in CD₂Cl₂containing MeI (20-

fold excess) showed the presence of 1c together with signals at

2.4194(9)

2.2923(10)

2.7340(4)

2.7935(4)

2.0126(13)

1.126(5)

2.4182(13)

2.2949(12)

2.7222(4)

2.7890(4)

2.0138(18)

1.135(6)

C(2)-Ir-P(2)

C(2)-Ir-S

C(2)-Ir-I(1)

C(1)-Ir-C(2)

P(2)-Ir-S

96.43(11)

86.31(11)

86.56(10)

90.74(16)

91.32(3)

94.24(12)

89.78(12)

84.79(2)

176.6(4)

94.0(2)

84.0(2)

85.3(2)

93.2(3)

89.99(7)

95.5(2)

87.7(2)

86.82(5)

178.1(8)

97.23(15)

86.69(15)

86.50(14)

94.5(2)

93.16(4)

93.41(16)

89.94(15)

83.37(3)

175.7(5)

δ 67.3 (dd, ²J_RhP) 4 Hz) and 47.8 (dd, ¹J_RhP) 107 Hz, ²J_PP

)

C(1)-Ir-P(2)

C(1)-Ir-I(1)

S-Ir-I(1)

20 Hz), which we assign to the acetyl complex [Rh(dppmo)I₂-

(COMe)] (3c). An IR spectrum of the residue from this NMR

experiment confirmed the presence of 1c and 3c as the major

species along with a small amount of the methyl complex 2c,

which was not detected by NMR. After exposure of the solution

to CO (1 atm) for 1 h the ³¹P NMR spectrum showed the

presence of another species, with signals at δ 45.1 (d) and 21.3

(dd, ¹J_RhP) 127 Hz, ²J_PP) 11 Hz, ²J_RhPnot resolved) attributed

to [Rh(CO)(dppmo)I₂(COMe)], by analogy with the pattern

reported in the literature for the dppms analogue.¹⁰

O-C(1)-Ir

spontaneous migratory insertion. Spectroscopic data for 5a

indicate the presence of two isomers in solution, in the

approximate ratio 3:1. An X-ray crystal structure (Figure 3)

showed a single isomer in the solid state, with methyl cis to

both P and S donor atoms from the chelate and CO trans to

sulfur. A solution of crystals from the batch used for the X-ray

study gave a ³¹P NMR spectrum identical to that of the major

isomer of the bulk product, allowing us to assign the geometry

of this species as that found in the crystal structure. The methyl

When the reaction of 1c with methyl iodide was carried out

in acetonitrile, a much smaller amount of the methyl complex

2c was observed in the IR spectrum, together with an increased

proportion of acetyl product which displayed two ν(CO) bands

at 1663 and 1685 cm^-1rather than the three bands observed in

CH₂Cl₂. A ³¹P NMR experiment in CD₃CN showed a product

1

ligands of both isomers each display H resonances with a

relatively weak coupling to phosphorus (³J_HP) 4 Hz) showing

that methyl is cis to coordinated P in each case. This leaves

four possible geometries for the minor isomer of 5a, which our

data do not distinguish. Complex 4a was also found to react

with ethyl iodide to give a stable ethyl complex, [Ir(CO)I₂-

(dppms)Et] (6). Again the solution NMR data showed the

presence of two isomers in the approximate ratio 3:1. An X-ray

crystal structure (Figure 4) indicated a single isomer in the solid

state with a coordination geometry about the iridium center

corresponding to that in the methyl analogue 5a. The reaction

of the Ir(dppe) complex 4b with methyl iodide to give [Ir(CO)-

(dppe)I₂Me)] (5b) (including an X-ray crystal structure of 5b)

has been reported by Cleary and Eisenberg.^23-25Their data are

with signals at δ 42.9 (d) and 29.0 (dd, ¹J_RhP) 148 Hz, ²J_PP

)

7 Hz), and an IR spectrum of this species displayed the same

pair of ν(CO) bands in the acetyl region. This suggests that

migratory insertion in 2c is promoted by coordination of

acetonitrile to the five-coordinate acetyl product 3c to give

isomers of the solvento species [Rh(dppmo)(NCMe)I₂(COMe)]

(3c-MeCN). A crystal structure of this complex (Supporting

Information) shows a structure with trans iodides and the acetyl

ligand trans to the oxygen atom of the dppmo chelate, with a

molecule of acetonitrile coordinated trans to phosphorus.

The iridium(I) dppms complex 4a reacts with methyl iodide

to give a stable methyl complex, 5a, which does not undergo

(22) Adams, H.; Bailey, N. A.; Mann, B. E.; Manuel, C. P. Inorg. Chim. Acta

1992, 198-200, 111.

(23) Cleary, B. P.; Eisenberg, R. Inorg. Chim. Acta 1995, 240, 135.

(24) Cleary, B. P. Ph.D. Thesis, University of Rochester, Rochester, NY, 1995.

9

13600 J. AM. CHEM. SOC. VOL. 124, NO. 45, 2002

Effects of Chelate Ligands in Catalysis

A R T I C L E S

Figure 6. Series of IR spectra for the reaction of 1b with MeI (1.6 M in

CH₂Cl₂, 25 °C).

Figure 4. X-ray structure of [Ir(CO)(dppms)I₂Et] (6). Hydrogen atoms are

omitted for clarity. Selected geometric data are given in Table 2.

Scheme 3. General Scheme for Reactions of [M(CO)(L-L)I] with

Methyl Iodide^a

Figure 7. Series of IR spectra for the reaction of 1c with MeI (1.6 M in

CH₂Cl₂, 25 °C). The absorptions in the region 1650-1700 cm^-1are due to

isomers or conformers of 3c, in equilibium with 2c.

Table 3. Rate Constants (25 °C) and Activation Parameters for

Methyl Iodide Oxidative Addition Reactions in CH₂Cl₂

^aAlternative isomers of methyl complexes 2 and 5 are omitted for

simplicity.

ν

(

ÌÃ),

10³k₁,

M^-¹s^-

∆

H^q,

∆

S^q,

J mol^-¹K^-

1

reactant

cm^-

kJ mol^-

1a

1b

1c

1987

2011

1983

1961

2059, 1988

1972

1994

1.19

1.41

1.14

1.37

0.0293

50.0

47 ( 1

40 ( 1

34 ( 4

56 ( 13

50 ( 1

34 ( 1

30 ( 1

54 ( 1

-144 ( 2

-167 ( 2

-188 ( 13

-112 ( 44

-165 ( 4

-156 ( 2

-169 ( 3

-113 ( 4

[Rh(CO)(PEt₃)₂I]²⁸

[Rh(CO)₂I₂]^{- 29}

4a

4b

51.8

3.12

[Ir(CO)₂I₂]^{- 30}

2045, 1969

second-order rate constants (k₁) calculated from these are given

in Table 3. Variable-temperature kinetic data over the range

10-35 °C were also measured, and satisfactory linear Eyring

plots of these data gave the activation parameters listed in Table

3. In all cases the large negative activation entropies are

consistent with the associative S_N2 mechanism commonly found

for oxidative addition of methyl iodide to metal centers.^26,27

Figure 5. Series of IR spectra for the reaction of 1a with MeI (8 M in

CH₂Cl₂,10 °C). Note the weak band due to intermediate 2a.

in agreement with ours and show that 5b is formed as a single

isomer with inequivalent phosphorus atoms and methyl trans

to iodide, analogous to the Rh isomer cis-2b (Scheme 2).¹⁴

The second-order rate constant for oxidative addition of MeI

to 1a (Table 3) represents a rate enhancement by a factor of ca.

41 at 25 °C compared to the Monsanto catalyst [Rh(CO)₂I₂]^-.²⁹

The activation parameters display a slightly lower ∆H^qand a

more favorable ∆S^qthan for oxidative addition of MeI to

Kinetics of Methyl Iodide Oxidative Addition to [M(CO)-

(L-L)I]. Kinetic measurements for the reactions of 1a-c, 4a,

and 4b with methyl iodide in CH₂Cl₂were carried out using IR

spectroscopy to monitor changes in the ν(CO) bands due to

reactants, intermediates (where possible), and products. A

generalized reaction scheme is given in Scheme 3. Typical series

of spectra for the reactions of the rhodium complexes 1a-c

are shown in Figures 5-7. Pseudo-first-order conditions were

employed (at least 10-fold excess MeI), and in all cases, the

disappearance of [M(CO)(L-L)I] was found to be first-order

in both [complex] and [MeI]. Observed pseudo-first-order rate

constants (k_obs) are given in the Supporting Information, and

[Rh(CO)₂I₂]^-, resulting in a lowering of ∆G^qby ca. 10 kJ mol^-1

.

Very similar oxidative addition rates were found for the other

Rh(I) complexes studied, with rate enhancements relative to

[Rh(CO)₂I₂]^-of 48 for 1b and 35 for 1c. The oxidative addition

rates measured here are also comparable with the data reported

by Rankin et al. for the monodentate phosphine-containing

(26) Fulford, A.; Hickey, C. E.; Maitlis, P. M. J. Organomet. Chem. 1990, 398,

311.

(27) Griffin, T. R.; Cook, D. B.; Haynes, A.; Pearson, J. M.; Monti, D.; Morris,

G. E. J. Am. Chem. Soc. 1996, 118, 3029. Rendina, L. M.; Puddephatt, R.

J. Chem. ReV. 1997, 97, 1735. Labinger, J. A.; Osborn, J. A. Inorg. Chem.

1980, 19, 3230.

(25) Albietz, P. J., Jr.; Cleary, B. P.; Paw, W.; Eisenberg, R. Inorg. Chem. 2002,

41, 2095.

9

J. AM. CHEM. SOC. VOL. 124, NO. 45, 2002 13601

A R T I C L E S

Gonsalvi et al.

complex trans-[Rh(CO)(PEt₃)₂I].²⁸It is noteworthy that the

complexes 1a-c and trans-[Rh(CO)(PEt₃)₂I] all have similar

nucleophilicity toward methyl iodide, varying by a factor of

less than 1.5 in rate (or 1 kJ mol^-1in ∆G^q₂₉₈) despite the

relatively wide range (50 cm^-1) of ν(CO) frequencies for the

four complexes. The ν(CO) values of metal carbonyl complexes

are often taken as a good measure of electron density at the

metal center, and so might be expected to be related to

nucleophilicity. However, it is important to consider the local

environment of the coordinated CO and, in particular, the trans

ligand. For example, complex 1b might be expected to have

electron density on the Rh center rather similar to that of trans-

[Rh(CO)(PPh₃)₂I], but the ν(CO) values are actually quite

different (2011 and 1981 cm^-1, respectively) since a CO trans

to iodide experiences considerably more back-donation than one

trans to PPh₃. The use of ν(CO) to judge electron density on

the metal should therefore be made with consideration of the

stereochemistry of the complexes being compared.³¹

1c. Somewhat surprisingly though, no spectroscopic evidence

for an anionic species was found in this case.

As expected, faster oxidative addition rates were found for

the iridium complexes 4a and 4b. The ratios k_Ir/k_Rhwere found

to be 42 for the pair of dppms complexes and 37 for the pair of

dppe complexes, corresponding to a ∆∆G^q₂₉₈of ca. 9 kJ mol^-1

between Rh and Ir. A bigger difference in reactivity toward

methyl iodide was found previously for the anionic iodocarbo-

nyls [M(CO)₂I₂]^-, for which k_Ir/k_Rh) 120 corresponding to

∆∆G^q₂₉₈≈ 12 kJ mol^{-1 30}

. The smaller difference in reactivity

found between 4d and 5d metal complexes in this study can be

ascribed to the presence of bulkier bidentate ligands which

moderate the intrinsic nucleophilicity of the metal center. For

all the oxidative addition reactions studied, addition of polar

solvents (THF, MeCN, and MeOH) was found to give moderate

rate enhancements (by factors of between 1 and 3; see the

Supporting Information) as expected for S_N2 reactions proceed-

ing via polar transition states. Coordination of solvent to the

metal center during the S_N2 process may also play a role in

accelerating oxidative addition.

Halide salts have previously been found to accelerate MeI

addition to square planar Rh(I)^26,32and Ir(I)³³complexes. Rate

enhancements were also found for the complexes studied here

(k_obsvalues are given in the Supporting Information). Addition

of 1 equiv of Bu₄NI (per Rh) increased the oxidative addition

rate by factors of ca. 1.3 (1a), 1.1 (1b), and 2.4 (1c), while 10

equiv of Bu₄NI gave rate enhancements of ca. 2.1 (1a), 1.6 (1b),

and 8.1 (1c). It has been proposed previously that coordination

of an additional halide ligand increases the nucleophilicity of

the metal center. We have therefore probed the behavior of

complexes 1a-1c in the presence of excess ionic iodide.

Addition of a 100-fold excess of Bu₄NI to solutions of 1a or

1c in CH₂Cl₂did not give rise to any new rhodium carbonyl

species detectable by IR spectroscopy. By contrast, under the

same conditions the dppe complex 1b gives rise (over 3 h) to

Kinetics of Migratory CO Insertion in [M(CO)(L-L)-

I₂Me]. In the series of spectra for the reaction of 1a with MeI

(Figure 5), a weak absorption is apparent at 2062 cm^-1in

addition to the bands at 1987 and 1701 cm^-1due to reactant 1a

and product 3a, respectively. This weak band occurs in the

region expected for the methyl intermediate 2a, on the basis of

the frequency shift (75 cm^-1) relative to the Rh(I) precursor

1a.³⁵A similar shift in ν(CO) (69 cm^-1) is found between the

stable iridium analogues 4a and 5a. The kinetic behavior of

the 2062 cm^-1band is also consistent with assignment to

intermediate 2a, since it decays in direct proportion to the band

of 1a as predicted by the steady-state approximation

d[2a]

dt

≈ 0 ) k₁[1a][MeI] - k_-1[2a] - k₂[2a]

(3)

(4)

significant amounts of a new species with ν(CO) at 1960 cm^-1

.

The shift of ν(CO) to low frequency is consistent with coor-

dination of iodide to form an anionic complex, [Rh(CO)I₂(dppe)]^-.

This complex could be five-coordinate (with the dppe ligand

remaining bidentate) or square planar with one arm of the dppe

ligand dissociating. We note that a phosphine ligand in [Rh-

(CO)I(PPh₃)₂] can be displaced by iodide to give an anionic

complex, [Rh(CO)I₂(PPh₃)]^-(ν(CO) 1963 cm^-1), which is more

reactive toward MeI than the neutral precursor.³²The iodide

effects observed in our kinetic studies are best explained by

iodide coordination to give reactive anionic intermediates,

possibly accompanied by chelate ring opening.³⁴The most facile

dechelation would be expected for the dppmo chelate ligand,

and the largest rate enhancement is indeed found for complex

k₁[MeI]

k_-1+ k₂

[2a]

[1a]

)

Since k_obsis given by the expression

k₁k₂[MeI]

k_-1+ k₂

k_obs

)

(5)

(6)

eqs 4 and 5 can be combined to give

k_obs

k₂)

[2a]/[1a]

The migratory insertion rate constant, k₂, can therefore be

obtained from the measured value of k_obsand the steady-state

ratio [2a]/[1a], which can be estimated on the basis of observed

IR absorbances and estimated relative extinction coefficients

for 1a and 2a.³⁶Values of k₂in the range 10-25 °C obtained

(28) Rankin, J.; Poole, A. D.; Benyei, A. C.; Cole-Hamilton, D. J. Chem.

Commun. 1997, 1835. Rankin, J.; Benyei, A. C.; Poole, A. D.; Cole-

Hamilton, D. J. J. Chem. Soc., Dalton Trans. 1999, 3771.

(29) Haynes, A.; Mann, B. E.; Morris, G. E.; Maitlis, P. M. J. Am. Chem. Soc.

1993, 115, 4093.

(30) Ellis, P. R.; Pearson, J. M.; Haynes, A.; Adams, H.; Bailey, N. A.; Maitlis,

P. M. Organometallics 1994, 13, 3215.

(31) An empirical method of predicting ν(CO) values for metal complexes

(Timney, J. A. Inorg. Chem. 1979, 18, 2502) takes this into account by

utilizing different ligand effect constants depending upon whether the ligand

is cis or trans to CO. The Timney method predicts ν(CO) values of 2006.6

and 1970.5 cm^-1, respectively, for the cis and trans isomers of [Rh(CO)-

(PPh₃)₂I], in agreement with the experimental trend.

(34) Our (unpublished) mechanistic studies of the reactions of [Rh(CO)I(Ph₂-

PCH₂CH₂PAr₂)] (Ar ) 3,5-(CF₃)₂C₆H₃or 3,4,5-F₃C₆H₂) with MeI indicate

that chelate ring opening and coordination of iodide is more facile in these

complexes.

(35) Attempts to observe 2a by low-temperature ³¹P NMR were hampered by

the low solubility of 1a under these conditions.

(32) Forster, D. J. Am. Chem. Soc. 1975, 97, 951.

(36) The band due to 2a at 2062 cm^-1occurs under the high-frequency tail of

the strong absorption of 1a at 1987 cm^-1. The value of the ratio Abs(2a)/

Abs(1a) at 8 M MeI was therefore obtained by plotting A(2062)/A(1987)

vs [MeI] and subtracting the y intercept of the linear plot from the measured

(33) Basson, S. S.; Leipoldt, J. G.; Purcell, W.; Schoeman, J. B. Inorg. Chim.

Acta 1990, 173, 155. de Waal, D. J. A.; Gerber, T. I. A.; Louw, W. J.

Inorg. Chem. 1982, 21, 1259.

9

13602 J. AM. CHEM. SOC. VOL. 124, NO. 45, 2002

Effects of Chelate Ligands in Catalysis

A R T I C L E S

Table 4. Rate Constants (25 °C, CH₂Cl₂) and Activation

Parameters for Migratory Insertion Reactions of Rhodium Methyl

Complexes

which precluded the determination of accurate rate data for the

migratory insertion step. However, since the IR bands due to

the acetyl product 3c grow-in simultaneously with those of the

methyl intermediate 2c, the equilibrium between 2c and 3c must

be attained quite rapidly, indicating a relatively fast methyl

migration. The replacement of sulfur by oxygen as the donor

atom in the bidentate ligand therefore appears to affect the

thermodynamics of migratory insertion more than the kinetics.

No spontaneous methyl migration was observed for 5a and

5b, in accordance with the behavior of other iridium(III) methyl

complexes.³⁹We therefore investigated the propensity for these

complexes to undergo CO insertion by studying their reactions

with carbon monoxide at high temperature and pressure. On

the basis of previous observations for the anionic complex

[Ir(CO)₂I₃Me]^-, we chose a solvent system comprising 1% (v/

v) methanol in chlorobenzene.⁴⁰When the reaction of 5a with

CO (10 bar) at 93 °C was followed by in situ high-pressure IR

spectroscopy, the ν(CO) band of the reactant at 2039 cm^-1was

replaced by a new terminal ν(CO) band at 2051 cm^-1in a first-

order process (k_obs) 7.3 × 10^-4s^-1). Under the same conditions

the rate constant for carbonylation of [Ir(CO)₂I₃Me]^-is

10³k₂,

1

s^-

∆

H^q,

∆

S^q,

J mol^-¹K^-

1

reactant

kJ mol^-

2a

2b

620

0.4

54

54 ( 7

83 ( 2

63 ( 2

-67 ( 15

-30 ( 5

-59 ( 9

[Rh(CO)₂I₃Me]^{- 29}

by this method are listed in the Supporting Information. Table

4 gives the value of k₂at 25 °C together with activation

parameters derived from an Eyring plot.

The series of spectra in Figure 6 show that a much higher

concentration of methyl intermediate 2b is formed during the

reaction of 1b with methyl iodide, as a consequence of a much

slower migratory insertion step in the dppe system. Infrared

spectroscopy does not distinguish between the cis and trans

isomers of 2b (Scheme 2), but ³¹P NMR spectroscopy (vide

supra) showed that a relatively fast equilibration leads to an

isomeric ratio cis-2b:trans-2b of ca. 2:3 which remains steady

during most of the experiment. Apparent rate constants for

migratory insertion were obtained by monitoring the slow

exponential decay of the combined ν(CO) band of cis- and trans-

2b at 2075 cm^-1once the relatively rapid oxidative addition

step was complete. The resulting rate constants (Supporting

Information) were found to be essentially independent of [MeI].

An estimate of the true rate constant, k₂, for migratory insertion

must take into account the proportion of 2b existing in the cis

form, since only that isomer possesses the appropriate geometry

for cis migration.³⁷On the basis of the mechanism shown in

Scheme 2, and assuming that cis-trans isomerization is fast

relative to migratory insertion, the observed rate constant is

given by k₂/(1 + K_e). Since ³¹P NMR gave an estimate of K_e)

1.5, k₂can be obtained by multiplying the observed rate constant

by 2.5, leading to the value given in Table 4. The activation

parameters for the reaction cis-2b f 3b given in Table 4 were

estimated with the approximation that K_efor the cis-trans

isomerization is invariant with temperature.

estimated as ca. 3 × 10^-4s^{-1 41}

. The recovered product showed

IR bands at 2057 and 1642 cm^-1(CH₂Cl₂), consistent with a

product containing terminal and acetyl CO ligands, and the ¹H

and ³¹P NMR spectra of this product gave evidence for two

isomers in an approximate 1:1 ratio [δ(¹H) 2.59, 2.77; δ(³¹P)

63.1, 4.5 (each d, ²J_PP) 36 Hz) and 56.3, -7.8 (each d, ²J_PP

)

25 Hz)]. An X-ray crystal structure determination was hampered

by disorder but was consistent with the presence of [Ir(CO)-

(dppms)I₂(COMe)] having distorted octahedral geometry with

CO trans to iodide and acetyl trans to S of the dppms chelate.

By contrast, in situ high-pressure IR spectroscopy revealed no

decay or shift of the ν(CO) absorption of 5b over 3 h under the

same conditions. The iridium methyl complexes 5a and 5b

therefore display the same trend in reactivity as their rhodium

analogues 2a and 2b. Evidence was obtained that carbonylation

of 5b could be induced at higher temperature, a product with

ν(CO) absorptions at 2057 and 1649 cm^-1(CH₂Cl₂) being

obtained after treatment of a solution of 5b with 9 bar of CO in

a Fisher Porter vessel at 120 °C for 3 h.

X-ray Crystallographic Studies. The X-ray crystal structure

of the iridium methyl complex 5a (Figure 3) displays a pseudo-

octahedral geometry with the methyl group occupying a position

cis to both P and S donor atoms of the dppms ligand. The

carbonyl ligand is trans to S, as for the Ir(I) precursor. The

five-membered chelate ring adopts an envelope conformation

with the methylene carbon at its apex. This conformation places

two of the dppms phenyl groups in pseudoaxial positions and

creates a crowded “pocket” surrounding the methyl ligand. The

other two phenyl groups are placed in pseudoequatorial posi-

tions, such that the iodide trans to methyl is in a much less

crowded environment. The adoption of this conformation can

be explained by the smaller cone angle of methyl (90°)

compared with iodide (107°).¹The closest nonbonded contact

Despite the indirect methods employed for obtaining these

kinetic parameters, there is clearly a dramatic difference in

reactivity between 2a and 2b, with the dppms complex

undergoing migratory insertion ca. 1500 times faster than the

dppe complex, corresponding to a lowering of ∆G^qby ca.

298

18 kJ mol^-1. Thus, while migratory insertion in 2a is accelerated

by an order of magnitude compared with that in [Rh(CO)₂I₃Me]^-,

the reaction of 2b is >10 times slower than for the conventional

Monsanto catalyst.^29,38By comparison, in the PEt₃system

studied by Rankin et al., it was found that migratory insertion

only proceeded slowly when [Rh(CO)(PEt₃)₂I₂Me] was sub-

jected to a high pressure of carbon monoxide, which trapped

the presumed five-coordinate intermediate to give [Rh(CO)-

(PEt₃)₂I₂(COMe)].²⁸

More complex behavior was exhibited by the dppmo system,

where reaction of 1c with MeI gave an equilibrium mixture of

methyl and acetyl products 2c and 3c (Figure 7 and vide supra),

(39) Atwood, J. D. In ComprehensiVe Organometallic Chemistry II; Abel, E.

W., Stone, F. G. A., Wilkinson, G., Eds.; Pergamon: Oxford, 1995; Vol.

8, p 303.

value of A(2062)/A(1987) at 8 M MeI. This was then converted into a

ratio of concentrations, [2a]/[1a], by dividing by 0.6 the ratio of extinction

coefficients measured for the ν(CO) bands of the Ir analogues 5a and 4a.

(37) In our preliminary communication of these results (ref 16) the influence

of isomeric forms of 2b on the migratory insertion rate constant was

neglected.

(40) Pearson, J. M.; Haynes, A.; Morris, G. E.; Sunley, G. J.; Maitlis, P. M. J.

Chem. Soc., Chem. Commun. 1995, 1045.

(41) Calculated rate at 93 °C using the activation parameters ∆H^q) 33 kJ mol^-1

and ∆S^q) -197 J mol^-1K^-1for carbonylation of [Ir(CO)₂I₃Me]^-in 25%

(v/v) MeOH-PhCl (ref 40) and dividing by 25 to compensate for the

different MeOH concentrations.

(38) Haynes, A.; Mann, B. E.; Gulliver, D. J.; Morris, G. E.; Maitlis, P. M. J.

Am. Chem. Soc. 1991, 113, 8567.

9

J. AM. CHEM. SOC. VOL. 124, NO. 45, 2002 13603

A R T I C L E S

Gonsalvi et al.

Table 5. Distances and Angles between Pseudoaxial Phenyl

Groups in Complexes of dppms, dppmo, and dppe

r_ipso

,

r_cent

Å

,

θ,

deg

complex

geometry

Å

ref

[Rh(CO)(dppms)Cl]

[Pd(dppms)I₂]

sq pl

tet

sq pyr

oct

3.49

3.43

3.55

4.10

3.67

3.42

4.25

3.73

4.13

3.77

3.89

3.78

4.11

5.09

4.03

3.70

5.04

4.23

4.91

4.44

15

9

10

42

12

43

[Pd(dppms)₂]²⁺

17

35

19

10

55

39

45

44

[Rh(CO)(dppmo)Cl]

[Hg(dppms)I₂]

[Rh(dppms)I₂(COMe)] (3a)

[Rh(dppe)I₂(COMe)] (3b)

[Ir(CO)(dppms)I₂Me] (5a)

[Ir(CO)(dppms)I₂Et] (6)

[{Ir(CO)(dppms)I₂Me}₂]²⁺(7)

Figure 8. Space-filling models of the X-ray structures of complexes 3a,

5a, and 6. Note the increasing separation and widening of the angle between

the axial phenyl groups (shaded blue) caused by the introduction of methyl

and ethyl ligands (shaded red) in 5a and 6, respectively.

oct

between hydrogens of the phenyl and methyl groups (H5A and

H2A) is ca. 1.9 Å, which is less than double the van der Waals

radius of hydrogen. This steric interaction causes the methyl

ligand to lean away from the phenyl group, giving a P-Ir-Me

bond angle of 96°, which is the biggest deviation from regular

octahedral geometry in the complex. By contrast the P-Ir-

Me angles in 5b are both close to 90°.²⁴The crystal structures

therefore suggest that steric interactions might play a role in

accelerating migratory insertion in 2a and 5a compared to

analogous dppe complexes.

Figure 9. Ligand conformation and quadrant diagram for dppms complexes.

to reorientation of the equatorial phenyl groups. The structural

rearrangements caused by axial methyl or ethyl ligands in 5a

and 6 are also evident from the larger values of r_ipsoand r_cent

for these complexes in Table 5.

It is informative to compare the structure of 5a with

complexes of dppms which have fewer than two “axial” ligands

(where the equatorial plane is defined as that containing the

metal atom and the two donor atoms of the bidentate ligand).

The structure of [Rh(CO)(dppms)Cl] determined by Baker et

al.¹⁰exhibits a ligand conformation where two axial phenyls

block one face of the square planar complex. Again the five-

membered chelate ring adopts an envelope structure, but in this

case the sulfur-bound phosphorus is at the apex. Very similar

dppms ligand conformations are displayed in the recently

reported⁴²square planar palladium(II) complexes [Pd(dppms)-

I₂] and [Pd(dppms)₂]²⁺and in the tetrahedral [Hg(dppms)I₂].⁴³

A measure of the spatial arrangement of the axial phenyl

groups is provided by the distances between the ipso carbons

(r_ipso) or the ring centroids (r_cent) along with the angles between

the phenyl planes (θ) given in Table 5. For the square planar

dppms complexes θ is less than 20° with r_ipso) ca. 3.5 Å and

r_cent) ca. 3.8 Å (except in the more crowded bischelate [Pd-

(dppms)₂]²⁺). In [Hg(dppms)I₂] the phenyl rings move apart

slightly to accommodate an iodide in the tetrahedral structure.

An approximate parallel stacking arrangement (with θ ) 10°)

is maintained in the square pyramidal rhodium acetyl complex

3a (Figure 1), with the axial phenyls flanking the vacant sixth

coordination site, allowing the acetyl ligand to occupy the

sterically less demanding axial site. The space-filling diagrams

in Figure 8 show how θ widens to 39° to accommodate the

methyl ligand in 5a. This is achieved by a twist of the phenyl

attached to the iridium-bound phosphorus.⁴⁴An even more

pronounced distortion is apparent for the ethyl analogue 6, in

which θ is 45°. The accommodation of the ethyl ligand in 6

forces significant rearrangement in the chelate ring, which has

the Ir-bound phosphorus at the apex of the envelope, leading

A feature of all the structures containing the dppms ligand is

the steric crowding of one axial coordination site due to the

disposition of two of the phenyl groups. This steric environment

can be expressed in terms of a quadrant diagram (Figure 9).⁴⁵

When the two phenyls flank a vacant site (e.g., in square planar

or square pyramidal complexes), they can stack approximately

parallel, whereas in an octahedral complex, an axial ligand is

placed in the sterically crowded region, which disrupts the

stacking. The conformation adopted by phenyl groups in dppe

complexes is noticeably different from that described above for

the dppms ligand. For example, the angle between the planes

of the phenyl groups flanking the vacant coordination site in

3b is 55°, compared with 10° in the dppms analogue 3a. This

suggests that there is much less preference for phenyl groups

to stack parallel in the dppe system. A detailed analysis of

structures of this type by Morton and Orpen⁴⁶revealed that the

five-membered M(dppe) chelate ring typically adopts a twist

(C₂) rather than an envelope conformation, with a weak

preference for the conformation of phenyl groups shown in

Figure 10a, in which diagonal quadrants are more hindered. The

X-ray structures of octahedral complexes 5b²⁴and [Ir(dppe)I₄]^{- 47}

display just such a conformation with little crowding of the axial

methyl or iodide ligands.

Morton and Orpen also considered the M₂(µ-dppm) fragment

and revealed a preferred conformation (Figure 10b) very similar

to that observed for M(dppms).⁴⁶It was concluded that the M₂-

(µ-dppm) ring is more rigid than the M(dppe) system and places

greater constraints on the conformation of the phenyl groups.

The M(dppms) chelate ring can be regarded as analogous to a

dinuclear M₂(µ-dppm) system, with a sulfur replacing one of

the metal atoms, so similar conformational preferences are to

be expected. The driving force for adoption of the stacked

(42) Wong, T. Y. H.; Rettig, S. J.; James, B. R. Inorg. Chem. 1999, 38, 2143.

(43) Lobana, T. S.; Sandhu, M. K.; Liddell, M. J.; Tiekink, E. R. T. J. Chem.

Soc., Dalton Trans. 1990, 691.

(44) As judged by the M-P-C_ipso-C_orthodihedral angles of 10.6° in 3a and

42.8° in 5a. The corresponding dihedral angle for the other axial phenyl,

S-P-C_ipso-C_orth, is small (<1°) in both structures.

(45) Knowles, W. S. Acc. Chem. Res. 1983, 16, 106.

(46) Morton, D. A. V.; Orpen, A. G. J. Chem. Soc., Dalton Trans. 1992, 641.

(47) Chan, Y. N. C.; Meyer, D.; Osborn, J. A. J. Chem. Soc., Chem. Commun.

1990, 869.

9

13604 J. AM. CHEM. SOC. VOL. 124, NO. 45, 2002

Effects of Chelate Ligands in Catalysis

A R T I C L E S

Scheme 4. Mechanism for Reaction of 1a with MeI

Figure 10. Conformations and quadrant diagrams for (a) M(dppe) and (b)

M₂(µ-dppm) complexes.

an explanation for the equilibrium between methyl and acetyl

products 2c and 3c formed in the reaction of [Rh(CO)(dppmo)I]

with MeI.

arrangement of axial phenyl groups in dppms complexes is

primarily the envelope conformation of the chelate ring. An

intriguing possibility is that attractive π-π stacking interactions

between the phenyls further stabilize this arrangement. Recent

crystallographic and theoretical studies by Pregosin and co-

workers⁴⁸have concluded that attractive intramolecular π-π

stacking interactions between aromatic rings in coordinated

ligands can result in significant structural distortion of square

planar Pd(II) complexes. The distances between the phenyl rings

for dppms complexes (Table 5) are close to the range (3.4-3.5

Å) for π-π stacking in biological systems,⁴⁹suggesting that

such interactions might play a role in tuning organometallic

reactivity.

Reaction Mechanism for Oxidative Addition to 1a. In light

of the above structural analysis we propose that the ligand

conformation plays a key role in determining reactivity in the

dppms system. The suggested sequence of reactions is shown

in Scheme 4. Oxidative addition of methyl iodide to a square

planar complex is generally believed to proceed in two steps:

(i) nucleophilic attack by the metal center at carbon to displace

I^-and (ii) coordination of I^-to the five-coordinate intermedi-

ate.^26,27In the reaction of 1a it is likely that the MeI molecule

will approach the less-hindered face of the metal complex. The

S_N2 step will generate an initial cationic intermediate, [Rh(CO)-

(dppms)IMe]⁺, in which the vacant coordination site is flanked

by the axial phenyl groups. To complete the oxidative addition,

iodide must coordinate to the vacant site, but since iodide has

a larger cone angle than methyl, we propose that a change in

ligand conformation accommodates the incoming iodide.⁵⁰This

generates an octahedral product in which the axial phenyl groups

flank the methyl ligand, as found in the crystal structure of 5a.

Migratory CO insertion then proceeds with another change in

ligand conformation to place the acetyl ligand in the less

hindered axial coordination site, as in the crystal structure of

3a. Steric congestion is therefore relieved by migratory insertion.

We attempted to model the product of the first (S_N2) step in

Scheme 4 by treating 4a with methyl triflate. The reaction gave

a product with a ν(CO) band at 2058 cm^-1and analytical data

consistent with a cationic Ir(III) complex with a triflate counter-

ion, [Ir(CO)(dppms)IMe][OTf]. A solution of this product in

CD₂Cl₂shows four sets of signals in the ³¹P NMR (see the

Experimental Section), indicating the presence of isomers.

Crystals grown from a CH₂Cl₂solution were suitable for an

X-ray diffraction study which revealed a centrosymmetric

iodide-bridged dimer, [{Ir(CO)(dppms)(µ-I)Me}₂]²⁺(7) (Scheme

5 and Figure 11), together with two noncoordinated triflate

anions and a molecule of solvent. Thus, the hoped-for five-

coordinate intermediate actually dimerizes, at least in the solid

state.

On the basis of the structural analysis, we propose that

migratory insertion in [M(CO)(dppms)I₂Me] (2a, M ) Rh; 5a,

M ) Ir) is accelerated (compared to that in the dppe analogues)

by relief of the steric tension created by phenyl groups on the

bidentate ligand. It is clear that the acetyl ligand of 3a occupies

a much less sterically demanding site than the methyl ligand in

5a. If the steric crowding in 2a (or 5a) is partially removed in

the transition state for methyl migration, then an acceleration

will result (compared with complexes without similar steric

strain). Although alternative ligand conformations are undoubt-

edly accessible in solution, the solid-state structures exhibit

trends which are consistent with steric effects exerting an

important influence on reactivity. A recent theoretical study on

this system supports this argument (vide infra).²¹

Conformational analysis⁴⁶suggested that the rigidity of the

envelope structure (and constraints on the phenyl groups) in

the M₂(µ-dppm) fragment increases for longer M-M bonds.

Conversely, a decrease in rigidity would be expected on

shortening the M-M bond, an effect which can be mimicked

in our system by replacing the dppms ligand with dppmo.

Consistent with this, θ increases from 15° in [Rh(CO)(dppms)-

Cl] (r_Rh-S) 2.403 Å)¹⁰to 35° in [Rh(CO)(dppmo)Cl] (r_Rh-O

) 2.109 Å)¹²with corresponding increases in r_ipsoand r_cent

(Table 5). This reduction in steric crowding around one of the

axial coordination sites when dppms is replaced by dppmo offers

1

(50) The H NMR spectra of 1a and 4a exhibit pseudotriplets indicating time-

(48) Drago, D.; Pregosin, P. S.; Tschoerner, M.; Albinati, A. J. Chem. Soc.,

Dalton Trans. 1999, 2279. Magistrato, A.; Merlin, M.; Pregosin, P. S.;

Rothlisberger, U. Organometallics 2000, 19, 3591. Magistrato, A.; Pregosin,

P. S.; Albinati, A.; Rothlisberger, U. Organometallics 2001, 20, 4178.

(49) Ranganathan, D.; Haridas, V.; Gilardi, R.; Karle, I. L. J. Am. Chem. Soc.

1998, 120, 10793. Hobza, P.; Kabela´c, M.; Sponer, M.; Mejzl´ık, P.;

Vondra´sek, J. J. Comput. Chem. 1997, 18, 1136.

averaged equivalence of the two methylene protons of dppms due to rapid

inversion of the chelate envelopes on the NMR time scale. The proposed

changes in ligand conformation are therefore feasible. The ¹H NMR spectra

of 3a, 5a, and 6 display pairs of signals with ddd patterns, due to

2

inequivalent methylene protons with J_HHand two J_HPcouplings. This

inequivalence is conferred by unsymmetrical coordination in the two axial

sites, and will be maintained even if the ligand conformation is fluxional.

9

J. AM. CHEM. SOC. VOL. 124, NO. 45, 2002 13605

A R T I C L E S

Gonsalvi et al.

bulky ferrocene fragment prevents complete oxidative addition

of MeI to give an octahedral methyl complex, and it would seem

that steric effects in this system are more extreme than for the

dppms ligand.

Relevance to Catalytic Methanol Carbonylation. To pro-

mote rhodium/iodide-catalyzed methanol carbonylation, ligands

are sought which act as good electron donors, thus binding

effectively to rhodium and increasing the nucleophilicity of the

metal center toward MeI. Using the experimental activation

parameters for oxidative addition of MeI to 1a (Table 3),

extrapolation to 185 °C gives a predicted rate constant of 1.25

(+0.8, -0.5) dm³mol^-1s^-1compared with a value of 2.16

dm³mol^-1s^-1, calculated from the carbonylation rate reported

by Baker et al. for a dppms-promoted Rh catalyst.¹⁰Thus, the

model kinetic data fully support rate-determining oxidative

addition of MeI to 1a in the catalytic system.

Figure 11. X-ray structure of [Ir(CO)(dppms)(µ-I)Me]₂OTf₂(7). Selected

H atoms, triflate counterions, and a CH₂Cl₂solvent molecule are omitted

for clarity. Selected geometric data are given in Table 2.

Most phosphine ligands found to accelerate the oxidative

addition step (e.g., PEt₃, dppe) tend to slow the subsequent

migratory CO insertion step, for electronic reasons considered

below. The dppms ligand is unusual in that it promotes both

MeI oxidative addition and migratory insertion relative to the

corresponding reactions in the catalytic cycle for [Rh(CO)₂I₂]^-.

Extrapolation using the experimental activation parameters in

Table 4 predicts a rate constant of ca. 2 × 10³s^-1for migratory

insertion in 2a at 185 °C, consistent with this step being fast

relative to oxidative addition. Thus, while dppms fulfills the

role of a good donor ligand, it also has the appropriate properties

to encourage methyl migration.

A common problem associated with ligand-promoted metha-

nol carbonylation processes is the long-term stability of the

metal-ligand bonds. For example, the PEt₃-promoted system

reported by Rankin et al. gave high catalytic activity for only

15 min (at 150 °C).²⁸After this time, “normal” catalytic rates

associated with the presence of the conventional [Rh(CO)₂I₂]^-

catalyst were observed, the phosphine ligand being lost from

the metal coordination sphere as PEt₃Me⁺, PEt₃H⁺, and Od

PEt₃. In the dppms-promoted system, the reported high catalytic

rates (compared to those of [Rh(CO)₂I₂]^-) can be sustained for

up to 1 h (at 185 °C). After more prolonged reaction times,

activity drops and a Rh(dppm) complex can be isolated from

the reaction mixture,⁵³indicating catalyst deactivation resulting

from extrusion of sulfur from the dppms ligand. GC analysis

of the reactor headspace indicated the presence of H₂S.⁵⁴The

exact mechanism by which dppms ligand degradation occurs

during catalysis is unclear, although dechelation of the sulfur

arm of the chelate followed by reaction with water, acid, or H₂

(formed by the water gas shift reaction) are all feasible.⁵⁵The

in situ IR spectroscopic data (at 185 °C and 70 bar of CO) gave

no evidence for detectable quantities of species with a mono-

dentate dppms ligand, the only complex observed during the

period of high catalytic activity being the chelate complex 1a.

Scheme 5. Addition of Methyl Triflate to 4a

The coordination geometry around each Ir center in 7 is very

similar to that in 5a. The methyl ligand is again flanked by

axial phenyl groups from the dppms ligand, which adopts an

envelope conformation with the methylene carbon at its apex.

The angle between the phenyl planes, θ, is 5° wider than in 5a,

and the methyl again leans away from the coordinated phos-

phorus, with the bond angle P(2)-Ir(1)-C(1) ) 97°. Reaction

of 7 with a stoichiometric amount of Bu₄NI in CH₂Cl₂resulted

in bridge cleavage to give 5a, with IR and ³¹P NMR spectra

identical to those of a sample synthesized by the addition of

MeI to 4a. These results support the two-step mechanism for

oxidative addition of MeI, but dimerization of the five-

coordinate cation precluded an investigation of its ligand

conformation.

A recent investigation⁵¹into the structure and reactivity of

[Rh(CO){κ²-(P,S)-Fe(η⁵-C₅Me₄P(S)Ph₂)(η⁵-C₅Me₄PPh₂)}Cl] (8)

makes an interesting comparison to the dppms system. The

ν(CO) frequency of 8 (1987 cm^-1) is similar to that of the dppms

analogue (1992 cm^-1), suggesting that the two P,S bidentate

ligands have similar donor ability.⁵²An X-ray crystal structure

of 8 shows that the carbonyl ligand is coordinated trans to S as

in 1a and that one face of the square planar Rh(I) complex is

effectively blocked by the ferrocene fragment. The reaction of

8 with MeI gave no rhodium(III) methyl or acetyl products,

but instead resulted in halide metathesis to give the analogous

iodide complex and MeCl. Further reaction with MeI resulted

in dissociation of the bidentate ligand and quaternization of its

PPh₂group. These results were interpreted on the basis that the

(53) ³¹P{¹H} NMR spectroscopic analysis of the residue recovered from catalytic

reactions shows a doublet at δ -43.0 (J_RhP) 84 Hz) characteristic of a

Rh(dppm) complex with equivalent P nuclei. In situ IR spectroscopy also

indicates formation of some [Rh(CO)₂I₂]^-after prolonged reaction.

(54) In a reaction which resembles the reverse of the dppms ligand degradation

found here, Pd₂(µ-dppm) complexes are reported to desulfurize H₂S, giving

dppms as a byproduct. Wong, T. Y. H.; Barnabas, A. F.; Sallin, D.; James,

B. R. Inorg. Chem. 1995, 34, 2278. See also ref 42.

(55) Reversible migration of alkyl groups from acetyl to sulfur (via intermediate

rhodium alkyl species) has been reported in complexes of the type

[Rh(PPh₃)(L)(mnt)(COR)] (mnt ) maleonitriledithiolate) (Cheng, C.-H.;

Spivack, B. D.; Eisenberg, R. J. Am. Chem. Soc. 1977, 99, 3003. Cheng,

C.-H.; Eisenberg, R. Inorg. Chem. 1979, 18, 2438).

(51) Broussier, R.; Laly, M.; Perron, P.; Gautheron, B.; Nifant’ev, I. E.; Howard,

J. A. K.; Kuz’mina, L. G.; Kalck, P. J. Organomet. Chem. 1999, 587, 104.

(52) Similarly the iodide M has ν(CO) 1981 cm^-1compared with 1987 cm^-1

for 1a.

9

13606 J. AM. CHEM. SOC. VOL. 124, NO. 45, 2002

Effects of Chelate Ligands in Catalysis

A R T I C L E S

Chart 1. Diphosphine Ligands for Catalytic CO/Alkene

This contrasts with the observations of Wegman et al., who

found that a dicarbonyl, [Rh(CO)₂I(dppeo)], with monodentate

P-coordinated dppeo was the only species present during

catalytic carbonylation reactions at 80 °C and ∼3 bar of CO.¹²

Thus, dppms shows much less propensity for hemilabile

behavior than dppeo. This can be rationalized by the greater

stability of five-membered compared with six-membered chelate

rings, coupled with the stronger bonding of sulfur to a soft Rh(I)

center. The dppeo ligand was found not to be an effective

promoter at the higher temperature and pressure employed for

the dppms system. It appears, therefore, that retention of the

Rh(dppms) chelate structure is key to achieving high catalytic

activity. This structure confers the appropriate electronic and

steric environment to the Rh center for promotion of both MeI

oxidative addition and CO insertion steps, as shown by our

model studies under mild conditions.

Copolymerization and Hydroformylation

Diphosphine ligands, while promoting oxidative addition to

a degree similar to that of dppms, are not found to be as effective

for catalytic methanol carbonylation. For example, dppe gave

a catalytic rate an order of magnitude lower than that of

[Rh(CO)₂I₂]^-, although some improvement over dppe was

attained using unsymmetrical derivatives with fluorinated aryl

did not detect [Rh(CO)(dppe)I₂(COMe)] and found that acetic

acid could only be generated from 3b at 140 °C under a pressure

of CO. Thus, the dppms ligand also favors the reductive

elimination step compared with diphosphine complexes.

Relationship to Other Catalytic Reactions. It has long been

appreciated that an orientation of phenyl groups which shields

two diagonal quadrants is required for effective asymmetric

hydrogenation catalysts.^45,57However, there are other catalytic

reactions where there may be a requirement for steric hindrance

concentrated in two adjacent quadrants, as found in the dppms

system. Recent studies of ligand effects on Pd-catalyzed CO/

ethene copolymerization by Bianchini et al.⁵⁸have shown that

modification of the backbone of the conventional dppp ligand

gives some noteworthy effects on catalytic behavior. For

example, meso-bdpp (9a) (Chart 1) gave catalysts with higher

activity than both dppp itself and rac-bdpp (9b). A similar

activity trend was shown by Sesto and Consiglio⁵⁹for CO/

propene copolymerization. Even more dramatic effects on

activity and regioregularity were achieved with the ligands

10a-c for which the meso ligand 10b again gave the most active

and regioselective catalyst. These effects are clearly steric in

origin, and Bianchini’s interpretation stressed the importance

of the spatial distribution of the phenyl groups, with the meso

ligands creating a coordination sphere with steric crowding

concentrated on one face of the Pd complex.

groups, Ph₂PCH₂CH₂PAr^F.¹⁵In situ IR data reported for a

2

catalytic reaction using the ligand with Ar^F) C₆H₂F₃-3,4,5

showed the absence of any terminal ν(CO) absorptions, sug-

gesting that the catalyst resting state is not [Rh(CO)(L-L)I]. It

is thought that acetyl complexes [Rh(L-L)I₂(COMe)] are the

resting states for these systems, the acetyl ν(CO) band being

masked by the solvent.

Despite the low activity for acetic acid formation, diphos-

phines (e.g., dppp) have been shown to be effective for the

rhodium/iodide-catalyzed reductiVe carbonylation of metha-

nol.^14,18

MeOH + CO + H₂f MeCHO + H₂O

(7)

Kinetic studies showed the reaction to be first-order in [Rh-

(dppp)I₂(COMe)] but zero-order in MeI, and it was proposed

that reaction of [Rh(dppp)I₂(COMe)] with H₂determined both

the rate and acetaldehyde selectivity. Consistent with this,

extrapolation of our kinetic data for the dppe system to 140 °C

predicts that neither oxidative addition nor migratory insertion

should be rate-limiting for reductive carbonylation.⁵⁶

Five-coordinate acetyl intermediates [Rh(L-L)I₂(COMe)] are

clearly important in these catalytic carbonylation reactions, and

we find a significant difference in reactivity toward CO for L-L

) dppms and dppe. Baker et al. reported that reaction of 3a

with 30 atm of CO gave two isomers of [Rh(CO)(dppms)I₂-

(COMe)]. We find that the same conversion is achieved by

simply bubbling CO (1 atm) through a solution of 3a in CH₂-

Cl₂for 1 h, as judged by product terminal and acetyl ν(CO)

bands at 2085 and 1685 cm^-1. On standing under CO (1 atm)

for 24 h, the formation of an IR band at 1987 cm^-1indicated

the regeneration of 1a by reductive elimination of acetyl iodide.

By contrast, similar treatment of 3b did not give [Rh(CO)(dppe)-

I₂(COMe)] or any sign of reductive elimination to regenerate

1b. This concurs with the results of Moloy and Wegman,¹⁴who

Earlier studies of CO/alkene copolymerization determined

that, of the R₂P(CH₂)_nPR₂series, the ligand of choice for high

catalytic activity and a high molecular weight polymer was dppp

(i.e., R ) Ph and n ) 3), the order of activity for the R ) Ph

ligands following the order n ) 3 (dppp) > 2 (dppe) > 4

(dppb).⁶⁰This trend was reproduced more recently by Koide et

al., who introduced the concept of a pocket angle (or interior

ligand cone angle) to quantify diphosphine ligand steric

requirements, and suggested that dppp gave the optimum size

(57) Zhu, G.; Cao, P.; Jiang, D.; Zhang, X. J. Am. Chem. Soc. 1997, 119, 1799.

Koenig, K. E.; Sabacky, M. J.; Bachman, G. L.; Christopfel, W. C.;

Barnstorff, H. D.; Friedman, R. B.; Knowles, W. S.; Stults, B. R.; Vineyard,

B. D.; Weinkauff, D. J. Ann. N. Y. Acad. Sci. 1980, 333, 16.

(58) Bianchini, C.; Lee, H. M.; Barbaro, P.; Meli, A.; Moneti, S.; Vizza, F.

New J. Chem. 1999, 23, 929. Bianchini, C.; Lee, H. M.; Meli, A.; Moneti,

S.; Vizza, F.; Fontani, M.; Zanello, P. Macromolecules 1999, 30, 44183.

(59) Sesto, B.; Consiglio, G. J. Am. Chem. Soc. 2001, 123, 4097.

(60) Drent, E.; Budzelaar, P. H. M. Chem. ReV. 1996, 96, 663. Drent, E.; van

Broekhoven, J. A. M.; Doyle, M. J. J. Organomet. Chem. 1991, 417, 235.

(56) The activation parameters in Tables 3 and 4 predict catalytic reductive

carbonylation rates of ca. 15 or 200 mol dm^-3h^-1if, respectively, oxidative

addition or migratory insertion is rate-limiting under the conditions in ref

14. The reported maximum catalytic rate is 6 mol dm^-3h^-1

.

9

J. AM. CHEM. SOC. VOL. 124, NO. 45, 2002 13607

A R T I C L E S

Gonsalvi et al.

and shape of the active site.⁷X-ray crystal structures of [Pd-

(dppe)Cl(CO^tBu)] and [Pd(dppp)Cl(CO^tBu)] showed that the

dppp complex has adjacent hindered quadrants blocking one

face of the coordination plane whereas steric hindrance in the

dppe complex is more dispersed.⁶¹

leads to stronger M(d) f CO(π*) back-bonding, which is

expected to inhibit CO insertion by decreasing the electrophi-

licity of the carbonyl carbon.⁶⁵Indeed, migratory insertion is

often found to be enhanced by coordination of π-acceptor

ligands.⁶⁶On the basis of these arguments, the rapid migratory

insertions observed for 2a and 2c are surprising, since the

observed ν(CO) frequencies suggest greater back-donation from

Rh to CO in 2a (2062 cm^-1) and 2c (2057 cm^-1) compared

with 2b (2076 cm^-1). Thus, the S and O donor atoms appear to

exert a π-pushing effect which raises the Rh d orbital energies

and enhances back-donation to CO. This is expected to inhibit

rather than promote migratory CO insertion.

Most of the experimental evidence suggests the chelate rings

in these systems are relatively robust, at least under the mild

conditions of the model studies. However, the behavior of

complexes 1a-c in the presence of excess iodide anion indicated

that hemilabile behavior might be possible. If one arm of the

chelate ligand were to dissociate (most probably the S or O

arm of dppms or dppmo, respectively⁶⁷), the consequent

reduction in electron density on the Rh center and decrease in

Rh-CO back-donation would be expected to favor methyl

migration (although an unstable 14-electron acetyl species would

be the initial product before chelate ring closure). We are

doubtful that chelate ring opening occurs under mild conditions

(where rapid migratory insertion still occurs for 2a and 2c), but

under the more forcing conditions of the catalytic reactions, this

mechanism cannot be excluded.

Ab Initio Calculations. Since publication of our preliminary

communication,¹⁶Cavallo and Sola have applied density

functional theory (DFT) to some of the same systems.²¹In

addition to simple models, in which hydrogens replaced phenyls,

their study reported calculations with the full-size ligands, using

both hybrid QM/MM and pure QM methods. They concluded

that steric crowding introduced by ligand phenyl substituents

results in less exothermic oxidative addition, but a lower

activation barrier for migratory CO insertion, in agreement with

the conclusions of our experimental study.⁶⁸The experimental

migration insertion barrier for 2a was reproduced extremely well

by the calculations, whereas that for 2b was underestimated.

The theoretical results support the experimental observation that

migratory insertion is much faster for 2a, and do not require

dissociation of the S donor arm of dppms to achieve the lower

activation barrier. Another feature of the DFT study is that the

CO insertion barrier was predicted to be lower for the P,S chelate

complex even in the simple models where hydrogens replaced

phenyls. This suggested to us that there might be an electronic

contribution to the difference in reactivity, in addition to the

steric effects described above. Independently of the study by

Cavallo and Sola, we have carried out ab initio MP2 calculations

Bidentate phosphine ligands are also commonly used in

rhodium-catalyzed hydroformylation reactions. A wide range

of ligands have been tested, with particular recent emphasis on

the control of catalytic activity and product linear:branched (l:

b) ratios by tuning the ligand bite angle.^5,6,62There remains,

however, considerable debate about the precise mechanism by

which a change in ligand bite angle affects the catalytic

properties. In the words of Casey et al.,⁶“The regioselectivity

of hydroformylation is governed by a complex web of electronic

and steric effects that have so far defied unraveling.” A recent

theoretical study of Rh hydroformylation catalysts containing

xantphos (11) and related ligands concluded that “the leading

role in determining the regioselectivity is played by the

diphenylphosphino substituents” and that electronic (or “orbital”)

effects have little influence.⁶³The calculated structures of [Rh-

(CO)(alkene)(diphosphine)H] in that study show that the non-

bonding interactions of phenyl groups with coordinated ligands

are crucial in controlling the product l:b ratio. Interestingly, the

axial phenyl rings in [Rh(CO)(ethene)(benzoxantphos)H] were

found to be virtually parallel when they flanked an axial hydride

ligand, but when the larger CO was placed in this site, the

phenyls moved apart (Figure 2 in ref 62), closely resembling

the ligand distortions we observe (Figure 8), in which alkyl

ligands force the dppms phenyls to move apart. While a number

of effects (e.g., bite angle, donor strength, hemilability) are at

work in determining the behavior of different hydroformylation

catalysts, the steric influence of phosphine ligand substituents

is clearly an important contributor. The “embracing” effect of

phenyl rings has also been invoked to explain the effect of

diphosphine bite angle on the ratio of syn and anti isomers in

[Pd(1-methallyl)(diphosphine)]⁺complexes and thus the influ-

ence on the regiochemistry of allylic alkylation reactions.⁶⁴The

steric effects we have identified are therefore closely related to

those believed to influence a range of catalytic processes.

Electronic Considerations. The structural studies described

above provide one explanation for the observed ligand effects,

involving steric (nonbonding) interactions. The contribution of

electronic (through-bond) interactions must also be considered,

since the IR data show clearly that dppms and dppe differ in

their donor properties. Good donor ligands, while promoting

oxidative addition, normally retard CO insertion, as for the dppe

28

PEt₃systems. One explanation for this is the formation of

stronger metal-alkyl bonds by the more nucleophilic precursors.

Another consideration is that higher electron density on the metal

(61) Dppms has recently been found to give Pd catalysts with moderate activity

for CO/ethene copolymerization, whereas catalysts using Ph₂PCH₂CH₂P-

(S)Ph₂(which forms a six- rather than five-membered chelate ring) give

ca. 50% lower activity (Suranna, G. P.; Mastrorilli, P.; Nobile, C. F.; Keim,

W. Inorg. Chim. Acta 2000, 305, 151).

(62) Kamer, P. C. J.; Reek, J. N. H.; van Leeuwen, P. W. N. M. CHEMTECH

1998, 27. van der Veen, L. A.; Boele, M. D. K.; Bregman, F. R.; Kamer,

P. C. J.; van Leeuwen, P. W. N. M.; Goubitz, K.; Fraanje, J.; Schenk, H.;

Bo, C. J. Am. Chem. Soc. 1998, 120, 11616. Casey, C. P.; Paulsen, E. L.;

Beuttenmueller, E. W.; Proft, B. R.; Matter, B. A.; Powell, D. R. J. Am.

Chem. Soc. 1999, 121, 63.

(63) Carbo, J. J.; Maseras, F.; Bo, C.; van Leeuwen, P. W. N. M. J. Am. Chem.

Soc. 2001, 123, 7630.

(64) van Harren, R. J.; Oevering, H.; Coussens, B. B.; van Strijdonck, G. P. F.;

Reek, J. N. H.; Kamer, P. C. J.; van Leeuwen, P. W. N. M. Eur. J. Inorg.

Chem. 1999, 1237.

(65) Margl, P.; Ziegler, T.; Blo¨chl, P. E. J. Am. Chem. Soc. 1996, 118, 5412.

(66) Cardaci, G.; Reichenbach, G.; Bellachioma, B.; Wassink, B.; Baird, M. C.

Organometallics 1988, 7, 2475. Wright, S. C.; Baird, M. C. J. Am. Chem.

Soc. 1985, 107, 6899. Bellachioma, B.; Cardaci, G.; Macchioni, A.;

Reichenbach, G.; Foresti, E.; Sabatino, P. J. Organomet. Chem. 1997, 531,

227. Kubota, M.; McClesky, T. M.; Hayashi, R. K.; Webb, C. G. J. Am.

Chem. Soc. 1987, 109, 7569. Ghaffar, T.; Adams, H.; Maitlis, P. M.; Sunley,

G. J.; Baker, M. J.; Haynes, A. Chem. Commun. 1998, 1023.

(67) Ab initio calculations indicate that displacement of the S or O donor by

CO in [Rh(CO)(H₂PCH₂P(X)H₂)I] is more favorable by 11 kJ mol^-1(X )

S) or 12 kJ mol^-1(X ) O) than displacement of a PH₂donor in

[Rh(CO)(H₂PCH₂CH₂PH₂)I].

(68) The dppms ligand conformations in the DFT-calculated structures of 1a

and 2a in ref 21 differ somewhat from the conformations found in related

X-ray structures. In particular, the parallel stacked arrangement of axial

phenyls, found consistently by experiment, is not well reproduced.

9

13608 J. AM. CHEM. SOC. VOL. 124, NO. 45, 2002

Effects of Chelate Ligands in Catalysis

A R T I C L E S

Figure 13. Important orbital interactions for migratory CO insertion.

For [Rh(CO)(PH₃)₂I₂Me], with CO trans to I, the migration

barrier is similar those for the dhpms and dhpmo complexes.

These theoretical results suggest that the nature of the donor

atom trans to CO has an influence on the migratory insertion

barrier, but in the sense opposite to that predicted on the basis

of simple arguments involving the strength of back-bonding.

Cavallo and Sola did not offer any rationalization for this effect,

which is also apparent in the results of their calculations. The

orbital interactions important for the migratory insertion process

have been identified previously,⁶⁵as illustrated in Figure 13.

In the octahedral methyl complex the Rh-CH₃bond is formed

2

by overlap of the Rh d_zorbital with a σ-donor orbital from the

methyl group. This orbital also mixes with a p orbital on the

iodine trans to methyl to give some Rh-I bonding character

(A). In the transition state for methyl migration, this orbital is

destabilized due to elongation of the breaking Rh-C bond and

redirection of the local C₃axis of the methyl group toward the

carbonyl carbon (B). This destabilization is the greatest of any

orbital involved in the migratory insertion and is therefore

expected to be a major contributor to the activation barrier. The

destabilization can be partially relieved by interaction with a

CO π* orbital, to initiate C-C bond formation. The calculated

Figure 12. Stationary points on the reaction coordinate for migratory

insertion in [Rh(CO)(dhpms)I₂Me].

Table 6. Calculated Activation Energies ∆E^qand Reaction

Energies ∆E for Migratory Insertion in [Rh(CO)(L-L)I₂Me] or

[Rh(CO)(PH₃)₂I₂Me]

a

∆

E^q,

∆

E,

∆ ,

E^q

orb

1

ligand

geometry

method

kJ mol^-

dhpms

dppms

dhpmo

dhpe

dppe

(PH₃)₂

(PEt₃)₂

CO trans to S

CO trans to P

CO trans to S

CO trans to O

CO trans to P

CO trans to I

MP2

99

110

74

12

20

-36

-56

42

58

106

MP2

differences in orbital energies (∆E^q) E_B- E_A) for the

DFT²¹

MP2

orb

different ligand systems (Table 6) show the smallest destabiliza-

tion for dhpms (with CO trans to S), and substantially larger

destabilizations for dhpe and the dhpms isomer with CO trans

56

104

116

82

61

102

68

98

117

MP2

34

DFT²¹

MP2

-40

-57

38

-38

-12

to P. ∆E^qtakes intermediate values for the (PH₃)₂complex

orb

85

(where CO is trans to I) and the dhpmo complex.

DFT²¹

Inspection of the orbitals concerned in the dhpms system

(Figure 14) reveals substantial mixing of transition-state orbital

B with a sulfur p orbital, corresponding to significant Rh-S

σ-bonding character, as well as Rh-Me and C(Me)-C(carbo-

nyl) bonding. Mixing of this sort is not found for the corre-

sponding orbital in the dhpe or dhpmo systems or when CO is

trans to P in the dhpms system. It appears that a sulfur trans to

CO can stabilize the migratory insertion transition state via this

type of orbital interaction. The magnitude of this effect in the

real system can only be speculated from the computational

studies, and it is likely that the large difference in reactivity of

the dppms and dppe systems results from a combination of steric

and electronic factors. The noteworthy and surprising feature

of the theoretical results is that an electronic effect reinforces

the steric effect of the dppms ligand to promote migratory

insertion, whereas simple bonding arguments predicted that

electronic effects should operate in the opposite sense.

78

^a∆E^q_orbis the difference in orbital energies of orbitals A and B as defined

in the text and Figure 13.

on the simple model systems [i.e., H₂PCH₂P(S)H₂(dhpms), H₂-

PCH₂CH₂PH₂(dhpe), H₂PCH₂P(O)H₂(dhpmo), and PH₃].

The optimized structures of stationary points for migratory

insertion in [Rh(CO)(dhpms)I₂Me] are illustrated in Figure 12.

In each case, the reaction proceeds via a transition state in which

the methyl ligand has moved toward the carbonyl carbon,

accompanied by some reduction of the Rh-C-O bond angle.

The five-coordinate acetyl products all optimized with ap-

proximate square pyramidal geometry. While the absolute values

of activation and reaction energies given by our calculations

differ somewhat from the DFT results (Table 6), the trends

are similar. We find the migratory CO insertion barrier to be

17 kJ mol^-1lower for [Rh(CO)(dhpms)I₂Me] than for [Rh(CO)-

(dhpe)I₂Me] (the DFT value is 8 kJ mol^-1). The barrier for

[Rh(CO)(dhpmo)I₂Me] is marginally higher than for [Rh(CO)-

(dhpms)I₂Me]. For an isomer of [Rh(CO)(dhpms)I₂Me] with

Conclusions

Our detailed mechanistic and kinetic study of the reactions

of [M(CO)(L-L)I] with MeI has yielded considerable informa-

tion regarding steric and electronic effects on oxidative addition

CO trans to P, the migration barrier is raised by 11 kJ mol^-1

.

9

J. AM. CHEM. SOC. VOL. 124, NO. 45, 2002 13609

A R T I C L E S

Gonsalvi et al.

Experimental Section

Materials. All solvents used for synthesis or kinetic experiments

were distilled and degassed prior to use following literature procedures.⁶⁹

Synthetic procedures were carried out utilizing standard Schlenk

techniques. Nitrogen and carbon monoxide were dried through a short

(20 × 3 cm diameter) column of molecular sieves (4 Å) which was

regularly regenerated. Carbon monoxide was also passed through a short

column of activated charcoal to remove any iron pentacarbonyl

impurity.⁷⁰The ligand dppe was purchased from Aldrich and used

without any further purification. The ligands dppms⁷¹and dppmo⁷²and

the complexes [Rh(CO)₂Cl]₂,⁷³[Rh(CO)₂I]₂,²⁶and Bu₄N[Ir(CO)₂I₂]⁷⁴

were synthesized according to literature procedures. Methyl iodide

(Aldrich) was distilled over calcium hydride and stored in foil-wrapped

Schlenk tubes under nitrogen and over mercury to prevent formation

of I₂.

Instrumentation. FTIR spectra were measured using a Mattson

Genesis Series spectrometer, controlled by WINFIRST software running

on a Viglen 486 PC. High-pressure/high-temperature IR spectra were

recorded on a Perkin-Elmer 1710 Fourier transform spectrometer using

a SpectraTech cylindrical internal reflectance (CIR) cell (vide infra).

1

³¹P{¹H} and H NMR spectra were obtained using a Bruker AC250

spectrometer fitted with a Bruker B-ACS60 automatic sample changer

operating in pulse Fourier transform mode, using the solvent as

reference. Elemental analyses were performed using a Perkin-Elmer

2400 elemental analyzer.

Figure 14. Important molecular orbitals A (reactant) and B (transition state)

for migratory insertion in [Rh(CO)(dhpms)I₂Me]. The molecules are viewed

along the equatorial I-Rh-P axis. Note the Rh-S bonding component

which stabilizes the transition-state orbital B.

Synthesis of Rhodium Complexes. (a) [Rh(CO)(dppms)I] (1a).

[Rh(CO)₂I]₂(100 mg, 0.17 mmol) was placed in a 50 cm³round-

bottomed flask, and dissolved in CH₂Cl₂(5 cm³) to which a CH₂Cl₂

solution of dppms (0.23 g in 5 cm³) was added. After 20 min, the

volume of the solvent was reduced and diethyl ether was slowly added.

The yellow precipitate obtained was filtered on a sinter, washed with

ether, and air-dried. The solution was recovered, concentrated, and then

left at -10 °C for 24 h to give a second crop of product. Yield: 219

mg (91%). Anal. Calcd for (C₂₆H₂₂IOP₂RhS): C, 46.5; H, 3.3; I, 19.1.

and migratory insertion. All the chelate ligands employed

accelerate oxidative addition due to their good electron donor

properties, but the promotion of migratory insertion by dppms

was less expected. The promotion of both fundamental steps

by the dppms ligand is unusual, and our investigations suggest

that steric and electronic effects each contribute. X-ray crystal

structures for several Rh and Ir complexes show that the

conformation adopted by the dppms ligand creates steric

congestion around one of the axial coordination sites, favoring

the decrease in coordination number which accompanies methyl

migration. This proposal is supported by recent DFT calcula-

tions. Our own ab initio calculations on model systems suggest

that the placement of a sulfur donor ligand trans to CO also

promotes migratory insertion, via an orbital interaction not

obvious from simple bonding arguments. While hemilability of

the chelate ligands is possible (and probably plays a role in

catalyst deactivation), the model experimental and theoretical

studies are consistent with intact chelate complexes giving rise

to the high catalytic activity of the (dppms)Rh system.

1

Found: C, 46.3; H, 3.3; I, 18.8. IR (CH₂Cl₂): ν(CO)/cm^-11987. H

NMR (CD₂Cl₂): δ 8.10-6.85 (m, 20H, arom), 3.80 (pst, 2H, PCH₂P,

2

²J_HP) 10 Hz). ³¹P{¹H} NMR (CD₂Cl₂): δ 61.9 (dd, PdS, J_RhP) 1

1

2

Hz), 51.3 (dd, RhP, J_RhP) 163 Hz, J_PP) 60 Hz).

(b) [Rh(CO)(dppe)I] (1b). To a solution of [Rh(CO)₂I]₂(50 mg,

0.10 mmol) in toluene (10 cm³) was added a solution of dppe (100

mg, 0.25 mmol) in toluene (10 cm³), dropwise. After 1 h, the product

was obtained as a bright yellow precipitate, filtered, washed with ethanol

and ether, and then air-dried. The solution was recovered, concentrated,

and then left at -10 °C for 24 h. Yield: 70 mg (70%). Anal. Calcd for

(C₂₇H₂₄IOP₂Rh): C, 49.4; H, 3.7; I, 19.3. Found: C, 49.0; H, 3.6; I,

1

19.6. IR (CH₂Cl₂): ν(CO)/cm^-12011. H NMR (CD₂Cl₂): δ 8.10-

7.15 (m, 20H, arom), 3.70-3.55 (m, 4H, PCH₂CH₂P). ³¹P{¹H} NMR

(CD₂Cl₂): δ 70.7 (dd, RhP trans to I, ¹J_RhP) 161 Hz), 53.1 (dd, RhP

1

2

trans to CO, J_RhP) 123 Hz, J_PP) 32 Hz).

Our results explain the high activity of dppms/Rh/iodide-

catalyzed methanol carbonylation and also add quantitative

understanding of the reaction steps in the reductive carbonylation

reactions studied by Wegman and Moloy. They also have

implications for the behavior of other catalytic systems, notably

CO/alkene copolymerization and alkene hydrofomylation, where

the influence of chelating diphosphine ligands is crucial. The

“embracing” effect of the ligand phenyl groups of dppms models

the steric effects of certain diphosphines (e.g., xantphos) used

in hydroformylation, and suggests a mechanism by which the

ligand bite angle can influence the behavior of the active site.

Although similar interpretations have been given to explain

catalytic behavior in various systems, quantitative reactivity data

for individual reactions from a catalytic cycle, as we report here,

are much more scarce.

(c) [Rh(CO)(dppmo)I] (1c). [Rh(CO)₂I]₂(57 mg, 0.1 mmol) was

dissolved in a mixture of CH₂Cl₂(2 cm³) and toluene (5 cm³). A solution

of dppmo (40 mg, 0.2 mmol) in toluene (5 cm³) was then added

dropwise. After 30 min, the product was recovered by filtration as a

yellow precipitate, washed with ether, and dried under vacuum. Yield:

55 mg (84%). Anal. Calcd for (C₂₆H₂₂I₂O₂P₂Rh): C, 47.5; H, 3.3; I,

19.3. Found: C, 47.7; H, 3.4; I, 18.9. IR (CH₂Cl₂): ν(CO)/cm^-11983.

¹H NMR (CD₂Cl₂): δ 7.75-7.20 (m, 20H, arom), 3.35 (pst, 2H, PCH₂P,

(69) Perrin, D. D.; Armerego, W. L. F.; Perrin, D. R. Purification of Laboratory

Chemicals, 3rd ed.; Pergamon Press: Oxford, 1988.

(70) Haynes, A.; Ellis, P. R.; Byers, P. K.; Maitlis, P. M. Chem. Br. 1992, 28,

517.

(71) Grim, S. O.; Mitchell, J. D. Synth. React. Inorg. Met.-Org. Chem. 1974, 4,

221.

(72) Grim, S. O.; Satek, L. C.; Tolman, C. A.; Jesson, J. P. Inorg. Chem. 1975,

14, 656.

(73) McCleverty, J.; Wilkinson, G. Inorg. Synth. 1966, 8, 214.

(74) Forster, D. Inorg. Nucl. Chem. Lett. 1969, 5, 433.

9

13610 J. AM. CHEM. SOC. VOL. 124, NO. 45, 2002

Effects of Chelate Ligands in Catalysis

A R T I C L E S

2

²J_HP) 13 Hz). ³¹P{¹H} NMR (CD₂Cl₂): δ 63.1 (dd, PdO, J_RhP) 4

Hz), 41.9 (dd, RhP, J_RhP) 166 Hz, J_PP) 33 Hz).

(both J_HP) 4 Hz). ³¹P{¹H} NMR (CD₂Cl₂): δ 65.0 (d, PdS), 18.7 (d,

IrP, J_PP) 36 Hz (major isomer)), 56.4 (d, PdS), 3.7 (d, IrP, J_PP)

1

2

23 Hz (minor isomer)). The two isomers identified by NMR spectros-

copy were present in a ratio of ca. 3:1. An X-ray crystal structure of

the major isomer was obtained (vide infra).

(d) [Rh(dppms)I₂(COMe)] (3a). 1a (50 mg) was dissolved in 10

cm³of CH₂Cl₂under nitrogen. Methyl iodide (0.5 cm³, excess) was

added, and the solution was stirred at room temperature for 90 min

and then concentrated in vacuo until cloudiness was observed. The

product was precipitated as an orange powder by addition of diethyl

ether after cooling to -10 °C, and recrystallized from CH₂Cl₂. Yield:

45 mg (75%). Anal. Calcd for (C₂₇H₂₅I₂OP₂RhS): C, 39.7; H, 3.1; I,

31.1. Found: C, 39.2; H, 2.9; I, 31.4. IR (CH₂Cl₂): ν(CO)/cm^-11701.

¹H NMR (CD₂Cl₂): δ 7.80-6.90 (m, 20H, arom), 4.88 (ddd, 1H,

(d) [Ir(CO)(dppe)I₂Me] (5b). 4b (230 mg, 0.31 mmol) was

dissolved in CH₂Cl₂(20 cm³) and stirred under N₂. Methyl iodide (0.5

cm³, excess) was added, and the solution was stirred for 60 min. After

the solvent had been reduced by half, hexane (10 cm³) was added and

the flask was cooled at -10 °C for 24 h. The product was recovered

as a cream-white precipitate. Yield: 220 mg (81%). Anal. Calcd for

(C₂₈H₂₇I₂IrOP₂): C, 37.9; H, 3.0; I, 28.6. Found: C, 37.7; H, 3.1; I,

2

PCHH′P′, J_HP) 8 Hz, J_HP′) 4 Hz), 3.58 (ddd, 1H, PCHH′P′, J_H′P

) 5 Hz, ²J_H′P′) 2 Hz, ²J_HH′) 13 Hz), 3.16 (s, 3H, COCH₃). ³¹P{¹H}

NMR (CD₂Cl₂): δ 58.2 (dd, PdS, ²J_RhP) 3 Hz), 55.1 (dd, RhP, ¹J_RhP

1

28.5. IR (CH₂Cl₂): ν(CO)/cm^-12056. H NMR (CD₂Cl₂): δ 8.01-

7.58 (m, 20H arom), 3.07 (m, 4H, PCH₂CH₂P), 0.47 (dd, 3H, IrCH₃,

2

³J_HP) 8, 5 Hz). ³¹P NMR (CD₂Cl₂): δ 24.5 (d), -3.0 (d, J_PP) 5

) 137 Hz, J_PP) 46 Hz). A crystal suitable for an X-ray diffraction

Hz). Spectroscopic data are in agreement with published data.²⁵

study was obtained by recrystallization from CH₂Cl₂.

(e) [Rh(dppe)I₂(COMe)] (3b). 1b (50 mg) was dissolved in 10 cm³

of CH₂Cl₂under nitrogen, then methyl iodide (0.5 cm³, excess) was

added, and the mixture was stirred at room temperature for 3 h. The

solution was concentrated in vacuo until cloudiness was observed. The

product was obtained as a yellow precipitate after addition of diethyl

ether and cooling at -10 °C for 24 h. Yield: 42 mg (70%). Anal.

Calcd for (C₂₈H₂₇I₂OP₂Rh): C, 42.1; H, 3.4; I, 31.8. Found: C, 41.9;

H, 3.3; I, 32.0. IR (CH₂Cl₂): ν(CO)/cm^-11711. ¹H NMR (CD₂Cl₂): δ

7.82-7.73 and 7.52-7.23 (m, 20H, arom), 3.17-2.96 2.28-2.07 (m,

4H, PCH₂CH₂P), 2.65 (s, 3H, COCH₃). ³¹P{¹H} NMR (CD₂Cl₂): δ

70.5 (d, ¹J_RhP) 139 Hz). Single crystals suitable for X-ray diffraction

(vide infra) were obtained by recrystallization from CH₂Cl₂/diethyl

ether.

(e) [Ir(CO)(dppms)I₂Et] (6). To 4b (50 mg, 0.07 mmol) was added

EtI (0.5 cm³, excess) in CH₂Cl₂(10 cm³), and the solution was stirred

for 24 h at room temperature. After reducing the solvent by half in

vacuo, diethyl ether was added to give an orange precipitate which

was recovered after cooling for 24 h at -10 °C and recrystallized from

CH₂Cl₂. Yield: 34 mg (57%). Anal. Calcd for (C₂₈H₂₇I₂IrOP₂S): C,

36.6; H, 3.0; I, 27.6. Found: C, 36.2; H, 2.8; I, 28.1. IR (CH₂Cl₂):

1

ν(CO)/cm^-12036. H NMR (CD₂Cl₂): δ 8.05-6.93 (m, arom), 4.35

(ddd, PCHH′P′, major isomer, ²J_HH′) 15 Hz, ²J_HP) 11 Hz, ²J_HP′) 4

Hz), 4.08 (ddd, PCHH′P′, major isomer, ²J_HH′) 15 Hz, ²J_H′P) 11 Hz,

²J_H′P′) 5 Hz), 3.42 (q, IrCH₂CH₃, major isomer), 1.13 (t, IrCH₂CH₃,

3

major isomer, J_HH) 7 Hz). ³¹P{¹H} NMR (CD₂Cl₂): δ 65.2 (d, Pd

2

S), 19.2 (d, IrP, J_PP) 36 Hz (major isomer)), 55.1 (d, PdS), 3.3 (d,

Synthesis of Iridium Complexes. (a) [Ir(CO)(dppms)I] (4a). A

solution of Bu₄N[Ir(CO)₂I₂] (472 mg, 0.63 mmol) in toluene (10 cm³)

was added slowly to a solution of dppms (264 mg, 0.63 mmol) in a

mixture of toluene (10 cm³) and CH₂Cl₂(5 cm³) under N₂, and the

resulting solution was stirred at room temperature for 1 h. The product

was recovered as a yellow precipitate, filtered, washed with a mixture

of EtOH/ⁱPrOH (1:2) followed by diethyl ether and hexane, and dried

in vacuo. Yield: 390 mg (81%). Anal. Calcd for (C₂₆H₂₂IIrOP₂S): C,

40.9; H, 2.9; I, 16.6. Found: C, 40.6; H, 2.9; I, 16.9. IR (CH₂Cl₂):

ν(CO)/cm^-11972. ¹H NMR (CD₂Cl₂) δ 7.70-7.10 (m, 20H, arom), δ

3.75 (pst, 2H, PCH₂P, ²J_HP) 10 Hz). ³¹P{¹H} NMR (CD₂Cl₂): δ 66.3

IrP, ²J_PP) 25 Hz (minor isomer)). The two isomers identified by NMR

spectroscopy were present in a ratio of ca. 3:1. An X-ray crystal

structure of the major isomer was obtained (vide infra).

(f) [{Ir(CO)(dppms)(µ-I)Me}₂][SO₃CF₃]₂(7). 4a (50 mg, 0.07

mmol) was dissolved in CH₂Cl₂(3 cm³) and the solution stirred under

N₂. Methyl triflate (8 µL, 0.07 mmol) was added by syringe, and the

solution was stirred for 30 min at room temperature. The product was

recovered as a cream-white precipitate, and recrystallized as yellow

blocks from CH₂Cl₂. Yield: 44 mg (72%). Anal. Calcd for (C₂₈H₂₅F₃I₂-

IrO₄P₂S₂): C, 36.3; H, 2.7; I, 13.7. Found: C, 36.1; H, 2.9; I, 13.7. IR

1

(CH₂Cl₂): ν(CO)/cm^-12058. H NMR (CD₂Cl₂): δ 8.35-7.95 (m,

2

(d, PdS) 25.8 (d, IrP, J_PP) 54 Hz).

20H, 7.75-7.05 arom), 5.01, 4.78, 4.65, 4.42 (each ddd, total 2H,

PCH₂P for two isomers), 1.16, 1.11 (each d, total 3H, IrCH₃for two

isomers, ³J_HP) 3 Hz). ³¹P{¹H} NMR (CD₂Cl₂): δ 65.1 (d, PdS), 17.5

(d, IrP, ²J_PP) 28 Hz), 64.3 (d, PdS), 17.0 (d, IrP, ²J_PP) 29 Hz), 63.4

(d, PdS) 16.0 (d, IrP, ²J_PP) 25 Hz), 62.2 (d, PdS), 15.9 (d, IrP, ²J_PP

) 26 Hz). The four isomers identified by NMR spectroscopy were

present in a ratio of ca. 3:3:1:1. An X-ray crystal structure for an iodide-

bridged dimer was obtained (vide infra).

(b) [Ir(CO)(dppe)I] (4b). This synthesis followed the method of

Fisher and Eisenberg.¹⁷Bu₄N[Ir(CO)₂I₂] (400 mg, 0.5 mmol) was

dissolved in THF (10 cm³) and the resulting solution slowly added to

a solution of dppe (200 mg, 0.5 mmol) in THF (15 cm³) under N₂. The

solution was stirred at moderate reflux for 1 h and then allowed to

cool to room temperature. Ethanol (18 cm³) was added, and a vigorous

stream of N₂was bubbled through the solution until an orange

precipitate was obtained. The product was filtered and dried under

vacuum. Yield: 280 mg (75%). Anal. Calcd for (C₂₇H₂₄IIrOP₂): C,

43.5; H, 3.2; I, 17.0. Found: C, 43.3; H, 3.4; I, 17.0. IR (CH₂Cl₂):

X-ray Structure Determinations. Data were collected on either a

Bruker Smart CCD area detector with an Oxford Cryostream 600 low-

temperature system (complexes 5a, 6, and 7) or a Siemens P4

diffractometer (complexes 3a, 3b, and 3c-NCMe), in each case using

Mo KR radiation (λ ) 0.71073 Å). The structures were solved by direct

methods and refined by full-matrix least-squares methods on F².

Hydrogen atoms were placed geometrically and refined using a riding

model (including torsional freedom for methyl groups). Complex

scattering factors were taken from the SHELXL program packages.^75,76

Crystallographic data are summarized in Table 7 for complexes 3b, 6,

1

ν(CO)/cm^-11994. H NMR (CD₂Cl₂): δ 8.00-7.10 (m, 20H, arom),

2.45-2.04 (m, 4H, PCH₂CH₂P). ³¹P NMR (CD₂Cl₂): δ 49.6 (d), 46.0

2

(d, J_PP) 14 Hz).

(c) [Ir(CO)(dppms)I₂Me] (5a). 4a (390 mg, 0.51 mmol) was

dissolved in CH₂Cl₂(20 cm³) and the solution stirred under N₂. Methyl

iodide (0.5 cm³, excess) was added, and the solution was stirred for 30

min. Diethyl ether was added until cloudiness appeared, and the flask

was left at -10 °C for 24 h. The product was recovered as a cream-

white precipitate, and recrystallized as yellow blocks from CH₂Cl₂.

Yield: 336 mg (80%). Anal. Calcd for (C₂₇H₂₅I₂IrOP₂S): C, 35.8; H,

2.8; I, 28.0. Found: C, 35.4; H, 2.7; I, 28.3. IR (CH₂Cl₂): ν(CO)/

(75) SHELXL97, An integrated system for solving and refining crystal structures

from diffraction data: Sheldrick, G. M., University of Gottingen, Gottingen,

Germany, 1997. SHELXTL, An integrated system for solving and refining

crystal structures from diffraction data (Revision 5.1): Sheldrick, G. M.,

Bruker AXS Ltd., Madison, WI.

(76) SHELXL93, An integrated system for solving and refining crystal structures

from diffraction data: Sheldrick, G. M., University of Gottingen, Gottingen,

Germany, 1993.

1

cm^-12041. H NMR (CD₂Cl₂): δ 7.95-6.71 (m, arom), 4.90 (ddd,

2

PCHH′P′, major isomer, J_HH′) 15 Hz, J_HP) 5 Hz, J_HP′) 2 Hz),

4.27 (ddd, PCHH′P′, major isomer, ²J_HH′) 15 Hz, ³J_H′P) 5 Hz, ³J_H′P′

) 3 Hz), 1.64 (d, IrCH₃, minor isomer), 1.25 (d, IrCH₃, major isomer)

9

J. AM. CHEM. SOC. VOL. 124, NO. 45, 2002 13611

A R T I C L E S

Gonsalvi et al.

Table 7. Summary of Crystallographic Data for Complexes 3b, 6,

and 8

external optical bench coupled to the spectrometer. In a typical

experiment, 14 mg of the iridium methyl complex [Ir(CO)(L-L)I₂Me]

(L-L ) dppms or dppe) was dissolved in 8 cm³of the required solvent

and the resulting solution filtered into the CIR cell. The cell head was

fitted to the body of the cell, ensuring that the stirrer was firmly

tightened into the cell head. The vent lines, stirrer, and gas inlet from

the cylinder were then fitted, tightened, and flushed out four times with

N₂. The cell was flushed with N₂at least four times, the last two with

the stirrer on, and then heated to the required temperature. The cell

was then pressurized with CO, and spectra were recorded over the range

2200-1600 cm^-1at regular intervals under computer control.

Computational Details. Ab intio quantum mechanical calculations

using second-order Møller-Plesset (MP2) theory were performed using

the Gaussian94 suite of programs.⁷⁸Geometries of stationary points

were optimized using the Berny algorithm⁷⁹as implemented in

Gaussian94. Transition states were located by employing a reaction

coordinate method followed by a transition-state search based on the

Berny algorithm, and characterized by frequency calculations to give

Hessians with a single negative eigenvalue. We employed the LANL2DZ

Gaussian basis set developed by Hay and Wadt,⁸⁰which uses a semicore

double-ú contraction scheme for the heavy elements Rh, I, S, and P,

with the light atoms C, O, and H being described by the split-valence

Dunning 9-5V all-electron basis. Molecular orbitals were visualized

(e.g., Figure 14) using Molekel software.⁸¹

3b

6

7‚CH₂Cl₂

empirical

formula

fw

cryst syst

space group

color

a (Å)

b (Å)

c (Å)

R (deg)

â (deg)

γ (deg)

temp (K)

Z

C₂₈H₂₇I₂OP₂Rh C₂₈H₂₇I₂IrOP₂S C₂₉H₂₇Cl₂F₃IIrO₄P₂S₂

798.15

triclinic

P1h

919.50

monoclinic

P2₁/n

1012.57

monoclinic

P2₁/n

yellow

9.155(4)

10.436(4)

14.964(7)

92.53(4)

93.83(2)

100.34(3)

293(2)

2

yellow

10.2276(10)

16.9486(17)

17.2942(16)

90

10.7366(9)

15.6551(13)

20.1261(16)

90

102.196(2)

90

94.248(2)

90

150(2)

4

final R indices R1 ) 0.0456

R1 ) 0.0405

R1 ) 0.0413

[I > 2σ(I)] wR2 ) 0.1117 wR2 ) 0.1001 wR2 ) 0.0974

R indices

(all data)

GOF

R1 ) 0.0653

wR2 ) 0.1659 wR2 ) 0.1074 wR2 ) 0.1098

1.101 1.079 1.066

R1 ) 0.0477,

R1 ) 0.0554

and 7, and full listings of data are given in the Supporting Information.

The structures of 3a and 5a were presented in a preliminary com-

munication.¹⁶

Acknowledgment. This work was supported by BP Chemi-

cals Ltd. (studentship to L.G.), the University of Sheffield, and

The Royal Society. We thank Dr. D. B. Cook and Dr. P. W.

Badger for advice on ab initio calculations, Prof. Luigi Cavallo

for helpful discussions, and Prof. Rich Eisenberg for providing

access to data prior to publication.

Kinetic Experiments. Samples for kinetic runs were prepared by

placing the required amount of freshly distilled methyl iodide in a 5

cm³graduated flask which was then made up to the mark with the

solvent of choice (usually CH₂Cl₂). A portion of this solution was used

to record a background spectrum. Another portion (typically 500 µL)

was added to the solid metal complex (typically 7-8 µmol) in a sample

vial to give a reaction solution with a complex concentration of ca. 15

mM. A portion of the reaction solution was quickly transferred to the

IR cell, and the kinetic experiment was started. To obtain pseudo-first-

order conditions, at least a 10-fold excess of MeI was used, relative to

the metal complex. The IR cell (0.5 mm path length, CaF₂windows)

was maintained at constant temperature throughout the kinetic run by

a thermostated jacket. Spectra were scanned in the metal carbonyl ν(CO)

region (2200-1600 cm^-1) and saved at regular time intervals under

computer control. After the kinetic run, absorbance vs time data for

the appropriate ν(CO) frequencies were extracted and analyzed off-

line using Kaleidagraph curve-fitting software. The decays of the bands

due to 1a (1987 cm^-1), 1b (2011 cm^-1), 1c (1983 cm^-1), 4a (1972

cm^-1), and 4b (1994 cm^-1) were all well fitted by exponential curves

with correlation coefficients g0.999, to give pseudo-first-order rate

constants. Each kinetic run was repeated at least twice to check

reproducibility, the k_obsvalues given being averaged values with

component measurements deviating from each other by e5%.

For kinetic experiments carried out under a pressure of CO, a high-

pressure CIR cell was used to record in situ IR spectra. The CIR cell

comprised a batch autoclave (Parr) modified (by SpectraTech) to

accommodate a crystalline silicon CIR rod as described by Moser et

al.⁷⁷Spectra were recorded using a Perkin-Elmer 1710 FTIR spec-

trometer fitted with an MCT detector. The cell was mounted in an

Supporting Information Available: Tables of kinetic data,

details of crystal structure determinations (including ORTEP

plots), and tables of Cartesian coordinates for optimized

structures from ab initio calculations (PDF) and tables of crystal

data (CIF). This material is available free of charge via the

Internet at http://pubs.acs.org.

JA0176191

(77) Moser, W. R. In Homogeneous Transition Metal Catalysed Reactions;

Moser, W. R., Slocum, D. W., Eds.; Advances in Chemistry Series, Vol.

230; American Chemical Society: Washington, DC, 1992; pp 3-18.

(78) Gaussian 94, Revision D.4: Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.;

Gill, P. M. W.; Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.;

Petersson, G. A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M.

A.; Zakrzewski, V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.;

Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala,

P. Y.; Chen, W.; Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts,

R.; Martin, R. L.; Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart,

J. P.; Head-Gordon, M.; Gonzalez, C.; Pople, J. A., Gaussian, Inc.,

Pittsburgh, PA, 1995.

(79) Schlegel, H. B. J. Comput. Chem. 1982, 3, 214.

(80) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270. Hay, P. J.; Wadt,

W. R. J. Chem. Phys. 1985, 82, 299. Wadt, W. R.; Hay, P. J. J. Chem.

Phys. 1985, 82, 284.

(81) MOLEKEL (4.0): Flukiger, P.; Luthi, H. P.; Portmann, S.; Weber, J., Swiss

Center for Scientific Computing, Manno, Switzerland, 2000.

9

13612 J. AM. CHEM. SOC. VOL. 124, NO. 45, 2002

Article Doi

Steric and electronic effects on the reactivity of Rh and Ir complexes containing P-S, P-P, and P-O ligands. Implications for the effects of chelate ligands in catalysis

DOI: 10.1021/ja0176191

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja0176191

Full text of DOI:10.1021/ja0176191

Products guided by the article

R&D Labs maybe for 477849-06-2

Relevant to this article

Hot Product