Organometallic Chemistry Volume 35

By Ian Fairlamb, Jason Lynam

The Royal Society of Chemistry

Copyright © 2009 The Royal Society of Chemistry
All rights reserved.
ISBN: 978-0-85404-358-3

Contents

Preface Ian Fairlamb and Jason Lynam, 7,
Ligand electronic effects in homogeneous catalysis using transition metal complexes of phosphine ligands Matthew L. Clarke and Jamie J. R. Frew, 19,
Abnormal NHCs: coordination, reaction chemistry and catalytic applications Martin Albrecht and Kingsley J. Cavell, 47,
Application of phosphine ligands in organic synthesis Luis A. Adrio and King Kuok (Mimi) Hii, 62,
Recent developments in aryl–aryl bond formation by transition metal-catalysed C–H activation Gerard P. McGlacken, 93,
Alkali/coinage metals — organolithium, organocuprate chemistry Joanna Haywood and Andrew E. H. Wheatley, 130,
Group 2 (Be–Ba) and Group 12 (Zn–Hg) Felipe García and Dominic S. Wright, 162,
Scandium, yttrium and the lanthanides John G. Brennan and Andrea Sella, 183,
Developments in multiply bonded Group 14 organometallic chemistry Richard A. Layfield, 224,
Organo-transition metal cluster complexes Mark G. Humphrey and Marie P. Cifuentes, 234,

CHAPTER 1

Ligand electronic effects in homogeneous catalysis using transition metal complexes of phosphine ligands

Matthew L Clarke and Jamie J. R. Frew

DOI: 10.1039/b801377m

This report provides a discussion on the relative importance of electronic and steric effects in catalysis. More specificially, electronic effects that arise from phosphorus based supporting ligands in palladium catalysed cross-coupling and rhodium catalysed hydroformylation are reviewed and analysed. The report identifies certain trends that emerge from the literature data, and also points out that these effects can sometimes be subtle, or an indirect effect on the catalytic reaction rather than changing the rate of any of the key steps in the catalytic cycle directly.

1. Introduction

The use of homogeneous catalysis represents one of the most efficient and important methods for carrying out chemical transformations. There are many processes in industry (with more being developed every year) that utilise organo-transition metal catalysts. The development of selective chemical processes generally requires modifying ligands on the transition metal centre. Although P, N, O, S, Se, C and Te ligands are all known, by far the most important class in catalysis are phosphorus based ligands. These ligands have a unique ability to stabilise metals in several oxidation states and geometries and, more importantly, they can be tuned to radically change the reactivity of a catalyst. Simple changes to the structure of a phosphorus ligand can completely alter the product distribution, activity, regiochemistry or enantioselectivity of a transition metal catalysed reaction. Indeed, there are many processes that do not work at all unless the correct choice of ligand is made. This review aims to provide the reader with a discussion on the significance of electronic effects in transition metal catalysis. A comprehensive treatment of this subject would require several hundred pages. On the other hand, a report on very recent examples would not enable the reader with a working knowledge of the magnitude and nature of such effects. The purpose is to highlight how electronic effects can often be quite subtle and in some cases an indirect effect on the reactivity, to discuss some of the fundamental processes that can be subject to electronic effects, and to evaluate the significance of these effects on catalytic reactions. To illustrate these points, the discussion is mainly focussed on transition metal phosphine complexes, and is further restricted to two main types of catalytic reaction, Pd catalysed cross-coupling and Rh catalysed hydroformylation.

2. Measuring ligand properties

In an attempt to rationalise what features can be tuned within phosphorus ligands, chemists have classified according to their co-ordinate modes (e.g. monodentate, bidentate, tridentate, tetradentate and hemilabile) and stereo-electronic properties. The stereo electronic properties may be quantified in the following ways:

Measuring steric effects: Cone angle (θ)

The cone angle (θ), introduced by Tolman, is a quantitative measure of the steric effect of the substituents surrounding phosphorus. y for symmetrical ligands is the apex angle of a cone centred 2.28 Å from the centre of the P atom that just touches the outer Van der Waals radii of the outermost atoms of the substituents (Fig. 1). Other methods to quantify steric effects have been used, but the simplicity of the Tolman cone angle has ensured that it remains a popular method for comparing mono-phosphines. In bidentate ligands, the bite angle of the ligand is at least as an important factor as cone angle, since the bite angle can influence both steric and electronics. The term bite angle can refer to any measurement of the P-M-P angle, but the concept of a natural bite angle as used by Casey and extensively developed by Kamer and Van Leeuwen is based on molecular mechanics calculations and refers to the preferred bite angle of a ligand binding to a point in space that itself has no preference on the geometry observed. This is a useful measure, since one ligand can exert a wide range of crystallographically determined bite angles depending on what the metal’s preferred valence angles are. Steric effects and bite angles can be as important if not more important than electronic effects in deciding the properties of phosphorus ligands, since in addition to causing ligand-reactant repulsive interactions, a change in steric effect can also effect a change on the type of complex formed, for example from bis-ligated to mono-ligated, and changes in both bite angle and sterics can have an indirect electronic effect by altering the bond angles away from the geometry preferred by the metal centre.

Quantifying electronic properties

The quantitative measurement of the electronic effect that various ligands have is made by observing differences in the CO stretching frequencies of metal complexes such as [Ni(CO)3L] where L is a phosphine ligand. The CO stretching frequency is sharp and readily measurable, making it a convenient and accurate measure of ligand effects on the metal centre. Due to the extreme toxicity and volatility of nickel carbonyl, most quantitative measurements are now made using rhodium(I) carbonyl complexes of type trans- [Rh(L)2(CO)Cl] for monodentate phosphines. A compilation for some monophosphines is shown in Table 1 (Fig. 2).

As can be seen from the table, the use of trans- [Rh(L)2(CO)Cl] complexes as probes to measure electronic effects can be very useful in understanding the properties of new phosphines. A variety of other methods using different metal carbonyl complexes, measurement of reduction potentials, or theoretical methods have been proposed, but since the synthesis of the Rh complexes can be very straightforward (and is often quantitative), and gives crystalline complexes that may be useful for X-ray analysis or as catalysts, the above is recommended as the method of choice. Measurements of v(CO) should generally be carried out in a relatively non-coordinating solvent such as CH2Cl2. Nujol mulls can be less accurate due to solid state packing effects. Although we have generally found that KBr discs give the same values as solution measurements, the latter are recommended with several control measurements of known complexes to confirm accuracy, and it is also noted here that in situ synthesis and IR monitoring of the solutions is not recommended for making measurements of electronic properties. pKa values of phosphines can also be recorded, and generally show a good agreement with the donor strengths proposed in Table 1, but have the disadvantage of being more time-consuming, can be difficult for acid sensitive ligands, and will be less effective for the electron deficient π-acceptor phosphines due to different nature of M–P and H–P bonding. A few relevent pKa values are P(OPh)3 = -2.0, P(4-ClC6H4)3 = 1.03, PPh3 = 2.73, P(4-MeOC6H4)3 = 4.59, PEt3 = 8.69, PCy3 = 9.70 and But3P = 11.40]. The last phosphine does not form the same type of Rh carbonyl complex, so is not listed in Table 1. For bidentate phosphines, cis-[Mo(L) (CO4)] complexes are more often used to determine electronic donor characteristics, perhaps due to the fact that the reaction of diphosphines with [Rh(CO2)]2 (μ-Cl)2 can give either cis complexes of type [cis-Rh(L)(CO)Cl] or dimeric species in which the phosphines adopts a bridging chelate mode trans to each other. Only strongly chelating ligands give the cis monomeric species and relatively few examples are known.

Palladium-catalysed cross-coupling

Palladium-catalysed cross-coupling procedures are now amongst the most important methods for C–C and C-heteroatom bond formation. Key fundamental organometallic reactions that can take place during cross-coupling catalysis include oxidative addition, transmetalation, isomerisation, outer-sphere nucleophilic attack and reductive elimination. In the following pages, the degree to which electronic effects impact on these key organometallic processes will be discussed.

Oxidative addition

It is firmly established that metal complexes of more basic phosphines will undergo oxidative addition more rapidly. For example, addition of MeI to trans-]Ir(L)2(CO)Cl] where L = P(4-X-C6H4)3 does show a predictable and substantial increase in rate as X becomes more electron donating. This electronic effect has been determined to be more significant in Me–I oxidative addition than oxidative addition of hydrogen to the same complexes, showing that the magnitude of electronic effects will be highly dependent on the reaction being studied. It has been proposed that there is a second electronic effect in this specific reaction. Complexes derived from para-substituted triarylphosphines are surprisingly more reactive than very basic, but isosteric trialkyl phosphines such as tri-isobutylphosphine. This has led to the proposal of a second electronic parameter related to the number of aryl substituents a phosphine possesses. The magnitude of the effect of this aryl-electronic parameter will vary depending on the reaction to be studied, but is very significant for oxidative addition of either MeI or hydrogen to trans- Ir(L) 2(CO)Cl]. To the best of our knowledge, the magnitude of the aryl effect on oxidative addition to Pd(0) has not been determined.

Amatore and co-workers have found that there is a strong linear correlation between the rate of oxidative addition of phenyl iodide to a series of palladium triarylphosphine complexes and determined the Hammett values of the substituents in the para positions of the aryl rings (Fig. 3).

The study above demonstrates direct electronic effects on oxidative addition. Given that PPhMe2 was also found to be more active than PPhMe2 and PPh3, it would seem likely that the aryl effect discussed above is less pronounced in oxidative additions to Pd(0). A number of other factors can play a significant role in oxidative addition, and it is worth appreciating these before making any conclusions about phosphine ligand effects based on catalytic data alone. For example, it is established that halides, acetate and other anions can play a role in the oxidative addition step. Electrochemical measurement of the oxidative addition of phenyl iodide to [Pd(PPh3)3] was more than an order of magnitude slower than the same reaction carried out using electrochemically reduced [Pd(PPh3)2Cl2], which is thought to react via [Pd(0)(PPh3) 2Cl]-. If this latter reaction is carried out in the presence of various cations that can start to sequester chloride ions, rates increase again.

Pd(0) species generated from phosphines and Pd(II) acetate also tend to form acetate-ligated species, [Pd(0)L2(OAc)] – which can be more reactive than some Pd(0) precursors. These effects are predominantly caused by the differing concentration of the more reactive co-ordinatively saturated species, [PdL2]. However, it has been proposed that direct oxidative addition on the anionic Pd(0) species could also occur under some conditions, thus the presence of an anionic ligand on Pd(0) could have a direct electronic effect on the reaction. Pronounced electronic effects on the reactivity of Pd-alkene precursors have also been observed. Many cross-coupling reactions make use of Pd2dba3 (dba = E,E-dibenzylideneacetone) as a Pd(0) precursor that can be combined with a phosphine (or N-heterocyclic carbene) ligand. It has been established that dba binds quite strongly to Pd(0), making the major Pd(0) species in many reactions [Pd(0)L2(dba)], from which dba must dissociate prior to oxidative addition. Fairlamb and co-workers have studied a range of para-substituted dba ligands as their Pd(0) complexes and shown that the more electron-donating dba analogues give significantly more active catalysts. It seems likely that this is a pure electronic effect, since the series were para-substituted, but it is also an indirect one. The electron rich dba ligands have lower backbonding ability and thus will dissociate more readily from the palladium revealing the true active species in oxidative addition, [Pd(0)L2]. Given that one can imagine that such well defined Pd(0) catalysts could behave more reproducibly in catalysis than Pd(II) catalysts reduced in situ, this is an interesting approach to maximising the activity of a wide variety of Pd catalysts (Fig. 4).

In addition to the various dissociation equilibria that must be controlled to maximise the concentration of the active catalysts, the greater nucleophilicity of Pd(0) complexes of electron rich alkyl phosphines has significant implication in cross-coupling chemistry. Even solutions containing a high concentration of the desired species [Pd(0)L2] L = triphenylphosphine, do not activate aryl chlorides or most alkyl halides. The use of highly electron-donating alkyl phosphines is generally required to activate less reactive C–X bonds.

A key study by Milstein and co-workers demonstrated that the bis-chelate species [Pd(0)(dippe)2] (dippe = 1,2-bisdiisopropylphosphine ethane would oxidatively add chlorobenzene at 80 °C, in contrast to the less electron rich bis-chelate [Pd(0)(dppp)2] which is essentially inactive below 150 °C. The most active of Pd complexes of bidentate ligands investigated in this study was [Pd(0)(dippp)2] (dippp = 1,3-bis-diisopropylphosphino-propane), which is actually a three co-ordinate complex with one end of the phosphine not co-ordinated. This complex oxidatively adds chlorobenzene at 60 °C. As in many cases, it is not possible to completely decouple electronic and steric effects. In particular, it seems clear that the ability of one bidentate phosphine to dissociate has a significant effect on the rate of oxidative addition. Nonetheless, the results are consistent with phosphine donor strength playing a significant role in facilitating oxidative addition of more unreactive Ar–X bonds (Fig. 5(a)).

An early study by Osborn and co-workers demonstrated that in contrast to PPh3/Pd(0), [Pd(Cy3)2] oxidatively adds chlorobenzene at 60 °C. However, it is already clear from this work that steric effects might be even more profound, since electron rich PEt3/Pd(0) did not take part in this oxidative addition. This communication did not describe what source of Pd(0)/PEt3 was used, but it can be envisaged that the steric effect could in part cause the formation of different Pd(0) species such as [Pd(PEt3)3], which are less reactive in oxidative addition. This work also notes that But3P/Pd(0) also did not undergo oxidative addition under the same conditions as [Pd(PCy3)2], despite being a more electron-donating ligand.

Following up from this work, electron-donating phosphines are firmly established as a requirement for carrying out cross-couplings of aryl chlorides at moderate temperatures. However, if one looks more carefully at the data, it becomes clear that the most active catalysts have some other structural features that account for their improved performance over catalysts of other ligands that are strongly electron donating.

Despite their good activity in oxidative addition, Pd complexes of PCy3 are not especially active in the Suzuki cross-coupling of aryl chlorides. Much better results can be obtained with the other di-cyclohexylphosphino substituted ligands 6, 7 and 8 (and many others). It would be envisaged that the donor strength of such ligands is broadly similar, but yet the most active of these, 8 can activate aryl chlorides at room temperature. A brief screening of the highly electron rich ligands, 9 and 10 compared to the less strongly donating 7 shows that the functionalised ligand gives more active catalysts with catalysts derived from 8 more active still. The likely reason for this is a higher concentration of the most active Pd species for catalysis, which was proposed to be a mono-ligated Pd species. The ligands of type 11 (which are also very strong bases), used by Verkade and co-workers are envisaged to be very strong donors, and consequently give active catalysts for a variety of aryl chloride cross-couplings. However, a crude comparison to Buchwalds results using what are likely to be slightly less electron donating ligands such as 8 also suggests other factors play a significant role towards the overall activity. In addition to considering the effect of the ligand on individual steps of the main catalytic cycle, and the effect of other species present in the catalytic reaction, ligand structure can also have an impact on catalyst activation and decomposition. The higher activity of catalysts derived from ligand 14 over ligands 12 and 13 has been convincingly demonstrated to be due to less catalyst decomposition (of 14 over 12), and more efficient catalyst activation of 14 over 13. Mechanistic studies such as these are vital to gain some predictability of the most challenging cross-coupling reactions. Some useful studies evaluating ligand-effects on oxidative addition are described below.

(Continues…)Excerpted from Organometallic Chemistry Volume 35 by Ian Fairlamb, Jason Lynam. Copyright © 2009 The Royal Society of Chemistry. Excerpted by permission of The Royal Society of Chemistry.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.