Read an Excerpt
Chiral Sulfur Ligands
Asymmetric Catalysis
By Hélène Pellissier The Royal Society of Chemistry
Copyright © 2009 Hélène Pellissier
All rights reserved.
ISBN: 978-1-84755-962-3
CHAPTER 1
Allylic Substitution
1.1 Introduction
Carbon–carbon bond formation is one of the most important reactions in synthetic organic chemistry. One useful and popular method is the palladium-catalysed allylation, e.g. the Tsuji-Trost reaction, of which asymmetric versions have been extensively studied over the last decade. Several classes of chiral ligands, such as N/N ligands, biphosphines, monodentate phosphines, and P/N mixed donor ligands have been extensively studied and proven to be effective ligands for Pd-catalysed asymmetric allylic substitution reactions. The Tsuji–Trost reaction is by far the most intensively studied reaction performed in the presence of sulfur-containing ligands over the past few years. Since sulfur is a soft complexation site and palladium a soft metal, the resulting complexes are expected to be strong complexes. In addition, ret-rodonation of π-electron density from the metal atom to the empty relatively low-energy d orbitals of the sulfur can also increase the strength of the Pd-S bond. Strategies for controlling enantioselectivity in palladium-catalysed asymmetric reactions have depended on the design and application of chiral ligands. Many of the efficient homo- and heterodonor chiral ligands, such as N/N-(e.g. bis(oxazolines)), P/P-(e.g. Trost's P/P ligands), and P/N-(phosphinooxazolines) types have been exploited. It must be noted that mixed P/N ligands have played a dominant role among the heterodonor ligands. A particular efficient method of C–C bond formation was opened up by the reaction of carbon nucleophiles with allylpalladium complexes, the generation of which is in situ accomplished and requires only a catalytic amount of the transition metal. Considerable efforts have been devoted to study the reaction between allylic substrates and nucleophiles catalysed by chiral palladium complexes. The palladium-catalysed allylic substitution is one of the catalytic homogeneous processes that has attracted most attention in recent decades and for which the catalytic cycle is well established (Scheme 1.1).
This is due in part to the relative ease of isolating catalytic intermediates, especially the palladium allylic species 1 (Scheme 1.1), although some related Pd(0) species (2) have also been characterised in solution. The enantioselec-tivity of the process with soft nucleophiles (derived from conjugated acids with pKa <25) is controlled by the external nucleophilic attack on the more elec-trophilic terminal allylic carbon of 1. The chemo-, regio-, diastereo-, and enantioselectivities of this process have been widely analysed and the results applied to the synthesis of target molecules. Since the first enantioselective catalytic process, described by Trost et al. in 1977, which implicated bidentate ligands with phosphorus as the donor atoms, the enantioselective allylic alkylation reaction catalysed by Pd has been of great interest in recent years, involving many chiral ligands and allowing excellent enantioselectivities. The catalysts often consist of a palladium complex containing a chiral chelate ligand but they can also be generated in situ. The mechanism of this palladium-mediated allylic reaction is reasonably well understood. A chiral Pd(0) olefin complex oxidatively adds the prochiral allylic acetate to afford an isolable Z3-allylic cationic compound, which is then attacked by the nucleophile. One of the most widely investigated sulfur ligands have been the S/N- donor type, often derived from a chiral oxazoline moiety. Oxazolines are known to have several advantages as sources of chirality, the main one being that they are readily accessible from homochiral amino alcohols and have proved to be effective catalysts in a variety of reactions. Furthermore, these ligands are easily modifiable and can incorporate different donor atoms in the side chains of the heterocyclic ring. Other combinations of donor atoms with sulphurhave also been explored, such as chiral S/P-donor ligands, chiral S/O-donor ligands, and chiral bis(sulfoxides) ligands. These processes are commonly catalysed by palladium systems, but other metals such as rhodium, platinum, molybdenum, tungsten, nickel or iridium are also efficient. The fact that a mixture of diastereomers can be obtained upon coordination of the ligand (thioether for example) to a metal can cause a decrease of stereoselectivity if the relative rates of the intermediates are similar. In spite of this feature, however, excellent enantioselectivities have been achieved. To evaluate the selectivity of a new chiral ligand for allylic substitutions, the reaction usually performed, called the test reaction in the text, is the transformation of rac-1,3-diphenylprop-2-enyl acetate with dimethyl malonate in the presence of N,O-bis(trimethylsilyl) acetamide (BSA) and a base (Scheme 1.2).
Two general strategies based on the type of ligands can be applied to catalyse the asymmetric allylic alkylation. The first strategy is based on the use of chiral C2-symmetric ligands, which results in catalytic systems with restricted numbers of diastereomeric transition states and consequently enantioselection. Indeed, the use of this type of ligands leads to the formation of only one (π-allyl)-palladium complex when a symmetric allylic substrate is used. The nucleophilic attack will take place preferentially at the site at which the allyl functionality has the strongest steric interaction with the ligand. The second strategy is based on the use of mixed heterodonating ligands containing strong and weak donor het-eroatom pairs, thus giving rise to different electronic properties associated with each metal-heteroatom. In this case, two different (π-allyl)palladium complexes may be obtained, with each one having the possibility of being attacked by the nucleophile at two different sites, leading to four possible transition states. Two effects are responsible for the stereochemical outcome. The first one is the electronic effect: if the two donor atoms exhibit sufficiently different donor-acceptor strengths, the number of possible transition states is reduced. Due to the trans-effect, the nucleophile would thus preferentially attack the allylic system on the carbon possessing a greater positive charge character, e.g. on the carbon situated trans to the best π-acceptor. The second effect is the steric effect: the final stereoselectivity is determined by the chiral moiety of the ligand favouring one of the two π-allyl conformations. The following sections of this chapter successively deal with homodonor S/S ligands, heterodonor S/P, S/N, S/C, S/O ligands, sulfur-containing P/N, P/O, N/N, N/O and P/P ligands, and then sulfur-containing ferrocenyl ligands.
1.2 S/S Ligands
Relatively few chiral dithiother ligands have been used for Pd-catalysed allylic substitutions, even though the coordinating capability of thioether donors in transition-metal complexes is known. These chiral homodonor ligands generally provided only modest asymmetric inductions for these reactions even if some of them were really performing. An inherent characteristic of thioether ligands is that, upon coordination to the metal, the sulfur atom becomes ste-reogenic. While the close proximity of the chiral sulfur centre to the coordination sphere of the transition metal may be beneficial, the low inversion barrier of the sulfur–metal bond may account for the scarce use of dithioethers in asymmetric catalysis. Hence, any attempts to incorporate a thioether into a chiral ligand must firstly address stereocontrol at the sulfur atom. Such a control may be accomplished by steric bias such as the involvement of efficient catalysts based on chiral mixed S/P ligands.
Unlike other homo- and heterodonor chiral ligands, the S/S-type ligand has hardly been involved in spite of having advantages such as lower cost, toxicity and oxidation potential. In 2001, Gomez et al. reported the first example employing C2-symmetric S/S-type ligands 3–10 (Scheme 1.3) for the test reaction. For unexplained reasons, the major focus in academia has been on this particular allylic alkylation, although this system does not seem to have any industrial importance. These workers showed that the enantioselectivity could be increased to high values with the appropriate combination of the chiral backbone rigidity and the substituent at the sulfur atom (Scheme 1.3). The modest asymmetric induction observed (≤ 1% ee) was due to the donor sites being insufficiently different for discrimination between both terminal allylic carbons in the intermediate. In all cases, solid structures of complexes and structural studies in solution provided these authors with proof of S/S-coordination.
In order to rationalise these results, these workers have studied more examples of palladium systems in which systematic changes of the chelate ring size and the electronic and steric effects of the sulfur substituents were performed. In the course of this systematic study, novel chiral dithioether ligands were shown to afford high activities and excellent selectivities in all palladium-catalysed allylic reactions (Scheme 1.4). The study of the allylic intermediates, which were fully characterised both in solution and in the solid state, has demonstrated that the selectivity in the palladium-catalysed allylic alkylation containing homodonor dithioether ligands could be controlled by the thermodynamics of the palladium diastereomer formation (high-energy barrier between Pd isomers, as is the case for these new ligands) or by the kinetics of the nucleophilic attack (low-energy barrier among the palladium species) depending on the nature of the metallacycle.
In 2003, Nakano et al. planned to synthesise novel chiral S/S-type ligands having a borneol backbone and without C2-symmetry. These ligands were readily prepared from the reactions of mercaptoisoborneol or mercaptoborneol with phenylthiobenzaldehydes with good yields. The chiral sulfideoxathiane ligands 11–14 were shown to give excellent enantioselectivities (up to 99% ee) in the palladium-catalysed allylic alkylation of 1,3-diphenyl-2-propenyl acetate with a range of alkyl malonate nucleophiles (Scheme 1.5). The presence of the bulky linked 2,6-dimethylphenylthio moiety proved necessary for achieving a high level of enantioselectivity. Semiempirical molecular orbital calculations were performed in order to propose an explanation for the high performance of ligand 12 compared to 11. Geometry optimisation and energy calculations were in accordance with a control of the catalyst conformation by the steric hindrance generated by the two methyl substituents of the phenyl ring. Bulkier ligands, containing linking 1-naphthylthio or 2-naphthylthio moieties, were also investigated for the test reaction and gave contrasted results, since the first ligand gave a moderate reactivity and a good enantioselectivity, whereas the second ligand gave, inversely, an excellent yield and a moderate enantioselectivity (Scheme 1.5).
As previously mentioned, an inherent characteristic of the thioether ligands is that, upon coordination to the metal, the sulfur atom becomes stereogenic. While the close proximity of the chirality to the coordination sphere of the transition metal may be beneficial, the low inversion barrier of the sulfur-metal bond may be responsible for the poor results observed. In this context, Khiar et al. have reported the synthesis of C-2 symmetric bis(thioglycosides) as new ligands for the test reaction (Scheme 1.6). The sugar residue was intended to provide a well-defined chiral environment, while the control of the sulfur configuration was expected, due to stereoelectronic factors acting at the anomeric centre. The best enantioselectivity was obtained with the ligand bearing a pivalate protecting group. Both enantiomers of the allylated product were prepared with an enantioselectivity of 90% ee by using inexpensive natural sulfur-modified D-sugars. Exploiting the fact that α-D-arabinose is almost the mirror image of β-D-galactose, these authors demonstrated that their sulfur derivatives behaved as pseudoenantiomers in the Pd-catalysed allylic substitution. They further succeeded in synthesising the corresponding Pd(II) complex and observed by in-depth NMR studies the formation of a unique anti diastereomer with C2-symmetry in solution, thus indicating a real control of the sulfur configuration by the sugar backbone. X-Ray analyses further indicated that the sugar residues were placed in a pseudoaxial orientation (in the exo- anomeric conformation). As a result of n-σ* hyperconjugative delocalisation, this effect is strong enough for an efficient stereochemical control over the sulfur configuration, as both atoms possess (S) absolute configurations.
In 2008, Skarzewski and Wojaczynska studied the test reaction in the presence of chiral C2-symmetric S/S-donor five- and six-membered cyclic ligands depicted in Scheme 1.7, providing moderate activity and enantioselectivity.The best enantioselectivity (42% ee) was observed when (1R,2S)-bis (phenylsulfenyl)cyclopentane was involved as the ligand, whereas the corresponding six-membered ligand gave the best activity combined with a lower enantios-electivity, as shown in Scheme 1.7.
Although the use of chiral sulfoxides as chiral controllers in asymmetric synthesis is well documented, their utilisation as ligands in asymmetric catalysis has met with little success. In 2005, Khiar et al. reported the synthesis of C2-symmetric bis(sulfoxides) but, surprisingly, when used as chiral ligands in the palladium-catalysed asymmetric alkylation of 1,3-diphenylpropenyl acetate with dimethyl malonate, they were completely inactive, whereas the corresponding C2-symmetric bis(thioethers) afforded the (R)-isomer with 42% ee. Another example of the application of this type of ligands for the asymmetric Pd-catalysed allylic substitution was reported by Shibazaki et al., providing moderate enantioselectivities of up to 64% ee (Scheme 1.8). These authors have studied the chelating ability of the chiral bis(sulfoxides), depicted in Scheme 1.8, for palladium, rhodium and ruthenium. X-ray crystal-structure analyses of the palladium complex indicated that this ligand coordinated Pd(II) through the sulfur atom and that the complex had a C2-symmetry. In addition, the corresponding unsymmetric monosulfoxide ligand, depicted in Scheme 1.8, was also investigated for the same reaction, giving a higher yield but a lower enantioselectivity.
In 2004, Shi et al. reported Pd-catalysed asymmetric allylic substitutions using axially chiral S/S- and S/O-heterodonor ligands based on the binaph-thalene backbone. The test reaction was performed in the presence of these ligands combined with [MATHEMATICAL EXPRESSION OMITTED] as the palladium precursor, BSA, and different additives (KOAc or LiOAc) in various solvents. As shown in Scheme 1.9, moderate to high enantioselectivities could be reached with the S/S-coordinating ligands 17–19, depending on the nature of the solvent and the additive used. The coordination of ligand 17 to Pd was studied by NMR analyses, suggesting the formation of a S,S-heterodonor complex. No explanation was proposed concerning the differences observed in the configuration of the product. As shown in Scheme 1.9, the corresponding S/O-chelate 20 proved to be less efficient in terms of both activity and selectivity for the test reaction.
In conclusion, S/S-coordinating ligands have proved to be efficient ligands for the Pd-catalysed asymmetric allylic substitution process, since high levels of enantioselectivity have been reached in some cases, thus becoming competitors of the generally most efficient nitrogen- or phosphorus-containing ligands.
1.3 S/P Ligands
The incorporation of the C2-symmetry into the chiral ligand design is a well-recognised strategy for restricting the number of diastereomeric transition states in metal-catalysed enantioselective processes. Equally powerful stereochemical restrictions may also be achieved with chiral ligands lacking C2-symmetry through the use of electronic effects such as the transinfluence. Such effects are a natural consequence of the use of chiral bidentate ligands equipped with strong and weak donor heteroatom pairs (e.g. PR3/NR3, PR3/SR2). Such electronic effects have the potential to influence both the stability and reactivity of the intervening diastereomeric reaction intermediates in the catalytic cycle. While mixed P/N-bidentate ligands have been applied in enantioselective palladium-catalysed nucleophilic alkylation, chiral thioether-containing donor ligands and, more generally, S/P ligands have been less well developed. In 1999, Evans et al. reported a new class of mixed S/P ligands incorporating a metal-bound thioether as a chiral control element and a dia-rylphosphinite moiety as a strong donor heteroatom in asymmetric catalysis. The utility of these thioether-phosphinite ligands 21–23 was illustrated in the palladium-catalysed allylic alkylation involving enol-malonate and aminenucleophiles (Scheme 1.10).
(Continues...)
Excerpted from Chiral Sulfur Ligands by Hélène Pellissier. Copyright © 2009 Hélène Pellissier. 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.
