Read an Excerpt
P-Stereogenic Ligands in Enantioselective Catalysis
By Arnald Grabulosa
The Royal Society of ChemistryCopyright © 2011 Arnald Grabulosa
All rights reserved.
1.1 Historical Overview
A century has passed since the discovery that compounds bearing tetrahedral sp phosphorus atoms bound to different substituents (P-stereogenic compounds) are in general configurationally stable and can be separated into a pair of enantiomeric forms. During the first 50 years, very few resolved P-stereogenic compounds were prepared (Figure 1.1).
In these early days, only quaternary phosphorus compounds were resolved. Ethylmethylphenylphosphine oxide (1) constitutes the first ever racemic P-stereogenic compound separated into individual enantiomers in 1911 by Meisenheimer and Lichtenstadt. It took 15 years for the same laboratory to repeat that feat and apply the same method to resolve benzylmethylphenylphosphine oxide (2). In 1944 sulfide 3 was resolved by Davies and Mann and 15 years later McEwen and co-workers resolved the first acyclic phosphonium salt (4).
In 1961 Horner and co-workers were the first to isolate optically pure trivalent phosphines, proving that, unlike amines, they do not racemise at room temperature. Soon afterwards, a few dialkylaryl- and triarylphosphines (5–7) were prepared in optically pure form (Figure 1.2).
Those compounds were synthesised by alkaline hydrolysis or cathodic reduction of phosphonium salts or by reduction of phosphine oxides with silanes. In both cases, the optically pure precursor had to be prepared by chemical resolution (see Chapter 2, Section 2.2), a step that is time-consuming, inefficient and very dependent on the structure of the compound being resolved.
Two key breakthroughs occurred in 1967-1968. The first one was the development of a new, relatively flexible route for the preparation of phosphine oxides, based on the separation of asymmetrically substituted menthylphosphinates. This method, described in Chapter 2, allowed the preparation of many optically pure phosphine oxides, which were reduced to the parent phosphines by known methods (see Section 1.3).
The second breakthrough was the discovery of enantioselective hydrogenation, which has had an enormous impact in homogeneous catalysis. In 1965 Wilkinson and co-workers reported that a soluble Rh(I) complex, [RhCl(PPh3)3], was an excellent precatalyst for alkene hydrogenation under mild conditions. The research groups led by Knowles and Horner independently replaced the achiral triphenylphosphine ligands in the Wilkinson catalyst by optically enriched P-stereogenic phosphines. The chiral precatalysts were tested in the hydrogenation of prochiral alkenes (Scheme 1.1 and Table 1.1).
The optical yields of 9 were low, but proved that enantioselective catalytic homogeneous hydrogenation was feasible. The extremely low enantioselectivity with the phosphine of entry 5 (Table 1.1) was justified by the location of the stereogenic centres further away from the metal in comparison to P-stereogenic phosphines.
At the same time, it was discovered that (S)-amino-3-(3,4-dihydroxyphenyl) propanoic acid (L-DOPA) was very efficient in the treatment of Parkinson's disease, thereby creating a sudden demand for this rare amino acid. Knowles and co-workers, who were working at Monsanto, produced vanillin that was used in the synthesis of an L-DOPA precursor. They decided to use the newly developed enantioselective hydrogenation for the efficient synthesis of L-DOPA (Scheme 1.2).
To optimise the hydrogenation of 10, a model reaction with a simpler substrate, 2-α-acetamidocinnamic acid (11), was developed and a range of phosphines was screened (Scheme 1.3).
After a quick screening with phosphines giving low enantioselectivities, it was found that the introduction of an o-anisyl group was very beneficial and finally led to the phosphine CAMP, which hydrogenated 11 (and 10) in high enantioselectivity. This was the first time that enzyme-like selectivity was achieved with an artificial, 'man-made' catalyst. Indeed, the Rh/CAMP system was so good that the process was scaled up for industrial production of L-DOPA. The first ever industrial catalytic enantioselective process was born. It seems common sense that to achieve high enantioselectivities the stereogenic elements of the ligands have to be as close as possible to the rhodium centre, as Knowles himself stated in his Nobel Lecture. Therefore, at that time a bright future was anticipated for P-stereogenic ligands.
At the same time, however, Dang and Kagan prepared the chelating C2-diphosphine DIOP (Figure 1.3), which was only marginally worse than CAMP in the hydrogenation of 11 in spite of having the stereogenic elements on the backbone and not at the phosphorus atoms.
Knowles and co-workers reported that DiPAMP, the most well-known P-stereogenic phosphine, gave 12 with an outstanding (at that time) 96% ee. With this phosphine many other enamides, enol esters and similar compounds could be hydrogenated in high ee. This work can be considered a milestone in the field of enantioselective homogeneous catalysis and have been profusely cited. William S. Knowles was awarded (along with Ryoji Noyori and Barry Sharpless) the Nobel Prize in Chemistry for the year 2001 for his contributions to enantioselective catalysis.
In spite of the success of DiPAMP, the good results with DIOP as well as those of Bosnich with Chiraphos, Noyori with BINAP, Burk with DuPHOS and many others showed that stereogenic phosphorus atoms were not required to obtain high enantioselectivities. This result prompted the preparation of many more phosphines bearing stereogenic carbon atoms or axes (Figure 1.4).
These ligands were configurationally stable and often more easily prepared than those bearing P-stereogenic atoms. This is the reason than soon after the discovery of DiPAMP, the attention was shifted away from P-stereogenic systems, leaving them in the shadow for decades.
In spite of this apparent oblivion, the development of methods to prepare P-stereogenic compounds kept improving slowly. An important event was the introduction of borane as a versatile protecting group in P-stereogenic chemistry (see Section 1.3), displacing phosphine oxides. Phosphine boranes are key intermediates in the majority of the modern methods of synthesis of P-stereogenic phosphines, described in Chapters 3 to 5, which were developed in the decade of 1990. Since 2000, those methods have reached a considerable maturity and nowadays are widely employed. In addition, novel exciting routes, such as the preparation of ligands by enantioselective catalysis, are also being explored (see Chapter 6).
In 1994 an excellent review on P-stereogenic phosphines and derivatives was published by Pietrusiewicz and Zablocka. It summarised much of the classic synthetic work but also described some of the emerging new methods. The review ended with the sentence 'the time has come for the P-chiral ligands to merge the stream and to bring in the P-chirality factor into play again'. More than 15 years later, this prediction is becoming progressively more accurate as the methods of synthesis develop and the use of P-stereogenic ligands in homogeneous catalysis spreads.
1.2 Configurational Stability of P-Stereogenic Compounds
Any atom bearing three different substituents and an electron pair in a trigonal pyramid structure is stereogenic. These compounds can be configurationally unstable and invert following a mechanism where the molecule turns inside out, like an umbrella. This mechanism is usually referred as pyramidal inversion (Scheme 1.4).
For ammonia and amines the energy barrier for pyramidal inversion is around 25 kJ mol-1. This precludes, for instance, the isolation of optically active N-stereogenic amines, except in especial cases.
In contrast, the energy barrier for phosphine pyramidal inversion is high enough (around 155 kJ mol-1) to prevent inversion occurring appreciably at room temperature. The same happens, in general, for substituted phosphines. In spite of that, in some compounds the phosphorus atom can racemise upon heating at moderate temperatures or under certain reaction conditions. As a result, the risk of racemisation is always a concern for the P-stereogenic chemist. The value of the pyramidal inversion energy depends on steric hindrance, electronic effects (electronegativity of the substituents and conjugative effects) and other parameters.
The evaluation of the value of inversion barriers has been done experimentally, often by variable temperature NMR methods, or by calculation. Munro and Horner showed that the pyramidal inversion follows first-order kinetics with little influence of the solvent polarity in the rate of racemisation. There are several classical studies that give reliable estimates of the inversion barrier values of many phosphines. Some examples are collected in Table 1.2.
Entries 1–6 show that typical acyclic tertiary phosphines have inversion energies in the range 120 to 150 kJ mol-1. In general, trialkyl phosphines have higher inversion barriers compared to diaryl and triaryl counterparts (compare entry 1 with entries 2–5). In spite of that, it seems that structural parameters do not profoundly affect the barriers for pyramidal inversion (compare entries 2–4). Entry 5 shows a small but interesting correlation between the inversion barrier and the electronegativity of the aryl group in phosphines of the type PPhArMe. It was observed that electron-withdrawing substituents in para position of the aryl group lower the energy of pyramidal inversion.
Phosphines of entries 8 and 9 are examples of compounds whose configurational stability is inferior due to other effects. Acylphosphines (entry 8) have inferior inversion barriers due to conjugation in the planar transition state. Comparison between entries 7 and 9 shows a dramatic lowering of inversion energies in phospholes compared to phospholanes, probably caused by an increase in the delocalisation of the planar transition state in the phosphole inversion. In spite of that, more conjugated fused phospholes (entries 10 and 11) are more configurationally stable.
Although intramolecular pyramidal inversion is the most important and general pathway, some compounds can racemise following other mechanisms. A typical example is secondary phosphines, which are relatively stable towards pyramidal inversion (Table 1.2, entry 12) but can racemise following acid- and base-catalysed mechanisms (Scheme 1.5).
The formation of achiral phosphonium salts (14) or phosphide anions (15) explains the easy racemisation of 13 in the presence of traces of acids or bases. It has been found that 13 can be stabilised by the presence of a mild base, capable of neutralising any trace of protons but unable to form 15. A different way to stabilise secondary phosphines is the quenching of the electron pair by the formation of phosphine boranes (see Section 1.3).
Another family of configurationally unstable compounds are chlorophosphines, which racemise extremely fast through the formation of an achiral pentacoordinate intermediate (see Chapter 4, Section 18.104.22.168).
Inversion at the phosphorus atom is usually regarded as a serious drawback of P-stereogenic ligands but it can also be an advantage from the synthetic point of view in molecules containing additional stereogenic elements. In the ideal case, even optically pure compounds can be obtained if the phosphorus atoms are left to epimerise until the most thermodynamically stable isomer is formed. The best part is that only heating is required. This strategy, however, has been seldom used in P-stereogenic chemistry.
The conclusion of this section is that, except in special cases, tertiary phosphines have racemisation barrier energies higher than 120 kJ mol-1, corresponding to equilibration temperatures exceeding 100 °C. For example, solutions of methyl(1-naphthyl)phenylphosphine (Table 1.2, entry 6) have an inversion half-life of only 7 hours at 95 °C but of 10 years at 20 °C. When coordinated to transition metal centres, even better configurational stabilities are in general expected. Therefore, it can be concluded that P-stereogenic phosphines are stable enough to be used in most transition metal-catalysed reactions.
1.3 Interconversions between P-Stereogenic Compounds
Optically pure, trivalent P-stereogenic compounds are prepared by several methods, discussed in Chapters 2–6. Most of the syntheses are multi-step processes that involve tetra- or pentacoordinated intermediates. Two families of compounds, phosphine oxides and phosphine boranes are by far the most important direct precursors of free phosphines. The interconversions between phosphines and phosphine boranes and oxides are briefly discussed in this section.
1.3.1 Phosphines, Phosphine Oxides and Sulfides
Phosphine oxides were the classic precursors to P-stereogenic phosphines until they were (partially) supplanted by phosphine boranes. Here some of the methods of oxidation and reduction of phosphines (Scheme 1.6) are briefly considered.
Many synthetic routes lead to phosphine oxides although phosphines themselves are rarely oxidised on purpose, except for purification or analysis. If required, however, there are several reagents that deliver the phosphine oxide quantitatively and stereoselectively, usually with retention of configuration at the phosphorus atom. The most typical reagents are peroxo compounds such as H2O2, t-BuOOH and MCPBA. The most available oxidant, air, can be also used but often leads to impure products. Treatment of P-stereogenic phosphines with halogens or pseudohalogens in water leads to inverted phosphine oxides.
Due to the stability of the phosphoryl group, reduction of phosphine oxides to free phosphines requires strongly oxophilic reagents under harsh conditions. In spite of that, is such an important transformation in P-stereogenic chemistry that dependable methods with either retention or inversion of configuration have been developed during the last 50 years. The reductants of choice have traditionally been silane reagents, sometimes mixed with amines. Some selected examples of reductions with those reagents are listed in Table 1.3.
Trichlorosilane is a very powerful reducing agent. On its own or mixed with mildly basic amines or phosphines it leads to products with retention of configuration in the reduction of acyclic phosphines (entry 1) whereas with more basic amines it affords inverted phosphines (entries 2 and 3) with the exception of phosphetane oxides (entry 4). Entry 3 shows that the famous ligand DiPAMP was prepared for the first time using this reagent. Hexachlorodisilane has comparable reducing capacity often under milder conditions. With this reagent, acyclic phosphine oxides invert upon reduction (entry 5) but with phosphetane oxides retention is observed (entry 6). Entries 7–11 show that non-halogenated silanes have also been used. Phenylsilane (entries 7–9) and triethoxysilane or polymethylhydrosiloxane with a catalytic amount of Ti(IV) isopropoxide (entries 10–12) are milder reductants than chlorosilanes and provide phosphines with retention of configuration. In the systems of entries 10–12 it is thought that titanium hydride species are the active reductant.
Reduction reactions are typically performed in refluxing benzene (bp 80 °C), acetonitrile (bp 82 °C) or toluene (bp 111 °C). Depending on the substrate and conditions, the reaction time can be anything from minutes to days. Although the reductions themselves are frequently very stereoselective, exposure of the free phosphines to the reducing agents and their oxidation products degrades the optical purity of the phosphines. Therefore, the general rule is to keep the temperature and reaction time as low as possible to maximise the optical purity of the free phosphines. In spite of that, variable amounts of racemic products are often unavoidable.
Other reductive systems have been explored in much lower intensity (Table 1.4).
Excerpted from P-Stereogenic Ligands in Enantioselective Catalysis by Arnald Grabulosa. Copyright © 2011 Arnald Grabulosa. 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.