1
Introduction: A Survey of How and Why to Separate Enantiomers

Matthew Todd

This book is about the separation of enantiomers by synthetic methods, which is to say methods involving some chemical transformation as part of the separation process. We do not in this book cover chromatographic methods for the separation of enantiomers [1]. Nor do we focus on methods based on crystallizations as these have been amply reviewed elsewhere (see below). We are concerned mainly therefore with resolutions that involve a synthetic component, so mostly with the various flavours of kinetic resolutions through to more modern methods such as divergent reactions of a racemic mixture (DRRM). This introduction briefly clarifies the scope of the book.

The reasons such methods are of continued importance are threefold:

1. Society: the need for enantiopure compounds. New molecules as single enantiomers are important to our continued well-being because they are the feedstocks of new medicines, agrochemicals, fragrances and other features of modern society in a chiral world. Of the 205 new molecular entities approved as drugs between 2001 and 2010, 63% were single enantiomers [2]. Nature provides an abundance of enantiopure compounds, but we seek, and need, to exceed this by obtaining useful unnatural molecules as single enantiomers, and we may reasonably want to access both enantiomers of some compounds.
2. Academia: the basic science involved in the behaviour of chiral compounds. If we seek the state of the art in our discipline, we cannot help but think that rapid and selective chemical distinction between enantiomers, which results in their facile separation, is something beautiful in itself. There have been many successful methods developed for the synthetic separation of enantiomers, as we shall see, and these are both de facto interesting and instructive to consider for the design of future examples of such processes. The relationship between kinetic resolution and asymmetric catalysis is strong, and one can inform the design of the other. It is hoped that the diverse examples described in this book stimulate thoughts in the reader of what is possible next.
3. Industry: the need for new methods. There remain many classes of compounds that still cannot be resolved, or where efficiencies are too low for widespread adoption. It is still the case that classical resolution techniques are overwhelmingly used over other more complex methods. Of the 128 drug candidate molecules assessed in a recent industry survey, half were being developed as single enantiomers, and the sources of the stereocentres were mainly the chiral pool (55%) with resolution (28%) and asymmetric synthesis (10%) responsible for fewer examples [3]. This is, predictably, a feature of economics as much as science and one must not be too quick to judge new fields like asymmetric catalysis versus older ones like classical resolution. Pasteur added something enantiopure to a racemate in 1853 [4], whereas the catalytic prowess of a metal centre surrounded by chiral ligands was first demonstrated only in 1968 [5]. In addition, many chiral acids and bases have proven to be useful in classical resolutions, while Nature does not seem to be so generous in its supply of molecules that can effect catalytic, asymmetric transformations. The great progress made in synthetic chemistry has not (yet) brought us to the position that allows us to make any enantiopure substance in quantity given that resources are always limited. That leaves us with the synthesis of a racemate from which we pick one enantiomer out. Such a process can be remarkably efficient and cost-effective, if such tools are available, but the great successes described in this book should not hide the fact that we require better separation methods with wide applicability if we are to avoid an overreliance on just using whatever Nature provides.

The various methods considered in this book may be classified as follows.

1.1 Classical Methods

A racemate can be resolved with ease if it happens that the enantiomers form separate crystals – a so-called conglomerate. It becomes possible to separate, physically, the enantiomorphic crystals – a process sometimes referred to as triage. This is what Pasteur famously achieved in 1848 with a sample of ammonium sodium tartrate [6]. Such good fortune is quite rare and is in any case not a ‘synthetic method’ in the strictest sense (nor is it practical on a large scale). Other physical processes (alone, without any attendant synthetic process) such as evaporation or sublimation can be used to increase the enantiomeric excess of organic compounds, including amino acids [7].

Another important but non-synthetic method involves harvesting crops of enantioenriched crystals by seeding a supersaturated racemate and is known as resolution by entrainment, or preferential crystallization. These approaches have been well reviewed and will not be covered here [8]. Rare (but spectacular) examples where crystallizations of conglomerates are observed combined with racemization events in solution, leading to total spontaneous resolutions [9], are briefly mentioned in Chapter 7 as there have been interesting recent developments on that front. Included therein are special cases where diastereomeric interactions in solution combined with racemization may yield more than a 50% yield of one enantiomer, often known as a deracemization. Racemization is treated as a synthetic process of sorts, but we essentially stop short of processes where stereochemistry is created from prochiral intermediates since such a subject is formally the province of asymmetric catalysis. We will not be covering separations based on physical partitioning of enantiomers using, for example, chiral solvents, macrocycles or membranes [10].

Clearly the best-known non-chromatographic method for the separation of enantiomers, the so-called classical resolution, involves combination of a racemate and an enantiopure compound to form diastereomers that can be separated based on physical properties and the enantiomerically pure compound re-isolated (Scheme 1.1).

Scheme 1.1 The classical resolution: the separation of intermediate diastereomers differing in their physical properties.

The process is subtle and complex, belying its apparent simplicity on paper. My own experience with obtaining an important drug as a single enantiomer was instructive, and the process was publicly laid out over time as the project was carried out in an open source manner, meaning anyone could contribute to the solution and alter the direction of the research [11]. The project began with an assessment of a number of catalytic, asymmetric synthetic methods but feedback from the community was heavily in favour of moving to a search for a resolution, partly because the drug was needed for a low price. Sure enough a resolution was the solution. Sensitivity to conditions was illustrated by the eventual switch from one resolving agent to another – the structural difference was merely the substitution of the resolving agent's methoxy group for a hydrogen atom – resulting in the opposite enantiomer of the desired compound being isolated in the solid. This important and endlessly surprising area of organic chemistry, now including the so-called Dutch resolution approach of using mixtures of resolving agents, has been widely reviewed [12].

1.2 Kinetic Resolution (‘KR’)

A conceptually simple and well-known method for the separation of enantiomers is the kinetic resolution. A racemate is subjected to some reaction using a chiral agent and one of the enantiomers of the racemate reacts more quickly than the other (Scheme 1.2). The ‘selectivity factor’, s, is an expression of the difference in the rates of the two reactions.

Scheme 1.2 Generic kinetic resolution showing the ratio of enantiomers with conversion.

The enantiomer ratio is shown using a formalism borrowed from Ed Vedejs in Chapter 6. As can be seen, the composition varies dynamically as the reaction proceeds. At the outset, the system displays its best synthetic selectivity between the enantiomers because an equal amount of each enantiomer is present. As the reaction proceeds, statistics begins to interfere as the fast-reacting enantiomer is depleted. Success is a curse – as the fast-reacting enantiomer concentration reduces it becomes difficult to stop the slow-reacting enantiomer from participating – a ‘mass action’ problem. This explains the two best-known features of kinetic resolutions – that the maximum yield for the process is necessarily 50%, but that high enantiomeric excesses are found at the extremes of conversion – that is, that the enantioenrichment of the product (whatever that may be) is largest at the start of the reaction while the enantiomeric excess of the starting material left behind is largest just before the end of the reaction. In a kinetic resolution, one cannot usually have one's cake (high enantiomeric excess) and eat it (high yield).

A simple approach to kinetic resolution is to use a stoichiometric reagent. In the example shown in Scheme 1.3, an enantioenriched alcohol is obtained from a racemate through the addition of an enantiopure-acylating agent [13]. If 1 equivalent of the reagent were used, a 1 : 1 mixture of diastereomers would result with no kinetic differentiation. Stoichiometric kinetic...