Continuous production of single enantiomers at high yields by coupling single column chromatography, racemization, and nanofiltration

Abstract

A process combining single-column chromatography, racemization and solvent removal by membrane filtration to produce single enantiomers at high yields is proposed and investigated. Shortcut design methods are developed for a basic design using only a single chromatogram as input without the need for dynamic process simulation. A detailed process model is used to elucidate the role of relevant parameters and process dynamics. Experimental investigation of the process in fully coupled operation demonstrate the capability of the concept to produce single enantiomers at high yield and purity, as well as the applicability of the proposed design methods.

Introduction

Preparative chromatography is a powerful technology applied in various fields for resolving difficult separation problems. A difficult task that is particularly relevant in the pharmaceutical industry, is the production of single enantiomers at high purity. These are often produced via a classical chemical synthesis of the racemate, that is, the 50:50 mixture of the two enantiomers of a component. Since usually only one enantiomer is of therapeutic interest, this requires a subsequent separation of the racemate.

Such separation is challenging and expensive due to the extreme chemical similarity of the enantiomers. Various classical techniques are established, for example, crystallization, dynamic kinetic resolution, or chromatography. Besides this also hybrid separations that combine chromatography and crystallization with recycles were proposed for improving process performance [1], [2], [3], [4], [5]. However, due to the racemic composition of the starting material, the overall yield of enantioseparations is inherently limited to 50% only. This limitation can be overcome if an isomerization reaction, viz. racemization, of the undesired enantiomer can be included. Different corresponding concepts utilize crystallization, e.g. [6], chromatography e.g. [7], [8], or both [9]. As regards chromatography, a particular focus has been on the continuous simulated moving bed (SMB) technology. One research direction is devoted to reactive SMB processes and such with side reactors [8], [10], [11]. These concepts are particularly suited for low purity requirements. An improved performance for higher product purity, as is typical for enantiomeric products, can be achieved by SMB reactors with internally distributed reaction and separation [12], [13].

An alternative to the above approaches are reactor-separator-recycle systems. In such process chromatography is used to separate the desired enantiomer in required purity from the undesired form. The undesired enantiomer is subjected to a racemization reaction, which delivers, in equilibrium, again a racemic mixture. Together with some fresh racemic feed this mixture is recycled back to the chromatographic separation. Since chromatography causes a significant dilution, it is useful to perform a partial solvent removal before recycling the reaction product. Obviously, such concept can improve the yield of an enantiomer production up to 100%.

Bechtold et al. [8] suggested a corresponding process for enantiomer productions using SMB chromatography, an enzymatic racemization and membrane filtration for solvent removal. They proposed it also for other biotransformations. Recently Wagner et al. [14] documented a first successful experimental implementation in fully coupled operation for the production of a rare sugar, increasing the yield drastically from 25% to almost 100%. As regards enantiomer productions, several theoretical studies were performed [7], [8], [12], [15]. In particular in [5] possible economic benefits were underlined by cost savings of almost 50% for industrially relevant problems. However, so far a fully coupled implementation seems not reported.

In this work we extend the scope of this concept by using a conventional single column for the chromatographic separation. This can be an attractive alternative to SMB-based processes due to significantly lower investment costs and process complexity, respectively. Furthermore, this approach has a much higher flexibility since it allows removing additional impurities by defining corresponding waste fractions. Racemization can be performed by a homogeneous, heterogeneous, or enzymatic reaction, while solvent removal is preferably conducted by membrane filtration. It is worth noting that, to the best of our knowledge, so far only a batch-wise “decoupled” experimental implementation of a similar process based on evaporation and an additional crystallization was performed [9].

When aiming at a more attractive realization in fully coupled operation, the discontinuous nature of single-column chromatography complicates process design. The interconnections reduce the degree of freedom and prohibit an independent design of the units. Due to the distinct nonlinear dynamics of the periodically operated chromatography in combination with a recycle, a detailed design will require a correspondingly detailed dynamic model. However, simple methods for basic design are certainly also desired, since they facilitate a fast evaluation of achievable performance that might justify the efforts of a detailed design.

To address these questions, in the next section shortcut design methods are developed that require as input only a single chromatogram without the need for a detailed model of the complete process. Such model is developed subsequently for further design studies. In the following section the design methods are applied for evaluating the achievable process performance, and process dynamics are investigated using the detailed model. The final section demonstrates experimental feasibility of the process in fully coupled operation for an enantiomeric model system, as well as applicability of the design methods.