Optimization in Normal-Phase HPLC

The normal-phase (NP) mode is not often used in high-performance liquid chromatography, probably because many analysts suppose that it is too complicated or even unreliable. This is by no means true rather, it is a valuable tool for the separation of a large group of analytes. Sihca, as the prototype of an NP stationary phase, is superior for the separation of isomers ds/trans, positional isomers, and diastereomers). An example is shown in Fig. 1. The silica surface is rigid (in contrast to the flexible hydrocarbon chains of the usual reversed-phase materials) and the analytes interact in a stericaUy defined maimer with the polar groups that 

HPLC Made to Measure A Practical Handbook for Optimization. Edited by Stavros Kromidas Copyright 2006 WILEY-VCH Verlag GmbH Co. KGaA, Weinheim ISBN 3-527-31377-X 

In NP-LC, the user has a much larger choice of solvents with different properties than in the reversed-phase mode. Table 1 gives an overview and lists the following data  

The solvents can be presented in a selectivity triangle according to their acidic, basic, and dipole properties (a, P, and t ), as shown in Fig. 3. It is obvious that more pronoimced changes in selectivity can be expected if the test separations are tried with solvents that are far apart from each other in the triangle. 

Solvent Strength e° Viscosity [mPa s7 UV cut-off [nm] Locali- zation Basicity 

---

Another important issue that must be considered in the development of CSPs for preparative separations is the solubility of enantiomers in the mobile phase. For example, the mixtures of hexane and polar solvents such as tetrahydrofuran, ethyl acetate, and 2-propanol typically used for normal-phase HPLC may not dissolve enough compound to overload the column. Since the selectivity of chiral recognition is strongly mobile phase-dependent, the development and optimization of the selector must be carried out in such a solvent that is well suited for the analytes. In contrast to analytical separations, separations on process scale do not require selectivity for a broad variety of racemates, since the unit often separates only a unique mixture of enantiomers. Therefore, a very high key-and-lock type selectivity, well known in the recognition of biosystems, would be most advantageous for the separation of a specific pair of enantiomers in large-scale production. 

The PRISMA model was developed by Nyiredy for solvent optimization in TLC and HPLC [142,168-171]. The PRISMA model consists of three parts the selection of the chromatographic system, optimization of the selected mobile phases, and the selection of the development method. Since silica is the most widely used stationary phase in TLC, the optimization procedure always starts with this phase, although the method is equally applicable to all chemically bonded phases in the normal or reversed-phase mode. For the selection of suitable solvents the first experiments are carried out on TLC plates in unsaturated... 

GuL. KelmM. Hammerstone J.F. Beecher G. Cunningham D. Vannozzi S. Prior R. L. 2002. Fractionation of polymeric procyanidins from lowhush blueberry and quantification of procyanidins in selected foods with an optimized normal-phase HPLC-MS fluorescent detection method. J. Agric. Food Chem. 50 4852-4860. 

In recent years several normal-phase HPLC methods have been reported for the quantitative analysis of tocopherols and tocotrienols (Table 11.5). The best of these methods have been able to achieve baseline separation of all four tocopherols and all four tocotrienols, as shown in Figures 11.2 and 11.3. Kamal-Eldin et al. (2000) reported the optimal baseline separation of all eight common tocols using a Diol-bonded phase column and an isocratic mobile phase of hexane/methyl tert-butyl ether (MTBE), 96 4, v/v (Figure 11.2). Similar separations were reported by Moreau et al. (2007) using the same type of column and mobile phase. Schwartz et al. (2008) reported that, with a normal-phase silica column, plastochromanol-8 in rapeseed oil eluted between y-tocopherol and 5-tocopherol. 

In addition, sometimes a normal-phase HPLC method at subambient temperature must be applied for analytes that are extremely prone to hydrolysis. In the synthesis of leukotriene D4 antagonist, accurate quantitation of mesylate intermediate is essential for process optimization. Owing to its inherent instability, analysis of mesylate intermediate must be carried out under normal-phase conditions with nonprotic solvents however, significant cycliza-tion of mesylation was stiU observed in such condition at room temperature. The authors concluded that the on-column reaction of the mesylate was silica-catalyzed cyclization. By conducting the normal-phase HPLC analysis at -30 C, it was demonstrated that on-column cyclization was adequately inhibited [30]. 

Examination of the synthetic route used in production allows for the prediction of potential residual synthetic impurities present in the drug substance. The API structure allows for the postulation of degradation pathways via hydrolytic, oxidative, catalytic, and other mechanisms. Both of these evaluations serve to facilitate the interpretation of (subsequent) identification tests. An examination of the physicochemical properties also allows for the rational establishment of method screening experiments by precluding certain conditions. For example, the use of normal-phase HPLC will be eliminated if the API is a salt or shows limited solubility in nonpolar organic solvents. Similarly, if the API (or suspected related substances) has no significant chromophore above 250 nm, the use of tetrahydrofuran (THE) and other solvents as mobile-phase components is severely limited. For compounds with an ionizable group, variation of pH will have considerable influence on elution behavior and can be exploited to optimize the selectivity of a reversed-phase separation. 

Thus, in order to evaluate the enantiopurity of the Michael product 14a obtained during the optimization of the catalyst A (Scheme 5), the purification of the analytical quantities of the crude 14a from its diastereomer and residual p-ketoester 12a was performed by semipreparative normal phase HPLC using achiral silica-based column ( -hexanes/lPA, Zorbax Rx-SIL column). This purification provided partial separation, and the pure fractions containing major diastereomer (i.e., 14a) along with some mixed fractions containing both diastereomers were collected. Remarkably, the first collected fraction contained highly enantiopure 14a (99% ee). 

Once again, the wide depth and breadth of knowledge regarding HPLC method development and assay optimization available in the literature cannot be covered here. The first suggestion is to check the literature for well-documented assays of known or related compounds. In many cases, the method used for HPLC/UV provides a helpful starting point for LC/MS methods with the same compound. In the interest of brevity, method development discussions are limited to reverse-phase separations. The unique requirements of normal phase and chiral chromatography are best discussed in other forums. 

---