| Summary: | This thesis describes work carried out in order to increase knowledge concerning the influence of support structure and support surface chemistry on the enantioselectivity of cellulose tris(phenylcarbamate)-coated phases using high-performance liquid chromatography and flash chromatography. In the past cellulose carbamates have been coated (25% w/w) onto large particle (7 to 10 gm), macroporous (4000A) aminopropylated silica (APS) and packed into long (250 mm) HPLC columns. Therefore, the aim of the first part of this project was to use a small particle support, coated with a cellulose carbamate that could be packed into a short HPLC column in order to achieve rapid, efficient chiral separations. Hypersil APS with a particle size of 2.5 gm and a mean pore diameter of 120A was chosen as a potentially suitable support. As a result of using a small pore diameter support, the cellulose carbamate coating was not easily able to gain access to the pore volume and a 15% w/w cellulose carbamate loading was found to be optimum. As anticipated, 2.5 gm Hypersil APS coated with 15% w/w cellulose tris(3,5-dimethylphenylcarbamate) (CDMPC) was significantly more efficient than similarly coated 5 and 10 μm Hypersil APS phases and high eluent flow rates (≥ 1 ml/min) could be used without significant loss in efficiency. A 30 mm column packed with 15% w/w CDMPC-coated 2.5 μm Hypersil APS permitted the baseline separation of a range of chiral analytes in less than three minutes. In the next phase of the work, the influence of the support surface chemistry on enantioselectivity was investigated. Hypersil supports which were both more polar (underivatised silica) and less polar (octadecylated silica) than APS, and Hypercarb, a very non-polar porous graphitic carbon (PGC) support, were chosen in order to span a wide polarity range. Underivatised silica, due its small polar surface group, was able to accept a 20% w/w CDMPC loading and for many analytes this phase was found to be more enantioselective than a 15% w/w CDMPC-coated APS phase. In contrast, the large non-polar surface groups on octadecylated silica (ODS) appeared to significantly exclude CDMPC from the pore volume and a 15% w/w loading was found to be too high. However, because the octadecyl groups were able to shield acidic silanol sites to some extent, CDMPC-coated ODS showed potential for the separation of basic analytes. Hypercarb, which has a larger pore volume (250A) than Hypersil supports and has virtually no surface functional groups, accepted a 25% w/w CDMPC loading. This phase showed superior enantioselectivity for basic and acidic chiral analytes over CDMPC-coated APS. However, since PGC interacts strongly with flat molecules, badly tailing peak shapes were observed for a few bi- and polyaromatic chiral analytes. The high efficiency of the CDMPC-coated underivatised silica in HPLC columns led us to consider whether a modestly efficient, inexpensive flash chromatography silica coated with CDMPC could be used for preparative scale separations. Initially, a standard flash chromatography column packed with 20% w/w CDMPC-coated Davisil irregular silica (40-63 gm, pore size 150A) was used to investigate the separation of a range of chiral analytes. Resolutions were monitored by fraction collection with subsequent HPLC analysis. The method was found to be extremely easy, rapid and sample loadings of tens to hundreds of milligrams were achieved. Later two modifications were made; (i) in order to reduce the tedious collection of fractions, the flash column was modified to allow on-line UV detection and (ii) a 20% w/w CDMPC-coated Bondapak ODS flash column was prepared and was shown to dramatically improve the preparative resolution of basic analytes.
|
|---|
