A generic chiral separation strategy for the analysis of acidic compounds in CEC is proposed in completion of an earlier defined strategy for nonacidic compounds. The screening step of this strategy uses a 45 mM ammonium formate (pH 2.9)/ACN (35/65, v/v) mobile phase, a temperature of 25°C, and an applied voltage of 15 kV. To update the screening step, eight chiral stationary phases, which all possessed chlorinated and nonchlorinated polysaccharide-based chiral selectors, were evaluated using the earlier defined screening conditions. A combination of the two types of polysaccharide-based chiral phases proved to have the highest cumulative success rate. In the updated screening step, amylose tris(3,5-dimethylphenylcarbamate) (ADH), cellulose tris(4-methylbenzoate) (OJH), cellulose tris(3,5-dichlorophenylcarbamate) (SP5), and cellulose tris(3,5-dimethylphenylcarbamate) (ODRH) were included as selectors and their preferred screening sequence was determined as ADH > OJH > SP5 > ODRH. New optimization steps were also defined for SP5 by investigating the influences of different parameters on the separation outcome using an experimental design approach. After application of the updated strategy, 15 of 17 acidic pharmaceuticals were separated under screening conditions, of which 9 were baseline resolved. When the optimization steps were applied, another three compounds were baseline separated, while the total number of separations was increased by one, which brings the total number of separations to 16 of 17 with 12 baseline separated compounds. This reflects the successful performance of the updated strategy on acidic compounds.

In an earlier part of this study (performance evaluation) it was observed, for home-made capillary electrochromatography (CEC) columns, that smaller particle diameters do not always generate higher efficiencies. This phenomenon was further examined in this study, evaluating Van Deemter curves. Naphthalene and trans-stilbene oxide were analyzed on four 3 µm and four 5 µm chlorinated polysaccharide-based chiral stationary phases (CSPs) applying voltages ranging from 5 to 30 kV. Neither the 3 nor the 5 µm packings generated systematically the highest efficiencies. The varying column efficiencies were optimized by evaluating nine packing procedures for both 3 and 5 µm CSPs. Again it was observed that smaller particle-size packings were not necessarily beneficial for the efficiency of the CEC analysis. This observation was statistically evaluated. A variability study evaluated different precision estimates related to column packing and replicate measurement conditions. The best columns with the highest efficiencies (for chiral separations) and good precision, that is, the lowest RSD values, were generated by the packing procedure in which an MeOH-slurry and a water rinsing step of 8 h were applied.

In this study, a test set of 44 nonacidic compounds was analyzed on four 3 µm chlorinated polysaccharide-based chiral stationary phases with cellulose tris (3-chloro-4-methylphenylcarbamate) (Lux Cellulose-2®; LC2), amylose tris (5-chloro-2-methylphenylcarbamate) (Lux Amylose-2®, LA2), cellulose tris (4-chloro-3-methylphenylcarbamate) (Lux Cellulose-4®; LC4) and cellulose tris (3,5-dichlorophenylcarbamate) (Sepapak-5®, Sp5) as selectors. The analysis times, retention factors, efficiencies and enantioselectivities were compared with the results obtained on their 5 µm analogs. All 3 µm packings, except for LA2, individually separated more compounds than their 5 µm analogs. When the cumulative success rates on the 3 and 5 µm packings were considered, it was observed that they were similar for both particle sizes; the combination of three or four 5 µm columns separated one compound more from the considered test set than that of the same number of 3 µm columns. Furthermore, it was observed that the 3 and 5 µm packings showed some complementarity. Only four compounds were not separated on any of the columns, while the use of only either the 3 or 5 µm columns resulted in 10 and nine not-separated compounds, respectively. The analyses on 5 µm LC2 and Sp5 were faster than on their 3 µm analogs. For LC4 the 3 µm packing showed the shortest analysis times and diverse analysis times for both particle sizes were obtained on LA2. Furthermore, three out of four 3 µm packings, that is, LC2, LC4, and Sp5, were found to be more efficient than their 5 µm analogs.