RESUMEN
Herein, commercially available columns employed in hydrophilic interaction chromatography (HILIC) were characterized by determining their ability to selectively distinguish the minute structural differences between small molecules such as nucleosides and xanthines in complex sample matrices. Principal component analysis (PCA) was applied to the data obtained from structurally similar analytes, and the results showed that HILIC columns could generally be classified into two groups: (i) silane-modified columns that were prepared from either native silica particles or silica particles modified with low-molecular-weight silanes and (ii) polymer-modified columns obtained from silica particles functionalized with organic polymers. These two groups could be further subdivided based on the functionalities attached to the respective stationary phases. These results were confirmed via cluster analysis by preparing a dendrogram using the morphology-based selectivity parameters associated with the respective columns. We were able to determine the selectivity of columns for the OH groups, i.e., α(OH) and the prevailing pH conditions (cation- and anion-exchanging natures) on the surface of the respective stationary phases; α(theobromine/theophylline) was employed to obtain a similar two-dimensional plot. This test scheme, in which five compounds were analyze for each column, was helpful for understanding the impact of factors such as the hydrophilicity, degree of hydration, acidity/basicity, or the weak ion-exchange nature of the respective stationary phases on the separation characteristics of new HILIC stationary phases. The selectivity of columns for the CH2 group was also examined. The cation-exchange nature of the HILIC columns significantly influenced native silica columns and some polymer-modified columns. Herein, 45 commercially available HILIC columns were classified according to this method, and the results proved useful for understanding distinct separation characteristics of each HILIC column, enabling improved column selection.