The effect of column history in supercritical fluid chromatography: Practical implications

•Long-term retention time stability of 71 analytes was observed using eight columns.•The retention time variations were monitored for acids, bases, and neutral analytes.•The reversibility of stationary phase changes was tested using a regeneration procedure.•Adsorption of additive on the stationary surface is a factor contributing to the variations. Long-term stability of retention times of a wide range of analytes has been evaluated using eight different stationary phases. These were from a single manufacturer to minimize the differences in silanol activity caused by the manufacturing process. The tested stationary phases included bridge ethylene hybrid, 2-ethylpyridine bridge ethylene hybrid with direct modification of silica particles, bidentate crosslinked charged surface hybrid fluorophenyl, bidentate crosslinked high strength silica C18, and propanediol linked phases including diol (pure propanediol linker), and three phases based on diol further modified with 2-picolylamine, diethylamine, and 1-aminoanthracene group. Retention times were monitored at the first injection, after three, nine, twelve months, and after the column regeneration via washing with pure water. The analyses were carried out using three different mobile phases, including methanol, methanol with 10 mmol/L ammonium formate, and methanol with 0.1% ammonium hydroxide. No overall decreasing or increasing trends were observed after evaluating individual contributing parameters such as analyte, stationary phase, and organic modifier. Our results suggest that the silyl-ether formation is not the only factor contributing to changes in the stationary phase pore surface. Indeed, the adsorption of mobile phase additives is probably another significant factor. That was also confirmed by the regeneration procedure using water, which is likely to reverse the silyl-ether formation to achieve the original retention. However, the retention times returned to the original values for all analytes only on three columns. Retention times on other columns remained shifted within ± 15 % RSD depending on the analyte properties and the nature of organic modifier. The retention time variations observed for each analyte group, i.e., acids, bases, and neutrals, were interpreted for each stationary phase. We concluded that the sterically protected surfaces exhibited significantly smaller changes in the retention times. Although the regeneration procedure effect depended on the column type, the results suggeste