Purpose: Interactions of certain analytes with metal surfaces in UPLC instruments and columns are known to cause a range of deleterious effects including peak tailing, low peak areas and the formation of new peaks due to on-column reactions [1,2]. To mitigate these effects, we have developed a novel surface modification technology in which a hybrid organic/inorganic surface based on an ethylene-bridged siloxane chemistry is applied to the metal components in UPLC instruments and columns . This technology has been shown to give reduced tailing, higher peak areas and lower injection-to-injection variability for metal-sensitive analytes such as nucleotides, oligonucleotides and phosphopeptides. Here we demonstrate the benefits of this technology for UPLC separations of several small molecule pharmaceuticals.
Methods: Standard UPLC instruments and columns were compared to versions employing hybrid surface technology (HST), designated as ACQUITY Premier Systems and Columns. Separations of hydrocortisone sodium phosphate were carried out using ACQUITY UPLC™ BEH™ C18 1.7 µm 2.1 x 50 mm columns and an acetonitrile gradient using a mobile phase containing 10 mM ammonium formate (pH 3.0). UV detection was employed to determine peak area as a function of mass injected over the 2 – 200 ng range. Triplicate injections were made for each mass to determine the variability in the peak areas. Separations of deferoxamine mesylate were carried out using ACQUITY UPLC HSS T3 1.8 µm 2.1 x 50 mm columns and an acetonitrile gradient using an aqueous mobile phase containing 10 mM ammonium formate (pH 3.0). Electrospray ionization mass spectrometry detection was used to determine peak area as a function of mass injected over the 2 – 60 ng range. Ten consecutive injections were made at the 10 ng mass load to determine the variability in the peak areas. For the other mass loads, triplicate injections were made. Separations of clozapine were carried out using ACQUITY UPLC BEH C18 1.7 µm 2.1 x 50 mm columns and an acetonitrile gradient using an aqueous mobile phase containing 10 mM ammonium hydroxide. UV detection was used to determine peak areas of clozapine and its N-oxide over a series of injections.
Results: The results of the comparison of a standard UPLC system and column vs HST versions for hydrocortisone sodium phosphate showed that the HST versions produced narrower, more symmetric peaks with higher and more reproducible peak areas (shown in Figure 1). Relative to the standard system and column, the slope of plots of peak area vs mass injected were greater for the HST system and column and the peak area relative standard deviations (RSD) were greatly reduced for the lowest mass loadings (2, 20 and 50 ng). The results of similar standard vs HST system and column studies for deferoxamine mesylate also found greater peak areas and reduced peak area variability when using an HST system and column (shown in Figure 2). The peak area RSD for ten consecutive injections using an HST system and column at a mass load of 10 ng was 2.1%, vs 16.8% when using a standard system and column. The results for clozapine showed that the N-oxide (formed by on-column oxidation) was observed with a 2.05% relative peak area after 13 consecutive injections when using a standard column. In contrast, the N-oxide was not observed after the same number of injections on an HST column (shown in Figure 3).
Conclusions: These results demonstrate that interactions with metal surfaces in UPLC instruments and columns can affect the ability to obtain accurate and reproducible results for certain small molecule pharmaceuticals. The effects caused by these interactions include peak tailing, reduced peak areas and greater peak area variability. In addition, under some conditions pharmaceutical analytes can undergo on-column reactions catalyzed by metal surfaces. This results in the observation of peaks from compounds that were not present in the sample. We have demonstrated that the use of hybrid surface technology applied to the metal surfaces in UPLC systems and columns mitigates these effects. This enables more accurate and reproducible results for analytes that are susceptible to interactions with metal surfaces.
Acknowledgements: The authors acknowledge Mathew DeLano and Andrew Bates-Harrison for providing the HST components and Mike Dion for packing the columns evaluated.