Comparative modelling study on enantioresolution of structurally unrelated compounds with amylose-based chiral stationary phases in reversed phase liquid chromatography-mass spectrometry conditions

• Published in: Journal of Chromatography A Vol. 1625; p. 461281
• Main Authors: Pérez-Baeza, Mireia, Escuder-Gilabert, Laura, Martín-Biosca, Yolanda, Sagrado, Salvador, Medina-Hernández, María José
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 16.08.2020
• Subjects: Amylose-based chiral stationary phases. Reversed phase liquid chromatography. Enantioresolution modelling and description. Discriminant partial least squares. Feature selection; Amylose-based chiral stationary phases. Reversed phase liquid chromatography. Enantioresolution modelling and description. Discriminant partial least squares. Feature selection
• ISSN: 0021-9673
• DOI: 10.1016/j.chroma.2020.461281

•Compared enantioresolution (Rs) in 3 amylose-based chiral stationary phases (CSPs)•Rs-discriminant-models of 2 Amylose-CSPs in RPLC after feature selection•Rs-topological/molecular descriptor relationships for unrelated compounds•Chiral carbon-topological parameters are key variables for enantioresolution•Aromatic and nitrogen related parameters connect with enantioresolution Polysaccharide-based chiral stationary phases (CSPs) are the most used chiral selectors in HPLC. These CSPs can be used in normal, polar organic and aqueous-organic mobile phases. However, normal and polar organic mobile phases are not adequate for chiral separation of polar compounds, for the analysis of aqueous samples and for MS detection. In these situations, reversed phase conditions, without the usual non-volatile additives incompatible with MS detection, are preferable. Moreover, in most of the reported chiral chromatographic methods, retention is too large for routine work. In this paper, the chiral separation of 53 structurally unrelated compounds is studied using three commercial amylose-based CSPs -coated amylose tris(3,5-dimethylphenylcarbamate) (Am1), coated amylose tris(5-chloro-2-methylphenylcarbamate) (Am2), and immobilised amylose tris(3-chloro-5-methylphenylcarbamate) (Am3)-. Chiral separations are carried out using acetonitrile/ammonium bicarbonate (pH = 8.0) mixtures, reversed mobile phases compatible with MS detection. To provide realistic conditions for routine analysis, maximum retention factors are set to 15. Retention and enantioresolution behaviour of compounds in those CSPs are compared. On the other hand, to compare and describe the resolution ability of these CSPs, 58 structural variables of the compounds are tested to model for the first time a categorical enantioresolution (CRs) for Am1 and Am3 CSPs. Discriminant partial least squares, for one response categorical variable (DPLS1) is used for feature selection, modelling. The final DPLS1 models showed good descriptive ability.