Re-evaluation of the Retention Time Prediction “Polarity” Model in the Context of the Development of a Stationary Phase Variable

The Abraham–Carr retention time prediction model was reviewed and developed in the context of the introduction of a new stationary phase variable. The new variable consists of a binary mixture dielectric constant and a stationary phase descriptor that was derived from the Snyder–Dolan column descriptors theory. In this work, data reported by Torres–Lapasio were investigated (in-silico) in the context of the linearity behavior of the Abraham–Carr theorem logk = ƒ(PNm). This proposed model has been replaced by several experiments that have shown that the stationary phase constant PNS, which is represented by a variable, does not require laboratory experiments. The new model, which contains stationary and mobile phase variables, shows an improved correlation for the majority of the Torres–Lapasio compounds between logk and ΔPN (mobile and stationary phase variables expression). The discussion regarding the model's behavior in the presence of secondary interactions led to research using a polar column and “bulky” compounds as reported by Szymański. Newly developed tacrine 4-fluorobenzoic acid derivatives, potential drugs in Alzheimer's disease treatment, are homologs and were found problematic to separate using regular C18 sorbent. The obtained results show that in the range of 30–70% acetonitrile, a single retention model using Agilent Zorbax SB-cyano sorbent is present, which is confirmed by two investigated models (R2 > 0.95).