A new computational application of the AGDC algorithm for kinetic resolution of multicomponent mixtures (static and dynamic)

A new multipurpose computational methodology is proposed for the resolution of static (SMM) and dynamic (DMM) multicomponent mixtures without the need for separation or calibration. The treatment is performed by implementation of an updated version of AGDC, a robust algorithm of unconstrained mathematical optimization suitably adapted to the characteristics of this problem. This algorithm has been used previously in other fields within of chemometrics (mainly chemical kinetics and thermodynamics) with excellent results. Spectrophotometrically monitored kinetic data are computed after the addition to the mixture of several reagents that react with the analytes. The methodology allows the determination of (a) analyte concentrations, (b) rate constants, and (c) both parameters simultaneously, all in a single optimization process. Likewise, the method permits the validation of the model that best fits the experimental data. A large set of experiments was conducted with a view to know the limits of applicability of the computational methodology after evaluation of: the maximum deviation of the initial estimates, the number of parameters that can be optimized, the maximum difference in the order of magnitude that can exist among those parameters and the capacity of discrimination between analytes with very similar chemical characteristics. The methodology is applied to the particular case of the resolution of a DMM mixture formed by the isomers 5-cholesten-3-one and 4-cholesten-3-one. The concentrations of all the species participating in the process were determined and validation of the model demonstrated the necessary presence of an enol intermediate in the reaction mechanism.