Computer-aided approach for designing solvents blend for herbal phytochemical extraction

Abstract

There are many methods to obtain herbs phytochemicals such as the use of solvents as a phytochemical transfer medium in the extraction process. Currently, solvent is preliminarily selected based on the target phytochemicals and solvent polarities. This is followed by performing experiments to determine the potential solvents blends which give the highest desired phytochemicals content. The method uses a trial-and-error approach and is time consuming as well as resource intensive. The combination of property predictive models with computer-assisted search is one way to reduce the needs for huge amount of experiments needed to be conducted. Thus, the main objective of this work is to design solvent blends for the extraction of herbal phytochemicals using computer-aided approach. The methodology is divided into four levels which are pure component constraints (level 1), linear constraints (level 2), non-linear constraints (level 3) and stability checks (level 4). The proposed method has been applied to design a solvent mixture for the extraction of kaempferol from Kacip Fatimah herb as a case study. From the analysis, 12 feasible binary solvents mixture have been identified to be suitable for the extraction as it was within range of the design target. Thus, the optimal search has been performed to find the mixture solvents that can produce the highest kaempferol extraction yield with the lowest cost. Here, five binary solvents showed the highest kaempferol yield with the lowest solvent cost.