Abstract
Polyphenolic extracts from natural sources have received great interest due to their beneficial properties for human health. A method to reduce their variability is to use the design of experiments which allows a limited number of experiments to be performed while exploring the experimental space. Firstly, a 23-full factorial model was used to investigate the polyphenols extraction from wild bilberry leaves. Spectrophotometric data (the content of polyphenols, flavonoids, chlorophyll and radical scavenger activity) and extraction yield were used as responses, and six statistical models were determined depending on the two numerical factors (temperature and alcohol % of ethanol-water mixture) being significant (p < 0.05) in all cases. Numerical optimisation performed by Design Expert 13 software correlates well with the chemical profile determined by high-performance liquid chromatography and the amount of the polyphenol. Afterwards, under the optimised conditions, an extract was prepared in three extraction steps for which composition, chemical stability and antimicrobial properties were evaluated. The antimicrobial potential of the extract was compared with that of the standard compounds (rutin and chlorogenic acid), and the results supported a synergistic effect of the extract components.