Optimizing bleaching conditions of black cumin Oil by Box–Behnken design and artificial neural network to minimize losing of bioactive compounds

Black cumin oil (BCO) is considered vulnerable to rancidity, wherein black cumin crude oil is characterized by high peroxide value (PV) due to its high linoleic acid content. In this study, the effects of bleaching conditions (temperature, time, and bleaching earth concentration) on physicochemical traits of BCO were investigated. Different bleaching parameters were applied using the Box–Behnken design (BBD). The changes in PV, oil loss, linoleic acid, thymoquinone (TQ), and total tocopherol contents of oil were evaluated. A minimum PV (9 meq O2/kg) and oil loss of 6.87% were obtained, while the maximum concentrations for TQ and total tocopherols were 56.6 and 325 ppm, respectively. In addition to linoleic acid concentration, the overall fatty acid composition did not significantly change during bleaching. The highest difference between BBD and experimental data was 15.2 for total tocopherol concentration. An artificial neural network (ANN) was applied to predict the responses alternatively. The highest absolute differences between ANN and experimental data were 11.6 and 11.3 in TQ and total tocopherol contents. In general, ANN and BBD helped achieve feasible predictions. The multi-objective optimization maximized linoleic acid, TQ, and total tocopherols, while minimizing PV and oil loss. Optimum conditions were 100.30°C, 15 min, 3%, while the maximized concentrations of TQ and total tocopherols were 47.51 and 266.9 ppm at this optimum point. Practical applications The experimental data regarding the changes in the physicochemical properties of black cumin oil (BCO) revealed that the bleaching process successfully reduced the PV and oil loss to levels that could be recognized as indicators of good oil quality. In addition, minor components such as TQ and total tocopherols were preserved, and no remarkable changes in the fatty acid composition were observed. The predictions from BBD and ANN were comparable to experimental data and could be safely used with negligible differences. As the crude cold-pressed BCO had higher initial PV values (21.3 meq O2/kg) due to its non-refined nature, the direct consumption of BCO would inevitably trigger oxidation mechanisms. This study revealed the efficiency of the bleaching step in removing oxidation products to a minimum of 9 meq O2/kg while maintaining TQ and total tocopherols at 43.24 and 247 ppm, respectively. The results might guide future research regarding refining such cold-pressed oils having high oxidative potential regarding their nature.