Optimization and modeling for heat reflux extraction of total yield, stevioside and rebaudioside-A from Stevia rebaudiana (Bertoni) leaves

Stevia rebaudiana (Bertoni) leaves consist of stevioside and rebaudioside-A (Reb-A). This research sought to improve extraction of target steviol glycosides from stevia leaf powder using response surface methodology (RSM) and artificial neural networking (ANN) under these independent variables: ethanol concentration, X-1 (0-100%), extraction temperature, X-2 (55-75 degrees C), and extraction time, X-3 (45-75 min). ANN outperformed as potential alternative to RSM in predicting optimum conditions. Maximum responses were obtained at 100% X-1, 55 degrees C X-2, and 60 min X-3. Heat reflux extraction proved superior to maceration extraction in terms of higher extraction yields with reduced energy consumption and CO2 emission.