Optimization of NaCl-assisted starch extraction from Alocasia macrorrhizos corms and characterization of the extracted starch
Abstract
Alocasia macrorrhizos is an underutilized starch source with limited industrial application due to low starch yields reported in previous studies and insufficient physicochemical characterization. In this study, a NaCl-based starch extraction process from A. macrorrhizos corms was optimized using Response Surface Methodology (RSM) and Artificial Neural Network (ANN) modeling. Preliminary screening by Plackett–Burman design identified corm size, soaking temperature, sample concentration, and water proportion for milling as significant factors influencing starch yield. These variables were further optimized using a Central Composite Rotatable Design. The RSM-optimized process significantly enhanced starch yield to 83.88 g per 100 g dry corm, representing a 3.35-fold improvement over previously reported values. Comparative evaluation of RSM and ANN models showed superior predictive performance of RSM, with R² values of 0.9994 (design) and 0.8417 (validation), compared to 0.8158 and 0.8696 for ANN, respectively. The extracted starch samples, categorized as high-, medium-, and low-yield, exhibited granule sizes ranging from 1.683 to 2.008 μm and low levels of ash, protein, phenolics, and free amino acids, indicating relatively low non-starch impurities. FTIR and XRD analyses confirmed chemical stability and a well-defined crystalline structure. Overall, the optimized NaCl-based extraction process demonstrates strong potential for scalable and industrial starch production from A. macrorrhizos.
