Green Integrated Enzyme–Ultrasound Extraction of Astaxanthin into Soybean Oil from Dried Haematococcus pluvialis: Enhanced Cell Disruption and Response Surface Modeling

Aims: This study developed and statistically modelled an integrated enzyme–ultrasound process for recovering astaxanthin from dried Haematococcus pluvialis directly into food-grade soybean oil. Study Design: Laboratory experimental study comprising treatment comparison, single-factor screening, and a three-factor, three-level Box–Behnken response surface design. Place and Duration of Study: Faculty of Pharmacy, Nguyen Tat Thanh University, Ho Chi Minh City, Vietnam, from December 2025 to June 2026. Methodology: Standardised dried H. pluvialis biomass (initial astaxanthin content: 21.94 mg/g dry weight) was subjected to enzymatic pretreatment followed by ultrasound-assisted extraction into edible oil. Soybean, sunflower, and rice-bran oils and five cellulase-to-pectinase ratios were screened. A Box–Behnken design comprising 15 randomised runs with three centre-point replicates was used to model total enzyme concentration (1.0–3.0%), sonication duration (15–55 min), and liquid-to-solid ratio (50–150 mL oil/g dry biomass). Astaxanthin was quantified by UV–Vis spectrophotometry at 475 nm using the specific absorption coefficient in acetone. Results: Under the preliminary comparison conditions (1% total enzyme, 15 min sonication, 1.0 mL oil), individual enzyme or ultrasound treatments produced <0.3% recovery, whereas sequential enzyme–ultrasound treatment yielded 20.822 ± 0.322%. Subsequent screening under intensified conditions (2% total enzyme, 35 min sonication, 1.0 mL oil) identified soybean oil and a 50:50 cellulase–pectinase blend as the best-performing conditions among those tested. The quadratic response-surface model was significant (F = 32.39, p = 0.000659), with R² = 0.9831, adjusted R² = 0.9528, and non-significant lack of fit (p = 0.4173). The liquid-to-solid ratio was the strongest linear factor. At the selected practical condition of 2.0% enzyme, 35 min sonication, and 100 mL/g, the model predicted 79.626% recovery and experimental validation yielded 79.708 ± 1.224% (95% CI; n = 6). Conclusion: Sequential enzymatic pretreatment and ultrasound markedly enhanced astaxanthin recovery compared with the individual treatments under the conditions tested. Response-surface modelling provided a statistically adequate description of the process and supported reproducible extraction into a food-compatible soybean-oil phase.