ULTRASOUND-ASSISTED EXTRACTION KINETICS FOR RECOVERING PHENOLIC COMPOUNDS FROM Amaranthus viridis

Amaranthus viridis (green caruru) is a South American non-conventional food plant (NCFP) renowned for its nutritional and medicinal relevance, yet often discarded as a weed. Containing more than 130 identified bioactive metabolites, notably health-promoting phenolic compounds, it represents a promising raw material for agro-industrial valorization. High-intensity ultrasound-assisted extraction (UAE) is a green extraction technology that minimizes processing time, operational energy demand, and solvent consumption compared with conventional solid-liquid extraction methods. This study investigated the influence of extraction time on global yield (X₀), total reducing capacity (TRC), and total flavonoid content (TFC) kinetics from A. viridis flour (CF). Aerial parts collected at ESALQ/USP were dehydrated in a forced-air oven (65 °C, 20 h), finely milled, and stored at -18 °C. Extractions were conducted using 1 g CF with 50 g of 50% (w/w) aqueous ethanol in a jacketed beaker at 2 ± 1 °C to prevent thermal degradation of sensitive phenolics, applying a probe sonicator (QR 850W, Ultronique; 20 kHz, 13 mm titanium probe, 550 W nominal power) over durations ranging from 1 to 30 min. Extracts were centrifuged, filtered, and analyzed. Kinetic data were modeled through segmented three-line spline regression using SAS OnDemand for Academics and statistically evaluated by ANOVA with Tukey’s test (p < 0.05). The spline model provided an excellent fit across all experimental responses (R² of 0.999 for X₀, 0.996 for TRC, and 0.989 for TFC), elucidating three distinct mass transfer periods: the constant extraction rate period (CER, t_CER ~1.0 min), dominated by acoustic cavitation and intense convective driven extraction of accessible surface compounds, recovering 61% (X₀), 64% (TRC), and 35% (TFC) of cumulative extractables; the falling extraction rate period (FER, t_FER ~6.0 - 6.5 min), representing combined convective and diffusive internal mass transfer mechanisms, which accumulated approximately 88% to 92% of extractable solutes; and the diffusion-controlled period (DC), where extraction rates decreased drastically. As negligible extraction yield increments occurred past 6 min, 9 min was selected as the optimal processing time, maximizing bioactive compound recovery while ensuring high energetic efficiency. These findings provide fundamental kinetic parameters to support further investigations regarding the combined effects of ultrasound nominal power and solvent composition on phytochemical profiles and biological properties of caruru extracts. Acknowledgments: The authors thank CNPq (Grant 133654/2025-7), CAPES, and FAPESP (Process 2025/17032-0) for financial support.