The degradability of the biocomposite produced from a binary mixture of thermoplastic banana starch (TPS) and polycaprolactone (PCL) reinforced with fique fibers (Fs) was evaluated in three different environments (soil, compost, water). An experimental design with two factors (soil and compost) and three levels (5, 10, and 20 cm) was used, with additional tests for a third aqueous environment (water from the lake of the Universidad del Valle) at a depth of 20 cm. The biocomposite was prepared from the implementation of a twin-screw extrusion process of the binary mixture TPS/PCL and fique fibers (54, 36, and 10% composition, respectively), followed by hot compression molding, and after that, generating ASTM D638 type V specimens using a stainless-steel die. The specimens were dried and buried according to the experimental design, for a total experimental time of 90 days, and removing samples every 30 days. After 90 days, all samples showed signs of degradation, where the best results were obtained in the compost at a depth of 20 cm (34 ± 4% mass loss and a decrease in tensile strength of 77.3%, which indicates that the material lost mechanical properties). TPS was the fastest disappearing component and promoted the degradation of the composite material as it disappeared. Finally, the aqueous media presented the lowest degradation results, losing only 20% of its initial mass after 90 days of the experiment, being the least effective environment in which the biocomposite can end up.

Grape marc (Vitis labrusca), a major by-product of the winemaking industry, is generated in large quantities and represents a promising source of bioactive compounds. This residue is particularly rich in phenolic metabolites associated with antioxidant activity. In this study, supercritical CO₂ extraction was investigated as a sustainable strategy for the recovery of bioactive compounds from Vitis labrusca grape marc. A 2⁴⁻¹ fractional factorial design was employed to evaluate the effects of temperature (30–60 °C), pressure (137.9–275.8 bar), ethanol concentration (0–10 wt%), and particle size (116–601 µm) on extraction yield, total phenolic content (TPC), and antioxidant capacity (AC). Extraction performance was strongly influenced by operating conditions, revealing a clear trade-off between recovery and selectivity. The highest extraction yield (8.3 wt%) was obtained using 10 wt% ethanol as co-solvent, whereas the highest antioxidant capacity (365.3 µmol TE/g extract) was achieved under neat CO₂ conditions. TPC values reached approximately 69 mg GAE/g extract and were significantly affected by the combined effects of temperature, particle size, and ethanol concentration. The results revealed two distinct extraction regimes: a high-recovery regime promoted by ethanol addition and a high-selectivity regime under neat CO₂ conditions. Representative extracts were further characterized by UHPLC-QTOF-MS/MS. Ethanol-modified extraction was associated with higher relative abundance and diversity of flavonoids, stilbenes, and phenolic acids, whereas neat CO₂ extraction favored lipophilic metabolites such as oxylipins and unsaturated fatty acids. Selected operating conditions were successfully reproduced at pilot scale, supporting the scalability of the process. Overall, the results demonstrate that supercritical CO₂ extraction can be tailored to recover bioactive compounds from grape marc as extracts with distinct chemical profiles and provide a viable strategy for the valorization of Vitis labrusca winemaking residues.