INTEGRATING COSMO-RS, TOPSIS AND ULTRASOUND-ASSISTED EXTRACTION FOR SUSTAINABLE VALORIZATION OF COCOA BEAN SHELLS

Cocoa bean shells (CBS) are an abundant cocoa-processing by-product containing valuable methylxanthines and phenolic compounds. However, their sustainable valorization requires extraction systems that simultaneously consider recovery performance, solvent sustainability, and process efficiency. This study developed an integrated framework combining the Conductor-like Screening Model for Real Solvents (COSMO-RS), multi-criteria decision-making by the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), and high-intensity ultrasound-assisted extraction (UAE) for the recovery of bioactive compounds from CBS. COSMO-RS was initially applied to screen potential bio-based solvents according to their affinity for theobromine and epicatechin. Experimental validation demonstrated that solvent-solute affinity alone was insufficient to predict extraction performance because the low moisture content of CBS imposed matrix permeability and mass-transfer limitations. Water addition markedly enhanced the recovery of all target compounds. Experimental extraction performance for theobromine, caffeine, catechin, procyanidin B2, and epicatechin was subsequently integrated with CHEM21 safety, health, and environmental criteria using TOPSIS under four weighting scenarios. Water was favored when conventional sustainability criteria predominated, whereas 1,3-propanediol containing 20% water (v/v) was selected under a solvent-retention scenario, in which solvent recovery was not required. UAE optimization and experimental validation identified 1,3-propanediol containing 30% water, 100% ultrasound intensity (100 W), and a 60 °C endpoint as a favorable operational condition, yielding 10.69, 0.89, 2.42, and 2.08 mg g^-1 of theobromine, caffeine, procyanidin B2, and epicatechin, respectively. Compared with water, this system increased the recovery of theobromine and procyanidin B2 by 30.94 and 12.59%, respectively, while reaching 60 °C faster than both water and aqueous ethanol. Furthermore, reuse of the spent solvent with fresh CBS increased the cumulative loading of the target compounds by 52.9–75.9%, enabling extract concentration without a solvent-removal step. These results demonstrate that combining computational screening, multi-criteria solvent selection, and process intensification provides an application-oriented strategy for the sustainable valorization of agro-industrial by-products.

Acknowledgments: The authors acknowledge the São Paulo Research Foundation (FAPESP) (grants 2023/12621-1, 2022/10469-5, 2021/12264-9, and 2018/14582-5), the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES, Finance Code 001), and CATI Regional de São José do Rio Preto for providing the cocoa bean shells. This work was also supported by project PID2022-141965OB-C22 funded by MCIN/AEI/10.13039/501100011033 and ERDF/EU, and by Comunidad Autónoma de Madrid through AgroSUSTEC-CM (TEC-24/BIO-27). J.E. acknowledges grants RYC2022-035654-I and PID2024-156902OA-I00.