abstract
Supercritical fluids are alternatives to conventional harmful organic compounds. In the case of supercritical fluid extraction, CO2 is the most common solvent and can be advantageously modified with small contents of cosolvents like ethanol and ethyl acetate. The rigorous estimation of the tracer diffusion coefficients (D-12) of solutes in supercriticalmixtures (CO2+ co-solvent) requires their individual D-12 values in pure CO2 and pure co-solvent under the same operating conditions. This essay focuses the diffusivity of quercetin (solute) in two compressed liquid co-solvents (ethanol and ethyl acetate). Quercetin is a natural compound possessing a wide variety of bioactive properties, used as one of the most noticeable dietary antioxidants. The tracer diffusivity measurements are accomplished by the chromatographic peak broadening technique over 303.15-333.15 K and 1-150 bar. The diffusion coefficients lie between 0.414 x 10(-5) and 0.813 x 10(-5) cm(2)s(-1) in ethanol, and between 1.06 x 10(-5) and 1.69 x 10(-5) cm(2)s(-1) in ethyl acetate. Influence of temperature, pressure and hydrodynamic coordinates is analyzed and discussed based on the most relevant transport theories. Modeling is also carried outwith eleven models from the literature and demonstrated the unreliability of predicting equations in opposition to the very good correlations available to fit D-12 data. The influence of the accurate estimation of auxiliary properties (like solvent volume and viscosity) upon the calculated tracer diffusivities is also assessed, being possible to detect D-12 differences as high as ca. 70%. (C) 2020 Elsevier B.V. All rights reserved.
subject category
Chemistry, Physical; Physics, Atomic, Molecular & Chemical
authors
Zezere, B; Iglesias, J; Portugal, I; Gomes, JRB; Silva, CM
our authors
Bruno Miguel Martins Zêzere
PhD Student
Carlos Manuel Silva
Associate Professor
Inês Portugal
Assistant Professor
José R. B. Gomes
Principal ResearcherGroups
G4 - Renewable Materials and Circular Economy
G6 - Virtual Materials and Artificial Intelligence
acknowledgements
This work was developed within the scope of the project CICECO-Aveiro Institute of Materials, UIDB/50011/2020 & UIDP/50011/2020, financed by national funds through the Portuguese Foundation for Science and Technology/MCTES. Bruno Zezere thanks FCT for the PhD grant SFRH/BD/137751/2018.