Extraction of phenolic compounds from spent coffee ground using natural deep eutectic solvents: New modeling approach

Abstract

AbstractThe extraction of bioactive compounds from food wastes or by‐products with natural deep eutectic solvents (NADES) could be an attractive alternative to traditional hydroalcoholic extraction but needs to be optimized taking advantage of modeling. This work aims to propose, develop, and discuss a new model for the extraction of phenolic compounds from spent coffee grounds (SCGs) using NADES to overcome some gaps related to the kinetic extraction models and the surface response methodology. The model presented is based on the same assumption as the kinetic model, but improving it to be able to predict phenolic extraction changing different process parameters (e.g., water content, time of extraction, temperature, etc.) using betaine: triethylene glycol (molar ratio 1:2). The model was validated on the extraction of total phenolic compounds from SCG as a function of water content of NADES ranging from 30 to 60% v/v. There was good agreement between experimental and simulated data and even, in some cases, better than with the second‐order kinetic models. In addition, the new model needs fewer parameters to be regressed concerning the specific case investigated in this work. This model and methodology can be the basis for further modeling applicable to other bioactives, and/or other NADES and feedstocks.Practical ApplicationsNatural deep eutectic solvents (NADES) can be a potential sustainable solvent for the valorization of food by‐products as they can be used to extract bioactive compounds with high added value. This extraction process needs to be optimized depending on several factors like water content in NADES solution, solids ratio, extraction time, and temperature. The model validated in this work on spent coffee ground (SCG) could be used for this purpose, improving the gaps related to the kinetic extraction models and the surface response methodology (RSM). The new model enables predicting the behavior of the process as a function of the parameters previously mentioned and reduces the number of experimental trials required compared with RSM. In particular, the new model was validated on the variation of water content in NADES solution. Certainly, this model will need to be further validated, but, once further studies are conducted, this model could be used for scale‐up calculation and design optimization of the industrial process.