Introduction
The separation of 2,3-butanediol (2,3-BDO) from aqueous solutions has been a major challenge throughout the past decades [1]. Costly removal of side products followed by energy-intensive distillation of the aqueous system remains state of the art [2, 3]. Due to high downstream costs, the biocatalytic production of 2,3-BDO is limited. Despite numerous conceptual studies, no simple and efficient separation strategy for the recovery of the hydrophilic, high boiling compound was developed [4]. To reduce energy consumption, solvent extraction is being explored as an alternative [5]. However, its effectiveness has been limited by the low distribution coefficients and selectivity of established solvents when dealing with hydrophilic compounds like diols. Therefore, identifying more efficient solvents for extracting hydrophilic compounds from water is essential [6-8].
A novel class of solvents is the family of terpenoids, which have recently been proposed for extraction [9]. Terpenoids are a group of natural products that form strong hydrogen bonds and can be divided into subgroups of molecules that act as hydrogen bond donors (HBD) and hydrogen bond acceptors (HBA) [10]. These solvents are particularly promising in extraction when the solute also forms hydrogen bonds targeting the corresponding free association side of the solvent [11, 12]. Preliminary experiments have shown that carvacrol shows promising distribution coefficients for 2,3-BDO from water. To better evaluate the potential of terpenoids as novel solvents, they should be considered in the process.
Methodology
To this end, the solvent extraction of 2,3-BDO from water with carvacrol is examined in more detail in terms of equilibrium experiments, operability in extraction columns and specific exergy demand. We gather experimental data of the relevant liquid-liquid equilibrium (LLE) and vapor-liquid equilibrium (VLE) to describe extraction and distillation as the proposed unit operations.
To analyze the operability of carvacrol as solvent, experiments are carried out in a DN32 Kühni extraction column (active height: 1350 mm, 48 rotor-stator compartments). The analysis of the flood points and maximum column load should serve as a measure of the operability.
For applicability in process design, a Gibbs-excess model is parametrized using the gathered equilibrium data. The separation of 2,3-BDO from water based on distillation is defined as a benchmark for process assessment. Alternative process concepts are extractive processes using conventional solvents. Finally, based on thermodynamic equilibrium data, operability and process modeling, the potential of terpenoids as new green solvents for the extraction of hydrophilic, high-boiling molecules from water is evaluated.
Results
The LLE experiments show that carvacrol offers higher distribution coefficients and selectivity for 2,3-BDO compared to conventional solvents.
Initial column experiments have shown that pure carvacrol exhibited a low maximum load in the column, primarily due to the small density difference between carvacrol and water. Despite the counter-current operation, no concentration gradient is achieved due to the axial dispersion with pure carvacrol. To overcome these limitations, a mixed solvent system (carvacrol and n-dodecane) is analyzed in addition to pure carvacrol. N-dodecane is chosen for its low density and minimal cross-solubility with water, resulting in an increased maximum column load. The extent to which the mixed solvent system influences the distribution coefficients and selectivity is discussed based on equilibrium data. Finally, considering equilibrium data, operability, and specific exergy demand, the potential of both pure carvacrol and the mixed solvent system as alternatives to conventional solvents is critically assessed.