Überkritische Extraktion natürlicher Produkte

A NEW PHYSICO CHEMICAL APPROACH IN SUPERCRITICAL NATURAL PRODUCTS EXTRACTION

Dr. Carsten Zetzl

Introduction

Several models have been developed in the past decades in order to describe the batch extraction of natural products with supercritical fluids. The theoretical spectrum can be distinguished in Diffusion- Desorption-Dispersion Models, Shrinking Core Models, Empirical Models and others.(Figure 1):

Figure 1:Theoretical shape of an Extraction yield curve
Figure 1:Theoretical shape of an Extraction yield curve

 

However, applicants of lab scale and pilot batch SFE plants often suffer from a lack of basic physicochemical parameters which have to be introduced into the model, as (effective) diffusion coefficients, desorption equilibrium, thermodynamic and/or operational solubility.

Some questions which are up to now not yet completely dissolved concern the following facts:

  • the effective diffusion coefficient is up to three dimensions smaller than the binary diffusion coefficient (according the established equations of Catchpole, Funakuzuri and others). The gap between the calculated values and the effective one is frequently explained by tortuosity factors and adsorption influences.
  • Additionally, the effective diffusion coefficient increases with reaction time, when applying the so called Simple Single Plate model

Figure 2: Influence of run time on the effective diffusion coefficient (Brunner, 1994)
Figure 2: Influence of run time on the effective diffusion coefficient (Brunner, 1994)

The focus of this work is to compile a program tool which allows the rapid simulation of Batch SFE processes and its comparison with experimental results.

Specifications of this tool are

  • comprehensive program allowing to use models from Goto et al.(1996), Brunner et al.(1994), Reverchon et al.(2003), Sovova (1996) , and others
  • direct information and development of the required physicochemical parameters like Diffusion coefficient, desorption and dispersion coefficients, mass transfer
  • parameter calculation using recent and established models from Catchpole (1994, 1996), delValle (2000,2002), Tan and Liou(1989), Liong (1992) and others
  • capacities for the subsequent treatment of the data

Recent activities

The recent “beta” version of the program focuses on the prediction of the mass transfer parameters, herein even dispersion and by-pass effects shall be taken into account. Therefore, the maximum of information on the co-ordinates of the Constant Extraction Rate section must be deduced. A matrix model of the pores will take into account the influence of the pore wetting on the local and time dependent effective diffusion coefficient.From Normal Temperature and Pressure Technologies, a matrix model has been adapted which allows to consider the mass transfer resistance as a sequence of parallel and serial influences. With this approach, it was possible to explain the increase of the diffusion coefficient from low to high values: 

    Figure 3: Deduction of the effective diffusion coefficient as a local function of time / pore wetting  Figure 3: Deduction of the effective diffusion coefficient as a local function of time / pore wetting
Figure 3: Deduction of the effective diffusion coefficient as a local function of time / pore wetting

Future activities

The publication of the program as a downloadable tool is in preparation. The number of experimental results and predictions of SFE yield curves on natural products is enormous. However, as the dimensions of the lab scale columns differ strongly, and have different flow rate capacity ratios compared to the industrial plants, there is the need of a standardization of the presentation.

Figure 4: The VTII Simulation tool, conclusion of different spice extraction yield curves (Majewski, 1998)Figure 4: The VTII Simulation tool, conclusion of different spice extraction yield curves (Majewski, 1998)
Figure 4: The VTII Simulation tool, conclusion of different spice extraction yield curves (Majewski, 1998)