Modeling and Optimization of Supercritical Fluid Extraction of Compounds from Campomanesia xa nthocarpa Fruits: Comparison between Artificial and Diffusion Based Models

The extracts of Campomanesia xanthocarpa fruits and leaves have many reported applications, especially in folk medicine for the treatment of several diseases. As it is hard to experimentally investigate the influences of the operating conditions on the supercritical fluid extraction (SFE) yield in a continuous range, modeling of this process can be an efficient method for this purpose. In this paper, the potential of two models, namely artificial neural network (ANN) and Hot Sphere Diffusion (HSD), were examined to simulate the SFE of compounds from Campomanesia xanthocarpa fruits. Noticeably, the numbers of adjustable variables were approximately the same for both models: 10 for the HSD model compared to 11 for the ANN model. A comparison of the models with experimental data points indicated the superiority of the ANN model over the HSD model. After verification of the ANN model, the influence of pressure and temperature on the extraction yield was studied, and the optimum temperature and pressure for maximizing the extraction yield were determined. The influence of operating conditions at low extraction times was slightly complicated, but for higher extraction durations, increasing the pressure and temperature had respectively a positive and negative effect on the extraction yield.