Coumarin Extraction from Cuscuta reflexa using Supercritical Fluid Carbon Dioxide and Development of an Artificial Neural Network Model to Predict the Coumarin Yield

Abstract

Coumarin found in Cuscuta reflexa (a medicinal herb) is a phytochemical that possesses blood-thinning, anti-fungicidal and anti-tumor activities, and increases the blood flow in the veins and decreases capillary permeability. Supercritical fluid carbon dioxide extraction is an effective method for extraction and separation of organic compounds. Coumarin was extracted using supercritical fluid carbon dioxide associated with a small amount of a co-solvent methanol to increase the selectivity of carbon dioxide and identified by high-performance liquid chromatography comparing with authentic coumarin standard. Determination of proper extracting temperature, time, and pressure is essential for the maximum extract of coumarin yield. Coumarin was extracted under different extraction conditions of temperature (35-75 A degrees C), time (30-150 min), and pressure (15,160-34,450 kPa). Central composite rotate design was used to plan the experimental extraction conditions. The highest yield of coumarin was 90.13 A +/- 0.11 (A mu g/g), while extraction was done at 55A degrees C for 150 min under 24,805 kPa. A feed-forward multilayer backpropagation artificial neural network (ANN) model was developed for the prediction of coumarin yield (A mu g/g), where input variables were temperature, time, and pressure. Several configurations were evaluated while developing the optimal ANN model. The optimal ANN model consisted of one hidden layer and five neurons. This model was able to predict coumarin yield quantitatively. Prediction of coumarin yields using the ANN model was proven to be a simple, convenient, and accurate method.