OPTIMISATION OF ORGANIC SOLVENT EXTRACTION OF TOTAL CAROTENOIDS FROM BULGARIAN TOMATO PROCESSING WASTE

The quantity of tomato wastes combined with beneficial characteristics of components of the wastes justifies the great interest of researchers and manufacturers in extracting of carotenoids from this low cost material. In this study the response surface (RSM) approach and Box-Behnken design (BBD) were used to explore the possibility of modelling and optimisation the organic solvent extraction of total carotenoids from Bulgarian tomato processing waste (TPW).  Bulgarian TPW consists of tomato pomace was used in this study. Dried and ground TPW was plased in the extraction flask and stirred with acetone at various extraction conditions. The extract obtained was vacuum filtered through filter paper and was subjected to spectrophotometrically total carotenoids content determination. A three-variable, three-level BBD of RSM was employed in optimisation the extraction conditions for the highest recovery of total carotenoids from Bulgarian TPW. A response surface quadratic model was developed and statistical analysis was carried out. Data concerning model adequacy tests indicated that the values of R2 and Radj2 for quadratic model were the higest in comparison to the other models. The obtained F-value of 38.65 implies the model was significant (p < 0.05) and could be used for optimisation. The effects of extraction temperature, solvent to solid ratio and extraction time were significant in total carotenoids yield. Total carotenoids content ranged from 9.78 to 25.28 mg/100 g dried TPW. The predicted values of total carotenoids content were closed to the experimental observed values. By use of RSM the optimal extraction conditions were determined as follow: extraction time of 90 min, solvent to solid ratio of 60 mL/g and extraction temperature of 50ºC. The results obtained showed that predicted (28.40 mg/100 g dried TPW) and experimental (28.86±0.92 mg/100 g dried TPW) values of total carotenoids content were not significant different (p > 0.05).