Abstract
In this study, speciality lipids based on fish oil (FO) and capric acid (CA) were produced in packed bed bioreactors using immobilized Lipozyme IM from Rhizomucor miehei in a solvent-free environment. Our goal was to compare the product quality and yield among reactors consisting of one or two enzyme columns. Response surface methodology (RSM) was used to optimize process variables for maximum incorporation (Inc) of CA for each reactor configuration. The studied process variables were substrate molar ratio (1:1-3:1 CA/FO), temperature (35-55 A degrees C) and flow rate (0.5-1.5 mL/min). All experiments were conducted based on a face-centered cube design. The maximum predicted Inc of CA into FO (31.7 mol%) using one column was obtained when substrate molar ratio, temperature and flow rate of substrates were 2.70:1 (CA/FO), 55 A degrees C and 0.5 mL/min. The corresponding optimal Inc of CA into FO (22.7 mol%) using two columns was predicted at 2.95:1 (CA/FO), 55 A degrees C and 0.86 mL/min. Analysis of variance (ANOVA) showed that Inc of CA into the one-bed design was significantly influenced by all experimental conditions, with substrate molar ratio having the greatest impact. In the two-bed design, only temperature and flow rate had an effect; the molar ratio of substrates was not significant. Coefficients of determination were low for both designs, indicating a poor fit of our data to the model. However, the main purpose of this study was to assess the effect of process factors on Inc of CA into FO, rather than generate a model useful for prediction of responses under conditions not examined with this design. Therefore, the significant ANOVA results are much more important, stressing the real relationship between parameters and response, than the low coefficients of determination.