Two approaches to shift the acylglycerol equilibrium were tested as follows: addition of monoacylglycerols and lowering of the temperature. None of these approaches were able to shift the equilibriumtoward higher diacylglycerol (DAG) contents. The glycerolysis reaction was optimized with five factorsusing response surface methodology. Evaluation of the resulting model enabled the determinationof optimal reaction conditions for glycerolysis aiming at high DAG yield. However, verification of themodel showed that the model was unable to take the molecular equilibrium into account but it providedgood insight in how process settings can be chosen to, for example, minimize
production costs. Optimalconditions were found to be the following: no extra water, low content of glycerol (molar ratio of 2),temperature of 60-65
C, 4-5 h reaction time, and only 5 wt % lipases. Up scaling of the glycerolysisprocess was performed and revealed that scale-up to a 20 kg
production in a pilot
plant batch reactorwas possible with a similar DAG yield (60 wt %) as in lab scale. Purification of DAG
oil using batchdeodorization and short path distillation yielded 93 wt % pure DAG
oil.Keywords: Diacylglycerol; glycerolysis; Novozym 435; optimization; response surface methodology;pilot
plant production; purification; short path distillation