Concentration of dairy flavours using pervaporation : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Food Technology at Massey University, Auckland, New Zealand

Abstract

The food industry could potentially benefit from using pervaporation, a membrane process, to concentrate flavours. This research aimed to investigate its application for concentrating flavours in dairy process streams. Pervaporation experiments were carried out at a range of operating conditions, using hydrophobic membranes. The feed mixtures were either aqueous model solutions of dairy flavour compounds (acids, esters and ketones), complex model mixtures containing flavour compounds plus non-volatile dairy components, or real dairy products. Flavour compound enrichment factors ranged from below one to above 30, with esters and ketones being concentrated more effectively than acids. Thus, the flavours could be partially fractionated based on their chemical structure. The permeation of acids was reduced by approximately 50% when the feed pH was increased to near their p Ka values. For flavour compounds with lower molecular weights than approximately 1 20 g mol- I , permeation was controlled mainly by sorption i n the membrane; for larger compounds it was controlled mainly by diffusion through the membrane. The mass transfer of each flavour compound increased with temperature, following an Arrhenius-like relationship. The activation energy was a function of each compound's heat of sorption, its molecular weight, and the elastic modulus of the membrane. The activation energy was also related to the Arrhenius preexponential factor. Thus, fluxes could be estimated through empirical correlations. The non-volatile feed composition was an important factor influencing the pervaporation performance. Milk protein isolate (4% w/v) or lactose (6% or 1 2% w/v) bound with the flavour compounds in the feed, thus lowering the enrichment of sorption-controlled compounds. Milk fat (up to 38% w/v, in the form of cream ) reduced the enrichment of all the flavour compounds tested. Esters and ketones became unavailable for pervaporation as they partitioned into the fat phase; acids remained mainly in the aqueous phase, but their permeation was reduced because the added cream increased the feed pH. Experiments with real dairy products showed that pervaporation could be used to concentrate diacetylin starter distillate, and to selectively recover short-chain esters from ester cream. Of these two products, starter distillate is the more promising for use as a pervaporation feed stream.