Process analyses and characterization of liquid virgin whey protein isolate
No Access Until
In the food industry, native whey proteins (WP) are desirable because of their high nutritional quality and excellent functional properties. In this study, virgin whey (VW) was harvested as permeate in the microfiltration of slightly acidified skim milk prior to cheesemaking. Free of cheesemaking remnants, bacteria and spores, VW did not require pretreatment before concentration by ultrafiltration (UF). Not exposed to extreme physicochemical conditions of cheesemaking, the WP in VW retained their native conformation. Therefore, both protein-protein and membrane-protein interactions were minimal during UF, enabling VW concentration by UF alone at reasonable flux. This allowed the production of liquid virgin whey protein isolate (LVWPI) containing up to 26% total solids, about 91% of which was WP.
The LVWPI is a novel ingredient rich in native WP and of low mineral content. It showed unique physicochemical properties and functional behavior not observed in commercial WP products. It exhibited low viscosity and thermal stability against rapid aggregation that led to controlled heat-induced aggregation and gelation suitable for fine-tuned food texturization. It was produced by concentrating VW at 45 ?C using pilot-scale two-stage UF system with polysulfone membranes (10-kDa molecular weight cut-off). VW was first concentrated ~13x in a spiral wound module (SWM), and then diafiltered to achieve ~99% lactose removal before further concentrating ~5x in a hollow fiber module. SWM flux data showed as much as six times increase compared to those observed in the UF of cheese whey, resulting in lower process energy requirements.
To understand the unique UF fouling behavior of VW, a two-parameter flux model was derived. One parameter, expressed as the ratio of feed stream (F) to membrane area (A), quantified membrane-protein interactions that give rise to initial flux decline. Another is the long-term fouling parameter, m, which indicated protein-protein interactions. Results showed that m was constant, regardless of F value, due to VW?s consistent composition. However, initial flux decline depended on F/A. The model proved to be a practical design equation for optimum F/A in UF systems. Finally, a technology transfer model was designed wherein a developing country benefits from the LVWPI technology developed in this study.