A novel approach for the production of soy protein isolates was investigated integrating electroacidification and membrane ultrafiltration. The effect of electroacidification on the ultrafiltration performance and physicochemical properties of the soy protein extracts was obtained by comparing an electroacidified (pH 6) and a non-electroacidified (pH 9) soy protein extract. The effect of membrane fouling on the permeate flux decline was studied in a hollow fiber and a dead end ultrafiltration system. Due to more significant membrane fouling, the permeate flux was always lower for the electroacidified extract, resulting in at least 1.5-fold increase in the total fouling resistance compared to the non-electroacidified extract. The total amount of protein deposited on the membrane surface during unstirred dead-end ultrafiltration was comparable (about 7 mg/cm2) for both soy protein extracts. The discrepancy between the total fouling resistance and the protein deposition estimates was attributed to the formation of denser (less permeable) fouling deposit for the electroacidified extract, which was supported by scanning electron microscopy studies of fouled membranes. The removal of carbohydrates and minerals was evaluated for direct ultrafiltration and two-stage discontinuous diafiltration using a hollow fiber system. The carbohydrate removal results were always consistent with the theoretical predictions, indicating that the carbohydrates were freely permeable across the membrane. In contrast, the minerals were partially retained by the membrane, but to a higher extent for the non-electroacidified extract, which demonstrated that the electroacidification pretreatment enhanced the mineral removal during the ultrafiltration. Incorporation of the diafiltration step improved the ash (mineral) and carbohydrate removal. Stronger electrostatic interactions between soy proteins, calcium/magnesium, and phytic acid (antinutrient) at alkaline pH resulted in less efficient removal of calcium, magnesium, and phytic acid during the ultrafiltration of the non-electroacidified extract compared to the electroacidified extract. Consequently, the soy protein isolates produced by electroacidification and the hollow fiber ultrafiltration had a lower mineral and phytic acid content. The protein content was at least 88 % (dry basis), with or without the electroacidification pretreatment. The study of the viscosity revealed that the electroacidification pretreatment reduced the viscosity of the soy protein extract, which resulted in a lower axial pressure drop increase during the ultrafiltration of the electroacidified extract compared to the non-electroacidified extract. Adjusting the pH of the electroacidified extract to 9 and the pH of the non-electroacidified extract to 6 had a great impact on the particle size distribution but only a marginal effect on the viscosity of the pH adjusted extracts. This indicated that the pH and the particle size distribution were not responsible for the viscosity difference between the electroacidified and the non-electroacidified soy protein extracts. It was proposed that the electroacidification pretreatment had some impact on the water hydration capacity of the soy proteins, which consequently affected the viscosity.

A novel approach for the production of soy protein isolates was investigated integrating electroacidification and membrane ultrafiltration. The effect of electroacidification on the ultrafiltration performance and physicochemical properties of the soy protein extracts was obtained by comparing an electroacidified (pH 6) and a non-electroacidified (pH 9) soy protein extract. The effect of membrane fouling on the permeate flux decline was studied in a hollow fiber and a dead end ultrafiltration system. Due to more significant membrane fouling, the permeate flux was always lower for the electroacidified extract, resulting in at least 1.5-fold increase in the total fouling resistance compared to the non-electroacidified extract. The total amount of protein deposited on the membrane surface during unstirred dead-end ultrafiltration was comparable (about 7 mg/cm2) for both soy protein extracts. The discrepancy between the total fouling resistance and the protein deposition estimates was attributed to the formation of denser (less permeable) fouling deposit for the electroacidified extract, which was supported by scanning electron microscopy studies of fouled membranes. The removal of carbohydrates and minerals was evaluated for direct ultrafiltration and two-stage discontinuous diafiltration using a hollow fiber system. The carbohydrate removal results were always consistent with the theoretical predictions, indicating that the carbohydrates were freely permeable across the membrane. In contrast, the minerals were partially retained by the membrane, but to a higher extent for the non-electroacidified extract, which demonstrated that the electroacidification pretreatment enhanced the mineral removal during the ultrafiltration. Incorporation of the diafiltration step improved the ash (mineral) and carbohydrate removal. Stronger electrostatic interactions between soy proteins, calcium/magnesium, and phytic acid (antinutrient) at alkaline pH resulted in less efficient removal of calcium, magnesium, and phytic acid during the ultrafiltration of the non-electroacidified extract compared to the electroacidified extract. Consequently, the soy protein isolates produced by electroacidification and the hollow fiber ultrafiltration had a lower mineral and phytic acid content. The protein content was at least 88 % (dry basis), with or without the electroacidification pretreatment. The study of the viscosity revealed that the electroacidification pretreatment reduced the viscosity of the soy protein extract, which resulted in a lower axial pressure drop increase during the ultrafiltration of the electroacidified extract compared to the non-electroacidified extract. Adjusting the pH of the electroacidified extract to 9 and the pH of the non-electroacidified extract to 6 had a great impact on the particle size distribution but only a marginal effect on the viscosity of the pH adjusted extracts. This indicated that the pH and the particle size distribution were not responsible for the viscosity difference between the electroacidified and the non-electroacidified soy protein extracts. It was proposed that the electroacidification pretreatment had some impact on the water hydration capacity of the soy proteins, which consequently affected the viscosity.