Processing factors that influence casein and serum protein separation by microfiltration1

Our objective was to demonstrate the effect of various processing factors on the performance of a microfiltration system designed to process skim milk and separate casein (CN) from serum proteins (SP). A mathematical model of a skim milk microfiltration process was developed with 3 stages plus an additional fourth finishing stage to standardize the retentate to 9% true protein (TP) and allow calculation of yield of a liquid 9% TP micellar CN concentrate (MCC) and milk SP isolate (MSPI; 90% SP on a dry basis). The model was used to predict the effect of 5 factors: 1) skim milk composition, 2) heat treatment of skim milk, 3) concentration factor (CF) and diafiltration factor (DF), 4) control of CF and DF, and 5) SP rejection by the membrane on retentate and permeate composition, SP removal, and MCC and MSPI yield. When skim milk TP concentration increased from 3.2 to 3.8%, the TP concentration in the third stage retentate increased from 7.92 to 9.40%, the yield of MCC from 1,000 kg of skim milk increased from 293 to 348 kg, and the yield of MSPI increased from 6.24 to 7.38 kg. Increased heat treatment (72.9 to 85.2°C) of skim milk caused the apparent CN as a percentage of TP content of skim milk as measured by Kjeldahl analysis to increase from 81.97 to 85.94% and the yield of MSPI decreased from 6.24 to 4.86 kg, whereas the third stage cumulative percentage SP removal decreased from 96.96 to 70.08%. A CF and DF of 2× gave a third stage retentate TP concentration of 5.38% compared with 13.13% for a CF and DF of 5×, with the third stage cumulative SP removal increasing from 88.66 to 99.47%. Variation in control of the balance between CF and DF (instead of an equal CF and DF) caused either a progressive increase or decrease in TP concentration in the retentate across stages depending on whether CF was greater than DF (increasing TP in retentate) or CF was less than DF (decreasing TP in retentate). An increased rejection of SP by the membrane from an SP removal factor of 1 to 0.6 caused a reduction in MSPI yield from 6.24 to 5.19 kg/1,000 kg of skim milk, and third stage cumulative SP removal decreased from 96.96 to 79.74%. Within the ranges of the 5 factors studied, the TP content of the third stage retentate was most strongly affected by the target CF and DF and variation in skim milk composition. Cumulative percentage SP removal was most strongly affected by the heat treatment of skim milk, the SP removal factor, and the target CF and DF. The MCC yield was most strongly affected by initial skim milk composition. Yield of MSPI was strongly affected by skim milk composition, whereas the heat treatment of milk and SP removal factor also had a large effect.