Production efficiency of micellar casein concentrate using polymeric spiral-wound microfiltration membranes

Most current research has focused on using ceramic microfiltration (MF) membranes for micellar casein concentrate production, but little research has focused on the use of polymeric spiral-wound (SW) MF membranes. A method for the production of a serum protein (SP)–reduced micellar casein concentrate using SW MF was compared with a ceramic MF membrane. Pasteurized (79°C, 18s) skim milk (1,100 kg) was microfiltered at 50°C [about 3 × concentration] using a 0.3-μm polyvinylidene fluoride spiral-wound membrane, bleed-and-feed, 3-stage process, using 2 diafiltration stages, where the retentate was diluted 1:2 with reverse osmosis water. Skim milk, permeate, and retentate were analyzed for SP content, and the reduction of SP from skim milk was determined. Theoretically, 68% of the SP content of skim milk can be removed using a single-stage 3× MF. If 2 subsequent water diafiltration stages are used, an additional 22% and 7% of the SP can be removed, respectively, giving a total SP removal of 97%. Removal of SP greater than 95% has been achieved using a 0.1-μm pore size ceramic uniform transmembrane pressure (UTP) MF membrane after a 3-stage MF with diafiltration process. One stage of MF plus 2 stages of diafiltration of 50°C skim milk using a polyvinylidene fluoride polymeric SW 0.3-μm membrane yielded a total SP reduction of only 70.3% (stages 1, 2, and 3: 38.6, 20.8, and 10.9%, respectively). The SP removal rate for the polymeric SW MF membrane was lower in all 3 stages of processing (stages 1, 2, and 3: 0.05, 0.04, and 0.03 kg/m2 per hour, respectively) than that of the comparable ceramic UTP MF membrane (stages 1, 2, and 3: 0.30, 0.11, and 0.06 kg/m2 per hour, respectively), indicating that SW MF is less efficient at removing SP from 50°C skim milk than the ceramic UTP system. To estimate the number of steps required for the SW system to reach 95% SP removal, the third-stage SP removal rate (27.4% of the starting material SP content) was used to extrapolate that an additional 5 water diafiltration stages would be necessary, for a total of 8 stages, to remove 95% of the SP from skim milk. The 8-plus stages necessary to remove >95% SP for the SW MF membrane would create more permeate and a lengthier process than required with ceramic membranes.