Effect of ceramic membrane channel diameter on limiting retentate protein concentration during skim milk microfiltration.

Our objective was to determine the effect of retentate flow channel diameter (4 or 6mm) of nongraded permeability 100-nm pore size ceramic membranes operated in nonuniform transmembrane pressure mode on the limiting retentate protein concentration (LRPC) while microfiltering (MF) skim milk at a temperature of 50°C, a flux of 55 kg · m(-2) · h(-1), and an average cross-flow velocity of 7 m · s(-1). At the above conditions, the retentate true protein concentration was incrementally increased from 7 to 11.5%. When temperature, flux, and average cross-flow velocity were controlled, ceramic membrane retentate flow channel diameter did not affect the LRPC. This indicates that LRPC is not a function of the Reynolds number. Computational fluid dynamics data, which indicated that both membranes had similar radial velocity profiles within their retentate flow channels, supported this finding. Membranes with 6-mm flow channels can be operated at a lower pressure decrease from membrane inlet to membrane outlet (ΔP) or at a higher cross-flow velocity, depending on which is controlled, than membranes with 4-mm flow channels. This implies that 6-mm membranes could achieve a higher LRPC than 4-mm membranes at the same ΔP due to an increase in cross-flow velocity. In theory, the higher LRPC of the 6-mm membranes could facilitate 95% serum protein removal in 2 MF stages with diafiltration between stages if no serum protein were rejected by the membrane. At the same flux, retentate protein concentration, and average cross-flow velocity, 4-mm membranes require 21% more energy to remove a given amount of permeate than 6-mm membranes, despite the lower surface area of the 6-mm membranes. Equations to predict skim milk MF retentate viscosity as a function of protein concentration and temperature are provided. Retentate viscosity, retentate recirculation pump frequency required to maintain a given cross-flow velocity at a given retentate viscosity, and retentate protein determination by mid-infrared spectrophotometry were all useful tools for monitoring the retentate protein concentration to ensure a sustainable MF process. Using 6-mm membranes instead of 4-mm membranes would be advantageous for processors who wish to reduce energy costs or maximize the protein concentration of a MF retentate.