Skimmilk crossflow microfiltration performance versus permeation flux to wall shear stress ratio

Comparative Assessment of Tubular Ceramic, Spiral Wound, and Hollow Fiber Membrane Microfiltration Module Systems for Milk Protein Fractionation

The fractionation efficiency of hollow fiber membranes (HFM) for milk protein fractionation was compared to ceramic tubular membranes (CTM) and spiral wound membranes (SWM). HFM combine the features of high membrane packing density of SWM and the more defined flow conditions and better control of membrane fouling in the open flow channel cross-sections of CTM. The aim was to comparatively analyze the effect of variations in local pressure and flow conditions while using single industrially sized standard modules with similar dimensions and module footprints (module diameter and length). The comparative assessment with varied transmembrane pressure was first applied for a constant feed volume flow rate of 20 m3 h-1 and, secondly, with the same axial pressure drop along the modules of 1.3 bar m-1, similar to commonly applied crossflow velocity and wall shear stress conditions at the industrial level. Flux, transmission factor of proteins (whey proteins and serum caseins), and specific protein mass flow per area membrane and per volume of module installed were determined as the evaluation criteria. The casein-to-whey protein ratios were calculated as a measure for protein fractionation effect. Results obtained show that HFM, which so far are under-represented as standard module types in industrial dairy applications, appear to be a competitive alternative to SWM and CTM for milk protein fractionation.

Prediction of the Limiting Flux and Its Correlation with the Reynolds Number during the Microfiltration of Skim Milk Using an Improved Model

Limiting flux (J_L) determination is a critical issue for membrane processing. This work presents a modified exponential model for J_L calculation, based on a previously published version. Our research focused on skim milk microfiltrations. The processing variables studied were the crossflow velocity (CFV), membrane hydraulic diameter (d_h), temperature, and concentration factor, totaling 62 experimental runs. Results showed that, by adding a new parameter called minimum transmembrane pressure, the modified model not only improved the fit of the experimental data compared to the former version (R² > 97.00%), but also revealed the existence of a minimum transmembrane pressure required to obtain flux (J). This result is observed as a small shift to the right on J versus transmembrane pressure curves, and this shift increases with the flow velocity. This fact was reported in other investigations, but so far has gone uninvestigated. The J_L predicted values were correlated with the Reynolds number (Re) for each d_h tested. Results showed that for a same Re; J_L increased as d_h decreased; in a wide range of Re within the turbulent regime. Finally, from dimensionless correlations; a unique expression J_L = f (Re, d_h) was obtained; predicting satisfactorily J_L (R² = 84.11%) for the whole set of experiments

Structural Characterisation of Deposit Layer during Milk Protein Microfiltration by Means of In-Situ MRI and Compositional Analysis

Milk protein fractionation by microfiltration membranes is an established but still growing field in dairy technology. Even under cross-flow conditions, this filtration process is impaired by the formation of a deposit by the retained protein fraction, mainly casein micelles. Due to deposition formation and consequently increased overall filtration resistance, the mass flow of the smaller whey protein fraction declines within the first few minutes of filtration. Currently, there are only a handful of analytical techniques available for the direct observation of deposit formation with opaque feed media and membranes. Here, we report on the ongoing development of a non-invasive and non-destructive method based on magnetic resonance imaging (MRI), and its application to characterise deposit layer formation during milk protein fractionation in ceramic hollow fibre membranes as a function of filtration pressure and temperature, temporally and spatially resolved. In addition, the chemical composition of the deposit was analysed by reversed phase high pressure liquid chromatography (RP-HPLC). We correlate the structural information gained by in-situ MRI with the protein amount and composition of the deposit layer obtained by RP-HPLC. We show that the combination of in-situ MRI and chemical analysis by RP-HPLC has the potential to allow for a better scientific understanding of the pressure and temperature dependence of deposit layer formation.

Technical Concepts for the Investigation of Spatial Effects in Spiral-Wound Microfiltration Membranes

Existing works on the influence of spatial effects on flux and permeation of proteins in microfiltration (MF) have focused on ceramic membranes. There is little information on spiral-wound membranes (SWMs). Since the inner core of a SWM is practically inaccessible by non-destructive techniques, three different prototypes were constructed in this study to optimize suitability for the investigation of spatial effects on filtration performance. To measure the pressure drop, shortened SWMs 0.25, 0.50, and 0.75 times the length of a standard industrial SWM (0.96 m) were designed. Second, a sectioned membrane (0.96 m) with separated compartments on the permeate side was constructed to analyze spatial effects on flux and protein permeation along the flow path of a SWM. Three different features characterized this sectioned module: sectioned permeate pockets, a sectioned permeate collection tube, and sectioned permeate drain and measurement systems. Crossflow filtration experiments showed that these modifications did not alter the filtration performance compared to an unmodified control SWM. Thus, it can be applied to assess spatially-resolved filtration performance in SWMs. The third prototype designed was a test cell with accessible flat sheet membranes and spacer material, as in SWMs. The flow path in this test cell was designed to match the characteristics of the channels between the membrane sheets in a standard SWM as closely as possible. The flow path length and the combination of membrane material and spacer architecture were the same as in the control SWM. This test cell was designed to assess the effects of length and processing conditions on the formation of a deposit layer. The combined results of these test modules can yield new insights into the spatial distribution of flux, permeation of target components, and deposit formation.

Milk Protein Fractionation by Means of Spiral-Wound Microfiltration Membranes: Effect of the Pressure Adjustment Mode and Temperature on Flux and Protein Permeation

Protein fractionation by means of microfiltration (MF) is significantly affected by fouling, especially when spiral-wound membranes (SWMs) are used. We investigated the influence of the mode of transmembrane pressure (Δp_TM) increase to target level and the deposit layer pressure history on the filtration performance during skim milk MF at temperatures of 10 °C and 50 °C. Two filtration protocols were established: No. 1: Δp_TM was set directly to various target values. No. 2: Starting from a low Δp_TM, we increased and subsequently decreased Δp_TM stepwise. The comparison of both protocols tested the effect of the mode of Δp_TM increase to target level. The latter protocol alone tested the effect of the deposit layer history with regard to the Δp_TM. As expected, flux and protein permeation were both found to be functions of the Δp_TM. Further, both measures were independent of the filtration protocol as long as Δp_TM was held at a constant level or, as part of protocol No. 2, Δp_TM was increased. Thus, we can state that the mode of Δp_TM increase to target level does not affect filtration performance in SWM. We found that after completion of a full cycle of stepping Δp_TM up from 0.5 bar to 3.0 bar and back down, flux and deposit layer resistance were not affected by the deposit layer history at 10 °C, but they were at 50 °C. Protein permeation, however, was lower for both 10 °C and 50 °C, when the Δp_TM cycle was completed. The processing history had a significant impact on filtration performance due to remaining structural compression effects in the deposited layer, which occur most notably at higher temperatures. Furthermore, temperatures of 50 °C lead to deposit layer aging, which is probably due to an enhanced crosslinking of particles in the deposit layer. Apart from that, we could show that fouling resistance does not directly correlate with protein permeation during skim milk MF using SWM.

Recovery of Lignosulfonates from Spent Sulfite Liquor Using Ceramic Hollow-Fiber Membranes

Spent sulfite liquor is an abundant but currently less used wastewater stream from the pulp and paper industry. The recovery of lignin from this resource would provide an inexpensive raw material for the manufacture of fuels and fine chemicals. Here we investigated the suitability of ceramic hollow-fiber membranes for the concentration of spent sulfite liquor as an alternative to common membrane technologies. We tested three ceramic hollow-fiber membranes (3, 8, and 30 nm) in different membrane processes (fed-batch and total recycle mode) and compared their performance with the widely-used tubular membrane geometry. We also evaluated backflushing as a strategy to reduce membrane fouling during filtration. The juxtaposition of the two membrane geometries revealed that wall shear stress is the most important process parameter for the assessment of membrane performance according to permeate flux. The higher the wall shear stress, the higher the permeate flux. Due to the smaller inner diameter of the hollow-fiber membranes, higher wall shear stress can be achieved more easily. Backflushing had no effect on the permeate flux during the concentration experiments.

Flux and transmission of β-casein during cold microfiltration of skim milk subjected to different heat treatments

Raw skim milk was subjected to different heat treatments: thermization (65°C, 20 s), pasteurization (72°C, 15 s), and no heat treatment (milk was microfiltered using 1.4-µm membranes at 50°C for bacteria removal; 1.4 MF). The milk (thermized, pasteurized, and 1.4 MF) was cooled and stored at 2°C until processing (at least 24 h) with cold (∼6°C) microfiltration using a benchtop crossflow pilot unit (Pall Membralox XLAB 5, Pall Corp., Port Washington, NY) equipped with 0.1-µm nominal pore diameter ceramic Membralox membrane (ET1-070, α-alumina, Pall Corp.). The flux was monitored during the process, and β-casein transmission and removal were calculated. The study aimed to indicate the conditions that should be applied to maximize β-casein passage through the membrane during cold microfiltration (5.6 ± 0.4°C) of skim milk. The proper selection of heat treatment parameters (temperature, time) of the feed before the cold microfiltration process will increase β-casein removal. It is not clear whether the difference in β-casein transmission between 1.4 MF, thermized, and pasteurized milk results from the effect of heat treatment conditions on β-casein dissociation from the casein micelles or on passage of β-casein through the membrane. The values of the major parameters (permeation flux and tangential flow velocity, through the wall shear stress) responsible for a proper membrane separation process were considerably lower than the critical values. It seems that the viscosity of the retentate has a great effect on the performance of the microfiltration membranes for protein separation at refrigerated temperatures.

Concentration of Immunoglobulins in Microfiltration Permeates of Skim Milk: Impact of Transmembrane Pressure and Temperature on the IgG Transmission Using Different Ceramic Membrane Types and Pore Sizes

The use of bioactive bovine milk immunoglobulins (Ig) has been found to be an alternative treatment for certain human gastrointestinal diseases. Some methodologies have been developed with bovine colostrum. These are considered in laboratory scale and are bound to high cost and limited availability of the raw material. The main challenge remains in obtaining high amounts of active IgG from an available source as mature cow milk by the means of industrial processes. Microfiltration (MF) was chosen as a process variant, which enables a gentle and effective concentration of the Ig fractions (ca. 0.06% in raw milk) while reducing casein and lactose at the same time. Different microfiltration membranes (ceramic standard and gradient), pore sizes (0.14–0.8 µm), transmembrane pressures (0.5–2.5 bar), and temperatures (10, 50 °C) were investigated. The transmission of immunoglobulin G (IgG) and casein during the filtration of raw skim milk (<0.1% fat) was evaluated during batch filtration using a single channel pilot plant. The transmission levels of IgG (~160 kDa) were measured to be at the same level as the reference major whey protein β-Lg (~18 kDa) at all evaluated pore sizes and process parameters despite the large difference in molecular mass of both fractions. Ceramic gradient membranes with a pore sizes of 0.14 µm showed IgG-transmission rates between 45% to 65% while reducing the casein fraction below 1% in the permeates. Contrary to the expectations, a lower pore size of 0.14 µm yielded fluxes up to 35% higher than 0.2 µm MF membranes. It was found that low transmembrane pressures benefit the Ig transmission. Upscaling the presented results to a continuous MF membrane process offers new possibilities for the production of immunoglobulin enriched supplements with well-known processing equipment for large scale milk protein fractionation.

Novel hybrid photocatalytic reactor-UF nanocomposite membrane system for bilge water degradation and separation

Bilge water degradation and separation via a hybrid photocatalytic membrane reactor based on a polyvinylidene fluoride (PVDF)/halloysite nanotube clay (HNTs) nanocomposite membrane has been achieved by this study.

Influence of casein on flux and passage of serum proteins during microfiltration using polymeric spiral-wound membranes at 50°C

Raw milk (approximately 1,800 kg) was separated at 4°C, pasteurized (at 72°C for 16s), and split into 2 batches. One batch (620 kg) was microfiltered (MF) using pilot-scale ceramic uniform transmembrane pressure Membralox membranes (model EP1940GL0.1 μA, 0.1-μm alumina; Pall Corp., East Hills, NY) to produce retentate and permeate. The permeate from the MF uniform transmembrane pressure was casein-free skim milk (CFSM). The CFSM was MF using polymeric spiral-wound (SW) membranes (model FG7838-OS0x-S, 0.3 μm; Parker-Hannifin Corp., Process Advanced Filtration Division, Tell City, IN) at a concentration factor of 3× and temperature of 50°C. Following the processing of CFSM, the second batch of skim milk (1,105 kg) was processed using the same polymeric membranes to determine how casein content in the feed material for MF with polymeric membranes affects the performance of the system. There was little resistance to passage of milk serum proteins (SP) through a 0.3-μm polyvinylidene fluoride (PVDF) SW membrane at 50°C and no detectable increase in hydraulic resistance of the membrane when processing CFSM. Therefore, milk SP contributed little, if any, to fouling of the PVDF membrane. In contrast, when processing skim milk containing a normal concentration of casein, the flux was much lower than when processing CFSM (17.2 vs. 80.2 kg/m(2) per hour, respectively) and the removal of SP from skim milk with a single-pass 3× bleed-and-feed MF system was also much lower than from CFSM (35.2 vs. 59.5% removal, respectively). Thus, when processing skim milk with a PVDF SW membrane, casein was the major protein foulant that increased hydraulic resistance and reduced passage of SP through the membrane.

Pilot-scale crossflow-microfiltration and pasteurization to remove spores of Bacillus anthracis (Sterne) from milk

High-temperature, short-time pasteurization of milk is ineffective against spore-forming bacteria such as Bacillus anthracis (BA), but is lethal to its vegetative cells. Crossflow microfiltration (MF) using ceramic membranes with a pore size of 1.4 μm has been shown to reject most microorganisms from skim milk; and, in combination with pasteurization, has been shown to extend its shelf life. The objectives of this study were to evaluate MF for its efficiency in removing spores of the attenuated Sterne strain of BA from milk; to evaluate the combined efficiency of MF using a 0.8-μm ceramic membrane, followed by pasteurization (72°C, 18.6s); and to monitor any residual BA in the permeates when stored at temperatures of 4, 10, and 25°C for up to 28 d. In each trial, 95 L of raw skim milk was inoculated with about 6.5 log(10) BA spores/mL of milk. It was then microfiltered in total recycle mode at 50°C using ceramic membranes with pore sizes of either 0.8 μm or 1.4 μm, at crossflow velocity of 6.2 m/s and transmembrane pressure of 127.6 kPa, conditions selected to exploit the selectivity of the membrane. Microfiltration using the 0.8-μm membrane removed 5.91±0.05 log(10) BA spores/mL of milk and the 1.4-μm membrane removed 4.50±0.35 log(10) BA spores/mL of milk. The 0.8-μm membrane showed efficient removal of the native microflora and both membranes showed near complete transmission of the casein proteins. Spore germination was evident in the permeates obtained at 10, 30, and 120 min of MF time (0.8-μm membrane) but when stored at 4 or 10°C, spore levels were decreased to below detection levels (≤0.3 log(10) spores/mL) by d 7 or 3 of storage, respectively. Permeates stored at 25°C showed coagulation and were not evaluated further. Pasteurization of the permeate samples immediately after MF resulted in additional spore germination that was related to the length of MF time. Pasteurized permeates obtained at 10 min of MF and stored at 4 or 10°C showed no growth of BA by d 7 and 3, respectively. Pasteurization of permeates obtained at 30 and 120 min of MF resulted in spore germination of up to 2.42 log(10) BA spores/mL. Spore levels decreased over the length of the storage period at 4 or 10°C for the samples obtained at 30 min of MF but not for the samples obtained at 120 min of MF. This study confirms that MF using a 0.8-μm membrane before high-temperature, short-time pasteurization may improve the safety and quality of the fluid milk supply; however, the duration of MF should be limited to prevent spore germination following pasteurization.

Relationship between types of surface shear stress profiles and membrane fouling

Shear stress has been recognized as an important parameter in controlling particle back-transport from membrane surfaces. However, little is known of the relationship between transient shear conditions induced by air sparging and fouling control near membrane surfaces. In this paper, the different types of surface shear stress profiles that had beneficial effects on minimizing reversible surface fouling were examined. The relationship between different statistical shear parameters (e.g. time-averaged shear, standard deviation of shear and amplitude of shear) and fouling control that have been used by others were examined as well. It was found that the fouling rate for membranes subjected to transient shear conditions was lower than for membranes subjected to constant shear conditions. The magnitude, duration and frequency of the shear conditions were found to have an impact on the fouling rate of membranes. It was also found that although some statistical shear parameters could generally be used to relate shear and fouling, they were inadequate to relate surface shear stress to fouling, for all transient shear conditions examined.

Microfiltration of Raw Whole Milk to Select Fractions with Different Fat Globule Size Distributions: Process Optimization and Analysis

We present an extensive description and analysis of a microfiltration process patented in our laboratory to separate different fractions of the initial milk fat globule population according to the size of the native milk fat globules (MFG). We used nominal membrane pore sizes of 2 to 12 microm and a specially designed pilot rig. Using this process with whole milk [whose MFG have a volume mean diameter (d43) = 4.2 +/- 0.2 microm] and appropriate membrane pore size and hydrodynamic conditions, we collected 2 extremes of the initial milk fat globule distribution consisting of 1) a retentate containing large MFG of d43 = 5 to 7.5 microm (with up to 250 g/kg of fat, up to 35% of initial milk fat, and up to 10% of initial milk volume), and 2) a permeate containing small MFG of d43 = 0.9 to 3.3 microm (with up to 16 g/kg of fat, up to 30% of initial milk fat, and up to 83% of initial milk volume and devoid of somatic cells). We checked that the process did not mechanically damage the MFG by measuring their zeta-potential. This new microfiltration process, avoiding milk aging, appears to be more efficient than gravity separation in selecting native MFG of different sizes. As we summarize from previous and new results showing that the physico-chemical and technological properties of native milk fat globules vary according to their size, the use of different fat globule fractions appears to be advantageous regarding the quality of cheeses and can lead to new dairy products with adapted properties (sensory, functional, and perhaps nutritional).