Elsevier

Journal of Membrane Science

Volume 618, 15 January 2021, 118700
Journal of Membrane Science

Build-up and relaxation of membrane fouling deposits produced during crossflow ultrafiltration of casein micelle dispersions at 12 °C and 42 °C probed by in situ SAXS

https://doi.org/10.1016/j.memsci.2020.118700Get rights and content

Highlights

  • Formation and removal of casein micelle deposits were observed via in situ SAXS.

  • Deposits were compressed during filtration but swelled after pressure release.

  • Compression was partially irreversible: higher compression resulted in lower swelling.

  • Swelling was more limited at 12 °C than at 42 °C.

Abstract

Skim milk filtration is performed either at low or high temperature. However, there is still a lack of knowledge concerning the influence of temperature on membrane fouling. We used in situ small-angle X-ray scattering (SAXS) to study external membrane fouling during crossflow ultrafiltration of milk protein (casein micelle) dispersions at 12 °C and 42 °C. Casein micelle concentration distribution was measured in the concentration polarization layer and the deposit in a three-step filtration experiment that consisted of (i) fouling development step (with applied pressure and average cross-flow velocity fixed at 110 kPa and 3.1 cm s−1, respectively), (ii) pressure relaxation step (with pressure reduced to 10 kPa), and (iii) deposit erosion step (with crossflow velocity increased to 15.6 cm s−1).

Despite a higher average filtrate flux obtained at 42 °C, filtration at 42 °C resulted in the formation of a thicker and more concentrated deposit than filtration at 12 °C. At both temperatures studied, subsequent pressure relaxation resulted in deposit swelling, which was more intense in the less-compressed external part of the deposit. Deposit swelling rate was significantly higher at 42 °C than at 12 °C. The swelled deposit obtained at 42 °C was more eroded by the crossflow compared to the poorly swelled deposit obtained at 12°. This suggests that deposit formation and compression were more reversible at 42 °C than at 12 °C.

Introduction

Milk filtration became an essential dairy industry process with the development of membrane technology, first with tubular ceramic membranes and, later, spiral-wound polymer membranes [1]. For example, microfiltration is used to separate casein micelles from serum proteins, and ultrafiltration is used to concentrate the milk proteins. Milk is filtered at high (50–55 °C) or low (9–15 °C) temperatures to curb the bacterial growth that leads to milk spoilage and equipment contamination [2]. High-temperature filtration generally uses ceramic membranes, which although expensive are preferred for their better cleanability and durability, while low-temperature filtration uses cheaper but less durable polymeric membranes.

Membrane fouling degrades the efficiency of both high- and low-temperature milk filtration, but it is apparently inevitable given that the composition of milk can give rise to different foulant–foulant and membrane–foulant interactions [2]. Along with bacterial development, membrane fouling obliges to interrupt a milk filtration cycle (which generally lasts about 6–8 h with ceramic membrane filtration and up to 15 h with polymeric membrane filtration) for 2–3 h of cleaning operations that include several physical and chemical membrane cleaning steps [1,3,4]. A recent environmental performance assessment of current cleaning protocols found that they demand excessive amounts of time, energy, water and chemicals [5]. Fouling control and optimization of membrane cleaning are therefore ongoing challenges for the dairy industry that require research into fouling mechanisms, foulant–foulant and foulant–membrane interactions (e.g. properties of fouling deposits), and conditions governing fouling development and removal.

Despite a large body of research into membrane fouling and membrane cleaning in milk filtration (e.g. Refs. [[2], [3], [4],[6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28]]), they are indeed not sufficiently studied. It is difficult to analyze and explain the published data on milk filtration: for example, different conclusions can be found about the role of transmembrane pressure ([10] and [6,7,13,17,[29], [30], [31], [32], [33], [34]]) and temperature ([[12], [19]] and [33]) in membrane fouling; even the qualitative composition of the fouling layer is under discussion ([13,25,31] and [21,35]), etc.

One reason for this uncertainty is that many of previous studies have focused on parameter averages to characterize membrane fouling (average filtrate flux and membrane resistance, e.g. Ref. [12]). Only few researchers studied local membrane fouling in milk filtration: e.g. Refs. [10,80,82,83] measured local fouled membrane resistance, while [50] directly characterized local deposit layer (also, there were reports on characterization of local membrane fouling by milk with the help of complex optical methods [3,15,81]). Moreover, very different flux reducing effects can occur during membrane filtration: for example, it can be expected that microfiltration with more open porous membranes will be affected more by pore blocking than ultrafiltration. Unfortunately, the membrane fouling mechanism is not always discussed or validated. Also, there are other difficulties of reaching sufficient precision in the data presentation and analysis (as detailed in Appendix). Therefore, apparent contradictions between conclusions made in different studies on the basis of raw data comparison should not be discussed as conflicting, but rather yielded by difference (usually, undefined) in fouling mechanisms, operating conditions or properties of studied fluids. It can be expected that these contradictions can be avoided, if membrane fouling phenomena are discussed in addition to basic filtration data. Therefore, rigorous characterization of membrane fouling needs to consider the conditions governing fouling development, the type of fouling and its properties (e.g. spatio-temporal evolution of the concentration polarization layer and deposit), and the kinetics of fouling removal.

The present study is focused on characterizing the local membrane fouling by deposit produced during the ultrafiltration of casein micelle dispersions. Casein micelles are the main colloidal constituents of skim milk. According to Ref. [31], during skim milk filtration, casein micelles play a central role in “irreversible membrane fouling” (where “irreversibility” is defined according to Refs. [[7], [76]]) via the formation of “irreversible” casein micelle deposit under critical filtration conditions (in line with previous theoretical considerations [36] and further experimental findings [37]).

Casein micelles represent about 80 wt% of milk proteins. They are roughly spherical colloidal agglomerates with diameter ranging from 50 nm to 500 nm and an average diameter of about 150–200 nm [38,[84], [85], [86]], whereas other milk proteins are dissolved in serum as monomers or dimers. The micelles are core-shell-type particles composed of assemblies of αs1-, αs2- and β-casein molecules connected with nanoparticles of calcium phosphate and covered with a κ-casein brush lending them stability against aggregation. Casein micelles are soft colloids, which means they are highly porous, deformable, and can shrink under compressive pressure and re-swell after pressure release [39,40].

Concerning the properties of casein micelles at high concentrations [[41], [42], [43]], studied osmotic compression of casein micelle dispersions and demonstrated that casein micelles form gels when the critical sol–gel transition concentration and the corresponding critical osmotic pressure are exceeded. Once the compressive pressure is released, these gels swell and redisperse over the course of time. In highly-compressed gels, concentration of gel–sol transition is lower than that sol–gel transition [40] (this demonstrates partial irreversibility of compression of casein micelles gels obtained via osmotic compression). In the later study on membrane filtration of casein micelle dispersions [37] it was confirmed that a critical minimal transmembrane pressure is required in order to obtain persistent membrane fouling (that was explained by deposit formation). The results of dynamic compression of fouled membranes also suggested that the deposit of casein micelles is reversibly compressible (i.e. it gets compressed and re-swells after the change of applied pressure [37] similarly to gels of casein micelles obtained via osmotic compression [40]).

Besides, a series of work from the Technical University of Munich [8,11,32,33] systematically demonstrated that removal of fouling during membrane rinsing after filtration of skim milk is a continuous process. It was also suggested that the kinetics of fouling removal is related to deposit compression (which slows down fouling removal) and diffusion of foulant into the bulk (which was assumed to be the mechanism of fouling removal). The evidence from Refs. [8,11,32,33] and [37,40] suggests that deposit swelling must take place in parallel with redispersion (diffusion, in terms of [8]) and thus play a role in fouling removal.

Therefore, in the case of formation and removal of casein micelle deposit, meaningful discussion of fouling should be based on quantitative characterization of deposit properties, i.e. concentration profile, permeability, compressibility, reversibility of compression, and cohesiveness (as previously done for gels obtained by osmotic compression in Refs. [41,42] and dead-end filtration in Ref. [44]).

Decreasing temperature (<12 °C) results in the partial release of β-casein and calcium phosphate from casein micelles into the serum phase [45]. It also increases casein micelle voluminosity υ (the ratio of particle volume to particle dry weight): according to Refs. [46,47], decreasing the temperature from 70 °C to 5 °C increases υ from 3.5 ml g−1 to 5.0 ml g−1. These temperature-related phenomena may be important for milk filtration due to their possible influence on the membrane fouling. For example, it was demonstrated that compressibility of casein micelles dispersions obtained by compression at constant osmotic pressure significantly changes in the temperature range 7–20 °C (i.e. at constant osmotic pressure, casein micelle concentration increases with temperature) [41]. Therefore, temperature can influence the compressibility of casein micelle deposits on the membrane surface.

However, to the best of the authors’ knowledge, the influence of filtration temperature on the properties (permeability, compressibility, reversibility of compression, cohesiveness) of casein micelles deposits produced in crossflow mode has never been investigated. Here we premise that this influence can be characterized via proper analysis of local solid concentration distribution in deposits obtained at different filtration temperatures. Previously, it was demonstrated that the required concentration distribution for casein micelles (as well as other colloids) can be measured in situ with a spatial resolution of tens of microns via small-angle X-ray scattering (SAXS) [[48], [49], [50]]. These previous studies used specific X-ray-adapted crossflow filtration cells to capture the time-course and spatial evolutions of casein micelle concentrations during the process under controlled crossflow, transmembrane pressure (TMP) and temperature. Using these purpose-designed filtration cells together with a highly collimated X-ray beam, it was managed to capture this time-resolved spatial information and its process time-course kinetics within the concentration polarization layers and deposits, in the vicinity of the membrane surface at distances down to a few hundred micrometers with a precision of 20 μm.

In the present article we combined data on local deposit concentration obtained via in situ SAXS–filtration method with the filtration flux measurements in a multi-step filtration experiment set up to vary TMP and crossflow velocity to shed light on the influence of filtration temperature on the formation and reversibility of casein micelle deposits.

The experiments were carried out at 12 °C and 42 °C (close to typical industry temperatures for skim milk filtration) using casein micelles dispersed in skim milk ultrafiltrate. SAXS–filtration data were analyzed, deposit removal was described, and the relationship between average (hydraulic resistance of fouled membrane) and local (deposit properties) data was discussed.

Section snippets

Casein micelle dispersions

Casein micelle dispersions were prepared from casein micelle powder and milk ultrafiltrate (UF permeate). Casein micelle powder Promilk 852B was provided by Ingredia (Arras, France) and was obtained by microfiltration of pasteurized skim milk with a membrane pore size of 0.1 μm. UF permeate was prepared at the STLO laboratory by ultrafiltration of a fresh skim milk at 12 °C through a membrane with a nominal molecular weight cut-off of 5 kDa. Thiomersal (0.02 wt%) and sodium azide (0.05 wt%)

Casein micelle accumulation during the filtration step

Fig. 4 presents average filtrate flux obtained during crossflow filtration of casein micelle dispersions at 12 °C and 42 °C (both at the constant transmembrane pressure of 110 kPa).

At both temperatures studied, the flux quickly dropped at the beginning of filtration (at tF < 10 min) and then gradually reached the steady-state value (practically constant within the experimental error at tF > 50 min), which was 1.6 times higher at 42 °C than at 12 °C. This result is in the agreement with a number

Conclusions

Crossflow ultrafiltration of casein micelle dispersions was studied at 12 °C and 42 °C by in situ SAXS, enabling us to time-resolved spatial information on the local concentration in the vicinity of the membrane surface at distances down to a few hundred micrometers with a resolution of 20 μm. Local solid concentration distribution and average filtrate flux were measured and analyzed at different experiment steps (filtration, pressure relaxation, erosion).

Filtrate flux was higher at 42 °C than

Author Statement

Each co-author made equally important input to this article

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

We thank the French National Research Institute for Agriculture, Food and Environment (INRAE) and the Brittany Regional Council for providing valuable financial support. We also thank the SOLEIL Synchrotron for providing synchrotron beam time and further financial support (project ID: 2017023). We thank Glen McCulley for invaluable help preparing the manuscript.

LRP is part of PolyNat Carnot Institute (‘Investissements d'Avenir’–grant agreement #ANR-11-CARN-030-01), Labex TEC 21

References (86)

  • H. Bouzid et al.

    Impact of zeta potential and size of caseins as precursors of fouling deposit on limiting and critical fluxes in spiral ultrafiltration of modified skim milks

    J. Membr. Sci.

    (2008)
  • D. Delaunay et al.

    Mapping of protein fouling by FTIR-ATR as experimental tool to study membrane fouling and fluid velocity profile in various geometries and validation by CFD simulation

    Chem. Eng. Process

    (2008)
  • W. Zhang et al.

    Membrane cleaning assisted by high shear stress for restoring ultrafiltration membranes fouled by dairy wastewater

    Chem. Eng. J.

    (2017)
  • W. Youravong et al.

    Critical flux in ultrafiltration of skimmed milk

    Food Bioprod. Process.

    (2003)
  • S. Methot-Hains et al.

    Effect of transmembrane pressure control on energy efficiency during skim milk concentration by ultrafiltration at 10 and 50°C

    J. Dairy Sci.

    (2016)
  • K.S.Y. Ng et al.

    Influence of processing temperature on flux decline during skim milk ultrafiltration

    Separ. Purif. Technol.

    (2018)
  • D. Tremblay-Marchand et al.

    A process efficiency assessment of serum protein removal from milk using ceramic graded permeability microfiltration membrane

    J. Dairy Sci.

    (2016)
  • M. Koutake et al.

    Osmotic pressure model of membrane fouling applied to the ultrafiltration of whey

    J. Food Eng.

    (1993)
  • T.J. Tan et al.

    A physicochemical investigation of membrane fouling in cold microfiltration of skim milk

    J. Dairy Sci.

    (2014)
  • S. Popovic et al.

    Twisted tapes as turbulence promoters in the microfiltration of milk

    J. Membr. Sci.

    (2011)
  • D.M. Krstic et al.

    The effect of turbulence promoter on cross-flow microfiltration of skim milk

    J. Membr. Sci.

    (2002)
  • G. Gésan-Guiziou et al.

    Cake properties in dead-end ultrafiltration of casein micelles: determination of critical operating conditions

    Desalination

    (2006)
  • A.J.E. Jimenez-Lopez et al.

    Role of milk constituents on critical conditions and deposit structure in skimmilk microfiltration (0.1 μm)

    Separ. Purif. Technol.

    (2008)
  • N.W. Diagne et al.

    Cleanability versus limiting and critical fluxes of a polyethersulfone membrane of skim milk ultrafiltration

    Procedia Engineer

    (2012)
  • P. Bacchin et al.

    A unifying model for concentration polarization, gel-layer formation and particle deposition in cross-flow membrane filtration of colloidal suspensions

    Chem. Eng. Sci.

    (2002)
  • P. Qu et al.

    Dead-end filtration of sponge-like colloids: the case of casein micelle

    J. Membr. Sci.

    (2012)
  • C.G. De Kruif

    Supra-aggregates of casein micelles as a prelude to coagulation

    J. Dairy Sci.

    (1998)
  • A. Bouchoux et al.

    How to squeeze a sponge: casein micelles under osmotic stress, a SAXS study

    Biophys. J.

    (2010)
  • P. Qu et al.

    On the cohesive properties of casein micelles in dense systems

    Food Hydrocolloids

    (2015)
  • A. Bouchoux et al.

    Casein micelle dispersions under osmotic stress

    Biophys. J.

    (2009)
  • S. Nobel et al.

    Apparent voluminosity of casein micelles determined by rheometry

    J. Colloid Interface Sci.

    (2012)
  • S. Nobel et al.

    Apparent voluminosity of casein micelles in the temperature range 35-70°C

    Int. Dairy J.

    (2016)
  • Y. Jin et al.

    Effects of ultrasound on cross-flow ultrafiltration of skim milk: characterization from macro-scale to nano-scale

    J. Membr. Sci.

    (2014)
  • Y. Jin et al.

    Modeling and analysis of concentration profiles obtained by in-situ SAXS during cross-flow ultrafiltration of colloids

    J. Membr. Sci.

    (2017)
  • C. Gaiani et al.

    The dissolution behaviour of native phosphocaseinate as a function of concentration and temperature using a rheological approach

    Int. Dairy J.

    (2006)
  • D.J. Kapsimalis et al.

    Ultrafiltration of skim milk at refrigerated temperatures

    J. Dairy Sci.

    (1981)
  • A. Makardij et al.

    Microfiltration and ultrafiltration of milk: some aspects of fouling and cleaning

    Food Bioprod. Process.

    (1999)
  • S.M.A. Razavi et al.

    Dynamic prediction of milk ultrafiltration performance: a neural network approach

    Chem. Eng. Sci.

    (2003)
  • G. Samuelsson et al.

    Minimizing whey protein retention in cross-flow microfiltration of skim milk

    Int. Dairy J.

    (1997)
  • G. Samuelsson et al.

    Predicting limiting flux of skim milk in crossflow microfiltration

    J. Membr. Sci.

    (1997)
  • N.D. Lawrence et al.

    Microfiltration of skim milk using polymeric membranes for casein concentrate manufacture

    Separ. Purif. Technol.

    (2008)
  • N.A. McCarthy et al.

    Pilot-scale ceramic membrane filtration of skim milk for the production of a protein base ingredient for use in infant milk formula

    Int. Dairy J.

    (2017)
  • M. Rabiller-Baudry et al.

    A dual approach of membrane cleaning based on physico-chemistry and hydrodynamics. Application to PES membrane of dairy industry

    Chem. Eng. Process

    (2008)
  • Cited by (13)

    • Assessment of pervaporative concentration of dairy solutions vs ultrafiltration, nanofiltration and reverse osmosis

      2022, Separation and Purification Technology
      Citation Excerpt :

      The dispersion characteristics and colloidal properties of casein micelles relevant to concentration change due to filtration have also been investigated by Bouchoux and co-workers [27–30]. Of particular interest are the findings that casein micelles are soft colloids with a high porosity and deformability, allowing for shrinkage under pressure and expansion after pressure release [30–32], which ultimately affect the packing behavior and gel formation of casein micelles relevant to membrane fouling and gel layer formation during ultrafiltration of dairy streams [33–36]. While detailed studies of the composition profile of the concentrated milk retentate and associated physicochemistry are beyond the scope of the present study, the RO and PV have shown to perform better than UF and NF in terms of solute retentions.

    • NaCl precleaning of microfiltration membranes fouled with oil-in-water emulsions: Impact on fouling dislodgment

      2022, Separation and Purification Technology
      Citation Excerpt :

      In situ investigation at a local scale can offer highly valuable insights into the fouling process [30], governed by short-range interactions between the oil, the membrane, the surfactants and the salts. New methods are being developed to describe the membrane fouling with online analytical methods [31-34], and Tanudjaja et al [32,35] showed that the on-ionic of oil droplets is far from solid particles. The resolution near 1 µm allowed on-ionic large droplets and the impact of hydrodynamics but the role of local interaction important for small droplets could not be considered.

    • Effect of flow channel number in multi-channel tubular ceramic microfiltration membranes on flux and small protein transmission in milk protein fractionation

      2022, Journal of Membrane Science
      Citation Excerpt :

      Deposit formation by suspended biopolymers and the effects on protein fractionation was not investigated in these studies. However, various works showed that high ΔpTM levels at first increase the flux and the transport of material towards the membrane, resulting in a more intense deposit layer formation up to the level of complete compaction at ΔpTM levels above the critical limit, leading to deposit controlled permeation and high whey protein retention [18–23]. This should also apply to the radial distribution of ΔpTM between outer and inner flow channels in multi-channel elements.

    View all citing articles on Scopus
    View full text