Review

Current developments of microfiltration technologyin the dairy industry

Luciana V. SABOYAa, Jean-Louis MAUBOISb*

a Escola Superior de Agricultura “Luiz de Queiroz” (ESALQ – USP),Av. Pádua Dias 11, 13418-900/Piracicaba, SP, Brasil

b Laboratoire de Recherches de Technologie Laitière, INRA,65 rue de Saint Brieuc, 35042 Rennes Cedex, France

(Received 2 May 2000; accepted 27 September 2000)

Abstract — This paper deals with the most recent developments of cross flow microfiltration(CFMF), some of them just patented, in the dairy industry. Combination of the use of uniform trans-membrane hydraulic pressure concept (UTP) with its different ways of carrying out microfiltraterecirculation, longitudinal porosity gradient and of new ceramic membrane materials allows nowa-days to get a differential separation of each category of milk particles. Pretreatment by CFMF ofincoming milk is used for the production of low heated fluid milks having a flavor similar to that ofraw milk and a shelf life 3 to 5 times longer than that of classical products. Results observed in theretention by the MF membrane of pathogenic bacterial species allow it to be said that the hygienic safetyof cheeses made from raw MF milk is, at least, equal or even higher than that of cheeses made frompasteurized milk. Numerous other applications of CFMF are under development such as removal ofresidual fat from whey or the clarification and the removal of bacteria from cheese brine but themost promising are undoubtedly the differential separation of micellar casein and of small milk fatglobules. With the products obtained on both sides of the MF membrane, dairy technology will havethe possibility not only to improve yield and quality of many dairy products but also to create adiversified range of new textures.

microfiltration / membrane / liquid milk / native casein / brine / whey / cheese

Résumé — Développements actuels de la technologie de microfiltration dans l’industrie laitière.Cet article traite des développements les plus récents, certains tout juste brevetés, de la microfiltra-tion tangentielle dans l’industrie laitière. La combinaison de l’utilisation du concept de pressiontransmembranaire uniforme, avec ses différentes possibilités de mise en œuvre : recirculation du micro-filtrat, gradient de porosité longitudinal, et de nouveaux matériaux membranaires céramiques

Lait 80 (2000) 541–553 541© INRA, EDP Sciences

* Correspondence and reprintsTel.: (33) 2 23 48 53 22; fax: (33) 2 23 48 53 50; e-mail: maubois@labtechno.roazhon.inra.fr

L.V. Saboya, J.-L. Maubois

1. INTRODUCTION

Microfiltration (MF) like all the othermembrane separation processes is a tech-nique which allows the differential concen-tration in the liquid retained by the mem-brane, named MF retentate, of thecomponents with a size bigger than the aver-age pore diameter of the membrane. Theliquid going through the MF membrane isnamed microfiltrate. It contains, at the sameconcentration as those existing in the incom-ing treated liquid, the components with asize smaller than the average pore diameter.

As shown in Figure 1, pore diameter ofMF membranes ranges from 10 µm to0.1 µm, which means MF will be used forthe specific separation of the particles sus-pended in liquids.

Milk contains particles of which the sizedistribution ranges (Tab. I) are relativelywell separated. By decreasing size, they aresomatic cells (15–6 µm), fat globules(15–0.2 µm), bacteria (6–0.2 µm) and caseinmicelles (0.3–0.03 µm) [40].

The breakthrough of MF in the dairyindustry came in the 1980s with the devel-opment of new ceramic membranes com-prising a multichannel geometry and ahighly permeable support [15] which made

industrially feasible the revolutionaryhydraulic concept proposed by Sandblom[47]: the uniform and low transmembranepressure (UTP). The UTP concept over-comes most of the heterogeneous foulingof the MF membrane which results from thehigh cross flow velocities required forobtaining both high permeation fluxes andaccurate selectivity.

542

permet maintenant d’obtenir une séparation différentielle de chaque catégorie de particule laitière. Leprétraitement par microfiltration tangentielle du lait est utilisé pour la production de lait de consom-mation peu chauffé ayant une flaveur similaire à celle du lait cru et une durée de vie 3 à 5 fois supé-rieure à celle des produits classiques. Les résultats observés sur la rétention par les membranes de micro-filtration d’espèces bactériennes pathogènes permet de dire que la sécurité hygiénique des fromagesfabriqués à partir de lait cru microfiltré est au moins équivalente et même supérieure à celle des fro-mages fabriqués à partir de lait pasteurisé. Plusieurs autres applications de la microfiltration tan-gentielle sont en cours de développement, comme l’extraction de la matière grasse résiduelle dulactosérum ou la clarification et la séparation des bactéries des saumures de fromage, mais la plus pro-metteuse est indubitablement la séparation différentielle de la caséine micellaire et des petits globulesgras du lait. Avec les produits obtenus des deux côtés de la membrane de microfiltration, la techno-logie laitière aura la possibilité non seulement d’améliorer le rendement et la qualité de beaucoup deproduits laitiers mais aussi de créer une gamme diversifiée de nouvelles textures.

microfiltration / membrane / lait de consommation / caséine native / saumure / lactosérum /fromage

Figure 1. Illustration of filtration membranes:microfiltration (MF), ultrafiltration (UF), nanofil-tration (NF), reverse osmosis (RO).Figure 1. Illustration de la filtration membra-naire : microfiltration (MF), ultrafiltration (UF),nanofiltration (NF), osmose inverse (OI).

Current developments of microfiltration

was offered for the envisaged applications inthe dairy industry by the commercial emer-gence of the ceramic-based membranes [14]originally developed for isotope enrichmentin the nuclear industry. Nevertheless, if thechemical and heat stability of this new mem-brane generation was corresponding fullyto requirements of the dairy industry, fluxdecline and changes in selectivity versustime were so high [42] that none of the pro-posed processes [16, 42] could be consid-ered as suitable for industrial scale-up. Mostof the encountered problems in controllingthe fouling of MF membranes over time bythe deposited milk components were solvedby the carrying out of the hydraulic conceptof the uniform transmembrane pressure(UTP) proposed as early as 1974 bySandblom [47]. The different membraneand system configurations currently usedfor obtaining hydraulic UTP are describedlater.

Pore sizes of MF ceramic membranesrange from 20 µm to 0.1 µm. Structures andmethods of manufacture of these membranesare described by Rios et al. [44]. In a firstapproach, membranes consist of two parts:a macroporous support and the active mem-brane coated onto the surface. The support-ing materials made from alumina (the mostdeveloped), carbon, stainless steel or SiC(not employed in the food industry becauseof its low stability at high pH) must drainaway the permeate without any hydrody-namic resistance and thus have a minimumpore diameter of 10 µm. The membranelayer (usually composed of alumina, tita-nium oxide or zirconia or a mixture of bothoxides) is formed by coating the supportwith a colloidal suspension of highly dividedpowder (more than 100 m2.g–1), then sin-tered to the support by firing. Membranethickness is around to 3–5 µm [18]. Narrowpore size distribution is obtained throughboth accurate control of the colloidal parti-cle size and sintering of two, or sometimesthree successive layers (Société desCéramiques Techniques – SCT, personalcommunication).

Cross flow microfiltration (CFMF) hasemerged as an industrial separation tech-nology in the dairy industry for at leastthree main applications: (i) removal of bac-teria [25, 50], (ii) whey defatting [39], and(iii) micellar casein enrichment of thecheesemaking [9], but numerous other appli-cations are currently being investigated, suchas selective separation of somatic cells fromraw whole milk, whey or milk protein frac-tionation or milk fat separation.

2. MICROFILTRATIONENGINEERINGCONSIDERATIONS

Efficiency of any membrane separationtechnologies is governed by many param-eters: amongst them the most active are thematerials constituting the membrane and itssupport and the hydrodynamics of the wholeequipment.

2.1. Membranes

In the early 80s, attempts were made touse MF as an alternative technology to cen-trifugation for clarification and microbialepuration of whey [32, 33, 41] but the avail-able organic membrane materials (polysul-fone and polycarbonate) were not satisfac-tory both in terms of flux and selectivity aswell as in terms of mechanical heat andchemical stability [20]. A novel opportunity

543

Table I. Relative diameters of milk particles(adapted from [40]).Tableau I. Diamètre moyen des particules dulait (adopté de [40]).

Component Size (µm)

Casein micelles 0.300–0.032Bacteria 15.000–0.200Fat globules 6.000–0.200Somatic cells 15.000–6.000

L.V. Saboya, J.-L. Maubois

Such a restriction of the width of theGaussian distribution curve of the porediameter can improve significantly the selec-tivity of the MF process. For example, sub-stitution of a 1.4 µm pore size Membraloxby a 1.4 µm Sterilox i.e. with the same aver-age pore diameter but with a much narrowerpore size distribution increases the MF reten-tion of milk bacteria by 2 log [26]. As saidbefore, ceramic membranes are the onlyones that satisfy all the requirements of theapplications in the dairy industry i.e. a strongmechanical resistance which allows the useof high recirculation velocities of viscousMF retentates, a wide tolerance to pH (0.5 to13.5) allowing its usage for cleaning in placeof caustic soda (up to 3%), of nitric acid (upto 2%) – but phosphoric and hydrofluoricacids should be avoided – and of sodiumhypochlorite for sanitation. Membranes wereconfigured originally in monotubes with aninner diameter ranging from 3 to 8 mm anda length around 85 cm. Recent designs con-cern mostly the multichannel configurationwith either classical cylindrical channels orvarious cross section forms: square, trifoliumor “daisy” like those proposed by the Tami®

Co. Monotubular and multichannel mono-liths, whose the length is now more than 1 mare assembled in bundles placed in stain-less steel housings named MF modules. Thearea developed in one module ranges from0.2 m2 to 10 m2 or more according to themanufacturer. Most of the proposed mod-ules are steam sterilizable but in practice,the operating temperatures are limited notby the membranes, which can withstandtemperatures of several hundred °C, but bythe other components of the MF modulesuch as gaskets.

2.2. MF equipment configurations

Except for some old equipment for wheydefatting, the first industrial application ofMF in the dairy industry [8, 29], most ofthe current MF plants operate according tothe hydraulic concept of UTP. To overcome

the rapid fouling of the MF membraneresulting from the necessary high cross-flowvelocity conditions of the MF retentate(more than 7 m.s–1), Sandblom [47] has sug-gested recirculating the MF permeate, insuch flow conditions that the transmem-brane pressure is low and uniform all alongthe membrane (Fig. 2). In the methodpatented by Alfa Laval [19], the permeatecompartment filled with plastic balls is pres-surized by pumping the permeate in a loopparallel to the direction of the retentate flow.A basic MF system comprises a feed tank, afeed pump and two recirculation loops: onefor the MF retentate including one or a mul-tiple number of two modules and a heatexchanger – one for the MF permeate. Bothloops have outlet valves regulated by sensorsfor a continuous production of the two sep-arated fluids.

Recent developments in ceramic mem-brane technology have allowed the sup-pression of the MF permeate recirculationloop, and thus a decrease in the investmentcost per m2 of installed equipment and asaving of some consumed energy. The firstone, known as Membralox GP®, was devel-oped by SCT [10]. The requested counter-pressure in the permeate side is obtained bya continuous variation of the porosity of themembrane ceramic support. The second onevery recently proposed by Tami Co, named“Isoflux®”, claims obtaining UTP by a con-tinuous gradient of MF membrane thickness.Both developments must be used for well-defined applications, the porosity gradient ofthe membrane support as well as the thick-ness gradient of the membrane being bothbuilt for precise flux and selectivity.

2.3. Starting procedure

MF equipment must be started with carein order to avoid dramatic and fast fouling.They are firstly filled with warm water(52 °C) with the air exhaust valve open forremoving air bubbles. Then, hydraulic recir-culation parameters which will be used on

544

Current developments of microfiltration

colloidal particles or microorganisms sus-ceptible to cause an inside blocking of themembrane pores or an undesirable bacte-rial contamination of the cleaned equipment.Consequently, dwell and tap water must benot only softened but also microfiltrated ona 0.2 µm dead end filter before being usedfor cleaning.

The cleaning cycle used includes:

– rinse with water heated at 50 °C;

– alkaline cleaning with a ternary deter-gent such as Ultrasil® 25 (Henkel) at a con-centration of 0.5–1.0% or similar productat 65–70 °C for 20 min in the same operat-ing conditions as the ones used for the appli-cations;

– rinse with water at 50 °C;

– facultatively sanitation with ClONa200 ppm at 20 °C for 15 min;

milk are adjusted on the recirculated water.Finally, milk of which the physico-chemicalequilibrium has been reestablished i.e. by apreheating of 50 °C for 20 min, is progres-sively introduced in the MF retentate loop.The hydraulic parameters are readjusted dur-ing the first ten minutes before running instationary conditions.

2.4. Cleaning and sanitationof the MF equipment

Cleaning of MF equipment is performedin the same way as for all membrane tech-nologies with differences coming from thenature of the membrane material and from ademand for a higher degree of water quality.Indeed, water utilized for the rinsing steps aswell as for the preparation of the differentsolutions must not contain suspended

545

Figure 2.Pressure profiles during cross flow microfiltration.Figure 2. Profil des pressions en microfiltration tangentielle.

L.V. Saboya, J.-L. Maubois

– acid cleaning NO3H 0.5% for 20 min at50 °C in the same operating conditions asthose used for the alkaline step;

– rinse with water at 50 °C.

Efficiency of the cleaning is checked bythe water flux in defined conditions and bythe bacteriological quality of the last rinse(no CFU in 100 mL).

3. APPLICATIONS OF MFIN THE DAIRY INDUSTRY

The major application of MF in the dairyindustry concerns the removal of microor-ganisms from skim milk in order to producea debacterized raw material which is after-wards transformed into fluid milks, cheesesor long storage dairy products such as pow-der or protein derivatives. Another emergingapplication is the selective separation ofmicellar casein, which opens new avenuesfor better added values of milk components.Moreover these two main industrial devel-opments, numerous other applications havebeen proposed which will be brieflyreviewed here.

3.1. Bacterial removal

3.1.1. Current developmentfor the fluid milk industry

Milk collected by the dairy plant alwayscontains a microbial flora formed by numer-ous species illustrating the contaminations ofmilk by the udder, the milking machine, thelocal farm atmosphere, the bulk tank andthe transportation equipment. Taking intoconsideration all these potential contami-nations, one can say that whatever are orwhatever will be the hygienic practices atfarm level, the collected milk will be alwayscontaminated. Cow husbandry cannot bedone in aseptic conditions and milk willalways have the risk of containingpathogenic bacteria for human beings suchas Listeria, Brucella, MycobacteriumorSalmonella.

Destruction of the potentially risky florahas been done for fifty years by applying tomilk heat treatments such as HTST pas-teurization or UHT. But if heating effec-tively kills most of the microbial flora, thedead cells remain in the milk with theirpotentially active enzymes that with themetabolic activity developed by the growthof the remaining thermoduric bacteria, willcause alterations of liquid milks during stor-age, thus reducing shelf-life.

Membrane MF offers an interesting alter-native to heat treatments. Initially proposedby Holm et al. [19] and Piot et al. [42], ithas led to the technology and the equipmentcalled “Bactocatch®” by the Tetra Laval Co.Numerous studies done in Sweden [25, 31]and in France [24, 26, 50, 51] have allowedimprovements of the original parametersdescribed in the patent of Holm et al. [19].Nowadays, skim milk heated to 50 °C is cir-culated at a velocity of 7.2 m.s–1 along aSterilox® membrane with an average poresize of 1.4 µm. The value of the uniformtransmembrane pressure (UTP) is around0.5 bar. The volumetric reduction factor(VRF) is 20 and in some huge equipment,the 5% flow represented by the MF retentateis concentrated 10 times more in cascadeby a second MF apparatus, thus leading to aVRF of 200. Permeation rates of proteinsand total solids are respectively higher than99.0% and 99.5%. Fluxes obtained indus-trially are in the order of 500 L.h –1.m–2 dur-ing 10 h.

Average observed decimal reduction(DR) is above 3.5 which means for an enter-ing 20 000 CFU.mL–1 skim milk, the micro-filtrated product contains less than10 CFU.mL–1. Spore forming bacteria whichrepresent the main surviving species to pas-teurization are better retained by MF mem-brane because of their large cellular volumeas shown by Trouvé et al. [50]. The observedDR are higher than 4.5.

Synthesis of the studies done by Madecet al. [24], Pasteur Institute and INRA (datanot published) shows for Listeria monocy-togenes, Brucella abortus, Salmonella

546

Current developments of microfiltration

of casein micelles permeation rate but thebacterial epuration of the skim milk isso high (remaining count is less than1 CFU⋅L–1) that to obtain a fluid milk witha long shelf life at room temperature, a mod-erate heat treatment (less than 100 °C, 2 s) isnecessary. This new fluid milk is under com-mercial development and the first results(Fauquant, Madec and Maubois, data non-published) indicate a room temperature shelflife ranging from 6 to 9 months. Sensorytests show a flavor close to the one of HTSTclassical milks and much better than the cur-rent UHT milks.

As well as the use for the fluid milk sec-tor, MF pretreatment of the skim milk can beexpanded to all skim milk used for the pro-duction of milk derivatives such as low-heatmilk powder as proposed by Schuck et al.[48], milk proteins concentrates or micel-lar casein powder [49]. On the other hand,MF milks, because of the removal of deadcells and somatic cells, offer to the dairyscientist the chance to study without anybias the exact effect of endogenous enzymeson milk components.

3.1.2. Current developmentsfor the cheese industry

Cheese consumers wish to have the pos-sibility to buy and to eat both cheeses withtypical flavor and texture, characteristic of ageographical area i.e. cheeses belonging totheir cultural patrimony, as well as moreclassical cheeses for daily consumption. Aswell as this diversified offer, the consumersalso want a constant high quality and nohygienic risk. Such high requirements callfor a complete control of the cheesemakingprocess starting from the raw material col-lected by the cheese plant.

Use of milks pretreated by MF with thesame technology as described before for thefluid milk production i.e. a mixture of MFskim milk and heat-treated cream but some-times with the variant of a lower MF tem-perature (35–37 °C) in order to respectthe raw milk labeling, offer to the cheese

typhimurium and Mycobacterium tubercu-losisrespective DR of 3.4, 4.0, 3.5 and 3.7.Such results will assure that MF skim milkwill contain less than 1 CFU⋅L–1 of thesepathogenic bacteria if we take into accountusually described contaminations at farmlevel.

Somatic cells are evidently totallyretained by the MF membrane and conse-quently MF milk will not be degraded bytheir thermoduric enzymes [22].

Several varieties of MF liquid milks arepresently offered to the consumers. InFrance, the Marguerite® milk is consideredto be a raw milk because no pasteurizationis applied. The MF skim milk is mixed withthe amount of heated cream (95 °C, 20 s)requested for fat standardization; the mixtureis slightly homogenized and asepticallyfilled. The claimed shelf life at 4–6 °C is15 days. In Scandinavia, Canada [7] and theUK [45], before aseptic filling a HTST(72 °C, 15 s) heat treatment is applied, lead-ing to a claimed shelf life of 35 days. TheseMF milks encounter great commercial suc-cess because of their flavor (no cooked taste)and their improved storage ability [7].

APV has patented a variant of the Bac-tocatch process which proposed the mixingof highly heat-treated MF retentate withcream and MF milk. Such a mixing is neverdone nowadays by the Bactocatch users. Inthe APV variant, the MF retentate, not heat-treated, is returned to the raw milk enteringthe cream separator. Consequently, most ofthe bacteria as well as somatic cells are dis-charged through the separator sludge [34].

Use of smaller pore diameter (0.5 µmeffective) recently proposed by Lindquist[23] has shown a bacteria removal increasedby 2 or even 3 log cycles compared to the1.4 µm pore diameter. Membranes utilizedin this patent were either Membralox® orPall® ones set respectively in a classicalcross flow equipment and in a DynamicMembrane Filtration equipment. Decrease ofpore size evidently leads to a slight decrease

547

L.V. Saboya, J.-L. Maubois

technologist a complete control of the rawmaterial. Consequently, it can be claimedthat cheeses made from MF milk are at leastas safe from a hygienic point of view thancheeses made from pasteurized cheese milk.Moreover, because MF membrane removes,at a very high level, spore-forming bacteriasuch as Clostridium tyrobutyricum[31, 50],addition of nitrate at a level of 15 g per100 kg as it is done in a few countries suchas The Netherlands to prevent late blowingin semi hard or hard cheeses, could be sup-pressed with the positive consequences forenvironment and consumer health.

On the other hand, use of MF milk opensfor the cheese scientist new ways to deter-mine and to characterize the exact roleplayed in cheese ripening by each micro-bial species or strain composing the ripeningecosystem, i.e., lactic acid bacteria, non-starter lactic acid bacteria, propionibacte-ria, surface yeasts and molds [30], withoutthe bias always introduced in the publishedstudies by the use of heat sterilized milks,i.e., milks with the enzymatic potential ofdead cells.

But, as described by Maubois et al. [30]use of MF milk in cheesemaking brings upnumerous scientific and legal questions, par-ticularly for AOP cheeses:

– Are they still raw milk cheeses? Withcurrent technology, removal of bacteria isdone by MF at 35 °C on about 90% of thecheese milk. The cream part is highly heat-treated (generally 90–95 °C, 2 or 3 min).What are the effects of this heating on fat, onthe globule membranes? What are the con-sequences on the casein network structureduring formation and ripening by compari-son with the use of a totally non-heatedcheesemilk?

– Many studies demonstrate that naturalflora plays a major role in the developmentof flavor and texture characteristics of AOPcheeses. Separation and concentration ofthis natural flora by MF is easy and as verywell shown by Demarigny [6], a typical floraisolated from milks produced in a defined

area A can be added to a MF milk obtainedin another area B, even very far from A.That will originate a cheese in area B whichhas the same typicity as cheeses produced inarea A. One sees easily the legal problemswhich arise from this possibility and theextent of delocalizations which could bethus created.

3.2. Selective separationof micellar casein

When whole or skim milk (evidentlyBactocatch® treated skim milk) is circulatedalong a MF membrane with a pore sizediameter of 0.1–0.2 µm, a microfiltrate witha composition close to that of sweet whey isobtained [9]. This microfiltrate is crystalclear and it can be sterile and claimed with-out any viral particle [12] if the downstreamequipment prevents recontamination. TheMF retentate is an enriched milk solutionof native micellar casein. Excepting the poresize of the MF membrane, engineering char-acteristics of the technology are the sameas those used for bacteria removal of milki.e. use of a device leading to a uniformtransmembrane pressure, a retentate veloc-ity of at least 7 m.s–1 and a running tem-perature of 50–55 °C. As proposed by Pierreet al. [38] and by Schuck et al. [48], puri-fied micellar casein can be prepared accord-ing to this application of MF by followingthe process described in Figure 3 with threemain steps:

(a) concentration of the MF retentate upto a VRF of 3–4;

(b) diafiltration of this MF retentate with4 diavolumes of R.O. treated water;

(c) concentration of the diafiltered MFretentate up to a VRF of 6–7;

(d) spray-drying of the 6:1 MF retentate.

Many variants of this process can beenvisaged such as use of UF instead of MFat step (c), diafiltration at lower or higherVRF or uses of saline or acid solutions fordiafiltration instead of pure water.

548

Current developments of microfiltration 549

Figure 3.Micellar caseinmaking process.Figure 3. Procotole defabrication de la caséinemicellaire.

L.V. Saboya, J.-L. Maubois

Micellar casein enriched milk improvescheesemaking process. Survey of the stud-ies of Caron et al. [2], Saint-Gelais et al.[46] and Daviau et al. [4] demonstrates thatsuch a MF milk coagulates slightly fasterand leads to a more rigid curd (increase ofthe Young modulus with a coefficient of30.5 × 103 Pa per g of casein per kg of milk)and firmer curd (increase of fracture stresswith a coefficient of 2.55 × 103 Pa per g ofcasein per kg of milk). The yield increasein Camembert cheesemaking was 5.6 g.kg–1

per g of casein increase in cheese milk [4].

Micellar casein powder which, to ourknowledge, is not yet produced at an indus-trial scale, has large potentialities to be usedfor fortifying cheese milk, for replacingcheese curd in process cheese formulations.It also constitutes an excellent raw materialfor preparing individual caseins (β-CNnotably) or κ-glycomacropeptide [28] in astrategy to produce purified nutraceuticalderivatives from milk proteins [27].

On the other hand, partial removal byMF of whey proteins significantly reducesthe detrimental effects of heating on rennetcoagulability and cheese abilities of milk.To decrease the κ-casein/β-lactoglobulinratio will evidently lower the extent of for-mation of their complex during the heattreatment and consequently rennetedmicelles will aggregate almost as well as inraw milk [43]. This result was used todevelop a new medium heat powder withsuch a cheesemaking ability that it can beused after reconstitution for the productionwith higher yields and good flavor charac-teristics of any cheese variety [11].

3.3. Selective fractionationof globular milk fat

Separation of milk fat in small globulesand in large globules was recently proposedby Goudédranche et al. [17] through the useof special ceramic MF membranes whichcause very low damage of the native glob-ular membrane. It was claimed by the

authors of this patented process that milkscontaining as fat only globules with a diam-eter lower than 2 µm led to smoother andfiner textural characteristic dairy products:cream, liquid milk, cheese etc. than thosemade from reference milks or from milkscontaining large fat globules with the excep-tion of butter. The observed results wereascribed to the ability of fat globule mem-brane components to bind water and to thedifference in triglycerides composition.

3.4. Fat removal of whey

The need to remove fat in whey proteinconcentrates was highlighted by de Wit [5]through his systematic approach to evalua-tion of functional properties notably the suit-ability of these dairy products as whippingingredients. Numerous studies have beendone to optimize the total removal of wheyresidual fat by MF after an aggregation pre-treatment developed according to the resultsof Pearce [36], Maubois et al. [29], Baumyet al. [1], Gésan [13], Daufin et al. [3], andPierre et al. [39]. The optimized procedurewas described by Maubois and Ollivier [28].It includes: (a) preconcentration by UF ofthe whey to 25% of its initial volume at50 °C; (b) removal of microorganisms byMF with a 0.8 µm MF membrane set in anUTP equipment; (c) adjustment of temper-ature to 55 °C and pH to 7.5, this pH valuebeing constantly maintained and (d) sepa-ration of the aggregated calcium phospho-lipid complex by MF with a 0.1 µm poresize membrane set up in a UTP equipment.The so-obtained defatted whey concentratedby UF to a protein content of 10% showsfoaming properties identical to egg white.

3.5. MF and microbial fermentations

As reviewed by Kulozik [21], membranetechnologies and especially MF can be prof-itably incorporated in the upstream anddownstream processing of the fermentationbroth either in discontinuous operations for

550

Current developments of microfiltration

carcinogenic [35] and anyway, disposal ofKieselguhr will be totally forbidden by lawin European countries in 2002 for environ-mental reasons.

MF in a UTP equipment with a pore sizeof 1.4 µm or 0.8 µm membrane completelyrejects yeasts and moulds, retains 99.9% ofthe contamination bacteria and only 6.7%of the calcium salts and 2–3% of thenitrogen matter [37]. The flux at 20 °C was600 L.h–1.m–2 during 8 h and the concen-tration factor higher than 1:100.

From these very satisfactory results, it islikely that MF of cheese brine will expandrapidly if the investment cost of UTPceramic membranes decreases as it can beexpected.

4. CONCLUSION

With the use of membrane microfiltra-tion technology, the world dairy industryhas today a new powerful and flexible toolfor a dramatic improvement of the hygienicsafety of all dairy products while applying aminimal heat treatment to the milk. Conse-quently, it can be expected that nutritionaland bioactive properties of milk componentswill be kept intact or at least little damagedin products derived from MF milk: liquidmilks, cheeses, powders, fermented prod-ucts, etc. Removal of bacterial and somaticcells with their endogenous thermoduricenzymes offers moreover to dairy researchthe chance to determine without any biasthe precise activity versus storage time andtemperature conditions of endogenous milkenzymes. The same research area is alsoopened for studying the metabolic conse-quences of each component of the micro-bial ecosystem added to MF cheesemilks.

Specific separations of micellar caseinand of small and large fat globules open todairy technology new and diversifiedavenues not only for improving current man-ufacturing processes, and consequently over-all quality of resulting products, but also for

separating the biomass from the producedmetabolites or in continuous fermentationprocesses. For example, MF equipment witha 0.1 µm pore size membrane are connectedto the fermentor to introduce aseptically ster-ile thermolabile components such as vita-mins or salts. MF with 1.4 µm ceramic mem-branes can achieve cell harvesting of starterbacteria with final concentration of thebiomass increased to 1012–1013 CFU.mL–1

(Rolland and Proudy, data not published).

As well as the applications of MF inbiofermentations of milk components, thisseparation technology has a great potentialalone or in combination with nanofiltration(NF) for the treatment of dairy wastesthrough the carrying-out of membrane biore-actors. Indeed, activated sludges confinedin a MF loop can easily degrade biologicalcontent of dairy effluents and thus restitutepurified water to the environment.

3.6. Cheese brine purification

Efficient sanitation of cheese brine isrequired in the dairy industry in order toprevent post contamination of cheeses dur-ing salting. It is effectively well demon-strated that brine may contain unwantedmicroorganisms such as gas-producing lac-tobacilli, pathogenic bacteria (Staphylococci,Listeria and others) yeast and molds [35,37]. As well as its bacterial load, the overallquality of brine and consequently the qual-ity of the cheese salting, results from a rel-atively delicate balance equilibrium betweenNaCl content (18 to 26%), soluble and pre-cipitated Ca salts, lactose and lactic acidsoluble and denaturated whey proteins.

Heat treatment and Kieselguhr filtrationare still the most well-used technologies forbrine treatment but both present many dis-advantages: heat treatment requires expen-sive plate heat exchangers to avoid corro-sion of usual stainless steel and modifies toan undesired extent the calcium phosphateequilibrium. Kieselguhr filtration needsmanual handling, its dust is suspected to be

551

L.V. Saboya, J.-L. Maubois

creating new milk derivatives answering tothe constant needs of novelty of the worlddairy consumers.

ACKNOWLEDGEMENTS

Funding support of Ms Luciana Saboya’sgrant by CAPES and FAPESP is greatlyacknowledged.

REFERENCES

[1] Baumy J.J., Gestin L., Fauquant J., Boyaval E.,Maubois J.L., Technologies de purification desphospholipides du lactosérum, Process 1047(1990) 29–33.

[2] Caron A., Saint Gelais D., Pouliot Y., Coagu-lation of milk enriched with ultrafiltered ordiafiltered microfiltered milk retentate powders,Int. Dairy J. 9 (1997) 445–451.

[3] Daufin G., Labbé J.P., Quemerais A., MichelF., Merin U., Optimizing clarified whey ultra-filtration: influence of pH, J. Dairy Res. 61(1994) 355–363.

[4] Daviau C., Effet des caractéristiques physico-chimiques du lait sur l’aptitude à la coagulationprésure et à l’égouttage dans une technologiepâte molle, Ph.D. thesis, INRA, Rennes, France,2000.

[5] de Wit J.N., Functional properties of whey pro-teins in food systems, Neth. Milk Dairy J. 38(1984) 71–89.

[6] Demarigny Y., Rôle de la flore naturelle du laitcru et de paramètres technologiques (teneur ensel et température d’affinage) sur l’évolutiondes caractéristiques biochimiques, microbiolo-giques et sensorielle des fromages à pâte cuite,Ph.D. thesis, INRA, Poligny, France, 1997.

[7] Eino M.F., Lessons learned in commercializationof microfiltered milk, Bull. Int. Dairy Fed. 320(1997) 32–36.

[8] Fauquant J., Vieco E., Brulé G., Maubois J.-L.,Clarification des lactosérums doux par agréga-tion thermocalcique de la matière grasse résidu-elle, Lait 65 (1985) 1–20.

[9] Fauquant J., Maubois J.-L., Pierre A., Microfil-tration du lait sur membrane minérale, Tech.Lait. 1028 (1988) 21–23.

[10] Garcera D., Toujas E., (Société des CéramiquesTechniques) Support macroporeux à gradientde perméabilité et son procédé de fabrication,Demande de Brevet No. 97.04.359, 1998.

[11] Garem A., Schuck P., Maubois J.-L., Cheese-making properties of a new dairy-based powermade by a combination of microfiltration andultrafiltration, Lait 80 (2000) 25–32.

[12] Gautier M., Rouault A., Méjean S., FauquantJ., Maubois J.-L., Partition of Lactococcus lac-tis bacteriophage during the concentration ofmicellar casein by tangential 0.1 µm pore sizemicrofiltration, Lait 74 (1994) 419–423.

[13] Gésan G., Daufin G., Merin U., Labbé J.P.,Quémerais A., Fouling during constant fluxcross-flow microfiltration of pretreated whey.Influence of transmembrane pressure gradient,J. Membr. Sci. 80 (1993) 131–146.

[14] Gillot J., Garcera D., New ceramic filter mediumfor cross-flow microfiltration, in: FILTRACongress, Paris, France, 1984, pp. 2–4.

[15] Gillot J., Garcera D., Ceraver: nouveaux médiasfiltrants céramiques, Tech. Lait. 1007 (1986) 23.

[16] Glimenius A.R., Jansson G.T., Kemi K.W.,Sandblom R.M., Filtering method for separatingskim milk from milk products, US Pat. No.4.140.806, 1979.

[17] Goudédranche H., Fauquant J., Maubois J.-L.,Fractionation of globular milk fat by membranemicrofiltration, Lait 80 (2000) 93–98.

[18] Grandison A.S., Finnigan T.J.A., Microfiltra-tion, in: Grandison A.S., Lexis M.J. (Ed.), Sep-aration process in the food and biotechnologyindustries, Woodhead Publ. Ltd., Cambridge,England, 1996, pp. 141–153.

[19] Holm S., Malmberg R., Svensson K., Methodand plant for producing milk with a low bacte-rial content, Eur. Pat. No. 0.194.286, 1984.

[20] Jost R., Jelen P., Cross-flow microfiltration –an extension of membrane processing of milkand whey, Bull. Int. Dairy Fed. 320 (1997) 9–15.

[21] Kulozik U., Membranes in microbial fermen-tations, Bull. Int. Dairy Fed. Special issue 9201(1992) 141–160.

[22] Law B.A., Goodenough P.W., Enzymes in milkand cheese production, in: Tucker G.A., WoodsL.F.J. (Eds.), Enzymes in Food Processing,Blackie Acad. & Professional, Glasgow, UK,1995, pp. 114–143.

[23] Lindquist A., A method for the production ofsterile skimmed milk, PCT Pat. WO No. 57549,1998.

[24] Madec M.N., Méjean S., Maubois J.-L., Reten-tion of Listeria and Salmonella cells contami-nating skim milk by tangential membrane micro-filtration (Bactocach process), Lait 72 (1992)327–332.

[25] Malmberg M., Holm S., Low bacteria skim milkby microfiltration, North Eur. Food Dairy J. 54(1998) 75–78.

[26] Maubois J.-L., Current uses and future per-spectives of MF technology in the dairy indus-try, Bull. Int. Dairy Fed. 320 (1997) 37–40.

[27] Maubois J.-L., Léonil J., Peptides du lait à acti-vité biologique, Lait 69 (1989) 245–269.

[28] Maubois J.-L., Ollivier G., Extraction of milkproteins, in: Damodaran S., Paraf A. (Eds.),Food proteins and their applications, MarcelDekker, Inc., New York, 1997, pp. 579–595.

552

Current developments of microfiltration

[41] Piot M., Maubois J.-L., Schaegis P., Veyre R.,Microfiltration en flux tangentiel des lactosé-rums de fromagerie, Lait 64 (1984) 102–120.

[42] Piot M., Vachot J.C., Veaux M., Maubois J.-L.,Brinkman G.E., Écrémage et épuration bacté-rienne du lait entier cru par microfiltration surmembrane en flux tangentiel, Tech. Lait. Market.1016 (1987) 42–46.

[43] Quiblier J.P., Ferron-Baumy C., Garric G.,Maubois J.-L., Procédé de traitement des laitspermettant au moins de conserver leur aptitudefromagère, Brevet Fr. No. 0.542.583 B1, 1992.

[44] Rios G.M., Tarodo de la Fuente B., BennasarM., Guidard C., Cross-flow filtration of biolog-ical fluids on inorganic membranes: A first stateof the art, in: Thorne S. (Ed.), Developments inFood Preservation, Elsevier Applied Science,London, UK, 1989, pp. 131–175.

[45] Russell P., A new fresh milk comes to Britain,Eur. Dairy Mag. 3 (1999) 25–27.

[46] Saint-Gelais D., Roy D., Audet P., Manufac-ture and composition of low fat Cheddar cheesefrom milk enriched with different protein con-centrate powders, Food Res. Int. 31 (1998)137–145.

[47] Sandblom R.M., Filtering process, SW Pat. No.74.16.257, 1974.

[48] Schuck P., Piot M., Méjean S., Le Graet Y.,Fauquant J., Brulé G., Maubois J.-L., Déshy-dratation des laits enrichis en caséine micellairepar microfiltration, comparaison des proprietésdes poudres obtenues avec celles d’une poudrede lait ultra-propre, Lait 74 (1994) 47–63.

[49] Schuck P., Piot M., Méjean S., Le Graet Y.,Fauquant J., Brulé G., Maubois J.-L., Déshy-dratation par atomisation de phosphocaséinatenatif obtenu par microfiltration sur membrane,Lait 74 (1994) 375–388.

[50] Trouvé E., Maubois J.-L., Piot M., Madec M.N.,Fauquant J., Rouault A., Tabard J., BrinkmanG., Rétention de différentes espèces micro-biennes lors de l’épuration du lait par microfil-tration en flux tangentiel, Lait 71 (1991) 1–13.

[51] Vincens D., Tabard J., L’élimination des germesbactériens sur membranes de microfiltration,Tech. Lait. 1033 (1988) 62–64.

[29] Maubois J.-L., Pierre A., Fauquant J., Piot M.,Industrial fractionation of main whey proteins,Bull. Int. Dairy Fed. 212 (1987) 154–159.

[30] Maubois J.-L., Caudron B., Daviau C., MadecM.N., Pierre A., Membrane technologies: toolsfor a total control of the cheesemaking process,IDF Banff 2000 Seminar.

[31] Meershon M., Nitrate free cheesemaking withBactocatch, North Eur. Food Dairy J. 55 (1989)108–113.

[32] Merin U., Bacteriological aspects of microfil-tration of cheese whey, J. Dairy Sci. 69 (1986)326–328.

[33] Merin U., Gordin S., Tanny G.B., Microfiltrationof sweet cheese whey, J. Dairy Res. 18 (1983)153–160.

[34] Osterland N., Removal of bacteria and sporesfrom milk by microfiltration, in: Proc. INRA –Int. Dairy Fed. Seminar, New applications ofmembrane technology in the dairy industry,Saint Malo, France, 1999, p. 18.

[35] Ottosen N., Konigsfeld P., Microfiltration ofcheese brine, Eur. Dairy Mag. 4 (1999) 22–24.

[36] Pearce R.J., Thermal separation of β-lactoglob-ulin and α-lactalbumin in bovine Cheddar cheesewhey, Aust. J. Dairy Technol. 42 (1983) 75–78.

[37] Pedersen P.J., Microfiltration for the reductionof bacteria in milk and brine, Bull. Int. DairyFed. Special issue 9201 (1992) 33–50.

[38] Pierre A., Fauquant J., Le Graet Y., Piot M.,Maubois J.-L., Préparation de phosphocaséinatenatif par microfiltration sur membrane, Lait 72(1992) 461–474.

[39] Pierre A., Le Graet Y., Daufin G., Michel F.,Gésan G., Whey microfiltration performance:influence of protein concentration by ultrafil-tration and of physicochemical pretreatment,Lait 74 (1994) 65–77.

[40] Pierre A., Goudédranche H., Garem A., DaufinG., Industrie Laitière, in: Daufin D., René F.,Aimar P. (Coords.), Les séparations par mem-brane dans les procédés de l’industrie alimen-taire, Tec. Doc. Lavoisier, Paris, 1998,pp. 282–371.

553

To access this journal online:www.edpsciences.org

L.V. Saboya, J.-L. Maubois554