CH688909A5 - Production of sterile milk through dynamic microfiltration. - Google Patents

Description

       

  
 



  The invention relates to a method for the treatment of raw milk for the production of treated milk with a lower bacterial content than the raw milk, as is also the subject of US patent application serial no. 07/901 238 with the filing date June 19, 1992.



  The invention thus relates to a method for producing milk, whole milk or skimmed milk, the process product and a method for distributing the milk to customers.



  The known pasteurization process has been used for decades to kill bacteria in milk. Unfortunately, the high temperatures required for pasteurization negatively affect the taste of the milk. In addition, the pasteurization process does not eliminate all undesirable bacteria even at these high temperatures and leads to the short shelf life of most milk products.



  Bacillus cereus is usually the predominant bacterium in conventionally sterilized milk of older age because it survives the pasteurization process and thrives at low temperatures, which promotes the deterioration of the milk. In general, there is a need for a method that can reduce the bacterial content in both whole and skimmed milk to increase the shelf life of the product and improve its taste by avoiding the pasteurization process.



  It is also of great economic importance to eliminate the very costly and labor-intensive distribution process that is currently required to transport the milk to the customer. currently After the raw milk has been sterilized by homogenization and other steps, each dairy must fill the milk in containers for distribution to the customers and transport this milk chilled. For every dairy, this means that they have to purchase and maintain a large park of refrigerated vehicles in order to transport the milk to the customer's place. By providing a sterile or almost sterile milk product, it would be possible to avoid transporting the milk under cooling conditions. Unfortunately, the pasteurization process only delivers milk with a reduced bacterial content, but not a sterile product.



  In addition, if a sterile milk product could be manufactured, it would no longer be necessary to store the milk in a refrigerated location. The avoidance of large cooling compartments, e.g. to be found in the typical grocery stores would also be of great economic advantage.



  Even with today's pasteurization process, it is very important in some cases to reduce the bacterial content of the milk before pasteurization, especially if a certain amount of raw milk is contaminated so that mere pasteurization would not lead to a shelf life that is reasonable according to today's standards .



  For some applications it is also valuable to be able to deliver treated milk in which the bacterial content is largely reduced, e.g. to a hundredth of the initial value. It is particularly important for the production of cheese to supply milk with a relatively low bacterial content, since incorrect bacterial cultures can destroy the cheese. Simple heat treatment of the milk to a sufficient degree is normally not suitable for cheese production because such heat treatment leads to a lower yield of cheese and can negatively affect the coagulation time.



  Conventionally, additives are used to reduce this problem. Sometimes, however, it is desirable to avoid using these additives.



  Various methods are known for producing milk with a reduced bacterial count using filtrations, but none of them have been widely accepted. The prior art processes generally either have low flow rates, making the process largely uneconomical, or damage the quality of the milk, making the product unacceptable to the customer.



  Conventional filters have tried to produce milk with a reduced bacterial content. Swedish patent application no. 380 422 discloses a process in which whole milk is divided into filtrate and concentrate fractions by microfiltration. The filtrate that passes through the pores of the filter (the pore size of the filter can be between 0.1 micron - 10 micron) consists of milk with a significantly reduced fat content. The concentrate, i.e. the fraction that was retained by the surface of the filter consists of cream, since not only bacteria, but also fat balls are retained to a high degree by the filter.



  Swedish Patent Application No. SE-A 6 715 081 discloses a method for sterilizing milk, in which the fat is first separated from the skimmed milk. The fat fraction is then heat sterilized. The skimmed milk fraction is sterilized by removing the bacteria by filtration (the pore size of the filter is not specified). Finally, the fat and skimmed milk fractions are mixed again to obtain a sterile milk product. In order to sterilize the skimmed milk fractions by filtering the bacteria, the pore size of the filter must be so small that the filter does not let bacteria through. This in turn results in low transmission rates and the undesirable retention of fat globules and proteins from milk.



  U.S. Patent No. 5,064,674 relates to a method of producing hypoallergenic milk by ultrafiltration using membranes that are permeable to molecules with a molecular weight of 5 kDa and less. The components excluded by the membrane include milk proteins, viable or non-viable bacteria, bacterial protein antigens and milk fat. The filtrate collected in the ultrafiltration process is therefore not only free from bacteria and bacterial protein antigens, but also from fat and milk proteins, which makes the product unusable as milk per se.



   Obviously, the bacterial filters used in the prior art to sterilize milk not only remove bacteria, but also fat globules and at least some proteins. Such a filter is quickly clogged by the retained material so that the flow rate decreases rapidly and the filter often has to be cleaned or replaced. The high cost of such an inefficient process is generally prohibitive. In addition, the quality of the milk is impaired because the filter also retains fat globules and proteins.



  From the discussion so far it is evident that there is still a need for an improved milk filtration process which provides a sterile or rather almost sterile product with improved durability and undiminished quality.



  So far, some attempts to treat milk have been made using cross-flow or tangential flow filtration devices. Such devices are known.



  Various types of filtration devices have been described which enable such tangential or cross-flow filtration. Perhaps the oldest known device of this type is described in Soviet patent application no. 142,626 by Zhevnovatyi, A.I. from 1961. This device is formed by a tube made of porous material which is fastened in a second tube. The suspension to be filtered passes under pressure at high speed through the annular space between the two tubes, the filtrate flowing inside the porous tube. Improved devices of a similar design use two concentric cylinders, the inner of the two cylinders being formed by a microporous membrane and the liquid, which is subjected to a forced helical flow, flowing around this inner cylinder.



  Other cross-flow devices include a series of superimposed filter elements in the form of plates or disks with microporous membranes, e.g. are arranged around a filtrate collecting tube, the suspension to be filtered passing in succession in a helical path between the disks.



  Many other variations of the cross flow filtration system have been developed. For example, the U.S. Patent No. 5 009 781 on a cross-flow filtration device with a filtrate network that includes a plurality of longitudinal filtrate chambers and one or more filtrate channels that cut these chambers.



  The U.S. Patent No. 5 035 799 relates to a cross-flow filtration device with filter blades arranged in parallel in the filter container. The input is done under pressure to create a turbulent cross flow in the medium.



  The U.S. Patent No. 5,015,397 relates to a cross-flow filtration device and method which includes a tube of spirally wound profile wire. The contaminated material is placed at one end of the pipe. As it flows through the tube, the contaminants are concentrated in it, while cleaned liquid penetrates the wall of the tube. The U.S. Patent No. 5 047 154 relates to a method and an apparatus for increasing the flow rate of cross-flow filtration systems. The U.S. Patent No. 4,569,759 relates to a tangential filtration device and a plant containing such a device.



  Cross flow filtration differs significantly from flow filtration. In cross-flow filtration, the liquid is supplied parallel to the filter surface, and the filtration takes place perpendicular to the direction of flow of the feed material. In general, in cross-flow filtration systems, an accumulation of the filtered out substances on the filtration devices is reduced due to the shear forces of the flow, since the flow of the feed material is tangential to the membrane surface. Crossflow filtration therefore opens up the possibility of working in the so-called "steady state", ie in equilibrium, if the pressure drop causing it is kept constant. Unfortunately, this theoretical possibility could not be achieved in practice. Therefore, conventional cross-flow filtration systems have always suffered from the problem of decreasing filtrate flow.

  Most of the suspended solids were retained on the pipe wall and quickly formed a dynamic membrane, also known as a filter cake or sludge layer. This dynamic membrane is largely responsible for the subsequent filtration.



  The particles that first penetrate into the wall matrix are ultimately retained in it due to the irregular and tortuous shape of the pore structure. As microfiltration progresses, the penetration of further small particles through the wall matrix through the dynamic membrane is prevented. The formation of the dynamic membrane together with the possible clogging of the pore structure of the tube by the retained particles leads to a decrease in the filtrate flow. In conventional systems, this decrease is approximately exponential to the filtration time.



  Cross-flow filtration of milk has been tried, but has not been widely accepted because of the problems just discussed. The U.S. Patent No. 5,028,436 relates to a process for separating into dissolved and undissolved milk components using a microporous membrane with a pore size in the range from 0.1 to 2 microns, which is treated with an aqueous solution or a dispersion or emulsion of lipids or peptides and which is pre-treated Membrane separated milk was pretreated.



  In the process disclosed in the patent, a first filtration step is used, in which a microporous membrane is used in the tangential flow process. A clear filtrate and a viscous concentrate are obtained. The filtrate contains all salts, lactose, amino acids, oligopeptides and polypeptides of low molecular weight in native, undenatured form. The concentrate contains practically all of the casein and the fat components of the milk. Therefore, once all the fat substances have been removed from it, the filtrate cannot be considered "milk".



   The U.S. Patent No. 4,876,100 relates to a cross-flow filtration process for the production of milk with a reduced bacterial content. Raw milk is separated by centrifugation into a fraction consisting of cream and another containing skimmed milk. The skimmed milk fraction is passed into a microfilter, in which part of the fat globules, proteins and bacteria are separated. A filtrate and a concentrate are obtained from the microfilter. The filtrate consists of skimmed milk with a reduced fat, protein and bacterial content, whereas the concentrate has an increased fat, protein and bacterial content. The concentrate is then sterilized.

  In addition to reducing the bacterial content of the filtrate, the filtration method of the '100 patent also reduces the fat and protein content of the filtrate, changing its properties compared to the original skimmed milk.



  Obviously, the use of crossflow filtration has not provided an acceptable method to reduce bacterial contamination of milk.



  Dynamic microfiltration is one way to overcome the problems associated with classic cross-flow filtration technology. The dynamic microfiltration process overcomes the disadvantage of classic cross-flow technology, since the liquid is not simply passed tangentially over the membrane surface. The membrane surface or a solid body in the vicinity of the membrane surface are moved such that the liquid is subjected to shear forces at the interface between the rotor and the stator. The shear forces mechanically clean the membrane surface, keeping it relatively clean from particulate material and preventing the formation of a "filter cake on the membrane surface".

  The particulate material that would otherwise accumulate on the membrane surface remains in suspension and is finally removed in the secondary stream, the so-called concentration stream.



  Dynamic microfiltration systems can take various forms. E.g. affect the U.S. Patents Nos. 5,037,562; 3,997,447; 5,037,562; 3 997 447 and 4 956 102 dynamic disc (disk) microfiltration systems.



  Cylindrical dynamic microfiltration devices are known in the U.S. Patents Nos. 4,956,102; 4,900,440; 4,427,552; 4,093,552; 4,066,554 and 3,797,662 and many others. The disclosure of all patents cited in this application is hereby incorporated by reference. So far, no dynamic microfiltration has been used for the production of milk and the use of cross-flow filtration of milk has been limited and has been used principally to fractionate milk based on the fat content in components.



  It has now surprisingly been discovered that the dynamic microfiltration of milk can be carried out successfully by the method according to the invention without the known problems of destroying the milk quality, the premature clogging of the filters and the inadequate removal of the bacteria.



  According to the invention, the milk, either whole or skimmed milk, is first homogenized and then fed to a filtration. When carrying out the homogenization as the first step, the particle size of the fat globules and other large, floating components of the milk is significantly reduced, which enables the microfiltration of the milk without fat or other components being significantly removed or entrained.



  Milk is an emulsion of fat and protein particles in water. Homogenization offers a method of reducing the size of the emulsified particles in order to allow passage through an appropriately dimensioned microporous membrane and thereby to retain the bacteria contained therein, without undesired removal of the fat and protein content of the milk.



  After homogenization, the milk is filtered using dynamic microfiltration. The invention therefore provides an improved process for the production of milk with a reduced bacterial content without the milk having to be pasteurized. The portion of the milk fraction retained by the microfilter (the concentrate fraction) can be recirculated as part of the feed, but can also be removed or used in other processes.



  Thus, this invention provides a method of treating raw milk to produce treated milk with a reduced bacterial content compared to raw milk. The method involves homogenizing the milk and adding dynamic microfiltration within about 5 minutes after homogenization, passing the milk through a microfilter whose average pore size is sufficient to reduce the bacterial content of the milk flowing through and around to obtain a filtrate with a bacterial content lower than that of the raw raw milk and to a concentrate with a bacterial content higher than the raw raw milk.

  The milk thus obtained has a very low bacterial content of around 10 <3> bacteria / ml or less and contains more organoleptic components than pasteurized milk of the same bacterial content.



  The milk which can be obtained as a product of the process according to the invention is generally more stable than milk which has been pasteurized in a conventional manner. There is significant bacterial residue in pasteurized milk because milk naturally contains certain bacteria that survive the pasteurization process. Therefore, pasteurized milk must be kept refrigerated to reduce bacterial growth and spoilage.



  Unfortunately, some of the bacteria in raw milk are both heat-stable (bacteria that survive pasteurization) and psychrophilic (bacteria that grow at low temperatures below 15 degrees Celsius), such as Bacillus cereus. The presence of thermostable, psychrophilic bacteria in packaged milk products is very disadvantageous since their rapid growth leads to spoilage of the milk even under cooling conditions.



   This invention can produce sterile milk that can be used even at room temperature for extended periods, e.g. can be stored for more than 30 days. The sterile milk according to the invention can generally be characterized by the absence of bacteria, in particular by the absence of the following bacteria and pathogens:

   
 <tb> <TABLE> Columns = 2
 <tb> <SEP> Thermostable bacteria
 <tb> <SEP> Micrococcus <SEP> M. luteus, M. roseus
 <tb> <CEL AL = L> Streptococcus <SEP> S. pneumoniae, S. lactis, S. faecalis
 <tb> <CEL AL = L> Lactobacillus <SEP> L. delbrueckii, L. lactis,
 L. helveticus, L. casei, L. trichodes
 <tb> <SEP> Staphylococcus <SEP> S. aureus, S. eoidermidis,
 <tb> <SEP> Bacillus <CEL AL = L> B. cereus, B. subtilis, <SEP> B. macerans, B. stearothermophilus
 <tb> <SEP> Clostridium <SEP> C. butyrium, C. pasteurianum,
 C. botulinum, C. perfringens
 C. tetani
 <tb> <SEP> Psychrophilic bacteria
 <tb> <SEP> Pseudomonas <SEP> P. aeruginosa, P. fluorescens
 P. pseudomallei, P. mallei
 <tb> <SEP> Archromobacter
 Alcaligenes
 <tb> <SEP> Acientobacter <SEP> A. lignieressii, A.

  A. equirli
 <tb> <CEL AL = L> Flavobacterium <SEP> F. aquatile, F. menigosepticum
 <tb> <SEP> Bacillus <SEP> B. cereus, B. subtilis,
 B. macerans,
 B. stearothermophilus
 <tb> <SEP> Coliform bacteria
 <tb> <SEP> Enterobacter <SEP> E. coli, Salmonella Typhi
 Shigella dysenteriae
 Klebsiella pneumoniae
 <tb> <SEP> others
 <tb> <SEP> Listeria <SEP> L. monocytogenes
 <tb> </TABLE>



  The milk according to the invention can therefore meet or exceed the conditions for pasteurized milk of class A, according to which the number of colonies of bacteria in milk determined using standard methods must not exceed 30,000 per ml and coliform bacteria may only be present with less than 10 colonies per ml.



  In another embodiment, the invention provides a method of treating raw milk to produce treated milk with a reduced bacterial content compared to raw milk. This procedure involves (1) separating the milk into a fat fraction with a minimum fat content of about 10% and into a skimmed milk fraction, (2) homogenizing the skimmed milk fraction and subjecting it to dynamic microfiltration within 5 minutes after homogenization, the skimmed milk fraction is passed through a microfilter with an average pore size sufficient to reduce the bacterial content of the milk flowing through and the resulting filtrate has a lower bacterial content than the original fraction of skimmed milk,

   while the concentrate has a higher bacterial content than the original fraction of skimmed milk; (3) separately, reduction in the bacterial content of the fat fraction; and (4) subsequent combination of the skimmed milk fraction after microfiltration and the fat fraction with reduced bacterial content.



  In a further embodiment, the invention provides a method for producing milk with a fat content of approximately 2%, with:



  (1) homogenize the skimmed milk fraction and, within 5 minutes after homogenization, (2) deliver the skimmed milk fraction to dynamic microfiltration by passing the skimmed milk fraction through a medium pore size microfilter sufficient to add the bacterial content to the milk flowing through decrease, the resulting filtrate having a lower bacterial content than the original skimmed milk fraction, while the concentrate has a higher bacterial content than the original skimmed milk fraction, (3) reducing the bacterial content of a cream fraction with a minimum fat content of about 10%, and (4 ) subsequent combination of the skimmed milk fraction after microfiltration and the cream fraction with reduced bacterial content.



  The invention also provides a method of treating milk for consumer consumption, comprising: obtaining raw milk; Homogenize the milk; within about 5 minutes after homogenization, subjecting the skimmed milk fraction to dynamic microfiltration, the milk fraction being passed through a microfilter with a medium pore size sufficient to reduce the bacterial content of the milk flowing through to a filtrate with a lower bacterial content than that to obtain original raw milk; Pack the milk in containers suitable for use by the customer, and transport the milk without cooling to a point from which the milk can be distributed to the consumer.



  In general, the invention provides a milk distribution method for consumer consumption, comprising: obtaining raw milk, the milk bacterial content at a level of 10 <3> to lower bacteria per ml or less, to pack the milk in a container for use by the consumer and to transport it to a point without cooling where it can be distributed to the consumer. This makes refrigerated transport and delivery vehicles unnecessary.



  The invention is explained in more detail below with reference to the accompanying drawings, in which:
 
   Fig. 1 is a schematic diagram of the equipment used in the inventive method, and
   2 shows the particle sizes of the milk after homogenization
 



  shows.



  The starting material is fresh, untreated raw milk from a domesticated animal, e.g. a cow. The method according to the invention can also be applied to treated milk, e.g. pasteurized, are used, although then not all advantages, such as the production of milk with improved organoleptic properties compared to pasteurized milk, come into play.



  The raw milk to be treated can first be passed through a heat exchanger to bring it to a suitable temperature and, if desired, then passed through a centrifugal separator to remove all or part of the cream fraction in a conventional manner.



  As an overview: The raw milk is homogenized and passed through a dynamic microfilter fairly quickly, resulting in a filtrate fraction and a concentrate fraction. The pores of the microfilter are dimensioned such that they retain at least some of the bacteria. The filtrate, which is the portion of the milk fraction that passes through the retention surface of the microfilter, consists of milk with or without a reduced bacterial content (compared to milk before microfiltration), the fat and protein content remaining essentially unchanged. The filtrate fraction can then either be used directly to make other products e.g. Milk powder, can be used or packaged without further treatment.



   The filtrate fraction is more desirable than the milk obtained by conventional pasteurization for many reasons. It contains more organoleptic components than pasteurized milk, which makes it more substantial in taste, according to the wishes of the consumer. In addition, the milk obtained according to the invention has a considerably longer shelf life, since e.g. psychrophilic bacteria such as Bacillus cereus can be completely removed with this invention, which was impossible with conventional pasteurization.



  The concentrate fraction, i.e. the portion of the milk fraction that is retained by the membrane surface of the microfilter from which it can be recovered consists of milk with an increased bacterial content (compared to the milk used before microfiltration), the fat globules and protein content being essentially different does not change. The concentrate fraction can then be discarded or used in other processes.



  The filtrate may still contain some bacteria, but the lower the bacterial content, the more durable the product. Complete sterilization is desirable, but the initial growth rate of a low concentration of remaining bacteria is usually low enough to give the milk product a significantly longer shelf life.



  The shelf life of the milk, which is produced according to the method according to the invention, is considerably longer than that of conventionally pasteurized milk, since the concentration of Bacillus cereus bacteria in particular is greatly reduced.



  Because the milk of the present invention can be made sterile (whereas milk obtained using conventional pasteurization techniques cannot be really sterile), it can have an extremely long shelf life at room temperature or under refrigerated conditions, especially when the milk is stored in containers under aseptic conditions is filled. One way to do this is through the form-fill-seal technique known in the packaging industry. The technique is often used for packaging sterile solutions or the like, e.g. in the pharmaceutical industry. The milk produced according to the invention can be filled using the form-fill-seal technique. This milk can then have an extremely long shelf life even at room temperature.



  The exact process or plant for filling is not critical. The following description is presented only by way of example and to explain the form-fill-seal technique.



  Some vertical form-fill-seal systems use flat sheet material of a synthetic, thermoplastic film, which is unwound from a roll and formed into an endless hose in a tube-forming process section by the longitudinal edges of the film being connected to one another in a watertight manner. In other devices, the tube is extruded from a molten resin at the time of use. The hose formed in this way is conveyed on to a filling station, where it is folded along its cross-section. The cross section lies at a sealing station below the filling station. The sealing station creates a transverse weld seam on the folded part of the tube, which seals the tube airtight. After applying the cross seal seam, a certain amount of the material to be packed, e.g.

  Liquid, filled into the hose at the filling station and the hose filled above the cross seal mentioned. The tube is then moved down a predetermined distance and sealed and cut off at the second cross station. Such a vertical form-fill-seal system, as just described, is sold under the trademark PREPAC, another is in U.S. U.S. Patent No. 5,038,550.



  After a possible separation by centrifugation, the milk fraction is preferably first heated or cooled to a temperature which can be used for the homogenization. The milk is then placed in a homogenizer, in which the size of the emulsified fat is reduced to such an extent that it can pass through the membrane. A size of all suspended particles of less than about 1 micron is preferred. It is important that the milk is filtered relatively soon after homogenization. The filtration is preferably carried out within less than about 5 minutes, particularly preferably within less than about 2 minutes and very particularly preferably in less than about 30 seconds after homogenization.



  The important point here is not the time that elapses before filtration, but the conduct of the filtration prior to substantial agglomeration of the beads, which leads to the formation of a substantial number of particles larger than about 1 micron in size.



  It is essential to homogenize the skimmed or whole milk in a cylindrical, dynamic microfiltration unit before filtration in order to thoroughly emulsify and suspend the fat components and other components of the milk, to reduce their size sufficiently and to achieve perfect filtration. A rotating disc filter develops significant shear forces directly on the surface of the rotating disc. Therefore, some degree of homogenization of the milk can occur substantially simultaneously with the filtration. Such "self" emulsification of the milk by the action of the dynamic microfilter enables skimmed milk to be produced using a rotating disc filter without the need for a separate homogenizer.



  In fact, the environment of the rotary disc filter causes the milk to be both homogenized and filtered at the same time, which is not achieved by a cylindrical rotary filter unit. A rotary disc filter can have a shear rate of approximately 200,000 sec <-> 1 cause, whereas a rotating, cylindrical unit only a shear rate of about 10,000 sec <-> <1> causes.



   Although the shear forces in a rotating disc filter unit are considerable, in most cases it cannot be assumed that they are sufficient to homogenize whole milk to the extent required.


 Dynamic filtration
 



  According to the invention, the filtration is carried out as dynamic filtration, i.e. that the filter medium itself is kept in constant motion so that the actual flow rate of the milk over the filter is extremely high. The special physical shape of the dynamic membrane is not critical. The membrane can e.g. have the shape of disks or cylinders. Such dynamic microfiltration devices have been previously discussed and are suitable for use in the invention. In general, the dynamic microfilter consists of a cylindrical or disc-shaped membrane element which rotates in an impermeable cylinder surrounding it.

  In a cylindrical, dynamic microfilter, when the liquid to be filtered is placed in the gap between the stator and the rotating membrane, the impulse of the rotating membrane is transmitted to the liquid. The liquid near the inner cylinder experiences a greater centrifugal force than the liquid near the outer cylinder. Under certain conditions, this phenomenon creates a flow pattern known as the Taylor vortex that prevents the build-up of significant residues on the membrane surface.



  The dynamic filtration process uses the development of Taylor vortices by keeping the surface of the membrane free of possible residues, the components of which remain suspended in the liquid to be filtered. The process divides the feed into a filtrate (the portion of the liquid that passes through the membrane) and a concentrate (the fraction that contains the suspended particles that would normally have settled on the surface of the membrane and thereby clogged the membrane). In this way, a high flow rate through the membrane can be maintained over a long period of time. The amount of feed and concentrate must be controlled so that there is a constant flow. Even with low concentrate flow rates, it is possible to maintain a steady flow of liquid to the membrane surface.



  The dynamic microfiltration enables a wide range of effective surface speeds of the filter relative to the milk. E.g. an effective surface speed of about 3 m / sec to about 50 m / sec can be used, preferably from about 5 m / sec to about 30 m / sec and be particularly preferably from about 8 m / sec to about 20 m / sec.



  In order to achieve the desired surface speeds, a representative filter medium in the form of a cylinder with a diameter of approximately 6.35 cm must be rotated at a speed of approximately 1000 to approximately 6000 revolutions per minute (rpm), usually taking a speed of 5000 rpm becomes.



  When a dynamic disc filtration device is used, the typical disc filter has a diameter of about 5.08 to about 121.92 cm. Such disks can e.g. at speeds of about 1000 rpm to about 8000 rpm (usually from about 3000 to about 6000 rpm), depending on the design of this particular dynamic microfilter being used.



  The shear rates of such a disc filter are preferably between approximately 100,000 sec <-> <1> up to about 400,000 sec <-> <1>. Preferred disk filters include those of the type described in U.S. Patent application No. 07/812 123, filed on December 24, 1991. The associated patent specification is introduced in this patent specification as a disclosure to avoid repetition.



  The pore size of the microfilter is selected so that the bacteria contained in the milk are retained and at the same time an acceptable flow rate through the microfilter is maintained. Suitable membranes are hydrophilic, microporous membranes with good flow properties, narrow pore size distribution and constant performance in removing the bacteria of interest. The pore size of the microfilter membrane should range from about 0.01 micron to about 5.0 micron, as can be determined using prior art methods. The tests are known as the bubble point (ASTM F316-86) and KL method (U.S. Patent No. 4,340,479). Preferably, the pore size will range from about 0.1 to about 1.0 micron. Filters whose pore size ranges from about 0.2 to about 0.5 micron are particularly preferred.

  Such microporous filters are well known and easy to obtain.



  Preferred microporous membranes that can be used in the present invention include those from Pall Corporation under the trademark Ultipor N66 <TM>, Fluorodyne <TM> and Posidyne <TM>; from Cuno Corporation under the trademark Zetapor <TM> and from Millipore under the trademark Durapore <TM> are sold.



  The cylindrical membrane elements which are used in accordance with the invention include those which are leak-tightly fastened to a holder using known methods.



  Finally, the bacteria should be concentrated in a stream that contains less than 5% of the feed; more than 95% of the solids and proteins normally contained in milk should pass the membrane over a longer period of time.



  The dynamic microfilter can be operated for a single pass without the concentrate having to be fed back. If desired, the concentrate can be returned to the feed. If a cylindrical, dynamic microfilter is used, it can be operated with different ratios of filtrate flow to the total feed flow (concentration factors). A filtrate / feed ratio of more than 90%, better still about 95% and in particular more than 98% is advantageous when operating a cylindrical dynamic microfilter in order mainly to obtain the desired filtrate, which contains very few bacteria.



   The situation is similar when a rotating, disc-shaped dynamic microfilter is used. It can also be operated with different filtrate flow to total feed flow ratios, with a wide variation of filtrate / feed ratio being possible. Choosing a large ratio leads to low flow, while a small ratio means higher flow. Operation at a ratio of about 40% is believed to be advantageous in order to maintain a constant flow rate through the filter, although other ratios can be used.



  The filtration of the freshly homogenized milk can be carried out warm at 40 ° C. to 60 ° C., which corresponds approximately to the crystallization temperature of the higher melting components of the milk fat, which is around 40 ° C. This value is lower than that used in conventional thermal pasteurization. Alternatively, with a slight decrease in flow rate, the milk can be filtered at much lower temperatures, e.g. between about 15 ° C. and about 35 ° C., in particular from about 20 ° C. to about 25 ° C.


 General
 



  After microfiltration, the concentrate can be removed in any suitable manner, subjected to further treatment or used directly.



  The method of this invention can advantageously be used when the desired end product is either whole milk, standardized milk or skimmed milk.



  The flow rate through a membrane that retains the bacteria is normally higher in milk with a reduced fat content than in milk with a high fat content. In certain situations it is economically more advantageous to produce milk with a higher fat content (e.g. milk with 2% fat) by mixing filtered, skimmed milk with a filtered fat fraction. This fat fraction can be a cream fraction with a minimum fat content of around 10%.



  The filtration of the cream fraction can be carried out using the 07/952 337 disclosed methods or using a filter cartridge that has at one end a filter impermeable to bacteria. The filtration can be carried out in an industrially acceptable manner by heating the fat mixture to a point where it is in the liquid state and can be easily filtered through a microporous membrane. The preheated fat can be homogenized before filtration. Alternatively, the fat mixture can be fed to pasteurization to reduce the bacterial content, or a combination of pasteurization and microfiltration can be used.



  Furthermore, if the method is to obtain a protein concentrate of the milk of a transgenic animal (e.g. a transgenic cow), the dynamic microfiltration can be performed using a microporous membrane with a pore size of about 0.2 micron or less to obtain a highly concentrated concentrate to reach.



  A suitable device for carrying out the method according to the invention can be constructed by connecting conventional devices such as centrifuges, microfilters, sterilization units, heat exchangers and pumps to one another. Those skilled in the art will readily be able to provide flow and pressure control valves and other necessary aids to make such a system operational and to provide other common modifications necessary for the particular application.



  All references mentioned so far are included here as cross-references. The following examples illustrate certain application forms, but are in no way intended to limit the scope of the invention as disclosed in the claims.


 General procedure
 



  The following general procedures were used for the examples.


 Procedure A:
 Milk temperature setting
 



  Unless otherwise specified, the milk used in the following examples is milk purchased from retailers.



  The temperature of the milk was brought to a process temperature suitable for the treatment before the filtration. The preferred operating temperature (40-60 ° C) was used because most of the milk fat is in this form in non-crystallized form. The process took place in a jacketed 35 l fermentation vessel (Type 3000 from Chemap A.G.). The jar was filled with the milk and the contents were heated to about 50 ° C. with a warm water jacket, unless stated otherwise. The milk was stirred during the heating process to speed up the heat exchange.



  As soon as the milk had reached the desired process temperature, it was pumped into the homogenizer at a rate of about 1 l / min.


 Procedure B:
 Homogenization of the milk
 



  Once in the homogenizer (Model 15 MR from APV Gaulin, Inc.), the milk is subjected to a two-phase homogenization process, with a first step at about 17.5 x 10 <6> Pa (2500 psi) and a second at about 35 x 10 <5> Pa (500 psi). The procedures for starting and operating this unit specified in the APV Gaulin operating manual were followed. After homogenization, the milk was usually transferred to a jacketed pressure vessel in which the desired process temperature was maintained. This container acted as a liquid buffer between the homogenizer outlet and the output to the filter.

  Whenever desired, the homogenized milk could be returned through the homogenizer to maintain a constant volume in the container.


 Procedure C:
 Introducing bacteria into the milk feed stream
 



  In some experiments, the milk flow was artificially inoculated with bacteria in order to demonstrate the extremely high titer reduction possible with this invention. The bacterial suspension was added to the feed stream via a measuring pump connected between the treatment vessel and the homogenizer. The flow rate of the inoculum was maintained so that a desired bacterial concentration of about 106 bacteria per ml milk was achieved. Since the bacteria were added before homogenization, the bacteria were mixed well in the liquid to be treated before reaching the filtration device. E.coli strain ATCC 15 224 was used in most of the examples.



   Another method of inoculating the milk with bacteria would be to add the bacteria directly to the process vessel to the desired concentration. Such a method is not preferred because it exposes the bacteria to temperatures above the ambient temperature that are too long. Depending on the strain used, this could lead to undesirable growth or excessive killing of the bacteria before they have reached the filtration device.


 Procedure D:
 Bacteria detection
 



  Mesophiles: the bacterial concentration was determined by passing the appropriate dilution through a sterile membrane with a pore size of 0.2 micron after a series of dilutions of the sample and subsequently culturing on Mueller-Hilton agar at 32 ° C. for 24 h. These procedures are described in "Manual of Clinical Microbiology, 2nd Edition, 1974, ASM, Washington, D.C." described in detail.



  Listeria monocytogenes, ATCC 43256, was the pathogen tested. The population density in the samples was determined using the method according to Agello et al. (Agello, G., Hayes, P. and Feeley, J .: Abstracts of the Annual Meeting. 1986, ASM, Washington DC, p. 5).


 Procedure E:
 Cleaning process
 



  Disinfection and sterilization were performed with 0.1 N NaOH before each experiment. In the sterilization process, the membrane and the entire associated device were first rinsed with water and then treated for about 1/4 h at 50 ° C. with 0.1 N NaOH. The caustic liquid was then neutralized with phosphoric acid. This neutralized solution was used to flush the system until all parts were neutralized. The filtration tests were carried out immediately after this procedure. The entire device and filter elements were disinfected using the sterilization procedure after each test.


 Procedure F:
 Integrity test
 



  Each membrane element was checked for integrity before each test (bacterial challenge). A "forward flow test" as described in the NM 900a publication "The Pall Ultipor membrane filter guide", copyright 1980, available from Pall Corporation, was used as an integrity test.


 Description of the filter device
 


 1. The cylindrical dynamic microfilter
 



  The cylindrical, dynamic microfilter (cylindrical DMF) used for these tests was of the BDF-01 type, manufactured by Sulzer Brothers Limited, Winterthur, Switzerland. The equipment is at Rebsamen et al. (Dynamic Microfiltration and Ultrafiltration in Biotechnology, Rebsamen, E. and Zeigler, H., Proceedings of the World Filtration Congress IV, 1986, (Ostend, B)). See also U.S. Patents Nos. 4,066,554 and 4,093,552, which are included here as references / cross references.


 2. Description of the membrane filter elements
 



  The filter elements typically used in these experiments were different classes of nylon membranes, Ultipor N66 <TM> and Posidyne <TM>, available from Pall Corporation, Glen Cove, N.Y. The pore sizes were 0.2, 0.30, 0.45 and 0.65 micron. The surface of the membrane elements had a size of 0.04 m <2>.


 3. The dynamic microfilter in disk form
 



  The disc-shaped version consists of a disc-shaped membrane holder with a diameter of 15.24 cm, which is attached to a hollow shaft and is located in a watertight housing with the necessary connections for the liquid supply and discharge. The retaining disc has devices to secure the membranes against leakage and contains drainage channels to allow the filtrate to flow through the membrane and the disc and through the shaft to the outside. The effective membrane area was 0.014 m <2>, and speeds of rotation of up to 4500 rpm can be achieved.



  Any of the dynamic disk microfiltration units listed above can be used according to the invention. For the description of another dynamic microfiltration unit in disk format which can be used according to the invention, reference is also made to the U.S. Patent application No. 07/812 123, filed December 24, 1991.


 4. Description of the membrane filter elements
 



  The membrane filter elements belonged to the same class of membranes as described in the section "cylindrical DMF". Usually the membranes consist of round, smooth rings ("donuts"), which are cut to fit the disc-shaped DMF. As soon as they were installed in the dynamic microfilter, the filtrate chamber was sealed against the feed material using O-rings. The membrane filter elements had a surface area of 0.014 m <2>.


 Procedure G1:
 Operation of the cylindrical dynamic microfilter
 



  Before filtration, a filter element was installed in the cylindrical dynamic microfilter (DMF) as described in the section on filter assembly. Disinfection and sterilization were carried out according to procedure E. After adhering to the procedures for starting up the system carried out in method G2, the milk to be filtered was pumped from the container into the cylindrical DMF via a displacement pump. The amount of concentrate is controlled by a second pump or a pressure relief valve, which are attached to the concentrate opening. The temperature and flow rate of the feed, the filtrate and the concentrate, and the feed pressure were taken at different times, usually at 10 min intervals, during the conduct of the experiment.

  Standard conditions for operating the cylindrical DMF were a rotation speed of 5000 rpm, a ratio of filtrate to feed of more than 95% and a feed pressure of about 1.3 x 10 <5> -2.0 x 105 (1.3-2.0 bar). All examples carried out with this device were carried out with a constant flow rate of the feed material.


 Procedure G2:
 Commissioning of the dynamic filter
 



  Before the milk was placed in the dynamic filter, warm, deionized, filtered (0.2 micron filter) water was sent through the system to start up the associated machines. The speed of rotation of the dynamic filter was brought up to operating speed with the water flowing through the system. As soon as the system reached equilibrium, the milk flow was switched on. The milk displaced the water in the system and the filtration started.


 Procedure H:
 Operating the dynamic disc microfilter
 



   A disc DMF filter element, as described in the section on filter assembly, was installed in the disc DMF. Disinfection and sterilization was carried out according to procedure E. After adhering to the procedures for starting up the system carried out in method G2, the milk to be filtered was pumped from the pressure container into the disk DMF. The amount of concentrate and feed pressure were controlled by a valve located at the concentrate opening. The temperature and flow rate of the feed, filtrate, concentrate, and feed pressure were measured at different times, usually at 10 min intervals, during the conduct of the experiment. A feed rate of 960 ml / min was observed in all examples.

  The filtrate fluxes listed were those which were achieved when the flow in the filter unit had stabilized.


 Examples
 


 example 1
 



  Skimmed milk at room temperature was placed at a rate of 600 ml / min in a cylinder DMF made with a 0.45 micron Ultipor N66 <TM> membrane was populated, pumped. The operating conditions in the DMF described in more detail in method G1 were adhered to and are summarized in Table 1. The feed pressure started to rise rapidly a few minutes after the start of the test, indicating a clogging of the microporous membrane.


 Example 2
 



  Skimmed milk was heated to 50 ° C. according to method A and homogenized according to method B. The homogenized milk was then stored in the intermediate container for about four hours. Meanwhile, the milk was kept at 50 ° C. After this four-hour break, the milk was fed into a 0.45 micron Ultipor N66 at a feed rate of about 600 ml / min <TM> Diaphragm-filled cylinder-DMF pumped. The preferred operating conditions for the DMF carried out in process G1 were observed. The feed pressure started to increase rapidly after only a few minutes of operation, which indicated that the microporous membrane was clogged, which is why the test had to be ended.


 Example 3
 



  Skimmed milk, heated to 50 ° C according to method A and homogenized according to method B, was placed in a 0.45 micron Ultipor N66 within no more than 5 minutes after homogenization <TM> Diaphragm-filled cylinder-DMF pumped. The preferred operating conditions for the DMF carried out in process G1 were observed. By the end of the milk supply there was a stable filtrate flow of 1080 l / h / m <2> complied with. No increase in feed pressure was observed during the course of the experiment. As soon as all the milk had been treated, the task was switched to 50 ° C. warm, non-homogenized, skimmed milk without disturbing the operation of the system. Within a few minutes, the filtrate flow dropped rapidly and the system pressure rose, indicating that the membrane had clogged.

  This example clearly shows that the milk has to be homogenized in order to obtain significant flow through the membrane.



  Examples 1-3 show that it is necessary to subject the milk to sufficient shear before filtration (in this case via homogenization) in order to reduce the size of the emulsified particles sufficiently. This enables the milk to pass through the microporous membrane, resulting in perfect filtration. Example 2 shows in particular that the particle size increases again within a short time after the homogenization. Therefore, for perfect filtration, the homogenization must take place shortly before the filtration, e.g. less than 5 minutes beforehand, with shorter intervals being even better.


 Example 4
 



  Skimmed milk was preheated according to procedure A and placed in a 0.45 micron Ultipor N66 <TM> Membrane-filled disc DMF pumped. The procedures described in Procedure H were used. A uniform flow of the filtrate was quickly established, which was then maintained for approximately 100 minutes until the milk supply was terminated.



  The operating conditions of the disc DMF produce a calculated shear rate of approximately 200,000 seconds <-> <1> in the space between the rotating disc and the membrane. This shear is in the range of the shear rates generated by the homogenizer under the conditions of method B.



  The example shows that the shear required before filtration can be achieved in one step, i.e. without the need for separate homogenizers.



  The example clearly shows that the membrane was not clogged by the solids in the milk and that the shear of about 200,000 seconds generated by the rotation of the disk <-> <1> was sufficient to reduce the particle size in the skimmed milk to such an extent that the milk could pass through the microfilter membrane and a perfect filtration was obtained.



  Table 1 summarizes the results of Examples 1-4; the data show that equilibrium filtrate flow through the membrane is achieved if the milk is sheared shortly before filtration.
 <tb> <TABLE> Columns = 8 Table 1
 <tb> Head Col 1: Example No.
 <tb> Head Col 2: filtration
 method
 <tb> Head Col 3: feed-
 temperetur
 (DEG C)
 <tb> Head Col 4: shear in the filter (sec <-> <1>)
 <tb> Head Col 5: homogeneous
 sation
 <tb> Head Col 6: break after homogeneous
 sation
 <tb> Head Col 7: pore sizes (micron)
 <tb> Head Col 8:

  flow
 (l / h / m <2>)
 <tb> <SEP> 1 <SEP> cylinder <SEP> 25 <SEP> 10,000 <SEP> no <SEP> no <SEP> 0.45 <CEL AL = L> 0
 <tb> <CEL AL = L> 2 <SEP> cylinder <SEP> 50 <SEP> 10,000 <SEP> yes <SEP> 4 hours <SEP> 0.45 <SEP> 0
 <tb> <SEP> 3 <CEL AL = L> cylinder <SEP> 50 <SEP> 10,000 <SEP> yes <SEP> 5 min. <SEP> 0.45 <SEP> 1080
 <tb> <SEP> 4 <CEL AL = L> disc <SEP> 50 <SEP> 200,000 <SEP> no <SEP> - <SEP> 0.45 <SEP> 1600
 <tb> </TABLE>


 Example 5
 



  In order to determine the relationship between the particle size and the time after homogenization, skimmed milk was heated according to method A and homogenized using the method described in method B. The distribution of the particle size as a function of the time after homogenization was determined. The particle size distribution was measured using an Integrated Micro-Optical Liquid Volumetric Sensor (IMOLV-.2), available from Particle Measurement Systems, Colorado. This laser particle counter is suitable for measuring particle sizes in the range of approximately 0.1-5.0 micron.



   The milk samples were diluted 1: 300,000 and then subjected to the test as described in the operating manual for the IMOLV device. Filtered (0.04 micron filter) deionized (DI) water at 18 mega-ohms with a particle number of less than 50 per ml was used to dilute the milk samples.



  Figure 2 shows the results of the particle analysis. In one figure, the number of particles in relation to the number of particles is plotted against the particle size after 5 seconds. The figure clearly shows that the number of larger particles increases with the time after homogenization. Since the number of smaller particles decreases accordingly in the same time, it is obvious that the small particles agglomerate into larger particles over time.


 Examples 6-9
 



  In the cylinder DMF, membranes with different pore sizes and different bacterial retention capacities were tested in order to determine the order of magnitude of the uniform filtrate flow of the milk. The procedure generally used in Examples 6-9 is given below.
 
   1. The desired membrane filter element was installed in the cylinder DMF.
   2. A test for integrity as described in Method F was performed and the membrane filter element was discarded if it failed the test.
   3. The device was disinfected according to procedure E.
   4. The milk to be filtered was preheated according to the procedure in procedure A.
   5. The milk was homogenized according to method B.
   6. The commissioning procedure specified in procedure G2 was carried out.
   7.

  The milk was transferred from the intermediate container into the cylinder DMF at the desired flow rate.
   8. The operating parameters were determined using the guidelines from Procedure G1.
   9. Appropriate measurements were made.
 



  Typically, the cylinder DMF was operated at 5000 rpm, corresponding to a shear rate of approximately 10,000 seconds <-> <1> in the filter. The temperature of the feed was 50 ° C and the feed pressure varied from 1.3 x 10 <5> -2.0 x 10 <5> Pa (1.3-2.0 bar). The filtrate to feed ratio was maintained above 95% for each of these examples. The flow shown in Table 2 is the equilibrium filtrate flow typically obtained 15 minutes after the start of filtration. The total duration of the experiment differed in each case, since the filtered volume of milk was constantly 30 liters.


 Example 6
 



  For this example, a 0.2 micron Ultipor N66 <TM> membrane used. The feed rate was 250 ml / min, around an equilibrium flow of the filtrate of 330 l / h / m To get <2>. The filtration was carried out for about 130 minutes without an apparent decrease in the filtrate flow rate until the milk was processed in the treatment vessel.


 Example 7
 



  For this example, a 0.30 micron Ultipor N66 <TM> membrane used. The feed rate was 550 ml / min, around an equilibrium flow of the filtrate of 775 l / h / m To get <2>. The experiment was ended after 60 min.


 Example 8
 



  For this example, a 0.45 micron Ultipor N66 <TM> membrane used. The feed rate was 740 ml / min, an equilibrium flow of the filtrate of 1080 l / h / m To get <2>. Filtration was carried out for about 40 minutes until the milk supply ran out, with no appreciable decrease in flow rate.


 Example 9
 



  In this example, a 0.65 micron Ultipor N66 <TM> membrane inserted. The feed rate was 110 ml / min and there was an equilibrium flow of the filtrate of 1680 l / h / m <2> received. The filtration was carried out over 30 minutes until the milk supply was exhausted.



  Examples 6-9 are summarized in Table 2. The data show that a stable filtrate flow can be obtained when using the filtration method according to the invention and when using different bacterial retention membrane stages. The table shows that membranes with smaller pores and consequently improved bacterial retention can be used at the expense of the filtrate flow rates in the process according to the invention.
 <tb> <TABLE> Columns = 10 Table 2:
 Milk flow when using different membranes in cylindrical DMF
 <tb> Head Col 1: Bei-
 game number
 <tb> Head Col 2: Fluid
 <tb> Head Col 3: Mem-
 bran
 <tb> Head Col 4: pore
 size (micron)
 <tb> Head Col 5: rpm
 <tb> Head Col 6: task-
 good temperature
 temperature (DEG C)
 <tb> Head Col 7: task-
 pressure
 (x 10 <5> Pa)
 <tb> Head Col 8: ratio of filtrate / feed material
 <tb> Head Col 9:

  Experimental
 duration (min.)
 <tb> Head Col 10: river
 (l / h / m <2>)
 <tb> <SEP> 6 <SEP> skimmed milk <SEP> Ultipor N66 <TM> <SEP> 0.20 <SEP> 5000 <SEP> 50 <CEL AL = L> 2.0 <SEP> 0.97 <SEP> 130 <SEP> 330
 <tb> <SEP> 7 <SEP> skimmed milk <SEP> Ultipor
 N66 <TM> <SEP> 0.30 <SEP> 5000 <SEP> 50 <SEP> 1.6 <SEP> 0.97 <SEP> 60 <SEP> 775
 <tb> <SEP> 8 <CEL AL = L> skimmed milk <SEP> Ultipor
 N66 <TM> <SEP> 0.45 <SEP> 5000 <SEP> 50 <SEP> 1.5 <SEP> 0.97 <SEP> 40 <SEP> 1080
 <tb> <SEP> 9 <CEL AL = L> skimmed milk <SEP> Ultipor
 N66 <TM> <SEP> 0.65 <SEP> 5000 <SEP> 50 <SEP> 1.3 <SEP> 0.97 <SEP> 30 <SEP> 1680
 <tb> </TABLE>


 Example 10
 



  In this example, a 0.2 micron Posidyne <TM> membrane with positive surface charge used. The pore surfaces of the membrane used are covered with quaternary ammonium groups and this has a high absorption capacity for biological material.



  The feed rate was 260 ml / min, an equilibrium flow of the filtrate of 360 l / h / m To get <2>. The filtrate flow was of the same order of magnitude as that obtained with uncharged membranes (described in Example 6). The filtration was carried out for about 120 minutes until the milk supply in the treatment tank was exhausted, without a noticeable decrease in the filtrate flow rate. During the experiment, a filtrate to feed ratio of over 97% was maintained. Other experimental conditions are given in Table 3.



   It was to be expected that a large part of the proteins from milk would bind to the membrane surface and eventually clog them. This example shows that in dynamic operation even a membrane that normally binds proteins behaves well.


 Example 11
 



  Whole milk was fed at a feed rate of 740 ml / min and a stable filtrate flow of 1130 l / h / m <2> received. Further experimental conditions are shown in Table 3. The filtration was carried out for about 40 minutes until the milk supply was exhausted. The example shows that whole milk can be filtered using the method according to the invention. The differences in filtrate flow observed between whole milk and skimmed milk (as in Example 9) seem to be mainly due to their viscosity differences.

  The ratio of the filtrate flow achieved for whole milk to skimmed milk is approximately equal to the ratio of the viscosities of whole milk to skimmed milk.
 <tb> <TABLE> Columns = 10 Table 3
 <tb> Head Col 1: Example No.
 <tb> Head Col 2: Fluid
 <tb> Head Col 3: Mem-
 bran
 <tb> Head Col 4: pore
 size (micron)
 <tb> Head Col 5: rpm
 <tb> Head Col 6: task-
 good temperature
 temperature (DEG C)
 <tb> Head Col 7: task-
 pressure
 (x 10 <5> Pa)
 <tb> Head Col 8: Filtrate / Auf- ratio
 good
 <tb> Head Col 9: experimental
 duration (min.)
 <tb> Head Col 10: river
 (l / h / m <2>)
 <tb> <SEP> 10 <SEP> skimmed milk <SEP> Posi-
 dyne <TM> <SEP> 0.20 <SEP> 5000 <SEP> 50 <SEP> 2.0 <SEP> 0.97 <SEP> 120 <SEP> 360
 <tb> <SEP> 11 <CEL AL = L> whole milk <SEP> Ultipor N66 <TM> <SEP> 0.65 <SEP> 5000 <SEP> 50 <SEP> 1.4 <SEP> 0.93 <SEP> 40 <CEL AL = L> 1130
 <tb> </TABLE>


 Examples 12-16
 



  The examples for determining the filtrate flow through various bacteria retention membranes were repeated using the disc-shaped dynamic microfilter. The general procedures for Examples 12-16 are described below. The generally described conditions apply to each example, unless expressly stated otherwise.
 
   1. The desired membrane filter element was installed in the disc DMF.
   2. As described in Method F, an integrity test was performed. The membrane filter element was removed if it failed the test.
   3. The device parts were disinfected according to procedure E.
   4. The milk to be filtered was preheated using the method described in Method A.
   5. The milk was homogenized according to method B.
   6.

  The general procedure for commissioning was carried out as described in procedure G2.
   7. The milk was transferred from the container to the disc DMF at a predetermined flow rate.
   8. Appropriate measurements have been made.
 



  A rotational speed of 3500 rpm was typically maintained, which corresponds to a shear speed of approximately 200,000 sec <-> Corresponds to <1>. The feed had a temperature of 50 ° C. and the feed pressure was approximately 0.2 × 10 <5> Pa (0.2 bar). The milk was pumped into the filter at a rate of 960 ml / min to maintain a high cross flow rate across the membrane. The ratio of the filtrate to the feed material was specially adapted to the pore size of the membrane, the feed material temperature and the rotor revolutions (rpm). The unfiltered part of the feed was returned to the process container. The flow shown in the following table is the steady state flow of the filtrate through the membrane, which was normally achieved half an hour after the start of the filtration.


 Example 12
 



  For this example, a 0.2 micron Ultipor N66 <TM> membrane used. The equilibrium filtrate flow was 850 l / h / m <2>.


 Example 13
 



  For this example, a 0.45 micron Ultipor N66 <TM> membrane used. The equilibrium filtrate flow was 1600 l / h / m <2>.


 Example 14
 



  For this example, a 0.45 micron Posidyne <TM> membrane used. An equilibrium filtrate flow of 1600 l / h / m <2> was received.



  The data presented in Table 4 summarize Examples 11-13. The data show that a stable filtrate flow can be obtained when using the filtration method according to the invention using different stages of bacteria retention membranes in disc DMF. The table shows that membranes with smaller pores and therefore higher bacterial retention (titer reduction) can be used according to the invention at the expense of the filtrate flow rate.
 <tb> <TABLE> Columns = 10 Table 4
 <tb> Head Col 1: Bei-
 game number
 <tb> Head Col 2: Fluid
 <tb> Head Col 3: membrane
 <tb> Head Col 4: pore
 size (micron)
 <tb> Head Col 5: rpm
 <tb> Head Col 6: task-
 good temperature
 temperature (DEG C)
 <tb> Head Col 7: task-
 pressure
 (x 10 <5> Pa)
 <tb> Head Col 8: ratio of filtrate / feed material
 <tb> Head Col 9: experimental
 duration
 (min)
 <tb> Head Col 10:

  flow
 (l / h / m <2>)
 <tb> <SEP> 12 <SEP> skimmed milk <SEP> Ultipor
 N66 <TM> <SEP> 0.20 <SEP> 3500 <SEP> 50 <SEP> 0.2 <SEP> 0.22 <SEP> 137 <SEP> 850
 <tb> <SEP> 13 <CEL AL = L> skimmed milk <SEP> Ultipor
 N66 <TM> <SEP> 0.45 <SEP> 3500 <SEP> 50 <SEP> 0.2 <SEP> 0.37 <SEP> 80 <SEP> 1600
 <tb> <SEP> 14 <CEL AL = L> skimmed milk <SEP> Posidyne <TM> <SEP> 0.45 <SEP> 3500 <SEP> 50 <SEP> 0.2 <SEP> 0.37 <SEP> 80 <CEL AL = L> 1600
 <tb> </TABLE>


 Example 15
 



  Skimmed milk at a temperature of 18 ° C is homogenized according to method B and into one with a 0.45 micron Ultipor N66 <TM> membrane-equipped disc-DMF pumped. Filtration was carried out at a feed rate of 860 ml / min and an equilibrium filtrate flow of 860 l / h / m <2> performed. The filtered milk was measured at 25 ° C. Other conditions of this example are shown in Table 5.



  The example shows that chilled, skimmed milk at around 18 ° C. can be processed with a method according to the invention through a bacteria retention membrane. The reduced flow of filtrate at this low temperature is probably due to the higher viscosity of the milk at this temperature compared to higher temperatures.


 Example 16
 



   A disc DMF was made with a 0.45 micron Ultipor N66 <TM> membrane equipped. Whole milk was added to the disk DMF at a rate of 900 ml / min and there was an equilibrium filtrate flow through the membrane of 850 l / h / m <2> reached. The experiment was performed without recycling the unfiltered portion of the feed stream.



  The example shows that whole milk can be filtered using a disc DMF using the method according to the invention. Skimmed milk gives an equilibrium filtrate flow of approximately 1600 l / h / m under essentially the same conditions <2>. The differences in filtrate flow observed between whole milk and skimmed milk correspond approximately to the ratio of the different viscosities of the liquids.


 Example 17
 



  Using the procedures described above, a disk DMF filtration experiment was performed using a high filtrate to feed ratio. The skimmed milk was fed in at 115 ml / min and a rotation speed of 2100 rpm. There was a filtrate flow of 460 l / h / m <2> received.
 <tb> <TABLE> Columns = 10 Table 5
 <tb> Head Col 1: Bei-
 game number
 <tb> Head Col 2: Fluid
 <tb> Head Col 3: Mem-
 bran
 <tb> Head Col 4: pore
 size (micron)
 <tb> Head Col 5: rpm
 <tb> Head Col 6: task-good temp-
 temperature (DEG C)
 <tb> Head Col 7: task-
 pressure
 (x 10 <5> Pa)
 <tb> Head Col 8: Filtrate / Auf- ratio
 good
 <tb> Head Col 9: experimental
 duration
 (min)
 <tb> Head Col 10:

  flow
 (l / h / m <2>)
 <tb> <SEP> 15 <SEP> skimmed milk <SEP> Ultipor N66 <TM> <SEP> 0.45 <SEP> 3500 <SEP> 18 <CEL AL = L> 0.2 <SEP> 0.23 <SEP> 130 <SEP> 860
 <tb> <SEP> 16 <SEP> whole milk <SEP> Ultipor N66 <TM> <CEL AL = L> 0.45 <SEP> 3500 <SEP> 50 <SEP> 0.2 <SEP> 0.23 <SEP> 90 <SEP> 850
 <tb> <SEP> 17 <SEP> skimmed milk <SEP> Ultipor N66 <TM> <SEP> 0.45 <SEP> 2100 <SEP> 50 <SEP> 0.5 <SEP> 0.92 <SEP> 50 <SEP> 460
 <tb> </TABLE>


 Example 18
 



  To demonstrate the long-term behavior, an experiment was carried out with a large amount (500 l) of raw, unpasteurized, skimmed milk. The milk was preheated to 50 ° C. using a plate heat exchanger. It was then homogenized according to method B and pumped into the cylinder DMF equipped with a 0.65 micron membrane. Typically, the dynamic microfilter was operated at 5000 rpm in this example. The feed pressure fluctuated between 1.3 x 10 <5> and 1.5 x 10 <5> Pa (1.3 and 1.5 bar) at a feed rate of approximately 1300 ml / min. The filtrate to feed ratio was maintained above 95%. There was an equilibrium filtrate flow of approximately 1680 l / h / m <2> received.

  During the six hours of continuous operation required to treat the 500 L, neither a decrease in the flow of filtered milk nor any increase in the feed pressure was observed.



  The example shows that it is possible to use the filtration method according to the invention over a long period of time.


 Example 19
 



  The example shows the capabilities of the method according to the invention for extracting protein from milk. Proteins in milk generally range in size from about 0.02 to about 0.3 micron (DG Schmidt, P. Walstra and W. Buchheim, Neth. Milk Dairy J. 27 (1973): 128), which makes it possible is to win them by means of the method according to the invention.



  This is particularly important for the extraction of biologically important proteins from transgenic animals, e.g. transgenic cows, sheep and the like, important, which have been genetically modified using techniques known per se to stimulate the production of such proteins.



  A disc DMF equipped with a 0.2 micron nylon filter was used. The filtration of the milk was carried out at a feed rate of 840 ml / min, at which an equilibrium flow through the membrane of about 850 l / h / m <2> was reached, and a rotor speed of 3500 rpm was carried out. The retentate was recycled in this process while the filtrate was discarded. Samples of the feed material, the retentate and the filtrate were taken at regular intervals, which were then examined for their total protein content using the Kjeldahl method. This indicated that the protein content of the retentate initially corresponded to that of the feed material, but increased with the duration of the recirculation of the retained material (4.9% retentate, 3.1% feed material).



  The use of membranes with a smaller pore size should result in an even better protein concentration in the concentrate stream.


 Examples 20 and 21
 



  The examples were carried out to demonstrate that no fractionation of the milk components takes place during the filtration process according to the invention. In these examples, samples of feed, filtrate, and concentrate were examined at various times during filtration to determine protein concentration by the Kjeldahl method and total solids by evaporation.


 Example 20
 



  Feed, filtrate and concentrate samples were taken at different times during the experiment described in Example 18 and examined for their total solids. The data in Table 6 show no noticeable loss of total solids in the filtrate when using a 0.45 micron membrane.


 Example 21
 



  Feed, filtrate, and concentrate samples were taken at different times while performing Example 13 and examined for total solids and protein levels in each stream. The results are shown in Table 6. Again, the filtered milk showed no significant loss of solids or proteins when using a 0.45 micron membrane.
 <tb> <TABLE> Columns = 8 Table 6
 <tb> Head Col 1: filtration
 method
 <tb> Head Col 2: membrane po
 ren size
 (micron)
 <tb> Head Col 3 to 5 AL = L: proteins%
 <tb> Head Col 6 to 8 AL = L: total solids%
 <tb> Head Col 3 AL = L: filtrate
 <tb> Head Col 3: task-
 Well
 <tb> Head Col 4: Concentrated
 kicked
 <tb> Head Col 5: filtrate
 <tb> Head Col 6: task-
 Well
 <tb> Head Col 7:

   Concentrated
 kicked
 <tb> <SEP> cylinder
 Example 20 <SEP> 0.65 <SEP> - <SEP> - <SEP> - <SEP> 9.14 <SEP> 9.17 <SEP> 9.35
 <tb> <SEP> disc
 Example 21 <SEP> 0.45 <SEP> 3.38 <SEP> 3.15 <SEP> 3.35 <SEP> 8.64 <SEP> 8.70 <SEP> 8.85
 <tb> </TABLE>


 Examples 22-28
 



  Experiments 22-28 were performed to demonstrate the ability of this invention to remove bacteria. The general operating procedure was the same as that used in Examples 6 to 18, except that Method C bacteria were added to the treatment stream. Unless otherwise noted, E. coli bacteria, which are commonly found in milk, were used in this experiment to inoculate.



  Feed, filtrate and bacterial concentrate samples were taken under sterile conditions at different times during the filtration. The samples were examined for bacteria using the method carried out in method D. The results are shown in Table 7.



  As can be seen from the table, it is possible according to the invention to achieve a significant reduction in the bacterial content of milk. The high removal rate of E.coli can be directly transferred to a high removal rate of Bacillus cereus, a bacterium that cannot be completely removed with conventional pasteurization. E.coli is known to be rod-shaped, with a size of approximately 1.1 to 1.5 μm to approximately 2 to 6 μm. Bacillus bacteria such as Bacillus cereus also have a rod shape and a similar size from approximately 1.0 to 1.2 μm to 3 to 5 μm. Therefore, the ability to remove E. coli, as shown in Table 7, means that the method also makes it possible to remove the highly undesirable bacterium Bacillus cereus. This leads to a milk with a very long shelf life, even at room temperature.


 Examples 22, 23, 24
 



  Examples 6, 8 and 9 were repeated with the change that E. coli was added to the treatment stream according to method C. Samples of the feed, the filtrate and the concentrate were taken for bacterial tests. The titer reduction data are given in Table 7.


 Example 25
 



  Example 13 was repeated with the change that bacteria were added to the feed stream according to method C and the bacterial concentrate was not returned to the treatment vessel. There was an equilibrium milk flow of about 1600 l / h / m <2> received. The microbiological results are shown in Table 7.



  The filtered milk contained only a very low content of approximately 7 to 10 bacteria per ml milk, i.e. significantly lower than the bacterial content of the feed material of 10 <6> per ml. In this case the titer reduction was greater than 10 <5>. For comparison: during conventional pasteurization of milk, titer decreases of only 10 <2> to 10 <3> reached.


 Example 26
 



  In this experiment, the experimental conditions and procedures of Example 12 were repeated with the change that according to procedure C. E. coli is added to the feed stream and the concentrate is not returned to the treatment vessel. There was an equilibrium milk flow of about 850 l / h / m <2> reached. Samples of the feed, the filtrate and the concentrate were taken for bacterial tests. The results in Table 7 show a titer reduction of more than 10 <6>. The sterile milk was treated until bacteria were no longer detectable in the filtered milk.



  This example shows the ability of the method according to the invention to remove almost completely the bacteria from the milk using a disc DMF and a suitably selected membrane. This means that sterile milk can be produced.


 Example 27
 



  Unpasteurized raw milk contains a wide range of organisms, including coliforms such as E. coli, pathogens such as Listeria and Campylobacteria and Bacillus cereus bacteria. In this example, the raw milk was not inoculated with external bacteria, but the milk was examined for its own or "natural" bacteria.



  For bacterial tests, 18 samples of the feed material, the filtrate and the concentrate were taken during the implementation of the example and examined for naturally occurring bacteria according to method D.



  Only 14 bacteria per ml milk were found. The feed contained 2500 bacteria per ml and the concentrate 2 x 10 <4> bacteria per ml. In addition, no psychrophilic bacteria could be detected in the filtrate. Psychrophilic bacteria are those that grow at low temperatures and cause chilled milk to spoil.



  Table 7 summarizes experiments 22 to 27. The results show that increased titer reductions are achieved at the expense of the filtrate flow in both the cylinder and disc operating modes. The table also shows that with the right choice of filter it is possible to get a sterile milk filtrate.
 <tb> <TABLE> Columns = 9 Table 7
 <tb> Head Col 1: Bei-
 game number
 <tb> Head Col 2: pore
 size
 (micron)
 <tb> Head Col 3: Mem-
 brand type
 <tb> Head Col 4: through
 sectional
 flow
 (l / h / m <2>)
 <tb> Head Col 5: task-
 good-
 temperature
 (DEG C)
 <tb> Head Col 6: operational
 pressure
 (x 10 <5> Pa)
 <tb> Head Col 7: task-
 well concentrated
 tration
 (Bakt./ml)
 <tb> Head Col 8:% bacteria
 rail removal
 nung
 <tb> Head Col 9:

  Titer
 reduction TR
 <tb> <SEP> 22 <SEP> 0.2 cylinder <SEP> Ultipor N66 <TM> <SEP> 330 <SEP> 50 <CEL AL = L> 2.0 <SEP> 10 <6> (E.coli) <SEP> 99.9992% <SEP> 10 <5>
 <tb> <SEP> 23 <SEP> 0.45 cylinder <SEP> Ultipor N66 <TM> <SEP> 1080 <SEP> 50 <SEP> 1.5 <SEP> 10 <6> (E.coli) <SEP> 99.992% <SEP> 4 x 10 <5>
 <tb> <SEP> 24 <CEL AL = L> 0.65 cylinder <SEP> Ultipor N66 <TM> <SEP> 1680 <SEP> 50 <SEP> 1.3 <SEP> 10 <6> (E.coli) <SEP> 99.96% <CEL AL = L> 4 x 10 <2>
 <tb> <SEP> 25 <SEP> 0.45 disc <SEP> Ultipor N66 <TM> <SEP> 1600 <SEP> 50 <SEP> 0.2 <CEL AL = L> 10 <6> (E.coli) <SEP> 99.9995% <SEP> 8x10 <5>
 <tb> <SEP> 26 <SEP> 0.2 disk <SEP> Ultipor N66 <TM> <CEL AL = L> 850 <SEP> 50 <SEP> 0.2 <SEP> 10 <6> (E.coli) <SEP> 100.00% (SEP) <> 10 <6>
 <tb> <SEP> 27 <SEP> 0.65 cylinder <SEP> Ultipor N66 <TM> <SEP> 1680 <SEP> 50 <SEP> 1.3 <SEP> 10 <3> (undefined
 acts) <SEP> 99.44% <SEP> 2 x 10 <2>
 <tb> </TABLE>


 Example 28
 



   In addition to the bacteria (E. coli) used in the titer reduction experiments, there are pathogenic organisms (such as listeria) in milk that are really worrying for the milk industry. These pathogens are a much more difficult challenge than the coliform bacteria (E. coli), which were also tested in the dynamic filter. In order to see whether the membrane filter elements used are able to effectively remove pathogens such as listeria, experiments were carried out using the methods described in method D. This test was carried out on a test facility located outside the facility (off-line test rig) and not with a dynamic filter.



  The data shown in Table 8 clearly show that a 0.45 μm Ultipor N66 <TM> Membrane with a specific bubble point (ASTM F316-86) enables complete removal of the listeria.
 <tb> <TABLE> Columns = 5 Table 8
 <tb> Head Col 1: membrane
 <tb> Head Col 2: bubble point
 <tb> Head Col 3: total allowance on
 Listeria bacteria / ml
 <tb> Head Col 4: filtrate
 Bakt./ml
 <tb> Head Col 5: titer reduction
 change TR
 <tb> <SEP> 0.45 µm, Ultipor N66 <TM> <SEP> 24 psi <SEP> 7.10 x 10 <7> <SEP> 0 (SEP) <> 5 x 10 <6>
 <tb> <SEP> 0.45 µm, Ultipor N66 <TM> <SEP> 22 psi <SEP> 6.6 x 10 <7> <SEP> 1 <SEP> 4.7 x 10 <6>
 <tb> <CEL AL = L> 0.65 µm, Ultipor N66 <TM> <SEP> 16.5 psi <SEP> 7.10 x 10 <7> <SEP> 7.5 x 10 <3> <SEP> 6.8 x 10 <2>
 <tb> </TABLE>


 Example 29
 



   Filtered, skimmed milk produced by the method according to Example 16 is collected in a disinfected container.



  Commercial cream is heated to 65 ° C. and through a 0.2 micron Ultipor N66 <TM> filter cartridge (available from Pall Corporation, East Hills, NY) with a minimum titer reduction of 10 <6>, achieved with E.coli, filtered. The filtered cream is essentially free of bacteria and is collected in a disinfected container.



  The filtered, skimmed milk and the filtered cream are then mixed and homogenized to obtain a milk containing 2% fat with a reduced bacterial content.


    

Claims (25)

      1. Process for the treatment of raw milk for the production of treated milk with a lower bacterial content than the raw milk, characterized in that the milk is homogenized and, within 5 minutes after homogenization, is fed to a dynamic microfiltration by passing the milk through a microfilter, the average Pore size is sufficient to reduce the bacterial content of the milk flowing through, is passed to obtain a filtrate with a lower bacterial content compared to the raw raw milk and a concentrate with a higher bacterial content compared to the raw raw milk. 2nd The method of claim 1, comprising:    (a) separating milk into a fat fraction with a minimum fat content of 10% and a skimmed milk fraction,    (b) Homogenize the skimmed milk fraction and perform dynamic microfiltration within 5 minutes after homogenization by passing the skimmed milk fraction through a microfilter with an average pore size sufficient to reduce the bacterial content of the milk flowing through to carry a filtrate obtain a lower bacterial content than the original skimmed milk fraction and a concentrate with a higher bacterial content than the original skimmed milk fraction,    (c) separate reduction in the bacterial content of the fat fraction,  and    (d) Combining the skimmed milk fraction after microfiltration and the fat fraction with reduced bacterial content. 3. The method according to claim 2, characterized in that the bacterial content of the fat fraction is reduced by dynamic microfiltration. 4. The method according to claim 2 or 3, characterized in that the bacterial content of the fat fraction is reduced by pasteurization. 5. A method according to claim 1 for the production of milk with a fat content of about 2%, comprising:    (a) homogenizing a skimmed milk fraction and    (b) subjecting the skimmed milk fraction to dynamic microfiltration within 5 minutes after homogenization by passing the skimmed milk fraction through a microfilter with an average pore size sufficient to reduce the bacterial content of the milk flowing therethrough to provide a filtrate with a obtain a lower bacterial content than the original skimmed milk fraction and a concentrate with a higher bacterial content than the original skimmed milk fraction,    (c) reducing the bacterial content of a cream fraction containing a minimum fat content of 10%,  and    (d) Combining the skimmed milk fraction after microfiltration and the cream fraction with reduced bacterial content. 6. The method according to any one of claims 1 to 5, characterized in that a rotary disk filter is used for microfiltration. 7. The method according to claim 6, characterized in that the homogenization takes place substantially simultaneously with the microfiltration. 8. The method according to claim 6 or 7, characterized in that the microfiltration is carried out with a shear rate of the rotary disk filter between 100,000 sec <-> <1> to 400,000 sec <-> <1>. 9. The method according to any one of claims 1 to 5, characterized in that a cylindrical, dynamic microfilter is used for microfiltration. 10th Method according to one of claims 1 to 9, characterized in that the concentrate is recirculated. 11. The method according to any one of claims 1 to 10, characterized in that the microfiltration is carried out at a temperature in the range from 15 ° C. to 60 ° C. 12. The method according to any one of claims 1 to 11, characterized in that the dynamic microfiltration is carried out with an effective surface speed of the microfilter of 3 m / sec to 50 m / sec. 13. The method according to any one of claims 1 to 12, characterized in that the microfiltration takes place within less than 30 seconds after the homogenization. 14. The method according to any one of claims 1 to 13, characterized in that a microfilter with a pore size of 0.01 to 5.0 microns is used. 15. A method according to claim 14, characterized in that a microfilter with a pore size of 0.1 to 1.0 micron is used. 16. The method according to claim 15, characterized in that a microfilter with a pore size of 0.2 to 0.5 micron is used. 17. The method according to any one of claims 1 to 16, characterized in that the bacterial content of the milk is reduced to a level of 10 3 bacteria per ml or less. 18. Milk obtained by treatment by the method according to any one of claims 1 to 17. 19. unpasteurized milk according to claim 18, characterized by the absence of Bacillus cereus bacteria. 20. unpasteurized milk according to claim 18, characterized by the absence of campylobacteria. 21. unpasteurized milk according to claim 18, characterized by the absence of listeria. 22.  Unpasteurized milk according to claim 18, characterized by the absence of Escherichia coli. 23. unpasteurized milk according to any one of claims 18 to 22, characterized in that it has a bacterial content of 103 bacteria / ml or less and contains more organoleptic components than pasteurized milk of the same bacterial content. 24. Use of the milk produced according to one of claims 1 to 17 for the production of ready-to-use packaged milk which can be transported and stored without cooling. 25. Use according to claim 24, characterized in that the milk is packaged using the form-fill-seal technique.