DK171112B1 - Process for the processing of raw milk for the preparation of treated milk with a lower bacterial content than raw milk and a method for producing milk with a low bacterial content and a fat content of approx. 2% - Google Patents

Description

DK 171112 Bl

The present invention relates to a process for producing treated milk, either whole milk or skimmed milk, with a lower bacterial content than raw milk, and a process for producing milk having a low bacterial content and a fat content of about 5%. 2%.

The well-known pasteurization process aimed at killing bacteria in milk has been used for many decades. Unfortunately, the higher temperatures required by the pasteurization process undesirably affect the taste of the milk.

Furthermore, even using such high temperatures, the pasteurization process does not eliminate all unwanted bacteria, leading to a short-term storage stability of most milk products.

Bacillus cereus is often the dominant bacterium in conventionally treated milk of relatively advanced age, as this bacterium can survive the pasteurization process and thrives at low temperatures, which promotes milk destruction. There is a general need for a method for reducing the bacterial content of 20 milk, both whole milk and skimmed milk, to increase the storage stability of the product and improve its taste, by eliminating the pasteurization process.

Also of great economic importance is the need to eliminate the extremely expensive and cumbersome distribution method currently needed to bring the milk to the consumer. It is currently necessary for any dairy, after treating the raw milk by homogenization or other steps, to fill the milk in containers for distribution to the consumer and to transport this milk 30 under cooled conditions. This requires that each dairy purchase and maintain a significant number of refrigerated transport vehicles for transporting the processed milk to the point where the distribution to the consumer takes place. By providing a sterile or almost sterile milk product, it will be possible to eliminate the need for transport of milk under such cooled conditions. Unfortunately, the pasteurization process only provides milk with a reduced bacterial count and not a sterile product.

Furthermore, if a sterile milk product could be produced, it would also be possible to eliminate the need for storage of the milk at the distribution site under cooled conditions. Eliminating the need for large cooling rooms, e.g. in the typical retail, would also be of enormous economic importance.

Even when using the pasteurization process used today, it is in some cases of particular importance to obtain milk with a reduced bacterial content prior to pasteurization. For example, a particular batch of raw milk may be so contaminated that pasteurization alone will not result in merely satisfactory storage life by today's standard.

Furthermore, for some applications it is of value to be able to provide treated milk in which the bacterial content has been greatly reduced, e.g. to approx. one hundredth of the original value. It is particularly important to provide milk with a relatively low bacterial content for the production of cheese, as incorrect bacterial cultures can destroy the cheese. Normally, it is not appropriate to simply heat-treat milk to a sufficient extent for use in cheese production, since such heat treatment can provide a lower yield of cheese and also undesirably affect the coagulation time. Conventionally, additives are used to reduce the problem. In many cases, however, it would be desirable to avoid the use of such additives.

Various methods for preparing milk with a reduced bacterial count using filtration are known in the art, but none of these methods have found widespread use. The methods of the prior art generally give either low flow rates, which makes the method 35 uneconomical on a large scale, or undesirably affects the quality of the milk, making the product unacceptable to the consumers.

Conventional filtration methods for producing milk with a reduced bacterial content have been tried. SE patent specification 380,422 discloses a process in which whole milk is divided into filtrate and concentrate fractions by microfiltration. The filtrate passing through the pores of the filter (the size of the pores can vary widely from 0.1 μα to 10 μτα) consists of milk of substantially reduced fat content, and the concentrate, which is the fraction retained by the surface of the filter, consists of cream since not only bacteria but also small globules of fat are essentially retained by the filter.

SE Publication No. 67.15081 discloses a method for sterilizing milk, in which fat is first separated from the skim milk. Next, the fat fraction is sterilized by heat and the skim milk fraction is sterilized by bacterial filtration (filter pore sizes are not indicated). Finally, the sterilized fat and skim milk fractions are re-mixed to form a sterile milk product. Thus, in order to sterilize the skim milk fraction by bacterial filtration, the pore size of the filter must be so small that no bacteria can pass through it, resulting in poor throughput rates and the undesirable retention of fatty globules and protein content of the milk.

U.S. Patent No. 5,064,574 discloses a process for the preparation of hypoallergenic milk by ultrafiltration methods employing membranes which allow molecules having a molecular weight of less than or equal to about 5,000. 5 kDa, to pass through them. The excluded constituents captured by the membrane include milk protein, viable or non-viable bacteria, bacterial protein antigen, and milk fat. The filtrate collected from the ultrafiltration process is therefore not only free of bacteria and bacterial protein antigen, but also of fat and milk protein, making the product as such unsuitable for use as milk.

Thus, it is clear that the pores in bacterial filters used in the art, which filters are effective in sterilizing milk, will also remove not only the bacteria but also the fat beads and at least some of the proteins. Such a filter is quickly blocked by trapped material, so the flow rate through the filter is rapidly diminished and the filter must be frequently cleaned or replaced. The high cost of such an ineffective approach is generally prohibitive. Furthermore, since the filter retains fat-beads and proteins, the quality of the milk is also adversely affected.

From the foregoing discussion, it is clear that there remains a need for an improved milk filtration treatment method capable of providing a sterile or near-sterile product having an improved storage life without adversely affecting milk quality.

Several different attempts have been made in the past to use cross-flow or tangential flow filtration means for treating milk, such means being known in the art.

Several types of filtering means have been described which enable such tangential or cross-flow filtration. The perhaps oldest known organ disclosed in USSR Patent No. 142,626 (Zhevnovatyi, AI, 1961) is formed of a tube of porous material attached within another tube, the suspension to be filtered being subjected to 30 loads. at high speed through the annular space between the two tubes as the filtrate flows within the porous tube. Enhanced members of similar construction comprise two concentric cylinders, the inner cylinder being formed of a microporous membrane, the liquid subjected to an forced helical flow around such an inner cylinder.

Other cross-flow means include a series of filtration elements superimposed in the form of plates or disks on the two surfaces of which microporous membranes are composed, e.g. about a filtrate-collecting tube, the suspension to be filtered being passed between the slices in a spiral-shaped path, one after the other.

Many other variations of the cross-flow filtration system have been developed. For example, US Patent No. 5,009,781 relates to a cross-flow filtration means having a filtrate network comprising a plurality of filtrate chambers along 10 and one or more filtrate channels transversely intersecting the chambers. U.S. Patent No. 5,035,799 discloses a cross-flow filter means having filter blade means arranged parallel within the filter tank with pressurized feed to produce turbulent cross-flow of liquid 15 over the media.

U.S. Patent No. 5,015,397 discloses a cross-flow filtration apparatus and method for filtration, which apparatus comprises a tube of helically wound wedge wire. Contaminated flowing material enters at one end and as it flows through the tube, it becomes more concentrated with contaminated material, while clarified liquid permeates through the tube wall. U.S. Patent No. 5,047,154 discloses a method and apparatus for increasing the flow rate in cross-flow filtration systems. U.S. Patent No. 4,569,759 discloses a tangential filtration apparatus and plant comprising such apparatus.

Cross-flow filtration is substantially different from flow-through filtration in that the liquid-fed material is introduced parallel to the filter surface and the filtration takes place in a direction perpendicular to the direction of the flow of the applied material. In cross-flow filtration systems - because the direction of the flow of feed material is tangential to the membrane surface - the accumulation of the filtered solids on the filtration medium is generally reduced by the shear effect of the flow. Thus, cross-flow filtration offers the possibility of a quasi-equilibrium operation with an almost constant flow when the driving pressure differential is kept constant. Unfortunately, this theoretical option has not been achieved in practice. Thus, the problem of diminished filtration flows has been a major disadvantage in conventional cross-flow filtration systems. Most of the suspended solids are retained on the wall of the tube and quickly form a dynamic membrane 10 (also referred to as a "filter cake" or a "mud layer"). The dynamic membrane is largely responsible for the subsequent filtration.

The particles that initially enter the wall matrix are eventually enclosed therein, due to the irregular and curved nature of the pore structure. As microfiltration progresses, the penetration of additional small particles into the wall matrix is inhibited by the presence of the dynamic membrane. The formation of the dynamic membrane, together with the possible clogging of the pore structure in the tube by enclosed particles, results in a reduction of the filtration flow.

In conventional systems, this decrease is approximately exponentially related to the filtering time.

Cross-flow filtration of milk has been attempted, but 25 have not been widely accepted due to the above mentioned problems. U.S. Patent No. 5,028,436 discloses a process for separating the dissolved and unresolved constituents of milk using a microporous membrane having a pore size in the range of 0.1 to 2 μη, which has been pretreated with a aqueous solution, dispersion or emulsion of lipids or peptides and the milk separated on the pretreated membrane. In the method of the patent, a first filtration step is used using a microporous membrane and a tangential flow. A clear filtrate and viscous concentrate are obtained. The filtrate contains all salts, lactic acid, amino acids, oligopeptides and low molecular weight polypeptides in genuine, non-denatured form, and the concentrate contains virtually all casein and fat components of the milk. Thus, the filtrate cannot be considered to be 5 "milk" since the fats have all been removed therefrom.

U.S. Patent No. 4,876,100 discloses a two stream filtration method for producing milk with a reduced bacterial content. Raw milk is divided by centrifugal separation into one fraction of cream and another 10 fraction of skimmed milk. The skimmed milk fraction is fed into a microfilter in which a portion of the fat-beads, protein and bacteria are separated. From the microfilter is obtained a filtrate consisting of skimmed milk with a reduced fat, protein and bacterial content as well as a concentrate with an increased fat, protein and bacteria content. The concentrate is then sterilized. The filtration method of the latter US patent, in addition to reducing the bacterial levels in the filtrate, also reduces the fat and protein content of the filtrate, changing its properties relative to the properties of the original skim milk.

Evidently, the use of cross-flow filters to date has not provided an acceptable method for reducing bacterial contamination in milk.

A method of overcoming some of the problems associated with the classical cross-flow filtration technology, known as dynamic microfiltration, has emerged. The dynamic filtration process overcomes the disadvantage of classical cross-flow technology, since the liquid to be filtered is not simply passed tangentially across the membrane surface. The diaphragm surface or a solid body adjacent to the diaphragm surface is moved so that the liquid at the interface between the rotor and the stator is subjected to a shear action. The shear stress tends to "clean" the membrane surface, keeping it relatively clean of particulate matter and preventing a filter cake from forming: on the membrane surface. The particulate material that would otherwise accumulate on the membrane surface remains suspended and is finally removed in the secondary stream, generally referred to as a concentrate stream.

5 Dynamic microfiltration systems may have different shapes. For example, U.S. Patent Nos. 5,037,562, 3,997,447, 5,037,562, 3,997,447, and 4,956,102 disclose dynamic microfiltration disk systems.

Cylindrical dynamic microfiltration means are disclosed in U.S. Patent Nos. 4,956,102, 4,900,440, 4,427,552, 4,093,552, 4,066,554, and 3,797,662, and in several other patents.

No one has ever used dynamic microfiltration 15 for the treatment of milk, and the use of cross-flow filtration of milk has been limited and mainly used for fractionation of milk into constituents based on fat content.

Surprisingly, it has now been found that dynamic microfiltration of milk can be carried out successfully without the problems known in the art of deteriorating milk quality, clogging of the filter prematurely and insufficient bacterial removal.

In accordance with the present invention, milk, either whole milk or skimmed milk, is first homogenized and then subjected to filtration. By first performing the homogenization step, the particle size of the fat gull balls and other larger, suspended constituents of the milk is significantly reduced, allowing microfiltration of the milk without significant removal or containment of the fat or other constituents.

Milk is an emulsion of fat and protein particles in water. Homogenization is a method of reducing the emulsion particle size to allow passage through a microporous membrane of appropriate pore size to retain bacteria contained therein without undesirable removal of the fat and protein content of the milk.

The milk is filtered after homogenization using dynamic microfiltration. Thus, with the present invention, there is provided an improved method for producing milk with a reduced bacterial content, without the need for pasteurization. The portion of the milk fraction retained by the microfilter (concentrate fraction) may be recycled as part of the feed material or may be discarded or used in other processes.

Thus, in a first aspect, the present invention relates to a process for treating raw milk for the production of treated milk having a lower bacterial content than the raw milk. The process is characterized in that (1) separates the milk into a fat fraction with a minimum fat content of approx. 10% and a skim milk fraction; (2) homogenizes the skim milk fraction and within ca. Five minutes from homogenization, the skim milk fraction undergoes dynamic microfiltration by passing the skim milk fraction through a microfilter with an average pore size sufficient to reduce the bacterial content of a filtrate having a lower bacterial content than the original skim milk fraction and a higher bacterial content. than the original skim milk fraction, (3) separately decreases the bacterial content of the fat fraction and (4) then combines the skim milk fraction after microfiltration and the fat fraction having the reduced bacterial content. The resulting milk has a very low bacterial content, e.g. ca. 103 bacteria per ml or less, and retains more organoleptic ingredients than are found in pasteurized milk with the same bacterial content.

The milk obtainable as a result of this process according to the invention is generally more storage stable than milk obtained as a result of conventional pasteurization. Significant residual bacteria remain in milk after pasteurization, since milk naturally contains certain bacteria that also survive the pasteurization process. Thus, pasteurized milk must still be refrigerated to reduce bacterial growth and destruction due to bacteria.

Unfortunately, some of the bacteria present in raw milk are both resistant to high temperatures (bacteria that survive pasteurization) and psychrotrophic (bacteria that thrive at low temperatures below 15 ° C), e.g. Bacillus cereus. The presence of bacteria that can survive high temperatures and psychrotrophic bacteria in the packaged milk product is extremely destructive, since their rapid growth, even under cooling conditions, results in the destruction of the milk.

With the present invention, 15 sterile milk can be prepared which can be stored even at room temperature for extended periods, e.g. for 30 days or more. The sterile milk can be characterized by the absence of bacteria in general and in particular by the absence of bacteria and pathogens such as the following:

Bacteria that can survive high temperatures

Micrococcus M. luteus, M. roseus

Streptococcus S. pneumoniae, S. laetis, 25 S. faecalis

Lactobaccilus L. delbrueekii, L. lactis, L. helveticus, L. casei, L. trichodes 30

Staphylococcus S. aureus, S. epidermidis

Bacillus B. cereus, B. subtilis, B. macerans, 35 B. stearothermophilus

Clostridium C. butyrium, C. pasteurianum C. botulinum, C. perfringens, C. tetani 40

Psychrotrophic bacteria

Pseudomonas P. aeruginosa, P. fluorescens, 45 P. pseudomallei, P. mallei 5 11 DK 171112 B1

Archnomobacter

Alcaligenes

Acientobacter A. lignieressii, A. equirli

Flavobacterium F. aquatile, F. menigosepticum 10 Bacillus B. cereus, B. subtilis, B. macerans, B. stearothermophilus Coliform bacteria

Enterobacter E. coli, Salmonella Typhi

Shigella Dysenteriae,

Klebsial Pneumoniae 20

Various

Listeria L. monocytogenes 25 The milk will thus be able to meet and typically exceed the requirements for quality A pasteurized milk, which requires that the milk does not exceed a bacterial plate number of 30,000 per day. and a coliform number exceeding 10 per ml. ml, as determined by standard methods.

In another aspect, the present invention relates to a process for the preparation of milk having a fat content of approx. 2%, which is characterized in that (1) homogenizes a skim milk fraction and within approx. Five minutes from the homogenization (2), the skim milk fraction undergoes dynamic microfiltration by passing the skim milk fraction through a microfilter having an average pore size sufficient to reduce the bacterial content of the milk flowing through the filter to obtain a filtrate having a lower bacterial content. the original skim milk fraction, and a concentrate having a higher bacterial content than the original skim-45 milk milk fraction, (3) decreases the bacterial content of a 12 DK 171112 B1 cream fraction with a minimum fat content of approx. 10% and (4) then combine the skim milk fraction after microfiltration and the cream fraction having the reduced bacterial content.

5 The invention is further illustrated in the drawing, in which FIG. 1 is a schematic diagram of the equipment used in the practice of the methods of the invention; and FIG. 2 shows the particle sizes in milk after homogenization.

The original material is fresh, raw raw milk from livestock, e.g. a cow. The methods of the invention can also be used in conjunction with treated milk, e.g. milk that has already been subjected to pasteurization, but the full benefits will not be obtained, e.g. the preparation of milk with improved organoleptic properties, in comparison with milk that has not been pasteurized.

The raw milk to be processed can first be passed through a heat exchanger to an appropriate temperature setting, whereupon the milk is passed through a centrifugal separator if desired to remove all or part of the cream fraction in a conventional manner.

In summary, the raw milk is homogenized and passed reasonably quickly through a dynamic microfilter to give a filtrate fraction and a concentrate fraction. The pores in the microfilter are of such size that they retain at least part of the bacteria. The filtrate, which is the part of the milk fraction passing through the retaining surface of the microfilter, consists of milk with no or a reduced bacterial content (relative to the milk before microfiltration no one), with essentially no change in the fat and protein content. The filtrate fraction can then be used directly to prepare other products, e.g. powdered milk, or packaged without further processing.

13 DK 171112 B1

The filtrate fraction is, for many reasons, more desirable than the milk obtained by conventional pasteurization. It contains more organoleptic ingredients than milk that has been pasteurized, which makes it more flavorful and desirable from the consumer's point of view. Furthermore, milk obtained in accordance with the present invention has a much longer storage life because bacteria, e.g. the psychrophilic bacteria, especially Bacillus cereus, can be completely removed by practicing the method of the present invention, which is an impossibility using conventional pasteurization.

The concentrate fraction, which is the part of the milk fraction retained by and recovered from the retaining membrane surface of the microfilter, consists of milk having an increased bacterial content (relative to the milk fed prior to the microfiltration) and substantially no change in the content. of fat-beads and protein. The concentrate fraction can subsequently be discarded or used in other processes.

The filtrate may contain some bacteria, but the lower the bacterial content, the more storage stable the product will be. Full sterilization is desirable, but the initial growth rate of a small, residual bacterial concentration is usually low enough to still result in a greatly increased storage life of the resulting milk product.

The storage life of milk produced by the practice of the methods of the invention is substantially increased over the storage life of pasteurized milk in conventional manner, as the concentration of especially Bacillus cereus bacteria is greatly reduced.

Since the milk produced by the methods of the present invention can be rendered sterile, while 35 milk obtained using conventional pasteurization methods cannot be truly sterile, the milk may have extremely long storage life under cooling conditions or room temperature conditions, especially if applied in a container under aseptic conditions. One method of carrying out this is using the mold-filling closure method, which is now well known in the packaging industry. This technique is often used for packing sterile solutions and the like, e.g. in the pharmaceutical industry. The milk produced by the methods of the invention can be packaged using the mold filling closure method and such milk can exhibit extremely long storage life, even at room temperature.

The exact method or apparatus used to perform the filling is not critical. As just one example and explanation of how such a mold filling closure method can be used, the following description is given.

In some vertical mold filling closure machines, a flat web of synthetic thermoplastic film is used which is unwound from a roller and formed into a continuous tube in a tube forming section by interlocking the longitudinal edges of the film. In other machines, the tube is extruded from a resin melt at the time of use. The tube thus formed is conveyed to a filling station where it is collapsed across a transverse cross-section, the position of the cross-section being at a closing device below the filling station. A transverse heat seal is produced by the closure device at the collapsed portion of the tube, thereby producing an airtight seal across the tube. After producing the transverse closure, a portion of the material to be packed is brought, e.g. liquid to enter the tube at the filling station and the tube is filled upwardly from the aforementioned transverse closure. The tube is then moved downward to a predetermined distance and closed and joined in the transverse direction by the second transverse section.

One such vertical mold-filling machine of DK 171112 B1 of the type described above is sold under the trademark PRE-PAC, and another is disclosed in U.S. Patent No. 5,038,550.

5 HOMOGENIZATION

Preferably, the milk fraction is first heated or cooled after centrifugal separation, if used, and prior to homogenization to a suitable temperature for homogenization. The milk is then passed to a homogenizer, where the fat emulsion size is reduced to a size sufficient to allow passage through the membrane. Preferably, the size of all suspended particles is less than ca. 1 μη. It is important that after homogenization the milk is filtered relatively quickly. Preferably, filtration will take place in less than ca. 5 minutes, preferably less than about 5 minutes. 2 min. and especially in less than approx. 30 sec.

Again, the important factor is not the holding time prior to filtration, but the fact that the filtration 20 takes place prior to any substantial agglomeration of small spheres forming a substantial number of particles having a size greater than ca. 1 μη.

Homogenization of skimmed milk or whole milk prior to filtration in a cylindrical dynamic microfiltration unit is absolutely necessary to properly emulsify and suspend the fat and other milk components and reduce the size sufficiently to achieve proper filtration. However, a rotating disk filter develops a significant degree of shear immediately at the surface of the rotating disk. Thus, some degree of homogenization of the milk can take place almost immediately simultaneously with the filtration. Such "self" emulsification of the milk by the action of the dynamic microfilter allows skimmed milk to be treated with a rotary disc filter without the need for a separate homogenizer.

In reality, the rotating disk functions so that the milk is both homogenized and filtered simultaneously, which does not take place in a rotating cylindrical filter unit. A rotating disk filter can produce a shear rate of about 5 cm. 200,000 sec. "1, while a rotating, cylindrical assembly can produce a shear rate of only 10,000 sec. -1. Although the shear force is substantial in a rotary disc filter unit, it is not believed that in most cases it will be sufficient to be required. extent to homogenize 10 whole milk.

DYNAMIC FILTERING

According to the present invention, the filtration is carried out as a dynamic filtration, that is, the filtration medium itself is kept in constant motion so that the effective flow rate of milk across the medium is extremely high. The particular physical shape of the dynamic membrane is not critical. Thus, the membrane medium may take the form of e.g. washers or cylinders. Several different dynamic microfiltration means have been previously described and are useful in the practice of the present invention. In general, the dynamic microfilter comprises a cylindrical or disc membrane element rotating in the interior of an outer, impervious cylinder.

25 In a cylindrical dynamic microfilter, in which the liquid to be filtered is introduced into the space between the stator and the rotating membrane, torque is transmitted from the rotating membrane to the liquid. The liquid near the inner cylinder is subjected to a higher centrifugal force than the liquid close to 30 at the outer cylinder. This phenomenon produces, under certain conditions, a flow pattern known as Taylor swirls, which prevents the development of substantial residue on the membrane surface.

Thus, the dynamic filtration process benefits from the generation of the Taylor vortices to keep the membrane surface free of potential residue, whose constituents thus remain suspended in the liquid being filtered. In the process, the applied material is then cleaved in a filtrate (the part of the liquid which penetrates the membrane) and a concentrate (the fraction containing the suspended particles which would normally be deposited on the surface of the membrane and clog it). . In this way, a high flow rate can be maintained through the membrane for a long period of time. The amount of material and concentrate added must be controlled in a manner that results in a stable liquid flow. Even with low concentrate fluid velocities, it is possible to maintain a steady flow of fluid to the membrane surface.

The dynamic microfiltration allows a wide range of effective surface velocities of the filtration medium in relation to the milk product supplied. For example, an effective surface velocity of from approx. 3 m / sec. to approx. 50 m / sec, especially from approx. 5 m / sec. to approx. 30 m / sec, in particular from approx. 8 m / sec to approx. 20 m / sec.

In order to obtain the desired surface velocities, it will be necessary that a representative filtration medium in the form of a cylinder having a diameter of approx. 64 mm is rotated at a speed of approx. 1,000 to approx. 6,000 rpm at a rate of approx. 5,000 rpm The minute is typical.

25 If a dynamic disk filtration device is used, a typical disk filtration medium will have dimensions of from approx. 51 mm to approx. 1,220 mm in diameter. Such slices can e.g. rotated at speeds from approx. 1,000 to approx.

8,000 rpm per minute, typically from approx. 3,000 to approx.

30 6,000 rpm. per minute, depending on the design of the particular dynamic microfilter used. Preferably, the shear rates for such disc filters will be from about 1 100,000 sec ”1 to approx. 400,000 sec. "1. Among the preferred disk filters are the filters of the type described in WO Publication No. 93/12859 with priority from US Patent Application 07 / 812,123 filed December 18, 171112 B1 December 24 , 1991.

The microfilter pores are of such size that they retain the bacteria present in the milk while still maintaining an acceptable flow rate 5 through the microfilter. Suitable membranes include hydrophilic microporous membranes with good flow properties, narrow pore size distribution and sustained bacterial removal ability of the bacteria of interest. The pore size of the microfilter membrane should be from approx. 0.01 to approx. 5.0 10 μιη as determined by the methods known in the art, the tests used being known as "bubble point" (ASTH F316-86) and the KL method, cf. U.S. Patent No. 4,340,479. Preferably, the pore size will be from about 1 0.1 to approx. 1 μιη. In particular, filters having pore sizes 15 in the range of from ca. 0.2 to approx. 0.5 μιη. Such microporous filters are well known and readily available.

Preferred microporous membranes which can be used in the practice of the present invention include those sold by Pali Corporation under the trademarks Ultipor 20 Ngg®, Fluorodyne® and Posidyne®, the membranes available from Cuno Corporation under the trademark Zetapor, and the membranes sold by Millipore under the trademark Dura-pore®.

The cylindrical membrane elements used in the practice of the present invention comprise those elements which can be connected to a support in a leak-proof manner according to methods known in the art.

Finally, the bacteria should be concentrated in a stream that amounts to less than ca. 5% of the feed material and more than approx. 95% of the solids and proteins normally found in the milk should pass through the membrane for extended periods.

The dynamic microfilter can be used with a single pass without the need for recirculation of the concentrate. If desired, the concentrate can be recycled to the supplied DK 171112 B1 material. When a cylindrical dynamic microfilter is used, it can be used for different ratios of filtrate flow to total flow of feed (concentration factors). However, the cylindrical dynamic microfilter 5 is advantageously used for filtrate to feed ratio greater than 90%, in particular about 95% and especially above 98%, to obtain predominantly the very low bacterial filtrate as the desired product.

When a rotary disk dynamic microfilter is used, this can also be used for different ratios of filtrate flow to the flow of total feed material. However, the rotary disc dynamic microfilter can be used for filtrate-to-feed material extends over a wide range. Evolution of a high ratio will simply decrease the flow rate, while operating at a low ratio will result in a higher flow rate. It is assumed that operating at a ratio of approx. 40% is advantageous for maintaining a steady flow rate through the filter, although other conditions can be used.

The filtration of the freshly homogenized milk can be carried out hot at 40 to 60 ° C, which is at or slightly above the crystallization temperature of approx. 40 ° C for the higher-melting components of milk fat. This is below the temperatures used in conventional heat pasteurization. Alternatively, the milk with some reduction in the flow rate can be filtered at much lower temperatures, e.g. from approx. 15 to approx. 35 ° C, especially from approx. 20 to approx. 25 ° C. 1 2 3 4 5 6

IN GENERAL

2

After microfiltration, the concentrate can be discarded in any acceptable manner, subjected to further processing 4 or used directly.

5

The process of the invention can be used with advantage when the desired end product is either whole milk, standardized milk or skimmed milk.

20 DK 171112 B1

The flow rates through a bacteria-retaining membrane for milk with a reduced fat content are usually higher than the flow rates for milk with a high fat content. In certain situations, it is economically more advantageous to produce milk with a higher fat content, e.g. 2% milk fat, by combining a filtered skim milk with a filtered fat fraction. This fat fraction can be a cream fraction with a minimum fat content of about 10%.

Filtration of the cream fraction may e.g. may be performed by the method described in U.S. Patent No. 5,395,531, or by using a bacteria-retained closed-end filtration cartridge. The filtration can be carried out in an industrially acceptable manner by heating the fat composition to a point where it is in a liquid state and is easily filtered through a microporous membrane. The preheated fat can be homogenized before filtration. Alternatively, the fat material may be subjected to pasteurization to reduce its bacterial content, or a combination of pasteurization of microfiltration may be used.

If the purpose of the process is to obtain protein concentrates, e.g. from the milk of a transgenic animal, e.g. a transgenic cow, the dynamic microfiltration 25 is used to obtain a high concentration of concentrate, using a microporous membrane having a pore size of approx. 0.2 μιη or less.

Suitable apparatus for carrying out the method according to the invention can be constructed by interconnecting 30 conventional equipment, including centrifugal separators, micro-filters, sterilization units, heat exchangers and pumps. Those skilled in the art will readily be able to provide flow and pressure control valves and other necessary equipment parts that serve to make such apparatus 35 operable, and those skilled in the art will also be able to make further conventional modifications to such apparatus. 21 appliances, as required in each case.

The following examples illustrate particular embodiments of the invention.

5 GENERAL PRACTICE

In the examples, the general procedure used is as follows:

Method A: Temperature recording of milk 10 Unless otherwise stated, the milk used in the following examples is commercial milk obtained from a retail outlet. The temperature of the milk is adjusted to a suitable process temperature prior to filtration. The preferred operating temperature (40-60 ° C) is used since the majority of the 15 fats in the milk are not in a crystallized form at such temperatures. A 35 liter, sheathed fermentation vessel (type 3000 from Chemap A.G.) is used as a processing vessel. The container is filled with milk and the contents are heated to approx. 50 ° C unless another 20 is indicated via a hot water jacket. The milk is stirred during the heating process to increase the heat transfer.

When the milk has reached the desired process temperature, it is pumped to a homogenizer at a rate of approx. 1 1 / min.

25

Method B: Milk homogenization

Upon entry into the homogenizer (model 15 MRI from APV Gaulin, Inc.), the milk undergoes a two-step homogenization process, the first of which occurs at ca. 17,237 30 k * Pa, and the other takes place at approx. 3,448 k * Pa. Normal startup and operation procedures are followed as described in the APV Gaulin Operations Manual for this unit. After homogenization, the milk is typically transferred to an intermediate stir fry equipped with a jacket and kept at the desired process temperature. This tank acts as a liquid buffer between the homogenizer outlet and the supply to the filter. Whenever desired, the homogenized milk can be recycled through the homogenizer to maintain a constant volume in the fryer.

Method C: Introducing bacteria into the milk feed stream In some experiments, artificial grafting of the milk stream with bacteria is used to detect the very high titer reduction possible using the present invention. Bacterial graft material is added to the feed stream via a metering pump between the treatment vessel and the homogenizer. The flow rate of the graft material is maintained so that a desired concentration level of bacteria of approx. 106 pr. milliliters of milk. Since the bacteria are introduced in front of the homogenizer, the bacteria are thoroughly mixed in the process fluid prior to entry into the filtering means. In most of the examples, the E. coli strain ATCC 15224 is used.

An alternative method of grafting the milk with bacteria would be to add the bacteria directly to the process container at the desired concentration. Such a method is not preferred because the bacteria are thereby exposed to long residence times at temperatures higher than that of the surroundings. This could cause unwanted growth or excessive eradication of the bacteria before entering the filtration means, depending on the strain used.

Method D: Bacterial assay test

Mesophiles: Bacterial concentration is determined by serial dilution of the samples and passage of appropriate dilutions 30 through sterile 0.2 µm membranes and culture on Mueller-Hin-ton-Agar for 24 hours at 32 ° C. These procedures are described in more detail in a publication entitled "Manual of Clinical Microbiology, 2nd Edition, 1974, ASM, Washington, D.C.".

Listeria monocytogenes ATCC 43256 is the pathogen being tested. The population levels in the samples are determined by the method used by Agello et. eel. (G. Agel lo, P.

23 DK 171112 B1

Hayes and J. Feeley, Abstracts of the Annual Meeting, 1986, ASM, Washington DC, page 5).

Method Purification Procedure 5 Sanitization and sterilization are performed prior to each experiment using 0.1 N sodium hydroxide. In the sterilization process, the membrane and all associated equipment are first rinsed with water and then treated with 0.1 N sodium hydroxide at 50 ° C for approx. half an hour. The basic solution is then neutralized using phosphoric acid. This neutralized solution is then used to rinse the system until all portions are neutral. Filtering experiments are performed immediately after this procedure. The assembled equipment and membrane elements are sanitized using the sterilization procedure at the end of each experiment.

Method F: Intearity testing

Each membrane element is tested for integrity prior to the bacterial influence. An early flow test as described in publication NM 900a, "The Pall Ultipor membrane filter guide", copyright 1980, available from Pali Corporation, is used for the integrity test.

25 Description of filtration apparatus 1. The cylindrical dynamic microfilter.

The cylindrical dynamic microfilter (cylindrical DMF) used for these experiments is a BDF-01 manufactured by Sulzer Brothers Limited, Winterthur, Switzerland. The apparatus is described by Rebsamen et. eel. (Dynamic Micro-filtration and Ultrafiltration in Biotechnology, E. Rebsamen and H. Zeigler, Proceedings of the World Filtration Congress IV, 1986 (Ostende, Belgium)). See. also U.S. Patent Nos. 35,066,554 and 4,093,552.

DK 171112 Bl 24 2. Description of membrane filter elements.

The membrane filter elements used in these experiments are typically various grades of nylon membrane, Ultipor Ng6® and Posidyne®, commercially available from Pali Corporation, Glen Cove, NY. The pore sizes used are 0.2, 0.30, 0.45 and 0.65 μια, the membrane elements having a surface area of 0.04 m2.

3. The dynamic microfilter in disk design: 10 The disk design consists of a membrane support disk with a diameter of approx. 15.2 cm mounted on a hollow shaft and contained in a leak-proof housing, with the required connections for fluid inlet and outlet. The support disc has means for sealing membrane sheets to its surface 15 in a leak-proof manner and contains of drainage space for passing the filtrate flow through the membrane and the disc and out through the shaft. The effective membrane area is 0.014 m2 and rotational speeds of up to 4,500 rpm can be used.

Any of the dynamic disk microfiltration units discussed earlier can be used in the practice of the invention. Reference may also be made to WO Publication No. 93/12859 with priority from U.S. Patent Application No. 07 / 812,123 filed on March 24.

December 25, 1991, with respect to the description of another dynamic microfiltration means of disc design that can be used in the practice of the invention.

4. Description of membrane filter elements.

The membrane filter elements are the same qualities of membranes as described in the section on the cylindrical DMF. Typically, the membranes are circular planar sheets ("donuts") which are cut to fit slice DMF. When assembled in the dynamic microfilter, the filtrate chamber is sealed off from the applied stream using O-rings. The membrane filter elements have a surface area of 0.014 m2.

DK 171112 B1

Method GI; Operation of the cylindrical dynamic microfilter

Prior to filtration, a filter element is assembled as described in the section of filter assemblies in the Cylindrical Dynamic Microfilter (DMF). Sanitary treatment and sterilization are performed using the procedure described in connection with Method E. After observing the start-up procedures specified in connection with Method 62, the milk to be filtered is pumped from the ladder tank into the cylindrical DMF via a positive displacement pump. The amount of concentrate is controlled by another pump or pressure relief valve connected at the concentrate opening. Temperatures and flow rates of the feed, filtrate and concentrate as well as the pressure of the feed are measured at various times during the course of the experiment, typically at 10 minute intervals. Standard operating conditions for the cylindrical DMF are a rotational speed of 5,000 rpm. a ratio of filtrate to feed material greater than 95% and a feed pressure pressure of from approx. 1.3-2.0 bar. All 20 examples of this apparatus are performed at constant flow rates for feed material.

Method G2: Starting the dynamic filter

Before introducing the milk into the dynamic filter 25, hot, ionized 0.2 μπι-f filtered water is passed through the system to start up the associated equipment. The rotational speed of the dynamic filters is brought to operating speed with water flowing through the system. When the system has reached equilibrium, the milk flow is initiated.

The milk displaces the water in the system and filtration begins.

Method H: Operation of the Dynamic Disc Filter with Disc Design 35 A disc DMF filter element described in the section under filter compositions is assembled into disc DMF. Sanitation and sterilization is carried out using the procedure specified in connection with method E. After observing the start-up procedures specified in connection with method G2, the milk to be filtered is pumped 5 the ladder tank into slice DMF. The amount of concentrate and pressure of feed material is controlled by means of a valve located at the entrance of the concentrate. Temperatures and flow rates of the feed material, filtrate and concentrate as well as pressure of feed material are measured at different times during the course of the experiment, typically at 10 minute intervals. A feed rate of approx. 960 ml / min. is maintained in all examples. The filtrate flows reported are those obtained when the flow is stabilized in the filtration unit.

15

EKSEMPT.TO

EXAMPLE 1

At room temperature, skimmed milk is pumped at a rate of approx. 600 ml / min. into a cylindrical DMF equipped with a 0.45 / m Ultipor Ng6® membrane. The operating conditions of the DMF are maintained as indicated for Method GI and are summarized in Table I. The pressure of the applied material begins to increase rapidly a few minutes after the commencement of the experiment, indicating clogging of the microporous membrane.

EXAMPLE 2

Skimmed milk is heated to 50 ° C according to method A and homogenized by method B. The homogenized milk is then stored in the fryer for approx. 4 hours while keeping the temperature of the milk at approx. 50 ° C during this time. After this four hour delay period, the milk is pumped into a cylindrical DMF equipped with a 0.45 μιη-Ultipor Ng6® membrane at a feed rate of approx. 600 ml / min. The preferred conditions for DMF operation as indicated in connection with method 27 DK 171112 B1 GI are maintained. The pressure of the applied material begins to rise rapidly after only a few minutes of operation, indicating clogging of the microporous membrane and the test must be terminated.

EXAMPLE 3

Skimmed milk heated to 50 ° C in accordance with Method A and homogenized by Method B is pumped into a cylindrical DMF equipped with a 0.45 μπι-Ul-10 tipor Ng6® membrane within no more than 5 minutes from homogenization. The preferred conditions for DMF operation as indicated in connection with Method Gi are maintained. A steady filtrate flow of 1,080 l / hr / m 2 is achieved until the milk supply runs out. No increase in pressure of the applied material is observed during the course of the experiment.

After all the milk has been processed, the feed material is switched to homogenized skim milk at 50 ° C without disturbing the operation of the system. In a few minutes, the milk filtrate flow decreases rapidly and the system pressure increases, indicating that the membrane clogging has occurred. This example clearly demonstrates the need to homogenize milk to achieve significant flow through a microfiltration membrane. 1 2 3 4 5 6 7 8 9 10 11

Examples 1-3 show that it is necessary to give 2 sufficient shear on milk (via homogenization 3 in this case) prior to filtration to reduce 4 the emulsion particle size of the milk to a sufficient degree to allow passage through the microporous 6 membrane and thus obtaining proper filtration. Example 7 2 shows in particular that the particle size distribution 8 returns to larger sizes within a short space of 9 after homogenization. Thus, to obtain the desired filtration 10, homogenization must take place within 11 a short time prior to filtration, e.g. within less than 5 minutes or preferably a shorter period of time.

EXAMPLE 4

Skim milk is preheated by Method A and pumped into a disc DMF equipped with a 0.45 μπι-Ultipor N66® membrane. The procedures set forth in connection with Method H. are used. A steady stream of filtrate is rapidly established and maintained for approx. 100 minutes until the milk supply runs out.

The operating conditions of the disc DMF used produce a calculated shear rate of approx. 200,000 sec. ”10 at the interface between the rotating disc and the diaphragm. This displacement is in the range of shear rates produced by the homogenizer under the conditions specified in connection with Method B.

This example shows that the required shear stress prior to filtration can be achieved in one step, i.e. without the need for separate homogenization equipment.

The example clearly shows that the membrane is not contaminated by the solids in milk and that the shear effect produced by the rotation of the disc is approx. 200,000 sec. ”1, 20 is sufficient to reduce the particle size in skimmed milk to allow passage through a microfilter membrane and thus obtain proper filtration.

Table I summarizes the results of Examples 1-4. The indicated data show that an equilibrium filtrate flow is obtained through the membrane when the milk is sufficiently sheared within a short period prior to filtration. 1 35 DK 171112 B1 29

Il o o u Ώ o o oo o +3 o vo

tt ri HH

ii in in in tn • tf Tf * 3 · 'd · ^ ^ ^ o o o o | in! 'S S · ί ΙΛ jta H | li? £ A $! Ω o o o o -P o o o o

HH O O O O

• Hl · · · ·

<H> {O O O O

Tu Η Η H O

UrH (9 cm

IN"

and I

• h oj u in o o o

And -P o <n m in in ΐ 'li 11 I i

rh

^^ CM Η -M · 30 DK 171112 B1 EXAMPLE 5

To determine the ratio of particle size to time after homogenization, skim milk is heated by Method A and homogenized using the procedure described in connection with Method B. The particle size distribution with respect to the time after homogenization is determined. The particle size distribution is measured using an integrated micro-optical fluid-volumetric sensor (IMOLV-2) available from Particle Measurement Systems, Colorado. This laser particle counter is designed to measure particle size distribution in the range of approx. 0.5 to approx. 5.0 μιη.

The milk samples are diluted in a ratio of 1: 300,000 and then subjected to analysis as specified in the operating manual of the IMOLV apparatus. 0.04 μιη-Filtered, 18 mega-ohm DI water with particle numbers less than 50 per. milliliters is used to dilute the milk samples.

FIG. 2 of the drawing shows the results of the particle analysis. A plot of the number of particles relative to the number of particles at 5 seconds relative to particle size is shown in the figure. The figure clearly shows that as the time after homogenization increases, the number of larger particles increases. As the number of smaller particles similarly decreases over this time, it is evident that the smaller particles agglomerate over time to form the larger particles.

EXAMPLES 6-9

Membranes with different pore sizes and bacterial retention properties are tested on the cylindrical DMF to determine the amount of stable milk filtrate flow that is obtainable. The general method used in Examples 6-9 is given below.

1. The desired membrane filter element is assembled into the cylindrical DMF.

2. An integrity test is performed as specified in Method F. Connection to Method F. The membrane filter element is discarded if the sample fails.

3. The equipment is sanitized according to Method E.

4. Milk to be filtered is preheated by the procedure specified for Method A.

5. Milk is homogenized according to Method B.

6. The startup procedure specified in connection with method 62 is performed.

7. The milk is transferred from the risers to the cylindrical DMF at the desired flow rate.

8. The operating parameters are determined using the guidelines specified for method GI.

9. Appropriate measurements are made.

Typically, the cylindrical DMF is used at 5,000 rpm. per minute corresponding to a shear rate of approx. 10,000 sec. "1 in the filter. The supply temperature is 50 ° C and the supply pressure varies from 1.3 to 2.0 bar. The ratio of filtrate to feed is maintained at over 95% in each of these examples. The flow reported in Table II, the equilibrium filtrate flow obtained is typically fifteen minutes after the start of filtration, the total time of the experiment varies in each case since the volume of milk being filtered is constantly 30 liters.

EXAMPLE 6

In this example, a 0.2 μιιι-Ultipor Ngg® membrane is used. A feed rate of 250 ml / min is used. to obtain an equilibrium filtrate flow of 330 l / h / m2. Filtration is continued for approx. 130 minutes, at which time there is no more milk in the processing vessel, without apparent decrease in filtrate flow rates.

EXAMPLE 7

In this example, a 0.3 μια-Ultipor Ngg® membrane is used. A feed rate of approx. 550 ml / min to achieve an equilibrium flow of 775 l / h / m2 for approx. 60 min, after which the trial is terminated.

EXAMPLE 8

In this example, a 0.45 μπι-Ultipor N66® “meinl ::> ran” is used. A feed rate of 740 ml / min is used. to achieve an equilibrium flow of 1,080 l / hour / m2. Filtration is continued for approx. 40 minutes without apparent reduction in flow rates, after which the milk supply is depleted and the experiment is terminated.

EXAMPLE 9

In this example, a 0.65 μιη-Ultipor Ngg® membrane is used. A feed rate of 1,100 ml / min is used. to achieve an equilibrium flow of 1,680 l / hour / m2. Filtration is continued for approx. 30 minutes, after which the milk supply is depleted and the test is completed.

The results of Examples 6-9 are summarized in Table II. The indicated data show that, using the filtration process of the invention, stable filters can be obtained to flow using various grades of bacterial retaining membranes. The table shows that in the method of the present invention, membranes with smaller pores and thus increased bacterial retention can be used, at the expense of filtrate flow rates.

DK 171112 B1 33 ίε o lo o o • h m oo and 4-1 co r- o »-Η Η« Η S I -7 • S y xJ § o o o o pi H ^^ n

B

P 43 ώ i3 p 'r> t "t» S ή' Q Ή σ \ σι σι σ \ ^ b IH iu o o o o o

Yes

§ Ih ^ jjj o vo in r> • H Lj Jj * »*. *. *»

CU Η -P s— 'O) Η Η H

• S S & U O O O O

• h 5} ° in in in in ίπ o o o o

• H O O O O

S o o o o O in in in in? fl s s s s ^ S s i .o o o o

V ®10 ®10 ®iD ®iO

g £ X K K

I i i g, a a S .3 * .3 * · ίτ -ίΤ ·· g 43 43 -p -p η i s d s s to ci vo r * co cn 34 DK 171112 B1 EXAMPLE 10

In this example, a 0.2 μιη-Posidyne® membrane with a positive surface charge is used. The membrane used has its pore surfaces populated with quaternary ammonium groups and has a high absorption capacity for biological material.

A feed rate of 260 ml / min is used. to achieve an equilibrium flow of 360 l / hr / m2. The filtrate stream is of the same order of magnitude as obtained with unloaded membrane as described in Example 6. The filtration is continued for approx. 120 minutes, at which time there is no more milk in the processing vessel, without apparent reduction in filtrate flow rates. A filtrate-to-feed ratio greater than 97% is maintained throughout the experiment. Other experimental conditions are given in the following Table III.

It is expected that a large amount of milk proteins will bind to the membrane surface and eventually clog it. This example shows that under dynamic conditions, a membrane that usually exhibits protein affinity functions satisfactorily.

EXAMPLE 11

A feed rate of 740 ml / min is used. for whole milk, and a stable filtrate flow of 1,130 l / h / m2 is obtained. Other experimental conditions are given in the following Table III. Filtration is continued for approx. 40 minutes, after which the milk supply runs out and the experiment ends.

This example shows that whole milk can be filtered using the method of the invention. The observed difference in filtrate flows between whole milk and skimmed milk (as in Example 9) appears to be primarily due to the differences in their viscosities. The ratio of the filtrate flow obtained for whole milk to that obtained for skimmed milk is approximately equal to the ratio of the viscosity of whole milk to the viscosity of skimmed milk.

DK 171112 B1 35 'i g I o o S -H vo r>

C 4J η H

ft r-H * H

fi 'T o o

Igi a * 4J Z) nj iH Γ "Γ0 JH3 ° i ° i

HBy O O

pt «<w <w i 5, <* o« * 1 f 1 ~ h 'k ø i.

in §f £ s s c o o

• H O O

g O O

0 ιή in 6 p $ o m 8 β i o o ®io @ z * 1 ® 5 I ss 1 *

In s a 36 DK 171112 B1 EXAMPLES 12-16

Examples for the determination of filtrate flows through various bacteria in retaining membranes are repeated using the dynamic disk microfilter. The general procedures used in Examples 12-16 are described below. The conditions described are generally used in each example, unless explicitly stated otherwise.

1. The desired membrane filter element is assembled in disk- -DMF.

2. An integrity test is performed as specified in Method F. The membrane filter element is discarded if it fails the sample.

3. The equipment is sanitized according to Method E.

4. Milk to be filtered is preheated by the procedure specified for Method A.

5. Milk is homogenized according to Method B.

6. The general start-up procedure specified in connection with method 62 is carried out.

7. The milk is transferred from the frying tank to the slice DMF at a desired flow rate.

8. Appropriate measurements are made.

Typically, disk DMF is maintained at 3,500 rpm, corresponding to a calculated shear rate of approx. 200,000 sec-1. The supply temperature is 50 ° C and the supply pressure is maintained at approx. 0.2 bar. Milk is pumped into the filter at a rate of 960 ml / min. for maintaining a high cross flow rate across the membrane. The ratio of filtrate to feed material is specially adjusted taking into account the membrane pore size, feed temperature and rotor number of revolutions / minute. The unfiltered portion of the feed material is recycled into the treatment container. The flow shown in the following table is the equilibrium flow obtained through the membrane as a filtrate, typically half an hour after the start of the filtration.

EXAMPLE 12 In this example, a 0.2 µm Ultipor N55® membrane is used. An equilibrium filtrate flow of 850 l / h / m2 is obtained.

EXAMPLE 13 In this example, a 0.45 µm Ultipor N55® membrane is used. An equilibrium flow of 1,600 l / hr / m2 is obtained.

EXAMPLE 14 In this example, a 0.45 μιη-Posidyne® membrane is used. An equilibrium flow of 1,600 l / h / m2 is achieved.

The data shown in the following Table IV summarizes Examples 11-13. These data show that using the filtration process of the present invention, stable filtrate flows can be obtained using various grades of bacterial retaining membranes using a slice DMF. The table shows that membranes with smaller pores and thus increased bacterial retention (titer reduction) can be used in connection with the present invention at the expense of filtrate flow rates.

DK 171112 B1 38 'h 5 § o o o H -H in o o η 4-> ω ιο 10 • H ^ · ·

& - (r ~~ t t-l iH

6 TJ · § t "o o S" Π o 3 æ 00

+ J * Ό S. Q

• P Φ .3 ni t · »ηόη µ η n

fn IH <W O cT cT

in

Η fN CM CM

E- «-P - O o O * H Jjo ooog ih w in in C ooo -dooog in in · · · 0 η n fo (S) ω otn tn £ H oj ·« u φ 6 ^ ^ ^ ft δ 3. ooo © © 10 10

KO KO

»S © i g i 1 § § § 1¾ a s a 39 DK 171112 B1 EXAMPLE 15

Skimmed milk at a temperature of 18 ° C and homogenized by method B is pumped into a disc DMF equipped with a 0.45 μιη-Ultipor N66® membrane. Filtration is carried out at a feed rate of 860 ml / min. at which equilibrium filtrate flow of approx. 860 l / hr / m2 through the membrane. The filtered milk is measured at 25 ° C. Other conditions in this example are shown in the following Table V.

This example shows that chilled skimmed milk at a temperature of approx. 18 ° C can be treated by the method of the present invention through a bacterial retaining membrane. The diminished filtrate flows at this diminished temperature are thought to reflect the higher viscosity of milk at this temperature, as compared to higher temperatures.

EXAMPLE 16

A disk DMF is equipped with a 0.45 μιη-Ultipor Ngg® mem-brane. Whole milk is fed into the disc DMF used at a rate of 900 ml / min, and an equilibrium filtrate flow of approx. 850 1 / hour / m2 through the membrane. The test is performed without recirculation of the unfiltered portion of the feed stream.

This example shows that whole milk can be filtered using the method of the present invention using a slice DMF. Skimmed milk under substantially identical conditions gives an approximate equilibrium filtrate flow of approx. 1,600 l / hr / m2. The observed difference in filtrate flows between skimmed milk and whole milk corresponds approximately to the ratio differences in the liquid viscosities.

EXAMPLE 17

Using the methods previously described, a filtration experiment is performed using a slice DMF, maintaining a high ratio of filtrate to feed material. In this experiment, a 0.45 µm Ultipor Ngg® membrane is used. The supply of skimmed milk is maintained at 115 ml / min and a rotational speed of 2 .100 rpm is used. A filtrate flow of 460 l / h / m2 is obtained.

DK 171112 B1 41

K

H S o o o 4J · η ιο in vo r-H + J CO CO ^ 1 *

fa rH

y Ό-Η o o o £ w H ** 10 |, 3 4J o) £ η η γμ nao M N O 'fa <w (Η o' o 'o i H H CM CM in • H U XX · »fa fa

H -D w O O O

► iJD V CO oo Η ^ h in C ooo • hooo B in mh '-fa * · · · O η n cm dj 9) in in m firj „" Μ * Μ * · Μ · ssiooo ®vo ®co ® io £ £ £ 1, »» åy in vo

S H H H

Example 17

For a demonstration of extended operation, a test is performed with a large quantity (500 liters) of raw, unpasteurized skim milk. The milk is preheated to 50 ° C when passing through a plate heat exchanger. It is then homogenized in accordance with Method B and then pumped into a cylindrical DMF equipped with a 0.65 µm membrane. In this example, the dynamic microfilter is typically held at 5,000 rpm. The supply pressure varies from 1.3 to 1.5 bar at a feed rate of approx. 1.300 ml / min. The ratio of filtrate to feed is maintained at over 95%. An equilibrium filter rate of approx. 1,680 l / hr / m2. There is no reduction in the flow of filtered milk, nor is there any reduction in the supply pressure during the 6 hours of continuous operation required to treat the 500 liters.

This example shows that it is possible to use the filtration process of the invention for extended periods of time.

EXAMPLE 19

This example shows that the process of the present invention can be used for filtering milk using a slice DMF for the purpose of extracting proteins from the milk. Proteins in milk generally have a size range of approx. 0.02 to approx. 0.30 μια (D. G. Schmidt, P. Walstra, and W. Buchheim, Neth. Milk Dairy J. 27, 128 (1973)), which allows them to be extracted according to the method of the present invention.

This is particularly important in the extraction of biologically significant proteins from transgenic animals, e.g. transgenic cows, sheep and the like that have been genetically altered to stimulate the production of such proteins, in accordance with the methods already known in the art.

A disc DMF equipped with a 0.2 μιη nylon filter is used. The filtration of the milk is carried out at a feed rate of 840 ml / min, at which an equilibrium permeate flow of approx. 850 l / hr / m2 through the membrane at a rotor speed of 3,500 rpm, while recirculating the retentate, discarding the permeate by the process of the present invention. The added material, permeate and retentate are sampled for a specified period of time and then analyzed for total protein according to the Kjeldahl method. It is determined that initially the protein content of the retentate is the same as in the feed material, but is increased (4.9% retentate and 3.1 feed material) with prolonged recycling of the retentate.

Use of a smaller pore size membrane should give an even better concentration of protein in the concentrate stream.

EXAMPLES 20 and 21

These examples are carried out to establish that there is no fractionation of the constituents in the milk during the filtration process of the invention. In these examples, samples of the feed material, filtrate and concentrate are analyzed at various times during the filtration to determine protein concentrations by the Kjeldahl method and the amount of total solids by evaporation.

EXAMPLE 20

Samples of feed material, filtrate and concentrate are taken at various times during the experiment described in Example 18 and the samples are analyzed for total solids in each stream. The data in Table VI shows that there is no significant removal of the total solids from the filtrate when using a 0.65 μιη-mem brane.

EXAMPLE 21

Samples of feed material, filtrate and concentrate are taken at various times during the execution of Example 13, and the samples are analyzed for total solids and proteins in each stream. The data obtained are shown in Table VI. Also in this case, there is no significant removal of solids or proteins from the filtrate mixture when using a 0.45 μιη membrane.

DK 171112 B1 45

C

□ in m S r> oo <#> & <n oo fc η ύ

5 S

J S

W t "O

(Bh h r «·

<H H

c B w “

rh

I i 2 a s • H »· * fe o \ oo

Id 0 in C n £ Kl «I li | p

S Ϊ1 H

(S · Η **

Eh I fo Η n • H »· fe I« S 8 m m

Hr g B <o o m ° h jT {n rsj

• 5 U Η H

li ύ i! EXAMPLE 22-28

Examples 22-28 are carried out to demonstrate that bacteria can be removed from milk by the present invention. The general operating procedure is maintained as the same as in the experiments in Examples 6-18, but except that bacteria are added to the process stream using Method C. The bacteria E. coli, commonly found in milk, are used in these experiments for grafting. , unless otherwise stated. Samples of feed material, filtrate and bacterial concentrate are taken at different times during filtration using sterile technique. These samples are analyzed for bacteria using the procedure set out in Method D and the results are shown in the following Table VII.

As can be seen from the above table, the present invention can achieve drastic reductions in the bacterial content of milk. The highest degree of elimination of E. coli can be directly translated to a high removal of Bacillus cereus bacteria that cannot be completely removed using conventional pasteurization. E. coli is known to have a rod-like structure with a dimension of from ca. 1.1 to 1.5 μιη times 2 to 6 μτ &, while Bacillus bacteria such as Bacillus cereus have similar dimensions and also have rod-like structures, with the dimensions ranging from approx. 1.0 to 1.2 / m times 3 to 5 pm. Thus, the ability to remove E. coli, as shown in Table VII, also shows that the process can also remove the highly undesirable Bacillus cereus bacteria, resulting in milk with extremely long storage life, even at room temperature.

EXAMPLE 22. 23 and 24

Examples 6, 8 and 9 are repeated but except that E. coli is introduced into the process stream by Method C. Samples of feed material, filtrate and concentrate are taken for bacterial analysis. The data obtained for titer-for-reduction are shown in Table VII.

EXAMPLE 25

Example 13 is repeated, except that bacteria are introduced into the feed stream by Method C and the bacterial concentrate is not recycled back to the treatment vessel. An equilibrium milk flow of approx. 1,600 l / hr / m2. The microbiological data are presented in Table VII.

The filtered milk contains only very low amounts of 7 to 10 bacteria per day. per ml of milk, which is drastically lower than the feeding levels of 106 bacteria per ml. ml. In this case, the titre reduction is greater than 105. By comparison, during conventional pasteurization of milk titre reductions of only approx. 102 to 103.

EXAMPLE 26

The test conditions and procedures of Example 12 are repeated in this experiment but with the exception that E. coli is added to the feed stream by Method C and the concentrate is not recycled in the process vessel. An equilibrium milk flow of approx. 850 1 / hour / m2. Samples of feed material, filtrate and concentrate are taken for bacterial analysis. The data shown in Table VII shows a titer decrease of more than 106. Sterile milk is produced since no bacteria can be detected in the filtered milk.

This example shows that the process of the present invention can substantially completely remove bacteria from milk using a disc DMF and a suitably selected membrane. Thus sterile milk can be prepared.

EXAMPLE 27

Unpasteurized raw milk contains a wide variety of organisms, including coliform organisms such as E. coli and 48 pathogens such as Listeria and Campylobacteria, as well as Baccilus cereus bacteria. In this example, no external bacterial grafting of the raw milk is performed, but the milk is tested for intrinsic or "native" bacteria.

Samples of feed material, filtrate and concentrate for bacterial analysis are taken during the experiment of Example 18, and these samples are analyzed for native bacteria by Method D.

There are only 14 bacteria per ml in the filtrate. The added material contains 2,500 bacteria / ml and the concentrate contains 2 x 10 4 bacteria / ml. Furthermore, no psychrophilic bacteria are detected in the filtrate. Psychrophilic bacteria are those that grow at low temperatures and cause destruction of chilled milk.

Table VII summarizes the results of the experiments in Examples 22-27. The indicated data show that by using both the cylinder method and the disc method, increased titer reduction is obtained at the expense of filtrate flow. The table also shows that by selecting the correct membrane a sterile milk filtrate can be obtained.

DK 171112 B1 49 6! Ή 0) 61 Ό> * 1, Ό, * **> IL% s s s s g h g η H oo Λ cm &

• H dP dP

b + J CM dP into dP

<D ω σι cm dp σι o dp +> p σι σι ιο σι ο · «# χ Η σι σι σι σι * · ^ (Π φ ^ kk k Ο k η τ'ι σι σι σι σι ο σι dp <H <ησι σι σι η σι å I „„ (U φ · Η · Η · Η · Η · Η ΟΙ · Η r — I rH ιΗ r — I ιΗ «ρ- *» IJ 8 8 8 8 8 .a Η cS ιη · ιθ · ιο · νο · Ό. · η Λ • Η 8¾ Ο W Ο Η Ο Η Ο W OW Ο Ζ Εη λ Λ Η ^ Η Η ι — I '- · τ τ-i ^ -¾ u ο in η cm οι η jj Λ οί Η Ο Ο Η 5 i 1 a & • Η J CJ Ο Ο Ο ο + + 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 + I JO JO JO JO JO JO JO JO JO ϊο ιο οο ιο S 2 H rH rH τΗ γΗ © © © © © © © V0 ΙΟ ΙΟ V0 ΙΟ VO VO ΙΟ 10 ΙΟ ΙΟ (ζ ζ ζ ζ a ζ

8 I & & tL I I

η ι ii ι ι ι LI * * ^ ί * 0} οι ιη ιη ιη ιη ΜΜ CM · Τ ΙΟ ^ 1 · CM Ό Π Q) β * - * Λ H 3. ο ο ο ο ο cm η> 3 · ιη νο r ~

HE CM CM CM CM CM CM

EXAMPLE 28

Apart from the bacteria (E. coli) being tested for titre reduction, there are pathogenic organisms in milk such as Listeria which are of real practical importance in the dairy industry. These pathogens pose a more serious challenge than the coliforms (E. coli) that have also been tested in the dynamic filter. Experiments are performed using the methods set out in connection with Method D to determine whether the membrane filter elements used can effectively remove these pathogens. This experiment is performed using an off-line test equipment and not on the dynamic filter.

The data presented in Table VIII clearly shows that a 0.45 μιη-Ultipor Ngg® membrane with a specific bubble point (ASTM F316-86) will provide absolute removal of Listeria.

DK 171112 B1 51 E * I 3

S I

«I C VO 'o

. if I s S

fl EH <W Λ * 3 * Ό * il λ |

h II Ο Η I

Έ pn i3 i i

«I

I fl p. R- <H 5 I o o

(g 10 ri rH

g -H O

I η H vo rH

Η Μ P 'vo tø £ il S £ £ jf * * Λ .Q in N ^ ο vo in η

n rH rH rH

© © © VO vo vo T T S? £ I k & & e 2 3 2 η d d d * I i i i g λ in in H e tj · vo N Qi ^^^ H S ο ο o 52 DK 171112 B1 EXAMPLE 29

Filtered skim milk prepared by the method of Example 16 is collected in a sanitized container.

Commercially available cream is heated to 65 ° C and filtered through a 0.2 μιη-Ultipor N66® filtration cartridge available from Pall Corporation, East Hills, NY, with a minimum titer reduction of 106 with E. coli bacteria . The filtered cream is essentially bacteria-free and collected in a sanitized container.

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

Claims (5)

53 DK 171112 B1 A process for treating raw milk for the preparation of treated milk with a lower bacterial content than the raw milk, characterized in that (1) the milk is separated into a fat fraction with a minimum fat content of approx. 10% and a skim milk fraction; (2) homogenizes the skim milk fraction and within ca. Five minutes from the homogenization, the skim milk fraction undergoes dynamic microfiltration by passing the skim milk fraction through an average pore size microfilter sufficient to reduce the bacterial content of the milk flowing through the filter to obtain a filtrate having a lower bacterial content and a concentrate having a higher bacterial content than the original skim milk fraction, (3) separately decreasing the bacterial content in the fat fraction and (4) then combining the skim milk fraction after microfiltration and the fat fraction having the reduced bacterial content. Process according to claim 1, characterized in that the bacterial content of the fat fraction is reduced by dynamic microfiltration. Process according to claim 1, characterized in that the bacterial content of the fat fraction is reduced by pasteurization. 4. Process for producing milk with a low bacterial content and a fat content of approx. 2%, characterized in that one (1) homogenizes a skim milk fraction and within approx. Five minutes from the homogenization (2), the skim milk fraction undergoes dynamic microfiltration by passing the skim milk fraction through an average pore size microfilter sufficient to reduce the bacterial content of the milk flowing through the filter to obtain a filtrate having a lower bacterial content than that. original skimmed milk fraction, and a concentrate having a higher bacterial content 54 (3) 17112 B1 than the original skimmed milk fraction, (3) reduces the bacterial content of a cream fraction with a minimum fat content of approx. 10% and (4) then combine the skim milk fraction after microfiltration and the cream fraction having the reduced bacterial content.