Membrane Processing in the Dairy Industry

Milk is a complex material which can be separated into a number of important components. To help with this had led to great strides with membrane processing in the dairy industry.

Milk is complex because in the main it is a multi-dispersed system of protein especially casein and whey, non-protein nitrogenous (NPN) compounds), fats, sugar in the form of lactose, vitamins and minerals. It also contains particulates as well as soluble molecules which have different shapes and charges.  The largest natural particles of milk are fat globules (1-15μm, with an average of around 3.5μm) and the following are casein micelles with average diameter of 0.4μm. 

Cross-flow filtration is a technology of great potential in dairy milk processing

Microfiltration (MF) can be used for separation of casein from whey proteins and relies on the physical and chemical differences between the two types of protein. Milk contains fat in the form of globules which are actually micelles. Generally the fat globules can be separated from whole milk using the relative difference in density between the two phases. Here, the cream centrifuge or decanter is employed. because of the relative size difference between fat globules and proteins, cross-flow microfiltration could be usefully used. There are many studies to support the application where a pose-size of 5 microns was enough to separate and fractionate the milk fat globules according to diameter.

Yogurt and cheese manufacture have also employed membrane fractionation. These foodstuffs are made with small fat globules which have a finer structure amongst other benefits. The downside is the structure of the fat globule itself. These globules are surrounded by a thin film which is known as the milk fat globule membrane (MFGM). This film is sensitive to mechanical impact and oxidation due to the high concentration of phospholipids which are rich in unsaturated fatty acids. Before employing any microfiltration at this point, the fracturing of the milk globule needs to be taken into account. Here, the centrifuge has the advantage because the relative shear forces on the globule are not as strenuous. 

There are as yet no known commercial application of this technology.

Microfiltration For Removal Of Micro-Organisms

Membrane CF is a viable alternative to heat treating milk because it can effectively remove bacteria from any liquid food. Removal of bacteria, spores and even virus is possible with MF. The process is often called cold pasteurisation.

Bactocatch Process

Commercial manufacturers of membranes such as Alfa-Laval/Tetra Laval have patented the Bactocatch process. This method relies on a milder pasteurization process using  an MF membrane of 1.4 microns which is operated under UTMP conditions. A typical operating condition would be a UTMP of less than 0.5bar but very high cross-flow velocities of 6 to 8 m/s.  

Cream is separated from milk using a cream separator. This is needed to increase the volume concentration ratio (VCR) while maintaining a high permeate flux during the microfiltration. The reasons for doing so is to retain as much as possible of the bacteria and spores while letting milk components pass through the membrane. Common industrial permeate fluxes of MF in the Bactocatch process ranges from 500-800L/h/m2, with a VCR of up to 20. The proteins and total solids transmissions are about 99 and 99.5%, respectively.

The operation time approaches 10h before membrane washing is needed. Microfiltration of skim milk with a 1.4 μm pore-size membrane produced a 3.79 log reduction in total bacteria and a further reduction of 1.84 log after the subsequent minimum pasteurization, leading to a total reduction of 5.63 log. Based on total microbial counts, the shelf life of this micro-filtered pasteurized milk when stored at 4.2ºC was beyond 92 days but the real shelf life was limited to 42d due to proteolysis. The extension in shelf-life is still much longer than the shelf life of normal pasteurized milk, which is about a week at refrigeration condition.

High temperature, short-time (HTST) treated milk can last only up to 14d due to spoilage bacteria. Ultra-pasteurization can extend the shelf life to 45 days under refrigeration conditions, however, this heat treatment creates a distinct cooked and caramelised flavor, which is not acceptable to a number of consumers.

The main advantage of membrane filtration is that it can reduce bacteria to an acceptable level while maintaining not only the typical flavor but also higher nutrition value of milk which is due to the less thermal impact. The other advantage of MF over thermal pasteurization is that the former can remove effectively spore-forming bacteria while the latter is aimed to destroy vegetative pathogens.

Skim Milk Processing Using Membranes

Membrane processing can produce fractions of considerable value to the dairy ingredient supplier. Skim milk is concentrated using ultrafiltration, or with microfiltration in combination with diafiltration to produce milk protein concentrate.

Microfiltration is employed to fractionate skim milk into two major portions, a micellar casein solution which is used in cheese manufacturing and a whey protein solution which is usually concentrated further by ultrafiltration to produce whey protein concentrate. If this material is passed through an ion-exchange column for chromatography, then three major proteins lactoferrin, alpha-lactoglobulin and beta-lactoglobulin can be isolated.

Ultrafiltration of skim milk is used as a method of standardizing milk for sale. The ice-cream industry makes use of skim milk concentrate which is prepared by reverse osmosis.

Nanofiltration 

Nanofiltration is the most popular membrane technology for cleaning whey. Whey is the main by-product obtained from
cheese production. It contains a high concentration of organic matter, mainly proteins and lactose, and mineral salts. A typical whey composition is 5% lactose, 1% proteins and 0.5% ash (Roginski et al., 2003).

Nanofiltration is a membrane process used to separate mineral salts from lactose, where the milk proteins were previously removed by ultrafiltration. Nanofiltration processes have been studied in some detail too, to reduce the losses of lactose and to improve the process capability to demineralize whey (Jeantet et al., 1996; Mucchetti et al., 2000; Atra et al., 2005).

The process involves rejection of mostly divalent and multivalent ions such as calcium, magnesium and iron. The rejection of monovalent ions is low (Cuartas-Uribe et al.,2007) but dependent on its concentration in the feedstock and the overall composition of the milk.  

The mechanisms of separation are not well understood. It is a complex mix of steric, electrical and osmotic pressure effects. A good understanding of the Donnan effect is needed to fully appreciate the phenomenon. The electroneutrality principle is often cited as a starting point for defining the mechanisms that lead to separation (Xu & Spencer, 1997).

The overall performance as we mentioned earlier is also dependent on the membrane configuration, the chemical and physical behaviour of the nanofiltration membrane such as its hydrophobicity, isoelectric points of all the components, membrane porosity, the nature of the solute etc. (Mohammad & Takriff, 2003).

The filtration is used to treat sweet whey for example which also contains lactose. Solute flux is the main characteristic of  performance. One of the solute characteristics relates to pH of the feedstock and how it influences the isoelectric points of all the proteins and the membrane itself. It’s worth checking the references by  Tanninen and Nyström, 2002; Mohammed & Takriff, 2003 and  Qin et al., 2004.

In those studies cited above, the membrane used is less negatively charged which implies that it’s neutrality increases especially when salts are present. This is due to the adsorption of solutes with charge and the dissociation of functional groups making up membrane composition. Referring back to the Donnan effect, counter ions such as calcium, magnesium, potassium and sodium probably decrease the zeta potential (Deshmukh & Childress 2001; Shim et al. 2002). Later studies from Tay et al., (2002) and Teixeira et al., (2005) state that the salt concentration alongside pH influenced membrane charge and understanding this component was crucial to defining the best approach to nanofiltration.

Fouling And Nanofiltration

Any proteins present will foul nanofiltration membranes. With whey, the main issues are caseins, lactalbumin and beta-lactoglobulin. If fats are present, that is an issue too. They not only bind to the membrane but also clog the pores. A discussion of cleaning membranes and ways to recover permeate flux are discussed elsewhere. The main ions in milk and whey such as phosphates and calcium also foul membranes (Alkhatim et al., 1998; Tay et al., 2002).

One approach to reducing fouling is to try ultrafiltration before nanofiltration and using membranes with hydrophilic properties  (Rektor and Vatai, 2004; Atra et al., 2005; Cuartas-Uribe et al., 2007).

 The key methodology is to understand the composition of the feedstock especially milk and whey and then the characteristics of the membranes. The important factors are: membrane surface area, molecular weight cut off which is between 150 and 300 Daltons in most cases, the maximum pressure and temperature a membrane can withstand, its pH tolerance, isoelctric point which is often around pH 4 to pH 4.5, the zeta potential and contact angle. You can also determine a mean pore radii, thickness and porosity, rejection level for salt and lactose (in the whey industry especially) and a water permeability coefficient. A high water permeability coeff. is regarded as a positive feature for nanofiltration membranes.

The higher the water permeability coefficient, is as much due to membrane strength and compression resistance as any other factor of performance.

The study by Cuartas-Uribe et al., (2007) is one of the most informative.

Nanofiltration Membrane Performance

In nanofiltration, the most effective models are those employing osmotic pressure and that of Kedem-Spiegler. Osmotic pressure is most dominat when low transmembrane pressures (TMP) are applied. As the TMP rises, concentration polarization takes over.

Any permeate flux differences can be explained by examining the thickness and porosity of the membrane and the fitted pore size.  The physical characteristics too of the membrane also have a bearing – a nonwoven layer for support is used which affects rigidity and ultimately might have a bearing on pore size.

In terms of membrane rejection of ions, as  transmembrane pressure increases so does ion rejection of any type until it is nearly 100 per cent. The monovalent ions such as potassium and sodium pass through more easily but it does depend on the membrane. Levels of 20 to 40 per cent rejection are possible at the highest operating transmembrane pressures. It is thought that anions such as chloride will pass through the membrane because it is maintaining electroneutrality of the permeate stream. This has been observed by Gilron et al., (2001); Szoke et al., (2002) and   Cuartas-Uribe et al.,(2007).

Explanations for ion rejection include Donnan exclusion mechanisms and steric effects. These effects are described more fully by Marcus (1985) – steric effects, and by Afonso and de Pinho, (2000) and  Labbez et al., (2003).

Most polyvalent ions are firmly rejected by nanofiltration membranes.

Nanofiltration coupled with diafiltration will achieve a very high level of salt removal. In one study, continuous diafiltration using water reduced the feed solution conductivity to 1000 μS cm^-1. 

References

Afonso, M. D., & de Pinho, M. N. (2000). Transport of MgSO4, MgCl2, and Na2SO4 across an amphoteric nanofiltration membrane. Journal of Membrane Science, 179(1-2), pp. 137-154 (Article).

Alkhatim, H. S., Alcaina, M. I., Soriano, E., Iborra, M. I., Lora, J., & Arnal, J. (1998). Treatment of whey effluents from dairy industries by nanofiltration membranes. Desalination, 119(1-3), pp. 177-183 (Article).

Atra, R., Vatai, G., Bekassy-Molnar, E., & Balint, A. (2005). Investigation of ultra-and nanofiltration for utilization of whey protein and lactose. Journal of Food Engineering, 67(3), pp. 325-332 (Article).

Cuartas-Uribe, B., Alcaina-Miranda, M. I., Soriano-Costa, E., & Bes-Pia, A. (2007). Comparison of the behavior of two nanofiltration membranes for sweet whey demineralization. Journal of Dairy Science, 90(3), pp. 1094-1101

Gilron, J., Gara, N., & Kedem, O. (2001). Experimental analysis of negative salt rejection in nanofiltration membranes. Journal of Membrane Science, 185(2), pp. 223-236 (Article). 

Jeantet, R., Schuck, P., Famelart, M. H., & Maubois, J. L. (1996). Intérêt de la nanofiltration dans la production de poudres de lactosérum déminéralisées. Le lait, 76(3), pp. 283-301 (Article)

Labbez, C., Fievet, P., Szymczyk, A., Vidonne, A., Foissy, A., & Pagetti, J. (2003). Retention of mineral salts by a polyamide nanofiltration membrane. Separation and Purification Technology, 30(1), pp. 47-55.

Marcus, Y. (1985). Ion solvation. Wiley.

Mohammad, A. W., & Takriff, M. S. (2003). Predicting flux and rejection of multicomponent salts mixture in nanofiltration membranes. Desalination, 157(1-3), pp. 105-111 (Article)

Mucchetti, G., Zardi, G., Orlandini, F., & Gostoli, C. (2000). The pre-concentration of milk by nanofiltration in the production of Quarg-type fresh cheeses. Le lait, 80(1), pp. 43-50 (Article).

Qin, J. J., Oo, M. H., Lee, H., & Coniglio, B. (2004). Effect of feed pH on permeate pH and ion rejection under acidic conditions in NF process. Journal of Membrane Science, 232(1-2), pp. 153-159 (Article).

Rektor, A., & Vatai, G. (2004). Membrane filtration of Mozzarella whey. Desalination, 162, pp. 279-286 (Article).

Roginski, H., Fuqua, J.W. and Fox, P.F. (2003) Encyclopedia of Dairy Sciences, Academic Press, London.

Szoke, S., Patzay, G., & Weiser, L. (2003). Characteristics of thin-film nanofiltration membranes at various pH-values. Desalination, 151(2), pp. 123-129 (Article).

Tanninen, J., & Nyström, M. (2002). Separation of ions in acidic conditions using NF. Desalination, 147(1-3), pp. 295-299 (Article).

Tay, J. H., Liu, J., & Sun, D. D. (2002). Effect of solution physico-chemistry on the charge property of nanofiltration membranes. Water Research, 36(3), pp. 585-598 (Article).

Teixeira, M. R., Rosa, M. J., & Nyström, M. (2005). The role of membrane charge on nanofiltration performance. Journal of Membrane Science, 265(1-2), pp. 160-166

Xu, X., & Spencer, H. G. (1997). Transport of electrolytes through a weak acid nanofiltration membrane: Effects of flux and crossflow velocity interpreted using a fine-porous membrane model. Desalination, 113(1), pp. 85-93 (Article).  .