Filtration technology has filled a supporting role in food and beverage production, but multifunction filters are taking a star turn by replacing other technologies in some applications.

Dairies will be in the first wave of food processors to leverage charged ultrafiltration membranes to boost throughput in protein separation, but the technology has broad potential in other processes, as well. Source: Fonterra Cooperative Group.

Dairy processors and other food manufacturers are facing stricter wastewater discharge limits of nitrogen and phosphorous elements, making membrane bioreactors a more viable option in pretreatment. Source: O’Brien & Gere.

Pore size, for the most part, dictates what permeates through a filter and what is retained, with the rate of restriction escalating from microfiltration to ultrafiltration and finally nanofiltration and reverse osmosis (R/O). That limitation would seem to categorize filtration as a one-trick pony, but filters are tackling a broader range of processes, from wastewater treatment and wine clarification to protein separation for exciting, new nutrition sources.

At the extreme, R/O permeates only water, deflecting salts, sugars, proteins and fats from the permeate stream. Desalination would be a logical application, though upstream filtration is needed to avoid overwhelming the system. Research conducted in Japan in the 1980s attempted to address this with purposefully charged UF filters. The concepts outlined by the Japanese scientists didn’t find an audience in North America for a decade, but in the years since, a community of chemical engineers has grown and begun applying the principles to food. Their work is preliminary, and industrial applications may be years away, but manipulation of membrane charge could drive cost reductions and throughput gains in protein fractionation, beginning with whey and extending to value-added products from foodstuffs currently discarded as fodder for feed.

Whey protein concentration is the likely first application for charged UF membranes. Alpha-lactalbumin and beta-lactoglobulin account for most of the protein content in whey, though immunoglobulin, bovine serum albumin, glycomacropeptide, lactoferrin and lactoperoxidase also are found. When isolated, those proteins have considerable commercial value, but segregating them is a challenge, particularly when there is little difference in size and molecular weight. By charging the membrane and the proteins to be retained, dairy processors could retain molecules that otherwise would be small enough to pass through a UF membrane.

Andrew Zydney, head of the chemical engineering department at Penn State University, pioneered North American research in this area in the late 1990s and is currently investigating optimal surface charge characteristics for different protein separation processes, including purification of whey proteins. “The advantages of using charged ultrafiltration membranes are largely in cost and throughput,” Zydney states. “They are much easier to scale to very large-scale manufacturing operations than chromatography, and they have the potential to provide significant cost savings.” For reasons that remain unclear, charged membranes appear less susceptible to fouling than conventional membranes.

Mark Etzel, a chemical engineer and professor of food science at University of Wisconsin-Madison, concurs. Because of UF’s relatively large pores, liquids would run at a higher flux, possibly as much as a 10-fold increase in flow rate, “and still reject proteins smaller than the pore,” he says. Etzel and his colleagues are working primarily with milk and whey, but the technology could be applicable to all types of animal and plant proteins. The growing market for soy protein isolates produces meal of little current value, yet it contains potentially valuable proteins. “It’s the cheese whey of plant proteins,” he says, a reference to the nuisance status whey used to hold in cheese making.

Charged membranes are following the path of chromatography, “a workhorse in the biopharma field,” adds Etzel. Originally, chromatography relied on uncharged beads in a tank to filter and exclude particles based on size. Once a charge was added to the beads, applications exploded. Etzel developed a process using ion exchange chromatography to isolate whey proteins without denaturing them, with patent rights assigned to Grande Cheese Co., Lomira, WI. But chromatography is not widely used in the food industry. Membrane systems, on the other hand, have a significant installed base, and the “human capital and experience” could be leveraged to exploit the advantages of charged membranes, once the chemistry is better understood.

Some membranes already are charged, though they are not as finely tuned as is necessary to separate protein molecules. Millipore Corp. and Pall Corp. have fabricated some of the prototypes used by university researchers, but most filter suppliers are taking a wait and see approach. “We’re pretty much into large utilizations,” says Carl Hoffman, business manager for Koch Membrane Systems’ food/dairy and beverage practice.

Large utilizations may not be too far off. Work in protein isolates from soy and canola meal (“Research’s long and winding slog,”Food Engineering, October 2010) and the polyphenol hydroxytyrosol in olive meal (“From waste to elixir,”Food Engineering, September 2010) are indicative of science’s growing knowledge base of molecules with key health benefits that often are found in byproducts of extraction processes.

Filtration is the primary step in Creagri Inc.’s fractionation of olives into oil, vegetation water, solid particles and the stone. The Hayward, CA company is one of the rising number of nutritional ingredient developers that rely on filtration technology to isolate beneficial molecules found in food. Source: Creagri Inc.

CIP with COW water

Geography, as much as technology, can dictate how and where filtration is used. Modifications to cross-flow filtration systems can produce permeate that is virtually pure water, but value is tied to water’s availability. The clarity would be a benefit to fruit juices, notes Bruce Blanchard, national sales manager-North American filtration at GEA Filtration, Hudson, WI, but the permeate has little value in the water-rich Northwest, where many fruit juices are manufactured. In arid regions, water availability often is a constraint on production, but applying the technology in the Southwest, home to a growing number of large cheese manufacturers, is another matter. Reverse osmosis polishers can yield potable water at acceptable operating costs, and that is driving installations in that region.

The core technology has been around for decades. Conventional reverse osmosis systems consume significant amounts of energy, making them economically unattractive. By lowering operating pressures from about 600psi to as low as 200psi, large-scale desalination can reduce pumping requirements sufficiently to lower energy costs and create a favorable economic outcome.

In recent years, GEA Filtration has installed recovery systems at dairies, primarily cheese operations in New Mexico and the Texas panhandle. Wastewater from R/O systems and condensate of whey (COW water) from evaporators retains trace amounts of organic materials, so it is further processed in a low-pressure R/O polisher. Depending on the need, any of the three grades of water specified in the Pasteurized Milk Ordinance is produced: category III, which can be used for general cleaning but not on product contact surfaces; category II, which can be used for any application except final rinse in a CIP cycle; and category I, which can be used as process water.

The system reduced one cheese plant’s reliance on municipal water by a third. The plant, which produces whey protein concentrate and ultrafiltered permeate powders as well as cheese, takes in 700,000 gallons of raw milk a day. Assuming solids content of 12 percent, that translates to 616,000 gallons of COW water. Half of it is recovered and used for final CIP rinse, process water for ultrafiltration and boiler feed. Membranes are lasting two to four years, according to Blanchard.

GEA also engineers R/O systems that operate at around 900psi to help concentrate whey that must be transported off site. Instead of trucking whey that is 80 percent water, the process can boost solids content to 30 percent, enhancing its value for fractionation and lowering fuel costs for transport.

GEA does not manufacture membranes, instead sources them from multiple suppliers to fit the application. Separators are a bigger part of its business, and separators are often viewed as competing technology. The centrifugal force of a separator relies on density differences to separate elements, while filters are based on molecular size disparities. A collaboration between the German companies Westfalia Separator AG and Seitz-Schenk resulted in Profi, a replacement for diatomaceous earth in beer filtration. Profi is jointly marketed by erstwhile competitors GEA and Pall Corp., the respective owners of those German firms (see “Filtration’s strange bedfellows,” Food Engineering, April 2005). It took years to convince brewers the process would have no adverse affect on beer’s flavor profile, color or other characteristics. Carlsberg Group installed the first system at its Frederica, Denmark brewery in 2005. With the recent installation of a system in Africa, Profi filtration now is in place on six continents.

Profi’s road to commercial acceptance was short, compared to wine filtration. Wineries began experimenting with cross-flow filtration as a replacement for diatomaceous earth in the 1970s, but the results were more than disappointing. Temperatures spiked during the clarification process, and oxygen pickup contributed to browning and other degradation issues. In time, mechanical separation was all but abandoned.

A decade ago, “a new generation of cross-flow systems came out” for wine clarification, according to Nicole Madrid, global marketing manager for wine & spirits at Pall, Port Washington, NY, and wineries big and small are embracing the technology. Filtration systems designed specifically for wine applications addressed the shortcomings of the early filters. Wider-pored microfiltration membranes mounted in larger manifolds minimized oxygen pickup and reduced heat transfer. Wineries that switched to hollow-fiber filtration discovered another benefit: higher yields.

“A good diatomaceous earth filter typically results in 0.5 percent wine losses with each pass, and two or three passes usually are needed for clarification,” says Madrid. “But with a good cross-flow filter, the losses are less than 0.3 percent, and a single pass is sufficient.”

Clarification occurs post-fermentation and prior to bottling, where cartridge filtration is the norm. Pall’s second generation clarification systems feature advanced controls that simplify the winemaker’s job. A greater benefit, however, is the dosing of a specially formulated bentonite for protein stabilization upstream of the filter modules. Aside from the efficiency of combining two steps in one, bentonite dosing eliminates settling time.

Wine remains in the sediment, or lees, and wineries typically use rotary vacuum drums to recover it. Those open systems introduce oxygen, however, and the recovered wine is devalued and sold at a discounted price. An Australian winery recently began applying hollow-fiber filtration to recover wine from lees, recovering more product that also commands a higher price.

Ultrafiltration plays a critical part in Select Milk’s manipulation of the protein and lactose content of fluid milk processes, enabling the dairy to offer “designer milk” with specific attributes.

Sweet smell of ethanol

Wastewater treatment is emerging as a filtration sweet spot. When sugars are integral to a product, filtering the waste stream not only minimizes disposal costs, it generates a revenue stream from animal-feed firms. In the case of distressed or out-of-date soft drinks, juices, beer and other beverages, a Louisville, KY company stands ready to purchase the goods and pay manufacturers based on the product’s sugar content. Each year, railcars and trucks deliver 13 million cases to Parallel Products’ Louisville and Ontario, CA facilities, where they are filtered and converted to 5.5 million gallons of ethanol.

 “Waste treatment used to be just tanks and bubbles; now it includes membrane bioreactors (MBRs),” reflects Paul Greene, vice president with O’Brien and Gere, an engineering firm that includes design/build and operation of treatment facilities among its specialties. Among its projects is Keystone Foods’ Gadsden, AL facility, Food Engineering 2010 Plant of the Year.

Nutrients such as nitrogen and phosphorous are the targets of more stringent regulations involving discharged water, and that is requiring more rigorous treatment by many food processors. “If you could put 1 part per million of phosphorous in your milk stream, new regulations in Wisconsin would limit you to 0.1 ppm in 2011,” says Greene. Chemicals could be added to knock the phosphorous out of solution, but that creates solid waste. “It solves one problem by creating another,” he points out. Some dairies are concluding MBR is a better approach.

Ken’s Foods Inc., a Marlborough, MA salad dressing and marinades manufacturer, commissioned an innovative anaerobic MBR system in 2008. Expansion plans meant 60 percent more high-strength, high-solids effluent loaded with fats, oil and grease would have to be treated. Four submerged MBR tanks with Kubota membrane cassettes help the system process 100,000 gallons of water a day, reducing BOD more than 99 percent and removing virtually all suspended solids.

Engineers continue to experiment with different membrane configurations. “Originally, they all were outside-in, with vacuum pulling water from outside to the middle,” says Greene. “Now, there are more inside-out membranes, using pressure.” These systems cost less and produce higher flux.

Removing potential adulterants in process also creates opportunities for filtration. Two- and three-stage sterile filters for compressed air and process air are subject to continuous upgrades, according to Allan Fish, senior filtration manager at Parker Hannifin Corp., Haverhill, MA. The firm collaborated with representatives of several major food companies recently in designing a line of 304 stainless steel filters to withstand caustic, high-pressure washdowns. Fish says the company also is working with food scientists at the University of Massachusetts to produce a portable device that can identify any microbial contamination in compressed air systems.

New applications drive production efficiencies for filtration suppliers, and that translates to lower costs for end-users, including food and beverage companies. Membranes used to be priced at $10-$15 per square foot, says Koch’s Hoffman. Today, they’re available for about $2 a square foot. Future benefits will come in process optimization that reduces energy requirements, as is the case with inside-out membranes.

Simplified maintenance and other user-friendly changes make the technology more attractive. “The economics of automated filtration is swinging to more applications,” says Hoffman. As the audience for filtration becomes more diverse, cost will continue to decline as functionality increases.

For more information:
Bruce Blanchard, GEA Filtration, 715-377-0533
Carl Hoffman, Koch Membrane Systems Inc., 978-694-7176,
Paul Greene, O’Brien & Gere, 518-758-2179,
Nicole Madrid, Pall Corp., 516-801-9137,
Allan Fish, Parker Hannifin Corp., 800-343-0051,