Development and Control of Bacterial Biofilms on Dairy Processing Membranes


  • Current affiliation: Diwakar Singh, Jerome Cheese Company, a division of Davisco Foods International Inc., Jerome, ID 83338.

  • Current affiliation: Mallika Avadhanula, Corporate QA Specialist at Davisco Foods, Le Sueur, Minnesota (Greater Minneapolis-St. Paul Area, MN).


Membrane fouling is a major operational problem that leads to reduced membrane performance and premature replacement of membranes. Bacterial biofilms developed on reverse osmosis membranes can cause severe flux declines during whey processing. Various types of biological, physical, and chemical factors regulate the formation of biofilms. Extracellular polymeric substances produced by constitutive microflora provide an effective barrier for the embedded cells. Cultural and microscopic techniques also revealed the presence of biofilms with attached bacterial cells on membrane surfaces. Presence of biofilms, despite regular cleaning processes, reflects ineffectiveness of cleaning agents. Cleaning efficiency depends upon factors such as pH of the cleaning agent, temperature, pressure, cleaning agent dose, optimum cleaning time, and cross-flow velocity during cleaning. Among different cleaning agents, surfactants help to prevent bacterial attachment to surfaces by reducing the surface tension of water and interfacial tension between the layers. Enzymes mixed with surfactants and chelating agents can be used to penetrate the biofilm matrix formed by microbes. Recent studies have shown the role of quorum-sensing-based cell-to-cell signaling, which provides communication within bacterial cells to form a mature biofilm, and also the role of applying quorum inhibitors to prevent biofilm formation. Major cleaning applications are also summarized in Table 1.


Membrane technology has emerged as one of the important nonthermal techniques in the dairy and other food industries. It has found increased applications in processes such as processing of whey, which is a by-product obtained during cheese making containing valuable constituents such as protein, fat, lactose, minerals, and lactic acid (Madaeni and others 2010). Because of unfavorable lactose-to-protein ratio and high biological oxygen demand, whey had traditionally been viewed as an ecological burden. However, advancements in process technologies such as membrane filtration systems have converted this former waste product into valuable material with nutritional value by separating protein and lactose from whey (Hanemaijer 1985).

With the help of reverse osmosis (RO) membranes, whey is processed for concentration (Pepper 1984), partial demineralization (Short and Doughty 1976), fractionation (McDonough and Hargrove 1972), and salt removal (Prouty and others 1994). As a part of the process modification, other membranes can also be associated with RO membranes. With the combination of ultrafiltration (UF) and RO membranes, both protein and lactose can be collected together. During typical membrane processing, permeate flows through the membrane, whereas retentate is collected from the feed side of the layers (Tang and others 2009). Membrane technologies thus offer economical solutions for the treatment of whey. They require less energy, smaller installations, and have several environmental advantages. Annually, the United States produces 10.6 billion lbs cheese (USDA 2012), which is about 4.81 billion kg. Though whey production varies with the type of cheese, on average about 9 L of whey is obtained for 1 kg of cheese produced, and a large cheese making plant can produce over 1 million L of whey every day (Jelen and others 2003; Onwulata and Huth 2008). Whey typically contains 7% solubilized solids with a composition of 10% to 12% protein; the rest consists of 74% lactose, 8% minerals, and 3% fat and also lactic acid (Morr 1989; Onwulata and Huth 2008). Although whey can be a significant biological pollutant, the nutritional quality of its solid components makes it a valuable commodity. Recovering whey solids is significant for 2 major reasons: to reduce organic pollution, and to optimally utilize the nutritional and functional properties of whey proteins (Ostojic and others 2005; Onwulata and Huth 2008). With the application of membrane processing technologies, whey yields many quality products with health benefits (Balagtas and others 2003). However, fouling is a serious problem in the use of membranes, thus affecting the operational performance of membranes and causing premature membrane replacements (Paul and Abanmy 1990; Eykamp 1995; Subramani and Hoek 2008).

Different types of fouling mechanisms include those that are particulate-based, crystallization-based, chemical binding-based, and combination thereof (James and others 2003). In addition, biofouling is also being recognized as a widespread problem on nanofiltration (NF) and RO membranes. Several investigations have shown the presence of a variety of microorganisms on the surfaces of RO membranes after prolonged operation (Bailey and others 1974; U.S. Dept. of the Interior 1979; Ridgway and others 1983; Avadhanula 2011). Membrane fouling can cause severe flux decline and affect the quality of the final product. One of the biggest challenges in using membranes is the effect of fouling due to solute adhesion and microbial fouling (Melo 1992; Ridgway and others 1999; Chang and others 2002; Ivnitsky and others 2007; Susanto and Ulbricht 2007). Membrane fouling, thus, is an important issue for dairy plants due to the cost of equipment and chemicals, and the down-time incurred when a fouling issue arises. The structural integrity of the membrane is also destroyed by biofouling, which causes membrane system failure and ultimately an increase in operation and maintenance costs.

Basic research has elucidated that when there are abundant nutrients available under favorable temperatures, bacteria on food contact surfaces develop into biofilms (Ridgway and others 1999; Sharma and Anand 2002a). Biofilms matrices may contain viable or nonviable, single-species or multispecies communities of microorganisms embedded in extracellular polymeric substances (EPS) attached to food-contact surfaces (Chmielewski and Frank 2003). In the case of membrane processing, bacterial cell attachment is governed by physicochemical interactions deriving from operating conditions, fluid quality and temperature, membrane properties, and module geometry (Bailey and others 1974; Ridgway and others 1983; Ridgway and Flemming 1996; Flemming and others 1997; Sadr Ghayeni and others 1998; Flemming 2002; Subramani and Hoek 2008; Subramani and others 2009). Further, biofilm development depends on nutrient availability, fluid dynamics, and quantity of initially deposited bacteria. Formation of biofilms on dairy separation and concentration membranes such as RO membranes affects not only the performance but also the quality and safety of the end product (Sharma and others 2003; Simões and others 2010; Avadhanula 2011). Frequent cleaning of membranes controls biofilm formation to some extent. However, a significant observation made during previous research is that the bacterial cells in biofilms are generally more resistant to cleaning agents than they are in their planktonic state (Stewart and Costerton 2001; Chmielewski and Frank 2003; Shi and Zhu 2009; Avadhanula 2011; Anand and Singh 2013). It may be due to this reason that the process of biofilm formation is observed to be continuous and thus affecting the membrane performance. The following text reviews the membrane processes and, fouling due to bacterial biofilms in dairy/food membrane processing environments.

Membrane technologies in the dairy industry

Until the 1970s, whey protein was available in a denatured form with an unappealing color that had limited applications (Wingerd 1971; Onwulata and Huth 2008). This was because of the heat treatment used in conventional methods of whey processing that involved 2 stages of whey concentration: evaporation and demineralization by electrodialysis (ED) and ion-exchange (Zadow 1992; Van der-Horst and others 1995; Jeantet and others 2000; Cuartas-Uribe and others 2009). New methods came up later, which resulted in producing more desirable forms of proteins. Amongst them, membrane application was found to be a good alternative. The conceptual origin of membrane applications for separation dates back to the 18th century when a diaphragm made of a pig's bladder was used for water permeation (Baker 2004; Pouliot 2008). Eventually, application of membrane separation spread to other industries. Developments in membrane materials and technology opened doors for innovative applications in the dairy industry (Britz and Robinson 2008). The immediate advantage of membrane processing over the other methods was that it did not damage the proteins and nutrients, because it was performed at room temperature.

Today, membrane filtration has become an integral part of the dairy industry for clarification, fractionation, and concentration of a variety of dairy products (Britz and Robinson 2008; Onwulata and Huth 2008). It is not only used for the recovery of serum proteins from whey or concentration of casein micelles, but also for the removal of bacteria and spores from skim milk. Among different types of membranes, spiral-wound membranes (Ridgway and others 1983) are the most common due to high surface area per unit volume (Whittaker and others 1984).

Common membrane processing techniques

Membrane separation is a molecular sieving technique that separates 2 or more components in the liquid phase based on size difference. Common types of membrane separation are UF, ED, microfiltration (MF), NF, and RO. Other than ED, all are pressure-driven processes with a selective permeability that separates the feed stream into permeate and retentate, with a dividing line generally referred to as a molecular weight cutoff. All the pressure-driven membranes are similar to each other; the only differences are the degree of semipermeability and the separation characteristics. The MF and UF are considered filtration methods, whereas RO is a concentration method. Membranes have optimum operational conditions such as pressure, because of differences in permeability. Milk and whey are the most suitable dairy products for membrane processing. Combinations of pressure-driven membrane processes are applied to produce whey products with different protein contents.

Membrane material, module design, and operational parameters determine the application of membranes. Commercial membranes are made of organic or inorganic materials. Organic membranes are made of a variety of organic polymers that include cellulose and its derivatives, polyamides, polysulfones, polyolefines, chlorine- and fluorine-substituted hydrocarbons, and so on. On the other hand, inorganic membranes are made of ceramic material (Berk 2008). Each membrane material exhibits different tolerances for both chemical and physical parameters, which ultimately decide the most effective chemical cleaning protocol for a membrane.

One of the critical design parameters for membrane separation technology is its throughput capacity, also known as flux, which is measured as quantity of permeated liquid (kg or L) per membrane areas unit (m2) and time unit (h). To accommodate better process conditions, it is important to pack large membrane surface areas into the smallest possible volume. This objective has led to different membrane configurations. The 4 types of membrane configurations used in the food industry are: plate and frame, spiral-wound, tubular, and hollow fiber configurations. In membrane processing, spiral-wound configuration membranes are the most common, because of their lower price and large surface area (Ridgway and others 1983; Tang and others 2009). Spiral-wound membranes are composed of 2 flat-sheet membranes, separated by a flexible spacer mesh to create space between 2 membranes for permeate to flow into a central perforated channel serving as a collector for permeate. The membrane sheets are rolled up along with mesh spacer, providing a flow channel for the retentate. Spiral-wound membranes are commercially available as cylindrical assemblies or cartridges, complete with central tube, spacers, and connections (Berk 2008). However, their design is more susceptible to fouling because of the close spacing (Whittaker and others 1984; Cartwright 2003; Tang and others 2009). Depending on the direction of the feed flow on the membrane, filtration processes are classified into dead-end flow or cross-flow. The dairy industry generally utilizes the cross-flow configuration, where permeate is collected and retentate circulates in the system until it reaches the target concentration. Fouling of membranes with cross flow configuration is drawing considerable research attention at present (Al-Akoum and others 2002; James and others 2003). The industrial application of RO started with water desalination, which is still the major user of this process. It is recognized for fulfilling strict regulations by producing safe water (Berk 2008; Kim and others 2009). Advanced technologies in the design and manufacture of polyamide, cellulose acetate, and composite polymer semipermeable membranes for a wide range of applications have made RO the preferred method for water purification (Ridgway and others 1983). The worldwide installed membrane surface areas for RO, UF, and NF of dairy liquids such as whey, milk, and permeate are continuously increasing. As per the 1995 IDF report, 76% of the RO surface area is used for whey processing, followed by permeate and milk. Whey is processed in 2 stages: the recovery of protein by the UF system and the removal of lactose from deproteinized whey by the RO system. In addition to whey concentration (Pepper 1984; Amjad 1993; Del Re and others 1998), RO is used for partial demineralization (Short and Doughty 1976), whey fractionation (McDonough and Hargrove 1972), and salt removal (Prouty 1994). Various components of whey can be separated by combining RO with other membrane processes.

Since RO concentration is limited to fluids with low osmotic pressure and viscosity, it is used for preconcentration followed by the evaporation process. Preconcentration of dairy fluids prior to evaporation results in savings in energy and cost (Berk 2008). RO technology can offer more than 75% reduction in operating cost when compared with 5 multieffect evaporation systems (Jevons and Awe 2010). Cheese whey concentration by RO gives 28% solids. In addition, dairy plant cleaning waste water contains considerable quantities of milk solids, which could be concentrated by RO and used as animal feed (Grandison and Glover 1994; Goulas and Grandison 2008). Economic applicability of RO depends on maintaining constant permeate flux throughout the membrane (Ridgway and others 1983). As the filtration process continues, the membrane pores gradually clog with feed components and develop fouling. Consequently, production efficiency begins to decline and flux drops. Flux decline over time in pressure-driven membrane processes is due to concentration polarization and membrane fouling. In membrane processing, concentration polarization is unavoidable, yet can be minimized by improving membrane properties (Marshall and Daufin 1995). One of the biggest challenges in using different types of membranes is the effect of fouling due to solute adhesion and microbial fouling (Melo 1992; Ridgway and others 1999). Some studies have been conducted in this area for the control of biofilms on membrane surfaces related to whey processing RO membranes (Avadhanula 2011).

The process of membrane fouling and its implications

Fouling is a key element affecting membrane life and, therefore, cost (Eykamp 1995; James and others 2003). It affects the separation characteristics of a membrane and the composition of the products due to loss of deposited solutes. It leads to increased system downtime for cleaning and premature replacement of membranes (Muthukumaran and others 2005; Madaeni and Ghaemi 2007). Fouling is a surface layer formed on the substrate during the process and it can be based on particles, crystals, chemicals, or a combination of these. According to reports on waste water, RO and the NF membrane processing system, a formed biofouling layer appears like a cake and is composed of proteins, polysaccharides, and bioparticles. RO membrane fouling during whey processing is a complex phenomenon. It may be due to the wide range of components present in the dairy fluids. Protein adsorption and protein–protein interaction lead to fouling. Previous studies indicated protein deposition on both whey and milk RO membranes (Lim and others 1971; Glover and Brooker 1974). Mineral contribution to fouling is less in milk than whey due to the influence of casein micelle stabilization.

Factors that could influence membrane fouling are membrane surface properties (Combe and others 1999; Childress and Elimelech 2000; Herzberg and Elimelech 2007; Li and others 2007), physicochemical properties of feed (Li and Elimelech 2004; Ang and others 2006; Ang and Elimelech 2007; Tang and others 2007; Ang and Elimelech 2008), and operating conditions of the processing system (Seidel and Elimelech 2002; Tang and others 2007). Additionally, continuous development of biofilms along with the concentration of organics and inorganics may lead to rapid flux decline with an increase in pressure by 10% to 15%. At this stage a chemical cleaning is necessary to bring back the water flux to the required flow rate. As the biofilm matures, the outer layers compress to form a more impermeable hydrophobic barrier, and may account for increased concentration polarization leading to deterioration of membrane (Paul and Abanmy 1990).

Fundamental investigations on membrane fouling revealed the effects of membrane surface properties such as roughness and hydrophobicity. Laboratory-scale experiments conducted to investigate the effect of membrane surface properties on membrane performance showed that the rate of colloidal fouling was significantly influenced by physical roughness of membrane surface. The experiments further demonstrated that more particles deposit on rough membranes than smooth membranes. These particles accumulate in valleys of rough membranes, leading to valley clogging and eventually to flux decline (Vrijenhoek and others 2001; Subramani and others 2009). It is known that the protein adsorption and biocontact properties of polymers depend on surface chemical composition and topography, surface hydrophilic or hydrophobic balance and charge, mobility of surface functional groups, thickness and density of the modifying layer, its adhesion to substrate, and so on. By changing some of the parameters, we can control foulant adsorption (Vladkova 2007). Membrane manufacturers are also working on increasing hydrophilicity of membranes to control fouling (Chen and Belfort 1999; Espinoza-Gomez and Lin 2001).

The physicochemical properties of the feed stream have a great impact on fouling and flux drop; even a small change in fluid composition could greatly affect the fouling (Mekmene and others 2009; Rice and others 2009). The initial feed whey characteristics like pH, minerals, and fat levels are influenced by cheese production method, starting milk composition, and pretreatment of milk. During whey processing, lactose and other components have little to contribute to fouling (Trägårdh 1989; Jeantet and others 2000; Madaeni and Mansourpanah 2004). Similarly, there is little evidence supporting the contribution of lipids toward fouling. This may be because of their low concentration in whey or skim milk, and the dominating effects of protein deposition and mineral precipitation. Severe fouling has, however, been experienced with high fat levels. Separation and clarification of feed before membrane processing is thus an important step for reduction of fouling (Marshall and Daufin 1995). Separation efficiency of membranes also depends on the operational conditions. Membrane systems are designed to process particular products at specific parameters of flow rates, time, and temperature. The recommended pressure drop across the membrane must be maintained at all times to maintain and achieve the required concentration with the correct permeation rate. If parameters are optimal, the average lifespan of a new membrane is approximately 2 y. Imbalance in these parameters leads to fouling and premature replacement of membranes (Tamime 2008). Membrane fouling depends on the solute and solute-membrane interactions and, hence, deposition of particles on the surface blocks the membrane pores. During this process of fouling of membranes, there are possibilities of deposition of suspended or dissolved materials at its pores or within the pores. These deposited materials may be protein or some inorganic salts based on calcium.

Role of bacterial biofilms in membrane fouling

Along with organic and inorganic compounds, bacteria are considered major fouling agents (Ridgway and others 1983; Tang and others 2009). The solid–liquid interfaces of aquatic ecosystems offer an ideal environment for microbial attachment and the formation of bacterial biofilms. Bacteria are known to attach to different food contact surfaces and eventually develop into biofilms (Sharma and Anand 2002a). Studies have also shown that food-borne pathogens and spoilage microorganisms form as biofilms on a variety of surfaces including stainless steel, aluminum, glass, Buna-N and Teflon seals, and nylon materials used in food processing environments (Herald and Zottola 1988a, 1988b; Mafu and others 1990; Notermans and others 1991; Blackman and Frank 1996; Kumar and Anand 1998, Stoodley and others 2002). Within a biofilm matrix different microbes compete for nutrients and are affected by desiccation and fluctuations in fluid flow and temperature. Previous reports have revealed that 60% of the bacterial species isolated from water RO are actively contributing to biofilm formation (Kim and others 2009). In another study, Tang and others (2009) isolated various types of organisms from whey RO and UF membranes. Several factors influence bacterial cell attachment to contact surfaces. Prominent among them are types of microorganisms and their growth stage, surface properties of bacteria and substrate, microbial and nutritional quality of the feed stream, and pH of feed. Biofilm formation is generally recognized as a 3-stage process involving development of a conditioning film, bacterial attachment, and development of mature biofilm structures. Biofilm attached to the membrane surface causes reduced flux rate along with increased module pressure drop (Flemming and others 1992; Avadhanula 2011).

Conditioning film and bacterial attachment

In a typical processing environment, bacteria and other components present in the feed stream are transported and adsorbed on substrate surfaces and form a conditioning film (Hood and Zottola 1997; Kumar and Anand 1998). The conditioning film formation happens immediately after the substrate surface comes in contact with the liquid phase (Schneider and others 1994; Sadr Ghayeni and others 1998). Conditioned film also plays a larger role in changing physicochemical properties such as surface charges and texture (Melo 1992; Carpentier and Cerf 1993; Subramani and others 2009). Conditioned surfaces tend to encourage bacterial cell attachment and biofilm formation (Loeb and Neihof 1975; Ciston and others 2008). On the other hand, the surface charge of the bacterial cell is influenced by its growth phase and surrounding nutritional environment. A combination of above parameters aggravates the process of biofouling that limits membrane performance, may lead to equipment corrosion (Oliveira 1992, Viera and others 1993; Mittelman 1998), and increases overall operation cost (Subramani and others 2009). In previous studies, a variety of surfaces have been known to attract milk proteins and form conditioned surfaces (Speers and Gilmour 1985; Mcguire and Swartzel 1989; Shi and Zhu 2009). It has been reported that in the presence of whey proteins the attachment of milk associated microorganisms is high on different contact surfaces (Speers and Gilmour 1985; Kumar and Anand 1998). It has also been observed that some pathogen attachments were more potent on solid surfaces conditioned with diluted milk rather than with whole milk (Hood and Zottola 1997; Shi and Zhu 2009). It may be because some proteins inhibit bacterial attachment to different surfaces.

It is generally believed that the attachment and formation of biofilms are supported to a greater extent on hydrophilic surfaces as compared to hydrophobic surfaces (Marshall 1992; Blackman and Frank 1996; Sinde and Carballo 2000; Chmielewski and Frank 2003; Smith and others 2004; Hassan and others 2010). Contradicting this belief, some reports stated that hydrophobicity was relatively more retention-determinant (Bos and others 2000). The primary mechanism in the attachment of microorganisms to surfaces involves cell surface hydrophobicity and secretion of exopolysaccharides ( Simões and others 2010). Exopolysaccharides act as a binding element that causes cell attachment with a conditioned surface. Therefore, it is no surprise that significant attention has been directed toward the development of efficient protein-resistant surfaces.

Role of EPS in biofilm formation

Biofilms contain single or multispecies organisms embedded in EPS (Marshall and others 1971; Chmielewski and Frank 2003). An EPS is a key component in biofilm formation and is always associated with cell surfaces. It helps with cell-to-cell attachment and also with attachment to the substrate surface to establish the biofilm structure (Watnick and Kolter 2000; Hall Stoodley and others 2004; Kolter and Greenberg 2006; Valle and others 2006; Jun and others 2009). Previous studies indicated dairy biofilms to be composed of EPS and milk residues (Flint and others 1997; Mittelman 1998; Flemming and others 2000). A key component of microbial EPS is the extracellular polysaccharide material which is associated with the cell surface or excreted into the growth medium. A relatively higher EPS production has been shown in biofilm cells as compared to planktonic cells (Beech and others 1991; Spenceley and others 1992; Evans and others 1994). In multispecies biofilms, each species will produce a different type of polymers. These polymers merge to give heterogeneous regions of polymers within the biofilm homogenous matrix (Gilbert and others 1997). The multispecies EPS components differ significantly from those of single purified components (Allison and Matthews 1992; Gilbert and others 1997). The EPS acts like a shield and protects cells from hostile environments (Costerton and others 1994; Sutherland 2001). In addition, the EPS in biofilms is also a hygiene issue in the food industry (Kumar and Anand 1998; Wingender and others 1999). Adherent microbial communities embedded in a polysaccharide matrix can thus survive and cause human infections, and may even be antibiotic-resistant. Reports also indicate that a formed EPS matrix can change the membrane surface properties. It may elevate the concentration polarization and lead to flux drop (Flemming 1997; Herzberg and Elimelech 2007; Khor and others 2007). It is for these reasons that EPS is a biological foulant known to take an active part in microbial aggregation and biofilm formation (Laspidou and Rittmann 2002; Yeo and others 2007).

Composition of EPS

Several studies on different environmental surfaces have indicated that the formation of EPS results in an irreversible attachment by different microbes (Donlan and Costerton 2002; Van Hullebusch and others 2003; Romani and others 2008). Many varieties of microbes present in a biofilm matrix produce EPS, which are composed of carbohydrates (Kennedy and Sutherland 1996), proteins, lipids, small quantities of nucleic acids, and a variety of humic substances (Nielsen and others 1996; Liu and Fang 2002; Vu and others 2009; Lee and others 2010). Environmental conditions responsible for the development of biofilm affect the composition and quantity of EPS for different types of microorganisms (Mayer and others 1999). It was observed that many bacterial exopolysaccharides contain 1,3- or 1,4-β-linked hexose residues which make the film structure more rigid (Sutherland 2001). The biofilm matrix contains 50% to 95% water with EPS as the remaining part of it (Flemming and others 1992). Lack of nutrients in the environment increases the production of EPS and promotes the hydrophobic interactions needed for sorption onto solid surfaces (Sheng and others 2008). Different bacterial strains produce similar types of EPS, but there may be some differences in their physical properties of the matrix such as viscosity and gel formation (Sutherland 2001). The secretion of surfactants by different bacteria also alters the internal matrix of the biofilm.

Development of mature biofilms

The mature biofilms consist of conditioning film, bacterial consortia with EPS, and other organic compounds. Biofilm could be either single-species (Donlan 2002; Stoodley and others 2002) or multispecies communities enclosed in EPS and thus protected from environmental stresses. It has been observed that bacterial cells attached to different surfaces as biofilm are different than their planktonic counterparts.

Biofilms on surfaces show distinctive stages of development that include reversible attachment, irreversible attachment due to EPS, formation of microcolonies, development of mature biofilm with 3-dimensional structure, cell detachment, and dispersion. Initial reversible attachment is influenced by physicochemical properties of bacterial cells, available nutrients in the surrounding medium, and growth stage of the cells. Bacteria can be easily removed during the reversible stage by rinsing, demonstrating that irreversible attachment is crucial in biofilm development. At this stage, the polymeric substance forms strong bridges between cell surface and substrate surface. Hence, at this stage, strong forces are required to remove adhering cells. Later stages involve cell division, growth, and formation of layers by using nutrients present in the conditioning film and surroundings (Marshall and others 1971; Kumar and Anand 1998). The EPS production continues at this stage, which gives stability to the microbial colony from environmental stress. Size of the microcolony further increases due to accumulation of other components from the fluid stream (Melo 1992). Although biofilm development is a slow process, it accumulates to a few millimeters thick in a matter of days. The 3-dimensional structures look like mushrooms. Microbial distribution inside the colony is not uniform and the microcolonies have been reported to contain water channels (Costerton and others 1994; Kumar and Anand 1998; Avadhanula 2011). Heterogeneous biofilms are thicker and more stable compared to their homogeneous counterparts. In membrane processing systems, the feed, process configurations, and operational conditions are responsible for differences in biofilm structures (Chen and others 2004). As a biofilm matures, large particles of it are detached due to shear effects of the fluid stream (Rittmann 1989; Applegate and Bryers 1991; Kumar and Anand 1998; Breyers and Ratner 2004; Simões and others 2010). The dispersed cells are transported to other locations and start to develop into new biofilms (Marshall 1992; Kumar and Anand 1998). Their resistance to sanitation also increases with the age of a biofilm (Lee and Frank 1991; Anwar and others 1992). The detached cells are more resistant to environmental stresses when compared to normal cells (Stewart and Costerton 2001; Chmielewski and Frank 2003; Kim and Wei 2007).

Constitutive microflora of biofilms

It has been demonstrated that biofilms act as reservoirs of different types of microflora leading to continuing product contamination. Previously, biofilms found on water filtration membranes have been reported to be constituted of several bacterial species such as Corynebacterium, Pseudomonas, Bacillus, Arthrobacter, Actinomycetes, Flavobacterium, and Aeromonas (Ridgway and others 1983, Ridgway and Flemming 1996; Dudley and Christopher 1999). Similarly, in water filtration systems employing MF and RO membranes, occurrence of biofilm was characterized using the polyphasic approach, 16SrDNA clone library, and fluorescence in situ hybridization techniques. Results revealed that out of 17 different bacterial groups, Proteobacteria was the largest microbial fraction identified on both MF and RO membranes (Chen and others 2004).

In some recent studies (Biswas and others 2010; Avadhanula 2011) the presence of multispecies bacterial biofilms on whey RO membranes has been reported. The identified bacterial species were from the genera Enterococcus, Staphylococcus, Micrococcus, Streptomyces, Corynebacterium, Bacillus, Klebsiella, Aeromonas, Pseudomonas, Streptococcus, and Chryseobacterium, as well as Escherichia coli. Bacillus isolates were observed to be the most resistant among the entire constitutive microflora against cleaning agents (Singh 2012; Anand and Singh 2013). Biofilms have also been reported to be a source of pathogens such as Listeria monocytogenes, Yersinia enterocolitica, Campylobacter jejuni, Salmonella spp., Staphylococcus spp., and even Escherichia coli 0157:H7 (Somers and others 1994; Kumar and Anand 1998; Wong 1998; Sharma and Anand 2002b). The presence of pathogens in biofilm matrices has a direct impact on food safety (Simões and Vieira 2009; Simões and others 2010). The presence of biofilms also results in lowering the shelf-life of the product (Zottoia 1994). The bacteria commonly found in dairy processing environments belong to the genera Enterobacter, Lactobacillus, Listeria, Micrococcus, Streptococcus, Bacillus, and Pseudomonas (Wiedmann and others 2000; Sharma and Anand 2002b; Waak and others 2002; Salo and others 2006). In another study, Tang and others (2009) isolated strains belonging to several genera such as Chryseobacterium, Bacillus, Lactococcus, Klebsiella, Enterobacter, Lactobacillus, Pseudomonas, and Blastoschizomyces from dairy UF and RO membranes. Within the dairy industry, Bacillus cereus strains are important postpasteurization contaminants because of their ability to form spores (Flint and others 1997; Svensson and others 2004; Lindsay and others 2006). Spore attachment has been found to be greater than that of vegetative cells to food contact surfaces, mainly due to hydrophobicity and hair-like structures on the cell surface (Rönner and others 1990; Husmark and Ronner 1992; Kumar and Anand 1998). Spores are also resistant to many cleaning protocols (Lindsay and others 2006). Cross-contamination of whey retentate with B. cereus from whey RO membrane has recently been reported by Anand and others (2012). Dairy manufacturing plants mostly observe biofilms on the surfaces of heat exchangers, which are in direct contact with flowing milk (Flint and others 1997). Within the regeneration section of the heat exchangers, thermophilic cells and spores are retained into the biofilms already formed during milk processing (Scott and others 2007; Burgess and others 2009). Multilayered films of milk constituents are formed by inconsistent milk flow and cleaning solutions in dairy manufacturing plants where bacterial cells get trapped, grow, and further develop into biofilms, including thermophilic spore former biofilms (Burgess and others 2009). Thermophilic bacilli were observed to form biofilms particularly in those sections of milk processing plant where temperatures fall between 40 °C to 65 °C such as plate heat exchangers of pasteurization units and membrane filtration plants operated at higher temperatures (Flint and others 1997; Scott and others 2007; Burgess and others 2009). During milk processing the spores are more resistant to the heat treatments so the attachment of spores to the surfaces of membranes plays an important role in biofilm formation (Parkar and others 2001). More specifically, the spores of C. butytricum had a D value of 23 min at 85 °C and pH 7.0, whereas its thermal death time was 10 to 15 min at 100 °C at pH 4.4 (Brown 2000; Carmen MatinezCuesta and others 2010). This indicates these spores can survive pasteurization in dairy plants and germinate and grow at later stages. Some strains of C. tyrobutyricum resisted low pH and grew at pH 4.5 to 7.5 (Bintsis and Papademas 2002). Similarly, Enterobacter agglomerans and Pseudomonas spp. were reported to be high in cottage cheese, due to higher pH and storage temperature (Brocklehurst and Lund 1988). Some of the spoilage microorganisms were also able to grow at relatively low pH values of 3.6 when grown in media at 20 °C. Rate of salt penetration into brined cheeses, types of starter cultures used, initial load of spores in the milk used for production, pH of the cheese, and ripening temperature affect the rate of butyric acid fermentation and gas production by C. tyrobutyricum (Stadhouders 1990). All these organisms have the potential to cross contaminate cheese processes and become a constitutive part of biofilms.

The nature of mixed-species biofilms on different surfaces

In previous studies, multispecies biofilms were obtained from different surfaces (Kawarai and others 2007; Macleod and Stickler 2007). Mixed-species biofilms are known to be more complicated than the biofilms developed by individual isolates, thus making biofilms formed by mixed cultures much more difficult to clean than the biofilms of individual cultures. Therefore, modification in a cleaning protocol should focus on the membrane material as well as on the prevention of regrowth of bacterial cells. Residual material left over after the cleaning process may act as a seed for further bacterial attachment to the surface and prevent the chemicals from adequate cleaning effectiveness. Individual cells of Klebsiella pneumoniae or Pseudomonas aeruginosa develop thinner biofilms (15 μ and 30 μ, respectively), but the mixed biofilms by these 2 species are thicker (40 μ). This noticeable increase with mixed-species biofilms indicated the possible enhancement of stability with each other (Jones and others 1969). Mixed-species biofilm formed by Klebsiella and Pseudomonas increases attachment with the surface, indicating that the microbes interact among themselves (Tang 2011).

Evaluation of biofilms by cultural techniques

To understand the implications of contamination due to biofilm formation on food contact surfaces such as membranes, it is important to study the source and nature of biofilms, and also other factors contributing to the development of biofilms. Identification and enumeration of biofilms from food contact surfaces helps find the source and types of organisms involved as contaminants. Swabbing, rinsing, agar flooding, and surface-contact method are the common conventional methods for bacterial enumeration from the surfaces of the substratum (Flemming and others 1992; Kumar and Anand 1998; Wehr and Frank 2004). There are also other methods such as scraping (Frank and Koffi 1990), and vortexing (Mustapha and Liewen 1989). These methods can be applied for biofilm isolation depending on the developmental stages of biofilms. After plating samples on the selective agars, followed by their incubation, colonies are identified on the basis of colony morphology and Gram's reaction. Further microbial identification is carried out by applying various biochemical tests as per standard procedures (Marshall 1992).

Studying biofilms by microscopic techniques

Imaging is vital to an investigation of biofilms, as that provides better understanding of the complex structure of biofilms, its relationship with surface substratum, microbial morphology, and species composition. Various microscopic techniques are used to achieve conclusive explanations for the contamination levels of biofilms (Wimpenny and others 2000; Chmielewski and Frank 2003). Scanning electron microscopy plays a key role in studying the biofilm surfaces (Notermans and others 1991). Environmental scanning electron microscopy (Little and others 1991; Hodgson and others 1995), epifluorescence microscopy (Holah and others 1988, 1989; Wirtanen and Mattila-Sandholm 1993), interference reflection microscopy, atomic force microscopy, and confocal scanning laser microscopy are other techniques used to study biofilms (Ladd and Costerton 1990; Beech 1996; Debeer and others 1997).

Confocal scanning laser microscopy has an ability to observe a hydrated sample and avoids many sources of artifacts caused by sample preparation (Hassan and others 1995). However, it has resolution limitations. Fluorescence microscopy has already been applied on various food systems to distinguish live/dead cells (Rodriguez and Kroll 1986; Duffy and Sheridan 1998; Couto and Hogg 1999; Mesa and others 2003). As a modification to the process, lectin conjugates were used to observe the bacterial EPS (Hassan and others 2002). Scanning electron microscopy has so far been the most commonly used microscopic technique. Most of the conclusions on development, composition, and distribution of biofilms and their relationship to substratum have been derived from scanning electron micrographs (SEM) (Little and others 1991; Hassan and others 2010).

Sample preparation for SEM requires extensive manipulation including dehydration, because SEM operates at high vacuum. The major challenge in studying the biofilms by SEM is that biofilms contain EPS, which contains about 95% moisture (Serp and others 2002; Hassan and others 2003). The SEM sample preparation method may thus collapse EPS and condense EPS around the cell (Donlan and Costerton 2002; Kalmokoff and others 2006). An air-drying technique was used by Hassan and others (2010) to keep the matrix intact. Similarly, biofilm samples must be coated with a conductive film before observing under SEM. Otherwise, the build-up of charges prevents the production of clear images. Elemental composition on substrate surface could be determined by energy-dispersive x-ray spectroscopy in the SEM. X-ray photoelectron spectroscopy determines the surface composition in terms of proteins, polysaccharides, and hydrocarbons (Rouxhet 1991; Rouxhet and others 1994; Christophe and others 2000).

Cleaning efficiency of existing protocols in membrane biofilm removal

Formation of biofilms is progressive. If not controlled early, the biofilm formation on membrane surfaces will impair their performance (Maxcy 1969; Mattila and others 1990). It is challenging to eliminate biofilms from all dairy/food processing equipment, because of their high resistance to antimicrobials (Simões and others 2006, 2010; Simões and Vieira 2009). Studies on cleaning of polyamide composite membranes indicate that fluid components and cleaning chemicals alter the surface properties of polyamide RO membrane. Strong bonds between bacteria, EPS, and membrane are not effectively removed by water, caustic solutions, or surfactants. This explains the complexity associated with membrane fouling and cleaning procedures being regularly followed in dairy processing units (Dunsmore 1981; Dunsmore and others 1981; Subramani and Hoek 2010). The cleaning agents remain ineffective for heavy fouling as they cannot penetrate into the foulant. Due to the complex distribution of microbes in biofilms, they are extremely resistant to disinfectants and antibiotics (Stoodley and others 2002). Among the different components of membrane fouling, some of them can be easily removed by rinsing, but some other components need specific cleaning agents to remove them from the surface. Membrane cleaning by using different chemicals and enzymes is thus an important factor for membrane regeneration.

Milk and whey membrane processing units generally operate continuously for up to about 24 h, and then are shut down for clean-in-place in order to restore production capacity and membrane permeability. A decision to shut down and clean is usually made based on flux data. Chemical cleaning is the most important method for reducing fouling with a number of chemicals being used separately or in combination depending on the nature of the foulant (Schafer 2001; Madaeni and Mansourpanah 2004; Madaeni and Ghaemi 2007; Madaeni and others 2010). However, even regular chemical cleaning of membranes will only partially restore the flux (Winfield 1979; Ridgway and others 1983). The cleaning compounds used on membrane systems are alkalis, acids, enzymes, disinfectants, and surfactants (Trägårdh 1989; Shorrock and Bird 1998; D'Souza and Mawson 2005; Madaeni and others 2010). Surfactants and alkali products are applied to dissolve food residues by reducing surface tension, denaturing proteins, and emulsifying fats (Mosteller and Bishop 1993; Forsythe and Hayes 1998; Maukonen and others 2003).

Bacterial cells attached to different surfaces as biofilm are different from their planktonic counterparts (Sternberg and others 1999; Loo and others 2000; Sauer and others 2002). Bacterial cells of the same strain attached to a biofilm matrix have different transcriptional programs as compared to the planktonic cells (Asad and Opal 2008). Based on multiple studies, bacterial biofilms have been shown to be more resistant than planktonic cells at rates up to 1000-fold against specific antibiotics (Costerton and others 1995). Due to their complex distribution, microbes in biofilms are extremely resistant to disinfectants and antibiotics (Stoodley and others 2002). Bacterial cells in biofilms were more resistant to cleaning agents than they were in their planktonic state (Stewart and Costerton 2001; Chmielewski and Frank 2003; Shi and Zhu 2009). Bacterial cells have been identified surviving at extreme conditions such as temperatures from 12 to 110 °C, and pH values between 0.5 and 13 (Lessel and others 1975; Characklis and others 1990).

Biofilm bacteria may be 150 to more than 3000 times more resistant to free chlorine (hypochlorous acid) in comparison to planktonic cells. To destroy these embedded cells, it is essential for the disinfectant to react with the surrounding polysaccharide network first. The high tolerance of mature biofilm against chlorine is due to the lower penetration power of chlorine in the matrix, only the outer surface is affected. This results in only some effect on the bacterial community on the surface (LeChevallier and others 1988; Marshall 1992). Ineffectiveness of different disinfectants on biofilm surfaces has been reported (Dolan and Costerton 2002). The main reason behind this ineffectiveness of the disinfectants might also be due to the presence of extracellular polysaccharide material within the matrix, which facilitates entrapment of the bacterial cells within the biofilm matrix.

Further, the maturation stages of biofilms provide resistance to the adhered cells even against antibiotics (Drenkard 2003). Most antibiotic substances have low penetration rates into the biofilm matrix, which might physically restrict the diffusion of antibiotics (Gilbert and others 1990; Stewart 2002). The lower susceptibility of biofilms to growth-dependent antimicrobial agents is also due to the lower growth rate or stationary state of some bacteria in the biofilm and to related changes with cell physiology (Stewart 2002; Shah and others 2006). Some microorganisms in biofilms express specific antimicrobial resistant biofilm genes that are not needed for biofilm formation. On the whole, penetration alone does not seem to be a prominent mechanism involved in biofilm resistance to antimicrobials, but this is a multifactorial mechanism that could differ from organism to organism (Patel 2005).

Bacterial cells attached with biofilms have been observed to be about 1000 times more resistant to antimicrobial stress than free-flowing bacteria of the same species. This resistance capacity depends on the type of organism and type of antimicrobial system (Lewis 2001; Mah and O'Toole 2001; Stewart 2002). Activity of microflora is enhanced with biofilm formation, which provides them with a protective shield against environmental stress such as desiccation, starvation, or exposure to heavy metals. The bacterial cells have also been observed to have greater resistance to disinfection (Keevil 2002).

Cleaning strategies for removal of biofilms

During whey processing using an RO membrane, the main deposit on the membrane surface is due to protein. Certain cleaning requirements need to exist to properly degrade these proteins from the surface. Cleaning-in-place (CIP), using general protocols and general chemicals, has been ineffective for removing attached bacterial cells (Dunsmore and others 1981; Czechowski 1990; Austin and Bergeron 1995), resulting in the formation of biofilms (Maxcy 1969; Mattila and others 1990). Chemicals used for the cleaning process may kill the attached microbes on the surface, but they leave a biomass which contributes to microbial recovery and biofilm regrowth (Simões 2005).

Chemical and physical interactions between the cleaning agents and the foulants affect the efficiency of cleaning. Determining the favorable chemical reaction with the target substances in the fouling layer is necessary for the selection of cleaning chemicals (Ang and others 2006). Proper selection of various chemicals and the correct order of cleaning steps are considered as the key points of an effective cleaning process. The concentration of the cleaning agent which provides maximum cleaning efficiency is known as the optimum concentration of the chemical.

Application of cleaning agents on membrane surfaces

Membrane cleaning includes the application of alkaline solution, acids, metal chelating agents, surfactants, sanitizers, and enzymes with a regular CIP protocol (Trägårdh 1989; Mohammadi and others 2003; Anand and Singh 2013). Mixtures of these compounds are available with commercial cleaning chemicals, but in most of the cases the actual composition remains unknown (Ang and others 2006). Of all the 6 typical individual cleaning steps, acid step was the most effective on membrane biofilms (Anand and Singh 2013). Alkaline solution allows hydrolysis and solubilization of the organic-based fouling materials by increasing the pH. It increases the negative charge and solubility of the organic foulant at higher pH (Thurman 1985). The cleaning efficiency of NaOH can be improved by increasing the concentration of NaOH, and/or with an application of more favorable physical conditions. Most polymeric membranes tolerate a limited pH range (3 to 12), hence, more favorable physical conditions can be applied (Ang and others 2006). Alkaline cleaning performed better in recovering membrane flux than acidic cleaning, due to an increase in membrane charge in an alkaline environment (Liikanen and others 2002). Repeated cycles of HCl or NaOH during cleaning of whey UF membranes revealed HCl to be more effective, mainly because of its ability to keep the pores open during the cleaning step (Norazman and others 2013). Similar findings were reported by Madaeni and Mansourpanah (2004) while observing maximum flux recovery in whey fouled RO membranes.

Chelating agents bind the metal ions from the complex organic molecules resulting in increased effectiveness of cleaning (Hong and Elimelech 1997). In a study conducted by Ang and others (2006), it was demonstrated that the ethylene diamine teteraacetate (EDTA) effectively cleaned organic fouled RO membrane, when optimized for chemical factors such as dose, and pH, and physical factors such as temperature, cross flow velocity, and time. Surfactants with hydrophilic and hydrophobic groups are semisoluble in both organic and aqueous solvents. These compounds can remove the foulant by solubilizing macromolecules by forming micelles around them (Rosen 2004). Formulated caustic agents and surfactants are useful for removal of proteins and lipids, whereas an acid blend is useful for cleaning of soluble mineral salts. Addition of a variety of enzymes provides enhanced cleaning effectiveness by breaking proteinaceous materials and polymeric foulants. In another study, Hijnen and others (2012) validated that a combination of NaOH and SDS was the most effective in cleaning RO membrane biofilms (Hijen and others 2012). Similarly, Madaeni and Samieirad (2010) showed that a combination of NaOH/SDS followed by HCl step resulted in a greater cleaning capability for eliminating foulants from RO membrane. This could be due to the increase associated with organic matter, and a simultaneous decrease in the surface tension resulting in an increased solubility of the foulant.

Factors responsible for effective and efficient cleaning

For an effective cleaning process of membrane surfaces, various physical and chemical methods, or a combination of both, can be applied. A physical method involves mechanical treatment, whereas chemical treatment methods are mainly related to multiple numbers of chemical reactions with membrane surfaces (Madaeni and Mansourpanah 2004). Chemical treatments using oxidizing or nonoxidizing biocides and physical treatments involving scrubbing and hot water are very important for the removal of biofilm. Chlorine, chlorine dioxide, ozone, and peroxide are considered oxidizing biocides, whereas quaternary ammonium compounds and formaldehyde are considered nonoxidizing biocides. In another study, NaOCl and a combination of H2O2/SDS showed higher removal of biofilms on RO membrane (Hijnen and others 2012). The cleaning process incurs huge costs associated with the consumption of time, energy, chemicals, and water. An effective cleaning process will succeed in removing deposited material from the membrane surface and restoring normal flux. With optimum cleaning parameters, membrane performance will be improved, as well as operational costs most likely reduced. Therefore, it is essential to have an effective, easy, and fast cleaning operation (Daufin and others 1991).

Cleaning efficiency depends on type of cleaning agent, cleaning solution pH, pressure, cleaning agent dose, optimum cleaning time, cross-flow velocity during cleaning, and cleaning solution temperature (Ang and others 2006). Cleaning for an extended duration may not be sufficient to remove strongly adsorbed fouling materials (Madaeni and others 2010). The chemicals used for cleaning may also affect the membrane surface causing swelling of the membrane and resulting in lower cleaning efficiency. The concentration of peptides increases with a higher concentration of acid or alkalis, resulting in decreased cleaning efficiency (Madaeni and others 2010). Cleaning agents remain ineffective for removal of heavy fouling when they cannot penetrate into the foulant.

Effective cleaning and sanitation of the membrane system also depends on mechanical design, turbulence, and nature of various chemicals. Chemical, mechanical, enzymatic, and a combination of mechanical and enzymatic methods have been developed to detach the bacterial aggregates from the surface (Bockelmann and others 2003). Studies based on countermeasures for biofilm have suggested the application of a 3-step protocol including detection, sanitation, and prevention of the biofilms (Flemming and others 1997).

Membrane cleaning using a combination of chemicals and enzymes is an important technique for membrane regeneration. While some components involved in membrane fouling can be easily removed by rinsing, other components need specific cleaning agents to remove them from the surface. Based on differing components of the feed and the nature of membrane fouling, suitable cleaning agents can be selected (Lindau and Jonson 1994). The nature of the cleaning agent should be such that it can dissolve the majority of the fouling material and can be removed from the surface without affecting the membrane surface (Munoz-Aguado and others 1996).

Role of surfactants in the cleaning process

Different forms of surface-active compounds have been used for preventing bacterial attachment to surfaces. Surfactants provide uniform wetting of the surface promoting additional cleaning effects (Cloete and others 1992; Lutey 1995). The surfactant binds the ionic sites on the proteins with electrostatic and hydrophobic interactions. Surfactants can be classified into 4 groups based on the nature of each hydrophilic group, namely, anionic, cationic, nonionic, and zwitterionic. Anionic surfactants interact with whey proteins to decrease the surface tension of molecules in contact with each other. The surface-active compounds diffuse between the structures of proteins and reduce their hydrophobic interaction with the membrane surface, which causes reduction of fouling (Madaeni and others 2010). In another study on organic fouled RO membrane, Ang and others (2006) determined the anionic surfactants to be efficient cleaning agents when used under favorable physical and chemical conditions. Cationic surfactants affect interchain hydrophobic bonding and result in an extended structure. Nonionic surfactants affect the hydrophobic interaction with nonpolar patches on the proteins and denature the tertiary protein structure along with reducing the hydrophobic interaction of proteins with the membrane surface (D'Souza and Mawson 2005).

Surfactants prevent the attachment of microorganisms and the formation of biofilm (MacDonald and others 2000) by reducing the surface tension of water by adsorbing at the liquid–gas interface as well as by reducing the interfacial tension between the layers (Whitekettle 1991). They also contribute toward the detachment of the microbes from the surface. One group of surfactants is capable of attaching to hydrophobic surfaces and another group of surfactants to hydrophilic groups or the aqueous phase (Rosen 1987). Surfactants also modify the bacterial cell surface charge.

Degradation of biofilm by using enzymes

Surfactants and detergents neutralize the charged colloidal particles and resuspend the particles, whereas enzymes hydrolyze the proteinaceous and glycoprotein exopolymers in which the microbes are embedded (Whittaker and others 1984). Enzyme-based formulas contain improved cleaning properties, are compatible with a lower cleaning temperature, and also are not considered pollutants. Different classes of enzymatic cleaning operations provide neutralization of cleaning effluents and biodegradability (Farone and Cahn 1970). Enzymes, when used as cleaning agents can reduce cleaning time. Enzymes used alone as a cleaning agent is not considered capable of breaking the biofilms due to their large molecular weight. Therefore, it is advisable to mix an anionic detergent with the enzyme to increase its performance (Coolbear and others 1992). Surfactants with chelating agents might be added with the enzymes to penetrate the biofilm matrix (Whittaker and others 1984). Purified enzymes and detergents with milder and environmentally friendlier cleaning protocols are considered to be the best chemicals for the removal of biofilms from polymeric membranes (Maartens and others 1996; Munoz-Aguado and others 1996). Application of enzymes alone or in combination with biodegradable detergents has been studied as an effective tool for reducing fouling (Maartens and others 1996, Leukes and others 1999). Enzymes can work with mild pH, temperature, and ionic strength without affecting the membrane surface (Maartens and others 1996, 1998). Also, the enzyme-based reactions are considered to be very specific on their substrates.

Utilization of a combination of enzymes (especially proteases and polysaccharide-hydrolyzing enzymes) may be considered effective for the removal of the biofilm matrix from a membrane surface (Meyer 2003). The QuatroZyme, which is a combination of several enzymes (lipase, protease, cellulose, and amylase), was found to be slightly better than other enzymes. It was also shown that using an enzyme cleaner followed by sanitizers would be more effective in removal of biofilms (Tang and others 2010). Some studies have used the application of α-glucosidase, β-galactosidase, and lipase for the degradation of EPS structures present in soil particulates (Bockelmann and others 2003). Scanning electron microscopy images showed convincingly the effect of enzymatic treatment as detachment of bacteria from soil particles. Another study based on the effect of commercial enzymes on marine biofilm found the application of savinase, among other enzymes, for the prevention of bacterial adhesion and the removal of adhered bacteria to be effective (Leroy and others 2007). The supernatant produced by a marine biofilm isolate (Bacillus licheniformis) was able to disperse the bacterial biofilm (Nijland and others 2010). Studies dealing with solubilization of complex primary sewage sludge concluded that celluloses, lipids, proteins, and polysaccharides can be hydrolyzed by using lipases, proteases, and α-and β-glucosidases (Whittington-Jones 1999).

Control of fouling can also be done via other techniques such as pretreatment of whey, maintaining the operating conditions such as moderate pressure, cross-flow, backwashing, and membrane regeneration, or cleaning the membrane. Studies indicated that sonication can be used for the removal of cake formed on the surface (Lim and others 1971; Popovic and others 2010). Turbulent flow, under optimum conditions, also results in higher cleaning efficiency (Madaeni and others 2010). Application of electrolyzed water may also be useful for decontamination because of its easy production and no need of high temperature for its operation (Mahmoud 2007). On-site generation by simple hydrolysis of a dilute salt solution helps to reduce the cost and hazards associated with handling, transportation, and storage of concentrated chlorine solution (White and others 2010).

Role of cellular signals in biofilm formation

Quorum sensing (QS) is a term used to describe cell-to-cell signaling or intracellular signaling in bacteria; it is a process of chemical communication used to check the species proportion and cell count among the local population. Bacteria have advanced directive mechanisms by which they can act together to simultaneously express a precise set of genes in response to variations in cell population density. QS thus involves producing, releasing, detecting, and responding to small hormone-like signal molecules called autoinducers, whose concentration increases as a function of cell density. Bacteria detect the minimal threshold stimulatory concentrations of these autoinducers and this leads to an alteration in gene expression. This process helps microbial cells with the formation of mature biofilms, an ability to monitor the environment, which allows them in their development and survival in the complex biofilm matrix. (Miller and Bassler 2001; Paresek and Greenberg 2005; Waters and Bassler 2005; Kociolek 2009).

Cell-to-cell signaling also performs a huge role in regulating various physiological processes, like formation or aggregation and dispersal of biofilms across far-off genera of bacteria. There are specific signals for biofilm dispersion and formation such as diffusible signal factor that act as an environmental signal in Xanthomonas campestri to trigger biofilm dispersal, and N-acylhomoserine lactones (N-AHLs) have been involved in the regulation of biofilm formation in a few bacteria (Dow and others 2003; Petrova and Sauer 2012). Individual bacterium-initiated QS-controlled processes (which act alone) are mostly ineffective, but when implemented together by a larger number of cells become useful. Thus QS enables bacteria to act as multicellular organisms. Two regulatory proteins, LuxI and LuxR, participate in the synthesis and recognition of the autoinducer, respectively, and manage expression of many genes engaged for bioluminescence and production of pigments or antibiotics (Jayaraman and Wood 2008; Bai and Rai 2011). There are 4 main classifications in the cell-to-cell signaling systems: autoinducer-1(AI-1), autoinducer-2(AI-2), autoinducer-3(AI-3), and autoinducing polypeptide (AIP). Autoinducer-1 (AI-1) and Autoinducer-2 (AI-2) types of QS processes are involved for intraspecies and interspecies interaction, respectively (Farah and others 2005). Cell-to-cell signaling systems for Gram-negative bacteria are possible by using autoinducer-1 (AI-1) and by autoinducer-3 (AI-3), whereas AIP is responsible for Gram-positive bacteria. Autoinducer (AI-2) is found in Gram-positive bacteria as well as in Gram negative cells (Bai and Rai 2011).

QS initiates phenotypic changes by modulation of gene expression in bacteria and thus helps in their growth by providing better response to the environmental conditions (Turovskiy and others 2007). QS affects regulation of virulence, toxin production, exopolysaccharide production, development of conjugative plasmids, sporulation, formation of biofilm, synthesis of antimicrobial peptides, symbiosis, and motility (Smith and others 2004). Deterioration of food due to proteolytic, lipolytic, chitinolytic, and pectinolytic activities are known to be regulated by QS (Bai and Rai 2011). Bacterial cells of the same strain attached to a biofilm matrix have different transcriptional programs as compared to the planktonic cells (Asad and Opal 2008).

Interference with QS can provide an alternative approach to control biofilms on different surfaces such as controlling the production of EPS and biofilm formation on different surfaces (Davies and others 1998; Von Bodman and others 1998; Rivas and others 2005; Waters and others 2005). The discovery of a quorum quenching (QQ) mechanism in biofilms has provided a novel opportunity to control the activity of microorganisms without the utilization of growth repressive agents like chemicals, disinfectants, and antibiotics (Manefield and others 2002). The extensive use of these growth-repressive agents has led to an evolution of “superbugs” that can withstand the conventionally used inhibitory agents and have shown to be damaging certain membrane material due to residual concentration. This has emphasized the development of novel strategies against unwanted microorganisms (Dong and others 2007). The QS interference approaches have many advantages of lower toxicity or nontoxicity, higher antibiofouling capability, low risk of bacterial resistance advancement, and eco-friendly substances (Xiong and Liu 2010). QS is thus a biochemical path to directly control the rate and extent of biofilm development, rather than detaching biofilm after deposition by physical or chemical ways (Kim and others 2009).

QS prohibitors or quorum quenchers are generally counterpart of the AHLs or compounds that can break down AHLs (Bai and Rai 2011). Many sets of QQ chemicals and enzymes have been determined as effective. Among the potent quorum quenchers are the halogenated compounds secreted by the marine seaweed Delisea pulchra, the artificial compounds that target R proteins, the artificial AHL, and the AIP counterpart that can fight with analogous QS signals, and also the AHL-degradation enzymes like AHL-acylase, AHL-lactonase, and paraoxonase (Dong and others 2007). Some other studies on quorum quenchers such as D-tyrosine (D-amino acid used as a biological control) and N-acetylcysteine (a nonantibiotic mucolytic agent) showed that they prevented irreversilbe biofouling by supressing or restricting the adhesion of bacteria and biofilm formation on NF and RO membranes, respectively (Kappachery and others 2012; Yu and others 2012)

Biofilms after treatment with quorum quenchers like furanones were efficiently destroyed upon chemical termination. Determination of crystal structures of different kinds of QQ enzymes has contributed beneficial hints that can help to interpret the catalytic mechanisms, protein tailoring, and molecular advancement (Dong and others 2007). This implies that biofilms treated with quorum quenchers are inclined to be more sensitive to bactericidal chemicals. A complete removal of biofilms could thus be achievable by a combination of QQ technique and chemical termination technique (Paul and others 2009). Further studies are necessary to examine the practicality of quorum quenchers with industrial-scale membrane systems for biofouling control. Major cleaning applications are summarized in Table 1.

Table 1. A summary of cleaning approaches to control membrane biofilms in dairy and water filtration processes
Membrane typeCleaning approachEffective treatmentReference
ROIndividual steps:Acid at pH 2.1, 30 minAnand and Singh (2013)
 Alkali, Surfactant, Acid, Enzyme, Sanitizer  
ROCombinations:NaOH/SDS (1%) at 12 pH, 20 °C;Hijnen and others (2012)
 Alkali, detergents, enzymes, chelating agents, acids, biocidesNaOCl and H2O2 (0.5%)/ SDS at pH 11 
RON-acetylcysteine at 1.5mg/ml Kappachery and others (2012)
UFSingle, double steps in repeated cycles: NaOH, HCLIn repeated cycles: HCl at 0.1 M, pH 1Norazman and other (2013)
NFD-tyrosine at 3 μM Yu and others (2012)
ROIndividual steps, combinations, 2 stages:NaOH (0.075%) + SDS (0.3%) at 25 °CMadaeni and Samieirad (2010)
 Acids, base, EDTA, surfactant.2nd Stage: HCl (0.5%) 
UFIn CIP single step replacement:QuatroZyme at 0.3%, 48 °C, pH 7.0 to 8.0, 30 min.Tang and others (2010)
 Enzymes: Reflux E2001, E1000, QuatroZymeMIOX EW anolyte at 20 °C, 10 min. 
 In CIP additional step:  
 Sanitizers : MIOX EW anolyte, Sodium hypochlorite, Perform  
ROIndividual steps: cids, base, EDTA, surfactant.HCl at 0.05% and pH 3Madaeni and Mansourpanah (2004)
ROIndividual steps:SDS 10 mM, pH 11, 60 min;Ang and others (2006)
 EDTA, SDS, NaOHEDTA 2.0 mM, pH 11, 40 °C 
ROCombinations: haotropic agents, bactericidal agents, enzymes, antiprecipitants, surfactants, and detergentsEnzyme-chelator-dispersant combination, Anionic detergent-denaturant combination, and Biz for 30 to 60 min.Whittaker and others (1984)


Membrane processes are now widely used in the dairy industry and the diversifications in product manufacture and greater demand have led to heavy use and long durations of continuous operations. The commencement of biofilm development is influenced by the physical and chemical components of membranes, characteristics of prior attaching bacteria, and the operational parameters of the membrane system. This provides a favorable condition for the growth of microflora and the development of irreversible membranes. The membrane processing temperatures are further conducive to biofilm development. Such biofilms enter the irreversible phase, generally after 8 to 10 h, and then are very difficult to clean by the regular cleaning and sanitation protocols. Studies have demonstrated recalcitrant biofilm development on processing membranes. This results in operational issues, such as difficulties in flux restoration, premature replacement of membranes, reduction in productivity, biodegradation of membrane material, increase in power consumption for bringing up operation pressure, increase in cost of cleaning, and several quality and safety issues. The industry is finding it increasingly difficult to deal with membrane biofilms in an effective and efficient way. The complex procedures for establishing the presence of biofilms on membrane surfaces, and the ineffectiveness of cleaning regimens further make the situation difficult. Future intensive research on membrane biofilms needs a discerning effort to understand the bacterial community that prevails in membrane processing plants. Much research is currently being focused on modifying cleaning protocols by changing the cleaning conditions and including enzymes in the protocols to effectively break down the complex biofilm matrices, and to modify sanitizers for killing the persistent bacteria in the embedded form. A recent approach in this area is to understand the role of quorum signals, in the development of biofilms on membranes, and the use of quorum inhibitors to prevent the colonization and formation of biofilms. The research in this area is currently in its infancy, and it will be a while to introduce food grade quorum inhibitors as an intervention strategy for the prevention of biofilm formation on membranes.


The financial support provided by the Midwest Dairy Foods Research Center for writing this review is acknowledged.

Author Contributions

Sanjeev Anand: Conceived, compiled, drafted, and reviewed the manuscript.

Diwakar Singh: Researched a part of the prior studies, helped in writing.

Mallika Avadhanula: Researched a part of the prior studies.

Sowmya Marka: Researched a part of the prior studies, helped in writing.