[Front page] [Contents] [Previous] [Next]

Cross-flow filtration of fruit juice

2. Theory

2.1 Sour cherry and Black currant juices
2.1.1 Cherries and cherry processing
2.1.1.1 Origins and botanical data
2.1.1.2 Anatomy, physiology and composition
2.1.1.3 Production figures
2.1.1.4 Harvesting, Handling and Processing
2.1.1.5 Applications
2.1.1.6 Cherry juice
2.1.2 Black currants and processing
2.2 Membrane processes
2.2.1 Microfiltration
2.2.2 Membranes
2.2.3 Factors that influence the permeate flux during filtration
2.2.3.1 Transmembrane pressure
2.2.3.2 Linear or cross-flow velocity
2.2.3.3 Temperature
2.2.4 Flux decline reasons
2.2.4.1 Concentration polarisation
2.2.4.2 Membrane fouling
2.3 Juice filtration
2.4 Fouling & Cleaning
2.4.1 Fouling analysis
2.4.2 Methods to reduce fouling
2.4.2.1 Backflush and Backshock techniques
2.4.2.2 Chemical cleaning
2.4.2.3 Ultrasounds

2.1 Sour cherry and Black currant juices

2.1.1 Cherries and cherry processing

2.1.1.1 Origins and botanical data

The cherry is part of the family Rosaceae and belongs to the genus Prunus. Other members of this genus include apricot, nectarine, almond, peach, and plum. Cherries have traditionally been used for a wide variety of food and beverage products in Europe, and with European migration this tradition has spread to North America and to many other parts of the world (Kaack, 1990).

The red cherry (Prunus cereus L.), also know as sour, tart or pie cherry, is a drupe fruit and originated in the territory between Switzerland and the Adriatic Sea on the west of the Caspian Sea and northward on the east (Hedrick, 1914).

Although there are 270 named varieties of red cherries, only a few are grown for commercial purposes, and therefore are well known. Based upon the colour of juice in the fruits, these can be divided into two groups. Those giving a colourless juice are known as Amarelles while those with a darker colour and a reddish huice are known as Morellos. The Morellos are more acid and sour (Webster & Looney, 1996).

Nutrition

Nutritional studies have shown that red or sour cherries are a good source of Vitamin A, calcium, iron, potassium, and phosphorous. More detailed nutritional information is given in Table 2.1.

Table 2.1. Nutritional Analysis of 100 g Fresh Sour Cherries (Cash et al., 1989).

Nutrient and Units

Red cherry

Water, g

82.53

Calories, kcal

63

Protein, g

1.08

Lipid, g

0.22

Carbohydrate, g

15.78

Fibre, g

1.48

Ash, g

0.39

Calcium, mg

17.2

Iron, mg

0.44

Magnesium, g

11.1

Phosphorous, mg

20.1

Potassium, mg

185.5

Sodium, mg

0.02

Zinc, mg

0.12

Cooper, mg

0.15

Manganese, mg

0.10

Ascorbic acid, mg

9.8

Thiamin, B1, mg

0.04

Riboflavin, B2, mg

0.04

Niacin, mg

0.19

Pantothenic acid, mg

0.16

Pyridoxine, B6, mg

0.05

Folacin, m g

7.5

Vitamin B12 , m g

0

Vitamin A, IU

850

Cyanidin-3-rutinoside and cyanidin-3-glucoside are important anthocyanin pigments in Morello sour cherry (Hong & Wrolstad, 1990).

2.1.1.2 Anatomy, physiology and composition

Although characteristics of berry and stone fruits differentiate, the anatomy and physiology of these fruits are similar. Three different tissues can be differentiated: exocarp or skin, mesocarp or fruit flesh, and the endorsperm or section of seeds or stones.

Skin

The skin provides protection and often contains small amounts of valuable juice.

Fruit flesh

The mesocarp consists of very big cells, which contain nearly all characteristic and desired components.

Seeds

The seeds or stones usually contain large quantities of tannins. Therefore, destruction during processing is avoided. An exception is the production of cherry juice where the destruction of approximately 10% of the stones is desired for sensoric enhancement (Hamatscheck et al., 1995).

Sour cherry juice contains a lot of small compounds such as phenols, proteins and different kinds of polysaccharides. All these compounds make the filtration of sour cherry juices difficult.

Phenols

The phenols in sour cherry juice are mainly hydroxycinnamic acids and anthocyanins (Macheix et al., 1990). The anthocyanins are responsible for the red/purple colour of the juice (Grassin and Fauquembergue, 1996). The amount of phenols in the juice depends on the species, the ripeness of the berries, place of growth, and method of production (Lee, 1992).

2.1.1.3 Production figures

Both the sweet and the sour cherry are deciduous trees originating around the Caspian and the Black Seas, and therefore thrive best in areas with a temperate or Mediterranean-type climate. Cherries are known to be produced in significant quantities in more than 40 nations of the world (Webster & Looney, 1996).
Total cherry production is close to 1.4 million tonnes and about 71% of this is produced in Europe. In 1987 10,000 tonnes of sour cherry were produced in Denmark and 50% were processed in different products like juice, wines, jams and jellies.
Nowadays, next to apples sour cherries is one of the most important fruits grown in Denmark.

2.1.1.4 Harvesting, Handling and Processing

The principal red cherry varieties are self-fertile when the conditions are proper (warm weather, adequate bee activity, required pollen tube growth, and proper nutrition). In selecting the proper location for a red cherry orchard, it is necessary to consider which sites are more susceptible to spring frosts, since cherries are very vulnerable to freezing three to four weeks before full bloom. Winter temperature should not go below –26°C, and the mean temperature for June, July, and August should be about 15°C. The blossoming period is sometime between mid-April and mid-May, depending on the location. The fruit ripens in late June to early July and may be harvested as late as mid-August in the more northern regions.

Time of harvest

Harvesting of cherries should be carried out at the optimum maturity stage for each commodity.
Sour cherries grown especially for processing are harvested when the various flavour components reach a minimum treshold level and this level is commonly different from that used for fresh market fruit. For nearly all cherry products, the balance of soluble solids (sugars) and free acids is very important and, partly because of ease of measurement, is the most commonly used indicator of harvest maturity. It is usually assumed that when the sugar/acid ratio is right the other components of flavour will also be present. The normal range of several processing-related chemical constituents in sour cherries is shown in Table 2.2.

Table 2.2. Chemical composition of sweet and sour cherries expressed as percentage of fresh weight.

 

Sour cherry

Soluble solids

12-17

Sugar

6-9

Non-soluble solids

1-4

Total acid

1.5-1.8

Minerals

0.5

Fibre

0.5

Harvesting

For many years, the harvesting and sorting of red cherries relied on workers, however, nowadays, the machines have simplified and speeded the harvesting operation as well as improved the processing. Mechanical harvesting is successfully carried out with the help of a trunk shaker. This machine shakes each trunk with a few firm movements loosening the fruits and dropping them into inclined frames. In this way, the fruits roll onto conveyor belts and then to tanks of cold water in order to be washed, to remove spray and to promote firming.

Sorting

As a first step of processing, the cherries go through an eliminator, where foreign matter and defective, undersized and immature fruit is removed. The next step is the sorting and an electronic sorter has taken the place of hand sorting. Once graded, the cherries move through mechanical pitters in order to separate them from the pits. From there, the cherries move along belts and are inspected again for blemishes, loose pits and foreign matter.

Cherries will be processed differently according to the final product.

2.1.1.5 Applications

Sour cherries are primarily grown for use in processing; they are used for juice and other beverages, nectar, jams and jellies, yoghurt, and are canned and frozen whole (usually pittet) or further prepared as fillings for use in a wide variety of bakery products (Poll, 1986). Small quantities of sour cherries are dehydrated and in some countries speciality wines are made from sour cherry juice. Furthermore, there is a considerable interest in using anthocyanin pigments from cherries as natural sources of food colorants (Chandra et al., 1993).

2.1.1.6 Cherry juice

Cold pressing frozen cherries

Juices from berries and stone fruits are an important product of the fruit juice industry as beverage products and as ingredients for the soft drink, spirit and candy production (Hamatscheck et al., 1995).

The cherry juice is the only deciduous fruit yielding bright red juice, but its strong flavour results in a limited commercial use compared to orange, grape and other popular juices. Consumers preferred when diluted or blended with another juice such as apple juice.

The most important prerequisite for the production of a high quality product is an optimal ripe, unimpaired, sound and clean raw fruit (Schobinger et al., 1987), and the two main quality demands in juice processing are to preserve the organoleptic quality and clarify the juice for storage.

Good quality

Good quality cherries should be used to produce juice: overripe or spotted cherries produce a juice with a higher benzaldehyde flavour which is considered undesirable; mushy fruit can result in a juice with so much suspended material that filtering becomes difficult and the juice will become cloudy during storage. On the other hand, less mature fruit results in a juice, which is low in sugar, poor in colour and lacks some soluble components that contribute positively to the flavour.

Assessing the suitability of raw fruit for juice processing mainly involves determining juice-soluble solids, acidity and anthocyanin-content. However, the incidence of cracking and decay, uniform maturity and the estimation of the yield of juice by pressing are also very important considerations. Juice yield depends mainly on the percentage of fruit flesh.

In industrial sour cherry production, the berries are milled after washing and sorting. Then, a mashing treatment is necessary to increase the juice yield and extract other fruit components such as colour and flavour. Next comes the pressing stage. Applied pressure is very important: the recommended procedure is to start with a low pressure and gradually increase. There are three methods for pressing cherries for juice: hot pressing, cold pressing and cold pressing of frozen cherries.

Hot pressing

Hot pressing is the simplest procedure. Pitted fruit is heated to about 66.5°C and the cherries are pressed before they cool. The resulting juice is strained, chilled, allowed to settle, and then filtered. Juice is a deep red colour since the heating extracts a greater level of the pigments present. However, the flavour is more similar to that of canned cherries rather than fresh fruit.

Cold pressing

Cold pressing yields a juice not rich in colour, but that has a flavour similar to fresh cherries. In this procedure the cherries are coarsely ground or pulped to aid in the extraction of the colour. The ground fruit is then pressed and the expressed juice is flash heated, cooled to inactive enzymes and reduce the microbiological load, clarified and filtered.

Cold pressing frozen cherries combines the best features of the hot and cold press methods. A dark juice is obtained, and its flavour resembles the fresh cherries. The cherries are either pitted or crushed and then frozen with or without sugar added. Prior to pressing they are thawed to a temperature of 4.5-10°C. After pressing, the juice is clarified and filtered (Tressler & Joslyn, 1971).

Centrifugation

Berry fruits can also de-juiced by means of a decanter after crushing. Centrifugal force is the effective principle, whereby liquid and solids are separated due to the different density. Separators and decanters are an indispensable integral part of modern technologies for the production of berry and stone juices since they completely avoid product looses and improve quality.

Decanters are an alternative to conventional pressing techniques and both methods are comparable regarding yield and sensoric quality. However, a slight increase in phenolic compounds is observed when using the decanter (Hamatscheck et al., 1995).

The prior removal of the stones is not required in case of juice extraction with the decanter. Depending on the pre-treatment of the fruits and the technology applied for phase separation, the recovered juices contain a certain quantity of cloud particles, which will be partly or totally removed during the further processing stages.

After the pressing stage, the cloudy juice leaving the press is screened to remove any coarse material. Then, an enzyme treatment is necessary for viscosity reduction and better clarification (Baumann, 1981).

Aroma development can be enhanced or suppressed during processing. With enzyme- or acid-catalysed conversion of amigdalins, occurring in the seeds and the fruit flesh, glucose, hydrocyanic acid and benzaldehyde are released. The latter is a very important aroma component in cherry. Because of these reactions, a pleasant cherry flavour can be obtained if about 20% of the stones are crushed. This is a practice often followed in the production of juices and wine.

Enzymatic treatment

In most cases, enzymatic decomposition of pectins is of advantage for an economic juice recovery. For juice extraction from fruits and vegetables, the cell walls have to be ruptured at least at one point. In practice, a combination of mechanical and enzymatic procedures achieve this, and in individual cases, this is supported by heat temperature treatment.

During depectinisation, pectic enzymes hydrolyse pectin and, thereby, reduce the viscosity of the juice, making it more easily filtered. In addition, since pectin is a high molecular weight colloid, which acts as a protective colloid in suspending the particles in sour cherry juice, these particles are released when it is hydrolysed and settle to leave the supernatant juice clear. Thus, the amount of gelatine used during the later clarification can be decreased.

There are two possibilities for using enzymes in juice processing: mash treatment and juice clarification (Höhn, 1996). Because of the low content of pectins in sour cherry, enzyme treatment before pressing is not always necessary. However, it is possible to obtain better quality and higher yield by treatment with pectin-degrading enzymes at 40-50°C (Baumann & Gierschner, 1974). Adding pectolytic enzymes to the mass prevents the formation of a pectin gel and increases the juice yield dramatically. The enzymes that are used for processing berry fruits have to meet special requirements, which are caused by the production technology and the character of the raw material. The mash is usually heated after crushing and special tubular heat exchangers are used for this production step. Heating the mash has several effects:

Improved extraction of the cell wall material
Better colour extraction
Support of the enzyme reaction by adjusting the temperature to the level at which the enzyme has its maximum activity

Cherry juice is generally clarified and filtered (Sahin & Bayindirli, 1993).

Clarification

In the clarification of sour cherry juice, the amount of gelatine used is very important as its use other than in low quantities may cause colour loss. During clarification, suspended particles in the juice settle. The reason for this precipitation action is the charge difference between the colloidal material in the juice and the clarifying agent. Settling of suspended particles makes the juice clearer and more easily filtered.
Depectinisation and clarification decrease the resistance during filtration (Shain & Bayindirli, 1993)

Current filtration of a wide variety of juices is accomplished using the minerals Diatomaceous Earth and Perlite (filter aids). The semi-concentrate is normally polished by means of ultrafiltration (McLellan, 1996).

If enzymatic treatment of the mash is carried out before phase separation, juice yield and clarification efficiency are increased. Subsequently, the product is de-pectinised, de-aromatised, and pre-concentrated.

During industrial processing, a small amount of anthocyanin is degraded but, more importantly, the juice contains much less benzaldehyde and cyanide (from the crushed stones) than is found in the macerated fruit (Table 2.3). Only about 18% of the benzaldehyde and cyanide is extracted to the juice.

Table 2.3. Composition of raw fruit and industrial processed juice from sour cherries (Kaack, 1990).

 

Soluble solids
(%)

Acid
(%)

Anthocyanin
(mg 100g-1)

Benzaldehyde
(mg kg-1)

Cyanide
(mg kg-1)

Raw fruit

21

1.9

204

93

34

Juice

21

1.6

181

17

6

2.1.2 Black currants and processing

Black currants contain large amounts of ascorbic acid (vitamin C), pectin and phenols. Pectin, which is primarily found at the cell wall, has the ability to form stable gels with water and enzyme treatment is therefore necessary to achieve an optimal pressing of the berries (Pilnik & Voragen, 1991).

Phenols

The phenols in black currants are mainly anthocyanins, which are responsible for the red/purple colour (Grassin & Fauqembergue, 1996). The amount of phenol in the juice is depending on the black currant species, the ripeness of the berries, place of growth, and method of production (Lee, 1992).

2.2 Membrane processes

Membrane technology is still evolving and finding more and more applications in food and pharmaceutical processes (Mulder, 1991) and the development of membranes will strongly influence separation processes in the future. Various pressure driven membrane processes can be used to concentrate, sterilise or purify aqueous solutions.

Development of membrane technology

Membrane technology can work as well or better than the existing technology regarding product quality, energy consumption and environmental issues. The costs of this technology are not currently at a level, which will make the implementation attractive for all applications, but this is on its way. Membrane technology demands that basic research within material science is coupled to the understanding of problems related to the specific industrial process where the membrane module is to be integrated. Too often research at laboratory scale shows promising results, but a stronger involvement of industry is necessary in order to develop the membrane to commercial level and to promote the incorporation of membrane modules in a process together with other unit operations. The argument for doing so is the obvious advantages of this technology: cleaner and simpler process solutions, less chemical additives and lower energy consumption. The demand on industry for better environmental solutions and cleaner technology is also pushing the development and implementation of membrane technology (Hägg, 1998).

Basic principles

In membrane separations, each membrane has the ability to transport one component more readily than other because of differences in physical and/or chemical properties between the membrane and the permeating components. Furthermore, some components can freely permeate through the membrane, while others will be retained. The stream containing the components that permeate through the membrane is called permeate and the stream containing the retained components is called retentate.

Driving force

Transport through the membrane occurs as a result of a driving force acting on the individual components in the feed and usually the permeation (selective transport) rate through the membrane is proportional to the driving force.

As it is seen in Table 2.4, membrane processes can be distinguished according to the type of the driving force that ensures the transport through the membrane.

Table 2.4. Classification of membrane processes according to their driving forces (Mulder, 1991).

See table 2.4

2.2.1 Microfiltration

Microfiltration is the membrane process, which most closely resembles conventional coarse filtration. The characteristics of microfiltration are shown in Table 1.5.

Table 2.5. Summary of Microfiltration (Mulder, 1991).

  Microfiltration
Membrane Symmetric or asymmetric pores
Thickness 10-150 mm
Pore sizes 0.05-10 mm
Driving force Pressure (< 2 bar)
Separation principle Sieving mechanism
Separation goal Solution or gas free of particles
Applications - Analytical applications
- Sterilisation (foods and pharmaceuticals)
- Ultrapure water (semiconductors)
- Clarification (beverages)
- Cell harvesting (biotechnology)

Dead-end vs. Cross-flow filtration

There are two possible methods of operation in filtration processes: Dead-end and Cross-flow filtration. The simplest method used is the Dead-end filtration, where the feed flow is perpendicular to the membrane surface. It is forced through the membrane, which causes the retained particles to accumulate and form a type of cake layer at the membrane surface. The thickness of the cake increases with filtration time. The permeation rate decreases, therefore, with increased layer thickness. However, to reduce fouling, Cross-flow microfiltration is generally used. The feed flows parallel to the membrane surface and part of the retained solutes accumulates. The feed composition inside the module changes as a function of distance in the module, while the feed stream is separated in two: permeate (filtrated product) and retentate (unfiltrated product) stream.

2.2.2 Membranes

General

A membrane can be defined as an interphase that separates components according to their structure. In a more general way, a membrane is a permselective barrier through which fluids and solutes are selectively transported when a driving force is applied across it.

The first membranes were produced in Germany in 1920 and used as filter for bacteria at laboratory scale. Since then, the membrane technology has been significantly developed. Nowadays, membranes are manufactured from a wide range of materials and they can offer a good selectivity, a high permeability and a considerable chemical stability.

Structure

The structure of the employed membrane is chosen according to the particles or molecular size, shape and chemical properties of the feed solution. Membranes can be classified as either biological or synthetic according to their nature.

Regarding to morphology or structure, solid synthetic membranes can be classified as symmetric or asymmetric.

Symmetric structure

In the symmetric membranes, the diameter of the pores is almost constant through cross section of the membrane. Furthermore, the entire membrane thickness causes resistance to mass transfer acting as a selective barrier.

Asymmetric structure

In asymmetric membranes, the pore size at the surface is smaller, so only the thin top of the layer determines the selective barrier. Large particles will not enter in the body of the membrane. In this way, the plugging of the membrane is avoided. These membranes combine the high permeation rate of a very thin membrane with the high selectivity of a dense membrane.
Selectivity is mainly the result of the pore sieving action, but it is also caused by hydrophilic-hydrophobic interactions and membrane charge. The smaller the pore sizes the better the selectivity. Nevertheless, selectivity shows a certain variation since the pore sizes are not uniform.
Asymmetric membranes are normally composed of two layers, a support layer and a skin layer. The support or porous layer has high porosity, no selectivity and a thickness of 50 to 200 mm. The skin or top layer is very thin (0.1-2 mm), and it is responsible for the membrane selectivity. Asymmetric membranes can be classified as normal or reverse. In normal asymmetric membranes the skin layer faces the feed solution and in reverse asymmetric membranes the porous layer faces the solution.

2.2.3 Factors that influence the permeate flux during filtration

2.2.3.1 Transmembrane pressure

Transmembrane pressure (TMP) is the driving force of the pressure driven membrane processes, and it is defined as the pressure difference between the retentate and the permeate side:

Dead-end filtration
Cross-flow filtration

where

Pin
is the inlet pressure
Pout is the outlet pressure
Pp is the permeate pressure

The permeate flux increases with the TMP, but the flux decreases with increasing resistance of the membrane. The relation between the flux and the membrane resistance is given by the Hagen Poiseuille´s equation:

J= Pi . TMP

where

J
is the permeate flux
Pi is the permeability coefficient
m is the viscosity
Rtot is the total resistance of the membrane

Permente flux

The permeate flux increases with the TMP but the relation between them is only linear when the feed is pure water. If the feed is another product, the flux becomes independent of the pressure and mass transfer controlled when the pressure increases above the level, where the concentration polarisation layer reaches a limiting concentration (see Fig.2.1).

See figure 2.1

Figure 2.1.Generalised correlation between operating parameters and flux, indicating the areas of pressure control and mass transfer control (modified from Mulder, 1991).

Gel layer

The gel layer model explains the flux independence from the pressure, where the main responsible for the limiting flux is the formation of a gel layer of a fixed concentration.
When the pressure is increased above this limit, a compaction of the gel layer occurs and, consequently, the flux does not increase. Field et al. (1995) introduced the concept of critical flux hypothesis for Microfiltration. According to this concept a critical flux exists, below which there is no flux decline with the time and the flux depends linearly on the TMP. When this critical flux is reached the flux increases more slowly and approaches a constant value, which is, named the limiting flux. It is, therefore, important to operate below this critical value.

If the feed is pure water, this kind of plotting can be used to measure the Water permeability, which is an indication of the cleanness of the membrane. This phenomenon occurs because the membrane is free of any type of particles when pure water is used as feed.

2.2.3.2 Linear or cross-flow velocity

Linear velocity is the velocity at which the feed flows across the membrane. For a tubular membrane the linear velocity can be defined as the relation between the feed flow (or retentate velocity) and the cross flow area of the membrane:

where Fret is the retentate flow [ m3/s]

Acout is the cross-flow area [ m2]

A higher flow rate tends to remove the deposited material and, consequently, reduces the hydraulic resistance through the membrane and, in this way, the obtained permeate flux will be higher.

Higher feed flow rates also reduce the concentration polarisation phenomena by increasing the mass transfer coefficient.

2.2.3.3 Temperature

Higher temperatures benefit higher permeate fluxes, since the viscosity will be lower and the diffusion higher. However, it is essential not to pass over certain limits, because high temperatures denature proteins and enhance microbial growth during processing.

2.2.4 Flux decline reasons

During an actual separation, the membrane performance can significantly change with time, and often a typical flux-time behaviour may be observed: the flux through the membrane decreases over time. This behaviour is mainly due to the concentration polarisation and fouling.

Mechanisms

These two phenomena are aspects of the same problem, which is the build-up of retained components in the boundary layer of the membrane-solution interface. Both phenomena induce additional resistances on the feed side to the transport across the membrane, and at the same time, they are responsible for the gradual reduction of the permeate flux through the membrane, and for the change of the selectivity of the process.

Concentration polarisation is a reversible phenomenon, while fouling is irreversible and can be caused by several mechanisms: adsorption, pore blocking and/or formation of a gel layer.
The two phenomena are not completely independent of each other since fouling can also result from polarisation phenomena.

The extent of these phenomena is strongly dependent on the type of membrane processes involved and the feed employed. The flux decline is very severe in Microfiltration and in Ultrafiltration and, very often, the process flux is less than 5% of that for pure water. Figure 2.2 shows a schematic representation of the various resistances that can arise during a separation process.

See figure 2.2

Figure 2.2. Overview of various types of resistance towards mass transport across a membrane (Mulder, 1991).

The flux through the membrane can be described as:

Flux = driving force / (viscosity * total resistance)

or:

 

Resistances, concentration polarisation and adsorption
The resistances shown in Fig. 2.2 contribute in different extents to the total resistance (Rtot). In the ideal case, only the membrane resistance (Rm) is involved. The membrane has the ability to transport one component more easily than other, or in some cases, completely retain the solutes, provoking an accumulation of retained molecules near the membrane surface. This results in a highly concentrated layer near the membrane and this layer provokes a resistance towards mass transfer, the concentration polarisation (Rcp). Polarisation phenomenon always occurs and it is inherent to membrane separation processes. When the concentration of the accumulated solute molecules becomes very high, a gel layer can be formed and can provoke the gel layer resistance (Rg). In porous membranes, it is possible for some particles with the same size as the pore size of the membrane to penetrate into the membrane and block the pores, leading to the pore blocking resistance (Rp). This is the most important factor responsible for the fouling in Cross-flow microfiltration. A resistance can arise due to adsorption phenomena. (Rad). Adsorption results by the deposition of solutes in the pores or in the membrane surface, due to chemical adsorption of the solute on the membrane surface. This binding of solutes, particularly macrosolute such as proteins or polysaccharides, is a result of various chemical interactions between the macrosolute and the membrane surface. The factors that may influence these physicochemical reactions are pH, temperature and ionic strength and specific interactions.

2.2.4.1 Concentration polarisation

Concentration polarisation describes the concentration profile of the solutes in the liquid phase adjacent to the membrane resulting from the balance between different transport phenomena (general convection and back diffusion). As it has been named before, this is a reversible mechanism that disappears as soon as the operating pressure has been released. Concentration polarisation is responsible for the decreasing flux in Cross-flow microfiltration during the first 15 seconds of operation.

Accumulation of solutes

When a driving force acts on the feed solutions (solvent and solutes), solutes are partially retained while the solvent permeates through the membrane. This means that the membrane has a certain retentivity for the solutes, whereas the solvent can pass more or less freely. Thus, the concentration of the solutes in the permeate (cp) is lower than the concentration in the bulk (cb), and this fact is the basic principle of membrane separations. As the membrane is to some degree impermeable for the solutes, the retained solutes can accumulate at the membrane solution interface and their concentration will gradually increase. These components can only be transported back to the bulk solution by diffusion and turbulent flow caused by the tangential flow along the membrane surface. This phenomenon of surface accumulation is called concentration polarisation.
The concentration build-up will generate a diffusive flow back to the bulk of the feed, but after a while, steady-state conditions will be established. The convective solute flow to the membrane surface will be balanced by the solute flux through the membrane plus the diffusive flow from the membrane surface to the bulk. The concentration profile that has now been established in the boundary layer is shown in Figure 2.3 for a normal asymmetric and a reverse asymmetric membrane.

See figure 2.3

Figure 2.3. Schematic representation of hollow fibre and concentration polarisation profile for a normal asymmetric membrane and a reverse asymmetric membrane, where (a) skin layer, (b) porous layer, (J) permeate flux, (Cp.) concentration of the solute in the permeate, (Cb) concentration of the solute in the bulk solution, and (Cm) concentration of the solute at the surface of the skin layer (Guerra et al., 1996).

2.2.4.2 Membrane fouling

One of the major problems in the application of membranes in the industry is fouling and this aspect will be reviewed in a forthcoming section.

2.3 Juice filtration

The two main quality demands in juice processing are to preserve the organoleptic quality and clarify the juice for storage. Current filtration of a wide variety of juices is accomplished using filters aids, i.e., Diatomaceous Earth (DE) and Perlite, because it can meet required turbidity standards while maintaining good flavour levels. However, a final polishing is achieved by means of ultrafiltration.

Filter aids

DE or Kieselguhr is a natural substance derived from the cell walls of certain microscopic algae. Deposits of DE are found in various locations, including the US, England, and France. After mining and processing, DE can be supplied in a variety of grades. Perlite or volcanic silicate is an ore of volcanic rock containing silica. When crushed and heated, perlite expands to become a light, fluffy powder and is then suitable as a filter aid (Munroe, 1995).

These filter aids have excellent filtration qualities, however they involve significant costs, can not be cleaned, and are discarded after use. As a result, it is a consumable in many filtration processes. DE filters cost between 60,000 and 100,000 Euros per year to small and medium-sized fruit juice manufacturers. In the last years 9000 tons of filter aids have been used annually in Denmark.

Disadvantages of filter aids

The use of filter aids in filtration has no known harmful effects. However, there are some disadvantages involved in their use (Casani et al., 1999):

Some filter aids are classified for provoking lung diseases due to the dust,
There are environmental effects of the deposition of sludge or filter cake obtained,
The use of kieselguhr results in high costs since it mostly is imported. A lot of water is needed and the deposition of the filter cake is quite expensive.

The handling and disposal involve risks of inhalation, with specific health risk of silicosis, a disease of the lungs caused by inhaling siliceous particles. As a result, a number of health authorities want to reduce or eliminate the use of these filter aids by finding economic substitutes, and some European governments are considering banning the use of Kieselguhr (Russ, 1992 and Bridge, 1987).

Disposal costs

The main driving force in substituting kieselguhr filtration with new filtration methods is that the costs of disposal of kieselguhr sludge will increase dramatically in the years to come in Europe. The European Community causes this trend through harmonisation of the cost of disposal from country to country. Today the costs of disposal in Germany range between 30 and 120 Euros per ton of disposed materials, and these costs will be set to a level of 600 Euros per ton in the coming years (Kvistgaard & Jensen, 1994).

For more than 30 years, centrifuges have been an integral part of the technology applied in fruit juice processing for the separation of insoluble solids (Hamatscheck & Schöttler, 1994). However, the use of centrifuges results in high energy requirements and costs.

Membrane technology

Membrane technology is currently a "proven technology" within a few main areas, i.e. food and dairy industry, water purification and treatment of liquid fluent streams, and it is presently being introduced into a wide variety of other applications. The recent development of membrane technology will strongly influence the way industry evaluates separation processes in the immediate future. Membrane technology can work as well or better than the existing technology regarding product quality, energy consumption and environmental issues.

Advantages of membrane technology

The advantages of using membrane technology in the beverage industry are related to economy, working conditions, environment and quality (Hägg, 1998):

Low energy requirements and costs,
Avoids dust and sludge (formation/deposition),
Possibility of lower temperature processing (hence reduction of thermal damage to food during processing),
Simpler process design.

Ultrafiltration

Nowadays, ultrafiltration is still the dominating membrane separation technique for clarification of clear juices (McLellan, 1996). Hollow fibre ultrafiltration membranes (cutoff 50,000 or 100,000) produced by Romicon, Inc., Massachussets, USA, have been successfully used to clarify apple juice. The apple juice presented an excellent quality. The UF membrane holds back essentially all the polysaccharide materials such as pectin and starch, which are responsible for cloud and sedimentation (Short, 1983). However, some studies have indicated some losses in flavour when clarifying apple juice by means of ultrafiltration compared to microfiltration. Furthermore, other advantages of microfiltration are better efficiency and shorter processing periods (Wu et al., 1990).

Microfiltration

Microfiltration is also being used for the clarification and biological stabilisation of some beverages, and improvement in the Microfiltration membranes has led to improvement of bacterial and hygienic qualities.

Cross flow microfiltration is one of the most recent developments in membrane technology, and it is replacing a number of traditional clarification and sterilisation operations in a wide variety of industries (Forbes, 1987). This technique is today used with success in some applications in the pharmaceutical and biotechnological industry in Europe when purifying products of high value and low volume. The increase of disposal and purchase costs of filter aids will make it economically feasible to invest in alternative filtration methods for several industry sectors with a large use of filter aids, such as the brewing and the beverage industries, and cross-flow microfiltration can be one possible alternative.

2.4 Fouling & Cleaning

Fouling may be observed in membrane filtration as serious flux decline. Fouling is very complex and difficult to describe theoretically. Even for a given solution it will depend on physical and chemical parameters such as concentration, temperature, pH, ionic strength and specific interactions.

Membranes are used to remove wide variation of substances from different process streams. However, membrane fouling is the main factor reducing the applicability of the membrane processes. The degree of membrane fouling determines the frequency of cleaning, lifetime of the membrane, and the membrane area needed, and this will have a significant effect on the cost, design and operation of membrane plants (Speth et al., 1998).

With concentration polarisation phenomena, the flux at a finite time is always less than the original value. When steady-state conditions have been achieved, a further decrease in flux will not be observed and the flux will become constant as a function of time. Polarisation phenomena are reversible processes, but in practice, a continuous flux decline can often be observed. Such continuous flux decline is the result of membrane fouling. Fouling should be expected from any feed stream and, it comprises the matter that has left the liquid phase (retained particles, colloids, emulsions, suspensions, macromolecules, salts, etc.) to form a deposit on the membrane surface (adsorption and constriction) or inside of the pores (blocking). Depending on the size of the particle and the membrane pore size different cases of fouling can occur, giving different flux declines (see Figure 2.4).

Figure 2.4. Fouling schematics: Case A- pore narrowing and constriction, Case B-pore
plugging, and Case C- solute deposition and gel/cake layer formation (Belfort, 1993).

Part of the fouling can be defined as permanent or irreversible, which means that requires mechanical or chemical cleaning to restore the membrane properties. Another fraction of fouling may be non-permanent or reversible, when the deposited material is swept away by cross-flow, just after the pressure difference has been released.

In normal asymmetric membranes the fouling cake layer or boundary layer is formed on top of the membrane. After some time the cake layer will be responsible for the separation. The resistance that this layer offers to the permeate stream is very dependent on the dynamic conditions during cross-microfiltration, such as linear velocity and transmembrane pressure.

In reverse asymmetric membranes the cake layer or the boundary layer is formed inside the support layer.

Fouling substances, foulants, can be divided into five categories:

Sparingly soluble inorganic compounds,
Colloidal or particulate matter,
Dissolved organics,
Chemical reactants,
Microorganisms.

Biofouling is fouling, in which the main reason for the flux decline and operational problems is caused by accumulation of microorganisms.

2.4.1 Fouling analysis

Flemming et al. (1997) presented a procedure for analysis of the fouled membrane. The procedure is initialized with an optical inspection of the membrane, followed by microscopically inspection in order to get more detailed information about the structure of the fouling layer. Then, a defined area of the fouling layer is scraped off and the total amount of organics is determined. Finally, a part of the material is suspended in water and more specific analysis of the organic foulants are carried out. The problem in analysing the foulants is that most of the methods used are destructive. Thus, the analysis must be done either before or after the filtration test, because after analysis the membrane is destroyed.

2.4.2 Methods to reduce fouling

The decline of the flux is prejudicial to the economics of a given membrane operation and, for this reason, measures must be taken to reduce this. It is possible to minimise fouling to some degree using one or more of the following alternatives:

Increasing the convective transport of solute back to the bulk solution is done by choosing the appropriate module configuration and optimising the flow conditions. The use of turbulent promoters, pulsation of feed flow (backflush, backshock techniques) (Bertran et al., 1993), ultrasonic vibration (Davies, 1993), rotating modules, replacement of membrane in some places with non-rejecting sections i.e. are all methods that may also be applied and nowadays are more commonly used in practice (Matthiasson & Sivik, 1980).
Pre-treatment of the feed solution is utilised to remove the foulants or to change the properties of the solution in order to stabilise the foulants. Pre-treatment methods include heat treatment, pH adjustment, addition of complexing agents (EDTA), chlorination, adsorption onto active carbon, chemical clarification, pre-microfiltration and pre-ultrafiltration.
Change of the properties of the membrane in order to turn the membrane less prone for fouling. The main purpose of the surface treatment is to create a surface of such nature that protein and other foulants will not stick to it. Fouling is more severe in porous membranes (Microfiltration and Ultrafiltration) than in dense membranes, but a narrow pore size distribution can reduce the fouling. The use of hydrophilic rather than hydrophobic membranes can also help to reduce fouling. Negatively charged membranes can also be useful to reduce fouling, when in presence of negatively charged colloids in the feed.
Module and process conditions. Fouling phenomena diminish as concentration polarisation decreases. Increasing the mass transfer (high flow velocities) and using low (er) flux membranes can reduce concentration polarisation. The use of various kinds of turbulence promoters will also reduce fouling, although fluidised bed systems and rotary module systems seem not very flexible from an economical point of view for large scale applications.
Cleaning. The frequency with which membranes need to be cleaned can be estimated from the process optimisation. Three cleaning methods can be distinguished and the choice depends on the module configuration, the chemical resistance of the membrane and the type of foulant encountered:

- Hydraulic cleaning methods include   backflushing, which consists in alternate pressurising and depressurising by changing the flow direction at a given frequency.

- Mechanical cleaning can only be applied in tubular systems using oversized sponge balls.

- Chemical cleaning

2.4.2.1 Backflush and Backshock techniques

Fouling is due to the deposition of solids on the membrane surface and within its pores, retaining partly the macromolecules, and affecting membrane performance. The fouling layer can be reduced by maintaining a shear at the membrane surface, which drags the suspended or oversized particles in a direction normal to the permeate flow. This requires high cross-flow velocities or backflushing techniques.

High cross-flow velocities require greater energy consumption and can create transmembrane pressure gradients along the membrane.

Backflushing

Backflushing is an in situ method of cleaning the membrane by periodically reversing the transmembrane flow. In this way, the stationary concentration polarisation profiles are disturbed and the fouling layer is removed from inside the membrane and from the membrane surface. The backflushing medium can be the permeate, another liquid or gas, but if the permeate flow is used for flushing, it results in a loss of permeate against an increased flux.

Backshock

Jonsson and Wenten (1994) and Wenten et al. (1995) introduced the novel "Backshock" technique to reduce the loss of permeate during the backflushing and optimises the operation time during the filtration process. In this technique, the permeate flow is reversed for a short period of less than a second in order to avoid or reduce fouling or concentration polarisation problems, allowing the use of low linear velocities, which reduces the cost of running the process.

Backshock and reversed asymmetric membranes

Jonsson and Wenten (1994) also reported the advantages in using the Backshock technique on reverse asymmetric membranes. In this type of membranes, the fouling deposited inside the porous structure is less compact and thicker than that formed on the surface of normal asymmetric membranes. The resulting fouling layer presents less resistance as long as the porous layer is not completely filled up. If the cake is not removed from the porous layer, the permeate flux will approach to zero. Therefore, it is important to apply a very frequent backflushing in order to remove the cake and to avoid compacting the porous layer. When the backshock technique is applied, the induced concentration profile across the porous layer for the reversed asymmetric membranes will permit a steady state with 100% protein transmission even if the skin layer is very selective. The advantages of this arrangement are that even at linear velocities as low as 1 m/s and with an appropriate pore size, it is possible to limit the extend of concentration polarisation and achieve highly stable fluxes.

2.4.2.2 Chemical cleaning

Chemical cleaning is the most important for reducing fouling, with a number of chemicals being used separately or in combination. The concentration of the chemical and the cleaning are also very important relative to the chemical resistance of the membrane.

In Table 2.6 several cleaning agents from three different companies are presented as well as their composition and function. The selection, combination and concentration of the cleaning agents will depend on the composition of the process fluid that is being filtered and on the type of membrane.

Table 2.6. Cleaning agents from Scan Diversey, NovoDan and Henkel-Ecolab.

Supplier Cleaning agent Composition Function
Scan Diversey Divos 124 KOH, tetrasosdium-EDTA, anionic tensid Emulsification of fat and proteins
  Divos 2S Acid agent based on Phosphoric-acid Removal of inorganic contaminates
  Divos 120 CL Strong oxidative step Emulsification of fat and proteins
NovoDan Ro-Dan 30 Organic acid Desegregation of portions and inorganic contaminates
  Enzyme PL1

+ Ro-TEN AlKA

+ pH adjustment solution

PL 1 is an enzymatic product based on proteases and lipases. The Ro-TEN AlKA is based on carbonate, potassium hydroxide anionic and nonanionic tensid. The pH adjustment is made with soda and NaHCO3 Emulsification and degradation of fat and proteins
  Ro-Dan Acid Solution based on Phosphoric acid and nitric acid Removal of inorganic contaminates
Henkel-Ecolab P3-Ultrasil 53 Neutral, enzymatic powder detergent containing a combination of organic and inorganic surfactants Emulsification and degradation of fat and proteins
  P3-Ultrasil 73 Organic acid with surfactants- Phosphor and Nitrogen free Disinfection and removal of fat and inorganic particle
  P3-Oxonia AKTIV

P3-Ultrasil 141

The P3-oxonia AKTIV is based on hydrogen peroxide and peracetic acid. The P3- Ultrasil 141 – Alkali cleaning agent is based on phosphate and potassium hydroxide Strong disinfecting and oxidative power given by the hydrogen peroxide. The P3-141 ultrasil removes and emulsifies proteins and fat

2.4.2.3 Ultrasounds

Background

Membrane resistance control in cross-flow microfiltration by the use of high-power ultrasound acting on the membrane, has been reported by E.S.Tarlton and R.J.Wakeman on china-clay and anatase and by Yutaka Matsumoto, on bakers yeast and Bovine Serum Albumin BSA.
Both groups are concluding that high-power ultrasonic irradiation is efficient in controlling fouling in some cases, and state that the effect is due to local cavitational events.E.S.Tarlton and R.J.Wakeman also reported very promising results from using electrostatic fields combined with ultrasound.

Yutaca Matsumoto recorded the important observation that major improvement in performance of the filtration can be obtained when the trans-membrane pressure is shifted to 0 (Bar) during ultrasonic irradiation.

[Front page] [Contents] [Previous] [Next] [Top]