K k1 r k

The k's appearing in the equation are mass transfer coefficients, and their reciprocals represent the mass transfer resistance at each step. For many pervaporation processes, the mass transfer resistance in the vapor boundary layer tends to be small enough to be ignored. This leaves only the liquid boundary layer (1/^bl) and membrane (1/km) resistances to deal with. km is strongly determined by polymer properties, the thickness of the membrane, and chemical structures of the components in the liquid. In situations where hydrophobic volatile compounds are being removed from the liquid by pervaporation, the mass transfer rate is often limited by diffusion of the compound in the liquid boundary layer, i.e., k°v = kbl. This situation arises because nowadays membrane can be made in such a way that the membrane provides minimal mass transfer resistance to the volatile compounds, almost to the point of being absent. This situation manifests itself as a phenomenon called concentration polarization, i.e., the steep discrepancy in volatile compound concentrations between the bulk and the membrane surface. When concentration polarization is severe, the function of membrane is to minimize the solvent flux, thereby maximizing the selectivity of the intended separation. The analysis of the liquid boundary-layer mass transfer resistance is very important to the process designers and operators alike. One common approach to the analysis is to find out the correlation between the mass transfer coefficient and process parameters. It is recommended that the boundary-layer theory would have to be adopted to provide more robust analysis that has broader application and scalability. However, in reality, it is exceedingly difficult to do it. Instead the semiempirical correlations that have the form of Sherwood number correlations shown in Equation 3.25 are commonly employed. Among these correlations is the frequently cited Lévêque's correlation (Equation 3.35):

d yLy

The only new parameter in this correlation is L, the length of the membrane channel. This correlation indicates that the mass transfer coefficient is mainly dependent upon flow conditions on the feed side and the shape and dimensions of the module. Temperature has a substantial impact upon the mass transfer coefficient through the diffusivity of the solute and viscosity. Nevertheless, temperature is not a commonly manipulated parameter due to the issues of membrane stability and the vapor pressure of the volatile solute because the feed has to be kept in the liquid phase. Flow velocity (flow rate) is the parameter that can be adjusted to minimize the liquid boundary layer resistance for a fixed membrane module (configuration).

Membrane distillation (MD)

Membrane distillation (MD) is a type of low-temperature, reduced pressure distillation using porous hydrophobic (water-hating) polymer materials. It is a process that separates two aqueous solutions at different temperatures and has been developed for the production of high-purity water, and for the separation of volatile solvents such as acetone and ethanol. MD can achieve higher concentration than RO. In MD, the membrane must be hydrophobic and microporous. The hydrophobic nature of the material prevents the membrane from being wetted by the liquid feed; hence, liquid penetration and transport across the membrane is avoided, provided the feed side pressure does not exceed the minimum entry pressure for the pore size distribution of the membrane. The driving force of MD is the temperature gradient, and the two different temperatures produce two different partial vapor pressures at the solution-membrane interface, which propels consequent penetration of the vapor through the pores of the membrane. The vapor is condensed on the chilled wall by cooling water, producing a distillate. This process usually takes place at an atmospheric pressure and temperature that may be much lower than the boiling point of the liquids (e.g., solvents). It is commonly observed that the effect of the osmotic pressure from the permeate to the feed solution will be prominent when the high solute concentrations of feed liquids are processed. A variation of MD is sometimes called low pressure membrane distillation or osmotic distillation, which uses an auxiliary device to condense the vapor coming out of the membrane. The driving force for vapor transport in this case is the pressure differential. Alternatively, the auxiliary cooling device can be replaced by using an inert sweep gas or absorbing strip liquid to remove the vapor permeate and maintain the pressure differential. The membrane distillation is very similar to a single-stage distillation and is thus unable to achieve a high separation factor. The primary advantage of MD is the high surface-area:volume ratio available and thus high permeation rates. Most food applications of MD are concentrated upon dehydration of liquid foods.

The performance of an MD can be evaluated by a phenomenological equation (Equation 3.36):

where the flux is related to two parameters; one is pressure difference and the other proportionality factor ("membrane permeability") F. AP is mainly determined by the temperature difference T, which can be related to the Clausius-Claperyron relationship (Equation 3.37):

Hvap is the enthalpy of the vapor of permeating species, T the temperature, and c the constant. MD sometimes experiences a temperature polarization phenomenon due to the difference in the heat transfer rate between the heat conduction in the membrane and the heat transfer in the bulk fluid.

Membrane contactor (MC)

Membrane contactors (MCs) are a motley group of several membrane processes that primarily use the membranes as mass transfer barriers for certain matters and interfaces between two phases. The driving force in the process is typically the difference in either vapor pressures or the osmotic pressures across the membrane barrier. The relevant membrane contactors for fruit juice processing are direct osmosis, MD, and osmotic distillation (OD). Direct osmosis (DO) originates from an old practice for juice concentration and is caused by the difference of osmotic pressures between the dilute juice on the upstream side and the brine on the other side. The disadvantages of DO are the high costs and low permeation rates, even though DO can reach concentrations greater than those achieved by RO. MD utilizes the vapor pressure difference across the membrane resulting from temperature difference to drive the solvent molecules across the mi-croporous hydrophobic membrane and condense at the cold side of the membrane. Because temperature difference (subsequently vapor pressure difference) drives MD, an increase in temperature on one side of the membrane would increase permeation rate. However, temperature polarization becomes prominent at high temperatures and exerts negative effect on permeation of MD. The other problem with MD operations is that the process is limited in terms of operating temperature due to concerns about thermal damage to the flavor compounds in fruit juices. This limitation with MD makes OD (also known as osmotic evaporation) or isothermal membrane distillation a better choice. In OD, a liquid mixture containing a volatile component is contacted with a microporous, nonwettable membrane whose opposite surface is exposed to another liquid phase where the mass transfer takes place across the membrane. This technology can remove se lectively the water from liquid wastes under atmospheric pressure and at room temperature. Like other membrane contactors, OD suffers from high costs and low permeation rate.

Design considerations

In many cases, it is still true to say that the design of a membrane process and/or the selection of a membrane module/material for desired separations remains a mixture of art and science, in which knowledge, experience, and science all play important roles. More often than not, there is no "right answer" in the absolute sense, because more than one solution is both technically and economically viable. However, a careful evaluation at the outset of as many as possible of the factors influencing the choices will help narrow down the items on the list.

When contemplating the use of any particular membrane process for the separation of components in a liquid food stream after the initial assessment, several process issues must be evaluated. The first step in doing so is to draw up the detailed requirements for the process. Accurate qualitative and, where possible, quantitative information on the following aspects should therefore be specified:

• The components and range of concentrations in the feed

• The intended use or fate of the treated feed liquids (i.e., final products, further processing, etc.)

• The intended use of or fate of the permeate (i.e., disposal, reuse, further processing, etc.)

• Permeation flux

• The minimum properties of the treated food fluids and permeate that will make the intended use or fate possible

• The membrane transport mechanism

• A cost-effective and environmentally friendly alternative solution

Any one or more of the above factors may, depending upon circumstances, influence the design of a membrane process. In the case of food-related wastes and by-products, it is normal for economical considerations to override operational simplicity. This is particularly true for aroma compound recovery in orange juice production effluents using pervaporation, in which the preservation of the permeate is the key to exerting a major influence on the ultimate cost of treating the wastewater stream. In other processing applications, a membrane process can be an important intermediate operation that is vital to subsequent processing operations.

The next issue to be addressed is whether the membrane processes are actually capable of separating the components from liquid foods. The answer to this question for the pressure-driven membrane processes such as MC, UF, NF, and RO is generally affirmative, provided that appropriate membranes (pore size and, for NF and RO, membrane properties such as charge, hydrophilic tendency) are used. For pervaporation, the answer is more complicated and conditional. It is well documented that PV works well when the compound to be removed has a high vapor pressure relative to the background material and a low solubility in the background material. In dilute aqueous solutions such as aromas in orange juice, it is generally the Henry's constant that determines whether an aroma compound can be effectively separated by pervaporation. The Henry's constant represents the vapor-liquid partitioning of organic compounds in an aqueous system. The general rule of thumb is: the more dissimilar the components, the easier it will be to separate them.

Once the process designer has determined that a particular membrane process will, theoretically, work. The subsequent questions to be answered are

• Does a membrane material exist which will do the job?

• Is this membrane material available in a membrane module?

The answer to the first question is usually positive. A great deal of membrane research has been performed on many membrane materials and feed mixtures. In addition, a wide array of membrane materials is available; these are materials that may achieve the desired separation, but they have not been tested in a membrane mode of interest.

Another variable in the selection of membrane materials is whether a single-layer membrane or a multilayer membrane is to be used. Membranes used in an MF are normally single-layer, isotropic; membranes in other pressure-driven membrane filtrations and pervaporation are composed of composite or multiplayer, nonhomogenous materials. This is because a membrane with desired selectivity may require a significant thickness to deliver the desired physical properties, such as burst pressure; however, improving membrane mechanical stability by increasing membrane thickness would inadvertently reduce permeation flux. To get around this problem, a composite or nonhomogeneous membrane is employed, where a thick layer of polymer material with large pore size supports a top thin layer of active membrane.

The second question is about the issue of commercial availability of membrane configurations or membrane modules for particular membrane materials. A module is the smallest unit into which the membrane material is packed. The reason for using modules is that although polymer membranes are made in two basic physical forms—flat sheet and tubular—many practical membrane systems that need large membrane areas can be accommodated only in membrane modules. For pressure-driven membrane processes, membrane distillation, and pervaporation there are four primary configurations (modules), each with inherent advantages and weaknesses: spiral wound, hollow fiber, plate-and-frame, and tubular.

Membrane modules

Spiral wound module

A spiral wound module is a logical step from a flat sheet membrane. In spiral wound modules, a flat membrane envelope or set of envelopes is rolled into a cylinder, as shown in Fig. 3.14. The envelope is constructed from two sheets of membrane, sealed on three edges. The inside of the envelop is the permeate side of the membrane. A thin porous spacer inside the envelope keeps the two sheets separated. The open end of the envelope is sealed to a perforated tube (the permeate tube) with a proper glue so that the permeate can pass through the perforations; for pervaporation, it is also the place to which the vacuum or sweep gas is applied. Another

_ FHH Sjixp - Hw^tiJflt ■pflimpaDe Spacf ■ UWlflH Srt -Fred Space*

Figure 3.14. A schematic illustration of a spiral wound module (courtesy of Dr. Leland Vane of USEPA).

_ FHH Sjixp - Hw^tiJflt ■pflimpaDe Spacf ■ UWlflH Srt -Fred Space*

Figure 3.14. A schematic illustration of a spiral wound module (courtesy of Dr. Leland Vane of USEPA).

spacer is laid on top of the envelope before it is rolled, creating the flow path for the feed liquid. This feed spacer generates turbulence due to the undulating flow path that disrupts the liquid boundary layer, thereby enhancing the feed-side mass transfer rate. It is the envelopes and spacers wrapped around the permeate tube that give the module its name, spiral wound module. The spiral wrapped envelopes and spacers are then wrapped again with tape or glass or netlike sieve before fitting into a pressure vessel. In this way, a reasonable membrane area can be housed in a convenient module, resulting in a very high surface-area:volume ratio. One noticeable drawback lies in the permeate path length. A permeating component that enters the permeate envelope farthest from the permeate tube must spiral inward several feet. Depending upon the path length, permeate spacer design, gel layer, and permeate flux, significant permeate side pressure drops can be encountered. The other disadvantage of this module is that it is a poor choice for treating fluids containing particulate matters.

Hollow fiber module

In a hollow fiber configuration, small-diameter polymer tubes are bundled together to form a hollow fiber module like a shell and tube heat exchanger (Fig. 3.15). These modules can be configured for liquid flow on the tube side, or lumen side (inside the hollow fibers), or vice versa. These tubes have diameters on the order of 100 microns. As a result, they have a very high surface-area:module-volume ratio. This makes it possible to construct compact modules with high surface areas. The drawback is that the liquid flow inside the hollow fibers is normally within the range of the laminar flow regime due to its low hydraulic diameter. The consequence

Epoxy Se?I

Figure 3.15. A schematic illustration of a hollow fiber module (courtesy of Dr. Leland Vane of USEPA).

Epoxy Se?I

Figure 3.15. A schematic illustration of a hollow fiber module (courtesy of Dr. Leland Vane of USEPA).

of prevalent laminar flows is high mass transfer resistance on the liquid feed side. However, because of the laminar flow regime, the modeling of mass transfer in a hollow fiber module is relatively easy and the scale-up behavior is more predictable than that in other modules. One noticeable problem with a hollow fiber module is that a whole unit has to be replaced if failure occurs.

Plate-and-frame module

Plate-and-frame configuration is a migration from filtration technology and is formed by the layering of flat sheets of membrane between spacers. The feed and permeate channels are isolated from one another using flat membranes and rigid frames (Fig. 3.16). This configuration was an early favorite; it is a natural scale-up from bench-scale laboratory membrane

Plate And Frame Membrane Module
Figure 3.16. A schematic illustration of a plate-and-frame module (courtesy of Dr. Leland Vane of USEPA).


Figure 3.17. A schematic illustration of a tubular module.

cells that have one feed chamber and one permeate chamber separated by a flat sheet of membrane. A single plate-and-frame unit can be used to test different membranes by swapping out the flat sheets of membrane. Further, it allows the use for membrane materials that cannot be conveniently produced as hollow fibers or spiral wound elements. The disadvantages are that the ratio of membrane area to module volume is low compared to spiral wound or hollow fiber modules, dismounting is time-consuming and labor-intensive, and higher capital costs are associated with the frame structures.

Tubular module

Polymeric tubular membranes are usually made by casting a membrane onto the inside of a preformed tube, which is referred to as the substrate tube. The tube is generally made from one or two piles of nonwoven fabric, such as polyester or polypropylene. The diameters of tubes range from 5-25 mm (Fig. 3.17). A popular method of construction of these tubes is a helically wound tape that is welded at the edges. The advantage of the tubular membrane is its mechanical strength if the membrane is supported by porous stainless steel or plastic tubes. Tubular arrangements often provide good control of flow to the operators and are easy to clean. Additionally, it is the only membrane format for inorganic membranes, particularly ceramics. The disadvantage of this type of modules is mainly higher costs in investment and operation. The arrangement of tubular membranes in a housing vessel is similar to that of the hollow fiber element. Tubular membranes sometimes are arranged helically to enhance mass transfer by creating a second flow (Dean vortex) inside the substrate tube (Moulin et



Figure 3.17. A schematic illustration of a tubular module.

Although the specification for the process is the most critical issue in process design of membrane systems, certain auxiliary steps must also be considered in the operation of a membrane system. For example, temperature, pH limits, and tolerance to certain chemicals, particularly cleaning agents such as alkalis and detergents should be considered before a process is put online. These cleaning chemicals, as well as seals and glues used in the membrane modules, have to be approved by FDA or other regulatory agencies for use in food processing—an aspect of process design that is often neglected by some designers.

Whose fault? Membrane fouling!

As described in the previous sections, concentration polarization phenomena of membrane processes cause noticeable decline of membrane performance. In membrane filtration processes such as microfiltration and ultrafiltration, concentration polarization phenomena always accompany the formation of a gel layer that is either irreversible or reversible. The cause of gel layer formation is thought to be the result of the rapid accumulation of retained solutes near the membrane surface to the point that the concentration of a macromolecule solute reaches the gel-forming concentration. High retention of solutes near the membrane surface inevitably also leads to concentration polarization and, as a result, the performances of membrane filtration processes (pressure-driven processes) suffer. The version of concentration polarization in pervaporation is slightly different from its kindred in membrane filtration, as stated in the previous section. concentration polarization also negatively affects the performance of an electrodialysis process. For membrane distillation, temperature polarization is the main culprit for the decline in the process performance.

Membrane fouling is suspected if the membrane flux is continuously declining after a period of time of operation. This is usually an irreversible, partially concentration-dependent, and time-dependent phenomenon, which distinguishes it from concentration polarization. The identification of membrane fouling is imprecise and often based upon operator's experience, performing fouling tests with membrane filtration index apparatus, and membrane vendor's recommendations. Membrane fouling is intimately related to concentration polarization, but the two are not exactly interchangeable in our description of membrane performance deterioration. We now know that all membrane filtration processes experience some degree of concentration polarization, but fouling occurs mainly in microfiltration and ultrafiltration. Relatively large pores in these membranes are implicitly vulnerable to fouling agents such as organic and inorganic precipitates and fine particulate matters that could lodge in these pores or deposit irreversibly on the membrane surface. The exact cause of membrane fouling is very complex and therefore difficult to depict in full confidence with available theoretical understandings. Even for a known solution, fouling is influenced by a number of chemical and physical parameters such as concentration, temperature, pH, ionic strength, and specific interactions (hydrogen bonding, dipole-dipole interactions).

Membrane fouling in membrane filtration processes like concentration polarization is unavoidable—this is particularly true for protein concentration or fractionation. However, certain steps that will greatly reduce the severity of membrane fouling can still be achieved. One effective way of reducing membrane fouling is to provide pretreatment to the feed liquids. Some simple adjustments such as varying pH values and using hydro-philic membrane materials can also do wonders in protein concentration operations. There are persistent interests in modifying membrane properties to minimize the membrane-fouling tendency around the world. Because membrane fouling is intimately associated with the concentration polarization phenomenon, any action taken to minimize concentration polarization will also benefit the fight against membrane fouling. Unfortunately, no matter how much effort put forward fighting membrane fouling, it will eventually occur. The only solution by then is employing a cleaning regimen. The frequency of cleaning depends upon many factors and should be considered as a part of a process optimization exercise. There are three basic types of cleaning methods currently used: hydraulic flushing (back-flushing), mechanical cleaning (only in tubular systems) with sponge balls, and chemical washing. When using chemicals to perform defouling, cautions must be observed because many polymeric membrane materials are susceptible to chlorine, high pH solutions, organic solvents, and a host of other chemicals.

Was this article helpful?

0 0
Project Earth Conservation

Project Earth Conservation

Get All The Support And Guidance You Need To Be A Success At Helping Save The Earth. This Book Is One Of The Most Valuable Resources In The World When It Comes To How To Recycle to Create a Better Future for Our Children.

Get My Free Ebook

Post a comment