Established membrane technologies

11.2.1 Micro- and ultrafiltration

Description of the process

Micro- and ultrafiltration (MF and UF, respectively) are some of the most important processes in food technology and biotechnology due to their easy handling and universal application. MF and UF are simple membrane filtration processes, i.e. the separation principle is a form of sieving.7 The only difference between membrane filtration and sieving is the cut-off point, which is determined by the pore size of the membrane. Using MF it is possible to separate particles with diameters between 200 nm and 20 |m. In contrast, UF can separate macromolecules with diameters between 5 and 200 nm. Hence, there is a functional overlap between the two processes.

A range of different polymeric materials such as polyethylene, polypropylene, polycarbonate, and cellulose acetate are used for MF and UF. In the food and biotechnology sectors, very stable (mechanically, chemically, and thermo-stable) membranes made of polysulfon or polyether sulfon (Fig. 11.1) are preferred since they can be easily cleaned and disinfected regularly to prevent fouling.

When choosing a membrane for processing a solution by MF/UF it is also important to consider the additional parameters of selectivity, Y, and molecular cut-off point. Notation for this chapter can be found in Section


The selectivity, Y, describes the separation efficiency. This relates to the ability of a membrane to hold back a specific component under defined conditions and is defined as:

Polyether sulfon

O PoIysuIfon

Fig. 11.1 Chemical structure of membrane materials.

O PoIysuIfon n

Fig. 11.1 Chemical structure of membrane materials.

where Y is selectivity, ciP is the concentration of i in the permeate solution and ciF is the concentration of i in the feed solution. The selectivity Y can be determined in the same way as the retentate, R, of the reverse osmosis (see below).

Molecular cut-off

This is defined as the molecular mass at which a dissolved globular protein (or another defined macromolecule) is nearly completely held back by the membrane at a given transmembrane pressure difference. Usual cut-off points are between 1000 and 100 000 Da for UF and between 10 000 and 300 000 Da for MF. When determining the molecular cut-off point it is important to avoid the formation of sediment layers which would influence the cut-off value.

The molecular cut-off point cannot be regarded as the absolute measure for the separation properties of a membrane, because these properties are also influenced by the molecular structure and the interaction between the membrane and the component that is held back (Fig. 11.2). Nevertheless, the cut-off point is still helpful when choosing the right membrane for a concrete separation problem.

As mentioned above, the separation efficiency of a UF/MF membrane is reduced by sediment layers and fouling. The formation of a sediment layer involves a concentration polarization (Fig. 11.3). An insufficient rediffusion of the component that is held back in the feed solution results in a higher concentration adjacent to the membrane. In the steady state, the convective transport towards the membrane is just compensated by the rediffusion in the laminar boundary layer. However, if the solubility limit of one component is reached, precipitation takes place, which results in the formation of a covering layer.

0 100

10 000 Molecular mass (Da) Fig. 11.2 Separation curve for a UF membrane.

1 000000

Fig. 11.3 Concentration polarization, schematic of membrane without a covering layer (a) and with a covering layer (b). YG, thickness of the laminar boundary layer; YB, solution (turbulently mixed); cg, concentration of dissolved component in the solution; cim, concentration of substance i at the membrane surface; cs, saturation concentration of substance i; Jv, permeate flow rate.

To decrease the thickness of the covering layer, the rediffusion of the molecules from the membrane into the solution has to be ensured through better mixing. This is usually achieved through the control of the direction and velocity of the flow. A cross-flow is the most widely used process mode. This is depicted schematically in Fig. 11.4. Here, the suspension or solution constantly flows over the membrane. There are two main flows cross-wise to each other: the filtrate flow (JF through the filter material) and the overflow (parallel to the filter medium). Hence cross-flow is the most commonly used process mode (see Fig. 11.4).

Membrane fouling is mainly caused by adsorption layers of bound proteins. Such protein adsorption depends on the local protein concentration. It will therefore exist, to some degree, at the beginning of the concentrating process of any dissolved protein(s) (for example in the process of whey concentrating) and is then strengthened by concentration polarization. The protein layer is a good food source for microorganisms which can reinforce the fouling and even result in the destruction of membrane material. Only by cleaning the whole plant with anti-fouling reagents such as acids, bases, or inorganic salts (K2CO3, Na2CO3) can such fouling be reduced.

Possible applications of MF/UF

Table 11.2 provides an overview of the types of applications to which MF and UF have been applied. Wastewater treatment is a very interesting field for MF and UF, as current research shows. In the food industry many aqueous effluents occur and need to be reprocessed. For example, Drouiche

Fig. 11.3 Concentration polarization, schematic of membrane without a covering layer (a) and with a covering layer (b). YG, thickness of the laminar boundary layer; YB, solution (turbulently mixed); cg, concentration of dissolved component in the solution; cim, concentration of substance i at the membrane surface; cs, saturation concentration of substance i; Jv, permeate flow rate.

Jo o O 0 o


w o o o o o 8°o88°o88°o8 8°o8 8°ocf 8°o8

: suspension flow : filtrate flow = concentrate flow lJp

Fig. 11.4 Principle of cross-flow filtration.

Table 11.2 Application fields of microfiltration (MF) and ultrafiltration (UF)

Pharmaceutical industry/biotechnology

Sterile filtration: diagnostics, antibiotics, blood products, culture mediums, solvents, acids

Food industry

Sterile filtration: wine, beer, whisky, cooking oil, vinegar, amino acids, sugar solutions

Clarifying whey, degreasing and sterilizing milk

Clarifying wine, beer, fruit juice

Chemical industry

Sterile filtration: solvents, reagents, inorganic solutions, fatty acids, waxes, polymers

Treatment of process water and other waste water

Recovery of catalysts and solvents

Filtration of strong acids, process chemicals

Separation of heavy metals as hydroxides, and lignin et al.8 developed a process for the treatment of olive mill waste water by combining UF and advanced oxidation processes. A process combining UF and clay support is used to purify water effluents from a milk factory.9 The recovery of proteins from effluents is a particularly interesting field of application. Lo et al.,10 for example, developed a process for the recovery of proteins from poultry processing. Membrane filtration of Mozzarella whey can be used to recover high-value proteins, this process is described in Rektor and Vatai.11 Even valuable enzymes can be recovered by UF and MF. Daufin et al.13 give a good overview of current developments in the use of membrane technology in the food industry.

11.2.2 Reverse osmosis

Description of the process Principle of reverse osmosis

The phenomenon of osmosis occurs when a semi-permeable membrane separates a solution from its solvent or a similar solution with a lower concentration. The difference in concentration results in a difference in the osmotic pressures which leads to a volume flux towards the solution with the higher concentration. The phenomenon of osmosis is described in detail in basic literature for physical chemistry (e.g. Atkins14).

However, the direction of the volume flow can be reversed. This can be achieved by applying hydrostatic pressure to the solution. This pressure has to be higher than the osmotic pressure difference across the membrane. The result is the active concentration of the solution. The process described is also known as reverse osmosis.

If, in the case of systems with more components, the couplings can be excluded, the driving force for the permeated component of a mixture i is exclusively the chemical potential difference, 4u» on both sides of the membrane. The chemical potential is a parameter that accumulates all effects that may influence the behavior of a system, in our case it includes the following parameters: pressure, temperature, and the concentration of the solution of one side of the membrane. This chemical potential is defined as an infinitesimally small change in the free molar (Gibbs) enthalpy, G, at an infinitesimal change of the concentration xt of those components in the case of an isobar-isothermal process.

where | i is the chemical potential of i, G is Gibbs enthalpy, p is hydrostatic pressure, T is temperature and xi is the mole fraction of i. Therewith, it follows that the work, W, needed to change the concentration x of a system beginning with a concentration 1 and ending with a concentration where W1,2 is the work needed to change the concentration x from x1 to x2.

For the thermodynamic principles behind these equations please refer to Atkins.14 The chemical potential of the components in liquids is:

where ^ is the universal gas coefficient, ai is the activity of component i, p0 is standard pressure and V is the molar volume of i.

The chemical potential of the components in an ideal gas mixture is:

m(T, p, xi) = mi0(T, p0) + KT ln a(T, p0, x,) = Vt(p - p0)

The osmotic pressure definition equation is:

and the definition equation of transport of the component i in the case of reverse osmosis is:

AmiiRo = V,[pf - Pp - (p,,F - p,,p)] = V,(Ap - Ap,)

where pF is the feed pressure and pP is the permeate pressure.

The notation demonstrates why reverse osmosis is so named. If the pressure difference on the transmembrane pressure (Ap) is higher than the difference of the osmotic pressures p, then it results in a change in flow direction (see Fig. 11.5).

By letting water flow through a selective membrane out of the more concentrated solution it would be possible to produce (provided that Ap > pW) pure water out of brine. Based on above-mentioned notations that describe the membrane separation behavior, it is possible to project separation efficiency (and also the manufacturing yield). Figure 11.6 shows a schematic diagram of a reverse osmosis unit depicting the necessary parameters required for the calculation of the separation efficiency. The separation performance of a reverse osmosis system can be described in terms of the medial permeate flux vp and the medial permeate mass fraction wP, provided that entry mass fraction wF and operation pressure p are given.16

Medial permeate flux, vP =

Medial permeate mass fraction, Wp =

p Reverse osmosis


Solution Solvent <-



Volume flow Volume flow

Fig. 11.5 Principle of reverse osmosis.



where v is volume flux; Amem is membrane area; m is mass; subscripts ges and R indicate overall and retentate, respectively; m is mass flux; subscripts P and F indicate permeate and feed, respectively.

Modeling of the mass transport through the membrane Mathematical models are necessary for dimensioning and optimizing modules and processes. They are also important for an economic evaluation and comparison of membrane processes with conventional alternatives. The basis for each plant dimensioning and process simulation is the preservation of mass/impulse/energy and matter equations. The differential balancing of the conserved quantities in each segment of the process (membrane) leads to a coupled differential equation system. The solution of such an equation system is in most cases only numerically possible. It is very important that these models contain relations that describe the mass transfer in the membrane not only qualitatively but also quantitatively. Such relations describe the mass transfer of the components through the membrane as a function of the operating conditions. This means that it is described as function of (along the module) changing internal state variables (pressure, temperature, concentration) and external state variables (e.g. feed flux) of the system. Unfortunately it is very difficult to model the mass transfer through the membrane because there are numerous interactions between all the components involved in the process. These, again, lead to non-idealities and cross-effects. However, there are different ways to get quantitatively reliable dimensioning equations. Figure 11.7 shows an overview of the different ways in which modeling of mass transfer could be conducted.

For engineering interpretations, the half-empirical models have been useful. They have enabled qualitatively and quantitatively precise characterization of the separation performance. They are based on real-system permeation experiments in conjunction with idealized physico-chemical models. The essence of half-empirical modeling is that the genuine, elementary physical and thermodynamic processes of molecular-level, descriptive mass values are combined into meaningful 'parameter groups'. The values

Mass transfer modeling

Origin / application

Modeling type Examples

Fundamental research, membrane developing

Engineering science, installation construction,

Regression analysis, installation construction

Fundamental research, membrane developing

Engineering science, installation construction,

Regression analysis, installation construction

Modeling type Examples

Examples Solution-diffusion model Friction model

Pores model Kedem-Speigler model

Fig. 11.7 Overview of possibilities for modeling the mass transfer.16

Examples Solution-diffusion model Friction model

Pores model Kedem-Speigler model

Fig. 11.7 Overview of possibilities for modeling the mass transfer.16

of those parameter groups are determined by conducting permeation experiments on real-system membrane-mass systems. Mass transport models of this kind cannot and should not be used to make any predictions about the transport coefficient and separation power of a membrane. They should rather be used to create qualitatively and quantitatively precise descriptions of the membrane separation characteristics. In practice, in order to have reliable and manageable notations, it is important that only a few readily determined free model parameters are introduced into the models.

In cases where modeling of the mass transport in the membrane is not possible, for example because few free model parameters are available, it is necessary to use (in order to determine the dimensions or a membrane unit) a regression analysis that is based on experiments in the field of interest.

Industrial application of reverse osmosis

Traditionally, reverse osmosis is used predominantly in the seawater and brackish water desalting industry. However, because of the increasing availability of more proficient membranes, it is frequently being applied to process organic/watery systems in many industries. Table 11.3 gives an overview of the applicability of reverse osmosis in industrial separation.

Guide to Alternative Fuels

Guide to Alternative Fuels

Your Alternative Fuel Solution for Saving Money, Reducing Oil Dependency, and Helping the Planet. Ethanol is an alternative to gasoline. The use of ethanol has been demonstrated to reduce greenhouse emissions slightly as compared to gasoline. Through this ebook, you are going to learn what you will need to know why choosing an alternative fuel may benefit you and your future.

Get My Free Ebook

Post a comment