Membrane Separations Technic.pdf

Mass Transfer (ChE - 392) Membrane Separation Processes Saeed GUL, Dr.Techn, M.Sc. Engg. Associate Professor

Department of Chemical Engineering, University of Engineering & Technology Peshawar, PAKISTAN

Introduction Whilst effective product separation is crucial to economic operation in the process industries, certain types of materials are inherently difficult and expensive to separate. Important examples include:  Finely dispersed solids, especially those which are compressible, and which have a density close to that of the liquid phase, have high viscosity, or are gelatinous.  Low molecular weight, non-volatile organics or pharmaceuticals and dissolved salts.  Biological materials which are very sensitive to their physical and chemical environment. The processing of these categories of materials has become increasingly important in recent years, especially with the growth of the newer biotechnological industries and with the increasingly sophisticated nature of processing in the food industries. 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

2

Introduction When difficulties arise in the processing of materials of biological origin, it is worth asking, how does nature solve the problem? The solution which nature has developed is likely to be both highly effective and energy efficient, though it may be slow in process terms. Nature separates biologically active materials by means of membranes

As STRATHMANN has pointed out, a membrane may be defined as “an interphase separating two phases and selectively controlling the transport of materials between those phases”. A membrane is an interphase rather than an interface because it occupies a finite, though normally small, element of space. Human beings are all surrounded by a membrane, the skin, and membranes control the separation of materials at all levels of life, down to the outer layers of bacteria and subcellular components. 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

3

Introduction  since the 1960s a new technology using synthetic membranes for process separations has been rapidly developed by materials scientists, physical chemists and chemical engineers.  Such membrane separations have been widely applied to a range of conventionally difficult separations.  They potentially offer the advantages of ambient temperature operation, relatively low capital and running costs, and modular construction.  In this chapter, the nature and scope of membrane separation processes are outlined, and then those processes most frequently used industrially are described more fully.

27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

4

Classification Of Membrane Processes Industrial membrane processes may be classified according to the size range of materials which they are to separate and the driving force used in separation Pressure Driven

Concentration Driven

Electric Potential Driven

Temperature Driven

Microfiltration

Dialysis

Electrodialysis

Membrane Distillation

Ultrafiltration

Forward Osmosis

Nanofiltration

Pervaporation

Reverse Osmosis

Gas Separation

27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

5

Size of Materials Retained, Driving Force, and Type of Membrane Process Microfiltration Ultrafiltration Nanofiltration Reverse Osmosis Dialysis

Electrodialysis Pervaporation Gas Permeation

Membrane Distillation 27 December 2015

Size of materials retained 0.1 - 10 µm microparticles 1 - 100 nm macromolecules 0.5 - 5 nm molecules < 1 nm molecules < 1 nm molecules

< 1 nm molecules < 1 nm molecules < 1 nm molecules < 1 nm molecules

Driving force Pressure difference (0.5 - 2 bar) Pressure difference (1 - 10 bar) Pressure difference (10 - 70 bar) Pressure difference (10 - 100 bar)

Type of membrane Porous Microporous Microporous Nonporous

Concentration diff.

Nonporous or microporous

Electrical potential difference

Nonporous or microporous

Concentration diff.

Nonporous

Partial pressure diff. (1 - 100 bar)

Nonporous

Partial pressure diff.

Microporous

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

6

Examples of Applications and Alternative Separation Processes Process Microfiltration Ultrafiltration Nanofiltration

Applications Separation of bacteria and cells from solutions Separation of proteins and virus, concentration of oil-in-water emulsions Separation of dye and sugar, water softening

Reverse Osmosis

Desalination of sea and brackish water, process water purification

Dialysis

Purification of blood (artificial kidney)

Electrodialysis

Separation of electrolytes from nonelectrolytes

Pervaporation

Dehydration of ethanol and organic solvents

Gas Permeation

Hydrogen recovery from process gas streams, dehydration and separation of air

Membrane Distillation

Water purification and desalination

27 December 2015

Alternative Processes Sedimentation, Centrifugation Centrifugation Distillation, Evaporation Distillation, Evaporation, Dialysis Reverse osmosis Crystallization, Precipitation Distillation Absorption, Adsorption, Condensation Distillation

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

7

Examples of Applications  This chapter is primarily concerned with the pressure driven processes, microfiltration (MF), ultrafiltration (UF), nanofiltration (NF) and reverse osmosis (RO).  These are already well-established large-scale industrial processes.  For example, reverse osmosis is used world-wide for the desalination of brackish water, with more than 1,000 units in operation.  Plants capable of producing up to 105 m3/day of drinking water are in operation.  it is now standard practice to include an ultrafiltration unit in paint plants in the car industry.  The resulting recovery of paint from wash waters can produce savings of 10–30 per cent in paint usage, and allows recycling of the wash waters. 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

8

The Nature of Synthetic Membranes  Membranes used for the pressure-driven separation processes, microfiltration, ultrafiltration and reverse osmosis, as well as those used for dialysis, are most commonly made of polymeric materials.  Initially most such membranes were cellulosic in nature.  These are now being replaced by polyamide, polysulphone, polycarbonate and a number of other advanced polymers.  These synthetic polymers have improved chemical stability and better resistance to microbial degradation.

27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

9

Membrane Synthesis Membranes have most commonly been produced by a form of phase inversion known as immersion precipitation This process has four main steps: 1. The polymer is dissolved in a solvent to 10–30 per cent by mass 2. The resulting solution is cast on a suitable support as a film of thickness, approximately 100 µm 3. The film is quenched by immersion in a non-solvent bath, typically water or an aqueous solution 4. The resulting membrane is annealed by heating The third step gives a polymer-rich phase forming the membrane, and a polymer-depleted phase forming the pores. The ultimate membrane structure results as a combination of phase separation and mass transfer, variation of the production conditions giving membranes with different separation characteristics. 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

10

The Nature of Synthetic Membranes Most microfiltration membranes have a symmetric pore structure, and they can have a porosity as high as 80 per cent. Ultrafiltration and reverse osmosis membranes have an asymmetric structure comprising a 1–2 µm thick top layer of finest pore size supported by a ∼100 µm thick more openly porous matrix. Such an asymmetric structure is essential if reasonable membrane permeation rates are to be obtained 27 December 2015

Electron micrograph of a section of an asymmetric ultrafiltration membrane showing finely porous “skin” layer on more openly porous supporting matrix.

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

11

The Nature of Synthetic Membranes Another important type of polymeric membrane is the thin-film composite membrane. This consists of an extremely thin layer, typically ∼1 µm, of finest pore structure deposited on a more openly porous matrix. The thin layer is formed by phase inversion or interfacial polymerisation on to an existing microporous structure.

27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

12

The Nature of Synthetic Membranes Polymeric membranes are most commonly produced in the form of flat sheets, but they are also widely produced as tubes of diameter 10–25 mm and in the form of hollow fibres of diameter 0.1–2.0 mm. A significant recent advance has been the development of microfiltration and ultrafiltration membranes composed of inorganic oxide materials. These are presently produced by two main techniques:

1. deposition of colloidal metal oxide on to a supporting material such as carbon, and 2. as purely ceramic materials by high temperature sintering of spraydried oxide microspheres. Other innovative production techniques lead to the formation of membranes with very regular pore structures. Zirconia, alumina and titania are the materials most commonly used. 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

13

The Nature of Synthetic Membranes The main advantages of inorganic membranes compared with the polymeric types are their higher temperature stability, allowing steam sterilization in biotechnological and food applications, increased resistance to fouling, and narrower pore size distribution.  The physical characterization of a membrane structure is important if the correct membrane is to be selected for a given application.  The pore structure of microfiltration membranes is relatively easy to characterize, atomic force microscopy and electron microscopy being the most convenient methods and allowing the threedimensional structure of the membrane to be determined.  The limit of resolution of a simple electron microscope is about 10 nm, and that of an atomic force microscope is <1 nm, 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

14

The Nature of Synthetic Membranes

AFM images of single pores in (a) microfiltration, (b) ultrafiltration and (c) nanofiltration membranes 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

15

The Nature of Synthetic Membranes

Surface SEM of Geopolymeric membrane for pore size. (B) Surface SEM micrograph of Ceramic membrane after curing and hydrothermal Treatment

Cross sectional view SEM of Ceramic membrane 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

16

The Nature of Synthetic Membranes A parameter often quoted in manufacturer’s literature is the nominal molecular weight cut-off (MWCO) of a membrane.

This is based on studies of how solute molecules are rejected by membranes. A solute will pass through a membrane if it is sufficiently small to pass through a pore, if it does not significantly interact with the membrane and if it does not interact with other, larger solutes. It is possible to define a solute rejection coefficient R by: where Cf is the concentration of solute in the feed stream and Cp is the concentration of solute in the permeate.

The nominal molecular weight cut-off is normally defined as the molecular weight of a solute for which R = 0.95. 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

17

General Membrane Equation The general membrane equation is an attempt to state the factors which may be important in determining the membrane permeation rate for pressure driven processes. This takes the form:

 J is the membrane flux∗, expressed as volumetric rate per unit area,  |ΔP| is the pressure difference applied across the membrane, the transmembrane pressure,  |ΔΠ| is the difference in osmotic pressure across the membrane,  Rm is the resistance of the membrane, and  Rc is the resistance of layers deposited on the membrane, the filter cake and gel foulants. If the membrane is only exposed to pure solvent, say water, then the equation reduces to: 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

18

General Membrane Equation For microfiltration and ultrafiltration membranes where solvent flow is most often essentially laminar through an arrangement of tortuous channels, this is analogous to the Carman–Kozeny equation used in conventional filtration In the processing of solutes, General Membrane Equation shows that the transmembrane pressure must exceed the osmotic pressure for flow to occur.

The separation of a solute by a membrane gives rise to an increased concentration of that solute at the membrane surface, an effect known as concentration polarisation. This may be described in terms of an increase in ΔΠ . 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

19

The Concept of Cross-Flow Microfiltration The concept of cross-flow microfiltration, described by BERTERA, STEVEN andMETCALFE, is shown in Figure below which represents a cross-section through a rectangular or tubular membrane module. The particle-containing fluid to be filtered is pumped at a velocity in the range 1–8 m/s parallel to the face of the membrane and with a pressure difference of 0.1–0.5 MN/m2 (MPa) across the membrane. The liquid permeates through the membrane and the feed emerges in a more concentrated form at the exit of the module.

Such a process allows the removal of particles down to 0.1 µm or less, but is only suitable for feeds containing very low concentrations of particles as otherwise the membrane becomes too rapidly clogged 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

20

Flow diagram for a simple cross-flow Microfiltration system

This is the system likely to be used for batch processing or development rigs and is, in essence, a basic pump recirculation loop. The process feed is concentrated by pumping it from the tank and across the membrane in the module at an appropriate velocity. The partially concentrated retentate is recycled into the tank for further processing while the permeate is stored or discarded as required. 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

21

Diafiltration In cross-flow filtration applications, product washing is frequently necessary and is achieved by a process known as diafiltration in which wash water is added to the tank at a rate equal to the permeation rate Diafiltration is an ultrafiltration membrane technique for completely removing, replacing, or lowering the concentration of salts or solvents from solutions containing proteins, peptides, nucleic acids, and other biomolecules. The process selectively uses permeable (porous) membrane filters to separate the components of solutions and suspensions based on their molecular size. Smaller molecules such as salts, solvents, and water pass freely through the ultrafiltration membrane, which retains the larger molecules 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

22

Membrane Fouling In practice, the membrane permeation rate falls with time due to membrane fouling; that is blocking of the membrane surface and pores by the particulate materials, as shown in Figure. The rate of fouling depends on the nature of the materials being processed, the nature of the membrane, the cross-flow velocity and the applied pressure. 27 December 2015

The time-dependence of membrane permeation rate during cross-flow filtration: (a) Low cross-flow velocity, (b) Increased cross-flow velocity, (c) Backflushing at the bottom of each “saw-tooth”

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

23

Membrane Fouling Ideally, cross-flow microfiltration would be the pressure-driven removal of the process liquid through a porous medium without the deposition of particulate material. The flux decrease occurring during cross-flow microfiltration shows that this is not the case. If the decrease is due to particle deposition resulting from incomplete removal by the cross-flow liquid, then a description analogous to that of generalized cake filtration theory, should apply. The general Membrane Equation (GME) may then be written as:

where Rc now represents the resistance of the cake 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

24

Membrane Fouling If all filtered particles remain in the cake, may be written as:

where r is the specific resistance of the deposit, V the total volume filtered, Vs the volume of particles deposited, Cb the bulk concentration of particles in the feed (particle volume/feed volume) and Am the membrane area. The specific resistance may theoretically be related to the particle properties for spherical particles by the Carman relationship as:

where e is the void volume of the cake and ds the mean particle diameter 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

25

Membrane Fouling Combing

and

gives Solution of this equation for V at constant pressure gives:

yielding a straight line on plotting t/V against V

SCHNEIDER and KLEIN have pointed out that the early stages of cross-flow microfiltration often follow such a pattern although the growth of the cake is limited by the cross-flow of the process liquid. 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

26

Membrane Fouling There are a number of ways of accounting for the control of cake growth. A useful method is to rewrite the resistance model to allow for the dynamics of polarization in the film layer . Equation

is then written as:

where Rsd is the resistance that would be caused by deposition of all filtered particles and Rsr is the resistance removed by cross-flow. Assuming the removal of solute by cross-flow to be constant and equal to the convective particle transport at steady state (=JssCb), then:

where Jss can be obtained experimentally or from the film-model 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

27

Membrane Fouling In a number of cases, a steady rate of filtration is never achieved and it is then possible to describe the time dependence of filtration by introducing an efficiency factor β representing the fraction of filtered particles remaining in the filter cake rather than being swept along by the bulk flow. Equation

becomes:

where 0 < β < 1. The layers deposited on the membrane during cross-flow microfiltration are sometimes thought to constitute dynamically formed membranes with their own rejection and permeation characteristics. 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

28

Ultrafiltration  Ultrafiltration is one of the most widely used of the pressuredriven membrane separation processes.  The solutes retained or rejected by ultrafiltration membranes are those with molecular weights of 103 or greater, depending mostly on the MWCO of the membrane chosen.  The process liquid, dissolved salts and low molecular weight organic molecules (500–1000 kg/kmol) generally pass through the membrane.  The pressure difference applied across the membrane is usually in the range 0.1–0.7 MN/m2 and membrane permeation rates are typically 0.01–0.2 m3/m2 h.  In industry, ultrafiltration is always operated in the cross-flow mode. 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

29

Ultrafiltration Concentration Polarization The separation of process liquid and solute that takes place at the membrane during ultrafiltration gives rise to an increase in solute concentration close to the membrane surface, as shown in Figure. This is termed concentration polarization and takes place within the boundary film generated by the applied cross-flow. With a greater concentration at the membrane, there will be a tendency for solute to diffuse back into the bulk feed according to Fick’s Law. 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

30

Ultrafiltration Concentration Polarization At steady state, the rate of back-diffusion will be equal to the rate of removal of solute at the membrane, minus the rate of solute leakage through the membrane: Here solute concentrations C and Cp in the permeate are expressed as mass fractions, D is the diffusion coefficient of the solute and y is the distance from the membrane. Rearranging and integrating from C = Cf when y = l the thickness of the film, to C = Cw, the concentration of solute at the membrane wall, when y = 0, gives:

Or: 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

31

Ultrafiltration Concentration Polarization If it is further assumed that the membrane completely rejects the solute, that is, R = 1 and Cp = 0, then:

where the ratio Cw/Cf is known as the polarization modulus. It may be noted that it has been assumed that l is independent of J and that D is constant over the whole range of C at the interface. The film thickness is usually incorporated in an overall mass transfer coefficient hD, where hD = D/l, giving:

The mass transfer coefficient is usually obtained from correlations for flow in nonporous ducts. 27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

32

Ultrafiltration Concentration Polarization One case is that of laminar flow in channels of circular cross-section where the parabolic velocity profile is assumed to be developed at the channel entrance.

where Sh is the Sherwood number (hDdm/D),dm is the hydraulic diameter, L is the channel length, Re is the Reynolds number (udmρ/µ), Sc the Schmidt number (µ/ρD), with u being the cross-flow velocity, ρ the fluid density and µ the fluid viscosity. This gives:

27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

33

Ultrafiltration Concentration Polarization or for tubular systems: where γ˙ is the shear rate at the membrane surface equals 8u/dm. For the case of turbulent flow the correlation will be like:

which for tubular systems gives: and for thin rectangular flow channels, with channel height b:

27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

34

Ultrafiltration Concentration Polarization  For both laminar and turbulent flow it is clear that the mass transfer coefficient and hence the membrane permeation rate may be increased, where these equations are valid, by increasing the cross-flow velocity or decreasing the channel height.  The effects are greatest for turbulent flow.  For laminar flow the mass transfer coefficient is decreased if the channel length is increased.  This is due to the boundary layer increasing along the membrane module. The mass transfer coefficient is, therefore, averaged along the membrane length.

27 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

35

Ultrafiltration Gel Polarization The boundary-layer theory applies to mass-transfer controlled systems where the membrane permeation rate is independent of pressure, for there is no pressure term in the model. In such cases it has been proposed that, as the concentration at the membrane increases, the solute eventually precipitates on the membrane surface. This layer of precipitated solute is known as the gel-layer, and the theory has thus become known as the gel-polarization model proposed by MICHAELS. Under such conditions Cw becomes replaced by a constant CG the concentration of solute in the gel-layer, and:

30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

36

Reverse Osmosis Osmosis is the spontaneous net movement of solvent molecules through a semi-permeable membrane into a region of higher solute concentration, in the direction that tends to equalize the solute concentrations on the two sides. Applying an external pressure to reverse the natural flow of pure solvent, thus, is reverse osmosis. Osmotic pressure is the minimum pressure which needs to be applied to a solution to prevent the inward flow of water across a semipermeable membrane. It is also defined as the measure of the tendency of a solution to take in water by osmosis. 30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

37

Reverse Osmosis The phenomenon of reverse osmosis has been extensively developed as an industrial process for the concentration of low molecular weight solutes and especially for the desalination, or more generally demineralization, of water. Reverse osmosis was originally developed for the desalination of brackish waters to produce potable water however, the improvements in membrane materials, flux, rejection coefficients and energy recovery techniques now make single pass extraction of potable water from seawater an economic proposition. Reverse osmosis is now used extensively in the food and dairy industries to perform a wide range of separation tasks including the concentration of fruit and vegetable juices, milk products and other heat sensitive biological materials. It is also employed in the micro-electronics and pharmaceutical industries for the production of ultra-pure water. 30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

38

Transport Through RO & NF Membrane  Several models on reverse osmosis ( RO) and Nanofiltration (NF) transport mechanisms and models have been developed to describe solute and solvent fluxes through RO/NF membranes.  The general purpose of a membrane mass transfer model is to relate the fluxes to the operating conditions .  The power of a transfer model is its ability to pr edict the performance of the membrane over a wide range of operating conditions.  To realize this objective, the model has to be integrated with some transport coefficients often determined based on some experimental results. 30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

39

Transport Through RO & NF Membrane  In both reverse osmosis and nanofiltration the solvent is transported through the membrane by application of a pressure.

 Because the solutes to be separated are small, reverse osmosis and nanofiltration membranes can be considered to be almost non-porous.  The separation mechanism is obviously not sieving, as the molecular volume of water and salt ions are approximately the same.

30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

40

Transport Through RO & NF Membrane From Fick’s law we can derive an expression for the flux of solvent which can be expressed as follows

Ds is the diffusivity of the solvent in the membrane cs is the concentration of the solvent in the membrane V~s is the partial molar volume and Δxm is the effective thickness of the membrane. For a given system operating isothermally these quantities can be taken as fixed and therefore the expression can be reduced to

Where A is called hydraulic permeability constant or permeability coefficient 30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

41

Transport Through RO & NF Membrane By a similar analysis we can derive the flux for the solute. In this case the partial molal volume of the solute is small and the difference in activity or concentration across the membrane is significant Thus the flux approximates to: Dsol is the diffusivity of the solute in the membrane Ksol is the solubility or distribution coefficient for the solute between the solution and the membrane. csolr and csolp are the solution concentrations in the retentate and permeate. For a given membrane solute system the term Dsol Ksol /Δx is constant and thus the flux for the solute can be written as: B is termed the permeation constant and the flux of the solute is almost independent of pressure. 30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

42

Transport Through RO & NF Membrane The result of the pressure dependence of solvent flux is that the rejection coefficient of the membrane for the solute is dependent on pressure. The rejection coefficient for the solute is normally written as:

30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

43

Membrane Modules and Plant Configuration  Membrane equipment for industrial scale operation of microfiltration, ultrafiltration and reverse osmosis is supplied in the form of modules.  The area of membrane contained in these basic modules is in the range 1–20 m2.  The modules may be connected together in series or in parallel to form a plant of the required performance. The four most common types of membrane modules are: 1. Tubular 2. Flat sheet

3. Spiral wound 4. Hollow fibre 30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

44

Membrane Modules and Plant Configuration The membrane module must satisfy a number of mechanical and economic requirements. The main issues to be satisfied can be summarized as follows: Hydrodynamic Optimization of solute transport and pressure drop. Minimization of fouling and long-term flux reduction. Ensuring properly distributed flow to avoid dead spots. Mechanical Housing and membrane support designed to withstand maximum anticipated working pressure. Design for reverse flow if necessary (backflushing). Avoidance of cross-contamination between retentate and permeate. Economic Optimization of initial membrane costs and operating life. Minimization housing costs and maximization of packing density. Designed for of ease of maintenance and replacement of membranes. 30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

45

Tubular modules  Tubular modules are preferred for solutions containing suspended solids and were amongst the first industrial designs.  The membranes are formed as tubes and are normally cast onto a supporting porous substrate.  The tubes are not self-supporting and are normally inserted in a perforated tube. The tubes are typically housed in a shell and tube configuration.  The membranes are normally sealed by means of elastomeric inserts at either end of the tube.  The permeate is collected in the shell side of the module.

 Tubular modules are normally operated in the turbulent region with a Reynolds number of >10,000 and fluid velocities in the region of 2-6 m/s. 30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

46

Tubular modules  The tubes are usually 8-12 ft in length  Tubular modules are preferred for applications where sanitary operation is of paramount importance because the design allows for easy cleaning and sterilization  The main drawbacks of tubular systems are the low surface to volume ratio and the high liquid hold-up. This limitation restricts the concentration factor that can be achieved in batch operation. 30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

47

Flat-Sheet Modules  Flat-sheet modules are similar in some ways to conventional filter presses.  This consists of a series of annular membrane discs of outer diameter 0.3 m placed on either side of polysulphone support plates which also provide channels through which permeate can be withdrawn.

 The sandwiches of membrane and support plate are separated from one another by spacer plates which have central and peripheral holes, through which the feed liquor is directed over the surface of the membranes, 30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

48

Flat-Sheet Modules  The flow is laminar.  A single module contains 19 m2 of membrane area.  Permeate is collected from each membrane pair so that damaged membranes can be easily identified, though replacement of membranes requires dismantling of the whole stack.

30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

49

Spiral-Wound Modules

 Spiral-wound modules consist of several flat membranes separated by turbulence promoting mesh separators and formed into a Swiss roll.  The edges of the membranes are sealed to each other and to a central perforated tube.  This produces a cylindrical module which can be installed within a pressure tube. 30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

50

Spiral-Wound Modules  These modules make better use of space than tubular or flat-sheet types, although they are rather prone to fouling and difficult to clean. The process feed enters at one end of the pressure tube and encounters a number of narrow, parallel feed channels formed between adjacent sheets of membrane.  Permeate spirals towards the perforated central tube for collection.  A standard size spiral-wound module has a diameter of some 0.1 m, a length of about 0.9 m and contains about 5 m2 of membrane area.  Up to six such modules may be installed in series in a single pressure tube.

30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

51

Hollow-Fibre Modules  Hollow-fibre modules, consist of bundles of fine fibres, 0.1–2.0 mm in diameter, sealed in a tube.  For RO desalination applications, the feed flow is usually around the outside of the unsupported fibres with permeation radially inward, as the fibres cannot withstand high pressures differences in the opposite direction.  This gives very compact units capable of high pressure operation, although the flow channels are less than 0.1 mm wide and are therefore readily fouled and difficult to clean. 30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

52

Hollow-Fibre Modules  The flow is usually reversed for biotechnological applications so that the feed passes down the centre of the fibres giving better controlled laminar flow and easier cleaning.  This limits the operating pressure to less than 0.2 MN/m2 however, that is, to microfiltration and ultrafiltration applications.  A single ultrafiltration module typically contains up to 3000 fibres and be 1 m long.  Reverse osmosis modules contain larger numbers of finer fibres.  This is a very effective means of incorporating a large membrane surface area in a small volume.

30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

53

Comparison Between Several Membrane Modules

30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

54

Membrane modules can be configured in various ways to produce a plant of the required separation capability. A simple batch recirculation system is most suitable for small-scale batch operation, but larger scale plants will operate either as feed and bleed or continuous single-pass operations. In Feed and bleed system,the start-up is similar to that in a batch system in that the retentate is initially totally recycled. When the required solute concentration is reached within the loop, a fraction of the loop is continuously bled off. 30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

55

 Feed into the loop is controlled at a rate equal to the permeate plus concentrate flowrates.  The main advantage is that the final concentration is then continuously available as feed is pumped into the loop.  The main disadvantage is that the loop is operating continuously at a concentration equivalent to the final concentration in the batch system and the flux is therefore lower than the average flux in the batch mode, with a correspondingly higher membrane area requirement.

30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

56

Continuous single pass. In such a system the concentration of the feed stream increases gradually along the length of several stages of membrane modules arranged in series as shown in Figure. The feed only reaches its final concentration at the last stage. There is no recycle and the system has a low residence time. Such systems must however, either be applied on a very large scale or have only a low overall concentration factor, due to the need to maintain high cross-flow velocities to control concentration polarization.

30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

57

30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

58

30 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

59

29 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

60

29 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

61

29 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

62

29 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

63

29 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

64

29 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

65

29 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

66

29 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

67

29 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

68

29 December 2015

Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan

69