13
CHAPTER 1
Membrane Separation Process
1.6 Introduction
1.7 Membrane Separation technology
1.8 Membrane Structure and morphology
1.4 Membrane processes that separate primarily based on size.
1.5 Membrane processes that separate based on principles other than size.
1.6 Various configurations of operating a filtration process
1.7 Membrane modules
1.8 Membrane characterization
1.9 Fundamental of membrane permeation
1.10 Mechanism of Separation through membranes
1.11 The phenomenon of liquid permeation
1.12 The phenomenon of gas permeation
1.13 Applications of Membrane
1.14 Reference
14
1.1 Introduction :
Membranes are the selective barriers normally used to separate two phases and restrict
the transport of various chemicals in a selective manner. (Fig.1.1). Membranes can be
homogenous or heterogeneous, symmetric or asymmetric in structure, solid or liquid, porous or
non porous. In case of porous membrane pore, size is the key parameters which determine the
effective ness and efficiency of the membrane. It can carry a positive or negative charge or be
neutral or bipolar. Transport through a membrane can be affected by convection or by diffusion
of individual molecules, induced by an electric field or concentration, pressure or temperature
gradient. Normally, separation occurs under a pressure gradient or sometimes under an electrical
potential gradient, associate with or without a catalytic reaction. System which can be separated,
are solid particles suspended in a fluid medium and mixture of two different liquids or gases.
Separation through a membrane is schematically shown in Fig 1.1. Porous membranes are
typically classified according to their pore sizes in the following manner:
Fig . 1.1 : A membrane as a selective barrier between two homogeneous phases [1]
1.2 Membrane Separation technology:
A membrane separation system separates an influent stream into two effluent streams known
as permeate and the concentrate. Permeate is the portion of the fluid that has passed through the
semi-permeable membrane. Whereas the concentrate stream contains the constituents that have been
Permeate (fluid)
Mixed phase feed
Porous
Membrane
15
rejected by the membrane. Membrane separation process enjoys numerous industrial applications
with the following advantages:
1. Appreciable energy savings
2. Environmentally benign
3. Clean technology with operational ease
4. It replaces the conventional processes like filtration, distillation, ion-exchange and
chemical treatment systems.
5. It produces high, quality products.
6. Greater flexibility in designing systems.
1.3 Membrane Structure and morphology: From a structural point of view membranes are broadly divided into two types as shown in Fig.
1.2:
a. Symmetrical, and
b. Asymmetrical (or anisotropic)
Symmetrical membrane has similar structural morphology at all positions within it. An
anisotropic membrane is constituted of two or more structural planes of non-identical
morphologies. From a morphological point of view, membranes can be porous or dense. Porous
membranes have been tiny pores or pore networks within themselves (see Fig. 1.3). On the other
hand, dense membranes do not have any pores and solute or solvent transport through these takes
place by a solubilization mechanism.
1.4 Driving force in membrane separation In order to drive the solutes and solvents through a membrane driving force is necessary. These
include:
1. Transmembrane (hydrostatic) pressure (TMP)
2. Concentration or electrochemical gradient
3. Osmotic pressure
4. Electrical field
5. Partial pressure
16
6. pH gradient
Fig. 1.2: Symmetrical and asymmetric membranes [2]
Fig. 1.3: Porous membranes (2)
1.5 Membrane processes that separate primarily based on size:
Membrane processes are divided into four types based upon the size of component in the feed
solution that is allowed to pass. With some overlap, the categorization, from largest to smallest
permeable species, is microfiltration, ultrafiltration, nanofiltration and reverse osmosis (see Fig.
1.4). The different applications are listed in Table 1.1. A fifth type of size based membrane
17
separation process called dialysis allows solutes similar those in reverse osmosis to pass through.
However, unlike reverse osmosis, which is a pressure driven process, dialysis is a concentration
gradient driven process.
Size, Microns 0.001 0.01 0.1 1 10
Microfiltration
Ultrafiltration
Nanofiltration
Reverse Osmosis
Salts
Virus PP Proteins
Flour Bacteria
Fig.1.4: “Size based” fractionation processes [3]
1.5.1 Microfiltration (MF)
Micro filtration (MF) is the process of removing particles or biological entities in the
0.025 µm to 10.0µm range from fluids by passage through a microporous medium such as a
membrane filter. Transmembrane pressures ranging from 1 to 50 psi are used as the driving
force. If the pore sizes of the membrane are smaller than the particles in the solution, surface
filtration results. Although micron-sized particles can be removed by use of non-membrane or
depth materials such as those found in fibrous media, only a membrane filter having a precisely
defined pore size can ensure quantitative retention. Membrane filters can be used for final
filtration or prefiltration, whereas a depth filter is generally used in clarifying applications where
quantitative retention is not required or as a prefilter to prolong the life of a downstream
membrane. Membrane and depth filters offer certain advantages and limitations. They can
complement each other when used together in a microfiltration process system or fabricated
Human Hair
Metal Ions
18
device. The retention boundary defined by a membrane filter can also be used as an analytical
tool to validate the integrity and efficiency of a system. For example, in addition to clarifying or
sterilizing filtration, fluids containing bacteria can be filtered to trap the microorganisms on the
membrane surface for subsequent culture and analysis. Microfiltration can also be used in sample
preparation to remove intact cells and some cell debris from the lysate. Membrane pore size cut-
offs used for these types of separation are typically in the range of 0.05 µm to 1.0 µm. [3]
1.5.2 Ultrafiltration (UF)
Ultrafiltration (UF) is the process of separating extremely small particles and dissolved
molecules from fluids. The primary basis for separation is the molecular size, although in all
filtration applications, the permeability of a filter medium can be affected by the chemical,
molecular or electrostatic properties for the sample. Ultra filtration can only separate molecules,
which differ by at least an order of magnitude in size. Molecules of similar size cannot be
separated by ultra filtration. Normal transmembrane pressure ranges from 10 to 100 psi. The
product can be the permeate, the retentate, or both Materials ranging in size from 1K to 1000K
molecular weights (MW) are retained by certain ultrafiltration membranes, while salts and water
will pass through. Colloidal and particulate matter can also be retained. Ultrafiltration
membranes can be used both to purify material passing through the filter and also to collect
material retained by the filter. Materials significantly smaller than the pore size rating pass
through the filter and can be dehydrogenated, clarified and separated from high molecular weight
contaminants. Materials larger than the pore size rating are retained by the filter and can be
concentrated or separated from low molecular weight contaminants. Ultrafiltration is typically
used to separate proteins from buffer components for buffer exchange, desalting, or
concentration. Ultrafilters are also ideal for removal or exchange of sugars, non-aqueous
solvents, the separation of free from protein-bound ligands, the removal of materials of low
molecular weight, or the rapid change of ionic and/or pH environment (see Figure 1.1).
Depending on the protein to be retained, the most frequently used membranes have a nominal
molecular weight limit (NMWL) of 3 kDa to 100 kDa. Ultrafiltration is far gentler to solutes
than processes such as precipitation. UF is more efficient because it can simultaneously
concentrate and desalt solutes. It does not require a phase change, which often denatures labile
species, and UF can be performed either at room temperature or in a cold room [3,4].
19
1.5.3 Nanofiltration (NF)
Nanofiltration is a liquid separation membrane technology positioned between reverse
osmosis (RO) and ultrafiltration. While RO can remove the smallest of solute molecules, in the
range of 0.0001 micron in diameter and smaller, nanofiltration (NF) removes molecules in the
range 0.001 micron. NF refers to a membrane process that rejects solutes approximately 1
nanometer (10 angstroms) in size with molecular weights above 200. Because they feature pore
sizes larger than RO membranes, NF membranes remove organic compounds and selected salts
at the lower pressure than RO system. It separates at the molecular level, removing all suspended
solids and most dissolved solids. Transmembrane pressures range from 40 to 200 psi (2.0 kg/cm2
to 14 kg/cm2. NF essentially is a lower-pressure version of RO where the purity of product water
is not as critical as with pharmaceutical grade water, or the level of dissolved solids to be
removed is less than what typically is encountered in brackish water or seawater. [3, 4]
1.5.4 Reverse Osmosis (RO)
It is used to remove dissolved solids from solvents. By applying transmembrane pressure
to concentrated solutions, it is possible to force the solvent through the RO membrane towards
the lower concentration. Hence the terms reverse osmosis. It is a separation process of small
(monovalent) ions and molecules (M < 300 Da) on so called “dense” membranes. The range of
the sizes of molecules that are separated with RO is 1–10 Å. The range of the transmembrane
pressure applied is 10–100 bars depending upon the concentration difference of the separated
species on both sides of the solution. Reverse osmosis (RO) is increasingly used in chemical,
textile, petrochemical, electrochemical, food, paper and tanning industries, as well as in the
treatment of tap water and wastewaters. Reverse osmosis is mainly used for water purification,
including ultrapure water production, desalination, water treatment, wastewater treatment and
landfill leachates treatment. After RO purification, pure water may be easily recovered from
wastewater and subsequently reused in various production steps. RO is used for industrial
effluents treatment, for water reuse, and for concentration of valuable products, for TDS and
COD removal from wastewater [2,3,5].
20
1.6 Membrane processes that separate based on principles other than size 1.6.1 Pervaporation (PV)
It is a process to separate a volatile or low-boiling-point liquid from a non-volatile liquid.
The driving force is a vacuum on the gaseous side of the membrane. It is a tool for separation of
liquid mixtures, especially dehydration of liquid hydrocarbons.
1.6.2 Dialysis
It is a membrane separation process in which one or more dissolved species flow across a
selective barrier in response to a difference in concentration. It is the earliest membrane based
molecular process to be developed. The mode of transport is diffusion, and separation occurs
because small molecules diffuse more rapidly than larger ones, and also because the degree to
which the membrane restricts the transport of molecules usually increases with solute size.
1.6.3 Electrodialysis (ED)
It is an electrochemical process used to separate charged particles from an aqueous
solution or from other neutral solutes. A stack of membranes is used, half of them passing
positively charged particles and rejecting negatively charged ones; the other half doing the
opposite. An electrical potential is imposed across the membranes, and a solution with charged
particles is pumped through the system. Positively charged particles migrate toward the negative
electrode, but are stopped by a positive-particle-rejecting membrane. Negatively charged
particles migrate in the opposite direction with similar results. Both types migrate in opposite
directions out of one set of cells and collect in the remaining cells. The result is a concentrated
solution of both positively and negatively charged particles in every other cell and a low
concentration (the product) in the remaining cells [2,3,5].
1.7 Various configurations of operating a filtration process:
1.7.1 Dead-end Filtration
The most basic form of filtration is dead-end filtration. The complete feed flow is forced
through the membrane, and the filtered matter is accumulated at the surface of the membrane.
The dead-end filtration is a batch process as the accumulated matter in the filter decreases the
21
filtration capacity, due to clogging. A next process step to remove the accumulated matter is
required. Dead-end filtration can be a very useful technique for concentrating compounds.
Table 1.1: Represents the characteristics of membranes used in different membrane separation processes, process driving forces and applications of such processes [3] Process Membrane Type
and Pore Radius Membrane Material
Process Diving Force
Applications
Microfiltration Symmetric microporous, 0.1-10 microns
Cellulose nitrate or acetate, Polyvinylidene difluoride (PVDF), Polyamides, Polysulfone, PTFE, Metal Oxides etc.
Hydro-static pressure difference at approx. 10-500 kPa
Sterile filtration, Clarification
Ultrafiltration Asymmetric microporous, 1-10 nm
Polysulfone, Polypropylene, Nylon 6, PTFE, PVC, Acrylic Copolymer
Hydrostatic pressure difference at approx. 0.1-1.0 Mpa
Separation of macromolecular solutions
Reverse Osmosis Asymmetric skin-type, 0.5-1.5 nm
Polymers, Cellulosic acetate, Aromatic Polyamide
Hydrostatic pressure difference at approx. 2-10 Mpa
Separation of salts and microsolutes from solutions
Electrodialysis Cation and anion exchange membrane
Sulfonated cross-linked polystyrene
Electrical potential gradient
Desalting of ionic solutions
Gas Separation Asymmetric homogeneous polymer
Polymers & copolymers
Hydrostatic pressure and concentration gradients
Separation of gas mixtures
Pervaporation Asymmetric homogenous polymer (A non-porous membrane)
Polyacrylonitrile, Polymers
Vapour pressure gradient
Separation of azeotropic mixtures
Nanofiltration Thin-film membranes
Cellulosic Acetate and Aromatic Polyamide
9.3-15.9 bar Removal of hardness and desalting
22
a) Co-current flow
b) Completely mixed flow
c) Counter current flow
d) Cross flow
.
e) Dead end flow
Fig. 1.5 : Types of ideal continuous flow used in membrane based separation [4]
R
P
F
S
R F
P
R
R F
S
P
P
P
F
F
23
1.7.2 Cross-flow Filtration
With cross-flow filtration, a constant turbulent flow along the membrane surface prevents
the accumulation of matter in the membrane surface. The membranes used in this process are
commonly tubes with a membrane layer on the inside wall of the tube. The feed flow through the
membrane tube has an elevated pressure as the driving force for the filtration process and a high-
flow speed to create turbulent conditions. The process is referred to as "cross-flow", because the
feed flow and filtration flow direction have a 90 degrees angle. Cross-flow filtration is an
excellent way to filter liquids with a high concentration of filterable matter.
1.7.3 Hybrid-flow Filtration
The hybrid flow process combines the dead-end and the cross-flow principle. As in the
cross-flow filtration tubular membranes are with the filtration layer on the inside wall are used.
The filtration process has two phases: the production phase and the flushing phase. During the
production phase, the tubes are closed at one side, and a dead-end filtration is performed. During
the flushing phase, the tube is open to both sides and the fraction that did not pass through the
membranes is removed in order to clean the membrane surface as in cross-flow filtration. This
filtration technique is especially suitable for treating water streams containing suspended solids in
low concentrations (polishing).
1.7.4 Submerged Filtration
With submerged membrane filtration, the membranes are submerged in the liquid that has
to be filtered. The filtration is performed from the outside to the inside of the membrane
(filtering layer is on the outer side of the tube or plate). Sheer forces along the membrane surface
are created by a flow of air bubbles on the surface. In some cases, the airflow also results in a
liquid flow created by the airlift principle. The driving force is a vacuum applied on the inner
side of the membrane [2,3,5].
1.8 Membrane modules
A key characteristic of membrane processes is the membrane geometry in the actual
equipment, which provides the membrane housing and the desired hydrodynamic conditions.
24
Membrane processes can be classified upon the basis of the flow pattern within the membrane
module. These are shown in Fig 1.5
Membrane modules can be of different types (see Table1.2 for comparison). Some of these are:
1. Stirred cell module
2. Flat sheet tangential flow (TF) module
3. Tubular membrane module
4. Spiral wound membrane module
5. Hollow fibre membrane module
1.8.1 Stirred cell modules
Stirred cell modules are useful for small scale and research applications. These are more
commonly used for UF and MF. Stirred cell modules provide uniform transmembrane pressure
and hydrodynamic conditions at all points of the membrane surface. The effects of process
parameters on process efficiency can be very easily determined using stirred cell units. These are
therefore very useful for small-scale process development work. However, these are not of much
use in intermediate and large-scale operations.
Fig.1.6 Stirred Cell module (9)
25
1.8.2 Flat sheet tangential flow (TF) module
The flat sheet TF design is similar to that of a plate and frame filter press. It consists of
alternate layers of membranes, support screens (corrugated structural sheets) and distribution
chambers for feed and permeates. The units may have square, rectangular or oval cross-section.
These units can be easily disassembled for cleaning and for replacement of defective membranes.
Other advantages include the ability to accommodate low levels of suspended solids and viscous
fluids. Disadvantages include relatively low packing density. TF designs are commonly used for
UF, MF and NF. Electrodialysis and electrochemical membranes use only this configuration.
The liquid flow pattern in a flat sheet TF unit can be quite complex. The pressure drop
distribution is also quite difficult to characterize. Consequently, design and analysis of such
devices are largely based on empirical methods.
Fig 1.7 Schematic Flat sheet tangential flow (TF) module [4]
26
1.8.3 Spiral wound membrane module
The spiral wound membrane module is like a huge envelope made of membrane and
containing a feed spacer. Feed flowing around the envelope at high pressure goes across the
membrane and is collected inside the envelope. Permeate flows inside the envelope into the roll
and then runs out the end of the module. The feed and permeate flow patterns are usually quite
complex. Design and analysis is therefore largely empirical. Advantages include high packing
density and relatively low cost. Disadvantages include problems handling suspended solids,
difficulty in cleaning and, in high-temperature applications (plastic components may deform).
Another major problem is that these devices cannot usually handle high transmembrane pressure.
Fig. 1.8 Schematic of Spiral wound module [4]
1.8.4 Tubular membrane module
The tubular membrane module has a cylindrical geometry with the wall acting as the
membrane. The tubes are generally more than 3 mm in diameter. Normally, a tubular membrane
module is made up of several tubes arranged as in a shell and tube type heat exchanger. The feed
stream enters the lumen of the tubes, the permeate passes through the walls and is collected in
the shell while the retentate passes out the other end of the tubes. Advantages include turbulent
flow (providing good membrane/solution contact and removing retentate film build-up),
27
relatively easy cleaning, easy handling of suspended solids and viscous fluids and ability to
replace or plug a failed tube while the rest of the system runs. The feed flow pattern is easy to
characterize and therefore design and analysis based on fluid dynamic principles is possible.
Tubular membrane modules can handle reasonably high transmembrane pressure. Disadvantages
include high capital cost, low packing density, high pumping costs, and limited achievable
concentrations.
Fig. 1.9: Schematic drawing illustrating the tubular membrane module [4]
1.8.5 Hollow fibre membrane module
The hollow fibre membrane module is similar in design to the tubular membrane module
except in terms of scale and number of tubes. The tubes or rather fibres are typically 0.25 to 2.5
mm in diameter. A hollow fibre membrane module usually consists of a bundle of several
hundred fibres. The hollow fibre membranes are spun separately, bundled, and potted into
permeate
feed
concentrate
porous tube
membrane
permeatefeed solution
28
cartridge housings. The fibres may be bundled together into either a U-shape or have a straight-
through configuration. Tube bundles are inside a pressure vessel and feed material normally
flows inside the tubes. Fibres in the straight-through design are somewhat larger and allow low
levels of suspended solids. The finer strands in the U-tube cannot tolerate suspended solids. U-
tubes tend to be used for reverse osmosis, and the straight-through design for ultrafiltration.
Advantages include low pumping power, very high packing density; cleaning can be
accomplished with back-flushing, and ability to achieve high concentrations in the retentate.
Disadvantages include the fragility of the fibres, inability to handle suspended solids well,
difficult cleaning and, in the straight-through design, damage of one fibre requires replacement
of the entire module [2,3,5,7].
Fig. 1.10 Schematic drawing illustrating the construction of a hollow fiber module [4]
epoxy resin
hollow fiber
permeate
feed solution
concentrate
shel l tube
29
Table 1.2: Comparison of different membrane modules
Type Fluid flow regime
Membrane area/module volume
Mass transfer coefficient
Hold-up volume
Special remarks
TF flat sheet
Laminar- turbulent
Low Low to moderate
Moderate Can be dismantled and cleaned easily.
Spiral wound
Laminar Moderate Low Low High pressures cannot be used.
Hollow fibre
Laminar-turbulent
High Low to moderate
Low Susceptible to fibre blocking.
Tubular Turbulent Low Moderate to high
Moderate to high
Flow easy to characterize. Excellently suited for basic membrane studies.
1.9 Membrane characterization
The performance of a membrane process depends on the properties of the membrane. Thus
membrane characterization is an important exercise form membrane developers and membrane
users. Some of these include:
1. Mechanical strength e.g. tensile strength, bursting pressure
2. Chemical resistance e.g. pH range, compatibility with solvents
3. Permeability to different species e.g. pure water permeability, gas permeability
4. Average porosity and pore size distribution
5. Sieving properties e.g. Nominal molecular weight cut-off
6. Electrical properties e.g. membrane zeta potential
30
1.10 Fundamental of membrane permeation: When two multi component mixtures containing atleast one fluid are separated by a permeable
membranous barrier (e.g. a solid porous film) such that the partial molar free energy of one or
more components common to both the mixtures differs across the barrier, there will generally be
a selective transfer of components through the barrier in the derection of the declining chemical
potential for each. A few different situations may arise: (i) If the upstream and down stream
fluids are gas or vapour mixtures at differing hydrostatic pressures, for example, components
pass through the barrier in the direction of declining partial pressure by a process termed as “gas
permeation”; (ii) If the upstream fluid is liquid mixture, and the downstream fluid is a gas in
which the partial pressure of the permeating component is lower than the vapour pressure over
the component of the upstream liquid, transmembrane permeation occurs by a process usually
termed as “pervaporation”. (iii) If the liquid phases in contact of the membrane are in the same
hydrostatic pressure, component can transfer across the barrier under the action of concentration
difference between the contacting liquids by the process of “dialysis”. (iv) If a liquid mixture is
confined at a higher hydrostatic pressure from one side of the barrier compared to that of the
other, certain component of the upstream liquid will permeate through the membrane by the
process of “ultrafiltration or reverse osmosis”[2,3,5,7].
All these membrane transport process have the common features of (a) transport the mass across
the membrane under a specific driving force e.g. pressure difference, concentration diference,
applied e.m.f. etc. (b) the capacity to alter mixture composition by virtue of the ability of the
membrane barrier to pass one component more rapidly than another despite equality of driving
potential. It is this unique characteristic of membrane separation which differentiates the process
from the common separation techniques. Various driving forces involved in different separation
processes (including membrane separation) and their useful size ranges are presented in Fig. 1.10
31
Fig. 1.11 : Different membrane separation process relative to size [8]
1.11 Mechanism of Separation through membranes
In a membrane separation process involving a mixture of gas or liquid and solids there are five
separation mechanisms which can be described from micro porous membrane.
i) Geometric exclusion
ii) Donan Exclusion
iii) Adsorption / Difusion
iv) Knudsen diffusion
v) Capillary condensation
vi) Surface Flow
32
1.11.1 Geometric exclusion (Sieving)
Generally, geometric exclusion mechanism takes place on the basis of the size effect. The
pore size of the membrane should be smaller than the particle size which is to be separated. Most
of the membrane based on the geometric exclusion principle. This is a very similar phenomenon
which occurs in the normal filters. However, pore sizes in this case, are much finer than that in
common filters. Obviously it increases the complexity of the process. Pore size of reverse
osmosis or gas separation membrane is less than 1 nm and ultrafiltration has pore size in the
range of 2 – 100nm, microfiltration membrane has pore size in the range of 100 – 5000nm and
the membrane with pore size greater than 5000nm are used as particulate filter. As the pore size
becomes finer, the problem of clogging and fouling is much of serious nature than in the
common filter and since the mass transport takes place through much finer constrictions, fluid
dynamic also become complex in nature. The pore surface also plays an important role in such
membranes. Some of the specific situations with reference to the liquid and gas separation
membrane are discussed later.
1.11.2 Donan Exclusion
The separation of charged particle takes place by the Donana exclusion process. In this
process soluble ions are separated from a solution. One familiar example of such process is
dialysis. For example, when we dialyse a colloidal electrolyte e.g. sodium salt of dye acid against
the pure water, then sodium ion penetration through the membrane, where is the colloidal –
dimensioned dye stuff anion is retained. However , it is not possible for a large amount of
sodium ion to be separated from dye stuff anion as this would result in an excess positive ion at
one side of the membrane and an excess negative ion on the other side. This would be against the
rule of electrochemistry. To retain the electrical neutrality, water molecule must be splited in a
quantity equal to the amount of sodium ion that passes through the membrane. The OH- ion will
migrate with Na+ ion and equivalent number of H+ ions associating with dye anion. The result is
high concentration dye acid in one compartment and that of sodium hydro-oxide I the other. This
phe omenon is known as membrane hydrolysis and is a special case of Donnan equilibrium and
the process is called Donnan exclusion.
33
1.11.3 Adsorption / Diffusion
In case of adsorption / diffusion mechanism, firstly molecules are absorbed on the surface
of the membrane so that interfacial fluid layer is enriched in one of the constituents of the bulk
solution. A continuous removal of this interfacial layer by flow under pressure through the
capillary of te membranes results in a product solution whose composition is different from the
bulk solution. In this capillary flow mechanism it is very important that the thickness of the
sorbed layer and the capillary diameter bear a critical relationship, Dp = 2t where Dp is the
capillary diameter and t is the absorbed layer thickness. If the capillary diameter is much larger
than the sorbed layer, separation can not occur efficiently. Thus depending on the layer thickness
and the pore size, adsorbed molecules are separated from the non absorbed component. The
schematic representation of the process is illustrated in Fig 1.5.
++ t ++ + ++ ∆P ++ + ++++ ++++++ + ++ + +++++++ ++ +++ t ++++ Dp < 2t Dp = 2t Dp>2t Fig. 1.12 An example of diffusion of adsorbed particles (o) through a membrane of varying pore (Dp) sizes. Effective separation for Dp < 2p (2) 1.11.4 Knudsen diffusion
For gases, pure Kudsen diffusion takes place if the mean free path of molecules is much
larger than the average pore size (50 to 100 times). In such cases total diffusion is determined by
Kundsen diffusion. Separation efficiency is proportional to the ratio of the square root of the
mass of the molecule as mean free path is related to the mass of the molecules. High separation
efficiency is obtained from mixture of gases like H2, He with heavy gases like CO2, SO2 etc. A
combined transport by Knuundsen diffusion and adsorption / diffusion is possible for gases [1-
5,7].
DP
34
1.11.5 Capillary condensation
Preferential sorption-capillary flow mechanism was the fundamental approach for the
practical development of reverse osmosis process for converting sea water into potable water. If
the surface of the membrane in contact with the solution is of such a chemical nature that it has
preferential sorption for water or preferential repulsion for solutes then a multimolecular layer of
preferentially sorbed pure water could exist at the membrane surface. Continuous removal of this
interfacial water can then be effected by letting it flow under pressure through membrane
capillaries.
Capillary condensation can occur according to the Kelvin equiation [3-5].
4γV RgT ln (p / p0 ) = ─ ------------ cos φ (1) Dp Where p is the critical vapour pressure at pore diameter Dp
P0 is the saturated vapour pressure at a temperature T
Rg is the gas constant
The γ is the surface energy of the condensed vapour
V is the molar volume of the condensed vapour
Φ is the angle of contact between the liquid qnd pore wall.
Separation occur in such cases by the difference between the vapour pressure of the
component in the mixture. For example in Fig.1.7 one of the gas in the mixture condensed at
high pressure side (p2). And then the condensd liquid evaporated through the membrane to the
lower gas pressure side (p1). Because of the difference in vapour pressure the other component
does not condense and the separation efficiency increases [5].
35
Fig. 1.13 Schematic representation of vapour condensation and transport through a Microsporus membrane (2). 1.11.6 Surface Flow
Permeation rates of gaseous molecules through the membranes may be enhanced in some
cases by surface diffusion mechanism. In surface diffusion, gas molecules adsorb to the surface
of the pores of the membranes and then diffuse along the pore walls. High selectivity can be
achieved in cases where preferential adsorption of one of the components occurs. Surface flow is
the result of solution diffusion mechanism and is used to describe the preferential permeation
rate of some gaseous components through membranes from a given feed mixture and is similar
to preferential sorption capillary flow mechanism described previously for liquid feed systems.
Separation of condensable gases and vapours from noncondensible permanent gases through
membranes is often explained by surface flow. Such a separation is carried out at a suitable
temperature and pressure near or below the critical temperature of the gas. At high pressures,
multilayer adsorption of gas molecules occurs at the pore walls reducing the size of the pores. At
a particular pressure, the entire pore volume is filled with the permeating gas and the gas
molecules are said to have condensed inside the membrane pores, effectively blocking the
passage of permanent gases. Such a phenomenon is known as capillary condensation [8- 12]
PpP p2<p1
p2
p1
membrane
36
1.12 The phenomenon of liquid permeation: The principle of separation by ultrafiltration and microfiltration, is based on the sieving
concept. The mean diameter of the pore for ultrafiltration membrane is in the range 20 A° to
1000 A° and the range is 0.1 µm to 10 µm. for the microfiltration membrane.
For membrane separation, the permeate flow rate is directly proportional to the
permeability of the membrane which in turn is related to the membrane pore diameter., the
transmembrane pressure differential and the effective membrane area. With reference eto a cross
flow membrane module, the transmembrane pressure differential (∆p) is schematically
represented by Fig.
For many MF application with membrane pore diameters in the range 0.2 – 1 µm and all UF
applications where pore diameters are typically in the range 40 – 500 A°, it can be shown that the
permeate flow through the membrane is laminar (13).
PERMEATE
Recirculative Loop
Fig.1.14 Simplified schematic of cross flow a membrane filtration module.
1 2
3
P + P Δp = - P2
(2)
Where P1 = tube side (feed) inlet pressure
P2 = tube side (retenate) pressure
P3 = permeate outlet pressure
RETENATE
Cross Flow Membrane
P1 P3 P2
FEED
37
The flow of liquid through the pore in a filter can be analysed by considering the
laminar flw through a capillary. The flow through such a system is described by the Hagen –
Poiseuille equation (14).
4π r g V pQ = 8 η l
(3)
where, Q is the flow rate of the fluid
r is the radius of a capillary
g is the acceleration due to gravity,
Δp is the pressure drop over the cylinder
η is the viscosity of the fluid
l is thwe length of each capillary.
If the process in a filter are considerd to be uniform and cylindrical, the porosity e of
the filter havig an over all surface area A is related to the number of cylindrical pores (N) in this
area and the pore radius r by the following equation :
e A = N π r2 (4)
For this system equation ( 4) becomes
2Nπr gVpQ =
8ηl (5)
Combination of equations ( 4 ) and ( 5) yields
2er g pQ = 8ηl
(6)
38
The so called Kozeny – Carman equation which can be derived from equation (6),
describes laminar flow through less uniform porous media (13).
3
2 20
V 1 t A 1 (1 )t v
pek k S e
(7)
Here A is the external surface area of the porous medium exposed to the high or
low pressure side of the permeant, l is the thickness of the porous medium, e is its volume
fraction composed of pores and Sv is the internal surface area of the porous medium per unit
volume of the non porous medium. The product of k0 kt is define as K, which is the so called
Kozeny – Carman constant. The numerical value of K depends on the shape of the pores and the
tortuosity of the membrane. If the porous medium is composed of a regular packing of more or
less isomatrically shaped particles, then K is close to 5 (15). So it can be concluded that the
permeability of a porous material of certain pore size and porosity can give an idea about the
structure of the porous medium [13].
Shimizu et al. [16] calculate the pore radius and tortuosity of the membrane from the
flux through the membrane and the rates of capillary rise of several solvents. Tortuosity values
by this method agreed well with that obtained by measurement of ionic conductivity. However,
pore radius obtained by this method was 1.17 times larger than that obtained by Hg intrusion
method. Pores in the membrane are irregular in shape because the membrane were prepared by
packing particles where as by Hg intrusion methods measured pores are considered to be
cylindrical.
The permeate flux through a membrane increases with Δp upto a point which is
generally termed as threshold pressure. Beyond the threshold value he flux becomes independent
of Δp. This phenomenon is often known as concentration polarosation [17,18] Below the
threshold pressure range, the Kozeny – Carman equation shows a linear relationship with the
tranmembrane pressure drop and flow rate is inversely proportional to the viscosity of the
permeating fluid and also the thickness of the membrane. Thus for a given size of the membrane
the permeability will reduce for a higher viscous fluid at a constant temperature and membrane
thickness.
39
Such membrane based separation devices have to satisfy certain performance criteria
to justify the cost effectiveness of the process. There are (1) Flux of filtration rate (2) long term
flux stability and (3) separation and rejection properties of membrane [19].
Flux stability is perhaps the most critical factor determining the efficiency of the
separation process and also its economic viability. In MF and UF applications, flux decay can be
a serious problem [18,20]. Flux decay is usually a direct result of an increase in the hydraulic
resistance of the membrane due to fouling i.e. excessive accumulation of debris / particulates at
he membrane surface or in the pores. Such clogging of the membrane can be prevented by
adding certain flocculating agents which absorb or flocculate the particles which are responsible
for clogging [21].
Fouling can be somewhat minimize by cross flow filtration. Due to continuous
removal of filtrate / permeate through the membrane, the cross flow velocity is somewhat higher.
An increase in cross flow velocity generally is due to higher shear rate, he removal of
particulates at the membrane surface is more effective [22].
In the filtration process, interaction takes place between the membrane surface and
the suspended particle. When a mixture is brought in contact with membrane surface by any
driving force, some molecule permeate through the membrane while the other are retained. This
leads to an accumulation of the retained components and a polarosation layer appear at the solid
liquid interface with a higher concentration of the retanate near the membrane surface. During
filtration process, a steady state is reached where the transport of the solutes towards the
membrane surface is counter balance bya diffusive flow of the rejected and accumulated
materials from the membrane surface back in to the bulk solution. Therefore a constant
concentration profile of the rejected materials is obtained in the boundry layer. Concentration of
the solute at the surface (Cw) is always higher than I the bulk solution (CB). However, interaction
phenomenon of the solution or the solute with materials of the pore wall, is called the
concentration – polarosation of the solute in the vicinity of the separating layer. The
concentration profile of a solute I a polarosation surface is schematically describe in Fig.
The membrane resistance to the flow can be described by the following equation
which is derived from the Hagen – Poiseuille equation [14].
40
m
1 pR = = Q J
(8)
where, Rm is membrane resistance
Q is permeability
Δp is transmembrane pressure
J is permeate flux
One of the most important parameter determining membrane rejection coefficient is
the mean pore diameter of UF and MF membrane having narrow pore size distribution. In
general smaller the pore diameter, higher is the rejection co-efficient to particulate / solutes of
greater normal size.
Membrane Concentration Boundry Layer
Cw
:
Flux
Cn
Distance
Fig. 1.15: Schematic representation of concentration polarization and the concentration profile during membrane filtration [22]
41
1.13 The phenomenon of gas permeation :
For non sorbable gases the flow of gas through the media is governed by the Knundsen
and Poiseuille equation. For a homogeneous media the total permeability Fₒ (mol / m² -sec-Pa)
i.e. given by [24,25].
Fₒ = Fₒk + F1op . p (9)
Where Fₒk is the permeability due to Knundsen diffusion and F1op the contribution due to laminar
kok
2εμ rvF = 3RTL
(10)
2
1 µ r8
popF
RTL
(11)
Where ε is the porosity, r is the modal pore radius of the medium, µk and µp shape
factor, which are both, in general, assumed to be equal to the reciprocal tortuosity of the medium,
R the gas contant, T the temperature, p the mean pressure, η the viscosity of the gas and v the
molecular velocity expressed as
128 R Tv = ( )
η M (12)
Where M is the molecular mass of the gas.
The relative contribution of laminar flow to the total flow is determined by the ratio of
the mean free path of the gas molecules to the modal pore size. In case of small pore, at low
pressure and high temperatures, Knundsen diffusion is the predominant transport mechanisim.
Pure Knundsen diffusion appear if the mean free path of the molecules is much larger
than the average pore size of the membrane [26]. In this case the total diffusion is determined by
Knundsen diffusion according to equation described below:
42
AAD KA
dCN DdX
(13)
where NAD is the molar diffusive flux of component A
CA is the molar concentration of component
DKA is the Knundsen diffusion co-efficient and can be expressed as
12
KA PA
8RT1 2-fD = D 3 ηM f
(14)
Where MA is the molecular mass of component A
R is the gas constant
T is the temperature
f is the fraction of molecules with diffuse reflection on the pore wall.
Separation of gases is possible by Knundsen diffusion process. The separation efficiency
is proportional to the ratio of the square root of the mass of the molecule. High efficiency factor
can be obtained only with light gases in mixture with heavier gases.
43
1.14 Applications of Membrane
The membrane processes are used for separation of chemical substances and are of immense
interest for a wide range of commercial applications. The analytical applications of membranes
are grossly over shadowed by the industrial applications of membranes. A complete and
comprehensive discussion of all membrane based applications would be very exhaustive and is
beyond the scope of this unit. However, a few major large scale applications of membranes will
be discussed here.
1.14.1 Desalination and Water Treatment
Desalination of sea water (containing approximately 35000 to 40 000 milligrams of dissolved
salts per litre) and brackish waters (containing approximately 5000 to 10000 milligram of
dissolved salts per litre) to produce potable water is one of the widely known industrial
applications of membrane processes. Reverse osmosis processes to a large extent has been
successfully used for the last two decades in many parts of the world. In addition to desalination,
reverse osmosis process is also used for the treatment of municipal waste waters and effluents
from the various Separation chemical industries.
1.14.2 Protein Recovery
The most prominent use of ultrafiltration process is in the food industry, where it is used ,for
example to recover proteins from cheese whey, to concentrate milk before cheese making and for
fruit juice clarification.Other important applications of UF process is in the removal of colour ,
odour and bacteria from the surface waters for drinking water needs. The pore size of most of the
UF membranes used for such applications are such that they remove virtually all the
microorganisms present in the water and viruses to a large extent.
1.14.3 Production of Table Salt
Electrdialysis was first developed for the desalination of saline waters, in particular brackish
water. The production of potable water is currently the most important industrial application of
44
electrodialysis. One significant feature of electrodialysis is that the salts can be concentrated to
comparatively high values (in excess of 18 to 20 %) . The production of table salt from sea water
by the use of electrodialysis to concentrate sodium chloride up to 200 grams per litre prior to
evaporation is a technique developed extensively and used in Japan.
1.14.4 Hemodialysis
The most important application of dialysis process is in the artificial purification of blood using a
dialysis membrane and this process is called hemodialysis. In hemodialysis, the impure blood
from the patients flows across a dialysis membrane and a physiological saline solution flows
along the other side of membrane. This process replaces kidney function in three principal areas,
namely, removal of waste metabolites, removal of excess body water and restoration of acid-base
and electrolyte balances. The waste metabolites include urea, uric acid, the end product of
protein metabolism and creatinine, the end product of muscle metabolism.
1.14.5 Ion Selective Membrane Electrode
The analytical applications of membranes are largely in the area of ion selective membrane
electrodes and particulate analysis. For most of the ion selective membrane electrodes, the role of
membranes is not to transport specific ions but to selectively adsorb on either side giving rise to
measurable electrical potential difference. A particular membrane permits only a particular kind
of ion to penetrate and adsorb. Ion selective membrane electrodes have many applications in
water analysis and environmental monitoring.
1.14.6 Specific Gas Probes
Membranes are also used in specific gas probes for measuring dissolved gases and gas phase
partial pressures. The specific gas analysis is based on the electrochemical oxidation or reduction
of the gas at the appropriate electrode of an electrochemical cell giving rise to a current whose
magnitude depends on the concentration of the gas consumed in the electrochemical reaction.
The whole detection system is protected by the membrane which is permeable to the particular
gas component of interest. The membrane can be polytetrafluoro ethylene, polyethylene,
polypropylene, nylon or cellophane. The membranes typically have pore sizes of several
45
microns. Portable devices for specific gas analysis are available for detection of gaseous oxygen,
hydrogen sulphide, carbon dioxide, sulphur dioxide, nitrogen dioxide and hydrogen cyanide.
1.14.7 Detection and Analysis of Particulate Contamination
Microfiltration membranes offer a general way of removing particulate material from fluid
streams and are routinely used in a range of analytical procedures to determine particulate
contamination in gases and liquids. In general, these methods involve passing a representative
sample of liquid or gas through a suitable membrane. All particulate matter which exceeds the
membrane pore sizes are retained on the surface where the contaminants may then be analysed.
1.14.8 Microbiological Analysis
Microporous membranes are important for the detection of microorganisms in foods, beverages
pharmaceutical products and potable water sources. The technique involves the filtration of the
samples through a microfiltration membranes to trap the microorganism, then culturing the
microorganism on the membrane and then counting the grown colonies [27].
46
1.15 Reference :
1. Keizer K and Burggraaf A. J., Science of ceramics, vol 14 etd. by Dereck Taylor(1991).
2. Ghosh R K, Membrane separation processes, J Chem. Engg, McMaster,page 3(2005-06).
3. Porter C. Mark, Handbook of Industrial Membrane Technology (1990)
4. Sirkar K K, Membrane Handbook, Van Nostrand Reinhold, New York(1992).
5. Coulson and Richardson’s Chemical Engineering Volume 2 (4th edition), J.M. Coulson
etal. Butterworth-Heinemann, Oxford(1991).
6. Keith Scott, Handbook of Industrial Membranes, (1995) Elsevier Advanced Technology,
Oxford,
7. Keizer K., Leenaars A.F.M.and Burgraff. Ceramic in Advance Energy Technologies,
Edited by H. Krockel, M. Merz, O Van der Biest, (Proceeding of European Colloquium,
20 – 22 Sept. 1982,, Petten, NetharlandD. Radial Publishing Co. (1984)..
8. Balakrishnan M., Agarwal G.P. and Coony C.P., Membrane Sci. 3, 85, 112(1993).
9. Van Halle E., Gaseous Diffusion for Aerodynamic Separation of gases and isotop, Van
Carman Instt. For Fluid Dyanamics, 8th Lecturer Series(1978).
10. Aeaeda M, L., Du D. and Usijima M., Proceedings 4th Int. Drying Sym., Edited by R.
Toel A.S. mazumdar, 2, 472(1984).
11. Aeaeda M, L., Du D. and M. Fuji, J. Chem. Eng., Jpn, 19, 84(1986).
12. Aeaeda M and L., Du D., J. Chem. Eng., Jpn, 19, 72(1986).
13. Leenaars A.F.M. and Burggraaf A.J, J Membrane Sci. 24, 245(1985).
14. McCabe W L and Smith J C, Unit operation of Chem. Engineering, 3rd editon, (1976).
15. Carman P C, Trans. Inst. Chem. Eng. 15, 150(1987).
16. Porter M. C. in Microfiltration in Bungay et al. (eds), Synthetic Membrane : Science,
Engineering and publication, Reidel, Dordrecht, 225(1986).
17. Lee S., Aurelle Y and H., Roques J Membrane Sci., 19, 23(1984).
18. Hsish H.P., Bhave R.R. and Fleming H.L., J Membrane Sci., 39,221(1998).
19. Cheriyan M and Kuo K. P., J Dairy Sci. 67, 1406 (1984).
20. Paul Hurly and High J, J Technol, 7(8), 21, (1987),.
21. L.Cot, British ceramic Proceedings, edt by B.C.Steele and D.P.Thomson, no. 43,
December p 111(1988).
22. Paulson D. J., Wilson R. L.and D. D. Spatz, J Food Technol, 38, 77(1984).
47
23. Tamon H, Kyotani S, Wade H, Okazaki M and Toei R, J Chem. Eng. Jpn, 14, 136(1981).
24. Rowell R. L., Carrano S. A., Bethune A. J. de and Malinauskas A. P. , J Colloid
Interface sci., 37, 242(1971).
25. Keizer K, Ulhorn R.J.R., Van Varun R. J. and Burggraaf A. J., J Membrane Sci39,
285(1988).
26. Korngold E., in “ Synthetic Membrane Processes” Edt. By George Delfort, Academic
Press. Inc. (1984).