1

CHAPTER 1
INTRODUCTION

1.1 Background of Study

The nanofiltration (NF) membrane is a type of pressure-driven membrane with

properties in between reverse osmosis (RO) and ultrafiltration (UF) membranes. NF
provides several benefits such as low operating pressure, high flux, high retention of
multivalent anion salts and an organic molecular above 300, relatively low investment and
low operation as well as economical maintenance costs. Because of these advantages, the
applications of NF worldwide have increased (Lu et.al, 2002).

Nanofiltration membranes have been applied in various fields such as desalination,

distillation, purification, concentration, pervaporation, gas separation, wastewater
treatment, removal or recovery of toxic or valuable components from various industrial
components, removal of natural organic matter (NOM) from water, blood heamodialysis
instrument, bio-separation, food processing and etc.

Membrane materials with high permeability and selectivity, and advanced of

fabrication technologies to produce membrane with a defect free ultrathin dense selective
layer are the primary focuses for most membrane technologists. Significant progresses have
been made in membrane materials, dope preparation, fabrication technology, and
fundamental understanding of membrane formation.

Choosing an appropriate membrane materials is one of the key factors to the

success of membrane performance. Basically, for producing asymmetric membrane
structure components that must consist in the dope formulation are polymer, solvent and
non-solvent(s). Among the polymers available, polyetherimide (PEI) was chosen in this
study because of their particular demand in the development of both UF and NF
 2

membrane. The aromatic imide units of PEI provide high-performance properties such as
considerable mechanical strength, thermal stabilities and chemical resistances. While the
flexible ether linkages provide good processability. In fact, PEI has been successfully used
in preparation of asymmetric membranes for gas separation and pervaporation. (Peinemann
et.al, 1998; Feng & Huang, 1996).

For the membrane fabrication technique, the phase inversion process induced by
immersion precipitation is a well-known technique for preparing asymmetric polymeric
membranes. Through the immersion of a substrate in a coagulation bath, a solvent in the
casting solution is exchanged with a non-solvent in the precipitation media, and phase
separation occurs. As reported by Kim and Lee (2001), they fabricated integrally skinned
uncharged PEI asymmetric NF membranes by dry/wet phase inversion method. The phase
inversion process must be carefully controlled in attaining a desired membrane morphology
and performance.

A key factor that determined the phase inversion path of a membrane-forming

system is the composition of the membrane-forming system, i.e. casting solution and the
coagulation media. Many researchers have been attempting to explain the membrane-
forming mechanism during the phase inversion process. For instance, Reuvers et al.(1987)
developed a model which explained the two types of demixing taking place during the
phase inversion process, instantaneous demixing and delayed demixing. Membranes
formed by instantaneous demixing generally show a highly porous substructure (with
macrovoids) and finely porous, thin skin layer. Whereas, delayed demixing mechanism
resulted a porous (often closed-cell, macrovoid free) substructure with dense, relatively
thick skin layer.

In membrane fabrication, addition of non-solvent additive into dope formulation

might be useful due to its influence to physical properties of membrane such as pore
formation, pore distribution by suppression of macrovoids or microvoids evolution. Kim et
al.(2001) studied the effect of non-solvent additives such as diethylene glycol dimethyl
ether (DGDE) and acetic acid (AA) on surface morphology of PEI asymmetric flat
 3

nanofiltration membrane. They reported that the macrovoids formation could be suppressed
by adding DGDE in the casting solution.

An introduction of appropriate additive(s) into the casting solution is very

convenient and one of the efficient method in order to produce a membrane with certain
structures and permeation properties. It have been reported that additives can affect the
final membrane structures either by changing solvent capacity or by changing phase
inversion kinetics and thermodynamics properties. Some researchers have also reported
their studies on the role of additives in membrane structures formation. Cranford et
al.(2000) studied on the effects of polyvinylpyrrolidone(PVP) towards physical and
chemical properties of PEI membranes. It was showed that PVP could leaks out into the
coagulation bath whilst phase separation occurred. As a result, a non-uniform PVP
distribution membrane was produced.

Appropriate procedures are required in developing desired membrane structures and

characteristics so that high performances membranes according to their applications can be
obtained. These include changing of the casting solution compositions in term of
(polymer/solvent/non-solvent) content in every dope formulation, coagulation bath
parameters (temperature, pH, bath compositions) and also controlling casting conditions
such as shear rates and evaporation time.

1.3 Problem Statements

Because of the economic competitiveness of the existing separation technologies

and the present challenges of aggressive environments for membranes, many applications
seek more robust membrane materials having higher selectivity and permeability. Not
surprisingly, membrane materials development is an important research area to synthesize
both higher performance and more durable membranes. Therefore, one of the empirical
rules for the formation of optimized membrane using this technique is the dope solution
must consist of at least three components: a polymer, a solvent and a non-solvent additive.
In this study, poly(etherimide) was selected because it was proved to provide high
performance properties such as adequate mechanical strength, thermal stability and
 4

chemical resistance as well as good processability besides gaining a huge demands in

membrane fabrication for in various applications worldwide.
PEI has been successfully used in gas separation and pervaporation because it
exhibits high-selectivities for all important gas pairs as reported by Barbari et al.(1989). In
this study, PEI was being used in preparation of nanofiltration membrane for ions selection.
Though several studies have been done for preparation of nanofiltration PEI membrane, but
there is still no study on multivalent salts rejection.
The additive can be a single component or a mixture. Generally, the additive being
a weak non-solvent for the polymer, reduces the solvent power in the solution. A number of
researchers (Boom et al., 1992; Wienk et al., 1996; Jimenez et al., 2004; Yeo et al., 2000)
have reported their observation on the role of additives in the membrane structures. DGDE
as reported by Kim et al.(2001) can suppress by the macrovoids formation in both sizes and
numbers. Hence, in this study non-solvent additive that is Diethyleneglycol Dimethylether
(DGDE) was introduced into the dope solution with different compositions which range
from (3%, 6%, 9%,12%,15% wt/wt). Based on the increment of DGDE contents in dope
formulation, the optimum content of DGDE can be studied whereby we can see how
DGDE can play roles in formation of final membrane structures and key performances i.e
pure water permeation(PWP) and multivalent salts rejection as the content of DGDE kept
increased.

1.4 Objectives

1) To develop newly dope formulations based poly(etherimide) for nanofiltration.

2) To determine the effect of non-solvent additive towards membrane morphologies and
performances
3) To investigate the role of polymeric additive in resulting of membrane final structures
and performances.

1.5 Scope of Study

1) Developing PEI based polymer solutions in ternary and quarternary components system.
2) Fabricating symmetric nanofiltration (NF) membranes based on phase inversion process.
 5

3) Analyzing the influence of different contents of non-solvent additive towards membrane

structures and key properties.
4) Characterizing the membranes morphologies via Scanning Electron Microscope (SEM).
5) Identifying the membrane selectivity through multivalent salts rejection (NaCl, MgCl2
Na2SO4 and MgSO4) and pure water permeation(PWP).

CHAPTER 2
LITERATURE REVIEW

2.1 Historical Development of Membranes

Synthetic membranes are a rather recent development and the technical utilization of
membrane processes on a large scale began just 40 years ago. The first recorded study of
membrane phenomena and the discovery of osmosis dates back to the middle of the 18th
century when Nollet discovered that a pig’s bladder passes preferentially ethanol when it
was brought in contact with a water-ethanol mixture and on the other side with pure water
(Nollet, 1752). Nollet was probably the first to recognize the relation between a
semipermeable membrane and the osmotic pressure. More systematic studies on mass
transport in semi-permeable membrane were carried out by Graham who studied the
diffusion of gases through different media and discovered that rubber exhibits different
permeabilities to different gases (Graham, 1866).

Most of the early studies on membrane permeation were carried out with natural
materials such as animal bladders or gum elastics. Traube was the first to introduce an
artificially prepared semi-permeable membrane by precipitating cupric ferrocyanide in a
thin layer of porous porcelain (Traube, 1867). This type of membrane was used by Pfeffer
in his fundamental studies in osmosis (Pfeffer, 1877). The theoretical treatment and much
of the interpretation of osmotic phenomena and mass transport through membranes is based
on the studies of Fick who interpreted diffusion in liquids as a function of concentration
 6

gradients, and van t’Hoff who gave a thermodynamic explanation of the osmotic pressure
of dilute solutions (Fick, 1855 and van’t Hoff, 1887). Little later Nernst and Planck
introduced the flux equation for electrolytes under the driving force of a concentration or
electrical potential gradient (Nernst, 1888 and Planck, 1890). With the classical
publications of Donnan describing the theory of membrane equilibria and membrane
potentials in the presence of electrolytes, the early history of membrane science ends with
most of the basic phenomena satisfactorily described and theoritically interpreted (Donnan,
1911).

With the beginning of the twentieth century, membrane science and technology
entered a new phase. Bechhold developed a method of making the first synthetic
membranes by impregnating a filter paper with a solution of nitrocellulose in glacial acetic
acid (Bechhold, 1908). These membranes could be prepared and accurately reproduced
with different permeabilities by varying the ratio of acetic acid to nitrocellulose.
Nitrocellulose membranes were also used in the studies of Zsigmondy as ultrafilters to
separate macromolecules and fine particles from an aqueous solution (Zsigmondy et al.,
1918). These studies were later continued by many others (Elford, 1931 and McBain et al.,
1931). The relation between the streaming potential, electro-osmosis and electrodialysis
were treated in a monograph (Prausnitz et al., 1931).

A milestone in membrane science and technology was the development of a reverse

osmosis membrane based on cellulose acetate which provide high salt rejection and high
fluxes at moderate hydrostatic pressures (Reid et al., 1959 and Loeb et al., 1964). This was
a major advance towards the application of reverse osmosis membranes as an effective tool
for the production of potable water from sea.

The membrane technology was developing successful in 20 years, from 1960 until
1980. The original Loeb-Sourirajan membrane technology and other process were
developed for making ultra thin, high performance membrane. Interfacial and multilayer
composite casting and coating was used. At that time, in order to make membrane with a
thickness of 0.1 pm is impossible except for a number of company.
 7

Methods of packaging membranes into large-membrane-area spiral-wound, hollow-

fine-fiber, capillary and plate-and-frame modules were also developed, and advances were
made in improving membrane stability. As a result, by 1980 microfiltration, ultrafiltration,
reverse osmosis and electrodialysis were all established processes with large plants
installed around the world. The principal milestone in the 1980s was the emergence of
industrial membranes gas separation processes. The first major development was the
Monsanto Prism membranes for hydrogen separation, introduced in 1980s, (Schäfer et al.,
2004).

Table 2.0 : Historical development of membrane based on year.

Year Membrane Development

1748 Nollet Osmosis
Semipermeable animal membranes (microporous and hydrophilic)

1803 Reuss Electro-Osmosis

1855 Fick Diffusion
Collodion Membrane (first synthetic membrane)

1861 Graham Dialysis

1887 Van’t Hoff Osmotic Pressure

1889 Nernst-Planck Electrolyte Transport

Einstein Osmotic Pressure
Henderson membrane potentiol

1892 Rubbery Membranes (homogeneous and hydrophopic)

1905 Donnan Membrane Equilibrium

1907 Sollner Anomalous Membrane

Nitrocellulose Membrane(Graded Pore Size)

1911 Pressure Retard-Osmosis

Asymmetric Cellulose Acetate Membrane

1920 Microfiltration
1930 Ultrafiltration
1950 Haemodialysis

1960 Electrodialysis
Reverse Osmosis
 8

1962 Asymmetric Membrane

Schmid Pore Model

1965 Lonsdale Solution-Diffusion Model

1970 Composite Membrane

Teorell Ionic Membrane

1980 Gas Separation

Membrane Distillation

1990 Pervaporation

2000-2020 Nanofiltration
Membrane Bioreactors
Membrane Contactors

2.2 Basic Concept of Membranes

Membrane can be defined and stated as a semi-permeable active or passive barrier

which, under a certain driving force, permits preferential passage of one or more selected
species or components (molecules, particles or polymers) of a gaseous and/or liquid
mixture or solution. The primary species rejected by the membrane is called retentate(s) or
sometimes just "solute" while those species passing through the membranes is usually
termed permeate(s) or sometimes "solvent". The driving force can exist in the form of
pressure, concentration, or voltage difference across the membrane (Baker et al., 1991).
The driving force will be explained further later.

A membrane can be divided into two types which are cell membrane and synthetic
membrane. The cell membrane is a semi permeable lipid bilayer which can be found in all
cells (Ghosh, 2003). Meanwhile, the synthetic membrane is a membrane that being
prepared for separation task in laboratory and industry. It can be homogeneous or
heterogeneous, symmetric or asymmetric in its structure, and it may be either neutral, or
may carry positive or negative charges, or both. Its thickness may vary between less than
100 nm to more than a centimeter. The electrical resistance of a membrane may vary from
thousands of megaohms to fractions of an ohm. Mass transfer through a membrane may be
 9

caused by diffusion of individual molecules or by convective flow induced by an electric

field, or a concentration, pressure or temperature gradient.

Separation in membrane processes are the result of differences in the transport rates
of chemical species through the membrane interphase. The transport rate is determined by
the driving force or forces acting on the individual components and their mobility and
concentration within the interphase (Kedem and Katchalsky, 1961). The mobility and
concentration of the solute within the interphase determine how large a flux is produced by
a given driving force. The mobility is primarily determined by the solute’s molecular size
and the physical structure of the interphase material (hardness, crystallinity, polymer chain
flexibility, etc.) while the concentration of the solute in the interphase is primarily
determined by chemical compatibility of the solute and the interphase material (Strathmann
and Michaels, 1977).

Membrane processes can be operated in two major modes according to the direction
of the feed stream relative to the orientation of the membrane surface: dead-end filtration
and crossflow filtration (Figure 1.0). The majority of the membrane separation applications
use the concept of crossflow where the feed flows parallel to and past the membrane
surface while the permeate penetrates through the membrane overall in a direction normal
to the membrane. The shear force exerted by the flowing feed stream on the membrane
surface help remove any stagnant and accumulated rejected species that may reduce
permeation rate and increase the retentate concentration in the permeate. Predominant in
the conventional filtration processes, dead-end filtration is used in membrane separation
only in a few cases such as laboratory batch separation. In this mode, the flows of the feed
stream and the permeate are both perpendicular to the membrane surface (Baker et al.,
1991).

(a) Crossflow Filtration

 10

(b) Dead-end Filtration

Figure 2.2: (a) Crossflow Filtration and (b) Dead-end Filtration

2.3 Membranes for Separation Process

2.3.1 Microfiltration

MF has the largest pore size (0.1 – 3 microns) of the wide variety of membrane filtration
systems. MF fills in the gap between ultrafiltration and granular media filtration. In terms
of characteristic particle size, this MF range covers the lower portion of the conventional
clays and the upper half of the range for humic acids. This is smaller than the size range for
bacteria, algae and cysts, and larger than that of viruses. MF is also typically used for
turbidity reduction, removal of suspended solids, giardia and cryptosporidum.This process
require low transmembrane pressure (1–30 psi) to operate, and now used as a pretreatment
to desalination technologies such as reverse osmosis, nanofiltration, and electrodialysis
(http://www.trusselltech.com/tech_desalination.asp)

2.3.2 Ultrafiltration

UF membranes were studied at the laboratory scale or small scale as early as 1907, but they
were first prepared with an initial goal of producing high-flux RO membrane. Most UF
membranes have an asymmetric porous structure and are often prepared by the phase-
inversion process. They have a finely porous surface layer or skin supported on a much
more microporous subtrate. The finely porous surface layer performs the separation as the
microporous substrate provides mechanical strength. The membrane discriminate between
dissolved macromolecules of different sizes and are usually characterized by their
 11

molecular-weight cut off, a loosely defined term generally taken of the mean molecular
weight of globular protein molecule that is 95% rejected by the membrane. Ultrafiltration
membranes are used to separate small colloidal particles and dissolved macromolecules
with diameter below 1000 armstrong. Pore diameter of the membrane can be inferred from
the cut-off value. The pore diameters usually have range from 10-1000 armstrong. While,
the driving force in UF is in the range of 1-10 bars. (Baker, 2012)

2.3.3 Reverse Osmosis

Reverse osmosis (RO) is a pressure-driven process aimed at the separation of ionic solutes
and macromolecules from aqueous streams.The method of dissolved salt removal is
different to that of microfiltration and is not just a physical process based on size difference
of solute and solvent.RO membranes can essentially separate all solute species, both
inorganic and organic from solution.The mechanism for separation of species are based on
processes relating to their size and shape, their ionic charge and their interactions with the
membrane itself. The membranes used for hyperfiltration are either asymmetric or
composite which typically have a <1 µm thick, dense top layer supported by a 50-150 µm
thick porous sublayer.The particle size range for applications of RO is approximately
0.0001-0.001 µm and with solutes of molar masses greater than 300 Da complete
separation is achieved (Wang et al., 2010).

2.3.4 Nanofiltration

The term NF was introduces by FilmTec in the second half of the 1980s to describe a “RO
process’ that selectively and purposely allows some ionic solutes in a feed water to
permeate through. It was derived from the membrane’s selectivity towards non-charged
solutes of approximately 1 nm cut-off. Within the few years, other membrane scientists
began using the word and Eriksson was one of the first who used it explicitly in 1988. As a
matter of fact, such membranes with selectivity between the RO and UF regions already
existed in the 1960s and were categorized as open, loose, low-pressure RO, intermediate
RO and UF, or tight UF membranes. Different from RO membranes which has non-porous
structure and transport mechanism of solution-diffusion, NF membranes operate at the
 12

interface of porous and non-porous membranes with both sieving and diffusion transport
mechanisms. NF membranes have a high permeability for monovalent salts (e.g.,
NaCl,KCl), but they are able to eliminate multivalent salts near completely and remove
relatively small organic compounds. NF was used as an alternative to RO for the
concentration and demineralization of whey in dairy industry. Until 1990, NF membranes
had found their more applications in seawater softening, food, textile and mining industries.
The NF-50 membrane was the first example of RO membrane capable of operation at UF
membrane pressures. Characteristics of this membrane include 30-40% NaCl rejection, 85-
90% MgSO4 rejection,98% sucrose rejection and 99% raffinose rejection. However, they
did not allow application in organic solvent due to the lack of chemical resistance of the
membrane polymers against the solvents. The most significance drawback of NF
membranes is to control the reproducibility of the membrane pore size and the pore size
distribution. Moreover, NF membranes are liable to fouling, possibly resulting in important
flux decline.

2.3.5 Forward Osmosis

In the FO process, a net water movement occurs through a semi-permeable membrane from
a low concentration solution to a high concentration solution under the osmotic pressure
gradient across the membrane (Cath et al., 2006). Compared with conventional pressure-
driven membrane processes, such as reverse osmosis (RO), the FO process offers the
advantages of no external hydraulic pressures for operation, nearly complete rejection to a
wide range of contaminants, and possibly lower membrane fouling propensity. As a result,
FO has been intensively studied recently for several potential applications, such as
wastewater treatment, water purification and seawater desalination. However, in these
applications FO requires an extra step to separate the water from the draw solution
(McCormick et al., 2008). Commercial FO membranes produced by the Hydration
Technologies Inc. (HTI) have been widely used in FO studies. The HTI’s proprietary
cartridge-type FO membrane is made of cellulose triacetate supported by an embedded
polyester screen mesh. In recent years, many researchers have worked on fabricating novel
FO membranes in both hollow fiber and flat sheet configurations (Shi et al., 2011; Widjojo
et al., 2011; Elimelech & Cutcheon, 2008). For example, thin-film composite FO hollow
 13

fibres with an ultra-thin RO-like skin layer on the surface of a porous hollow fiber substrate
were reported by Wang and her team.

Figure 2.3: Relationship between membrane pore sizes and separated species.

Table 2.3 : Classification of membrane separation process based on their membrane type and
pore radius, membrane materials, process driving force and applications.

Process Membrane type Membrane Process Applications

and pore radius materials driving force
Symmetric Cellulose nitrate
Microfiltration microporous, or acetate, Hydrostatic Bacteria and cyst
0.1-1.0 microns PVDF, polyamides, pressure filtration,
PTFE,metal oxides, differenceat clarification, potable
etc. approximate water treatment, RO
10-30 psi pretreatment

Ultrafiltration Asymmetric Polysulfone, Hydrostatic Separation of

microporous, polypropylene, pressure macromolecular
0.001-0.1 microns nylon-6,acrylic difference solutions, RO
copolymer at approximate pretreatment, potable
30-100 psi water treatment

Reverse Osmosis Asymmetric Thin-film Hydrostatic Separation of salts

skin-type, composite, pressure and micro-solutes
cellulose-acetate, difference at from solutions
 14

5 to 10 angstroms aromatic polyamide approximate

200-1000 psi

Electrodialysis Cation and Sulfonated Electrical Desalting of ionic

anion-exchange cross-linked potential solutions
membrane polystyrene gradient

Gas separation Asymmetric Polymer Hydrostatic Separation of gas

homogenous and copolymers pressure and mixture
polymer concentration
gradient

Pervaporation Asymmetric Polyacrylonitrile, Vapour Separation of

homogenous polymers pressure azeotropic mixture
membrane gradient
(anon-porous
membrane)

Nanofiltration Thin-film TFC 50-150 psi Removal of hardness

membranes and desalting

2.4 Types of membrane

2.4.1 Isotropic membrane

In isotropic membrane, the structure and the transport properties are identical over the cross
section and the thickness of the entire membrane determines the flux. Today, isotropic
membranes are mainly used in dialysis and electrolysis.

2.4.2 Anisotropic membrane

In asymmetric membrane, structural as well as transport properties vary over the cross
section of membrane. An asymmetric membrane consists of a 0.1-1.0 µm thick skin layer
on a highly porous 100-200 µm thick sub-structure. The skin represents the actual selective
barrier of the asymmetric membrane. Its separation characteristics are determined by the
nature of the materials or the size of pores in the skin layer. The mass flux is determined
 15

mainly by the thickness of the skin. The porous sub-layer serves only as a support for the
mostly thin and fragile skin and has little effect on the separation characteristics or the mass
transfer rate of the membrane. Asymmetric membranes are primarily used in a pressure-
driven membrane processes such as reverse osmosis, ultrafiltration, or gas and vapor
separation, since here the unique properties of asymmetric membranes, that is, high fluxes
and good mechanical stability, can be utilized.

Two techniques are used to prepare asymmetric membranes: one utilizes the phase
inversion process, which leads to an integral structure with the skin and the support
structure made from the same material in a single process, and the other resembles a
composite structure where a thin barrier layer is deposited on a porous substructure in a
two-step processs. In this case, barrier and support structures are generally made from
different materials.

Figure 2.4: Classification scheme of synthetic membranes based on their geometry, bulk
structure, production method, separation regime and application.
 16

2.5 Module types of membranes and characteristics

In order for a membrane to be used in certain application, it has to be installed in a

proper device which is generally referred to as membrane module. Membrane modules
must meet certain requirements as far as their production costs, their packing density,
energy consumption, and especially the control of concentration polarization and
membrane fouling is concerned.

However, on a large industrial scale they are several types used. These modules are
shown schematically in figure 2.5. They are quite different in their design, their mode of
operation, their production costs, and the energy requirements for pumping the feed
solution through the module. A very important criteria is the control of concentration
polarization and membrane fouling. In some modules, such as the tubular, the plate-and-
frame and the capillary type module, concentration polarization and membrane fouling can
effectively be controlled by the proper feed flow. Other modules, such as the spiral-wound
and the hollow fiber module, are more sensitive for membrane fouling. There is not one
module type that serves all membrane processes and applications. The different modules
which are commercially available today are designed for a certain application and process
in which they provide the technical and commercially best solution.

2.5.1 Plate and frame

Another module type used on an industrial scale for various membrane separation
processes including ultrafiltration, reverse osmosis, and gas separation is the plate-and-
frame module. Its design has its origin in the conventional filter press-concept. The
membranes, porous membrane support plates, and spacers forming the feed flow channel
are clamped together and stacked between two endplates and placed in a housing as
indicated in the schematic diagram of Figures 2.5.1.1 The feed solution is pressurized in the
housing and forced across the surface of the membrane. The permeate is leaving the
module through the permeate channel to a permeate collection manifold which in circular
devices is central tube as indicated in the Figure 2.5.1.2. Often the device contains one or
more baffels to extend the path-length of the feed solution in the device.
 17

Figure 2.5.1.1 : Schematic drawing illustrating the concept of a plate-and-frame membrane

module.

Figure 2.5.1.2 : Circular plate-and-frame filter device with one baffel to extend the feed flow
path length.

2.5.2 Spiral-wound

A variation of the basic plate-and-frame concept is spiral-wound module, which is widely

used today in reverse osmosis, ultrafiltration, and gas separation. Its basic design is
illustrated in Figure 2.5.2(a). The feed flow channel spacer, the membrane, and the porous
membrane support form an envelope which is rolled around a perforated central collection
tube and inserted into an outer tubular pressure shell. The feed solution passes in axial
direction through the feed channel across the membrane surface. The filtrate is moves along
 18

the permeate channel and is collected in a perforated tube in the center of the roll. Small
spiral wound units consist of just one envelope which limits the total membrane area that
can be installed in one unit to about 1 to 2 m2. The main reason for limitation of surface
area which can be installed in a module containing one single envelope is the pressure drop
encountered by the permeate moving down the permeate channel to the central collection
tube. Because the channel in a practical unit is very narrow its length is limited to 2 to 5 m.
A significantly longer path would result in an unacceptable pressure drop in the permeate
channel. To install larger membrane surfaces in a spiral wound module a multi-leaf
arrangement in used as indicated in the Figure 2.5.2.2.

Figure 2.5.2.1 : Cross section of a spiral-wound membrane module.

 19

Figure 2.5.2.2 : Schematic drawing illustrating the construction of a multi-leaf spiral wound
module.

2.5.3 Tubular

The tubular membrane module consists of membrane tubes placed into porous stainless
steel of fiber glass reinforced plastic pipes. The pressurized feed solution flows down the
tube bore and the permeate is collected on the outer side of the porous support pipe, as
indicated in Figure 2.5.3. The diameters of tubular membranes are typically between 1-2.5
cm. In some modules, the membranes are casted directly on the porous pipes and in others
they are prepared separately as tubes and then installed into the support pipes. Today,
tubular modules are used in ultrafiltration at low hydrostatic pressures. This allows the
membrane tubes to be made by a welding or glueing procedure of flat sheet membranes
that are cast on a relatively thick and mechanacilly strong porous polyester support
material. These tubes which have a diameter of 0.5 to 1 cm do not need additional support
when operated at hydrostatic pressures of less than 2 to 4 bars.
 20

Figure 2.5.3: Schematic drawing illustrating the tubular membrane module

2.5.4 Hollow-fiber

The same basic spinning process is used for the preparation of hollow fiber membranes,
which have an outer diamter of 50 to 100 μm. In hollow fiber membranes, the selective
layer is on the outside of the fibers, which are installed as a bundle of several thousand
fibers in a half loop with the free ends potted with an epoxy resin in a pressure tube as
indicated in Figure 2.5.4. The filtrate passes through the fiber walls and flows up the bore
to the open end of the fibers at the epoxy head.
The hollow fiber membrane module has the highest packing density of all module types
available on the market today. Its production is very cost effective and hollow fiber
membrane modules can be operated at pressures in excess of 100 bars. The main
disadvantage of the hollow fiber membrane module is the difficult control of concentration
polarization and membrane fouling.
 21

Figure 2.5.4 : Schematic drawing illustrating the construction of a hollow fiber

module

When operated with liquid solutions the modules will not tolerate with any discrete
particles, macromolecules or other materials that may easily precipitated at the membrane
surface. Therefore, an extensive pretreatment is required when hollow fiber membranes are
used for the treatment of liquid mixtures. The main application of the hollow fiber module
today is in reverse osmosis, desalination of sea water and in gas separation. Both
application require high operating pressures and low-cost membranes which have a long
useful life. In reverse osmosis of sea water, an extensive pretreatment of the sea water is
required.

2.6 Membrane Formation by Phase Inversion

Since Loeb and Sourirajan first introduced the phase inversion method, tremendous amount
of researches have been conducted in order to understand the mechanism of membrane
formation. The common technique for the preparation of asymmetric nanofiltration (NF)
membranes is the phase inversion method (Barth et al., 2002). The development of the first
integral asymmetric membranes by phase inversion was a major breakthrough in the
development of ultrafiltration and reverse osmosis. These membranes were made from
cellulose acetate (CA) and yielded fluxes 10 to 100 times higher than symmetric structures
with comparable separation characteristics.
 22

The phase inversion process induced by immersion precipitation is a well-known

technique to prepare asymmetric polymeric membranes (Mulder, 1997). By immersion of a
substrate into a coagulation bath, solvent in the casting solution film is exchanged with
non-solvent in precipitation media and phase separation occurs. This process results in an
asymmetric membrane with a dense top layer and a porous sub-layer. The sub-layer
formation is controlled by numerous variables in the polymer dope solutions such as
composition, coagulant temperature and organic/inorganic additives.

A number of methods can be used to achieve phase inversion. Among these, the
dry–wet phase inversion technique is the most commonly used in membrane
manufacturing. The dry–wet phase inversion technique, also called the Loeb-Sourirajan
technique, was used by Loeb and Sourirajan in their development of the first cellulose
acetate membrane for seawater desalination (Loeb and Sourirajan, 1963). In this method, a
polymer solution is prepared by mixing polymer and solvent (sometimes even non-solvent).
The solution is then cast on a suitable surface by a doctor blade to a precalculated
thickness. After a partial evaporation of the solvent, the cast film is immersed in a non-
solvent medium called a gelation bath. Due to a sequence of two desolvation steps, i.e.,
evaporation of the solvent and solvent–nonsolvent exchange in the gelation bath,
solidification of the polymer film takes place. It is desirable to choose a suitable solvent in
which it has a strong dissolving power with a high volatility. During the first step of
desolvation by solvent evaporation, a thin skin layer of solid polymer is formed instantly at
the top of the cast film due to the loss of solvent.

In the solvent–nonsolvent exchange process that follows, the non-solvent diffuses

into the polymer solution film through the thin solid layer while the solvent diffuses out.
The top skin layer can also be made porous by reducing the polymer concentration in the
casting solution and the solvent evaporation period. This is called, hereafter, the porous
skin layer.

The phase inversion process involves a phase separation of a polymer solution in

polymer rich and lean phases, which can be achieved by the immersion-precipitation
technique. A polymer solution is immersed in a non-solvent bath for the polymer, where a
 23

mass transfer process involving interchange of solvent/non-solvent occurs. The simplest

system is formed by three components, polymer/solvent/nonsolvent, and is described by the
ternary diagram shown in Fig. 2.6.

(a)

(b)

Figure 2.6: Schematic phase diagram for ternary (a) and quaternary (b) systems

The binodal curve delimitates the two-phase region, rich and lean polymer phases which
have their compositions given by the tie lines. The spinodal curve represent the line where
all possible fluctuations lead to instability. The region between binodal and spinodal
 24

corresponds to metastable compositions where phase separation by nucleation and growth

takes place.

The point where binodal and spinodal meet is referred to as the critical point (C). If the
precipitation path crosses the binodal below the critical point, nucleation of a polymer rich
phase may initiate the phase separation process. On the other hand, if the precipitation path
crosses the binodal above the critical point, nucleation of the polymer lean phase may
occur. At high polymer concentrations, phenomena such as vitrification, gelation or
crystallization can occur in the polymer solution interrupting the polymer lean phase
growth. These phenomena in a region of very high viscosity are not very well identified or
predicted. The experimental methods are not precise enough and reliable theoretical models
are not available yet (Machado et al., 1999).

If an additive is used in the polymer solution, the phase diagram can be represented by a
tetrahedron, in which binodal and spinodal are now surfaces intercepting each other at a
critical curve. Due to complexity in three dimensional representation it is very usual to
analyze quaternary system in a pseudo-ternary diagram, where the additive is considered
with the polymer as one single component. In addition to the recognition of regions where
different separation mechanisms take place, in order to predict morphologies, one must
know how the polymer solution, in contact with a non-solvent bath, changes its
composition with time and where it enters the two-phase region (Machado et al., 1999).
Reuvers (1987) identified two different mechanisms occurring in phase separation of
ternary systems by immersion-precipitation: instantaneous and delayed demixing.
Instantaneous demixing happens when phase separation begins immediately after
immersion. The precipitation path crosses the binodal and two distinct phases are formed.

In this case, the instant of immersion is the onset of demixing. Delayed demixing occurs in
the polymer solution when the precipitation path does not cross the binodal for a
measurable period of time after contact with non-solvent bath. Fig. 2.6.1 illustrate these
two mechanisms.
 25

(a) (b)

Figure 2.6.1: Precipitation paths in instantaneous (a) and delayed (b) demixing

Membranes formed by instantaneous demixing have a porous top layer and are used in
microfilltration and ultrafilltration processes. Membranes formed by delayed demixing
have a dense skin and are appropriate for use in gas separation, pervaporation and reverse
osmosis

2.6.1 Immersion Precipitation

Formation of asymmetric ultra- or hyperfiltration membranes are the result from immersion
precipitation method. Since their cross-section reveals an asymmetric structures which give
a very dense and thin skin supported by a porous layer, the membrane is called asymmetric
membrane. In immersion precipitation method, solvent was added into the coagulation bath
to obtain microporous membrane. By immersion precipitation also, the formation of
integrally-skinned asymmetric membranes has been extensively studied over the past
decade. This is because increasing importance of such membranes in reverse osmosis,
nanofiltration, gas separation, and pervaporation. A polymer solution is casted as a thin
film in immersion precipitation and after an optional partial evaporation period, the
polymer solution is immersed into a non-solvent gelation bath. A major advanced in
understanding of immersion precipitation has made by Reuvers membrane model. By this
model, it shows that two types of membranes can form where it is depends on the
conditions. Since this study is about formation of polyethermide/DGDE membrane, the
polyetherimide (PEI) has become the primary polymer to undergo immersion precipitation
 26

method. In Mulder membrane model, he used PSf-N,N-dimethylacetamide (DMAc)-2-

propanol solutions immersed in pure 2-propanol. From this model, it shows the membrane
performance that the membrane is sensitive to the initial polymer concentration of the film
(Baker, 2004).

2.6.1.1 Dry Phase

Most literature about dry phase inversion process of membrane are concerning on gas
separation membrane. Among the most important breakthroughs in membrane technology
is the development of asymmetric membranes by Loeb and Sourirajan where these
membranes are liable for existence of many high-flux commercial membrane separation
processes. Nowadays, to have an adapting membrane to replace conventional separation
processes, the membrane industry is faced with the challenge of inventing new membrane
materials as well as membrane manufacturing techniques. Through “phase inversion”
processes, many polymeric membranes are fabricated. One of the several processes in
phase inversion is dry-cast or complete evaporation process which the thermodynamic state
of polymer solution in the system can be altered to promote phase inversion. In dry-cast
process, the evaporation of non-solvent and/or solvent from an initially homogenous single-
phase polymer solution has been characterized. The evaporation takes to the formation of a
cloudy two-phase solution. Then, it is follows by solidification process where the polymer
from the polymer-rich phase precipitates to form a solid matrix which envelops the
polymer-lean phase. There is no model are available for dry-cast membrane because in dry-
cast process, it involves coupled heat and mass transport (Altinkaya et al., 2005). In
preparation of membrane (mostly gas separation membrane) through dry phase separation
process will give amorphous polymer gas separation membranes which this process is not
studied widely. This is because, dry phase inversions are different with wet or dry/wet
process by the absence of the immersion coagulation step. It is relatively simple technique
in the sense there is no exchange of non-solvent in the formation of membrane through dry
phase process (Altinkaya and Ozbas, 2004).
 27

2.6.1.2 Wet Phase

Because of thermal and chemical resistance, high mechanical strength, polysulfones has
belonged to a class of high-Tg polymer. In making asymmetric membrane, polysulfone are
widely used which the membrane will have different pore size in the active layer. Thus,
asymmetric polysulfone membrane has various applications as ultrafiltration membranes,
used as support layer for a thin film (Blanco et al., 2006). In preparing asymmetric
membrane, it can be done by different methods. Because of low cost and simple, wet phase
is frequently employed during membrane formation. There is interplay between the
polymer and its solvent and non-solvent in the phase inversion technique. Because of the
difference in diffusion exchange between the top phase and the bottom phase of the
membrane, the characteristics of the asymmetric membrane will generate (Chen et al.,
2009). According to Stropnik and Keiser (2002) they have experimental work on the
formation of polymeric asymmetric porous membrane by wet phase separation. After
immersion into the coagulation bath, a thin cast layer of the polymer solution is
transformed into a protomembrane which in general consists of non-solvent polymer
(Stropnik and Keiser, 2002). By the demixing of a stable, homogenous polymer solution in
a non-solvent precipitation bath shows how wet phase separation process been
characterized. Consists of either binary mixture of polymer and solvent or a combination of
polymers, solvents, and non-solvents, is how the casting solution generally prepared for
wet phase process. To concentrate the outermost region of the nascent membrane prior to
coagulation bath, a brief evaporation time is always used. During the wet phase separation
step, the selective skin, the transition layer supporting the skin and the open-celled
substrate are formed primarily due to the results of the detailed mass transfer happening at
the interface between the nascent membrane and the quench bath. However, membrane
produces from wet phase process have either too many defects that allow their used in
application without any further treatment or rather thick selective skin.

2.7.1.3 Dry/Wet Phase

Phase inversion process is the mostly common technique in making asymmetric membrane.
The asymmetric membrane has been invented by Loeb and Sourirajan on 1960s where it is
give a new future in membrane technology and industry. This asymmetric membrane
 28

usually used in application such as ultarfiltration, microfiltration, and nanofiltration.

Additional complexity is introduces in dry/wet phase process due to the presence of a non-
uniform concentration profile in the film at the beginning of the quenching in wet phase
inversion process (Altinkaya, 2006). Period for loss of a volatile solvent from a casting
solution containing a carefully selected amount of a less volatile non-solvent component is
required in the dry/wet phase process. Destabilization in the outermost region of the
nascent membrane is because from the selective loss of volatile solvent. The almost
instantaneous onset of milky in the outermost region is the result from the interfacial dry
phase separation. For wet process, the nascent membrane is immersed in a non-solvent
coagulant, where the bulk of the membrane structure is formed and extractions of the
remaining solvents and non-solvents take place (Pesek and Koros, 1993). Basically, in
dry/wet phase process, the dope solution is exposed to the non-solvent solution (mostly
water) for a time interval prior to immersion in the coagulation bath. By varying the casting
solution and the formation conditions, the properties of the membrane can be modified.
This can happen by changing the main factors such as the composition of the polymer
solution, the solvent evaporation temperature, and the nature and temperature of the
coagulation media. The morphology and properties of the membrane will have a great
impact with the factor of evaporation time and relative humidity (Gao et al., 2009).
Besides, solvent with strong dissolving power and high volatility is needed to be choosing.
A thin skin layer of solid polymer is formed directly at the top of the cast film during the
first step of desolvation by solvent evaporation due to the loss of solvent. The non-solvent
diffuse into the polymer solution through the thin solid layer in the solvent-non-solvent
exchange process whiles the solvent diffuse out (Khulbe et al., 2002).

2.8 Advantages of Membrane Separation Process

First of all, membrane separation process gives a new alternative and fresh technique and
technology in separation process in industry. In water desalination and purification, the
membrane processes compete directly with the more conventional water treatment
techniques. However, compared to these conventional procedures membrane processes are
very energy efficient, simple to operate and yield a high quality product. For the
desalination of brackish water either reverse osmosis or electrodialysis can be used. They
 29

are in direct competition to distillation which, however, has significantly higher energy
consumption and is affected by scaling and fouling. In this application both electrodialysis
and reverse osmosis seem to have substantial cost advantages. In seawater desalination,
reverse osmosis is the only cost effective membrane process today and is competing
directly with distillation. Depending on local conditions, including water quality, energy
cost and the required capacity of the desalination plant, either distillation or reverse
osmosis can be the more efficient process. For very large capacity units and if a power
plant can be coupled with the desalination unit, distillation is generally considered to be
more economical. For small and medium size capacity units reverse osmosis is generally
preferred. For surface water purification and waste-water treatment membrane processes,
especially micro- and ultrafiltration, are competing with flocculation, sand bed filtration,
carbon adsorption, ion-exchange and biological treatment. The membrane processes are
usually more costly but generally provide a better product water quality. Very often a
combination of conventional water treatment procedures with membrane processes results
in reliable and cost-effective treatment combined with high product water quality.

These are the list of advantages of membrane separation process include:

i. Because membrane processes can separate at the molecular scale up to a scale at which
particles can actually be seen, this implies that a very large number of separation needs
might actually be met by membrane processes.

ii. Membrane processes generally do not require a phase change to make a separation (with
the exception of pervaporation). As a result, energy requirements will be low unless a great
deal of energy needs to be expended to increase the pressure of a feed stream in order to
drive the permeating component(s) across the membrane.

iii. Membrane processes present basically a very simple flowsheet. There are no moving
parts (except for pumps or compressors), no complex control schemes, and little ancillary
equipment compared to many other processes. As such, they can offer a simple, east-to-
operate, low maintenance process option.

iv. Membranes can be produced with extremely high selectivities for the components to be
separated. In general, the values of these selectivities are much higher than typical values
for relative volatility for distillation operations.
 30

v. Because of the fact that a very large number of polymers and inorganic media can be
used as membranes, there can be a great deal of control over separation selectivities.

vi. Membrane processes are able to recover minor and valuable components from a main
stream without substantial energy costs.

vii. Membrane processes are potentially better for the environment since the membrane
approaches require the use of relatively simple and non-harmful materials.

2.9 Limitation of Membrane Separation Process

The major disadvantage of membrane processes is that until today the long-term reliability
is not completely proven. Membrane processes sometimes require excessive pre-treatment
due to their sensitivity to concentration polarization, chemical interaction with water
constituents, and fouling. Membranes are mechanically not very robust and can easily be
destroyed by a malfunction in the operating procedure. However, significant progress has
been made in recent years, especially in reverse osmosis seawater desalination, in
developing membranes which are not only have significantly better overall performance
but which also show better chemical and thermal stability and are less sensitive to
operational errors. Practical experience gained by the nations in the Middle East has
resulted in more reliable operation.
These are the list of limitation of membrane separation process include:
i. Membrane processes seldom produce two pure products, that is, one of the 2 streams is
almost always contaminated with a minor amount of a second component. In some cases, a
product can only be concentrated as a retentate because of osmotic pressure problems. In
other cases, the permeate stream would contains significant amount of materials which one
is trying to concentrate in the retentate because the membrane selectivity is not infinite.

ii. Membrane processes cannot be easily staged compared to processes such as distillation,
and most often membrane processes have only one or sometimes two or three stages. This
means that the membrane being used for a given separation must have much higher
selectivities than would be necessary for relative volatilities in distillation. Thus the trade-
off is often high selectivity/few stages for membrane processes versus low selectivity/many
stages for other processes.
 31

iii. Membranes can have chemical incompatibilities with process solutions. This is
especially the case in typical chemical industry solutions which can contain high
concentrations of various organic compounds. Against such solutions, many polymer-based
membranes (which comprise the majority of membrane materials used today), can dissolve,
or swell, or weaken to the extent that their lifetimes become unacceptably short or their
selectivities become unacceptably low.

iv. Membrane modules often cannot operate at much above room temperature. This is again
related to the fact that most membranes are polymer-based, and that a large fraction of
these polymers do not maintain their physical integrity at much above 100 °C. This
temperature limitation means that membrane processes in a number of cases cannot be
made compatible with chemical processes conditions very easily.

v. Membrane processes often do not scale up very well to accept massive stream sizes.
Membrane processes typically consist of a number of membrane modules in parallel, which
must be replicated over and over to scale to larger feed rates.

vi. Membrane processes can be saddled with major problems of fouling of the membranes
while processing some type of feed streams. This fouling, especially if it is difficult to
remove, can greatly restrict the permeation rate through the membranes and make them
essentially unsuitable for such applications.
 32

CHAPTER 3

MATERIALS AND METHODS

3.1 General Preview

Polymeric membranes have been well-established in a wide variety of industrial

applications since Loeb and Sourirajan discovered a method to prepare asymmetric
membrane structures. In 1960, Loeb and Sourirajan had successfully developed the first
integrally-skinned asymmetric membranes. An integrally-skinned asymmetric membrane
consists of a very thin and dense skin layer (0.1-1 µm) overlaying on a thick and highly
porous sub-layer (100-200 µm) (H.Strathmann, 1986; Chung et al.,1997; Wang et al.,
1991). The most commonly applied in membrane manufacturing are the dry-wet phase
inversion technique. This technique which is also called as Loeb-Sourirajan technique, was
used by Pinnau and Koros to produce an ultrathin defect-free skin layer by using dry/wet
phase inversion technique in the last few decades (Pinnau and Koros, 1990). The skin layer
achieved was of the order of 200 Å. According to Loeb-Sourirajan method, a polymer
solution is prepared by mixing a polymer, a solvent and sometimes non-solvent(s) or
additive(s). The solvent and non-solvent should exhibit a strong dissolving power with high
volatility.

In this chapter, development of the high-selective polymeric nanofiltration

membrane for ion selection is briefly described. The materials selection, dope preparation,
membrane fabrication, performance evaluation and membrane characterization are well
discussed in this section. In addition, characterization of developed NF membranes is done
by Scanning Electron Microscope (SEM) will portrays the effects of rheological and
preparation factor on the performances and membranes morphological structures. Flat sheet
 33

membranes were casted on the glass subtrates and immersed in the coagulation bath for
solvent/non-solvent exchanges to occur and resulting the membranes to solidi

3.2 Research Design

Fabrication of high-selective polymeric nanofiltration membrane can be influenced by

several factors as shown in figure 3.2. The parameters that determine the performance of
NF membrane including of polymer selection, solvent and non-solvent selection, solution
preparation method, casting parameters, flux and rejection test. The multivalent salts were
used to study the rejection rates of membrane using membrane permeation cell.

High-selective Polymeric
Nanofiltration Membrane for Ion
Selection

Materials Casting System

Selection Conditions Components
s

Figure 3.2 : Parameters controlled in fabrication of high-selective polymeric nanofiltration

membrane for ion selection
 34

Polymer Non-Solvent
Solvent
(Polyetherimide) (DiethyleneGlycol-
(N-methyl-2-
DimethylEther)
pyrrolidone)

Shear Rates Dope

Effect Formulation

Membrane
Fabrication

Pure Water
Permeation(PWP)

Multivalent Salts
Rejection

NaCl Na2SO4 MgSO4 MgCl2

Scanning Electron Fourier Transformed Infrared

Microscopy (SEM) Spectroscopy (FTIR)

Figure 3.2.1 : Schematic diagram of research framework.

 35

3.3 Materials Selection

Material selection is one of the most important parameter in membrane preparation

especially in dope formulation. System components which can either be a binary, tertiary or
quarternary can influences the membrane final structures and properties. In Loeb-
Sourirajan technique, composition of solution consists of polymer, solvent and sometimes
non-solvent(s) or additive(s).

3.3.1 Polymer Resin

Aromatic polyimides have attracted much attention as promising membrane materials

because of their excellent chemical, mechanical and thermal stabilities as well as good
separation properties. There are, however, some restrictions on selecting suitable solvents
in preparing asymmetric polyimide membrane through traditional phase inversion
technique since common polyimide materials are normally resistant to solvent dissolution.
As such, aromatic polyetherimide (PEI) appears to be of particular interest as a membrane
material in view of the fact that aromatic imide units provide the high performances
properties while the flexible ether linkages provide good processability.

Figure 3.3.1 : Diagram of monomer of polyetherimide(PEI)

 36

Table 3.3.1: Physical properties of polyetherimide (PEI)

Properties Unit Values

Density g/cm³ 1.27
Water absorption: mg 20/41
after 24h/96h immersion in water of 23°C
Melting point °F 640-800
Glass transition temperature °C 215
Thermal conductivity at 23°C W/(K·m) 0.22
Specific heat kJ.kg-1.K-1 2.0
Upper working temperature °C 170 – 200
Elongation at break % 15
Compressive strength MPa 140
Tensile strength MPa 85

3.3.2 Solvent

The selection of the proper solvent is essential in designing and engineering of a membrane
suitable for certain given nanofiltration process. Appropriate solvents selection for casting
of polyetherimide membranes involves the following conditions; the solvent must be
miscible with the coagulating agent and the solvent must provide good mechanical
properties. As such a casting procedure, it is recommended that good solvents being used
must have the power to open and solvate the polymeric chains. Lau et al. (1991) reported
that NMP shows high solvent miscibility and high ranking in solvent dissolution power
compared to the other solvent (i.e., Tetramethylurea, Dimethylacetamide,
Dimethylsulfoxide and Dimethylformamide). The solvent mixture of NMP and water is
easily miscible with water as the coagulant; fast coagulation took place from both sides of
the nascent fibre and the fast solidification across the overall membrane wall restricted the
growth of macrovoids in the polymer-lean phase (Wang et al., 2000). NMP is a widely used
industrial solvent for natural and synthesis plastics, waxes, resins and various types of
paints. It dissolves polymers, such as cellulose derivatives, polyamides, polyimides,
 37

polyesters, polystyrene, polyacrylonitrile, polyvinyl chloride, polyvinyl pyrrolidone,

polyvinyl acetate, polyurethanes, polycarbonates, polyethersulfones, polysulfones,
polyethers and many copolymers (http://www2.basf.us/diols/bcdiolsnmp.html).

Figure 3.3.2.1: Chemical structure of N-methyl-2-pyrrolidone (NMP)

Table 3.3.2.2: Physical and Chemical Properties of NMP

Physical and Chemical Properties of NMP

Property Unit value
Molecular formula C5H9ON
Molecular weight 99.1 g/mol
Density --- Liquid (20°C) 1.03 g/mol
(40°C) 1.02 g/mol
(60°C) 0.99 g/mol
Boiling Point (760 mm Hg) 204.3°C
Freezing Point (760 mm Hg) -23.6°C
Viscosity (20°C) 1.7 cp
(50°C) 1.0 cp
(100°C) 0.9 cp
Specific Heat---Liquid (50°C) 0.401 cal/g°C
Flash Point (ASTM-D 93-72) 91°C
Chemically stable and powerful polar -
solvent
Very weak base -
Solubility Parameter 11.0 MPa1/2
(http://www2.basf.us/diols/bcdiolsnmp_properties.html)

In this study, the organic chemical used in dope solution is N-methyl-2-pyrrolidone (NMP,
>99%) supplied by Merck, Darmstadt, Germany. Formula structure of N-methyl-2-
pyrrolidone is shown in Figure 3.1. Table 3.3.2.2 shows the physical and chemical
properties of NMP
 38

3.3.3 Non-Solvent

DiethyleneGlycol DimethylEther(DGDE) was used as a non-solvent in this study. DGDE

acts as a weak non-solvent and also as additive. The ratio of NMP to DGDE in the casting
solution is an important factor for reducing the pore size in the dry/wet phase separation
process. By adding a non-solvent additive, the polymer solution become unstable,
indicating that the aggregation of polymer solution become unstable, indicating that the
aggregation of polymer chains are promoted. Therefore, solvents located between polymer
chains can be easily diffused out into the coagulation bath. In addition to this aggregation
of polymer chains by adding non-solvent additive in the polymer solutions, a phenomenon
forming sharp interface between casting solution and the water should be investigated.
DGDE can create a sharp interfaces with water because of low miscibility between DGDE
and water, which is due to larger dipole moment difference of DGDE and water than that of
NMP and water. These two phenomena (sharp interface formation and the aggregation of
the polymer chains) can explain the formation of dense skin layer. In addition, a study done
by Kim et al. (2001) proved that by increasing the DGDE content, the macrovoids decrease
in both size and number. This explains that high-selectivity of membrane can be achieved
for multivalent salts rejection.

Figure 3.3.3.1: Molecular structure of DGDE

Table 3.3.3.2: Physical Properties of DGDE (MP Biomedicals/www.mpi.com)

Properties Unit Values

Boiling Point 155-165°C at 1013 hPa (Lit.)
Melting Point - 64°C (Lit.)
Density 0.943 – 0.945 g/cm3 at 20°C (Lit.)
Auto-Ignition 190°C at 1013 hPa (DIN 51794)(Lit.)
Solubility Soluble in water
Vapour density 4.6 (vs air)(Lit.)
Vapour Pressure 2.3 hPa at 20°C (Lit.)
 39

3.4 Formulation of Polymer Solution

In general, a polymer solution formulation for asymmetric membrane fabrication consists

of polymer, solvent and non-solvent. This can be called as multi component casting
formulation. Sometimes, this formulation consists of four components, where polymeric
additive was added. According to Wang (2000), the multi components dope solution
incorporates certain amounts of non-solvent so that the dope condition is bring nearer to the
binodal boundary thus facilitating the necessary rapid coagulation in the dry gap for
effective phase separation as discussed before.

Table 3.4: Composition of solution components relative to each membrane samples.

Solution component (wt %) S1 S2 S3 S4 S5

Polyetherimide (PEI) 20% 20% 20% 20% 20%

N-methyl-2-pyrrolidone (NMP) 77% 74% 71% 68% 65%
DiethyleneGlycol DimethylEther 3% 6% 9% 12% 15%
(DGDE)

3.4.1 Multi Component Dope Preparation

Apparatus for preparing the dope solution is shown in Figure 3.4.1.1. Polymer was first
dried for at least 24 hours in a vacuum oven at a temperature of about 100 ± 2°C in order to
remove all absorbed water vapours. The existence of water in the polymer solution will
influences the quality of a polymer solution. Round bottom flask was used to prepare the
dope solution. Thermometer was used to measure the temperature during the mixing
process. Processing temperature was controlled at the desired range. The function of stirrer
is to ensure that the polymer, solvent and non-solvent are well-mixed to form a
homogeneous solution.
Firstly, the solvent was poured into the round bottom flask until the temperature
increase to 50-60 °C. Thermometer measured the temperature while heater controlled the
 40

required temperature range. The mixing temperature must be controlled and maintained not
exceed the boiling point of the solvents and non-solvent (100°C) because they tend to
vaporize. However the low temperature will not encourage the mixing process. Diagram
below shows apparatus setted up for making of multi components of dope solutions

Figure 3.4.1.1: Schematic diagram of apparatus used for dope solution preparation

Figure 3.4.1.2: Apparatus set up for dope preparation

 41

3.5 Membrane Casting

After all of the newly formulation of dope solution have been prepared, membranes are
casted using using a developed casting system which consist of casting blade calibrated to
specific microscales, glass medium as membrane subtrate and sufficient amount of tap
water in a container as external coagulant. The speed of casting was fixed at 10 s and
temperature of tap water is equivalent to room temperature.

Figure 3.5: Schematic flow diagrams of membrane fabrication step by step.

 42

The schematic flow diagrams shows step by step of fabrication of flat sheet membranes.
Casting solution were prepared first. In this study, there are five different compositions
which consist of multi-components (polymer,solvent,non-solvent additive or/and polymeric
additive). Membrane was casted on a glass plate which acts as a substrate using a stainless
steel blade which was setted up to ± 100 µm thick. Tap water was used for coagulation
medium for membrane. The membranes produced were left in the tap water for about 24
hours.

3.6 Membrane Post-Treatment

For post-treatment of membrane, 100% concentration of ethanol was used. Membranes

fabricated were immersed in the ethanol solution for about 24 hours. After that, we proceed
the treatment with 100% concentration of n-hexane for 1 to 2 hours. Finally, the membrane
sheets were left to dry at room temperature for two days.

3.7 Nanofiltration Membrane Performances Evaluation

In the term of key performances evaluation of membrane, parameters that can be used to
measure the membrane efficiency and selectivity are pure water permeation(PWP) and salts
rejection. Those two parameters are the basic things studied by most researchers.

3.7.1 Nanofiltration Permeation Rig

The schematic diagram of nanofiltration permeation rig was being shown as in Figure
3.7.1. Using the well-known Sterlitech filtration cell, circular disk membranes were being
cut and mounted in a stainless steel, cylindrical membrane test cell by a porous support and
being tightened by a rubber O-ring. Effective area of the membrane being mounted under
the cell is 1.46 x 10-3 m2. The operating pressure for filtration test was being supplied by
pressurized nitrogen gas. The nitrogen gas outlet pressure was being regulated using a
pressure regulator where the equilibrium pressure was shown in the pressure gauge. A
pressure relief valve was being installed between the nitrogen gas and cell.
 43

NF
Membrane

Figure 3.7.1: Nanofiltration permeation rig (Sterlitech filtration cell)

3.7.2 Nanofiltration Permeation Test

Nanofiltration permeation test was out using dead end filtration unit at different pressure.
For each operating pressure, distilled water was used to determine the pure water
permeability (PWP) parameter/constant. PWP parameter was obtained from the slope of
volume flux versus pressure. Prior to the each nanofiltration test, membrane was subjected
to be pressurized till about 500 kPa for 1 hour. The flux was equilibrated for the passage of
the first 20 ml permeate whilst the following 20 ml permeate was collected for
concentration analysis. All the results presented are averaged data obtained through three
membranes samples with a variation of about ±10%. The permeate was collected and
weight was measured by an electronic balance for every minute. The volume flux was
calculated based on the weight and time as stated in Equation 3.7.

(Eqn 3.7)
 44

where,
J v = the permeate flux of salt aqueous solution or pure water flux (m3/m2s)
A = the effective area of membrane (m2),
t = the time (s)
V = Volume of permeate solution collected, (m3)

For pure water flux, the permeability of each membrane was determined by the
measurement of water flux as a function of applied pressure.

(Eqn 3.7.1)

where,
P m = Permeability (m3/m2s Pa),
J v = Flux (m3/m2 s)
ΔP = Pressure (Pa)

Concentration polarization is another aspect that needed taken into account for all
experimental result for this study. It was necessary to calculate the concentration
polarization of the both the polymer and the salt when analyzing flux-rejection curves. If
the membrane exhibits rejection for the solute, partial permeation will occur and non-
permeated solute accumulates in the boundary layer (surface of membrane) and hence
concentration profile develops. This phenomenon is called concentration polarization.
The equation for if this phenomenon was taken into consideration (Gilron et al., 2001):

(Eqn 3.7.2)

or

( ) (Eqn 3.7.3)
 45

( ) (Eqn 3.7.4)

[( ) ]
(Eqn 3.7.5)

The rejection characteristics of a membrane are described by observed rejection, Robs and
real rejection, Rreal. In the membrane separation processes, the concentration on the
membrane surface is always higher than in the bulk due to concentration polarization. A
concentration on the membrane surface is not directly obtained from experiment thus the
following equation was applied (Sching and Widmer, 1992):

( ) ( )
(Eqn 3.7.6)

[ ] (Eqn 3.7.7)

[ ] (Eqn 3.7.8)

where,
Cp = Salt concentration in permeate
Cb = Salt concentration in bulk
Cw =Salt concentration in wall

Mass transfer coefficient, k has been found to be a function of the stirring speed and thus
from the mass transfer correlation for a stirred cell, k can be written as a function of ω as
follows (Bowen et al., 1997):

ω (Eqn 3.7.9)

(Eqn 3.7.10)
 46

Where,

( ) (Eqn 3.7.11)

For this experiment,

r = radius of stirred cell = 5.1 cm,
υ = kinematic viscosity = 8.9 x 10-7 m2/s
Di,∞ = bulk diffusivity (m2/s)

The salt or bulk diffusivity coefficient was calculated from the ionic diffusion
coefficient as (Newman, 1991):

( ) (Eqn 3.7.12)

For single ions, the size can be presented by a Stokes radius calculated from the
Stokes-Einstein equation (Schaep et al., 2001):

(Eqn 3.7.13)

rsi = ion radius (m)

kB = Boltzmann’s constant,
T = temperature (K),
Di, ∞ = ion diffusion coefficient (m2/s),
μvis = ion dynamic viscocity (kg/ms)

3.7.3 Pure Water Permeation (PWP) Test

As PWP parameter and flux were obtained, the membrane permeation test was carried out
according to the procedure stated in Section 3.7.2. At five different operating pressures
(150 kPa, 200 kPa, 250 kPa, 300 kPa and 350 kPa) the permeation test was carried out
using 0.01 M NaCl. The flux and observed rejection were determined for each of operating
 47

condition. In order to reduce the polarization effects, the stirring speed was fixed at 400
rpm or 41.881 rad/sec. After each permeation test, NF membranes surface was rinsed with
distilled water. The pure water permeability was measured to ensure the original flux (pure
water) was recovered before next permeation test. The feed concentration of each salt
solution was also fixed at about 0.01 M. For the flux and rejection calculation purposes, the
salt concentration for all sample in feed and permeate steam were measured using a
conductivity meter (Hanna HI 8633, Italy). The concentration of NaCl was related to the
measured conductivity using the calibration curve.

3.7.4 Multivalent Salts (Na2SO4, MgSO4, NaCl and MgCl2) Permeation Test

The permeation test is conducted similar to the procedure as listed in Section 3.7.2. in order
to study the effect of rejection under different volume fluxes at pressure of 500 kPa to
obtain its respective fluxes. The salt permeation test was carried out using 0.01 M of
(Na2SO4, MgSO4, NaCl and MgCl2) solutions. After each permeation test, NF membranes
surface was rinsed with distilled water. The pure water permeability was measured to
ensure the original flux (pure water) was recovered before the next permeation test will be
run. The feed, permeate and retentate concentration for each were measured by using
conductivity meter (Hanna HI 8633, Italy). The flux and observed rejection were
determined for each operating condition. To reduce concentration polarization, the stirring
speed was fixed at 400 rpm or 41.881 rad/sec.

3.8 Structural Characteristics of Asymmetric NF Membrane

In order to characterize the membrane and determine the morphologies of membrane, the
techniques known as Scanning Electron Microscopy (SEM) and Fourier Transform Infrared
Spectroscopy (FTIR) were being used in this study. SEM characterize the membrane in
term of membrane surface structures in which formation of pores and finger-like structures
can be observed. On the other hand, FTIR determine the functional groups exist in the
membranes and identify the molecular orientations occur.
 48

3.8.1 Scanning Electron Microscopy

Scanning Electron Microscope (SEM) is a very useful instrument for membrane

morphological determination. This method is used for the determination of membrane
structure, membrane thickness and membrane surface. The principle of the Scanning
Electron Microscope is illustrated in Figure 3.8.1.1. A narrow beam of electrons with
kinetic energies in the order 1-25 kV hits the membrane sample. The incident electrons are
called primary (high-energy) electrons, and those reflected are called secondary electrons.
Secondary electrons (low energy) are not reflected but liberated from atoms in the surface,
determine the imaging (what is seen on the screen or on the micrograph). When a
membrane is placed in the electron beam, the sample can be burned or damaged, depending
on the type of polymer and accelerating voltage employed. This can be avoided by coating
the membranes with a conducting layer (gold palladium) to prevent charging up the
surface. Scanning electron microscopy allows a clear view of the overall structure of the
membranes; the top surface, the cross section and the bottom surface can all be observed
nicely. In addition, the porosity and the pore size distribution can be estimated from the
micrographs obtained.

Figure 3.8.1.1 : The principle of scanning electron microscopy

 49

Figure 3.8.1.2: JEOL JSM 6360LA Scanning Electron Microscopy (SEM)

3.8.2 Attenuated Total Reflection-Fourier Transform Infrared Spectroscopy(ATR-FTIR)

Molecular analysis PEI/NMP with addition of 3 wt%, 6 wt%, 9 wt%, 12 wt%, 15 wt% of
DGDE in run through VARIAN 3100 FTIR Excalibur Series in Figure 3.8.1.2. Attenuated
total reflection fourier transform infrared spectroscopy (ATR-FTIR) allows for infrared
(IR) analysis of surfaces. The IR spectrum can provide determination of vibrational
frequencies and transition intensities of most molecules (with the exception of diatomics
such as N2 and O2), including characteristic functional group frequencies. Knowledge of
vibrational frequencies of functional groups (or reference spectra) allows for chemical
identification of at least a class of compounds (e.g., aromatic amides), which may be less
apparent in XPS spectra. The IR radiation typically penetrates 1 μm into the surface. This is
a disadvantage because this is deeper than the active layer thickness of most composite
membranes, so the top layer is not necessarily isolated (Boussu et al., 2007). The
penetration depth can be decreased by careful selection of crystal and incident angle (for
Ge ATR element and 3000 cm-1 wavenumber it can be estimated as 200 nm (Müller and
Mehnert, 1997) at 45° incident angle). Alternatively, it is possible to perform ATR-FTIR
 50

measurements of both sides of an asymmetric membrane and then to essentially subtract

out the support layer spectral regions.

Figure 3.8.2 : VARIAN 3100 FTIR Excalibur Series

 51

CHAPTER 4
RESULT AND DISCUSSIONS

In this study, nanofiltration membranes were prepared through dry/wet phase inversion
process. This process is the most common process in making of asymmetric membrane. Six
different content of diethyleneglycol dimethylether (DGDE) with the same composition of
polymer, polyetherimide(PEI) were applied in the dope formulations. With different
contents of DGDE , the result of membrane structures and morphologies will be different.
Therefore, based on that outcomes, we can choose the best membrane structures for basic
or industrial application. In order to achieve membrane of high-selectivity, membrane
structures with thin skinned top-layer and highly-dense substructure is desired.

4.1 Membrane Permeation Performances

4.1.1 Pure water permeability (PWP)

Table 4.1: Pure water permeability (PWP) data of asymmetric polyetherimide(PEI)

membranes at different DGDE content (wt %)

Pure water permeability at different DGDE concentrations,

Pressure PWP ( x 10-6 m3/m2.s)
(kPa)
3.0 6.0 9.0 12.0 15.0

150 3.53 2.86 2.38 2.04 1.95

200 4.12 3.68 3.25 2.92 2.20

250 7.05 6.14 4.74 5.02 3.32

300 8.83 8.06 6.33 6.22 4.52

350 10.25 9.18 8.10 7.93 6.07

 52

The graph below shown how the differences in amount of DGDE affects the pure water
permeability (PWP) performance.

PWP vs DGDE content (wt%)

12

10
PWP (m3 /m2 . s)

0
3 6 9 12 15
DGDE content (wt%)

150 kPa 200 kPa 250 kPa 300 kPa 350 kPa

Figure 4.1.1.1: Graph Pure Water Permeability (PWP) vs DGDE content (wt%)

PWP vs Pressure (kPa)

12

10
PWP (Jv)

0
150 200 250 300 350
Pressure (kPa)

150 kPa 200 kPa 250 kPa 300 kPa 350 kPa

Figure 4.1.1.2: Graph Pure Water Permeability (PWP) vs Pressure

 53

In this study, pure water permeability is one of key paramaters that used to testify the
membrane performances. Therefore, based on the result achieved, PWP are decreasing
although the pressure is kept increased from 150 kPa to 500 kPa. Due to this occurrence,
this membrane is not a really good criteria for nanofiltration.

4.1.2 NaCl volume fluxes

Table 4.1.2: NaCl volume fluxes of membranes at different DGDE content (wt %).

NaCl volume flux at different DGDE concentrations,

Pressure Jv x 10-6 (m3/m2.s)
(kPa)
3.0 6.0 9.0 12.0 15.0

300 2.08 2.03 1.80 1.78 1.73

350 2.32 2.18 1.99 1.95 1.82

400 3.04 2.90 2.36 2.28 2.00

450 3.63 3.76 2.89 2.41 2.28

500 3.93 4.07 3.26 2.64 2.45

NaCl fluxes vs DGDE content

4.5

3.5
NaCl fluxes (Jv)

3
350 kPa
2.5
400 kPa
2
450 kPa
1.5 500 kPa
1 300 kPa
0.5

0
3 6 9 12 15
DGDE content (wt%)

Figure 4.1.2 : NaCl fluxes vs DGDE content (wt%)

 54

NaCl fluxes also showed decrement in value as well as pure water permeability. Therefore,
desalting process is not favourable for these membranes.

4.1.3 Real Rejection of NaCl solution

Table 4.1.3: Real rejection, Rreal of NaCl solution (0.01M) at different DGDE content (wt%).

Rreal of 0.01M NaCl at different DGDE concentration,

Pressure Rreal (%)
(kPa)
3.0 6.0 9.0 12.0 15.0

300 6.33 9.42 14.34 15.50 16.16

350 11.60 13.96 23.96 24.57 25.12

400 13.71 18.82 27.58 28.90 34.23

450 18.70 20.44 32.51 41.37 42.98

500 23.26 25.78 43.09 45.79 46.89

Rreal of 0.01M NaCl vs DGDE content

Rreal (%)
50
300kPa
45
40
350kPa
35 400kPa
30 450kPa
25 500kPa
20
15
10
5
0
3 6 9 12 15

DGDE content(wt%)

Figure 4.1.3.1: Graph of Rreal of 0.01M NaCl vs DGDE content

 55

The real rejection for NaCl salt definitely increase but still not very high. However, we can
see that the higher the DGDE content, the higher the salt rejection. We expect that for
achieving a very high NaCl rejection, addition of extra amount of DGDE might be
required.

Pure Water Permeability(PWP) vs Rreal NaCl Rejection at

350 kPa
12 50
45
10

Rreal Rejection (%)

40
35
8
30
PWP

6 25
20
PWP
4 Rreal rejection(%)
15
10
2
5
0 0
3 6 9 12 15
DGDE content (%)

Figure 4.1.3.2: Graph Intersection of PWP and Rreal NaCl Rejection at 500 kPa

Based on the graph plotted, it shows that by increasing DGDE content to certain amount
can actually give significant increase in NaCl rejection. From the intersection, we can
conclude that 8 wt% of DGDE is the optimum content for membrane preparation. At that
point, NaCl rejection performance shows sudden increase but for PWP it tends to decrease.
However, based on this we can determine the optimum DGDE content for a membrane to
have a good starting performance.

4.2 Multivalent salts rejection

Multivalent salts permeability test were conducted as to determine the membrane
performance. Four different salts were used as shown in the figure below.
 56

Table 4.2: Multivalent salts rejection at 500 kPa relatively to DGDE content (%)

Multivalent Salts Rejection at 500 kPa

DGDE NaCl MgCl2 Na2SO4 MgSO4

content (%)

3 23.26 56.22 73.45 78.38

6 25.78 64.14 76.51 79.86

9 43.11 79.18 81.15 85.11

12 45.79 80.53 82.49 87.93

15 46.89 82.06 83.36 89.71

With increasing the amount of DGDE in the polymer solution, it seems that the
performance of membrane in term of multivalent salts rejection also increase. From the
table above, we can see that the salts rejection is much more better for multivalent salts
such as MgCl2, Na2SO4 and MgSO4 compared to NaCl. This maybe happen due to the size
of ions.
 57

Percentage of Percentage of rejection vs. DGDE content

rejection (%)
100

89.71
90 87.93
85.11 83.36
81.15 82.49
82.06
79.86 80.53
80 78.38 79.18
76.51
73.45

70
64.14

60
56.22

50 46.89
45.79
43.11

40

30
25.78
23.26

20

10

0
3 6 9 12 15
DGDE content (%)
NaCl MgCl2 Na2SO4 MgSO4

Figure 4.2: Bar chart of multivalent salts rejection vs. DGDE content

From the bar chart above, this separation membrane show a good selectivity towards
multivalent salt especially MgSO4. It achieves a significant real rejection when the
membrane having 9 wt% of DGDE composition and above. Therefore, the optimum
content of DGDE is 9 wt%. As we can observed from the bar chart, percentage of NaCl
rejection eventually fluctuate at 9 wt% until 15 wt% while for other multivalent salts the
percentage of rejection increase gradually. Although NaCl salt rejection seems slightly
lower than other salts but the value still reliable because the percentage of NaCl rejection
by most commercial membrane is approximately 55 %.
 58

4.3 Fourier Transform Infrared Spectroscopy

Fourier-Transform Infrared (FTIR) Spectroscope is the best instrument to study distribution

of functional groups and molecular orientations in the membrane. FTIR will show the
bands that indicates all functional groups exist in the membrane. Besides, the molecular
orientations of molecules can also be observed based on the peaks formation of every
membrane

(a) PEI/NMP/DGDE-3%

(b) PEI/NMP/ DGDE-6%

 59

(c) PEI/NMP/DGDE-9%

(d) PEI/NMP/DGDE-12%
 60

(e) PEI/NMP/DGDE-15%

Figure 4.3: FTIR absorption bands of membrane based PEI/NMP/DGDE; (a)

PEI/NMP/DGDE-3%; (b) PEI/NMP/ DGDE-6%; (c) PEI/NMP/DGDE-9%; (d)
PEI/NMP/DGDE-12%; (e) PEI/NMP/DGDE-15%

Table 4.3 : Infrared absorption of a polyetherimide (PEI) membrane in the 4000-400cm-1

region, obtained from reflection spectra.

Membrane Sample Assignment Wavenumber (cm-1)

PEI/NMP/DGDE-3% 3° amide C-N stretch (br,m) 3436.432 (br)

Alkyl C-H stretch (w) 2969.525 (vw)
3° amide (C=O) stretch (s) 1721.946 (vs)
Arene (C=C) (m) 1600.948 & 1480.439 (m)
C-H (sp3) bends 1358.541 (m)
C-O-C (aliphatic) 1236.287 (m)
C-O-C (aromatic) 1102.086(s) & 846.406(w)

PEI/NMP/DGDE-6% 3°amide C-N stretch 3421.604 (br)

Alkyl C-H stretch 2970.364 (vw)
3°amide C=O stretch 1721.675 (vs)
Arene (C=C ) 1601.259 & 1480.676 (m)
C-H(sp3) bend 1358.125 (m)
C-O-C (aliphatic) 1235.684 (m)
C-O-C (aromatic) 1101.875(s) & 846.689(w)

PEI/NMP/DGDE-9% 3°amide C-N stretch 3447.338 (br)

Alkyl C-H stretch 2970.503 (vw)
3°amide C=O stretch 1721.404 (vs)
 61

Arene (C=C ) 1601.615 & 1480.573 (m)

C-H(sp3) bends 1357.790 (m)
C-O-C (aliphatic) 1235.802 (m)
C-O-C (aromatic) 1101.674(s) & 846.422(w)

PEI/NMP/DGDE-12% 3°amide C-N stretch 3430.647 (br)

Alkyl C-H stretch 2970.669 (vw)
3°amide C=O stretch 1721.307 (vs)
Arene (C=C) 1601.321 & 1480.393 (m)
C-H(sp3) 1357.745 (m)
C-O-C (aliphatic) 1235.689 (m)
C-O-C (aromatic) 1101.598(s) & 845.893(w)

PEI/NMP/DGDE-15% 3°amide C-N stretch 3436.780 (br)

Alkyl C-H stretch 2970.139 (vw)
3°amide C=O stretch 1721.686 (vs)
Arene (C=C ) 1601.337 & 1480.653 (m)
C-H(sp3) bend 1358.097 (m)
C-O-C (aliphatic) 1236.077 (m)
C-O-C (aromatic) 1102.159 (s) & 847.297(w)

Qualitative band shapes and intensities are given as br broad, sh shoulder, vs very strong, s
strong, m medium, and vw very weak.

1.05

0.95

0.9
Transmittance

0.85

0.8

0.75
3%.spc
6%.spc
9%.spc
0.7 12%.spc
15%.spc

0.65

0.6
3500 2500 1500 500
Wavenumbers [1/cm]
 62

A very strong molecular interactions occur in the membrane having 9%, 12% and 15% of
DGDE content. Infrared spectrum having many strong bands in the 'fingerprint' region of
the spectrum below 1800 cm-1 and from the peak formed, we can observed the strongest
values of transmittance depict by molecular interactions from membrane having 9 wt% of
DGDE content. Because of better alignment and interactions between molecular
orientations, this membrane exhibits a better performances compared to the membranes
having only 3 wt% and 6 wt% of DGDE and we expect the optimum permeability and
selectivity starts here because membrane performances suddenly increase especially for
multivalent salt rejection which changes drastically from 25.78% to 43.11% for NaCl real
rejection.

4.4 Membrane Morphological Test using SEM

The prepared five membrane samples were characterized using Scanning Electron
Microscope. SEM analysis showed the images of cross-section structure of asymmetric
nanofiltration membrane PEI/DGDE at different composition of DGDE. It was found that
all the membranes exhibits typical morphological structure of asymmetric membrane which
consists of dense top layer and a porous sub-layer containing macrovoids, pores, and micro
voids. This sub-layer formation is controlled by variables of polymer dope solution such as
composition, coagulant temperature, and additives (Rahimpour and Madaeni, 2007).

Top skin layer

Finger-like structures
Bottom sub-layer

Macropores

(a) PEI/NMP/DGDE-3%
 63

(b) PEI/NMP/DGDE-6% (c) PEI/NMP/DGDE-9%

(d) PEI/NMP/DGDE-12% (d) PEI/NMP/DGDE-15%

Figure 4.4: SEM micrograph of cross sectional view of PEI/NMP/DGDE membrane;

(a) PEI/DGDE-3%; (b) PEI/DGDE-6%; (c) PEI/DGDE-9%; (d) PEI/DGDE-12%;
(e) PEI/DGDE-15%

Figure 4.4 shows the images of cross-section structure of asymmetric nanofiltration

membrane PEI/DGDE at different composition of DGDE. It was found that all the
membranes exhibits typical morphological structure of asymmetric membrane which
consists of dense top layer and a porous sub-layer containing macrovoids, pores, and micro
voids. This sub-layer formation is controlled by variables of polymer dope solution such as
composition, coagulant temperature, and additives (Rahimpour and Madaeni, 2007).
 64

SEM micrograph showed the final membrane structures from the first sample; (a)
PEI/NMP/DGDE-3% revealed that membrane top skin layer is very thin while membrane
sub-layer has finger-like structures and macropores. There were smaller and larger
macropores formed and eventually smaller macropores positioned very close to the short
finger-like structures. No spongy structure was observed. Based on the membrane
permeability, this membrane shows the highest water flux. This fact was supported by the
membrane porosity where the formation of larger and smaller macropores in the sub-layer
make the membrane become more porous compared to others due to lesser resistance for
water to pass through. However, multivalent salts rejection for this membrane portrayed the
lowest percentages as less resistance occur and this was a result of very thin skin layer and
higher-porosity of membrane.

From the second membrane sample (b) which contained 6 wt% of DGDE revealed that
membrane top skin layer increased in thickness while in the membrane sub-layer
macrovoids disappeared and the dominant structures formed were long finger-like
structure. Underneath the finger-like structures, spongy structure slightly formed.
Membrane performance in term of pure water flux slightly decrease from previous
membrane as the membrane porosity was decreased due to significant changes to
membrane structures where the formation of spongy structure and slightly increase in upper
skin thickness lead to higher resistance and difficulties for water permeation. However,
multivalent salts rejection for this membrane portrayed the better percentages as membrane
became dense and more salts retention occurred.

While the final membrane structures from the third sample (c) which contained 9 wt% of
DGDE revealed that membrane top skin layer slightly decrease in thickness. While
membrane sub-layer showed the long finger-like structures decreased because of the
formation short finger-like structures altogether and smaller macropores suddenly
appeared. These structures provided a very good resistance for water as well as salts
retention due to denser membrane sub-layer. Hence, pure water flux experienced
significant decrement as the membrane porosity was greatly decreased. However,
multivalent salts rejection for this membrane was much better because at this stage
membrane started to achieve rejection rate up to 80% and above.
 65

The final membrane structures from the fourth sample (d) which contained 12 wt% of
DGDE revealed that there were no changes in membrane thickness. Further increase in
content of DGDE, the finger-like structures shorten and likely to form macrovoids however
the area of spongy structures became larger. Membrane performance in term of water flux
tends to drop drastically compared to previous membrane especially when the pressure at
250 kPa was applied. As the membrane became very dense, therefore pure water flux was
affected as well. Multivalent salts rejection performances also kept increased.

The final membrane structures from the fifth sample (e) which contained 15 wt% of DGDE
revealed that membrane top skin layer was very thick while membrane sub-layer became
spongier with the presence of large macrovoids. Irregular short and long finger-like
structures with different length was also formed. Membrane performance in term of water
flux constantly decrease from previous membrane but multivalent salts rejection for this
membrane portrayed the best percentages among other membranes due to larger area of
denser region in the membrane structures for salts retention.
 66

CHAPTER 5
CONCLUSIONS AND RECOMMENDATIONS

5.1 Conclusions

A newly dope formulations for asymmetric nanofiltration membrane have been

successfully developed. The influences of different contents of DGDE were being studied
on their morphological structures, permeability and rejection performance. In this study,
various multivalent salts were used to test for the membrane selectivity. In order to know
the optimum content of DGDE that should be used in polymer solution, we start with 3
wt% of DGDE and increased by 3 wt% until 5 different formulations are produced. Based
on the graph plotted between pure water permeability (PWP) and NaCl flux at five different
formulations, intersection occurs when the DGDE content is about 8 wt%. From the graph
in Figure 4.1.3.1, it clearly explained that for a membrane to reach an optimum
performances the content of DGDE should be started at that point. However, pure water
permeability drastically decreases at that point and we got an intermediate water flux
values. The main goal of this study which is to prepare high-selective polymeric
nanofiltration membrane for ion selection can be achieved because percentage of
multivalent salt rejection is high. The best membrane fabricated was from 15 wt% DGDE
content because it showed the highest percentage of multivalent salt rejection.
 67

5.1 Recommendations

Based on the membrane performances and morphologies, several recommendations can be

implemented for further development and improvement regarding this study.

(i) For water flux enhancement, membrane should be either treated with chemicals or
coated with thin film composite (TFC) and also can be prepared using dual-coagulation
bath during phase inversion process.

(ii) Although the percentage of salts rejection did not achieved up to 99% but by increasing
or adding more DGDE will actually improves the membrane selectivity as denser
membrane will be produced when amount of DGDE increases.

(iii) Instead of using ATR-FTIR, we can also use CFM, XPS and AFM to characterize the
membrane.
 68

REFERENCES

Altinkaya, S. A., Ozbas, B. 2004. Modeling of asymmetric membrane formation by dry-

casting method. J.Membr. Sci. 230. 71-89.

Altinkaya, S. C. 2006. Modeling of asymmetric membrane formation by a combination of

dry/wet phase inversion process. Desalination 199. 459-460.

Altinkaya, S. C., Ozbas, B., Yenal, H. 2005. Membrane formation by dry-cast process
model validation through morphological studies. J. Membr. Sci. 249. 163-172.

Aroon, M. A., Ismail, A. F., Rahmati, M. M. M., Matsuura, T. 2010. Morphology and
permeation properties of polusulfone membrane for gas separation: Effect of non-
solvent addities and co-solvent. Sep. Purif. Technol. 72. 194-202.

Baker, R. W. 2004. Membrane Technology and Applications. Second Edition, John Wiley
& Sons, England.

Bechhold H., Durchlassigkeit von Ultrafilter, 1908 Z.Phys. Chem. 64. 328.

Blanco, J. F., Sublet, J., Nguyen, Q. T., Schaetzel, P. 2006. Formation and morphology
studies of different polysulfone-based membranes made by wet phase inversion
process. J. Membr. Sci. 283: 27-37.

Bowen, W. R., Doneva, T. A., Yin, H. B. 2001. Polysulfone-sulfonated poly(ether ether)

ketone blend membranes: Systematic synthesis and characterization. J. Membr. Sci.
181. 253-263.

C. Barth, M.C. Goncalves, A.T.N. Pires, J. Roeder and B.A. Wolf, 2002. J. Membr.
Sci.169. 287–299.

Chen, S. H., Liou, R. M., Lin, Y. Y., Lai, C. L., Lai, J. Y. 2009. Preparation and
characterizations of sulfonated polysulfone membranes by wet phase inversion method.
European Polymer Journal 45. 1293-1301.

Chen, S. H., Liou, R. M., Lin, Y. Y., Lai, C. L., Lai, J. Y. 2009. Preparation and
characterization of sulfonated polysulfone membranes by wet phase inversion method.
European Polymer Journal 45: 1293-1301.

D. Wang, K. Li, W.K. Teo, 1998. Preparation and characterization of poly(ether imide)
asymmetric hollow fiber membranes for gas separation, J. Membr. Sci. 138. 193.

D. Wang, K. Li, W.K. Teo, D. Bo, Z. Kun, 1996. Polyethersulfone hollow fiber gas
separation membranes prepared from NMp/alcohol solvent systems, J. Membr. Sci.
115. 85–108.
 69

D. Wang, K. Li, W.K. Teo, D. Bo, Z. Kun, 1996. Polyethersulfone hollow fiber gas
separation membranes prepared from NMP/ alcohol solvent systems, J.Membr. Sci.
115
Donnan F.G., 1911. Theorie der Membrangleichgewichte und Membranpotentiale bei
Vorhandensein von nicht dialysierenden Elektrolyten, Z. fur Elektrochemie und
angewandte physikalische Chemie 17. 572.

F.-C. Lin, D.-M. Wang, C.-L. Lai, J.-Y. Lai, 1997. Effect of surfactants on the structure of
PMMA membranes, J. Membr.Sci. 123. 281–291.

F.G. Paulsen, S.S. Shojaie, W.B. Krantz, 1994. Effect of evaporation step on macrovoid
formation in wet-cast polymeric membranes, J. Membr. Sci. 91. 265–282.

Fick A., Uber Diffusion, Poggendorff’s Annalen der Physik und Chemie 94 (1855) 59

Gao, L., Tang, B., Wu, P. 2009. An experimental investigation of evaporation time and the
relative humidity on a novel positively charged ultrafiltration membrane via dry-wet
phase inversion. J. Membr. Sci. 326. 168-177.

Gao, L., Tang, B., Wu, P. 2009. An experimental investigation of evaporation time and the
relative humidity on a novel positively charged ultrafiltration membrane via dry-wet
phase inversion. J. Membr. Sci. 326. 168-177.

Graham T., 1866. On the Absorption and Dialytic Separation of Gases by Colloid Septa,
Philos. Magazine and J. of Sci. 32. 1059.

H. Hachisuka, T. Ohara, K. Ikeda, 1996 New type asymmetric membranes having almost
defect free hyper-thin skin layer and sponge-like porous matrix, J. Membr. Sci. 116.

H. Kawakami, M. Mikawa, S. Nagaoka, Gas permeability and selectivity through

asymmetric polyimide membranes, J. Appl. Polym. Sci. 62. 965–971.

H. Yanagishita, T. Nakane, H. Yoshitome, 1994 Selection criteria for solvent and gelation
medium in the phase inversion process, J. Membr. Sci. 89. 215–221.

H.Strathmann, 1986. Synthetic membranes and their preparation, in: P.M Bungay, H.K
Lonsdale, M.N. De Pinho(Eds.), Synthetic Membranes: Science, Engineering and
Applications, D riedel Publishers, Dordretch, pp. 1-38.

I. Cabasso, E. Klein, J.K. Smith, 1976. Polysulfone hollow fibers. I. Spinning and
Properties, J. Appl. Polym. Sci. 20. 2377.

I. Pinnau, J. Wind, 1991. Process for increasing the selectivity of asymmetric membranes,
US Patent 5,007,944.

I. Pinnau, W.J. Koros, 1991. Structures and gas separation properties of asymmetric
polysulfone membranes made by dry, wet and dry/wet phase inversion, J. Appl. Polym.
Sci. 43. 1491–1502.
 70

I. Pinnau, W.J. Koros, 1993. A qualitative skin layer formation mechanism for membranes
made by dry/set phase inversion, J. Polym. Sci., Polym. Phys. 31. 410–427.

I.C. Kim, H.G. Yoon, K.-H. Lee, 2002. Formation of integrally skinned asymmetric
polyetherimide nanofiltration membranes by phase inversion process, J. Appl. Polym.
Sci. 84. 1300.

I.C. Kim, H.G. Yoon, K.-H. Lee, 2002. Preparation of asymmetric polyacrylonitrile
membrane with small pore size by phase inversion and post-treatment process, J.
Membr. Sci. 199. 75.

I.-C. Kim, K.-H. Lee, T.-M. Tak, 2001. Preparation and characterization of integrally
skinned uncharged polyetherimide asymmetric nanofiltration membrane, J. Membr.
Sci. 183. 235.

I.C. Kim, T.M. Tak, 2000. Synthesis of soluble anion-exchange copolyimides and
nanofiltration membrane performances, Macromolecules 33. 2391.

I.F Wang, B.S. Minhas, 1991. Asymmetric polyimide membranes, US Patent 5,067,970.

J.H. Kim, K.-H. Lee, 1998. Effect of PEG additive on membrane formation by phase
inversion, J. Membr. Sci. 138. 153.

J.J. Shieh, T.S. Chung, R. Wang, M.P. Srinivasan, D.R. Paul, 2001. Gas separation
performance of poly(4-vinylpyridine)/polyetherimide composite hollow fibers, J.
Membr. Sci. 182. 111.

J.Reuvers, J.W.A van den Berg, C.A. Smolders, 1987. Formation of membranes by means
immersion precipitation. Part I. A model to describe mass transfer during immersion
precipitation. J.Membr.Sci. 85-100.

K. Kneifel, K.V. Peinemann 1992, Preparation of hollow fiber membranes from polyether-
imide for gas separation, J. Membr. Sci. 65. 295.

Kang, Y. S., Kim, H. J., Kim., U. Y. 1991. Asymmetric membrane formation via
immersion precipitation method. I. Kinetic effect. J. Membr. Sci. 60: 219-232.

Khulbe, K. C., Feng, C. Y., Matsuura, T. 2002. Membrane Characterization.Water and

Wastewater Treatment Technologies.

Loeb S., Sourirajan S., Seawater Demineralization by Means of Semipermeable Membrane,

in Advances in Chemistry, ACS Series No.38

M. Leeman, G. Eigenberger, H. Strathmann, 1996. Vapor permeation for the recovery of

organic solvents from waste air streams: separation capacities and process
optimization, J. Membr. Sci.113. 313.
 71

M. Mulder, 1997. “Basic Principles of Membrane Technology”, Kluwer Academic

Publishers.

M. Mulder, 1996. Basic Principles of Membrane Technology, Kluwer Academic

Publishers, London.

Ma, Y., Shi, F., Ma, J., Wu, M., Zhang, J., Gao, C. 2011. Effect of PEG additive on the
morphology and performance of polysulfone ultrafiltration membranes. Desalination
272: 51-58.

McBain J. W., Kistler S.S., 1931. Ultrafiltration as a Test for Colloidal Constituents in
Aqueous and Nonaqueous Systems, J.Phys. Chem. 35. 130.

Nollet J.A., 1752. Recherches sur les Causes du Bouillonnement des Liquides, Histoire de
I’Academie Royale des Sciences, Paris Annee MDCCXLVIII, 57.

P. Radovanovic, S.W. Thiel, S.T. Hwang, 1992. Formation of asymmetric polysulfone

membranes by immersion precipitation. Part II. The effects of casting solution and
gelation bath compositions on membrane structure and skin formation, J. Membr. Sci.
65. 231–246.

P.S.T Machado, A.C. Habert, 1999. C.P. Borges, Membrane fabrication mechanism based
on precipitation kinetics and membrane morphology: flat and hollow fibre polysulfone
membranes, J.Membr.Sci. 73. 171-183

Pesek, S. C., Koros, W. J. 1993. Aqueous quenched asymmetric polysulfone membrane

prepared by dry/wet phase separation. J.Membr. Sci. 81: 71-88.

Pfeffer W., 1877. Osmotische Untersuchungen, Leipzig.

Prausnitz, P.H., Reitstotter, J., 1931. Elektrophorese, Elektroosmose, Elektrodialyse,

Steinkopff Verlag Dresden.

P.S.T. Machado, A.C. Habert, C.P. Borges, 1999. Membrane formation mechanism based
on precipitation kinetics and membrane morphology: flat and hollow fibre polysulfone
membranes, J. Membr. Sci. 155. 171-183

R.E. Kesting, 1985. Synthetic Polymeric Membranes, Wiley, New York.

R.J. Cranford, H. Darmstadt, J. Yang, C. Roy, 1999.

Poly(etherimide)/polyvinylpyrrolidone vapor permeation membranes, physical and
chemical characterization, J. Membr. Sci. 155. 231.

Rahimpour, A., Madaeni, S. S. 2007. Polyethersulfone (PES)/cellulose acetate phthatlate

(CAP) blend ultrafiltration membranes: Preparation, morphology, performances and
antifouling properties. Journal of Membrane Science 305: 299-312.
 72

Reid C.E., Brenton E.J, 1959. Water and Ion Flow Across Cellulose Membranes, J.Appl.
Polymer Sci.1. 179

A.J. Reuvers, Membrane Formation, 1987. Ph.D. Thesis, University of Twente, Enschede,
The Netherlands.

S. Loeb, S. Sourirajan, 1963. Sea water demineralization by means of an osmotic

membrane, Adv. Chem. Ser., ACS 38 117.

S.P. Nunes, K.V. Peinemann, K. Ohlrogge, A. Alpers, M. Keller, A.T.N. Pires, 1999.
Membranes of poly(ether imide) and nanodispersed silica, J. Membr.Sci.157. 219–226.

Stropnik, Č., Kaiser, V. 2002. Polymeric membranes preparation by wet phase separation:
Mechanism and elementary processes. Desalination 145: 1-10.

Traube M., 1867. Physiologie und wissenschaftliche medizin, Arch.An.Physiol. 87.

Trussell Technologies. "Desalination Technologies". Pasadena, 2013. Trussell

Technologies. 6/4/2013. http://www.trusselltech.com/tech_desalination.asp. (11.11.2013)

Van’t Hoff J.H., 1887. Die Rolle des osmotischen Druckes in der Analogie zwischen
Losungen und Gasen, Z. phys. Chem. 1. 481

W. Albrecht, Th. Weigel, M. Schossig-Tiedemann, K. Kneifel, K.-V. Peinemann, D. Paul,

2001. Formation of hollow fiber membranes from poly (ether imide) at wet phase
inversion using binary mixtures of solvents for the preparation of the dope, J. Membr.
Sci. 192. 217.

W.R. Browall, 1979. Ultrathin poly(ether imide) membrane and gas separation process, US
Patent 4,156,597.

X. Lu, X. Bian and L. Shi, 2002. Preparation and characterization of NF composite

membrane, J. Membr.Sci., 210. 3-11.

Z.-L. Xu, T.-S. Chung, Y. Huang, 1999. Effect of polyvinylpyrrolidone molecular weights
on morphology, oil/water separation, mechanical and thermal properties of poly
(etherimide)/polyvinylpyrrolidone hollow fiber membranes, J. Appl.Polym. Sci. 74.
2220.

Zsigmondy R., Bachmann W., 1918. Uber neue Filter, Z. Anorg. Allg. Chem. 103: 119.