Membrane Technology
Membrane Technology
A Practical Guide to Membrane Technology
and Applications in Food and Bioprocessing
Edited by
Z.F. Cui and H.S. Muralidhara
AMSTERDAM • BOSTON • HEIDELBERG • LONDON • NEW YORK • OXFORD
PARIS • SAN DIEGO • SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO
Butterworth-Heinemann is an imprint of Elsevier
Butterworth-Heinemann is an imprint of Elsevier
The Boulevard, Langford Lane, Oxford OX5 1GB, UK
30 Corporate Drive, Suite 400, Burlington, MA 01803, USA
First edition 2010
Copyright © 2010 Elsevier Ltd. All rights reserved
No part of this publication may be reproduced or transmitted in any form or by any means,
electronic or mechanical, including photocopying, recording, or any information storage and
retrieval system, without permission in writing from the publisher. Details on how to seek
permission, further information about the publisher’s permissions policies and our arrangements
with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency,
can be found at our website: www.elsevier.com/permissions.
The
book
and the individual contributions contained in it are protected under copyright by
the Publisher (other than as may be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and
experience broaden our understanding, changes in research methods, professional practices,
or medical treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in
evaluating and using any information, methods, compounds, or experiments described herein.

the last two decades. For example, one of the major applications in the food industry
is the use of reverse osmosis membranes to extend the evaporation capacity and reduce
the overall energy costs, thus lowering the carbon footprint for the overall process.
The idea for this book, which focuses on the practical applications of membranes
in food and bioprocessing, came out of a North American Membrane Society (NAMS)
meeting in Chicago, Illinois, USA, in 2006, when both editors held a workshop on
“Fundamentals and Applications of Membrane Technology in Food/Bio-processing”.
With our combined experience in food technology and bioprocessing, we felt that a
book dedicated to the practical aspects and challenges of utilizing membranes effec-
tively in an industrial setting would be an extremely useful tool for any one in mem-
brane processing and practice. The idea gained even more momentum when Elsevier
conducted a market study and subsequently expressed its interest in publishing a book
on the topic.
In many ways editing this book has been a privilege and a unique experience.
Thanks firstly to our excellent contributors without whose support, this book would
not have materialized. It is most fitting that this technological work is published from
contributors around the globe and is founded on the spirit of free enquiry coupled
with hard work and imagination. It has indeed been a great pleasure to be in touch
with all contributors during the last 3 to 4 years. Thanks also for their patience and
understanding.
We would be utterly remiss if we did not acknowledge those people who have
provided us with the inspiration, motivation and never-ending encouragement
throughout the course of this work.
Dr Murali would like to acknowledge Mr Ronald Christenson, former Chief
Technology Officer, Cargill Inc; his wife Ponnamma; his children Shubha and
Shilesh, their spouses Chuck Harris and Nupur Parikh and his two lovely grandchil-
dren Reya and Azad; all his teachers and mentors during his entire career; and his
parents who would have been both excited and extremely proud to see this book
published.
xi

Oxford University spin-off in 2009.
H.S. Muralidhara Ph.D is a Chemical Engineer. He retired from Cargill Inc. in
Minneapolis, USA in Oct 2009 as a Vice President, Manager Process Technology,
Corporate Plant Operation after 20 years of service. Prior to that he was a Research
Leader at Battelle Memorial Institute in Columbus, Ohio, USA for 10 years. He is
currently an industrial consultant.
He has over 30 years of industrial experience in separation purification process
technologies including application of membranes in food and bioprocessing. He is a
coinventor of 27 US patents and 15 patents pending. He has edited two books and
has served as a key note speaker in many major international conferences.
xiii
About the Contributors
Vicki Chen is the director of the UNESCO Centre for Membrane Science and
Technology at The University of New South Wales, Australia.
Jim Davies is a biochemical engineer with a PhD from UCL and is currently a
Principal Group Leader in Purification Development at Lonza Biologics UK.
R.W. Field is a Reader in Engineering Science at Oxford University and has many
years of experience in membrane technologies.
Val D. Frenkel, PhD, PE, DWRE, Director Membrane Technologies with Kennedy/
Jenks Consultants, is the company-wide leader for Membrane Technologies.
Dr. Frenkel formed and leads the firm’s Membrane Technology Group and has
25 years of experience in engineering, with expertise in water and wastewater
treatment, water reuse, and membrane technologies, including desalination.
Yu Jiang was awarded her DPhil in 2009 from Oxford University. Her DPhil thesis
was entitled “Emergency drinking water device based on gravity driven
ultrafiltration”.
Bassam Jirjis is a Principal Chemical Engineer at Cargill Inc., in Minneapolis,
Minnesota. He has more than 25 years of experience in the area of separation
technologies and food processing.
N.S. Krishna Kumar, PhD, is a Senior Chemical Engineer at Cargill Inc.,

appropriate food/feed regulations.
Martin Smith, also a Biochemical Engineer from UCL, is currently a Bio-Process
Consultant at eXmoor Pharma Concepts in the fields of Biopharmaceuticals and
Regenerative Medicine.
Yinhua Wan, DPhil, is a Professor of Biochemical Engineering at Institute of
Process Engineering, Chinese Academy of Sciences. He has published more than
80 papers in refereed journals and holds a number of patents in membrane separa-
tion technologies.
xvi About the Contributors

Chapter 1
Fundamentals of
Pressure-Driv en Membrane
Separation Processes
Z.F. Cui, Y. Jiang and R.W. Field
Department of Engineering Science, Oxford University, Oxford, UK
Table of Contents
1.1 Introduction
1.2 Processes
1.2.1 Process Classification
1.2.2 Definitions
1.3 Membranes
1.3.1 Membrane Structures
1.3.2 Membrane Materials
1.3.3 Membrane Modules
1.4 Operation
1.4.1 Concentration
Polarization
1.4.2 Membrane Fouling
1.5 Prediction and Enhancement

using porous membranes for its close relevance to food and bioproduct
processing.
Membrane separation processes can be used for a wide range of applica-
tions
and
can
often offer significant advantages over conventional separation
such as distillation and adsorption since the separation is based on a physical
mechanism. Compared to conventional processes, therefore, no chemical,
biological, or thermal change of the component is involved for most mem-
brane processes. Hence membrane separation is particularly attractive to the
processing of food, beverage, and bioproducts where the processed products
can be sensitive to temperature (vs. distillation) and solvents (vs. extraction).
1.2 PROCESSES
1.2.1 Process Classification
There are four major pressure-driven membrane processes that can be divided
by the pore sizes of membranes and the required transmembrane pressure
(TMP): MF (0.1À5 μm, 1À10 bar), UF (500À100,000 Da, 1À100 nm,
1À10 bar), NF (100À500 Da, 0.5À10 nm, 10À30 bar), and RO (,0.5 nm,
35À100 bar). Figure 1.2 pres
ents
a
classification on the applicability of differ-
ent membrane separation p rocesses based on particle or molecular sizes. RO
process is often used for desalination and pure water production, but it is the
UF and MF that are widely used in food and bioprocessing.
Mixture
A + B
Component A
Component B

F
m
R
Permeate
m
P
FIGURE 1.3 A realistic membrane separation process.
0.001
Ionic
range
Macromolecular
range
Micron
particle
Fine
particle
Coarse
particle
µm 0.01 0.1 1.0 10 100 1000
Membrane process
Reverse
osomosis
Nano-
filtration
Ultrafiltration
Microfiltration Cloth, Fiber filter
Screens
FIGURE 1.2 The applicability ranges of different separation processes based on sizes.
3Chapter | 1 Fundamentals of Pressure-Driven Membrane Separation Processes
containing both the material that has been rejected by the membrane and a

tot
can be defined as:
J
tot
5
X
n
i 5 1
J
i
ð1:1Þ
The retention factor R
i
of a component i can be defined and used as a
measure of performance.
R
i
5 1 2
C
P;i
C
R;i
ð1:2aÞ
where C
P
and C
R
are the concentration of component i in the permeate and
the retentate.
Actually pressure-driven membrane processes can be operated in two dif-

4 Membrane Technology
which is the volume of m
p
produced per unit of membrane area per unit
time. Usually there is only one species, microparticle or macromolecule, to
be interested, and the rejection will only be referred to the concerned spe-
cies. Often the permeate flow rate is much less than the retentate flow rate in
a single pass, hence the change of concentration in the retentate is not signif-
icant. The rejection can then be conveniently calculated by:
R 5 1 2
C
P
C
F
ð1:2bÞ
where C
F
is the feed concentration.
The driving force in pressure-driven membrane separation is of course the
pressure, or the pressure difference between the upstream and the downstream
of the membrane, or between the feed and the permeate. This is referred to as
transmembrane pressure. As the pressure may vary in the membrane module
due to crossflow, an averaged pressure difference over the module is used:
TMP 5
ðP
F
2 P
P
Þ
in

R
C
J
Time
FIGURE 1.4 The schematic diagrams of the dead-end mode and the cross-flow mode, and their
effects on the permeate flux and the height of the cake layer (R À resistance as refereed later).
5Chapter | 1 Fundamentals of Pressure-Driven Membrane Separation Processes
bodies. The pores are of uniform size (isotropic) or nonuniform size (anisotro-
pic). Microporous membranes are designed to reject all the species above their
ratings. However, they tend to be blocked by the species that are of similar
sizes as the pores. The asymmetric membrane has a selective skin layer on the
top of its membrane body. The membrane body is usually void, only giving
mechanical support to the selective skin layer. Compared to the microporous
membranes, the asymmetric membranes rarely get blocked. Most UF, NF, and
RO membranes are of asymmetric structure, while most polymeric MF mem-
branes are of microporous structure.
1.3.2 Membrane Materials
In terms of materials, membranes can be classified into polymeric or organic
membranes and ceramic or inorganic membranes. Organic membranes are usu-
ally made up of various polymers, among which the typical ones are cellulose
acetate (CA), polyamide (PA), polysulfone (PS), polyethersulfone (PES), poly-
vinylidene fluoride (PVDF), polypropylene (PP), etc. Polymeric membranes are
relatively cheap, easy to manufacture, available in a wide range of pore sizes,
and they have been widely used in various industries. Nevertheless, most of the
polymeric membranes have limitations on one or more operating conditions
(either pH, or temperature, or pressure, or chlorine tolerance, etc.), which hinder
their wider applications. For example, CA is the classic material usually used to
produce the skinned membranes. However, it has many disadvantages such as
low temperature limit (30À40


characteristics: (1) due to their large internal diamet ers, tubular modules are
capable of dealing with the feed stream containing fairly large particles.
Furthermore, they can be easily cleaned by using either mechanical or chemi-
cal cleaning methods; (2) they need large pumping capacity, because they are
usually operated under the turbulent flow conditions with the Reynolds num-
bers greater than 10,000; (3) they have the lowest surface area-to-volume ratio
among all the four membr ane configurations. The holdup volumes of tubular
modules are also high, which need large floor space to operate.
Hollow fiber modules are actually the “thin” tubular membranes in com-
pact modules, but in the form of self-support that enables them to withstand
high backpressure. Normally, hollow fiber modules are composed of
50À3000 individual hollow fibers, bundled and sealed together on each end
with epoxy in a hydraulically symmetric housing. The fiber diameters typi-
cally range from 0.2 to 3 mm (except those used in RO, which may be as
thin as 0.04 mm and can withstand much higher pressure). The fiber lengths
range from 18 to 120 cm. In MF and UF, hollow fiber modules are oper-
ated in the inside-out mode with selective skin layers on the inner sides of
the fibers , while in RO, they are operated in the outside-in mode with
selective skin layers on both sides of the fibers . Hollow fiber modules have
some very different characteristics from tubular modules: (1) they are
recommended to operate with the Reynolds numbers in the range of
500À3000, therefore, most of them are run in the laminar flow region.
Additionally, the pressure rating of hollow fiber modules is low, normally
with a maximum of 2.5 bar; (2) due to the combination of low cross flow
rate and low pressure drop, hollow fiber modules are one of the more eco-
nomical modules in terms of energy consumption; (3) hollow fiber modules
have the highest surface area-to-volume ratio amo ng all the four membrane
configurations, and their holdup volumes are low; (4) because the fibers are
self-supported, hollow fiber modules have good backwash capacity and are
hence easy to clean; and (5) one distinct disadvantage of hollow fiber mod-

brane modules; and (4) suspended particles can easily block the mesh-like
spacers and then partially block the feed channel. Therefore, spiral-wound
modules require relatively clean feed that are with minimum content of sus-
pended particles. The pretreatment to reduce suspended particles is needed
for spiral-wound modules.
1.4 OPERATION
1.4.1 Concentration Polarization
Concentration polarization refers to the reversible accumulation of rejected
molecules close to the membrane surface. In membrane processes all compo-
nents in the feed are transported to the membrane surface by convection, and
the rate increases as the permeation through the membrane increas es. The
selectivity of the membrane holds back the less permeable components. At
steady state, these less permeable components have to be transported back
into the bulk of the feed stream. As the flow next to the membrane surface is
laminar, this transport can only be diffusive. The transport has to be based on
the established concentration gradient, i.e., an enrichment of the less
8 Membrane Technology
permeable components at membrane surface, as shown in Figure 1.5.Itisa
natural consequence of membrane selectivity and is equivalent to the mass
transfer boundary. If driving force is removed, permeation ceases, and such a
concentration polarization phenomenon disappears.
Under steady-state conditions, the following relationships describe the
relevan
t
flu
xes based on Figure 1.5:
Component 1
J
1;con
5 J

i
UC
i;P
2 D
ji
dC
i
dz
ð1:6Þ
Membrane
J
1
J
2
J
1,con
J
2,con
C
1F
C
1F,M
J
2,diff
I
b
Z
Concentration boundary
layer
Bulk

i;b
2 C
i;P

ð1:7Þ
In Equation (1.7), the term (D
ji
/l
b
) can be described as a mass transfer
coefficient k
i,b
.
For one interested species to be rejected and the solvent to be just water,
Equation (1.7) can be rewritten as:
J 5 kUln
C
M
2 C
P
C
B
2 C
P

ð1:8Þ
and for a total rejection operation where C
P
5 0, we have
J 5 kUln

FIGURE 1.6 The influences of operating parameters on permeate flux, showing the pressure
control region and the mass transfer control region.
10 Membrane Technology
macromolecule’s gelation concentration or the solubility of the rejected salt,
gelation or salt precipitation occurs and C
M
reaches its maximum value.
Further increase in TMP does not have any effect on C
M
, and hence the flux,
J, does not change 2 a region known as pressure-independent region.
On the other hand, increase in the mass transfer coefficient, k, by
increas-
ing
cross-flow velocity leads to a higher permeate flux, as indicated in
Equation (1.9).
The mass transfer coefficient can be estimated on the basis of heat and
mas
s
transfer
analogy (so-called Colburn analogy) using the semiempirical
Sherwood correlation. This correlation can be written as:
Sh 5 aURe
b
USc
c
U
d
l


4
Re . 2300.
TABLE 1.2 Dimensionless Numbers
Reynolds number: Re 5
ρud
h
μ
Sherwood number: Sh 5
kUd
h
D
Schmidt number: Sc 5
μ
DUρ
5
ν
D
Hydraulic diameter:
For tubes d
h
5 d
tube
For noncircular channel d
h
5 4
Cross-section area
Wetted perimeter
d
h
, equivalent hydraulic diameter; D, diffusivity of the rejected species; ρ, density of the feed

Foulants Fouling mode
Large suspended
particles
Particles present in the original feed or developed in the process
by scaling can block module channels.
Small colloidal
particles
Colloidal particles can rise to a fouling layer. Fouling of
membranes in recovery of cells from fermentation broth.
Macromolecules Gel or cake formation on membrane. Macromolecular fouling
within the structure of porous membranes.
Small molecules Some small organic molecules tend to have strong interactions
with plastic membranes (e.g., antifoaming agents such as
polypropylene glycols used during fermentation foul certain plastic
ultrafiltration membranes).
Proteins Interactions with surface or pores of membranes.
Chemical
reactions
Concentration increase and pH increase can lead to precipitation
of salts and hydroxides.
Biological Growth of bacteria on the membrane surface and excretion of
extracellular polymers.
12 Membrane Technology
Generally speaking, four fouling mechanisms for porous membranes can
be observed, as shown in Figure 1.7:
(a) complete pore blocking,
(b) intern
al
pore blocking,
(c) partial pore blocking, and

tion polarization. In this model, it can be assumed that (1) membrane pores
(A) Complete pore blocking (B) Internal pore blocking
(C) Partial pore blocking (D) Cake filtration
FIGURE 1.7 Fouling mechanisms of porous membranes.
13Chapter | 1 Fundamentals of Pressure-Driven Membrane Separation Processes
TABLE 1.4 Fouling Mechanisms, Phenomenological Background, Effect on Mass Transport, and Transport Equations
Fouling Mechanism n Phenomenological
Background
Effect on Mass Transport Transport Equation
Complete pore
blocking (see
Fig. 1.7a)
2 Particles
larger than the pore
size; the active membrane area
(pores) reached by particles is
blocked.
Reduction of the active
membrane area. Depending on
feed velocity, permeate might
be increased by increasing
transmembrane driving force
(pressure).
J 5 J
0
UK
b
AUt ð1:12Þ
Internal pore blocking
(see Fig. 1.7b)

Particle pore blocking
(see Fig. 1.7c)
1 An
y
particles reaching a pore
might seal it over time. Particles
might bridge a pore and not
block it completely.
Reduction of active membrane
area. The effect is similar to
pore blocking but not so severe.
J 5 J
0
U½1 1 K
i
UðAUJ
0
ÞUt
2 1
ð1:14Þ
Cake filtration (see
Fig. 1.7d)
0 Forma
ti
on of a cake on the
membrane surface of particles
that do not enter the pores.
The overall resistance becomes
the resistance of the membrane
plus the resistance of the cake.

where J
v
is the permea te flux, n
p
the number of cylindrical pores per unit
area, d
p
the pore diameter, ΔP the TMP, l
p
the pore length, and μ the viscos-
ity of the permeate.
The equation shows that the permeate flux is directly proportional to the
TMP and inversely proportional to the viscosity. The viscosity is primarily
controlled by the solvent type, feed composition, and temperature. Therefore,
in the pressure control region, increasing the temperature and pressure, and
decreasing the feed concentration can increase the permeate flux.
The resistance model is developed to express the entire TMP-flux behav-
ior in MF and UF, both in the pressure control region and in the mass trans-
fer control region. This model is based on the resistance-in-series concept,
which is a common concept in heat transfer. With the ideal membrane and
the ideal feed that lead to no fouling, the model can be expre ssed as:
J
v
5
ΔP
μR
m
ð1:17Þ
where J
v

to the permeate to pass through. Therefore, the resistance of the polarized
15Chapter | 1 Fundamentals of Pressure-Driven Membrane Separation Processes