Membrane Processing E-Book
159,99 €
Mehr erfahren.
- Herausgeber: John Wiley & Sons
- Kategorie: Fachliteratur
- Serie: Society of Dairy Technology Series
- Sprache: Englisch
In the last two decades, there have been significant developments in membrane filtration processes for the dairy and beverage industries. The filtration systems can be classified into four main groups: reverse osmosis, nanofiltration, ultrafiltration and microfiltration. The primary objective of this book is to assess critically the pool of scientific knowledge available to the dairy and beverages industry, as a tool for process and product innovation, quality improvement and safety.
The book is divided into three main parts. Part I reviews the principals, developments and designs of membrane processes that are mainly used in commercial dairy and beverage applications. Part II provides information on the applications of membrane processes in the manufacture of dairy products, from on-farm concentration of milk as a pre-treatment for cheesemaking to fractionation of milk and whey to provide ingredients for food and other applications. Part III considers membrane applications during the manufacture of fruit juices, beer and cider, wine and vinegar. These include concentration, deacidification and dealcoholisation processes.
Membrane Processing: Dairy and Beverages Applications is an ideal new reference for dairy and beverage processors involved in the application of membranes, both to aid the creation of novel products, and to improve their process economics. Students and lecturers of food and dairy science and technology will value its in-depth discussion of membrane processes, whilst readers based in the dairy industry will prize it as the most up-to-date and advanced volume yet published on this crucially important topic.
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 735
Veröffentlichungsjahr: 2012
Ähnliche
Table of Contents
Title Page
Copyright
Preface to the Technical Series
Preface
Contributors
Chapter 1: Development of Membrane Processes
1.1 Historical background
1.2 Basic principles of membrane separations
1.3 Types of membrane separations
1.4 Theory of membrane transport
1.5 Factors affecting membrane separations
1.6 General characteristics of membrane processes
1.7 Conclusion and future development
Chapter 2: Principles of Membrane Filtration
2.1 Introduction and definitions
2.2 Membrane properties based on materials
2.3 Flux behaviour in pressure-driven membrane operations
2.4 Effects of feed characteristics and operating parameter on separation efficiency
2.5 Cross-flow systems
2.6 Recent membrane processes following different operating principles
2.7 Conclusions
Chapter 3: Commercial Membrane Technology
3.1 Introduction: polymers used in membrane manufacture
3.2 Other materials used for membranes
3.3 Membrane configuration
3.4 Modes of operation
3.5 Conclusion and future developments
Chapter 4: Membrane Fouling, Cleaning and Disinfection
4.1 Introduction
4.2 Flux reduction
4.3 Membrane cleaning and disinfection
4.4 Recent developments
4.5 Conclusions
4.6 Nomenclature
Chapter 5: General Application for the Treatment of Effluent and Reuse of Wastewater
5.1 General wastewater quality issues
5.2 General wastewater treatment
5.3 Water reuse
5.4 Conclusions and future applications
Chapter 6: Liquid Milk Processing
6.1 Introduction
6.2 On-farm concentration of milk
6.3 Protein standardisation by ultrafiltration
6.4 Removal of bacteria by microfiltration
6.5 Fractionation of fat
6.6 Removal of somatic cells by microfiltration
6.7 Conclusions and future trends
Chapter 7: Membrane Processing of Fermented Milks
7.1 Introduction
7.2 Microflora of the starter cultures
7.3 Patterns of production and consumption
7.4 Manufacturing practice of gel-type (set and stirred) products
7.5 Manufacturing practice of concentrated products
7.6 Quality control
7.7 Conclusion
Chapter 8: Cheese
8.1 Background
8.2 Properties of membrane processed concentrates
8.3 Applications of ultrafiltration in cheesemaking
8.4 Cheese quality
8.5 Applications of microfiltration in cheesemaking
8.6 Nanofiltration
8.7 Milk protein concentrates
8.8 Future potential
Chapter 9: Whey Processing
9.1 Introduction
9.2 Whey: components, their functionality and uses
9.3 Problems of traditional whey processing
9.4 Membranes in whey processing
9.5 Conclusions
Chapter 10: Concentrated Milk and Powders
10.1 Introduction
10.2 Concentrated milks and powders
10.3 Milk protein concentrates
10.4 Conclusion and future trends
Chapter 11: Further Applications of Membrane Filtration in Dairy Processing
11.1 Introduction
11.2 Fractionation of milk proteins using membranes
11.3 Fractionation of milk fat using membranes
11.4 Fractionation of milk carbohydrates using membranes
11.5 Fractionation of milk salts using membranes
11.6 Conclusions and future trends
Chapter 12: Fruit Juices
12.1 Introduction
12.2 Fruit juice clarification by microfiltration and ultrafiltration
12.3 Membrane fouling and membrane cleaning
12.4 Performance of microfiltration and ultrafiltration membranes
12.5 Process configurations
12.6 Quality of the clarified juices
12.7 Integrated processes
12.8 Conclusions and future development
Chapter 13: Beer and Cider
13.1 Introduction
13.2 Beer brewing process
13.3 Cidermaking process
13.4 Membrane applications in the brewing process
13.5 Membrane applications in cidermaking
13.6 Membrane applications common to brewing and cidermaking
13.7 Future opportunities
Chapter 14: Wine
14.1 Background
14.2 Clarification and filtration methods
14.3 Membrane fouling
14.4 Must correction, wine correction and alcohol reduction using membrane technologies
14.5 Wine stabilisation and pH adjustment
14.6 Conclusions and future developments
Chapter 15: Application of Membrane Technology in Vinegar
15.1 Introduction
15.2 Process of vinegar making
15.3 Membrane technology in the production of vinegar
15.4 Conclusions
Index
Food science and Technology
The Society of Dairy Technology (SDT) has joined with Wiley-Blackwell to produce a series of technical dairy-related handbooks providing an invaluable resource for all those involved in the dairy industry, from practitioners to technologists, working in both traditional and modern large-scale dairy operations. For information regarding the SDT, please contact Maurice Walton, Executive Director, Society of Dairy Technology, PO Box 12, Appleby in Westmorland, CA16 6YJ, UK. email: [email protected]
Other volumes in the Society of Dairy Technology book series:
Probiotic Dairy Products (ISBN 978 1 4051 2124 8)
Fermented Milks (ISBN 978 0 6320 6458 8)
Brined Cheeses (ISBN 978 1 4051 2460 7)
Structure of Dairy Products (ISBN 978 1 4051 2975 6)
Cleaning-in-Place (ISBN 978 1 4051 5503 8)
Milk Processing and Quality Management (ISBN 978 1 4051 4530 5)
Dairy Fats (ISBN 978 1 4051 5090 3)
Dairy Powders and Concentrated Products (ISBN 978 1 4051 5764 3)
Technology of Cheesemaking, Second Edition (ISBN 978 1 4051 8298 0)
Processed Cheese and Analogues (ISBN 978 1 4051 8642 1)
This edition first published 2013 © 2013 by Blackwell Publishing Ltd.
Blackwell Publishing was acquired by John Wiley & Sons in February 2007. Blackwell's publishing program has been merged with Wiley's global Scientific, Technical and Medical business to form Wiley-Blackwell.
Registered office: John Wiley & Sons, Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
Editorial offices: 9600 Garsington Road, Oxford, OX4 2DQ, UK
The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
2121 State Avenue, Ames, Iowa 50014-8300, USA
For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com/wiley-blackwell.
The right of the author to be identified as the author of this work has been asserted in accordance with the UK Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher.
Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought.
Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with the respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. It is sold on the understanding that the publisher is not engaged in rendering professional services and neither the publisher nor the author shall be liable for damages arising here from. If professional advice or other expert assistance is required, the services of a competent professional should be sought.
Library of Congress Cataloging-in-Publication Data has been applied for
9781444333374 (hardback)
A catalogue record for this book is available from the British Library.
Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books.
Cover image: Adapted with permission from Tetra Pak
Cover design by www.hisandhersdesign.co.uk
1 2013
Preface to the Technical Series
For more than 60 years, the Society of Dairy Technology (SDT) has sought to provide education and training in the dairy field, disseminating knowledge and fostering personal development through symposia, conferences, residential courses, publications, and its journal, the International Journal of Dairy Technology (previously published as the Journal of the Society of Dairy Technology).
In recent years, there have been significant advances in our understanding of milk systems, probably the most complex natural food available to man. At the same time, improvements in process technology have been accompanied by massive changes in the scale of many milk processing operations, and the manufacture a wide range of dairy and other related products.
The Society has embarked on a project with Wiley-Blackwell to produce a Technical Series of dairy-related books to provide an invaluable source of information for practicing dairy scientists and technologists, covering the range from small enterprises to modern large-scale operation. This eleventh volume in this series, on Membrane Processing – Dairy and Beverages Applications, provides a timely and comprehensive update of the principles and practices involved in this technology. The commercial exploitation of membrane technology is developed further in its applications to the processing of milk and milk products, plus a wide range of non-dairy beverages, where membranes provide novel opportunities for separation, fractionation and concentration.
Andrew Wilbey Chairman of the Publications Committee, SDT
Preface
In the last two decades, there have been significant developments in membrane filtration processes for the dairy and beverage industries. The filtration systems can be classified into four main groups: reverse osmosis (RO), nanofiltration (NF), ultrafiltration (UF) and microfiltration (MF). The primary objective of this book is to assess critically the pool of scientific knowledge available to the dairy and beverages industry, as a tool for process and product innovation, quality improvement and safety.
Appraisals of the key technical aspects of membrane processing stages to control and maintain the consistency of dairy and beverages products are also included. Although this has produced some overlap in the coverage of membrane processes, I have felt justified in allowing this overlap because it emphasises the prime importance of processing in the preparation of milks and beverages, to achieve the desired quality and consistency of these products for the end-user.
Generally, the book is divided into three main parts. Part I, which consists of five chapters, reviews the principals, developments and designs of membrane processes that are mainly used in commercial dairy and beverage applications. Successful applications of membrane processes in the food industry and requirements pertaining to ensure hygienic conditions in the equipment are reviewed in Chapter 4, and Chapter 5 details the aspects of food industrial effluent treatment, recovery of detergent from cleaning-in-place (CIP) systems and reuse of wastewater.
The application of membrane technology in the manufacture of different dairy products could be briefly summarised as follows: (a) MF for the removal of bacteria and spores from skimmed milk for the production of an extended shelf-life product; (b) UF is used to concentrate the fat and protein components in milk, which can be used for standardisation of the protein content prior to cheesemaking and also to concentrate the fermentate for the production of concentrated yoghurt (labneh); in addition, a wide range of cheeses are commercially manufactured from ultrafiltrated milk (e.g. Quarg, Feta); (c) reduction of salts in milk is achieved using NF, electrodialysis (ED) has been used for whey; (d) combined applications of MF, UF and NF have been used to fractionate milk components and, in some instances, the composition of cow's milk has been modified to be similar to mare's milk; (e) RO is widely used as a lower energy alternative to evaporation where low degrees of concentration are needed.
Part II of the book provides information on the applications of membrane processes in the manufacture of dairy products. This ranges from on-farm concentration of milk as a pre-treatment for cheesemaking to fractionation of milk and whey to provide a wide range of ingredients for food and other applications.
Part III considers membrane applications during the manufacture of fruit juices, beer and cider, wine and vinegar. These include concentration, deacidification and dealcoholisation processes.
Dairy and beverage processors are encouraged to use this book as a reference in the application of membranes, both to aid the creation of novel products and to improve their process economics.
A.Y. Tamime
Contributors
Editor
Contributors
Chapter 1
Development of Membrane Processes
K. Smith
1.1 Historical background
The ability of membranes to separate water from solutes has been known since 1748, when Abbé Nolet experimented with the movement of water through a semi-permeable membrane. Depending on the reference, either Abbé Nolet or Dutorchet coined the word osmosis to describe the process. Throughout the 18th and 19th centuries, membranes were used exclusively for laboratory applications, and often consisted of sausage casings made from animal intestines or the bladders of pigs, cattle or fish.
The first synthetic membranes were produced by Fick in 1855, and appear to have been made of nitrocellulose. Membranes based on cellulose were known as collodion and had the advantages of reproducible characteristics compared with the previously used animal-based membranes. Bechhold further advanced the process for manufacturing collodion membranes when he developed methods for controlling pore size and measuring pore diameters in 1907. He is generally credited with first using the term ultrafiltration (UF). In addition, Richard Zsigmondy at the University of Göttingen, Germany, patented a membrane filter in 1918 that was referred to as a cold ultrafilter. His work becomes the basis of the membrane filters produced by Sartorius GmbH.
Collodion membranes produced by the Sartorius GmgH of Germany became commercially available in 1927. The primary use of membranes until the 1940s was the removal of micro-organisms and particles from liquids and gases and research applications. There was a critical need to test drinking water in Europe for microbial content following the Second World War, and membranes were developed that could rapidly filter water and capture any micro-organisms on the membrane surface, where they could quickly be enumerated to determine the safety of the water for human consumption.
In addition to the separation of relatively large particles from water, there was interest in developing membranes that could desalinate sea or brackish water. The term reverse osmosis (RO) had been coined in 1931 when a patent was issued for desalting water; however, the available membranes could not withstand the pressures required.
Although many improvements were made in the following years, including the use of other polymers for constructing membranes, membranes were limited to laboratory and small specialised industrial applications. Factors limiting the use of membranes included a lack of reliability, being too slow, not sufficiently selective and cost.
A breakthrough came in the early 1960s when Sourirajan and Loeb developed a process for making high-flux, defect-free membranes capable of desalinating water. Researchers at the time believed the best approach to improving flux would be to reduce the thickness and thereby the resistance to flow of the membrane. Sourirajan and Loeb attempted to produce such membranes by taking existing cellulose acetate membranes and heating them while submersed in water in a process known as annealing. They expected the membrane pores would increase in size by such a process, but instead the pores became smaller and the membrane more dense. When they attempted the same process with cellulose acetate UF membranes, they discovered not only did the pores become smaller but the ability of the membrane to reject salt increased, as did flux. The flux improvement was such that the membranes could be a practical way to desalinate water.
The annealing process of Sourirajan and Loeb had created an anisotropic or asymmetric membrane. Anisotropic membranes have different behaviour depending on which side of the membrane is used for the separation. Although this type of membrane had been seen over 100 years earlier with natural membranes, it had not been reproduced with the synthetic variety.
The key to the anisotropic membrane of Sourirajan and Loeb was the thin ‘skin’ on one surface of the membrane. The skin typically was approximately 0.1–0.2 µm thick and had a dense structure whereas the remainder of the membrane had a very porous open structure. The thickness of the membrane essentially determined the flux and so by reducing the effective separating distance from 100–200 µm to 0.1–0.2 µm the rate of liquid crossing the membrane dramatically improved, but because of the small pores in the skin the rejection of salt remained high.
Many changes in the production of membranes occurred during the 1960s, 1970s and 1980s. By continuing the work of Sourirajan and Loeb, others were able to develop additional methods for producing membranes. Initial membrane modules were plate-and-frame (Danish Sugar Corporation) or hollow fibre (Amicon) designs, but membranes in formats, such as spiral-wound and tubular (Abcor), were introduced shortly afterwards. The thickness of the separating layer was further reduced to less than 0.1 µm. Large plants using RO, UF and microfiltration (MF) were operating around the world by 1980.
Cellulose acetate remained the material of choice until the mid-1970s, when methods of producing composite membranes for water desalination were developed. By combining polysulphone and polyamide, composite membranes had the advantage of high salt rejection combined with good water flux and increased resistance to temperature and chemicals. Nanofiltration (NF) or ‘loose RO’ membranes became available in the mid-1980s. The NF membranes operated at lower pressures than RO systems, and were able to permeate monovalent ions. They found immediate application in producing ultrapure water by permeating trace salts from water produced by RO.
In addition, membranes made from inorganic materials, such as zirconium and titanium dioxide, became commercially available in the mid-1980s. Membranes made from these materials are referred to as mineral or ceramic and are available in tubular form for UF and MF. Union Carbide (USA) and Societé de Fabrication d'Elements (France) used carbon tubes covered with zirconium oxide for their inorganic membranes. Later Ceravèr (France) used a ceramic base with aluminium oxide. Chemical and temperature resistance were the significant advantages of ceramic membranes. It was originally thought that such membranes had an unlimited life, but subsequent experience has shown this is not the case.
Advancements in membrane composition and design along with operation of membrane systems have continued. A wide variety of membrane polymers and designs have been adapted for RO, NF, UF and MF, resulting in many commercial applications. The feasibility of membrane-based applications depends chiefly on the ability of the filtration process to economically produce an acceptable product. Membrane pore size distribution, selectivity, operating conditions, membrane life, capital and operating costs become important economic considerations. These parameters are in turn influenced by many factors, such as the membrane polymer, element configuration and system design.
1.2 Basic principles of membrane separations
Membrane filtration is a pressure-driven separation process using semi-permeable membranes. The size of membrane pores and the pressure used indicate whether the term RO, NF, UF or MF is used for a given separation. RO and NF systems use the highest pressures and membranes with the smallest pores, whereas MF has the lowest operating pressures and membranes with the largest pores. UF is intermediate in pressure used and membrane pore size.
1.2.1 Depth versus screen filters
In the past, filtration processes relied on depth filters. This type of filter has fibres or beads in a mesh-like structure. Particles in the feed solution become trapped or adsorbed within the filter network, which eventually clogs the filter, thereby resulting in replacement of the filter (Fig. 1.1). By contrast, screen type filters generally rely on pores, with the size and shape of the pores determining passage of particles. Pores are more rigid and uniform and have a more narrowly defined size than mesh openings in a depth filter. Components not able to pass through pores remain on the membrane surface and, therefore, do not typically become trapped within the membrane structure (Fig. 1.2). Because the fouling materials remain on the surface, internal fouling decreases and the membrane can be reused.
Figure 1.1 Depth filter with particles entrapped within the membrane structure.
Figure 1.2 Screen-type membrane separates particles at the membrane surface.
1.2.2 Isotropic versus anisotropic membranes
Membranes can have several types of internal structure. Terms, such as microporous, non-porous, isotropic and anisotropic, refer to the structure of the membrane. Typically, membranes are either isotropic or anisotropic. Microporous and non-porous refer to isotropic membrane structure. An isotropic membrane will have a relatively uniform structure (Fig. 1.3), i.e. the size of the pores is similar throughout the membrane. The membrane, therefore, does not have a top or bottom layer, rather the membrane properties are uniform in direction. Isotropic membranes generally act as depth filters and, therefore, retain particles within the internal structure resulting in plugging and reduced flux.
Figure 1.3 Structure of an isotropic membrane.
Microporous and non-porous membranes typically are isotropic. Microporous membrane structure can resemble a traditional filter; however, the microporous membrane has extremely small pores. Materials are rejected at the surface, trapped within the membrane or pass through pores unhindered, depending on particle size and size of the pores. A non-porous membrane will not have visible pores and materials move by diffusion through the membrane.
An anisotropic or asymmetric membrane has pores that differ in size depending on their location within the membrane (Fig. 1.4). Typically, anisotropic membranes will have a thin, dense skin supported by a thicker and a more porous substructure layer. The thin top layer provides high selectivity, whereas the porous bottom layer has good flux. Membranes used for commercial separations in the food industry are typically anisotropic.
Figure 1.4 Structure of an anisotropic membrane.
1.2.3 Cross-flow filtration
RO, NF, UF and MF systems all involve cross-flow filtration, which can be compared to the traditional method of perpendicular filtration. In traditional filtration (Fig. 1.5), the entire feed stream passes through the filtering media, i.e. the incoming stream flows perpendicularly to the filter with the filter retaining any trapped solids. The result is a filtered stream with solids trapped on and within the filter.
Figure 1.5 Traditional filtration with perpendicular flow.
In cross-flow filtration (Fig. 1.6), the feed stream passes parallel to the membrane. Some of the incoming feed stream and particles will cross the membrane into the permeate section, whereas the other portion with the concentrated solids is the retentate stream. At any time only some of the water and particles will cross the membrane into the permeate stream, unlike traditional filtration where most particles are trapped after one pass through the filter. Because the feed stream flows parallel to the membrane rather than perpendicular to it, cross-flow filtration is self-cleaning by comparison. Solutes and particles are continually swept along and away from the membrane surface by the retentate stream, thereby allowing longer operating times without cleaning than is possible with traditional filtration.
Figure 1.6 Cross-flow filtration.
The affect of cross-flow permeate flow and thickness of the fouling or cake layer can be seen in Fig. 1.7. In perpendicular filtration, the flow of permeate is reduced as the thickness of the material on the surface of the filter, i.e. the thickness of the cake layer, increases over time. With cross-flow filtration, however, the thickness of the material on the membrane is limited by action of the feed stream sweeping across the surface of the membrane. Because the thickness of the deposited material is limited, permeate flow is maintained at a higher level throughout filtering.
Figure 1.7 Effect of perpendicular and cross-flow filtration on flux and cake thickness.
1.2.4 Requirements of membrane processes
The shared characteristics of membrane processes are pressure-driven processes using semi-permeable membranes. Pressure is used to reverse the direction of the osmosis process, while differences in membrane permeability determine separation of molecules. The process of osmosis is illustrated in Fig. 1.8. Solutions containing two different concentrations of dissolved materials are separated by a membrane that will allow only water to cross (Fig. 1.8a). Nature will try to equalise the concentration of the two solutions. Since the dissolved material cannot cross the membrane, water must flow from the solution of lower concentration to the solution at the higher concentration (Fig. 1.8b). The flow of water will continue until the solutions are of equal concentration or no more water is available. The difference in the height of water in the corresponding tubes is a result of the movement of water from lower to higher concentration. The final water level in the more concentrated solution compared with the original level is equal to the apparent osmotic pressure.
In the process of ‘reverse’ osmosis, pressure is used to force water to flow in the opposite direction (Fig. 1.8c). Enough pressure must be applied to overcome the apparent osmotic pressure of the more concentrated solution before water can flow from the more concentrated to the less concentrated side. In doing so, the more concentrated side becomes even more concentrated through the loss of water. It is this ability to concentrate and separate that is taken advantage of in commercial membrane separations.
Figure 1.8 The processes of osmosis and reverse osmosis.
Another shared characteristic is the use of semi-permeable membranes. Membranes can be distinguished from filters by the size of the particulates that are separated. By convention, filters generally separate particulates that are greater than 1–10 µm in size, whereas membranes separate smaller particles. Semi-permeable refers to the ability to separate some particles from other particles.
1.3 Types of membrane separations
The classification of membranes as RO, NF, UF and MF is somewhat arbitrary, and has considerable overlap between categories. Generally, RO/NF membranes will retain molecules in the ionic size range, UF membranes will separate macromolecules, and MF will retain particles of micron size. Because RO, NF, UF and MF membranes differ in the size of molecules they separate, the osmotic pressure involved is considerably different between the processes. RO, which retains the smallest molecules, has the highest osmotic pressure to overcome and, therefore, requires the highest operating pressure. A range from 1.38 to 8.28 MPa is common for RO, 1.03 to 2.76 MPa for NF, 0.21 to 1.03 MPa for UF, and MF requires only from 0.07 to 0.69 MPa.
1.3.1 Reverse osmosis
RO membranes generally retain all compounds allowing only water to cross into the permeate. There are exceptions to this general statement and, at times, relatively large molecules may pass into the permeate. RO membranes can, therefore, either concentrate a feed stream (retentate stream) through removal of water or produce very pure water (permeate stream).
1.3.2 Nanofiltration
NF membranes are very similar to RO membranes with the exception that NF membranes will allow the passage of monovalent ions into the permeate. NF membranes are very effective at concentrating materials in the feed stream since only monovalent ions are removed into the permeate. The loss of monovalent ions into the permeate allows NF systems to operate at lower pressures than RO systems.
1.3.3 Ultrafiltration
UF membranes both fractionate and concentrate materials in the feed stream. Larger components, such as micro-organisms, lipids and proteins, typically are retained by UF membranes, whereas smaller molecules, such as sugars and minerals, pass into the permeate. UF systems operate at much lower pressures than RO and NF systems because of the permeation of sugars and minerals into the permeate stream.
1.3.4 Microfiltration
MF membranes also fractionate materials in the feed stream. Because MF membranes allow the passage of many larger components, such as smaller proteins, they are not as good at concentrating the feed stream compared with UF membranes. Operating pressures are the lowest of the four types of membranes and, in fact, MF membrane systems operated at higher pressures often lose their ability to permeate larger molecules and behave like UF membranes.
1.4 Theory of membrane transport
Although each of these membrane processes separates molecules of different sizes, the method of separation varies considerably. Generally, UF and MF membranes separate molecules based on size, shape and charge, whereas separation during RO and NF is based neither on size nor on shape.
1.4.1 Transport models
Two models, pore flow and solution diffusion, have been proposed for describing the movement of water and solutes into the permeate. The most significant difference between the two models is the size and permanence of membrane pores. In the pore flow model (Fig. 1.9), pressure-driven convective flow drives solutes through small pores or openings in the membrane. Solutes that are too large to move through the pores remain behind thereby resulting in a separation of solution components.
Figure 1.9 Pore flow model.
Separations in the solution diffusion model (Fig. 1.10) are due to differences in the solubility of solution components in the membrane and the rate at which the components will move across the membrane. Components flow from high to low pressure. Movement of individual components is not related and is determined by mobility, concentration and pressure gradients.
Figure 1.10 Solution diffusion model.
The size of the pores within a membrane probably determines the model that applies. A pore diameter of 5–10 Å probably represents the transition zone between the two models. It is very difficult to directly measure the size of pores within a membrane. Pore size often is inferred from an indirect technique, such as the size of molecules that will permeate the membrane. Given the limitations in determining membrane pore size, it is possible to assign RO, NF, UF and MF membranes into three general categories of separation models. RO membranes, which have no visible pores, are thought to follow the solution diffusion model. Separation in UF and MF membranes is probably based on the pore flow model. NF membranes are in the transition zone between the two models and, therefore, probably use a combination of the two models.
1.4.2 Reverse osmosis/nanofiltration membranes
Pores are not discernible in RO/NF membranes; however, if pores are present, they are thought to be 1–5 Å in diameter. It is the structure of water within the membrane that is important. The RO/NF membrane can be considered a water-swollen gel. Water is thought to move through the membrane based on diffusion and the ability of several water molecules to form a tetrahedral or ice-like structure through hydrogen bonding. One theory holds that water is absorbed into the voids of the RO/NF membrane where the water molecules form this tetrahedral structure. When pressure is applied, water on the retentate side of the membrane joins the structured water within the membrane, while water in the membrane on the permeate side is released into the permeate stream. Only those molecules or ions able to fit into this tetrahedral structure can cross the RO/NF membrane. Because ions are surrounded by a water shell, they do not readily fit into the ice-like structure and, therefore, do not cross into the permeate stream; however, small molecules, such as methanol, urea and lactic acid that exhibit hydrogen bonding, may be able to cross the membrane.
Based on this theory of RO/NF systems, water and solute passage across the membrane is not connected. Water flow into the permeate is based on pressure, whereas solute passage is based on the concentration gradient. An increase in pressure, therefore, will increase the flow of water across the membrane while solute passage remains unaffected and results in less solute in the permeate. Because each solute has its own electrochemical potential, the flux of each component is not related to other solutes. NF membranes are sometimes referred to as ‘loose RO’ membranes. In addition to the passage of compounds that resemble the structure of water, NF membranes also permit the passage of monovalent ions while rejecting polyvalent ions. Additional factors affecting the separating ability of both RO and NF membranes are discussed in Section 1.5.
1.4.3 Ultrafiltration/microfiltration membranes
Both UF and MF membranes separate compounds largely on size. Fundamentally, the membrane acts as a sieve by rejecting those molecules too large to fit through pores, while permitting smaller molecules to cross into the permeate stream. The change in pressure across the membrane is the driving force. Although molecular weight often is used to indicate the size of molecules retained by the membrane, other factors are important and are discussed in Section 1.5.
1.5 Factors affecting membrane separations
The methods of separation for RO and UF differ significantly; therefore, the factors affecting these processes also will differ. However, NF will more closely resemble RO, whereas MF is more similar to UF.
1.5.1 Factors affecting reverse osmosis/nanofiltration separations
Permeability of components during RO or NF is not based on size, but instead depends on the ability of a compound to cross the membrane using a pressure-driven solute diffusion process. Non-electrolytes and electrolytes, therefore, will be affected by different factors. Pressure, temperature, concentration of components and the type of compound affect RO/NF separations. Permeability of solution components during RO/NF also is affected by membrane composition. In general, as pressure increases the rate of water permeation increases. In turn, the concentration of dissolved solids in the permeate increases. A similar trend occurs with temperature during RO/NF. Increasing the temperature of the process increases the rate of permeation; however, the concentration of dissolved solids in the permeate also increases.
Permeation of components that typically do not cross a membrane is less affected by the presence and concentration of other solutes than components in the form of electrolytes. Conversely, electrolyte retention is affected by concentration and other solutes, with less retention of these components as their concentration increases. An example of the effect of membrane composition on permeability is illustrated by the membrane polymer cellulose acetate, which has a strong sorption of higher aliphatic alcohols and flavour compounds. Cellulose acetate is hydrophobic as are many of these compounds which, therefore, are drawn to the membrane as a result of being repelled by the water phase. Interactions between cellulose acetate and water also have a large affect on salt permeability during RO/NF. Increased interactions between cellulose acetate and water result in less water available for hydration of ions. This causes the salts to be less soluble and, therefore, more energy is required to hydrate these salts. Different amounts of energy needed to hydrate salts account for differences in permeability.
The amount of charge on the ions present in solution is very important to their permeability with NF membranes. NF membranes typically reject polyvalent ions while permitting the passage of most monovalent ions. The concentration of ions also is a factor with rejection of ions increasing as the concentration of ions increases in the solution.
1.5.2 Factors affecting ultrafiltration/microfiltration separations
Size, shape, deformability and hydrodynamic radius of a molecule are very important in determining whether a component is retained during UF/MF. Generally, linear, easily deformed molecules are less likely to be retained than highly structured, rigid molecules of equivalent molecular weight. Under pressure, the more flexible molecules can deform to fit through pores generally considered too small to allow their passage. Globular proteins, therefore, are used to define the molecular weight cut-off of a UF membrane since they are less likely to deform under pressure.
Membrane–molecule interactions may be more important than physical factors in determining the ability of a molecule to cross a UF membrane. An example of two materials having opposite effects is styrene and erythrocytes. Styrene, although small enough to cross UF membrane pores, will interact with pore walls and form agglomerates effectively blocking the pore. Erythrocytes, in contrast, are too large to pass through pores; however, their flexibility and lack of interactions with pore walls allow them to cross into the permeate.
Ionic strength and pH of the solution also influence the apparent size of a molecule. For example, a highly charged polyelectrolyte is more easily retained when in solution of lower ionic strength since these conditions give the molecule a larger effective radius and, therefore, limit passage through membrane pores.
The type of compound also is important. Acids, for example, are retained less readily than corresponding salts. Small organic molecules (alcohols, esters, aldehydes and ketones) and small non-ionic inorganic molecules, such as ammonia, are more likely to cross a membrane than ionic compounds.
Composition of the feed material also influences permeability. An example is small molecular weight proteins, which are more likely to cross a membrane when in a pure solution than when ultrafiltered in the presence of compounds such as larger molecular weight proteins.
1.5.3 System parameters
System parameters, such as operating conditions and membrane composition can influence passage of a molecule. Operating parameters including temperature, feed velocity, pressure and feed concentration can alter permeation of a molecule. For example, use of higher pressures and lower feed velocities can result in the rejection of proteins by a MF membrane that otherwise would permeate the protein. The affects of system parameters on separations will be discussed in a later section.
1.6 General characteristics of membrane processes
Certain characteristics of a particular membrane are especially important since they influence the economic viability and possible applications of the membrane system. Pore size and retention can be used as an indication of the ability of a membrane to retain a certain size molecule. Flux and membrane life, in contrast, affect the economics of processing with membranes.
1.6.1 Retention and rejection
The terms retention and rejection may be used interchangeably depending on whether the component is desired or undesired in the retentate stream. Retention (R) can be defined in several ways, and one common definition is:
1.1
where Cf is the concentration of a component in the feed stream and Cp is the concentration of a component in the permeate stream.
If a component freely permeates the membrane, R will be near zero, whereas a completely retained component has an R value of 1, and expressed as a percentage would be equivalent to 100% retention. Components having a rejection or retention value of zero will be found in the same concentration in the permeate and feed streams.
An alternative system that measures membrane performance according to permeability (P) is as follows:
1.2
or
1.3
Although easily defined, rejection characteristics of a membrane are affected by many factors, and there is no uniform set of conditions used by manufacturers to determine this property. Pore diameter, chemical composition of the membrane and interactions between the membrane and feed material are major factors influencing rejection characteristics of a membrane. Rejection values also change depending on operating conditions and during operation of a membrane system.
Rejection is one method of classifying a membrane and, typically, RO and NF membranes are rated according to their ability to reject sodium chloride or other salts. Because RO membranes are used extensively for water desalination, the ability of the membrane to limit the passage of sodium chloride into the permeate is extremely important in the production of water for human consumption. Even small amounts of salt in drinking water can affect taste of the water; therefore, RO membranes often are rated on salt rejection.
1.6.2 Pore size
Pore size is used by some membrane manufacturers to indicate the separating ability of a membrane. Units of microns (µm) are often used. Reference to pore size is typical of membranes in the MF range or to specific categories of membranes, such as ceramic and track etched.
1.6.3 Molecular weight cut-off
Manufacturers often use molecular weight cut-off to indicate the separating ability of a given membrane. Membranes with the same molecular weight cut-off, however, may not have the same retention for a compound since manufacturers use a variety of methods to determine molecular weight cut-off. Generally for a given molecular weight cut-off, 80% of the molecules of that molecular weight will be rejected. Furthermore, molecular weight cut-off is typically used for membranes in the UF range.
1.6.4 Flux
Flux is the amount of permeate produced in a given time period, and the term generally is given as a volume or mass per unit membrane per unit time. A unit, such as L m−2 h−1, is a possible volume flux unit, whereas kg m−2 h−1 is a mass flux unit.
Flux determines the area of membrane required to process a given amount of product to a certain concentration in a specific time period. The lower the flux for a given membrane the greater the membrane area required to process the same amount of product within a certain time than with a higher flux membrane. Flux, therefore, affects the economics of an operation, and is used as an indication of membrane fouling and cleaning adequacy. Factors influencing flux include pressure, feed velocity, temperature, viscosity and turbulence.
1.6.5 Concentration factor
Concentration factor (CF) expresses the degree of concentration for a feed and can be defined as:
1.4
Concentration factor is often expressed as 1×, 2×, 3× etc. Volumes or weights may be used. An example would be 100 L of product that is processed to a final volume of 33 L:
The ability to concentrate a feed material generally is determined by feed constituents, osmotic pressure and feed viscosity.
1.6.6 Membrane life
Although the cross-flow design allows reuse of a membrane, the operating life of a membrane is not indefinite. As membranes are used for processing, their characteristics flux and retention change, and it is the decreasing flux and retention with time that result in membrane replacement. The life of a membrane ends when membrane performance no longer meets specific performance criteria. In the case of a protein concentration operation, the loss of valuable protein into the permeate might be a criteria. In a water desalination system the criteria might be the permeation of excessive amounts of salt into the permeate, which is to be as drinking water.
Membrane life is influenced by many factors. Membrane composition is important, but often the processing and cleaning conditions have a greater influence. Composition, however, affects the resistance of a membrane to processing and cleaning treatments, and determines acceptable operating conditions.
1.7 Conclusion and future development
Although membranes have been known about for many decades, it has not been until relatively recently that technological improvements have permitted their widespread use by the food industry. Thus, the use of membrane processes for food applications can be considered a relatively new tool for the food processor. Knowledge about the theory of membrane systems operate will continue to increase as new applications for membrane systems are found.
Suggested literature
Baker, R.W. (2000) Membrane Technology and Applications, McGraw-Hill, New York.
Cheryan, M. (1998) Ultrafiltration and Microfiltration Handbook, Technomic Publishing Co., Pennsylvania.
Glimenius, R. (1985) Microfiltration—state of the art. Desalination, 53, 363–372.
Gregor, H.P. & Gregor, C.D. (1978) Synthetic-membrane technology. Scientific America, 239, 112–128.
Jonsson, G. (1980) Overview of theories for water and solute transport in UF/RO membranes. Desalination, 35, 21–38.
Lonsdale, H.K. (1972) Theory and practice of reverse osmosis and ultrafiltration. Industrial Processing with Membranes (eds. R.E. Lacey & S. Loeb), pp. 123–178, John Wiley & Sons Inc., New York.
Lonsdale, H.K. (1982) The growth of membrane technology. Journal of Membrane Technology, 10, 81–181.
Paulson, D.J., Wilson, R.L. & Spatz, D.D. (1984) Cross-flow membrane technology and its applications. Food Technology, 38, 77–87, 111.
Pusch, W. (1990) Performance of RO membranes in correlation with membrane structure, transport mechanisms of matter and module design (fouling)—state of the art. Desalination, 77, 35–54.
Strathmann, H. (1981) Membrane separation processes. Journal of Membrane Science, 9, 121–189.
Chapter 2
Principles of Membrane Filtration
A. Hausmann, M.C. Duke and T. Demmer
2.1 Introduction and definitions
This chapter provides a fundamental understanding of pressure-driven membrane operations, such as microfiltration (MF), ultrafiltration (UF), nanofiltration (NF) and reverse osmosis (RO), and the way these are used in the food industry. The three systems determining membrane performance are the membrane itself, the feed system and the operating conditions. All three influence each other and interact, resulting in a complex system of membrane operation. Each of these three systems will be discussed separately as well as pointing towards influencing correlations. Some novel membrane processes driven by other means than pressure will also be discussed briefly.
A membrane is a semi-permeable barrier, which means that at least one component of a fluid in contact with the membrane can pass through. The membrane enables selective separation of the compounds in a feed-stream, which is split into retentate (concentrate) and permeate (filtrate). The retentate contains the compounds that cannot pass through the membrane, the permeate is the product that passes through (Melin & Rautenbach, 2007).
2.1.1 Membrane processes
The many and various applications of membrane processes can be categorised into product concentration with the focus on the retentate, such as the pre-concentration of milk before cheese making, and solute removal for purification/clarification purposes as, for example, product sterilisation, demineralisation or desalination focusing on the permeate quality. Also possible are solute separation applications, where both permeate and retentate are considered valuable, as in fractionation of constituents: an example is the separation of milk proteins into a casein-enriched retentate and whey protein containing permeate, which is often referred to as ‘ideal whey’ (Lawrence et al., 2008). Membrane processes can be driven by differences in applied external pressure, osmotic pressure, vapour pressure and electric fields which are used for electrodialysis. This electro-membrane process works through an electric field introduced between the two membrane sides that allows for selective separation of ions in solution. The transport of ions across the membrane results in ion depletion on one side and ion concentration on the other. Electrodialysis can be utilised for several types of separation and concentration of salts, acids and bases from aqueous solutions or for separation and concentration of monovalent ions from multiple charged components as well as the separation of compounds from uncharged molecules. Desalination is commonly carried out by electrodialysis. In the food industry it is, for example, used for desalination of cheese whey for infant formulas, fractionation of single whey proteins or demineralisation of sugar cane juice (Kariduraganavar et al., 2006; Pouliot, 2008).
However, most membrane processes employed in the food industry to date are commonly driven by the application of pressure to the feed, while permeate is withdrawn from the system. As outlined in Fig. 2.1, the four pressure-driven membrane processes (MF, UF, NF and RO) differ by their average pore size, and can fractionate all major liquid food ingredients.
Figure 2.1 Separation by the four pressure driven membrane processes. (Adapted from Koros et al. (1996) for pore sizes by International Union of Pure and Applied Chemistry (IUPAC) definitions, and Ayral (2008) for pressure ranges.)
Based on the pore size differences, MF membranes retain suspended solids like bacteria or fat droplets. UF membranes are able to reject colloids, macromolecules, viruses and proteins. NF membranes can fractionate inorganic salts or diluted organic molecules, whereas RO membranes are mainly used for water desalination and the production of ultrapure water (Ayral, 2008).
2.1.2 Definitions of membrane processes
Flow regimes
There are two modes of operation for membrane processes, dead-end filtration and the cross-flow mode, with the ideal concepts shown in Fig. 2.2. In the case of dead-end filtration, the feed is pumped perpendicularly onto the membrane surface, whereas during cross-flow filtration it is pumped across the membrane surface tangentially rather than perpendicularly. Dead-end operation has to be run batch-wise to relieve the concentrated species, whereas the cross-flow mode allows for continuous process operation. For feed containing relatively high solids, cross-flow is advantageous in that it limits the build-up of solids on the membrane surface. Today the cross-flow mode is the standard operation for most filtration processes in the food industry (James et al., 2003).
Figure 2.2 Schematic of a dead-end filtration (left) and cross-flow filtration (right).
Membrane geometry
Membranes are available in flat sheet, hollow fibres or tubular geometries. Flat sheets are simply made with some kind of support (e.g. a porous backing or imbedded mesh) which supports the active (selective) layer. Hollow fibre and tubular membranes are similar in geometry as both appear tube-like. However, the difference comes from the tube diameters. Hollow fibre membranes have much smaller diameters than tubular membranes and, in general, the same material that forms the active layer also makes up the supporting layer. Both membranes are available as ‘inside-out’ membranes where the inner wall provides the active membrane layer or ‘outside-in’ where the outer wall acts as the membrane (Fig. 2.3). The remaining material in the membrane is more porous and, thus, acts merely to mechanically support the active layer and ideally offers no resistance to permeation.
Figure 2.3 Hollow fibre membranes for inside-out filtration with active layer on hollow fibre lumen (left) and for outside-in filtration with active layer on the outside (right).
Membrane module
For continuous membrane operations, membranes are packed into distinct units or membrane elements, known as the membrane module. The term ‘module’ stems from the modular design of membrane plants where identical membrane units are arranged and connected by pipelines. There are various module designs that meet the requirements of different membrane types. Flat sheet membranes can be arranged in spiral wound modules or plate on frame configuration as discussed in Chapter 3. All membrane module types consist of at least one entry for the feed to be separated and two exits for the filtered and retained components, as shown in Fig. 2.4. The area of membrane contained in these basic membrane units or modules is in the range of 1 up to 800 m2.
Figure 2.4 Schematic membrane module (top) and an example of a hollow fibre module (bottom).
Transmembrane pressure and flux
As mentioned above, typically the driving force for all four membrane processes (MF, UF, NF and RO) is the pressure across the membrane from feed to permeate side. For dead-end filtration this transmembrane pressure (TMP) is simply the pressure of the feed solution Pf minus the pressure on the permeate side of the membrane Pp:
2.1
In cross-flow mode, as a consequence of flow, the pressure on the feed/retentate side declines while passing along the module from its inlet value Pf to the outlet pressure Pr (Piry et al., 2008).
2.2
This pressure drop (ΔP) along the module is considered to be a source of energy loss in membrane systems, and the amount depends on factors influencing flow resistance, such as the cross-flow velocity, flow spacer design and channel width. In many module configurations, the filtrate-sided pressure Pp of the module is almost constant due to the locally marginal flow velocity. To realise filtration over the entire length of the module, the permeate-sided pressure Pp needs to be below Pr which is the lowest pressure on the feed channel side, as shown in Fig. 2.5.
Figure 2.5 Pressure relations along a membrane module in a cross-flow configuration.
The TMP in cross-flow set-ups is, therefore, calculated as the average pressure applied over the length of the membrane module:
2.3
Performance of a membrane process is assessed by measuring the filtrate throughput per unit membrane area and time that passes through the membrane; this flux is measured in or more commonly: . or as mostly seen in product sheets and other membrane media expressed simply as litre per square meter per hour (LMH) or (L m−2 h−1).
Wall shear stress
The tangential force induced by the fluid flowing along the membrane in cross-flow mode is represented by the wall shear stress (τW). An approximation that allows for experimental determination was defined by Le Berre & Daufin (1996):
2.4
where dhydraulic is the hydraulic diameter and L the membrane length.
Volume concentration ratio (VCR) and concentration factor (CF)
A measure for the volume reduction of the initial feed volume is the volume concentration ratio (VCR). It is defined as the quotient between initial feed volume Vf and retentate volume Vr (Cheryan, 1998).
2.5
For concentration applications, the concentration factor (CF) is commonly used to express degree of concentration of a target compound. It is defined as the ratio of the concentration of a component i in the retentate (ci,retentate) to the concentration of the same component in the feed (ci,feed) (Koros et al., 1996).
2.6
If the target compound is not retained completely but some gets lost into the permeate, the VCR needs to be increased above the CF to obtain the desired concentration. The relation between the retention factor, CF and VCR will be explained next.
Retention factor
The retention factor (Rf) of the various compounds in the feed stream reflects the membranes selectivity and can be calculated for each solute compound as the ratio between the concentration of that solute i in the bulk of the retentate (ci,retenate) and its concentration in the permeate (ci,permeate):
2.7
Retention is a measure of the membrane's ability to separate components of the feed solution. How the membrane selects which component diffuses varies on the scale of separation (i.e. bulk properties down to molecular properties). The most fundamental membrane property for separation is the pore size as determined by the structure of the membrane material. Other properties include material surface chemistry and charge. These will be discussed in more detail later.
Another way to express the Rf is given by Cheryan (1998):
2.8
2.9
This equation enables to calculate retention using only retentate data, as opposed to equation (2.7), which is based on both permeate and retentate data.
A disadvantage of Equations (2.5) to (2.7) is that the behaviour of dissolved components is not correctly reflected as their concentration changes with the water content of the system. A better way is given by
2.10
Diafiltration
A technique to improve separation between compounds with different retention factors is diafiltration (DF). It increases separation quality through the addition of a solvent (often water) to the feed in order to dilute the concentration of permeating species. For substances that are freely permeating through the membrane and not retained, the concentration in the final permeate equals its concentration in the retentate (the normal process is shown in Fig. 2.6). During diafiltration, the initial feed stream is diluted, so that more dissolved particles small enough to pass the membrane are ‘washed out’ through the membrane. This process results in a lower concentration of permeating compounds in the retentate while the concentration of fully retained particles remains unaffected, as schematically shown on the right-hand side of Fig. 2.6. Therefore, a more distinct separation can be achieved; however, dilution of the feed means a bigger volume needs to be processed in order to achieve the same degree of concentration, which results in higher energy demands. The DF is a measure of the extent of washing that has been performed. It is based on the volume of diafiltration medium Vd fed to the system relative to the volume of retentate produced Vr:
2.11
Figure 2.6 Principle of a diafiltration: normal filtration process (left) and feed stream diluted (right).
A common application is the dealcoholisation of wine, where diafiltration is often employed as alcohol is a freely permeating solvent and can be reduced by such solvent dilution and removal.
Table 2.1 Relation between the diafiltration factor and resulting protein purity and lactose level during ultrafiltration of whey.
2.2 Membrane properties based on materials
The variety of commercially available membranes has been largely extended in the last two decades, and overall thermal, chemical and mechanical strength has been improved. Such variety results from application-dependent interactions between feed components and membrane materials, so that membrane materials have been specifically optimised for each application (Melin & Rautenbach, 2007). This also means that laboratory and pilot plant testing is necessary for every new application. In the following, membrane properties will be outlined with considerations regarding membrane-feed systems.
2.2.1 Membrane structure
From a structural point of view membranes can broadly be divided into symmetric (or isotropic) and asymmetric (or anisotropic) membranes. As illustrated by Fig. 2.7, symmetric membranes have a similar structural composition and morphology at all positions within, and can be porous or dense. Asymmetric membranes, as shown in Fig. 2.7, are composed of two or more structural planes of non-identical morphologies (Koros et al
