membrane technology a practical guide to membrane technology and applications in food and bioprocessing

Thus the effi-ciency and productivity of separation technology can be critical to both the profitabil-ity and sustainability of an enterprise.bio-Recent developments in membrane technolo

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

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.

In using such information or methods they should be mindful of their own safety and the safety

of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for the injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

Library of Congress Cataloging-in-Publication Data Control Number

A catalog record for this book is available from the Library of Congress

ISBN: 978-1-85617-632-3

For information on all Butterworth-Heinemann publications

visit our website at elsevierdirect.com

Typeset by MPS Limited, a Macmillan Company, Chennai, India

www.macmillansolutions.com

Printed and bound in the United States of America

10 11 12 13 10 9 8 7 6 5 4 3 2 1

Separation and purification are key to many manufacturing processes in the food, processing and pharmaceutical industries Furthermore, these processes are becomingincreasingly important for the growth of industries such as bioenergy, biotechnologyand nanotechnology The multiple steps involved in separation/purification processingoften require substantial up-front capital and significant operating costs Thus the effi-ciency and productivity of separation technology can be critical to both the profitabil-ity and sustainability of an enterprise.

bio-Recent developments in membrane technology, especially the development ofnovel membrane processes, are the most growing areas of process technology Thereare several excellent books published in the general areas of membrane science andmembrane technology, but there is minimal focus on the applications of membranetechnology in food and pharma areas, even though they have grown steadily duringthe 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 reducethe 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 abook 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 Elsevierconducted 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 wouldnot have materialized It is most fitting that this technological work is published fromcontributors around the globe and is founded on the spirit of free enquiry coupledwith hard work and imagination It has indeed been a great pleasure to be in touchwith all contributors during the last 3 to 4 years Thanks also for their patience andunderstanding

We would be utterly remiss if we did not acknowledge those people who haveprovided us with the inspiration, motivation and never-ending encouragementthroughout the course of this work

Dr Murali would like to acknowledge Mr Ronald Christenson, former ChiefTechnology Officer, Cargill Inc; his wife Ponnamma; his children Shubha andShilesh, 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 hisparents who would have been both excited and extremely proud to see this bookpublished

xi

Dr Cui wishes to thank his wife, Dr Jing Yu, for her unreserved support over ades, his mother for her love, and his children, Jenny and Michael, for “keeping himhappy” He would like to thank those who inspired and encouraged him to embark on

dec-a cdec-areer in the “membrdec-ane world” Such dec-a list would be long, dec-and here just ndec-ame dec-afew: Tony Fane, John Howell, Norman Li and Bill Eykamp

In setting our goal of bringing this book to fruition, we kept in mind these wordsfrom Robert Frost in The Road Not Taken:

“Two roads diverged in a wood and

I-I took the one less travelled by,

And that has made all the difference”

ZFC and HSM

Zhanfeng Cui is the Donald Pollock Professor of Chemical Engineering, University

of Oxford since the Chair was established in 2000 He is the founding Director of theOxford Centre for Tissue Engineering and Bioprocessing (OCTEB)

He was educated in China and got his BSc from Inner Mongolia University

of Technology (1982) and MSc (1984) and PhD (1987) from Dalian University ofTechnology After a postdoctoral experience in Strathclyde University in Scotland, hejoined Edinburgh University as a Lecturer in Chemical Engineering (1991) He thenheld academic appointments at Oxford Engineering Science Department as UniversityLecturer (1994 1998) and Reader (1999 2000) He was a Visiting Professor ofGeorgia Institute of Technology, USA (1999), the Brown Intertec Visiting Professor

to University of Minnesota, USA (2004), and a Chang-Jiang Visiting Professor toDalian University of Technology, China (2005) He is a Chartered Engineer, aChartered Scientist, and a Fellow of the Institution of Chemical Engineers In 2009

he was award a Doctor of Science (DSc) by Oxford University to recognise hisresearch achievement in membrane science and technology Apart from membraneresearch, he also works on tissue engineering and stem cell technologies, and biopro-cessing He published widely and is also the academic founder of Zyoxel Limited, anOxford University spin-off in 2009

H.S Muralidhara Ph.D is a Chemical Engineer He retired from Cargill Inc inMinneapolis, 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 ResearchLeader at Battelle Memorial Institute in Columbus, Ohio, USA for 10 years He iscurrently an industrial consultant

He has over 30 years of industrial experience in separation purification processtechnologies including application of membranes in food and bioprocessing He is acoinventor of 27 US patents and 15 patents pending He has edited two books andhas served as a key note speaker in many major international conferences

xiii

Vicki Chen is the director of the UNESCO Centre for Membrane Science andTechnology at The University of New South Wales, Australia.

Jim Davies is a biochemical engineer with a PhD from UCL and is currently aPrincipal Group Leader in Purification Development at Lonza Biologics UK.R.W Field is a Reader in Engineering Science at Oxford University and has manyyears 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 wastewatertreatment, water reuse, and membrane technologies, including desalination

Yu Jiang was awarded her DPhil in 2009 from Oxford University Her DPhil thesiswas entitled “Emergency drinking water device based on gravity drivenultrafiltration”

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 separationtechnologies and food processing

N.S Krishna Kumar, PhD, is a Senior Chemical Engineer at Cargill Inc.,Minneapolis, Minnesota, USA His research interest is focused on application ofmembranes in food, bioprocessing and water treatment He is currently involvedwith vegetable oil process development research work concentrated towards pro-cess improvement

Abhay R Ladhe is a Research Chemical Engineer at Cargill Inc., Minneapolis,Minnesota, USA He graduated from University of Kentucky with a Ph.D inChemical Engineering His research interests include membrane functionalizationand membrane based separation He is currently working in the area ofvegetable oil processing for process improvement and process development.Hongyu Li is a Research Fellow in the UNESCO Centre for Membrane Science andTechnology, The University of New South Wales, Australia

Frank Lipnizki, PhD, is currently with Alfa Laval – Business Centre Membranes(previously Danish Separation Systems), Denmark Since 2010 also Docent at theUniversity of Lund, Sweden His main research interests are the integration andoptimization of membrane process for the food, biotech and process industry.Jianquan Luo, MEng, Institute of Process Engineering, Chinese Academy ofSciences Mr Luo is an experienced engineer specializing in membrane applica-tions in bioseparation and wastewater treatment

xv

Susana Luque is a Full Professor in Chemical Engineering at the University ofOviedo in Spain Her main interests are in applied membrane research and mem-brane-based hybrid processes.

Alan Merry has a PhD in Chemical Engineering He has 30 years’ experience atITT-PCI Membranes Ltd, working with all aspects of membranes including mem-brane development, applications and production

Joseph Scimeca, PhD, currently holds the position of Director of Global RegulatoryAffairs, in the department of Corporate Food Safety and Regulatory Affairs, atCargill, Incorporated, where he has responsibility for ensuring that the company’sfood and feed products and processes are safe, including being protected againstintentional acts of adulteration and bioterrorism, and are compliant with theappropriate food/feed regulations

Martin Smith, also a Biochemical Engineer from UCL, is currently a Bio-ProcessConsultant at eXmoor Pharma Concepts in the fields of Biopharmaceuticals andRegenerative Medicine

Yinhua Wan, DPhil, is a Professor of Biochemical Engineering at Institute ofProcess Engineering, Chinese Academy of Sciences He has published more than

80 papers in refereed journals and holds a number of patents in membrane tion technologies

separa-Fundamentals of

Pressure-Driven Membrane

Separation Processes

Z.F Cui, Y Jiang and R.W Field

Department of Engineering Science, Oxford University, Oxford, UK

of Permeate Flux1.5.1 Flux Prediction Models1.5.2 Flux Enhancement andFouling Control1.6 Summary

Further Readings

1.1 INTRODUCTION

Membrane processes are one of the fastest growing and fascinating fields inseparation technology Even though membrane processes are a relativelynew type of separation technology, several membrane processes, particularlypressure-driven membrane processes including reverse osmosis (RO), nano-filtration (NF), ultrafiltration (UF), and microfiltration (MF), are alreadyapplied on an industrial scale to food and bioproduct processing

The concept of membrane processes is relatively simple but neverthelessoften unknown Membranes (lat.: membrana5 thin skin) might be described

as conventional filters (like a coffee filter) but with much finer mesh or muchsmaller pores to enable the separation of tiny particles, even molecules! In

1Membrane Technology DOI: 10.1016/B978-1-85617-632-3.00001-X

© 2010 Elsevier Ltd All rights reserved.

general, one can divide membranes into two groups: porous and nonporous.The former group is similar to classical filtration with pressure as the drivingforce; the separation of a mixture is achieved by the rejection of at least onecomponent by the membrane and passing of the other components through themembrane (see Fig 1.1) However, it is important to note that nonporousmembranes do not operate on a size exclusion mechanism It should bepointed out that this chapter focuses on pressure-driven membrane processesusing porous membranes for its close relevance to food and bioproductprocessing.

Membrane separation processes can be used for a wide range of tions and can often offer significant advantages over conventional separationsuch as distillation and adsorption since the separation is based on a physicalmechanism 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 theprocessing of food, beverage, and bioproducts where the processed productscan be sensitive to temperature (vs distillation) and solvents (vs extraction)

applica-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

1
10 bar), NF (100
500 Da, 0.5
10 nm, 10
30 bar), and RO (,0.5 nm,

35
100 bar).Figure 1.2presents a classification on the applicability of ent membrane separation processes based on particle or molecular sizes ROprocess is often used for desalination and pure water production, but it is the

differ-UF and MF that are widely used in food and bioprocessing

While MF membranes target on the microorganism removal, and henceare given the absolute rating, namely, the diameter of the largest pore on themembrane surface, UF/NF membranes are characterized by the nominal rat-ing due to their early applications of purifying biological solutions The nom-inal rating is defined as the molecular weight cut-off (MWCO) that is thesmallest molecular weight of species, of which the membrane has more than90% rejection (see later for definitions) The separation mechanism inMF/UF/NF is mainly the size exclusion, which is indicated in the nominalratings of the membranes The other separation mechanism includes the elec-trostatic interactions between solutes and membranes, which depends on thesurface and physiochemical properties of solutes and membranes.

1.2.2 Definitions

In contrast to Figure 1.1, real membrane separations split the feed mixturestream into two streams with different compositions as shown inFigure 1.3.The feed stream _mF to a membrane module is split into (i) the retentatestream _m , which is the stream that has been retained by the membrane

Micron particle

Fine particle

Coarse particle

containing both the material that has been rejected by the membrane and aquantity of material that would not be rejected by the membrane but has yetnot been given the opportunity to pass through the membrane; and (ii) thepermeate stream _mP, the stream that has passed through the membrane, con-taining much less or no bigger molecules or particles than the pores.

Like any separation processes, the membrane separation processes can beevaluated by two important parameters, efficiency and productivity The pro-ductivity is characterized by the parameter permeate flux, which indicatesthe rate of mass transport across the membrane In general terms, the localmass transport of a component i through a membrane element is related toits concentration on the feed side CRi and the permeate side CPi (see

Fig 1.3) The flow of a component i through a membrane element can bereferred to as its flux Ji This flux is a velocity and is commonly expressed in[kg/(m2s)] or [kmol/(m2s)] When n components are permeating through themembrane a total flux Jtotcan be defined as:

dif-Fig 1.3) The dead-end mode is employed mostly in MF for clarification andsterilization, where the feed is relatively clean In most applications, the accu-mulation of the rejected particles or molecules is so severe that dead-end oper-ation becomes impractical and cross-flow operation has to be adopted Thetangential flow in the cross-flow mode can help to shear away the accumu-lated rejected species at the membranes, limit the heights of cake layers, andhence maintain the permeate flux The schematic diagrams of the dead-endmode and the cross-flow mode, and their effects on the permeate flux and theheight and resistance of the cake layer, are shown inFigure 1.4

In most applications dealing with aqueous solutions in food and duct processing, the solvent permeate is largely water and permeate flux isoften conveniently presented as [m3/(m2s), i.e., m/s] or [L/(m2h), LMH],

biopro-which is the volume of mp produced per unit of membrane area per unittime Usually there is only one species, microparticle or macromolecule, to

be interested, and the rejection will only be referred to the concerned cies Often the permeate flow rate is much less than the retentate flow rate in

spe-a single pspe-ass, hence the chspe-ange of concentrspe-ation in the retentspe-ate is not icant The rejection can then be conveniently calculated by:

signif-R5 1 2CP

where CFis the feed concentration

The driving force in pressure-driven membrane separation is of course thepressure, or the pressure difference between the upstream and the downstream

of the membrane, or between the feed and the permeate This is referred to astransmembrane pressure As the pressure may vary in the membrane moduledue to crossflow, an averaged pressure difference over the module is used:

Feed

Retentate Feed

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
resistance as refereed later).

bodies The pores are of uniform size (isotropic) or nonuniform size pic) Microporous membranes are designed to reject all the species above theirratings However, they tend to be blocked by the species that are of similarsizes as the pores The asymmetric membrane has a selective skin layer on thetop of its membrane body The membrane body is usually void, only givingmechanical support to the selective skin layer Compared to the microporousmembranes, the asymmetric membranes rarely get blocked Most UF, NF, and

(anisotro-RO membranes are of asymmetric structure, while most polymeric MF branes are of microporous structure

mem-1.3.2 Membrane Materials

In terms of materials, membranes can be classified into polymeric or organicmembranes and ceramic or inorganic membranes Organic membranes are usu-ally made up of various polymers, among which the typical ones are celluloseacetate (CA), polyamide (PA), polysulfone (PS), polyethersulfone (PES), poly-vinylidene fluoride (PVDF), polypropylene (PP), etc Polymeric membranes arerelatively cheap, easy to manufacture, available in a wide range of pore sizes,and they have been widely used in various industries Nevertheless, most of thepolymeric membranes have limitations on one or more operating conditions(either pH, or temperature, or pressure, or chlorine tolerance, etc.), which hindertheir wider applications For example, CA is the classic material usually used toproduce the skinned membranes However, it has many disadvantages such aslow temperature limit (30
40C), narrow pH range (2
8, preferably 2
6),and low chlorine tolerance (under 1 mg/L free chlorine)

Inorganic membranes have been commercialized since the early 1980s.Due to their obvious advantages of high mechanical strength, and chemicaland thermal stability over the conventional polymeric membranes, they haveextended the application of membrane technology into many new areas.Inorganic membranes (such asγ-alumina/α-alumina, borosilicate glass, pyro-lyzed carbon, zirconia/stainless steel, or zirconia/carbon) have strong toler-ance to even extreme operating conditions For instance, they have wide limits

of temperature, pH, and pressure, and have extended lifetime However, ganic membranes are very brittle, so the membranes can be easily damaged bydropping or unduly vibrating Additionally the availability of such membranes

inor-is only limited to mostly UF membranes and MF membranes today In tion, cost is the biggest disadvantage in the applications of inorganic mem-branes They are far more expensive than polymeric membranes

addi-1.3.3 Membrane Modules

Membrane module is the way the membrane is arranged into devices andhardware to separate the feed stream into permeate and retentate streams Sofar, there are four kinds of membrane modules that have been widely used in

industry They are (1) tubular modules, (2) hollow fiber modules, (3) flatsheet modules, and (4) spiral-wound modules These membrane modules aredesigned and developed by industry manufacturers in order to achieve differ-ent characteristics on the hydrodynamic conditions, filtration areas, energyconsumptions, etc.

Tubular modules are composed of a number of membrane tubes assembled

in a shell-and-tube arrangement The membrane tubes are usually made up ofporous fabric or plastic support with selective membranes on the inside Theinternal diameters of the tubes generally range from 5 to 25 mm, and the tubelengths are in the range of 0.6 to 6 m Tubular modules have some importantcharacteristics: (1) due to their large internal diameters, tubular modules arecapable 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 areusually operated under the turbulent flow conditions with the Reynolds num-bers greater than 10,000; (3) they have the lowest surface area-to-volume ratioamong all the four membrane configurations The holdup volumes of tubularmodules are also high, which need large floor space to operate

Hollow fiber modules are actually the “thin” tubular membranes in pact modules, but in the form of self-support that enables them to withstandhigh backpressure Normally, hollow fiber modules are composed of

com-50
3000 individual hollow fibers, bundled and sealed together on each endwith 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 asthin as 0.04 mm and can withstand much higher pressure) The fiber lengthsrange 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 ofthe fibers, while in RO, they are operated in the outside-in mode withselective skin layers on both sides of the fibers Hollow fiber modules havesome very different characteristics from tubular modules: (1) they arerecommended to operate with the Reynolds numbers in the range of

500
3000, therefore, most of them are run in the laminar flow region.Additionally, the pressure rating of hollow fiber modules is low, normallywith a maximum of 2.5 bar; (2) due to the combination of low cross flowrate and low pressure drop, hollow fiber modules are one of the more eco-nomical modules in terms of energy consumption; (3) hollow fiber moduleshave the highest surface area-to-volume ratio among all the four membraneconfigurations, and their holdup volumes are low; (4) because the fibers areself-supported, hollow fiber modules have good backwash capacity and arehence easy to clean; and (5) one distinct disadvantage of hollow fiber mod-ules is that, their thin fibers are susceptible to get blocked by the feed withlarge particles, when they are operated in the inside-out mode Therefore,the pretreatment to reduce particle size to 100μm is usually required forhollow fiber modules

Flat sheet modules comprise a selective flat sheet membrane on the topand a flat plate at the bottom, between which a net-like material is placed toprovide space for the permeate removal, and on the other side of the flat plate,another sheet membrane and another net-like material are placed in mirror toform a sandwich-like module Flat sheet modules have channel gaps rangingfrom 0.5 to 10 mm and are of lengths ranging from 10 to 60 cm The superfi-cial Reynolds numbers for flat sheet modules are in the laminar flow region;however, good mixing can be achieved when a screen is placed in the feedchannel The pretreatment to 150μm is recommended for flat sheet modules.With regard to packing density, energy consumption, and cost, flat sheet mod-ules lie in between tubular modules and spiral-wound modules.

The design of spiral-wound modules is similar to that of flat sheet ules In the spiral-wound modules, two membrane sheets are separated by amesh-like spacer with the active membrane sides facing away Three edges

mod-of the two membrane sheets are glued together with the fourth edge open

to a perforated center tube for the permeate removal On the other twosides of “the envelope,” another two mesh-like spacers with thicknesses inthe range of 0.56
3 mm are placed as the feed channel spacers The wholeassembly is rolled around the perforated center tube in a spiral configura-tion The characteristics of spiral-wound modules are as follows: (1) spiral-wound modules are operated in the turbulent flow region because ofthe presence of feed spacers; (2) due to the additional drag generated byfeed spacers, the pressure drop in spiral-wound modules is relatively high;(3) spiral-wound modules have fairly high surface area
volume ratio andare the lowest in terms of capital cost, among all the four kinds of mem-brane modules; and (4) suspended particles can easily block the mesh-likespacers and then partially block the feed channel Therefore, spiral-woundmodules require relatively clean feed that are with minimum content of sus-pended particles The pretreatment to reduce suspended particles is neededfor spiral-wound modules

1.4 OPERATION

1.4.1 Concentration Polarization

Concentration polarization refers to the reversible accumulation of rejectedmolecules close to the membrane surface In membrane processes all compo-nents in the feed are transported to the membrane surface by convection, andthe rate increases as the permeation through the membrane increases Theselectivity of the membrane holds back the less permeable components Atsteady state, these less permeable components have to be transported backinto the bulk of the feed stream As the flow next to the membrane surface islaminar, this transport can only be diffusive The transport has to be based onthe established concentration gradient, i.e., an enrichment of the less

permeable components at membrane surface, as shown inFigure 1.5 It is anatural consequence of membrane selectivity and is equivalent to the masstransfer boundary If driving force is removed, permeation ceases, and such aconcentration polarization phenomenon disappears.

Under steady-state conditions, the following relationships describe therelevant fluxes based onFigure 1.5:

G the process is steady state,

G the diffusion is described by Frick’s law,

G there is no chemical reaction,

G the concentration gradient parallel to membrane can be neglected,

G the density is constant, and

G the coefficient is independent from the solute concentration

Integration of Equation (1.6) taking the following boundary conditionsinto account,

J5 kUln CM

CB

ð1:9ÞAccording to Equation (1.9), a higher mass transfer coefficient, k, and ahigher membrane surface concentration lead to a higher permeate flux.Equation (1.9)helps to explain the commonly observed UF behaviors asshown in Figure 1.6 In the pressure control region, the increase in TMPincreases flux, leading to a higher CM But if CMreaches a certain value of

Higher flow rate

Higher temperature

Lower concentration

Mass transfer controlled region

Transmembrane pressure

Water

Pressure controlled region

FIGURE 1.6 The influences of operating parameters on permeate flux, showing the pressure control region and the mass transfer control region.

macromolecule’s gelation concentration or the solubility of the rejected salt,gelation or salt precipitation occurs and CM reaches its maximum value.Further increase in TMP does not have any effect on CM, 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 ing cross-flow velocity leads to a higher permeate flux, as indicated inEquation(1.9)

increas-The mass transfer coefficient can be estimated on the basis of heat andmass transfer analogy (so-called Colburn analogy) using the semiempiricalSherwood correlation This correlation can be written as:

Laminar 1 1.62 0.33 0.33 0.33 Hollow fiber

Turbulent 2 0.04 0.75 0.33
Tubular

Laminar 3 1.615 0.33 0.33 0.33 Flat sheet

Turbulent4 0.026 0.8 0.3
Flat sheet, tubular

1 Re , 2100 hydrodynamic fully developed profile, not fully developed concentration boundary layer.

2 Re , 10,000.

3 Re , 2300 hydrodynamic fully developed profile, not fully developed concentration boundary layer.

4 Re 2300.

TABLE 1.2 Dimensionless Numbers

Reynolds number: Re 5ρudh

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 solution; μ, viscosity of the feed solution; ν, kinematic viscosity of the feed solution.

1.4.2 Membrane Fouling

Fouling is generally defined as a process resulting in a loss of performance

of a membrane due to deposition of suspended or dissolved substances ontoits external surface Fouling cannot be removed simply by stopping the filtra-tion process Fouling is often the main limitation to the successful membraneapplication of food and biotech industries

Fouling can be seen as a reduction in the active area of the membraneand leads therefore to a reduction in flux below theoretical capacity of themembrane Several parameters influence the fouling rate such as:

G nature and concentration of solutes and solvents,

G pore size distribution,

G surface characteristics and material of membranes, and

G hydrodynamics of membrane module

Fouling can be related to different modes such as adsorption, chemicalinteractions, cake formation, and pore blocking by particles These modescan lead to blockage or partial blockage of the active membrane area or todeposition of a layer onto the membrane surface In Table 1.3, examples offoulants in membrane processes are given

TABLE 1.3 Examples of Foulants and Fouling Modes in Membrane

Processes

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.

Generally speaking, four fouling mechanisms for porous membranes can

be observed, as shown inFigure 1.7:

(a) complete pore blocking,

(b) internal pore blocking,

(c) partial pore blocking, and

(d) cake filtration

The following differential equation can be used to describe the influence

of fouling on the flux through the membrane in the absence of any flow effect

cross-J5 J0U½1 1 KUð2 2 nÞðAUJ0Þðn 2 2ÞUtð2 2 nÞ ð1:11Þ

In this equation, the phenomenological coefficients n and K depend onthe fouling mechanism InTable 1.4, the different values of n, their phenom-enological background, their effect on the mass transport, and the relevanttransport equations are given

1.5 PREDICTION AND ENHANCEMENT OF PERMEATE FLUX 1.5.1 Flux Prediction Models

Numerous different models have been developed to predict the permeate flux

as a function of operating parameters, membrane properties, and feed ties in UF However, due to the limitations of application conditions andmodel assumptions, or not enough understanding of the phenomena that takeplace around the membrane surface, no model so far is universally applicable

proper-or fully satisfactproper-ory Nevertheless, these models can help to understand theoperation and performance links

The pore model is applied to predict the permeate flux in the pressurecontrol region, under the conditions of no fouling and negligible concentra-tion polarization In this model, it can be assumed that (1) membrane pores

(A) Complete pore blocking (B) Internal pore blocking

FIGURE 1.7 Fouling mechanisms of porous membranes.

TABLE 1.4 Fouling Mechanisms, Phenomenological Background, Effect on Mass Transport, and Transport Equations

Fouling Mechanism n Phenomenological

Reduction of the active membrane area Depending on feed velocity, permeate might

be increased by increasing transmembrane driving force (pressure).

Increase in membrane resistance due to pore size reduction Internal pore blocking is independent from feed velocity No limiting might

Reduction of active membrane area The effect is similar to pore blocking but not so severe. J5 J0U½1 1 KiUðAUJ0ÞUt2 1 ð1:14Þ

Cake filtration (see

Fig 1.7d )

0 Formation 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.

are ideal cylindrical channels and are uniformly distributed on the membranesurface; (2) the permeate passing through membrane pores are laminar flow

in the steady state; and (3) the applied feed is of constant density (e.g.,incompressible) and Newtonian (with no dependence on the shear rate) Thepermeate flux can be calculated using the Hagen
Poiseuille equation, which

is based on the momentum balance:

The equation shows that the permeate flux is directly proportional to theTMP and inversely proportional to the viscosity The viscosity is primarilycontrolled by the solvent type, feed composition, and temperature Therefore,

in the pressure control region, increasing the temperature and pressure, anddecreasing the feed concentration can increase the permeate flux

The resistance model is developed to express the entire TMP-flux 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 andthe ideal feed that lead to no fouling, the model can be expressed as:

behav-Jv5 μRΔP

where Jvis the permeate flux,ΔP the TMP, μ the viscosity of the permeate,and Rmthe hydraulic resistance of the membrane, which is a constant value foreach membrane and can be determined by measuring the pure water flux

By consideration of the effect of concentration polarization, the increase

in solute concentrations near the membrane surface results in the increase inosmosis pressure, which effectively reduces the TMP The permeate flux isthen calculated by:

In the filtration of real feeds, both concentration polarization and brane fouling occur to add additional resistances to the membrane and hence

mem-to the permeate mem-to pass through Therefore, the resistance of the polarized

layer Rpand the fouling resistance caused by the physiochemical interactionsbetween the solutes and the membrane Rfneed to be taken into account Theequation is expressed as follows:

Jv5μðRΔP 2 Δπ

In MF, the rejected particle exerts insignificant osmotic pressure, and theconcentration layer resistance and fouling layer resistance can be combinedinto one parameter
the cake resistance

1.5.2 Flux Enhancement and Fouling Control

Flux enhancement can mean one of the following:

Increase permeate flux

Increase flux with same energy consumption

Decrease energy consumption while maintaining same flux

Fouling control is an important issue in order to maintain anacceptable flux level In addition, chemical cleaning represents a significantfraction of operational cost of membrane process and downtime too It isgenerally accepted that fouling is a fact of life in practical operation of mem-brane processes, but fouling can be controlled to an acceptable level withbetter understanding of the process operation and the feed characterization

To control fouling, different approaches have been developed and usedseparately or in combination:

(i) hydrodynamic management,

(ii) back flushing and pulsing,

(iii) membrane surface modification,

(iv) feed pretreatment,

(v) flux control, and

(vi) effective membrane cleaning

Hydrodynamic management aims at promoting local mixing close to themembrane surface and enhancing the back diffusion of the rejected mole-cules or particles In doing so, the concentration at the membrane surface isreduced, also leading to reduced adsorption of the molecules (adsorption iso-therm) Simply increasing cross-flow velocity leads to limited effect butmuch higher energy consumption Instead much effort has been directed toincrease local turbulence or mixing close to the membranes using variousmethods to introduce flow instability and secondary flow These include tur-bulence promoters, corrugated membrane surface, various spacers, vibrating/rotational membranes, reverse or pulsatile flow, use of sponge balls, gas bub-bles, etc It has been demonstrated that these techniques generate secondaryflows which improves local mixing and therefore improves mass transfer

Back flushing or pulsing is an approach to remove cake layers on thefeed side and, therefore, reduce the influence of fouling It is carried out byreversing the flow of the permeate through the membrane and, therefore, dis-lodges the foulant and reestablishes the flux at a high level (Fig 1.8) Inorder to maintain a high flux, back flushing is carried out periodically andrequires module types with a high-pressure resistance, e.g., pressurestable capillary modules Such back flushing can be carried out with ratherhigh frequency in a very short period of time, which may be termed as backpulsing Back flushing or pulsing is widely used in MF of high solid contentfeed or feed with high fouling tendency.

Membrane surface modification and feed pretreatment both act to alterthe interactions between the filtered molecules or microparticles and themembrane surface As such interactions are often dominated by electrostaticinteractions and nonspecific interactions, membrane surface modificationoften focuses on introducing charge groups and increasing hydrophilicity.Feed pretreatment can adjust the pH and salt concentration to alter the chargeeffect, but more importantly can charge the particle size distribution by pro-moting or demoting aggregation and hence improve the “filterability,” i.e.,reducing fouling tendency

Flux control is based on the understanding of critical flux, a level of fluxunder which fouling is minimal This flux is defined as a critical flux Thecritical flux hypothesis is that if flux is controlled on start-up of a membraneoperation, there exists a critical flux below which a decline in flux over timedoes not occur The region of operating under which no fouling is found istermed subcritical Experiments showed that MF can be operated at a con-stant flux with no increase in transmembrane pressure if the operating pres-sure is low Fouling is slight or negligible This can be understood byanalyzing the forces acting onto the rejected particles (or macromolecules).The convective force due to permeation pushes the particle toward themembrane pore; diffusional force drives the particle away from themembrane due to concentration gradient and more importantly velocity

Feed

FIGURE 1.8 Back flushing.

gradient-induced diffusion Electrostatic interaction also repels the particles.Whether the particle arrives at the membrane surface depends on the balance

of the forces At a higher flux, the particle will arrive at the membrane at thesurface; and fouling may occur Below the critical flux, it will not The con-cept of critical flux helps to understand fouling and to guide the operation intheory, but difficulty appears in practical applications as (i) its value may betoo low to be practically applied and (ii) it cannot be predicted largelybecause the feed is often a complex mixture

Chemical cleaning is required (i) as often fouling is inevitable and (ii) as

an integrated part of regulatory requirement for food and bioprocessing(cleaning in place) However, chemical cleaning, including the selection ofcleaning agents and formulation, the operational procedure, is largely based

on experience, and the outcome cannot be predicted with confidence This isbecause of the lack of fundamental understanding of fouling, the complexity

of foulant composition and fouling process, and poor characterization of thefeed mixtures This is an area that urgently needs research, as chemicalcleaning represents significant operational cost (cost of chemicals, loss ofproductivity, etc.) in membrane applications in food and bioprocessing.1.6 SUMMARY

This chapter presents some basic concepts related to membranes and brane processes Common issues to membrane application in food and bio-processing are outlined It provides a basis for detailed discussion on specificapplications later in this book

mem-FURTHER READINGS

[1] Scott K Handbook of industrial membranes Elsevier; 1995.

[2] Baker RW Membrane technology and applications McGraw-Hill; 2000.

[3] Li NN, Fane AG, Ho WS, Matsuura T, editors Advanced membrane technology and applications Wiley; 2008.

2.1 HISTORY OVERVIEW

Membrane technology has a relatively short but intense history Asymmetricmembranes (the foundation of most of today’s commercially available ones)were first synthesized in 1960s At that time, membranes were not consid-ered good for any application Later in the 1970s and 1980s membrane tech-nology blossomed and many thought they were going to solve all separationand even reaction issues (Table 2.1)

During the first years the main problems to be dealt with were the production ofusable membranes, the development of reasonable equipment in which the membranescould be used, and the resolving of all the practical difficulties connected with liquidpumping, cooling, high pressure tubing, gaskets, instrumentation, etc., which are asimportant as the more theoretical aspects of the process[2]

Today, membrane technology has a unique place in many industrial andwater management applications Some of those are well settled (i.e., precon-centration steps and protein fractionation in the dairy industry, municipalwaste water treatment by membrane bioreactors, etc.), while others areemerging (i.e., industrial waste water membrane bioreactors) Nevertheless,

in the past 20 years, some membrane applications have not left the

19Membrane Technology DOI: 10.1016/B978-1-85617-632-3.00002-1

© 2010 Elsevier Ltd All rights reserved.

TABLE 2.1 Early Version of the Membrane Filtration Spectrum[1]

Exhibit A Spectrum Membrane Processes

Membrane

Solid particles

Water Distinct Pores

(10 PSI)

Suspended colloidal

.100 A˚

Water Tighter membrane

stream

Dialysate feed Impurities

Ions low MW organics

Concentration

“promising technology” group (e.g., pervaporation for aroma recovery orsome enzymatic membrane reactors in a number of full-scale applications),either because of the inherent membrane limitations (low flux or selectivity)

or because material or system engineering drawbacks have not yet beenovercome The membrane lifetime picture has remained quite stable in thepast decade, as shown inFigure 2.1

From a global perspective of the industrial-scale processes, membraneshave succeeded in being alternate technologies to conventional separationtechniques, allowing more compact systems (with the added advantages of amodular design) and separations, which have challenged the way processengineers think and design a process Membranes can be fitted in at variousplaces in a production facility, and they can even been synergistically com-bined with other separation or reaction processes leading to hybrid technolo-gies This is one of the powerful features of this technology

Membranes are currently considered amongst the best available gies (BAT) in many process and waste management applications However,for some of the latter they are still more expensive than the less environmen-tally friendly, but still regulation-compliant, alternatives

technolo-Out of the 100 quads (1 quad = 1015BTU/year) of energy consumed inthe United States, the food processing industry approximately uses 2 quads

At $8/million BTU, this amounts to 8 billion dollars a year Half of theenergy consumed in the food processing industry is toward concentrationand drying For instance, corn wet milling uses 93.7 trillion BTU/year, grainmilling 153.3 trillion BTU/year, and vegetable oil processing 2.0 trillionBTU/year As an example, drying and evaporation steps are used in corn

Life time of process

Application know-how availability

High profit availability

Hemodialysis

Moderate profit high growth rate

Low profit high production efficiency

Microfiltr ation Ultr

afiltr ation

Re verse osmosis

Gas separationas

Electrodialysis

P er vaporation Membrane reactor

milling for concentrating the steep water and further drying it, and also toreach dryness of the germ, gluten, starch, and to concentrate dextrose solu-tion products (Fig 2.2) If the drying costs can be reduced for large-scaleprocessing in developing countries, it could be an enormous advantage This

is so because of the broad relation between wealth and energy consumption,

Efficient separations are, thus, needed not only to make industrialprocesses more economical, but also to accomplish the high purity andhigh selectivity requirements in the food processing and biotech industries.The cost of recovery or separation of a product from its raw materialdepends upon the efficiency of the separation processes involved, and theyincrease substantially with dilution (Fig 2.4) In such cases, processing costscan only be controlled using highly efficient downstream separation/purifica-tion processes This is especially relevant for new products (for which newprocesses have to be designed), owing to their contribution to the overalleconomics of the process

In such a scenario, membranes find several applications in food sing, owing to the different tasks they are able to perform (Table 2.2)

proces-Steeping

Drying

Grinding

and separation

Drying Drying Evaporation

Heavy step water

Fructose and fermentation products

2.2 MEMBRANE INDUSTRY

Membranes market has been steadily growing in the past decades More thanhalf (close to 60%) of the membrane area in the market (as well as sales)goes into different microfilters (Fig 2.5a), while ultrafiltration and reverseosmosis have an equivalent portion close to 17% The rest of the membranetechnologies share the remaning 8%, with pervaporation occupying about1.5% of that Among the different choices for membrane materials, polymersare used in close to three-quarters of the commercially available membranes,the rest being ceramic, metallic, and other inorganic materials (Fig 2.5b).With respect to the module type, the shares among tubular, capillary/hollowfiber, spiral wound, and flat/plate and frame are fairly even (Fig 2.5c).Membrane industry has been performing an excellent research and devel-opment (R&D) work on membrane materials, to make them suitable for awide variety of applications with extended life Significant efforts have alsoled to a full range of new nanofiltration membranes However, membranesthemselves are only a small part of the membrane system: modules, connec-tors, spacers, seals, etc., together with additional pieces of equipment forfluid flow and control are needed for operation Better systems and advancedcontrols, easier to operate, are also amongst the positive steps taken by mem-brane industry

1 0.1

Energy consumption per capita (TOE per person)

Affluence

Poverty

USA

France UK

Threonine Cephalosporin Gentamicin Gibberillic acid Bulk enzymes Glucose oxidase Insulin Monoclonal

antibodies

Glycerophosphate dehydrogenase Luciferase

Factor VIII Urokinase Therapeutic enzymes

Aamino acids Antibiotics Microbial proteases Amylases

Renin Research/diagnostic enzymes

FIGURE 2.4 The selling price of a product is a strong function of product concentration and, consequently, cost of separation and/or purification Reactor cost is typically under 25% of total production cost Three distinct categories are evident [5]

TABLE 2.2 Typical Examples of Membrane Applications in Food Processing(Sources: Cuperus and Nijhuis[6]; Muralidhara[7,8])

MF/UF Beverage, corn milling

Drying/thicken Whey UF/RO Potato, dairy

Desalting Water, cheese RO/ED Dairy, beverage Concentration Fruit juice, sugar RO Beverage, sugar

refining Dealcoholization Wine PV Beverage

Fractionation Egg UF Poultry

Product recovery Lactic acid, citric acid UF/ED Biotech

2.3 CURRENT CHALLENGES OF MEMBRANE TECHNOLOGYMain drawbacks that prevent membrane technology from expanding furtherthe current application often involve economic factors: membranes are stillexpensive and membrane systems are still energy intensive; both influenceoperating costs negatively However, membrane cost has been continuouslydeclining, e.g., for high volume applications (membrane bioreactors; MBRs),

as depicted in Figure 2.6 Other membranes have followed a differentapproach, which makes them highly specialized for applications in whichthere is currently no competitive process

Other current challenges can be associated with the membrane nies’ policies: (i) membrane designs are too often manufacturer specific,which makes difficult the exchange of membranes for those of a competitor,

compa-if desired and (ii) application-speccompa-ific membranes are not being developed.Membrane system costs and application are currently material limited.Membrane performance is measured as water flux and selectivity However,for an efficient and economically feasible industrial application, membranesneed to keep their integrity for their whole lifetime Often integrity and waterflux follow the opposite trend, such as polyethersulfone membranes, which

as a resistant, show only a fraction of the flux given by the regenerated lose membranes, which in turn have more integrity problems (Fig 2.7) Lessintegrity means lower life, and thus, higher replacement costs

cellu-Microfiltration

Reverse Osmosis 17%

Pervaporation 2%

Other 7%

(A)

Inorganic

27%

Polymeric 73%

(B)

Flat/plate and-frame 20%

Tubular

fiber 26%

Spiral wound 25%

(C)

FIGURE 2.5 Membrane processes (a), materials (b) and module types (c) according to the market use (adapted from Freedonia, 2000 [9] , and Mulder et al., 1997 [10] ).

Moreover, membranes need to be placed in modules, and those are just apart of the membrane system Module design and fabrication still have someissues to be tackled, e.g., compatibility of adhesives, seals, spacers, and feeddistributors System components, i.e., housings and connectors also face the

FIGURE 2.6 Membrane cost decline for MBR applications [11]

FIGURE 2.7 Membrane material limitations.

same problems This is especially important in food applications, as all thesecomponents have to be food-contact approved All the module componentsshould be pH compatible (both acidic and basic), solvent resistant, andenvironmentally friendly and recyclable As mentioned before, standardiza-tion is also needed; system components should be adaptable for membranesfrom different manufacturers, especially at lab and pilot scales Other issuesthat arise from the system design and configuration are fluid dynamics andcontrol-related issues Longer tubular systems, for instance, are not recom-mended, as membrane performance changes along the membrane length,since both transmembrane pressure and tangential velocity decrease Flowbypass in between the module and housing is another flaw in most designs,which is overcome by additional pumping capacity, but still results in anunnecessary poorer performance Control has to be kept simple, for the ease

of operation and maintenance

Another key issue is the use of both concentrate and permeate streamsleaving the membrane unit Most applications focus on one of the streams,leaving the other as a waste The ability to recycle or find an application forthe other stream is often the key for an economically feasible application,and the path to an integral usage of raw materials and an environmentallyfriendlier process

Membranes need periodic cleaning to maintain their performance at thedesired level Cleaning is performed using aggressive chemicals at moderatetemperatures (following manufacturer’s recommendations) Despite this, theynever result in a complete (100% effective) performance recovery, theyalways result detrimental to membranes, and they add to waste water.Typical clean in place systems (CIP) designed for large membrane systems,involve several steps, i.e., water flush, caustic/ultraclean wash or acid wash,water flush, and again caustic/hypochlorite wash with a final rinse before theprocess stream is again fed to the system Alternating the pH between acidand alkaline is detrimental to membranes Moreover, the use of hypochloriteraises several issues: leads to corrosion, makes membranes brittle, isunstable and thus bulk storage is not possible, needs operation interventionand makes automation not possible, and results in chlorine-containing com-pounds sent to the wastewater treatment facilities

Therefore, current CIP methods require high water consumption, cals, and energy, while they also imply downtime from processing Moreover,they reduce the choice of membrane materials (they are one of the reasons cel-lulose acetate membranes cannot be used)

chemi-Substances used for membrane cleaning must target the foulants, andnot the membranes An example of successfully implemented alternativechemicals are enzymatic cleaners They can be used to supplement theeffectiveness of enzymatic cleaners, thus allowing mild alkaline detergents,which are less detrimental to membranes They require a pH of between

9.5 and 10 They are stable under plant conditions and they allow bulk age and automation.

stor-Cleaning protocols also need optimization in frequency, temperature, anddosage Operating conditions can also be modified to reduce fouling and theneed of cleaning A successful approach following this idea is operation atlow transmembrane pressures, below the so-called critical flux In such con-ditions, the lower flux results in a need for larger membrane area, but the ini-tial investment is overcome by sustained reduced operating costs (lowerpressure results in lower pumping costs, and if fouling is under control,cleaning can be minimized, with no need of chemicals for periods of years).This is the approach currently used in membrane bioreactors for waste watertreatment, as for this application, low unit cost is paramount

Feed pretreatment is often used, as it acts as a guard against process turbances and inconsistencies, improves membrane throughput, cleanability,and life However, it may add to the total cost of a membrane system It isoften most important for a successful membrane operation Examples of typi-cally used feed pretreatment include clarification or particle filtration beforemicrofiltration, microfiltration prior to reverse osmosis, pH adjustment (ifallowed), and heat treatment (e.g., for proteins)

dis-Last, but not least, for successful design and operation of membrane tems, there is an outstanding need of engineering expertise More emphasisneeds to be given at the academic levels; i.e., undergraduates should knowthe practical aspects of membrane separation processes basic principlesand typical applications; and advanced degrees should develop a deeperunderstanding of principles and engineering issues (interfacial phenomena,rheology, material science) This would avoid many of the mistakes led bythe idea that buying any membrane module (or a few hundreds of them) andplacing them in front of a pump will result in the expected separation.Operator training is also relevant for large-scale application

sys-2.4 EMERGING APPLICATIONS AND HYBRID PROCESSESOne can easily find in the literature a list of state of the art membrane pro-cesses and emerging applications (e.g.,Table 2.3)

However, some of those “emerging applications” have been consideredlike that for the past three decades, as one could conclude after comparing

Tables 2.3 and 2.4, taken from a text edited in 1979

The above comparision suggests that R&D efforts have to the focused.While keeping doors open for real innovative solutions, potential applica-tions have to be fully explored at different levels, from membrane materialdevelopment to full system design In such a process, decision-making toolsshould be used alike in any other process for the food, chemical, or processindustry

A new field in membrane technology has been recently opened with aqueous applications For that purpose, reliable solvent-resistant membranes,spacers, adhesives, and compatible cleaning techniques have to be devel-oped Applications have already emerged in petrochemical and vegetable oilprocessing.

non-As David H Koch stated in 1997 [13], “While technical advances andefficiency improvements in specific unit operations are occurring all the

TABLE 2.3 State of the Art and Emerging Membrane Processes[3]

No Alternative to Membrane Processes

State of the art processes

High Water desalination, (waste)

water treatment

Production of ultrapure water

Artificial kidney, fuel cell separators Medium Natural gas treatment air

separation

Downstream process

of bioproducts

Therapeutic devices for controlled drug release

Low Dehydration of solvents Biosensors Diagnostic devices Emerging processes

High Membrane reactors MBR Artificial liver

Medium Organic/organic separation Effluent recycling Immune isolation

of cells Low Organic vapour recovery Affinity membranes

TABLE 2.4 Potential UltraFiltration Applications[12]

G Pyrogen removal from water

G Depyrogenation of human chorionic gonadotropin

G Ultrafiltration of prothrombin complex

G Manufacture of drugs and USP, purified water

G Hemofiltration and continuous-flow plasmapheresis

G Ultrafiltration in patients with end-stage renal disease

G Hybrid artificial pancreas and prototype liver assist device

G Production of human plasma protein solutions for clinical use

G Production of protein hydrolyzates in ultrafiltration enzyme reactors

G Concentration of proteins and oil emulsions

G Isolate oilseed protein without an effluent waste stream

G Vegetable protein isolates and concentrates

G Ultrafiltration of gelatin and salt solutions

G Fermentation products

G Surfactant micelle-enhanced ultrafiltration

time, the big story is the hybridization of the processes Combining ual unit operations, such as reaction, separation, heat exchanger into largerconcurrent operations will be a major trend in upcoming years.”

individ-Two concepts of hybrid processes combining a membrane unit with aseparation unit can be distinguished (see Fig 2.8): (i) an interlinked combi-nation achieving a single separation and (ii) a combination of consecutiveseparation processes achieving a separation that could be achieved neithertechnically nor economically alone

Most hybrid processes realized on an industrial scale are of the first type.The importance of this is further highlighted when water management, effec-tive material usage, and wastewater treatment is considered as a part of cleanproduction, which is a strategic element in production in the industry nowa-days While the first type achieves a clear separation of components A and

B, the second scheme commonly requires a regeneration or disposal step todeal with the component A retained in the second process This reduces thepotential economic and environmental benefits of the process and is there-fore the less favorable option[14]

A number of industrially relevant examples of the second type can also

be found, i.e., in the preconcentration by reverse osmosis, prior to tion This results in a significant energy consumption reduction (typicalvalues are given in Fig 2.9) Conventional evaporation results in energycosts close to $4.2 per metric ton of feed (or approximately $105 per metricton of soluble solids) By using reverse osmosis as a preconcentration step,most of the water is removed by filtration and not by evaporation, whichresults in energy savings of 78% ($3.3 per metric ton of feed are saved).Moreover, the permeate water can be directly recycled in the plant

evapora-Successful hybrid processes have economic benefits, often lower energyconsumption and lower capital costs resulting from a more intensiveapproach (smaller equipment size) and the fact they allow a better usage ofmaterials and energy, which results in an improved performance (better sepa-ration efficiency) They are the closest we have approached the “zero dis-charge” paradigm

B A

A + B

B A

A + B

FIGURE 2.8 Two concepts of hybrid processes combining a membrane unit with a separation unit (a) with internal recycle and (b) without internal recycle (adapted from Field and Lipnizki,

2001 [14] ).