Christian Kennes, MarÃ A C. Veiga (Auth.), C. Kennes, M. C. Veiga (Eds.) Bioreactors For Waste Gas Treatment 2001 PDF

Bioreactors for Waste Gas Treatment
ENVIRONMENTAL POLLUTION
VOLUME 4
Editors
Brian J. Alloway, Department (}fSoil Science, The University ofReading, UK.
Jack T Trevors, Department (~fEnvironmental Biolo!?y, University (~f Guelph, Ontario,
Canada
Editorial Board
T Anderson, The Institute ()(Environmental and Human Health, Texas Tech University,
Lubbock, Us.A.
TH. Christensen, Department (~!"Environmental Science and Engineering, Danislt
Tecltnical University, Lyngby, Denmark
I. Colbeck, Institutelor Environmental Research, Department ()!" Biolo!?ical Sciences,
University ()f Essex, Colcltester, UK.
K.c. Jones, Institute ()!" Environmental and Natural Sciences, Lancaster Universi(v, u.K.
S. Parry, TH. Huxley Scltool (~!" Environment, Earth Sciences and Engineering, Imperial
College at Silwood Park, Ascot, Berks, u.K.
W. Salomons, GKSS Research Center, Geesthacht, Gennany
Bioreactors for
Waste Gas Treatment
edited by
C. Kennes
and
M.C. Veiga
University (~t"La Comiia,
La CoruFia, Spain
SPRINGER-SCIENCE+BUSINESS MEDIA, B.V.

Library of Congress Cataloging-in-Publication Data
ISBN 978-90-481-5772-3 ISBN 978-94-017-0930-9 (eBook)

DOI 10.1007/978-94-017-0930-9
Printed on acid-free paper
AII Rights Reserved

© 200 l Springer Science+Business Media Dordrecht
Originally published by Kluwer Academic Publishers in 2001
No part of the material protected by this copyright notice may be reproduced or
utilized in any form or by any means, electronic or mechanical,
including photocopying, recording or by any information storage and
retrieval system, without written permission from the copyright owner.
This book is dedicated to David and Maria
and to the memory of Henri Kennes
CONTENTS
Preface Xlll
Acknowledgments xv
Contributors XVll
• PART 1 PRINCIPLES 1
Chapter 1 Fundamentals of air pollution 3

1. The atmosphere 3
2. Sources of air pollution 4
2.1. Mobile sources 4
2.2. Stationary sources 5
3. Types of air pollutants 9
3.1. Particles 9
3.2. Vapours and odours 10
4. Pollution quantification 11
4.1. Particles and vapour concentrations 11
4.2.0dours 11
References 12
Chapter 2 Non-biological treatment technologies 17
1. Introduction 17
2. Removal of particulate matter 17
2.1. Gravity settling chambers 18
2.2. Cyclones 19
2.3. Filters 20
2.4. Electrostatic precipitators 22
2.5. Scrubbers 23
3. Removal of volatile compounds 25
3.1. Absorption 25
3.2. Adsorption 32
3.3.lncineration 37
3.4. Condensation 42
3.5. Innovative processes 44
References 45
Chapter 3 Conventional biofilters 47
1. Introduction 47
2. Design of biofilters 47
2.1. Conventional biofilters fundamentals 47
2.2. Reactor configuration 48
2.3. Biofilter engineering and modelling 49
3. Parameters affecting biofilter performance 62
3.1. Feed conditions and composition 62
3.2. Carrier material 63
3.3. Microflora 68
3.4. Role ofnutrients and oxygen 74
viii Contents
3.5. Water content 76

3.6. Temperature 80
3.7. pH 81
3.8. Pressure drop and clogging 82
4. Costs 84
5. Applications 87
5.1. Lab-scale research 87
5.2. Full-scale applications 87
References 91
Chapter 4 Biotrickling filters 99
1. Introduction 99
2. Biotrickling filtration fundamentals 99
2.1. Biotrickling filtration principle 99
2.2. Definitions, performance reporting 102
2.3. Biotrickling filter construction 104
3. Biotrickling filter performance 106
3.1. Examples of pollutants treated in biotrickling filters 106
3.2. Toluene as model pollutant: comparison of biotrickling
filter performances 107
3.3. Pilot-plant studies at US waste water treatment plants 111
4. Factors affecting biotrickling filter performance 112
4.1. Temperature 112
4.2. Oxygen 113
4.3. Packing material 114
4.4. lnoculation and microbial ecology 115
4.5. Nutrients 117
4.6. Liquid recycling 119
5. Biomass growth and long term performance ofbiotrickling
filters 120
5.1. Biomass growth kinetics and pollutant elimination 120
5.2. Stages ofbiomass formation 121
5.3. Biofilm architecture and mass transfer 122
5.4. Steady state versus non-steady state 123
5.5. Prevention of clogging 124
6. Conclusions 126
References 126
Chapter 5 Bioscrubbers 133
1. Introduction 133
2. Absorbers 134
2.1. Construction 134
2.2. Interphase mass transfer and kinetics 137
3. Bioreactors 140
4. VOC removal: cases 144
4.1. Experiences in the nineteen eighties 144
4.2. Experiences in the nineteen nineties 145
5. VOC removal: developments 148
5.1. Thermophilic bioscrubbing 148
5.2. Bioscrubbers with two liquid phases 148
Contents IX
5.3. Anaerobic bioscrubbers 150

5.4. Cometabolic bioscrubbing 151
5.5. Foams 151
6. H2 S and SOx removal 151
6.1. H2S removal from aerobic gases 151
6.2. H2S removal from anaerobic gases 152
6.3. SOx removal from f1ue gases 153
7. NO x removal 155
8. NH 3 removal 155
9. Costs 156
9.1. Bioscrubbers for VOC and (aerobic) H2S 156
9.2. Other bioscrubbers 157
10. Conclusions and future opportunities 158
References 159
Chapter 6 Membrane bioreactors 163
1. Introduction 163
2. Membrane bioreactor applications for pollution control 164
3. Membrane fundamentals 166
4. Research overview 167
4.1. Treatment of low solubility compounds 170
4.2. Cometabolism 171
5. Theoretical models 172
5.1. Membrane mass transfer 173
5.2. Suspension mass transfer and degradation 173
5.3. Biofilm mass transfer and degradation 174
6. Conclusions 174
References 175
Chapter 7 Combined advanced oxidation and biodegradation 179
1. Introduction 179
2. Photochemical treatment of gases 180
2.1. Technology available 180
2.2. Photocatalytic oxidation 181
2.3. Photolysis 185
3. Treatment of gases using a non-thermal plasma 186
3.1. Dielectric barrier discharge and corona discharge 187
3.2. Electron beam discharge 189
4. The used of advanced oxidation technologies in water
treatment 190
5. Combined advanced oxidation and biological treatment of
gases 192
6. Conclusions and future opportunities 193
References 195
Chapter 8 Rotating biological contactors 201
1. Introduction 201
1.1. RBC for wastewater treatment 202
1.2. History of RBC for waste gas treatment 202
x Contents
2. Experience with the RBC for waste gas treatment 203

2.1. Results with the RBC 204
2.2. Comparison ofthe RBC and a biotrickling filter 206
2.3. Parameters and models 208
3. Design improvements 211
4. Conclusions 212
References 213
Chapter 9 Activated sludge and suspended growth bioreactors 215
1. Overview 21 5
2. Process description 216
2.1. Activated sludge 216
2.2. Sparged suspended growth reactors and bubble columns 219
2.3. Airlift reactor 222
3. Process variables 223
4. Modelling 225
4.1. Gas-liquid mass transfer 225
4.2. Biodegradation kinetics 227
4.3. Model predicted gas treatment 229
4.4. Power 232
5. Experience treating contaminated gases 232
5.1. Odours 232
5.2. lnorganics 236
5.3. Aromatic organics 240
5.4. Chlorinated organics 242
5.5. Other organics 246
6. Costs to construct and operatc suspended growth gas treatment
systems 246
7. Variations 248
7.1. Solid carrier addition to liquid 248
7.2. Co-solvent addition to liquid 249
8. Conclusions 249
References 250
PART2 APPLICA TIONS 255
Chapter 10 Biofiltration ofwaste gases from a dairy industry 257

1. Background 257
2. Technology selection 257
3. Biofilter design 258
3.1. Pre-treatment 258
3.2. Biofilter 258
4. Operation ofthe treatment system 259
References 260
Chapter 11 Treatment of high VOC levels in a c10sed biofilter 261
1. Introduction 261
2. Technology selection 261
Contents xi
3. Biofilter design 262

4. Biofilter performance 264
4.1. Start-up and operation 264
4.2. Optimization ofbiofilter performance 265
5. Conclusions 266
References 267
Chapter 12 New bioreactor system for treating sulphur- or nitrogen-
compounds 269
1. Introduction 269
2. Why deve10p a new bioreactor? 269
3. The different types of Bioway' s bioreactor system 270
4. Operating strategies 272
4.1. Introduction 272
4.2. Control per layer of media 273
4.3. Irrigation 274
4.4. Media composition 275
4.5. pH control 275
5. Start-up and process control 276
5.1. Start-up 276
5.2. Process control and overcoming operating problems 277
6. Examples 277
6.1. Case 1: Potato processing plant 277
6.2. Case 2: Brewery 278
6.3. Case 3: Sponge manufacturing plant 278
6.4. Case 4: Municipal waste water treatment plant 279
7. Future deve10pments 280
References 281
Chapter 13 Bioscrubber for treating waste gases from wastewater
treatment plants 285
1. Introduction 285
1.1. Damhusaaen waste water treatment plant 285
1.2. Assessment of methods to reduce odour emission 286
1.3. Air cleaning methods 286
2. Plant description 287
2.1. Covers and enclosures 287
2.2. Air cleaning process 288
2.3. Full-scale plant 290
3. Process experience 293
3.1. Start-up 293
3.2. Cleaning efficiency 293
3.3. Pressure drop 294
3.4. pH measurement 294
3.5. Maintenance 294
3.6. Consumption of chemicals 294
3.7. Operating costs 295
3.8. Effect of odour emissions on the neighbourhood 296
4. Future measures 296
5. Conclusions 297
xii Contents
References 298
Chapter 14 Odour control at waste water treatment plants by diffusion
into activated sludge basins 299
1. Introduction 299
2. Summary ofreported problems 299
2.1. Corrosion 299
2.2. Organic film accumulation 299
3. Concord, New Hampshire, case study 303
4. Valley Forge, Pennsylvania, case study 303
5. Annapolis, Maryland, case study 305
Index 307
Preface X11l
PREFACE
Air pollution is, and will certainly remain, one of the major problems to be solved
over the next century. Severa! conventional physical and chemical technologies have
been deve10ped and set up over the past decades, and many books describing such
techniques have been published. However, more recently new treatment alternatives
have been tested, based on the use of bioreactors. It very quickly appeared that such
systems present several advantages, among which are their high efficiency and low
cost. The treatment of polluted air in bioreactors allows the complete destruction of
the contaminants, contrary to such conventional technologies as adsorption or
absorption transferring the contaminant from one phase (gas) to another (liquid or
solid). In bioreactors pollutants are oxidised by microorganisms into innocuous
products. In more c1assical technologies, such as thermal or catalytic incineration,
pollutants are oxidised as well, but usually at rather higher investment and operation
costs.
An ever increasing number of research studies related to biological waste gas
treatment in bioreactors have been published in scientific journals, as well as some
review papers mainly on conventional biofilters. The conventional biofilter
represents the first developed reactor design for air pollution control. A large number
of such bioreactors have been installed relative1y recently, in Europe, Canada and the
United States, as well as in many other developed countries. Most of those reactors
operate successfully for solving air pollution problems. Over the past decade the
efficiency of biofilters has been improved further and the application fields have
been broadened. New bioreactor designs have been studied, as well. Biotrickling
filters (or trickling biofilters) and bioscrubbers are two such new designs that were
first studied at lab-scale. Full-scale systems were installed later, many of which have
now been operating successfully for several years in the field. The same holds true
for systems such as Rotating Biological Contactors (RBC) and airlift bioreactors,
which are being used at industrial scale, though in a much more limited range of
applications. Odour diffusion into activated sludge basins is another alternative
technology for air pollution control. Finally, the lab-scale efficiency of some
relatively new bioreactor designs, such as photobioreactors and membrane
bioreactors is at present actively being evaluated in research laboratories.
As bioreactors appear to be of great interest for solving air pollution
problems, no doubt a book on this topic was highly needed. This is the first book
presenting in a single volume a complete comprehensive overview of all major
bioreactor designs presently available on the market. A few chapters are also
dedicated to systems still at the lab-scale or pilot-scale research stage. A complete
chapter focuses on c1assical non-biological alternatives, allowing easy comparison of
physical, chemical and bioreactor technologies, and a more efficient selection of the
most suitable techniques for industrial-scale applications. The book is divided in two
parts. The first part is dedicated to explaining the principles of the different
techniques, while full-scale bioreactor applications are detailed for each system in
the second part of the book. Each chapter has been written by experts with much
experience in the specific bioreactor being described.
As different fundamental and applied aspects are presented, the book will be
of value to a broad range of potential readers in industry, environmental agencies, as
well as in academia. It will certainly be of interest to professors teaching graduate
XIV Preface
and post-graduate courses related to environrnental engineering and technology, and

it will serve as an instructional text for students interested in learning more about
innovative treatment alternatives available for air pollution control. In this sense
some of the chapters have evolved from a course on Air Pollution Control for post-
graduate students specialising in Environmental Technology and from lectures for
engineers on the same topic. The book can also be used by consultants and industrial
researchers to select the most suitable air pollution control system for their specific
application. The treatment of subjects such as bioreactor design and performance,
costs and examples of full-scale applications will certainly be useful to practising
environmental engineers.
Christian Kennes
Maria C. Vciga
xv
ACKNOWLEDGMENTS
The editors would like to thank the many colleagues, students, researchers, friends,
and all people in general who helped to some extent preparing this book. Special
thanks are due to the students and technician that helped preparing some of the
Figures and that collaborated in offering suggestions and in revising the chapters. We
are grateful to the companies who provided information and illustrative material
which has enabled us to improve the quality of the manuscript. We would especially
like to acknowledge the collaboration of Tecnium, Plastoquimica and Kalfrisa.
We are indebted to the organisations that make our work on air pollution
control possible. These include regional (Xunta de Galicia), national (Ministry of
Science and Technology) and European organisations and funding agencies.
Last but not least, the editors would like to thank the contributing authors and
the Publisher for their efficient cooperation.
Contributors XVII
CONTRIBUTORS
Angela R. Bielefeldt, University of Colorado, Department of Civil, Environmental

and Architectural Engineering, Campus Box 428, Boulder, CO 80309-0428. (Chapter
9)
Robert P.G. Bowker, Bowker and Associates Inc., 477 Congress Street, Suite 1004,
Portland, ME 04101. (Chapter 14)
Huub H.J. Cox, University of California, Department of Chemical and

Environmental Engineering, Riverside, CA 92521. (Chapter 4).
Mare A. Deshusses, University of California, Departrnent of Chemi cal and

Environmental Engineering, Riverside, CA 92521. (Chapter 4).
Sarina J. Ergas, University of Massachusetts, Department of Civil and

Environmental Engineering, 18 Marston Hali, Amherst, MA 01003. (Chapter 6).
Johan W. van Groenestijn, Institute of Environmental Sciences, Energy Research

and Process Innovation, Department of Environmental Biotechnology, PO Box 342,
NL-7300 AH Apeldoom. (Chapters 5 and 7)
Niels G. Hansen, Kriiger AlS, 363 Gladsaksevej, DK-2860 Soeborg (Chapter 13).
Christian Kennes, University of La Corufia, Department of Chemical Engineering,

Campus daZapateira, E-15071 LaCoruna (Chapters 1,2,3 and 10)
N.J.R. Kraakman, Bioway B.V., P.O Box 361, NL-6710 BJ Ede. (Chapter 12)
MiehaeI S. MeGrath, Monsanto Enviro-Chem Systems Inc., S1. Louis, MO.

(Chapter Il).
Jan-Carel Nieuwland, Monsanto Europe S.A./N.V., Avenue de Tervuren 270-272,

B-1150 Brussels. (Chapter Il).
Osear Prado, University of La Coruna, Department of Chemical Engineering,

Campus da Zapateira, E-15071 La Coruna (Chapter 2)
Kim Rindel, Lynettefaellesskabet IlS, 250 Refshalevej, DK-1432 Copenhagen.

(Chapter 13).
PhiIipp Rudolf von Rohr, ETH Centre, Institute of Process Engineering, CH-8092
Zurich. (Chapter 8).
Patrik Ruediger, ETH Centre, Institute of Process Engineering, CH-8092 Zurich.

(Chapter 8).
XVIIl Contributors
Maria C. Veiga, University of La Corufia, Department of Chemica1 Engineering,

Campus daZapateira, E-15071 LaCorufia (Chapters 1,2 and 3).
PART 1. PRINCIPLES
CHAPTER 1 FUNDAMENTALS OF AIR POLLUTION
Christian KENNES and Maria C VEIGA
1. The atmosphere
The atmosphere is a gas layer several hundreds of kilometres thick sUITounding the
earth. The average chemical composition of clean, dry atmospheric air is almost
constant, as presented in Table 1.1. In fact, atmospheric air also contains 1 to 3%
water vapour. Except for nitrogen and inert gases characterised by a basically
constant concentration and with an almost permanent residence rime, other
compounds listed in the table present a hmited residence time. The residence time of
compounds, such as ammonia, hydrogen sulphide or nitric oxide, is only a few days.
Although they are naturally present in the atmosphere, they may become
contaminants if their concentration increases, for extended periods of time, above
trace levels or above concentrations listed in Table 1.1.
Table 1.1. Composition of elean dry air
Compound % volume (or 10 4 ppmv)

Nitrogen 78.08
Oxygen 20.95
Argon 0.93
Carbon dioxide 0.032
Neon 0.0018
Hehum 0.00052
Krypton 0.0001
Methane 0.0001
Hydrogen 0.00005
Carbon monox.ide 0.00001
Xenon 0.000008
Ozone 0.000002
Sulphur dioxide < 0.000001 (traces)
Ammonia < 0.000001 (traces)
Nitric oxide < 0.000001 (traces)
Hydrogen sulphide < 0.000001 (traces)
A compound is considered to be a contaminant when it is present in the air at

concentrations adversely affecting human health or the environment in general,
including animals, plants, microbes or even buildings and other man-made materials.
Although minor air pollution problems were reported several centuries ago already
(Stern, 1976), it is typically a problem of the twentieth century and will remain a
concern for at least several decades in the future.
C. Kennes and M. C. Veiga (eds.), BioreactorsJor Waste Cas Treatmel1t, 3-15.

© 2001 Kluwer Academic Publishers.
4 C. Kennes and M C. Veiga
2. Sources of air pollution
Alterations in the composition of clean atmospheric air may originate from either
natural or anthropogenic sources. Natural sources have been present on earth for
ages, while pollution from anthropogenic sources appeared more recently and has
increased exponentially with the industrial revolution.
The intensity of natural pollution depends on many factors, such as location,
vegetation and temperature. Among the main natural sources of pollutants are
volcanic eruptions. Volcanoes are major sources of particlcs and ash, volatile
hydrocarbons, sulphur dioxide and hydrogen sulphide, the latler being oxidised to
S02 in the atmosphere. Among other natural sources are forest fires that may
produce smoke, ash and volatile gases. Decay of organic matler is a source of
ammonia, nitrous oxide (N20) aud methane, among others. It is interesting to note
that on an annual basis, some contaminants, such as hydrogen sulphide, some
nitrogen compounds, hydrocarbons, carbon monoxide, etc., are produced to a larger
extent from natural sources than from non-natural ones. However, anthropogenic
sources may be more harmful, since contaminant release may often be more
intensive locally and for prolonged periods of time. They are, in most cases, located
near regions with high population densities.
Anthropogenic sources are related to human and industrial activities. They
can be classified into stationary sources and mobile sources. Stationary sources
include domestic sources, mainly heating devices, and industrial sources releasing
contaminants largely through combustion processes and via ali kinds of waste gases
in general. Waste gases released from waste water treatment plants, composting or
other waste treatment processes are considered stationary sources, as well. Air
pollution may also result from remediation technologies transferring volatile or semi-
volatile contaminants from soil or aquifers into the atmosphere. Anthropogenic
mobile sources refer to ali types of vehicles, aeroplanes, boats and any other kind of
transportation related source.
Bioreactor technologies described in this book are, at this stage, only applied
to stationary sources. Such technologies are only suitable for the removal of
vapours/odours, and particulate matter often needs to be removed by other non-
biological processes first, to avoid problems during the biological treatment.
Treatment of VOCs from industrial waste gases is a relatively new application of
bioreactor technologies. Conversely, the treatment of odours from some specific
sources using conventional biofilters has been applied for several decades. especially
in the case of air pollution problems related to waste water treatment plants and
composting facilities (Chapter 3).
2.1 MOBILE SOURCES
For some pollutants, their release from motor vehicle engines, as a result of
inefficient combustion of fuel, is much more important than the release from
industrial or domestic sources. This is particularly true for carbon monoxide and
some specific hydrocarbons. Vehicle combusters are usually optimi sed to achieve
maximal power aud performance, rather than for minimising pollution. Major
compounds released from vehicles are water vapour, nitrogen oxides (NOx), carbon
dioxide, carbon monoxide, sulphur dioxide, hydrocarbons, particles and lead. Water
vapour and carbon dioxide are usually not considered to be contaminants. Several
Fundamentals 5
methods have been adopted to reduce air pollution from mobile sources, mainly cars
and trucks. Emissions of carbon monoxide, nitrogen oxides and hydrocarbons have
been reduced over the past decades by using catalysts in so called catalytic
converters. Exhaust gas recirculation is another way to decrease emissions of
nitrogen oxides. Since lead is a contaminant and does also poison catalysts, its
concentration has been drastically reduced in gasoline, reducing at the same time, air
contamination by lead. It does, however, affect fuel octane number which can be
compensated for by adding oxygenates to gasoline as methyl-ter-butyl-ether (MTBE)
or ethyl-ter-butyl-ether (ETBE). This does also allow reduction, to some extent, of
carbon monoxide and VOC emissions, at least at high engine loading (Osman et al.,
1993, Poulopoulos and Philippopoulos, 2000). Nevertheless, MTBE and ETBE are
aIso toxic and have been detected in the exhaust gas of vehic1es. The automotive
industry continues searching for alternative ways of reducing air contamination.
2.2 STATIONARY SOURCES
2.2.1. Combustion processes

Combustion processes are used at both domestic and industrial levels. Stationary
combusters represent the most important domestic source of air pollution. The nature
of pollutants generated by such sources is dependant on the type of fuel used.
Combustion of gasoil generates pollution by sulphur oxides (S02, S03), nitrogen
oxides, hydrocarbons and partic1es. Combustion of carbon is a source of sulphurous
anhydride, nitrogen oxides, flying ashes, soot and heavy metals. Natural gas used in
domestic combustion devices hardly contaminates. At industrial scale, the use of
fossil fuels, mainly in power stations, is one of the major stationary sources of
pollution via combustion. The amount of sulphur oxides released into the atmosphere
is highly dependent on the composition ofthe prime matter.
2.2.2. Industrial waste gases

Other important sources of air pollution are industrial waste gases re1eased during
industrial production processes. Typical examples are chemical reactors, distillation
units, boilers, stripping systems, condensers, etc.. Many such waste gases can be
treated in biological reactors. Chief parameters to be considered are waste gas flow
and concentrations (see also Figure 2.1) as well as the nature of the pollutants since
they need to be biodegradable. Their potential toxic and inhibitory effect also needs
to be considered. More detailed information on this aspect is given in the following
chapters but as a general rule the use of bioreactors will be interesting in the case of
low pollutant concentrations combined with relatively high gas flows. The type of
pollutants released to the atmosphere depends on the kind of industry being
considered. It would be very difficult to present an exhaustive list of applications,
although some examples of industry related waste gases treatable in bioreactors, or
air pollution problems solved via biological processes are given in Table 1.2. More
detailed examples can also be found in other chapters of this book.
2.2.3. Fugitive emissions

Fugitive emissions from industrial storage tanks are also stationary sources of
pollutants. Gasoline and petroleum storage tanks are typical examples. These are
usually low flow and high concentration waste gases treatable in gas phase
bioreactors. The complex nature of volatile pollutants re1eased from petroleum
related products, such as gasoline, might appear as a potential problem for biological
treatment. However, gasoline mainly composed of alkylbenzenes and alkanes is
biodegradable by many microorganisms when the VOCs are present at relatively low
concentrations (Hodge et al., 1991; Mikesell et al., 1993; Kennes et al., 1996).
2.2.4. Site remedia/ion

Polluted air originating from site remediation processes is also often treatable in
bioreactors. Contaminants found in polluted soils and degradable in vapour-phase
bioreactors are in many cases, though not exclusively, petroleum hydrocarbons.
Pollutants, such as alkylbenzenes (BTEX), appear in contaminated soils, among
others, as a result of leaking underground storage tanks. Biological air pollution
control can be combined to air sparging technology in groundwater pollution, and to
soil vapour extraction systems. Air sparging is used to transfer volatile pollutants
from contaminated groundwater to the air. Mass transfer is made possible by passing
air through the water phase. Soil vapour extraction is a similar technology used to
remove pollutants from soils in situ in the vadose zone or from excavated soils. As in
the previous case, air is passed through the soil resulting in the mass transfer of the
VOCs from the soil to the air stream. Polluted air generated during site remediation
needs to be treated prior to its reiease into the atmosphere. This can be accomplished
mainly by means of activated carbon columns, vapour phase bioreactors or a
combination of both technologies.
2.2.5. Waste water treatment

A waste water treatment plant may represent a stationary source of air pollution.
Volatile compounds are partly transferred into the air mainly in the case of aerated
bioreactors, as in the activated sludge process (Lee et al., 1998), and in other aerobic
waste water treatment technologies through stripping (Parker et al., 1996). In some
cases nearly 50% of the volatile compounds initially present in waste water may be
removed by stripping (Parker et al., 1993). In fact, for water pollution control, when
dealing with waste water containing low concentrations of volatile compounds, the
pollutants could be directly stripped from the waste water, followed by polluted air
treatment through adsorption processes, incineration (Dvorak et al., 1996) or
biological treatment.
2.2.6. Composting
Composting is a fermentation process used for (bio )degrading the organic fraction of
organic waste into a stable humus-like product. Many different substrates have been
used for composting: waste water sludge, yard waste, forest sub-products, bark, etc.
As in basically ali fermentation processes the prime matter is degraded into smaller
volatile products. Odour problems are often linked to composting. Although
fermentation usually takes place under aerobic conditions, anaerobic zones may
sometimes also appear in composting piles resulting in the release of highly odorous
metabolites. Unbalanced C:N ratios are also favourable to the formation and
volatilisation of odorous compounds, such as ammonia formed in the presence of
excess nitrogen. Other volatile compounds typically formed during composting
include sulphides, mercaptans, amines, fatty acids (typical of anaerobic
biodegradation), ketones and some aromatic compounds (Derikx et al., 1989; Hentz
et al., 1992; Miller et al., 1991; Prokop and Bohn, 1985). Treatment of composting
odours is one ofthe original applications ofbiofiltration (Prokop and Bohn, 1985).
Fundamentals 7
Table 1.2. Waste gases and air pollution problems treated with bioreactors
Sources Typical pollutants Treatment References

technology
Aerosol can filling Propane, butane Biofilter Leson and Winer,
1991
Animal rendering Odours Biofilter Huber, 1992
Animal rendering Odours Biofilter Luo and Oostrom,
1997
Asphalt processing Odours Biofilter Blake et al., 1999
plant (mainly H 2S)
Board industries Formaldehyde, Biofilter Pond,1999
phenol, methanol
Broiler chicken Odours, ammonia Biofilter, Pearson et al.,
house Bioscrubber 1992
Chemical industry VOCs Biofilter
(synthesis of organic
chemicals)
Cocoa roasting Odours Biofilter Hofman, 1989
Composting Odours Biofilter Kuter et al., 1993
Composting Odours, ammonia, Biotrickling filter Smits et al., 1995
hydrocarbons
Fibreglass industry Styrene, acetone Biofilter
Fish feed production Odours Bioscrubber Hansen and
Rinde1, 1992
RBC
Fish processing VOCs Biofilter Liebe,1989
Flavour Odours, Biofilter Yavorski, 1997
manufacturing VOC
Flexographer (print VOCs Biofilter Rothenbiihler et
shop) al., 1995
Food industries Odours, VOCs Biofilter Gibson et al., 1994
Food industries H2S, ammonia, Biofilter Chapters 10 and
VOCs Biotrickling filter 12
Foundry Alcohols, Biofilter Maier, 1989
aromatics
Foundry Ethanol Biofilter Leson and Winer,
1993
Foundry Phenol, ammonia Bioscrubber Biiren, 1989
Gelatine works Odours Biofilter Kirchner, 1990
Hardboard Odours Biofilter Allen and Van Tii,
manufacturer 1997
Lacquering industry VOCs (toluene, Biotrickling filter Bronnenmeier et
ethylbenzene, al., 1994
xylenes, butyl
acetate)
Table 1.2. (Continued)
Landfill gas Hydrogen sulphide Biofilter Sabo,1989

Odours
Latex production Styrene, butadiene Biofilter Windsperger et al.,
1990
Leather industries VOCs: toluene, iso- Biofilter Windsperger et al.,
propanol, MEK, 1990
dimethylformamide
Lithographic Acetone Biofilter
operations
Livestock buildings Odours Biofilter Noren,1986
Metal working 2-Butoxyethanol. Biofilter Schmidt. 1993
company methanol
Mineral wool Phenolic Biofilter Lehtomăki et al.,
production compounds, 1992
ammonia,
formaldehyde
Paint manufacturing VOCs Biofilter Hsu et al., 2000
(alkylbenzenes,
ketones, aliphatics)
Pighouse ventilation Ammonia Combined Gerards et al.,
air biotrickling filter- 1995
biofilter
Plastic dashboard Styrene, Biofilter McGrath et al.,
manufacturing Butylacetate 1999
Plywood production Formaldehyde Biofilter Mackowiak. 1992
Printing industry Ethylacetate Biofilter Nolte,1992
Pulp and paper Sulphur Biofilter,
industry compounds Biotrickling filter,
Bioscrubber
Rendering industry Odours Biofilter (soil bed) Prokop and Bohn,
1985
Rendering plants Odours Trickling filter Schirz, 1986
Resins processing Styrene Biofilter Demiriz, 1992
Resins production Phenol Biofilter Demiriz. 1992
Resins/glue Phenol, Biofilter,
production formaldehyde, Biotrickling filter
ammonia
Sponge H2S, CS 2 Biotrickling filter Chapter 12
manufacturing
Storage tanks VOCs Biofilter Mildenberger,
1992
Surface coating VOCs Bioscrubber Schippert. 1989
Surface coating Aromatics Biofilter Demiriz, 1992
Tabaco industry Odours Biofilter Kersting, 1992
Vegetable oii Odours Biofilter Eitner, 1992
production
Fundamentals 9
Viscose rayon H2S, CS 2 Biotrickling filter Chapter 12

manufacturing
Waste water Odours Biofilter Mildenberger,
treatment 1992
plants Odours Bioscrubber Chapter 13
Heist et al.,
1995
Odours Activated sludge Chapter 14
Bowker, 1997
Waste water VOCs Biofilter Mildenberger,
treatment 1992
Waste water Hydrogen sulphide, Combined Gerards et al.,
treatment (brewery) dimethylsulphide biotrickling filter- 1995
biofilter
Woodlpartic1eboard VOCs (phenols, Biofilter
industries a!cohols, terpenes)
(press vents, dryers)
Wood lacquering VOCs (Isobutyl Biofilter Windsperger et al.,
acetate, 1990
ethylacetate,
toluene, xylenes)
3. Types of air pollutants
The most easily detectable air pollutants are those related to smell or odour
problems. Other contaminants leading to air pollution problems, such as carcinogenic
compounds, are sometimes more difficult to detect, though they are not less harmful.
Their effect is often visible only after several years. This is also the case of, for
example, acid rain affecting monurnents, since the degradation process is only
detectable after quite a long period of exposure. The most important air pollutants
are, on the one hand, partic1es, and on the other hand, gases or vapours comprising
volatile organic compounds (VOCs), volatile inorganic compounds (VICs) and
odours. Although both VOCs and VICs may contribute to odour problems, we will
detine and explain odour problems separate1y.
3.1 PARTICLES
Particulate matter refers to either solid or liquid dispersed matter present in the air
and larger than 0.0002 Ilm -the mean size of single small molecules- but smaller
than about 500 Ilm (Wark and Wamer, 1989). Different types of partic1es may be
considered. Aerosols and partic1es represent virtually the same concept, though the
former is slightly more restrictive, since aerosols refer to all liquid or solid airbome
suspensions generally smaller than 1 Ilm. Dust and fumes are solids. Dust results
from grinding or crushing operations and fumes re suit from vapour condensation.
Liquid partic1es are often called mist or fog. Smoke is another type of particu1ate
matter re1eased from incomplete combustion processes. Soot is carbon particles also
10 C. Kennes and MC. Veiga
re1eased during combustion processes. In 1973, it was estimated that about 9% of the
total mass of air contaminants were partic1es. This value has slightly decreased over
the last decades. Particles are released both from mobile sources and from stationary
sources. Some particles may carry substances, such as carcinogenic organic
compounds. Particles cannot be removed in biological reactors. They need to be
removed in a pre-treatment step, should waste gases containing both particles and
volatile pollutants need to be treated in bioreactors. Possible pre-treatment processes
for the removal of particulate matter include filters, cyc1ones, electrostatic
precipitators, etc., as described in Chapter 2.
3.2 V APOURS AND ODOURS
Different definitions have been proposed for the different forms in which gases or
vapours may be found in polluted air. Definitions may vary depending on the
country, region or state. Using the temperature as reference, VOCs are organic
chemicals (vapours) containing carbon atoms and which have a normal boi ling
temperature be10w 100 aC, at 101 kPa. In some cases, a higher maximal normal
boiling temperature of 260 °c is mentioned. If one uses pressure as the key
parameter, VOCs are pollutants with a vapour pressure greater than 0.014 kPa at 25
aC. Here again other definitions can be found and vapour pressures greater than 0.07
kPa are sometimes used for defining VOCs. In the United States, VOCs are
sometimes defined as any organic carbon compound, excluding carbon monoxide,
carbon dioxide, carbonic acid, metallic carbides and carbonates, and ammonium
carbonate, which partlclpates in atmospheric photochemical reactions.
Photochemical reactions are reactions initiated in the presence of solar radiation and
involving pollutants, such as VOCs, NO x and atmospheric oxygen. Free radicals are
then generated, initiating new reactions leading to environmental problems, such as,
among others, smog formation. VOCs include most hydrocarbons. Howcver,
methane is excluded from the list of hydrocarbons as it is relatively nonreactive.
Some other volatile compounds are not significantly involved in photolysis and
should also be excluded from the group of VOCs, such as ethane, acetone, t-
butylacetate, dichloromethane and several halogenated compounds, including many
CFCs, HFCs and HCFCs. When the gases or vapours are inorganic, they are called
Volatile Inorganic Compounds or VICs. Some examples of inorganic air pollutants
are sulphur dioxide, NO x, hydrogen sulphide, carbon disulphide, ammonia, etc ..
Once again it is interesting to note that several VOCs and VICs detected in the
atmosphere are emitted from biogenic sources.
Odour may be defined as a physiological stimulus of olfactory cells in the
presence of specific molecules. The nature and concentration of molecules detected
by olfactory cells varies between individuals and with environmental conditions,
such as temperature, pressure and humidity. According to the definition given above,
the terrn odour includes some VICs, as well as some VOCs. Odours may be very
unpleasant, as in the case of hydrogen sulphide and several other sulphur-
compounds. Other pollutants, such as esters, may seem pleasant initially, though they
often lead to major complaints when the odour problem remains for long periods of
time.
Many VICs as well as a relatively wide range of halogenated and non-
halogenated aromatic and aliphatic organic compounds (VOCs) are biodegradable by
microorganisms (van Agteren et al., 1998) and can be treated in bioreactors.
Fundamentals 11
4. Pollution quantification
4.1 PARTICLES AND VAPOUR CONCENTRATIONS
While the concentration of particulate matter in air is generally expressed in g/m3 ,

concentrations of VOCs or VICs in polluted air are expressed either in glm 3 or in
ppmv. In the case ofliquids, ppm represents a mass to mass ratio and one ppm of any
contaminant in water is equivalent to 1 g of pollutant per 106 g of liquid or I mg/l
since the density of pure water is approximate1y 1000 gll. In the case of air pollution,
common units are parts per million by volume or ppmv, meaning that l ppmv
represents l volume of contaminant per 106 volumes of air plus contaminant. The
advantage of using a volume to volume ratio is that such a ratio does not vary when
modifying temperature or pressure. Concentrations given in ppmv can easily be
converted to glm3 by using the ideal gas law, leading to the following relationships:
At25°C: glm3 = (MW/24.5) ppmv (1.1)
At OaC: glm3 = (MW/22.4) ppmv (1.2)
where MW is the molecular weight ofthe contaminant and 24.5 or 22.4 represent the
volume occupied by one mole of an ideal gas at l atm and, respective1y, 25 °c or O
aC. From the above equations it can easily be concluded that for any temperature or
pressure the following equation applies:
(1.3)
where P is the pressure in atmospheres and T is the temperature in Kelvin.
4.20DOURS
In the case of odours, concentrations are commonly expressed in terms of odour units
per cubic meter or OU/m3 , allowing evaluation ofthe nuisance level to people living
in the surroundings of the odour producing source. One OU/m3 represents the
threshold amount of pollutant that will be detected by 50% of trained members of a
group, afler diluting the contaminant (odour) in one cubic meter of clean air under
standard conditions. This means that if, for a given pollutant, this minimal detectable
odour leve1 is, let us say, 100 ppmv, one OU/m3 ofthat compound will correspond to
100 ppmv. It can easily be understood that such a definition is less rigorous than
concentrations expressed in glm3 as mentioned previously for VOCs or VICs, since
the definition of an odour unit may vary depending on the persons included in the
group.
Another definition has recently been proposed by the European
Standardisation Committee. The European Odour Unit (European OUl m3) is defined
as the amount of pollutant (odour) that, diluted in one cubic meter of pure air under
standard conditions, leads to the same physiological response as a European
reference odour mass (EROM) diluted in one cubic meter of pure air under the same
conditions. One EROM is equivalent to 123 Ilg n-butanol.
12 C Kennes and MC Veiga
Dynamic olfactometry is a normali sed method allowing measurment of odour

concentrations. Samples of contaminated air are taken near the source of odour
discharge, according to reference methods and using normalised sampling devices.
Samples should then be brought to the laboratory as quickly as possible and, in any
event, in less than 30 hours, for analysis in an olfactometer. The olfactometer allows
diluting samples with elean air. A group of at least eight persons is exposed to the
different dilutions to determine the minimal detectable level.
References
Allen, P.J., Van Tii, T.S. 1997. Operating experience with a full scale biofilter at a hardwood miiI.
AIChE Symp. Series 315: 124-129.
B1ake, c., Cook, L.L., Apel, W.A. and Gostomski, P.A. 1999. Biofiltration controls odors from
asphalt processing p1ant. Oii & Gas J. 97: 69-72.
Bowker, R.P.G. Biological odor control by diffusion into activated sludge basins. In: Proceedings of
the Annual Conference ofthe Water Environment Federation, 1997, Houston, TX.
Bronnenmeier, R., Fitz, P. and Tautz, H. 1994. Reinigung von Lackiererei-Abluft mit einem
Gittertriiger-Biofilter. VOI Berichte. 1 \04: 203-215.
BUren, E. 1989. Bio-catalytische Abgasreinigung in einer Kernmacherei. VOI Berichte. 735: 89-107.
Demiriz, A.M. Neue Einsatzgebiete biologischer Filteranlagen: Giei3erei-Bereich und

Ulsemittelabscheidung. In: Biotechniques for Air Pollution Abatement and Odour Control Policies.
Dragt AJ. and van Ham J. (eds.), Eisevier, The Netherlands, 1992, pp. 293-296.
Derikx, P.J.L., de Jong, G.A.H., Op den Camp, HJ.M., Van der Drift, c., Van Griensven, L.J.L.D.
and Vogels, G.D. 1989. Isolation and characterization of thermophilic methanogenic bacteria from
mushroom compost. FEMS Microbio!. Eco!' 62: 251-258.
Dvorak, 8.1., Herbeck, CJ., Meurer, C.P., Lawler, D.F. and Speitel, G.E. 1996. Selection among
aqueous and off-gas treatment technologies for synthetic organic chemicals. J. Environ. Eng. 122:
571-580.
Eitner, D. Emissionsminderung in Olmiihlen durch Biofilter - Erfahrungsbericht. In: Biotechniques

for Air Pollution Abatement and Odour Control Policies. Dragt AJ. and van Ham J. (eds.), Eisevier,
TheNetherlands, 1992, pp. 197-205.
Gerards, R., Gevaert, D. and Vriens, L. 1995. Experiences in odour and off-gas treatment with full-
scale seghobioclean systems. Meded. Fac. Landbouww. Gent. 60: 2273-2278.
Gibson, N.B., Swannell, R.P.J., Woodfield, M., Van Groenestijn, J.W., Van Kessel, R.8.M. Wolsink,
J.H. and Hesselink, P.G.M. 1994. Biological treatment of volatile organic carbons from the food and
drink industry. VOI Berichte. 1104: 261-272.
Hansen, N.G. and Rindel, K. Recent experience with biological scrubbers for air pollution control in
Denmark. In: Biotechniques for Air Pollution Abatement and Odour Control Policies. Dragt AJ. and
van Ham J. (eds.), Elsevier, The Netherlands, 1992, pp. 143-154.
Heist, J.H., Hansen, N.G. and Rasmussen, H.H. Control of odor emissions from wastewater treatment
plants in a bioscrubber. In: Proceedings of the 68'h Annual Conference & Exposition of the Water
Environment Federation, 1995, Alexandria, VA.
Fundamentals 13
Hentz, L.H., Murray, C.M., Thompson, J.L., Gasner, L.L. and DunsonJr., J.B. 1992. Odor control
reserach at the Montgomery county regional composting facility. Water Environ. Res. 64: 13~18.
Hodge, D.S., Medina, V.F., Islander, R.L. and Devinny, J.S. 1991. Treatment of hydrocarbon fuel
vapors in biofilters. Environ. Technol. 12: 656~662.
Hofmann, W. 1989. Biofilter nach Kakaorăstereien. VOI Berichte. 735: 233~242.
Hsu, S.-K., Shen, K.-P., Lin, S.-S. and Wang, Y.-M. Biofilter application for control of volatile
organic compounds (VOC) from paint manufacturing lndustry. In: Proceedings of the 93,d Annual
Meeting & Exhibition ofthe Air & Waste Manage. Assoc., 2000, Salt Lake City, Utah.
Huber, J. Planung, Durchftihrung und erste Erfahrung zum Biofilter Tierkorperbeseitigungsanlage

Platting. In: Biotechniques for Air Pollution Abatement and Odour Control Policies. Dragt AJ. and
van Ham J. (eds.), Elsevier, The Netherlands, 1992, pp. 161~165.
Kennes, C., Cox, H.H.J., Doddema, HJ. and Harder, W. 1996. Design and performance of biofilters
for the removal of alkylbenzene vapors. J. Chem. Technol. Biotechnol. 66: 300~304.
Kersting, U. Behandlung gror3volumiger Abluftstrome durch Biofilter, vorgestellt an Beispielen der

Tabakindustrie. In: Biotechniques for Air Pollution Abatement and Odour Control Policies. Dragt A.J.
and van Ham J. (eds.), Elsevier, The Netherlands, 1992, pp. 155~160.
Kirchner, K. 1990. Biological purification of exhaust air. Chimica Oggi. March: 51~55.
Kuter, G.A., Harper, J.E., Naylor, L.M. and Gormsen, PJ. Design, construction and operation of
biofilters for controlling odors at composting facilities. In: Proceedings ofthe 86 th Annual Meeting &
Exhibition ofthe Air & Waste Manage. Assoc., 1993, Denver, Colorado.
Lee, K.-C., Rittmann, B.E., Shi, J. and McAvoy, D. 1998. Advanced steady-state model for the fate of
hydrophobic and volatile compounds in activated sludge. Water Environ. Res. 70: 1118~1131.
Lehtomăki, J., Torronen, M. and Laukkarinen, A. 1\ feasibility study of biological waste-air

purification in a cold climate. In: Biotechniques for Air Pollution Abatement and Odour Control
Policies. Dragt AJ. and van Ham J. (eds.), Elsevier, The Netherlands, 1992, pp. \31~134.
Leson, G. and Winer, A.M. 1991. Biofiltration: an innovative air pollution control technology for
VOC emissions. J. Air Waste Manage. Assoc. 41: 1045~I054.
Liebe, H.G. 1989. Einsatz von Biofiltern zur Minderung der Emissionen von Anlagcn zum
Verarbeiten von Fleish und Fisch ftir die menschliche Ernăhrung. VOI Berichte. 735: 215~231.
Luo, J. and van Oostrom, A. 1997. Biofilters for controlling animal rendering odor - a pilot-scale
study. Pure Appl. Chem. 69: 2403~241 O.
McGrath, M.S., Nieuwland, J.-C. and van Lith, C. 1999. Case study: biofiltration of styrene and
butylacetate at a dshboard manufacturer. Environ. Prag. 18: 197~204.
Mackowiak, J. Abscheidung von Formaldehyd aus der Abluft im Biofilter. In: Biotechniques for Air
Pollution Abatement and Odour Control Policies. Dragt AJ. and van Ham J. (eds.), Elsevier, The
Netherlands, 1992, pp. 273~278.
Maier, G. 1989. Biofiltration von Gier3ereiabgassen. VOI Berichte. 735: 285~292.
Mikesell, M.D., Kukor, J.J. and OIsen, R.H. 1993. Metabolic diversity of aromatic hydrocarbon-
degrading bacteria from a petroleum-contaminated aquifier. Biodegradation. 4: 249~259.
Mildenberger, H.-J. Biofiltersysteme zur Geruchsbeseitigung und zur Reduzierung von Organika-
Emissionen auf KIăranlagen und in der chemischen Industrie. In: Biotechniques for Air Pollution
14 C. Kennes and M C Veiga
Abatement and Odour Control Policies. Dragt AJ. and van Ham J. (eds.), Eisevier, The Netherlands,
1992, pp. 187-196.
Miller, F.C., McCauley, B.J. and Harper, E.R. 1991. Investigation of various gases, pH and redox
potential in mushroom composting phase I stacks. Aust. J. Experim. Agric. 31: 415-425.
Nolte, F.-J. Ermittlung der Eliminationsleistung eines Biofilters bei der Behandlung von
Ethylacetatabliiften. In: Biotechniques for Air Pollution Abatement and Odour Control Policies. Dragt
AJ. and van Ham J. (eds.), Eisevier, The Netherlands, 1992, pp. 297-301.
Non!n, O. Design and use of biofilters for livestock buildings. In: Odour prevention and control of
organic sludge and 1ivestock farming. Nielsen, V.C., Voorburg, J.H. and L'Hermite, P. (eds.),
Eisevier, The Netherlands, 1986, pp. 234-238.
Osman, M.M., Matar, M.S. and Koreish, S. 1993. Effect of methyl tertiary butyl ether (MTBE) as a
gasoline additive on engine performance and exhaust emissions. Fuel Sci. Technol. Intern. 11: 1331-
1343.
Parker, WJ., Thompson, DJ., BeII, J.P. and Melcer, H. 1993. Fate of volatile organic compounds in
municipal activated sludge p1ants. Water Environ. Res. 65: 58--65.
Parker, WJ., Monteith, H.D. and Melcer, H. 1996. VOCs in fixed film processes. 1: Pilot studies. J.
Environ. Eng. 122: 557-563.
Pearson, C.C., Phillips, V.R., Green, G. and Scotford, I.M. A minimum-cost biofilter for reducing
aerial emissions from a broiler chicken house. In: Biotechniques for Air Pollution Abatement and
Odour Control Policies. Dragt AJ. and van Ham J. (eds.), Eisevier, The Netherlands, \992, pp. 245-
254.
Poulopoulos, S. and Philippopoulos, C. 2000. lnfluence of MTBE addition into gasoline on

automotive exhaust emissions. Atmos. Environ. 34: 4781-4786.
Pond, R.L. 1999. Biofiltration to reduce VOC and HAP emissions in the board industry. TAPPI J.82:
137-140.
Prokop, W.H. and Bohn, H.L. 1985. Soi! bed system for control ofrendering plant odors. J. Air PolI.
Control Assoc. 35: 1332-1338.
Rothenbiih1er, M., Heitz, M., Beerli, M. and Marcos, B. 1995. Biofiltration of volatile organic
emissions in reference to flexographic print ing processes. Water, Air and Soil PolI. 83: 37-50.
Sabo, F. 1989. Praktische Erfahrungen mit Biofiltern zur Reinigung geruchsintensiver Deponiegase.
VOI Berichte. 735: 293-312.
Schippert, E. 1989. Biowăscher nach einer Dosenlackieranlage. VOI Berichte. 735: 77-89.
Schippert, E. 1989a. Das Biosiolv-Verfahren von Keramchemie zur Absorption von schwer
wasserliislichen Liisemitteln. VOI Berichte. 735: 161-177.
Schirz, S. Design and experience obtained with bioscrubbers. In: Odour prevention and control of
organic sludge and 1ivestock farming. Nielsen, V.c., Voorburg, J.H. and L'Hermite, P. (eds.),
Eisevier, The Netherlands, 1986, pp. 241-250.
Schmidt, M. 1993. Abluftreinigungsanlage mit Biofiltern. Chemie-Umwelt-Tech. 22: 74-75.
Smits, M.CJ., Hoek, A.P., Osinga, B., Ottengraf, S.P.P. and Wijngaard, M.H. 1995. Removal of
ammonia and odour from composting faci!ity waste gas with a bio-trickling filter. Meded. Fac.
Landbouww. Gent. 60: 2313-2320.
Fundamentals 15
Stern A.C. Air Pollution, Third Edition , Academic Press, New York, USA, 1976.
Van Agteren, M.H., Keuning, S. and Janssen, O.B. Handbook an Biodegradation and Biological
Treatment of Hazardous Organic Compounds. Kluwer Academic Publishers. Oordrecht. The
Netherlands, 1998.
Wark, K. and Warner, C.F. Air Pollution, lts Origin and Control. Harper & Row Publishcrs, New
York, USA, 1981.
Windsperger, A., Buchner, R. and Stefan, K. 1990. Reinigung 15sungmittelhaltiger Abluft mit
Biofiltern. Staub - Reinhaltung der Luft 50: 465-470.
Yavorski, 1. Odor and VOC control from flavor manufacturing through advanced biofiltration. In:
Proceedings of the 90'h Annual Meeting & Exhibition of the Air & Waste Manage. Assoc. 1997.
Pittsburgh, Penna.
CHAPTER 2 NON-BIOLOGICAL TREATMENT TECHNOLOGIES
Christian KENNES, Maria C. VEIGA and Oscar PRADO
1. Introduction
In order to make an adequate choice of the most suitable treatment technology, it is

necessary to be aware of ali possible major altematives available. Therefore non-
biologica! treatment processes are also briefly described in this chapter. Most often
the major criteria used to evaluate the application range of the different technologies
available for the removal of volatile compounds are gas flow rate and contaminant
concentration (Figure 2.1). However, many other criteria also need to be taken into
account, such as biodegradability in the case of biologica! treatment systems, and
costs, to mention only two examples. Air treatment in bioreactors is often preceded
by a pre-treatment step used, among others, for the removal of particulate mat1er. A
brief description of major technologies used for partide removal will therefore also
be presented in this chapter.
1
10 5 i
i
i INCINERATION :
i
i ,1
L._._._._._._._._._._._~_._._._._j
,,
: ABSORPTION
ADSORPTION ,,
,,,
~------------------
BIOREACTORS ,,
------------------~
(CRYO-)CONDENSAnON
o ~--------+---------~--------_+----------~--------~
10 10 2 103
Conccntration (glm 3)
Figure 2.1. Application range of major biological and non-biological air pollution control
technologies.
2. Removal of particulate matter
Several mature technologies are available on the market for the removal of
particulate matter. They wide1y differ in regards to costs and efficiencies. These
factors, together with the characteristics of the parti des to be removed, are the major
factors to be taken into account and to be compared, in order to se1ect the most
appropriate technology (Caputo and Pelagagge, 1999).
l7
C. Kennes and M. C. Veiga (eds.), Bioreactors for Waste Gas Treatment, 17-46.
18 C Kennes el al.
2.1 GRA VITY SETTLING CHAMBERS
As a result ofthe increasingly stringent emission standards, gravity settling chambers

are no longer widely used, as they generally do not achieve high partide removal
efficiencics. They may be used as a pre-treatment step before ducting the waste air to
other more efficient devices, allowing the removal of relatively large parti des of
more than about 50 Ilm. The removal of smaller partides would reguire too large a
settling chamber. In gravity settling chambers (Figure 2.2) partides are removed by
gravity. By considering the set of forces acting on a given partide, namely the force
due to gravity and viscous [orce, one can easily calculate the terminal settling
velocity (V t) for spherical partides:
4 g d p ( Pp - Pg ) (2.1 )
3 Pg C D
where
g = gravitational acceleration (9.81 m/s 2 )

d p = partide diameter (m)
Pp = partide density (kg/m 3)
Pg = gas density (kg/m 3)
C D = drag coefficient related to the partide Reynolds number (Re p)
(2.2)
with Ilg representing the dynamic viscosity of air.
The numerical value of CD depends on the flow regime. In the laminar region
and turbulent region CD is basically constant and equal to (24/Re p) and 0.45,
respectively. Empirical equations are used for the transition region.
Assuming Pp » Pg, the theoretical collection efficiency (Il) representing the
fraction of partides that will be removed from the air stream can be calculated from:
7J
~)
= 1 - exp ( - V H
( g
= 1 - exp - 18
d/ PV HLJ
j1 g
p
(2.3)
where
)1g = gas dynamic viscosity (kg/m.s)

V = horizontal velocity ofthe air and particles through the chamber (m/s)
L = length ofthe chamber (m)
H = height ofthe chamber (m)
Gravity settlers are characterised by their very simple design, minimal

maintenance reguirements and very low pressure drop.
Non-Biological Technologies 19
L
~---------------------------------------------»
Air - - - .
inlet
, - - - . Treated
H:,, alf
,,,
\;. ~. .I\;. ~. . Hoppers
Figure 2.2. Gravity settling chamber.
2.2 CYCLONES
Although the collection mechanisms are to some extent similar to those in settling
chambers, in cyclones (Figure 2.3) a higher farce is applied on the particles than in
settling chambers. In inertial collectars ar cyclones. particulate matter is removed
from air by a centrifugal force. Its magnitude can be calculated from:
(2.4)
where
V tg 2 = tangential velocity of the particle at radius r

r = radius of rotation of the particle
mp = mass ofthe (spherical) particle
(2.5)
where V and Pp correspond respectively to particle volume (m 3) and partide density

(kg/m3 ).
The removal efficiency is an increasing function of particle size (r\ particle
density (pp) and the square ofthe tangential velocity (V tg ). since collection efficiency
is directly related to the farce Fc acting on the particle. It will decrease as the radius
of rotation increases. In designing cyclones. it should be taken into account that
optimising parameters that increase efficiency might result in increased pressurc
drop, leading to higher operation costs.
20 C. Kennes et al.
Treatcd air
Particlcs
Figure 2.3. Schematic of a cyclone.
Cyc10nes allow the removal of smaller partic1es than gravity settling

chambers, as the magnitude of the centrifugal force applied on the partic1es is higher
than the gravitational force. They are quite cheap and easy to maintain and operate.
2.3 FILTERS
2.3.1. Filtration technology

Filtration is used to remove particulate matter from dry air streams. Polluted air is
forced through a porous material, retaining partic1es, forming a cake on the surface of
the filter. Different types of materials may be used, inc1uding paper and, most
frequently, different types of fabrics. Fibrous filters are made of such materials as
cotton, fibreglass, dacron (polyester), nylon. teflon. wooL etc .. each presenting
different characteristics concerning their resistance to high temperatures, acids, bases
and solvents. Fabric filters are, in most cases, tubular or pocket shaped bags arranged
in parallel in a chamber called baghouse. Fabric filters are used in many applications
because of their high efficiency, allowing retention of small partic1es of less than 1
Ilm with a relatively low pressure drop. Both the fabric filter and the cake operate as
filtering layers. Pressure drop is the sum of pressure drops across the c1ean c10th and
across the cake, according to the equation of Darcy (Wark and Warner, 1981):
(2.6)
where
K = permeability constant (m2)

V = air velocity (m/h)
x = depth ofthe filter or cake layer (m)
]lg = dynamic viscosity (kg/mh)
Subscripts f and c refer, respectively, to the fabric filter and to the cake.
2.3.2. Cleaning methods

Baghouses are often classified into different categories according to the method of
cleaning (i.e., removal of collected particles) of the cloth: shaker, reverse air flow,
pulse-jet. Efficient cleaning usually takes only a few seconds or minutes and allows
maintaining optimal performance ofthe filters.
2.3.2.1. Shakers. Dust laden air is usually flowing upwards through the bag (Figure
2.4). Particles are retained on the inside surface of the filter. Cleaning of the filter
consists of shaking the bags from the top, every few hours, allowing filtered particles
to fall into the hopper located beneath the baghouse. If continuous use of the shaker
is required, the collector needs to be compartmented in order to allow operation of a
portion of the collector to be stopped for cleaning while polluted air remains flowing
through the other portion ofthe baghouse (Cheremisinoff, 1993).
Treated air
Shaker
(or blower)
Filler
medium
. . . ţ . . . . . ţ .... ţ ..ţ . . ţ.
Air
inJet
J
Partieles
Figure 2.4. Shaker and Reverse flow cleaning system, equipped either with a shaking device
(in the case of a shaker) or with a blower (Reverse flow system) for cleaning the cloth.
22 C. Kennes et al.
2.3.2.2. Reverseflow cleaning. As in the case ofshakers, dust laden air is fed upflow
and particles accumulate upon the inner surface of the bag. Cleaning consists in
introducing clean air into the bags at low pressure and in the opposite direction to the
original flow of polluted air (Figure 2.4). As in the case of shakers, reverse flow
filters can be operated either in continuous mode by taking only few bags off line for
cleaning or they may be stopped temporarily for cleaning ofthe complete system.
2.3.2.3. Pu/se jet. In pul se jet collectors dust laden air flows inwards through the
bags. Particles accumulate on the outside surface of the filter. In order to maintain
their shape, the bags are held open by wire cages (Figure 2.5). A pulse of reverse
flow compressed air is passed through the filters, every few minutes, for cleaning the
cloth.
Blowpipe
Treateu _
+_ Compressed
air ţ:::~;:;::::==~===;;~==:::;:;:;-;==:=C mr
Rctainer
Filter
medium
t t t
Air
inlet
-J
Particles
Figure 2.5. Pulse jet collector.
2.4 ELECTROSTATIC PRECIPITATORS
The operation mechanism of a typical wire and plate electrostatic precipitator is

represented in Figure 2.6, though several other similar designs are possible.
Electrostatic precipitators allow removal of small particles of a mean size down to
0.01 ~m, as well as quite larger particles of about 1 mm, with over 99% efficiency.
Although the mechanism is a little complex, it will briefly be summarised hereafter.
The dirty air stream flows in between two positively charged grounded plates, also
called collectors or collecting plates. A high DC voltage, reaching up to 50000 volts,

is applied to the wires suspended in between the plates. The wires or discharge
electrodes produce a corona generating electrons. They are held in place by
suspending weights at their bottom. The air stream flows through the electric field
created between wires and plates. The negative ions generated by the discharge
electrodes attach to the particles. They then start flowing in the direction of the
positively charged collecting plates, allowing their removal from the air stream.
Although in the above description the corona is negative, positive coronas might be
used as well. A complete electrostatic precipitator consists of a series of several
wires and plates mounted in a metallic frame. The particles are removed from the
collecting plates by mechanically wrapping their surface. Finally, particles fall in
hoppers.
Dischargc
electrode
~
Treated
~ air
Air~
inlel
~
Collecting
plate
Weight
Figure 2.6. Schematic of operation of an electrostatic precipitator.
The collection efficiency (ll) may be ca1culated with the Deutsch equation:
17 = 1 - exp
Vp A)
(-O (2.7)
where
Vp = migration velocity ofthe particle toward the collecting plate (m/s)
A = total area ofthe collecting plate (m2 )
Q = volumetric gas flow rate (m 3/s)
2.5 SCRUBBERS
In wet scrubbers, particles present in dust-Iaden air are transferred to a solvent,

usually water. Scrubbers can be used both for the removal of particles and volatile
pollutants. They are usually not suitable for particles larger than about 1 or even 10
24 C. Kennes et al.
Jlm, depending on the required efficiency and costs. Higher efficiencies are reached
when increasing the velocity of the fluids and decreasing droplet size. Different
designs are available. The most common ones are spray towers, cyclone scrubbers
and venturi scrubbers. However, scrubbers or absorption towers are best suited for
the removal of volatile compounds rather than for removing particles. Particulate
matter may lead to clogging problems. More data on absorption are also given in the
next section dealing with volatile compounds.
In spray towers, the solvent is introduced by means of a spray nozzle located
in the upper zone of the scrubber. Polluted air flows upwards and counter currently
(Figure 2.7). The particles may be captured by the water droplets through several
mechanisms, such as inertial impaction or diffusion. Other possible, though less
frequent, mechanisms include interception and electric attraction.
The collection efficiency of spray towers can be increased by increasing the
relative velocity between the water droplets and the air stream or particles in a
cyclone-scrubber, in which the velocity of the streams is increased by applying a
spinning motion to the fluids generating centrifugal forces. Such device is therefore
also suitable for slightly smaller particles than spray towers.
Trealed air
i
Waler ---. 0 - - - ' - - - - - - - - - 1 . - eli~~~lor
supply
Air ---.
inlel
Drain
water/sludge
Figure 2.7. Spray tower.
In venturi scrubbers polluted air is fed through aventuri throat, where it is

mixed with the scrubbing liquid. The gas stream may reach high velocities of about
60 to 120 mls. Such characteristics, combined with the small droplet size of the
atomised scrubbing liquid, allows reaching a relatively high efficiency. Particles as

small as 0.2 Ilm can be removed with high efficiency, though with relatively high
pressure drops (Calvert, 1977).
3. Removal of volatile compounds
Conventional non-biological alternatives for the removal of volatile compounds

include absorption, adsorption, thermal and catalytic oxidation (incineration) and
condensation, as well as few other innovative or less used methods, such as
membrane technology and UV oxidation. The different technblogies will be
explained hereafter. Although each one will be described in separate sections, it
should be mentioned that they are nowadays quite often combined to reach higher
overall removal efficiencies. Combining biological and non-biological processes is
also possible. The combination of a scrubber and a bioreactor has been used
successfully at industrial scale for the removal of volatile pollutants, as described in
Chapter 5, which is dedicated to bioscrubbers. Combining UV oxidation and
biological oxidation is also being studied (Chapter 7).
3.1 ABSORPTION
In the case of absorption, contaminated air is put in contact with a non-volatile liquid
phase, the purpose being the mass transfer of the contaminant from the gas phase to
the liquid phase. Maximisation of the driving force and high diffusion rates are
requisites in order to reach an efficient separation. Solubility of the compound to be
transferred to the other phase and a high gas-liquid interfacial area will enhance the
process. In some cases, reaction between components ofthe gas phase and the liquid
phase may also take place (Shi el al., 1996; Thomas and Vanderschuren, 2000).
Often, the main disadvantage of this technique is that the pollutant is
transferred to a new phase instead of being destroyed, meaning that the pollution
problem remains present. However, such a characteristic becomes an advantage
whenever the contaminant is a relatively valuable chemical and when it needs to be
recovered for eventual resale or reuse.
3.1.1. Key elements of absorption columns

One of the key steps in designing absorption towers is the correct choice of the liquid
phase to be fed to the column. Characteristics of the gas phase are imposed by the
polluted air stream released from the emission source. Whenever possible, the
scrubbing liquid will be water, since it presents many advantages, such as its
relativeIy low cost and the ease of obtaining it in high quantities. In any case, several
factors should be taken into account when seIecting the solvent. It should be cheap
and easy to obtain in relatively large quantities. Corrosive, toxic, viscous, flammable,
and/or highly volatile chemicals should be avoided. A high solubility of a gas in the
liquid phase is required and is often reached by sclecting a liquid solvent with as
similar a molecular structure as the volatile compound to be absorbed.
Absorption of volatile pollutants can either be achieved by a continuous
contact between phases in which case packed columns are used (Figure 2.8), or by
equilibrium-staged contacts when plate columns are selected. Packed columns are
most frequently used. In case of packed columns, the packing material is a key
26 C. Kennes el al.
Figure 2.8. Thr~c-pha s c absorptioll systelll pa~kcd with Rasehig rillgs IIsed for the treatillent
of odours from a waste water treatillent plant (Photo courtesy of Plastoquilllica).
parameter. Packings used in absorption towers are produced from various materials,
sizes and shapes. They may either be stacked in regular geometric patterns or
dumped at random. Some examples of common tower packings dumped at random
are PalI rings, Raschig rings, Hiflow rings and Intalox saddles made either of
ceramic, metal or plastic. The relative void fraction reached with such packings often
exceeds 0.8 or even 0.9. Honeycomb tubes as well as grids, sheets or spirals are
typical forms used in arranged packings.
3.1.2. Design of absorplion columns
3.1.2.2. Packed lowers. In packed towers (Figure 2.9) both phases flow in continuous
mode through the system reaching an intimate contact between the gas and the liquid
phase. The absorption tower is filled with a packing material characterised by its
high superficial area. The liquid phase flows through the column, reaching a large
gas- Iiquid superficial contact area thanks, among others, to the presence of the
packing material.
Counter-current, co-current or cross flows may be used, the first alternative
being the most popular one. In counter-current operation (Figure 2.1O-A) the liquid
phase is flowing downwards while the contaminated air is flowing counter-currently.
In such case a high relatively constant driving force is reached. Contrarily, in co-
current mode (Figure 2.10-8), the gas phase is put in contact with fresh liquid at the
entrance ofthe column. While passing through the absorption tower, the liquid phase
in contact with the contaminated air gets progressively more and more concentrated
in pollutant, with a concomitant decrease in the driving force . In case of cross flow
(Figure 2.1O-C), the gas phase flows horizontalIy while the liquid phase flows
vertically and downwards. Although less frequently used, this configuration does
sometimes present some advantages, such as a lower pressure drop and a lower
consumption of scrubbing liquid.
Treated air
(G" y,)
Water inlet
~~~~*.- (L2. X2)
. - Airinlet
Water oudet .- r (G" y,)
(L"xl) --------'
Figure 2.9. Packed tower.
Treated Water Air Water

air supply inlet suppIy
Water
t + ++
suppIy
~
PACKED PACKED PACKED Treated ~
COLUMN COLUMN ~ COLUMN air
iniet
t
Air
+
Drain Treated Drain
Drain
water
inlet water air water
A B c
Figure 2.10. Mode of operatiOl1 of absorption columns,
A key parameter in designing packed columns is the choice of the packing

material. Several factors should be considered when selecting the packing materiaL
Good packings are characterised by a high porosity, low pressure drop, negligible
corrosivity, high wetted area per unit volume and low cost.
Design equationsfor packed columns. A mass balance equation applied to a pollutant

flowing through a counter-current column leads to the following expression (Figure
2.9):
(2.8)
28 C Kennes el al.
where
Oi = Total molar flow rate of gas (air + pollutant) per unit cross sectional area of the
column (kmol/s.m2)
Li = Total molar flow rate ofliquid (water + pollutant) per unit cross sectional area of
the column (kmol/s.m2 )
Yi = mole fraction ofpollutant in the gas phase
Xi = mole fraction of pollutant in the liquid phase
Subscripts 1 and 2 refer, respectively, to the inlet and the outlet ofthe column for the
gas phase and to the outlet and the inlet ofthe liquid phase (Figure 2.9).
O and L vary along the height of the column, since there is a continuous mass
transfer of pollutants from one phase to the other. A mass balance equation with
constant molar flow rates can be written by modifying the above equation:
(2.9)
where
Os = molar flow rate of pollutant-free air per unit cross sectional area of column
(kmol/s.m2)
Ls = molar flow rate of pollutant-free liquid per unit cross sectional area of column
(kmol/s.m2)
Yi = moles ofpollutant per mole ofpollutant-free air in the gas phase
Xi = moles ofpollutant per mole ofpollutant-free liquid in the liquid phase
y, -y_l- (2.10)
1 - YI
~ (2.11)
Xi
1 - XI
Dropping the subscript 2 and solving for Y yields the equation of thc
operaling line valid both for packed towers and plate columns:
Y = Y - h (X - X) (2.12)
I Os I
A mass balance for a differential column height dz yields:
Os dY = Ls dX (2.13 )
The equation of the rate of mass transfer of the pollutant from one phase to
another (Figure 2.11) is:
(2.14)
or (2.15)
where
N A = molar flux of pollutant per unit time and area (kmolls.m 2)
k G (or ky) = individual gas phase mass transfer coefficient (kmolls.m 2 L'1p units of
concentration)
h (or k x) = individualliquid phase mass transfer coefficient (kmol/s.m 2 .L'lC units of
concentration)
PG (or YG) = partial pressure (or mole fraction) ofpollutant in the bulk gas phase
Pi (or Yi) = partial pressure (or mole fraction) ofpollutant at the interface
C L or XL = pollutant concentration (or mole fraction) in the bulk liquid
Ci or Xi = pollutant concentration (or mole fraction) at the interface
As the concentrations (i.e., Pi or Ci) and mole fractions (i.e., Yi and Xi) at the
interface level are usually hard to estimate, it is more convenient to use equilibrium
concentrations and overall mass transfer coefficients. The latter are obtained
experimentall y.
N A = KG (PG - p*) = K L (C* - C L ) (2.16)
or NA = Ky (ye; - y*) = Kx (x* - XL) (2.17)
where
KG or Ky = overall mass transfer coefficient for the gas phase (kmolls.m 2 .L'lp units of
concentration)
Mass transfer, N A
Gas phasc Liquid phase
Interface
Figure 2.11. Concentration profiles ofpollutant A in the gas phase and in the liquid phase.
30 C. Kennes el al.
K L or Kx = overall mass transfer eoeffieient for the liquid phase (kmol/s.m 2 6C units
of eoneentration)
p* = partial pressure of pollutant in equilibrium with the liquid phase eoneentration
CL
C* = pollutant eoneentration in equilibrium with pressure po in the gas phase
y* = equilibrium mole fraetion referred to the gas phase
x* = equilibrium mole fraetion referred to the liquid phase
The relationship between individual and overall mass transfer eoeffieients ean
easily by shown by eombining the above equations,
Kx kx
+
[:y (;: ~ :J] k,
+
mk y
(2.18)
and
Ky ky
+ [ ~k, ( y~ ~ y)]
XL XI
=
ky
+
m
k,
(2.19)
where m is the slope of the equilibrium curve, a eurve representing pollutant

equilibrium data in the gas phase and liquid phase.
Considering onee again the differential eolumn height dz with the interfacial
contact area dA for mass transfer and reealling the mass transfer expression (N A )
(Equation 2.17),
N A dA = Ky (y - y*) dA (2.20)
in whieh subseript G has been deleted from the parameter Yo for simplifying the
expression.
The mass transfer area (dA) is usually not known and ean advantageously be
substituted for another parameter (a) representing the interfaeial area per unit volume
of eolumn. Then, if S is the eross-seetional area of the tower,
dA=aSdz (2.21 )
and
N A a S dz = Ky a (y - y*) S dz (2.22)
The height of the eolumn ean be ealculated by eombining the mass balanee
equation and the mass transfer rate equation:
- G dy = Ky a (y - y*) S dz (2.23)
Whenever relatively low pollutant concentrations are present in the gas phase,
(O/K y a S) will basically be constant and the total height (Z) of the system can be
estimated according to:
(2.24)
often written as follow:
Z = (HTU) (NTU) (2.25)
calling HTU the height of one transfer unit and NTU the total number of transfer
units needed for reaching the transfer of a given amount of pollutant from the gas
phase to the liquid phase.
In the case of dealing with air contaminated with high pollutant

concentrations, O will significantly vary and the above equations cannot be
integrated direct/y. The following expression needs to be considered for integration,
instead of(O dy):
(2.26)
Leading finally to the following expression:
z f,l'( 0(1 - y)lm dy

k y • aS (l - y) (y - y')
J (2.27)
with
(1 - y') - CI - y) (2.28)
(1 - y)lm
1-Y'
In (- -)
1 - Y
3.1.2.3. Plate columns. In plate columns (Figure 2.12) the phases flow in opposite
directions and the contact is intermittent. The liquid flows downwards from one plate
to another. Different plate designs are available. Each plate or stage allows an
intimate contact between the liquid and the gas phase, leading to diffusion and mass
transfer of the contaminant prior to phase separation. Calculation of the number of
theoretical equilibrium stages is an important design parameter which depends,
basically, on the composition of each phase and on equilibrium data. Since
equilibrium is not yet reached when both phases leave a given plate, the actual
number of stages should, afterwards, be estimated on the basis of a parameter known
as efficiency. Another design parameter is the column diameter. which is strongly
dependant on liquid and gas flows.
32 C. Kennes el al.
Clean
air outlet
i
..rf--------- Liquid in
1- -----------
----------1
1-----------------------
-----------------------1
1~~:=C-
Polluted
air intel
u,"WO"' _~)--:r
Figure 2.12. Schematic of a plate column.
3.2 ADSORPTION
In adsorption systems, the phases considered for mass transfer are a fluid and a solid.
As a matter of fact, when dealing with air pollution problems the fluid will be a gas.
Most solids are characterised by their capability to adsorb volatile compounds upon
their surface. However, the extent of adsorption does significantly vary from one
solid to another. Two adsorption mechanisms are possible, physical adsorption,
involving only weak forces of van der Waals and chemical adsorption or
chemisorption, involving a chemi cal reaction and formation of new chemi cal bonds
between the gas molecules and the surface of the adsorbent.
Physical adsorption is a reversible process because the forces between the
solid surface and the adsorbate are weak. The adsorbed volatile compound can thus
be desorbed and recovered, usually by thermic treatment. In the case of
chemisorption, true strong chemi cal bonds are formed and regeneration is more
difficult, leading most often to a modification of the chemi cal structure of the
adsorbent. Adsorption is an exothermic process, therefore adsorbed gas concentration
decreases as temperature is increased, when working at a constant pressure.
3.2.1. Nature of adsorbents

Several solids are suitable for adsorption. Factors to be considered when selecting
one of them include adsorptive capacity and selectivity, which depend on, among
others, properties such as porosity, pore size, surface area and bulk density. The most
widely used adsorbent is activated carbon (Ruhl, 1993). Other ones are activated
aluminas, silica gel, and molecular sieves. Table 2.1 summarises some typical
characteristics of these adsorbents. The data are only indicative and some variations
are possible, as a result of the wide variety of adsorbents belonging to the same
category. In some cases the adsorbent may be impregnated with a specific compound
which will react with the contaminant to be removed and enhance the removal
process, as in the case of NaOH impregnated activated carbon reacting with H2S
polluted air (Van Stone and Brooks, 1996). Adsorption is an efficient process only at
relatively low or moderate temperatures. At higher temperatures desorption will take
place or even destruction ofthe adsorbent (Table 2.1).
Activated carbon is obtained through thermic treatment of wood, nutshells,
carbon, peat, coal or other similar products. The porosity is significantly increased
after thermic treatment. Activated carbon, sold either in granular or in powder form,
is mainly used for the removal of organic compounds from contaminated air, the
process being suitable even in the presence of water. Most of the adsorption surface
is located in the pores (Figure 2.13). More information on activated carbon is also
given in Chapter 3.
Other adsorbents do usually preferentially adsorb water molecules. This is of
interest for drying air streams, though it becomes a major drawback when one wants
to remove organic contaminants from emis sion sources. In the latter case water must
be removed first. Activated aluminas are hydrated aluminium oxides obtained by
heat treatment. Silica gels are often used to remove sulphur. They are obtained by
neutralising sodium silicates with an acid solution. Salts are removed by washing
before drying and toasting the gel. Finally, molecular sieves or zeolites are
aluminosilicates possessing a crystalline and defined framework structure.
Adsorption does usually take place at the level of the cavity of each crystalline unit.
The pores leading to the cavity play a size-selective role for the molecules to be
Table 2.1. Major characteristics of adsorbents.
Parameter Activated carbon Activated Silica gel Molecular sieves

aluminas
Porosity (-) 0.25-0.8 0.25-0.6 0.35-0.5 0.3-0.55
Pore diameter
(10- 10 m) 10-100 20-130 20-40 3-20
Surface area
(m2/g) 500-1600 100-400 200-900 800-1000
Bulk density
(kg/m 3) 150-550 600-930 450-800 500-850
Thermal stability
(max. TO, 0c) 150 500 400 600
Desorption ~ (0C) 120 ± 20 220 ± 20 175 ± 50 250 ± 50

34 C. Kennes el al.
• • •• • • • •
Figure 2.13. Porous activated carbon granule.
adsorbed. More recently new or improved adsorbents have been used, among others
synthetic polymeric adsorbents (Anonymous, 1995) carbon fibers (Kenson, 1992),
catalytic/adsorptive carbon (Van Stone and Brooks, 1996), etc.
3.2.2. Adsorption equipment

In the case of absorption the gas is continuously removed from the column.
Contrarily, in adsorption systems. the pollutants remain adsorbed onto the solid
inside the column. The adsorbent will thus finally get exhausted. Whenever possible
adsorption processes are combined with a desorption process which aims at
recovering the volatile solvent, thus reducing or at least compensating overall costs
(Figure 2.14). The adsorbed solvent may be recovered in a vapour phase afier
steaming. The latter is afierwards liquefied in a condenser. Since the organic solvent
and the water phase are usually characterised by different densities, both phases can
easily be separated. If at least two adsorption columns are connected in series, one
can be regenerated while the other is still operating. Regenerated wet carbon is not
suitable for adsorption and a drying step is required before reuse (Gu et al., 1999).
3.2.3. Adsorplion isotherms

In the case of solid-fluid systems, an equilibrium will be reached between the
adsorbate in the fluid, i.e. air, and the amount of pollutant adsorbed onto the solid.
Adsorption isotherms represent such an equilibrium or, in other words, the
adsorption capacity of a given adsorbent in presence of a specific adsorbate or
contaminant. Such information is plotted in diagrams representing the amount of
solvent adsorbed (X) onto a given amount of adsorbate (M) against the solvent
concentration (C) or mass of adsorbate per gas volume (C) (Figure 2.15). The most
common adsorption isotherms are mentioned below. They are useful for calculating
the amount of solid required for adsorbing a given mass of contaminant.
Condenser Air inler Condenser

r--~===;""]
Activated Saturared
carbon bed activated
carbon bed
..-
Airout Water Vapor in
Decanter Decanter
(a) (b)
Figure 2.14. Schematic of an activated carbon contactor with steam recovery of the solvent.
The Freundlich isotherm IS mathematically represented by the empirica!

equation,
x kC" (2.29)
q= M
where k and n are constants that can be obtained experimentally by measuring and
p!otting severa! values of q against C.
q=X/M
c
Figure 2.15. Typical adsorptioll isotherm.
36 C. Kennes el al.
Another mathematical expression is the Langmuir isotherm. It is deduced

from some theoretical considerations and assumes that adsorption is reversible and
that it reaches equilibrium conditions. It also assumes that a monolayer of adsorbate
is present on the solid surface. The following equation is then used:
x abC
(2.30)
q = M
(1+bC)
The empirical constants a and b can be evaluated by linearizing the equation

and by plotting (l/q) versus (1/C).
Many other equations may be useful in some specific cases for describing
some ofthe many possible isotherm shapes. The BET (Brunauer, Emmet and Teller)
equation is another example. Most isotherm equations are empirical or involve
numerous assumptions. Finding the equation that best agrees with the observed
isotherm is useful for modelling purposes and for predicting performance and
operation characteristics of adsorption beds.
3.2.4. Breakthrough curves

When contaminated air is fed to an adsorption column, the pollutant gets first
adsorbed onto the adsorbent present at the entrance ofthe column. Afterwards it will
gradually be adsorbed onto the fresh virgin zone it will encounter deeper in the bed.
The zone where mass transfer is taking place is called the mass transfer zone (MTZ).
After a given operation period, the MTZ reaches the outlet of the column (Figure
Air inlet
Ca
Saturated_
~
zone
MTZ
C/Co
o
Volume of effluent treated (V)
Figure 2.16. Typical breakthrough curve and movement ofthe MTZ.
2.16) and part ofthe pollutant will then progressively appear in the effiuent leaving
the column. The breakthrough point is reached when a specified percentage of the
influent concentration is detected in the column effiuent. A plot of the variation with
time of pollutant concentration in the effiuent leaving the bed is called a
breakthrough curve (Figure 2.16). Column exhaustion is reached shortly after
reaching the breakthrough point.
The time elapsed when reaching the breakthrough point or the service time
(ts) can be calculated from the breakthrough curve according to the following
equation:
(2.31)
Service time is a key parameter in sizing adsorption columns.
The time corresponding to exhaustion is:
(2.32)
Total height of the bed (HE) and bed height corresponding to breakthrough
(Hs) are related through the expression:
H
S
= (.!.s..)
tE
H
E
(2.33)
The height of the mass transfer zone is related to the shape of the adsorption
isotherm. It corresponds to the following equation:
MTZ = HE - Hs = H E [1 - ( ~:)] (2.34)
The service time is also equal to:
HE - MTZ Hs (2.35)
t = =-
s v v
where v is the velocity of the MTZ flowing down the column. Such velocity is
usually quite lower than the linear gas velocity of air fed to the system.
3.3 INCINERATION
Incineration is a combustion process in which most often, though not exclusively,

under optimal conditions hydrocarbons are converted to carbon dioxide and water. It
is a process in which volatile organic compounds are combined with oxygen, with
38 C Kennes et al.
the concomitant rele ase of heat. If only partial combustion is reached, other
contaminants may be formed, such as organic acids, aldehydes, ketones, carbon
monoxide, etc.. In the case of halogenated pollutants an sulphur-compounds,
potential combustion products are S02, HCI and phosgene, among others
(Klinghoffer and Rossin, 1992; Freidel el al., 1993). It is worth underlining that some
combustion products might be more toxic than the original pollutant, as is the case
for phosgene generated during trichloroethylene incineration (Nimlos et al., 1993).
Reaching complete combustion is not always easy and depends on several
design factors, such as temperature, residence time (i.e., space velocity) and
turbulence (i.e., mixing) (Chu and Wu, 1998; Wang and Chou, 2000). A sufficiently
high temperature is required in order to allow fast ignition and combustion.
Combustion will start when the ignition temperature is reached. The addition of heat
is therefore needed. Once combustion has started, the oxidation process itself will
produce heat. If heat production is high enough combustion will go ono Such a
process is called direct flame incineration or direct combustion. Concentration of air
contaminants is often relatively low, meaning that heat production is also quite low.
If not enough heat is generated through oxidation, the addition of an external fuel
source is required and the process is called thermal incineration. In order to avoid
addition of fuel, a third alternative consists of using a catalyst which increases
reaction kinetics and allows lowering incineration temperature. Such a process is
called catalytic incineration. It allows reduc ing costs originating from the use of fuel.
However, catalysts are expensive and global costs will thus not necessarily be lower.
Although many VOCs are destroyed at temperatures around 600-700 aC,
thermal incinerators often need to be operated at temperatures close to 1000 aC, to
make sure combustion is as complete as possible and to avoid, whenever possible,
the formation of toxic or carcinogenic incineration products. Destruction of
halogenated compounds usually requires higher temperatures than the temperatures
used for non-halogenated aliphatic or aromatic molecules (Pope et al., 1976).
The level of completeness of oxidation also rei ies on residence time.
Residence time (t) is related to the volumetric flow rate ofthe feed gas (Q) and to the
volume ofthe oxidation chamber (V) through the following equation:
V (2.36)
Q
The volume of the oxidation chamber will be smaller when operating at

higher flow rates. If the reacting mixture passes too quickly through the incinerator,
oxidation will not be complete. Residence time is often less than one second and is
shorter for catalytic ineinerators than for thermal incinerators. Residence time may
be lowered whenever operating temperature is increased, since this will speed up the
reaction. Investment costs will then also be reduced, owing to the possibility ofusing
smaller devices. However fuel costs and operating costs will be higher.
The amount of oxygen (air) required for oxidation of a given amount of
organic contaminant is another key factor which should properly be calculated.
Oxygen concentrations foreseen by the stoichiometric equation are usually slightly
increased in practice, in order to assure complete or ne ar complete oxidation. As an
example, if toluene contaminated air is being incinerated, 9 moles of oxygen will be
required to burn one mole of toluene, according to the following stoichiometric
equation when air is used as the oxygen source:
Efficient mixing ensuring a good contact between air, volatile contaminants

and fuel is necessary for optimal oxidation. Other parameters to be considered in
combustion processes are discussed below. They are specific of a given process, such
as the amount of fuel to be added or the nature of the catalyst. Several thermal and
catalytic incinerators have been used for treating industrial waste gases (Figures 2.17
and 2.18).
3.3.1 Thermal incinerators

Figure 2.19 shows a sketch of a thermal incinerator which is basically composed of a
combustion chamber equipped with a burner and the inlet ports for feeding waste air,
fuel (usually natural gas or gasoil) if necessary and oxygen (combustion air).
Addition of oxygen or air is not required whenever the waste air already contains
enough oxygen for complete oxidation of the pollutants. The external part of the
walls of the incinerator are made out of steel while the inner part is formed by a
refractory material of up to 20 cm width, minimising heat loss. The refractory
material also protects steel from high temperatures and corrosive materials. Reat
recovery, either recuperative or regenerative, is often advantageously used mainly for
preheating the influent air stream with the hotter c1ean effluent air (Klobucar, 1995).
3.3.2. Catalytic incinerators

As already mentioned, (thermal) incineration temperatures as high as 1000 °c are
often required. When a catalyst is added to the process (Figure 2.20), shorter
residence times and temperatures around 200-500 °c can be used while maintaining
equally fast oxidation reactions as with thermal incinerators. Optimal temperature
also depends on the nature ofthe pollutants. Reat generated during the reaction and
major end-products are also the same in both processes. In catalytic incinerators,
since lower temperatures are used, concentrations of by-products such as NO x are
often lower (Prasad et al., 1984). The design of thermal and cata1ytic incinerators is
very similar except that in the latter a smaller layer of refractory material is needed as
a result of the lower operation temperature. External energy recovery is not always
interesting from an economical viewpoint since out1et temperatures are quite a bit
lower compared to thermal combustion. Nevertheless, internal heat recovery is often
used for preheating the contaminated feed air and for fuel savings.
3.3.2.1 Characteristics of the catalysts. A catalyst is a substance that increases the

rate of reaction though it is not consumed during the process. The catalyst is located
at a given distance from the burner. This ensures that the air flow possesses a
relatively uniform composition and temperature when reaching the catalyst. Two
major types of catalysts may be used in incinerators, namely metal oxides and noble
metals. Noble metals are chosen among Ag, Au, Pd or Pt. In the case of selecting
oxides the metals are most often Co, Cr, Cu, Fe, Mn, Ni or V. The catalyst is coated
onto a solid surface in such a way as to obtain as high as possible a surface area.
3.3.2.2. Deactivation of catalysts. Although catalysts are not consumed during

oxidation, their activity decreases with operation time. Such a deactivation process
40 C. Kennes el al.
Figure 2.17. Thermal oxidation chamber used for treating 12000 m3/h of toluene
contaminated air (Courtesy ofKalfrisa).
Figure 2.18. Catalytic oxidation system for the treatment of industrial waste gases (Courtesy
of Kalfrisa).
Polluted
air inlet
Fuel ~ ~ Air
andair ~ ~ outlet
Figure 2.19. Schematic of a thermal incineratoL
Preheat
Catalyst
bumer
Polluted
air inlet
----. Air
outlet
Figure 2.20. Schematic of a catalytic incinerator.
takes place for different reasons, among which are thermal aging, clogging and
poisoning. Therefore catalysts need to be replaced periodically.
With thermal aging, the active surface area decreases with time as a result of
exposure of the catalyst to the reJatively high temperature of the gas. This may be
due to several factors, such as changes in surface structure owing to crystal growth
and narrowing or even closing of catalyst pores. Temperatures above 650 De are not
recommended (Siebert et al., 1984). Oxidation is an exothermic process and the
temperature reached in the incineration chamber should be checked carefully in order
to avoid destruction of the catalyst by heat produced dur ing operation. Clogging
occurs when the waste stream is contaminated with particles. Particulate matter
should therefore first be eliminated in order to avoid or slow down catalyst
deactivation. Poisoning results from the reaction of compounds, such as arsenic,
lead, zinc, phosphorus, sulphur, etc. at the level of active sites of the catalyst. Stable
products are then formed. impeding the reaction between air, volatile contaminants
42 C. Kennes et al.
and the catalyst. Exposure of catalysts to halogenated air streams is not

recommended, though studies have been recently undertaken in order to develop
catalyst based systems able to cope with halogenated gases.
3.4 CONDENSATION
Condensation is the conversion of a gas or a vapour into a liquid, when its partial
pressure is equal to the vapour pressure of the system. It is achieved either by
pressure increase or by lowering the temperature. For economical reasons the latter is
usually preferred. Combination of both temperature and pressure variations is also
possible. By pressurisation of air the molecules are brought closer together, while
lowering the temperature reduces the kinetic energy ofthe molecules.
Condensers are not very expensive and allow recovery and reuse of valuable
compounds. Cooling is often achieved with water or air, though other substances
could be used as well, depending on the characteristics of the waste gas and the
desired efficiency. However, temperatures below 30-40 aC are most oilen not
reached unless admitting increased costs. meaning that highly volatile pollutants will
most often hardly be removed. Therefore this technology is, in most cases, simply
used as a pre-treatment process useful mainly for VOC concentrations gre ater than
5000 ppmv and relatively low mass flow rates. Efficiency will drop at lower
contaminant concentrations or increased flow rates. In case VOCs get dissolved in
the coolant during condensation, pollutants will then be transferred from one phase to
another as was also the case in absorption and adsorption processes. The resulting
waste water then needs to be treated as well.
3.4.1. Equipment and design

The two most common designs of condensers are the direct-contact condenser,
where intimate contact is reached between the coolant and the waste stream, and the
surface condenser, where a physical separation exists between both fluids.
3.4.1.1. Direct contact condensers. This is the cheapest of both alternatives. It

requires minimum investment and maintenance. Upon mixing, the coolant may either
be transformed to a gas or it may remain in liquid form. If the volatile pollutants and
the coolant are both liquids with different densities they can then be separated by
phase separation. In case of a gas-liquid mixture, separation is possible by venting
the coolant. Most often direct contact condensation is accompanied by absorption of
the VOC into the coolant liquid phase following a similar mechanism as in spray
towers. One key parameter to be evaluated when designing contact condensers is the
required amount of coolant.
3.4.1.2. Surface condensers. A surface condenser is an heat exchanger. Several

designs of heat exchangers have been developed, such as the double pipe process,
double plate, cross flow, shell and tubes, etc. The most common design for air
pollution control is the shell and tubes condenser (Figure 2.21). In such a system,
several tubes are placed in parallel in a single shell. One fluid passes through the
tubes and the other one is fed to the shell.
The heat exchange surface area is a key design factor. It can be obtained from
the heat exchange equation:
Trcated air Coolant inlet
Coolant outlct Air inlet
Figure 2.21. Schematic of a surface condenser.
Q=U A T m (2.37)
Meaning that the surface area can easily be calculated from:
(2.38)
where
A = contact area (m2)

Q = heat transfer rate (W)
U = overall heat transfer coefficient (W/m 2 .K)
T m = mean temperature difference (K)
The overall heat transfer coefficient englobes the different individual heat
transfer coefficients and resistances. Conceming Tm, there are severa! means of
calculating the mean temperature difference in paralle! and counter-current shell and
tubes condensers. When the overall heat transfer coefficient remains constant and
when the heat capacity of each fluid is constant, the log mean temperature (Tim) may
be used:
(T2 -TI )
(2.39)
In(~)
where
44 C. Kennes et al.
TI = lowest temperature difference between the hot and the cold fluid
T2 = highest temperature difference between the hot and the cold fluid
If condensation is not linear and if the temperature range between hot and
cold fluid is large, lets say more than 20 aC, the log mean temperature should not be
used.
3.5 INNOV A TIVE PROCESSES
3.5.1. UVoxidation
In UV oxidation technology, a combination of oxygen containing oxidants, such as
ozone or peroxide, and UV light are used for the photochemical stimulation of
oxidation reactions, allowing the conversion of halogenated and non-halogenated
aliphatic or aromatic organic pollutants mainly into water and carbon dioxide. A key
design parameter is the UV light frequeney (200-280 nm) which is selected
aeeording to the nature of the VOC pollutants present in the waste air. Other faetors
playing a role in UV oxidation efticieney are waste air flow rate and temperature.
The reaction may take place in bateh or in continuous photoreactors. As aIready
mentioned earlier for ineinerators (Equation 2.36), flow rate is related to reaetor
residenee time through the following equation:
V=Qt (2.40)
where t is the reactor residence time, Q is the volumetric flow rate of the feed gas
and V is the active volume ofthe reactor.
Shorter residence times or smaller reactor volumes may be used by increasing

the intensity of UV radiation at the expense of increasing treatment costs. Global
costs wilI also be inereased when higher removal efficiencies are required. UV
oxidation is a re1atively new technology for which removal efficiencies still need to
be optimised. Only a very few pilot-scale and industrial-scale results have been
reported (Ayer and Darvin, 1995). The process is usuaIly followed by a post-
treatment step aimed at increasing VOC removal levels. Recent1y the possibility of
using UV oxidation combined with biological processes (bioreactors) has been
tested, as well, for solving air pollution problems, though fuIl-scale data are not yet
available (Chapter 7).
3.5.2. Membrane technology

Semi-permeable membranes have been used for several decades and are still used for
treatment of contaminated waters. Their use for the removal of VOCs from
contaminated air is more recent. It is therefore considered as a relative1y new Of
emerging technology. Compared to other technologies it is quite effective for the
removal of some chlorinated compounds. It is best suited for relative1y high pollutant
concentrations, typically > 10000 ppmv.
As shown schematicaIly in Figure 2.22, in membrane processes contaminated
air is tirst compressed and then fed to the separation unit containing the semi-
permeable membrane (Baker et al., 1992). If required, several membranes may be
placed in series. The membrane is permeable to the pollutants but not to air. A
Z
Treated air
Condenser
t
Air inlet
--.----+.\d~.
Compressor
~ I ~ I- - - - ' . 1 Membrane
unit
L;": vac 1P<~~"

nWJ Vacuum
•
pump
Recycling
VOC
Figure 2.22. Schematic of a membrane process.
vacuum pump (Figure 2.22) creating a lower pressure on one side of the membrane
respective to the other enhances the separation process. Parameters goveming
treatment efficiency include waste gas flow rate and temperature, as well as VOC
concentration. Studies on the use of porous membranes in combination with gas
phase bioreactor technology have also been undertaken recently (Chapter 6) though
full-scale applications have not been reported so far.
References
Anonymous. 1995. Newly improved synthetic absorbents and adsorbents increase emission control
options. The Air Pollution Consultant 5: 1.6-1.12.
Ayer, J. and Darvin, C.H. Cost-effective VOC emission control strategies for military, aerospace, and
industrial paint spary booth operations: combining inproved ventilation systems with innovative, low
cost emission control technologies. In: Proceedings of the 88 1h Annual Meeting & Exhibition of the
Air & Waste Manage. Assoc., 1995, San Antonio, Texas.
Baker, R.W., Kaschemekat, J., Wijmans, l.G. and Simmons. V.L. Membrane vapor separation
systems for the recovery of VOCs. In: Proceedings of the 85 1h Annual Meeting & Exhibition of the
Air & Waste Manage. Assoc., 1992, Kansas City, Missouri.
Calvert, S. 1977. How to choose a particulate scrubber. Chem. Engin. 29: 54-68.
Caputo, A. and Pelagagge, P.M. 1999. Cost-effectiveness analysis of waste gas treatment plants for
the glass industry. J. Air Waste Manage. Assoc. 49: 1456--1462.
Cheremisinoff, P.N. 1993. Air pollution control and design for industry. Marcel Dekker, New York.
Chu, H. and Wu, L.W. 1998. The catalytic incineration of ethyl mercaptan over a Mn02/Fe203
cata1yst. J. Environ. Sci. Health. A33: 1119-1148.
46 C Kennes el al.
Freidel. I.M., Frost. A.C.. Herbert. K.J .. Meyer. F.J. and Summer, J.C. 1993. New catalyst
technologies for the destruction of halogenated hydrocarbons and volatile organics. Catalysis Today.
17: 367-382.
Gu, J .• Faqir, N.M. and Bart, H.-J. 1999. Orying of an activated carbon column after steam
regeneration. Chem. Eng. Technol. 22: 859-864.
Kenson, R.E. Recovery and recycling of chlorinated solvents from industrial air emlsslons. In:
Proceedings of the 85'h Annual Meeting & Exhibition of the Air & Waste Manage. Assoc., 1992,
Kansas City, Missouri.
Klinghoffer, A.A. and Rossin, J.A. 1992. Catalytic oxidation ofchloroacetonitrile over 1% platinium
alumina catalyst. Ind. Eng. Chem. Res. 31: 481--486.
Klobucar, J.M. 1995. Choose the best heat-recovery method for thermal oxidizers. Chem. Eng. Prog.
91: 57--63.
Nimlos, M.R., Jacoby, W.A., Blake. O.M. and Miine. T.A. 1993. Direct mass spectrometric studies of
the destruction of hazardous wastes. 2. Gas-phase photocatalytic oxidation of trichloroethylene over
Ti0 2: products and mechanisms. Envrion. Sci. Technol. 27: 732-740.
Pope, O., Walker, O.S. and Moss, R.L. 1976. Evaluation of cobalt for the oxidation of low
concentrations of organic compounds in air. Atmos. Environ. 10: 951-959.
Prasad, R., Kennedy, L.A. and Ruckenstein. E. 1984. Catalytic combustion. CataI. Rev.-Sci. Eng. 26:
1-58.
Ruhl, M.J. 1993. Recover VOCs via adsorption on activated carbon. Chem. Eng. Prog. 89: 37--41.
Siebert, P.C., Meardon, K.R. and Serne, J.C. 1984. Emission controls in polymer production. Chem.
Eng. Prog. 80: 68-76.
Thomas, O. and Vanderschuren, J. 2000. Nitrogen oxides scrubbing with alkaline solutions. Chem.
Eng. Technol. 23: 449--455.
Van Stone, G. and Brooks, D. 1996. Carbon clean. Water Environ. Technol. February: 40--43.
Wang, J.B. and Chou, M.-S. 2000. Kinetics of catalytic oxidation of benzene, n-hexane, and emission
gas from a refinery oil/water separator over a chromium oxide catalyst. J. Air Waste Manage Assoc.
50: 227-233.
Wark, K and Warner, C.F. 1981. Air pollution, its origin and control. Harper & Row Publishers, New
York.
CHAPTER 3 CONVENTIONAL BIOFILTERS
Christian KENNES and Maria C. VEIGA
1. Introduction
Most of the recent literature dealing with biological waste gas treatment indicates
that it is not really a new technology, which is indeed basically correct, taking into
account that biofiltration has been applied for several decades, though at small scale
and with low flow rates, mainly for the removal of volatile pollutants emitted from
waste water treatment plants and composting processes (Pomeroy. 1957; Ottengraf
and Diks, 1992). Biofiltration was initially mainly exc1usive1y used for the removal
of odours by means of soil biofilters designed and set up according to empirical
considerations. Applications have recently been extended to a much wider range of
compounds, mainly volatile organic compounds, over the last two decades.
Development of scientific studies, engineering concepts and non-empirical design of
biofilters is also quite recent, as well as the search for new support media and new
bioreactor configurations, as will also be described in following chapters.
2. Design of biofilters
2.1 CONVENTIONAL BIOFILTER FUNDAMENTALS
It is sometimes considered that the main difference between a conventional biofilter

and a biotrickling filter (or trickling biofilter) lies in the carrier material which
usually is natural, often organic, in conventional biofilters and inert or synthetic in
biotrickling filters. However, the main and most relevant difference is that for the
latter re actor an aqueous phase is trickling over the filter bed, usually in a continuous
mode, while this is not the case in conventional biofilters. The aqueous phase fed to
the biotrickling filter is a medium containing nutrients, such as nitrogen, phosphorus
and potassium, as well as a buffer solution, trace minerals, vitamins, etc., if required.
In many cases addition of these elements is only necessary when using inert or
synthetic carriers. The aqueous phase provides chemicais necessary for an optimal
microbi al activity and allows absorbing the pollutants through continuous mass
transfer from the gas phase to the liquid phase. This is also the reason why
biotrickling filters are sometimes called fixed film bioscrubbers by a few authors.
The aqueous trickling phase is also advantageously used in biotrickling filters for
removing inhibitory or toxic intermediate metabolites or end products. Bioreactors
filled with inert or synthetic carriers to which a nutrient solution is added, though not
by means of a continuous trickling phase, should be considered as conventional
biofilters. Typical examples are biofilters filled with perlite as inert carrier that have
heen used for the removal of styrene or alkylbenzenes (Kennes el al., 1996; Cox el
al., 1997; Veiga el al., 1999; Veiga and Kennes, 2001a) to which an aqueous
nutrient solution was added occasionally, i.e., once a week or sometimes even less
frequently.
47
C. Kennes and M. C. Veiga (eds.), Bioreactorsfor Waste Gas Treatment, 47-98.
2.2 REACTOR CONFIGURA nON
Three types of biofilters are widely used nowadays at industrial-scale, the

conventional biofilter, the biotrickling filter (Chapter 4) and the suspended growth
bioscrubber (Chapter 5). The first two configurations are characterised by the use of
packed beds with attached biomass contrary to the bioscrubber where suspended
biomass develops its activity in a stirred liquid phase. The latter reactor configuration
is characterised by the presence of biomass growing in suspension. A typical
conventional biofilter is depicted in Figures 3.1 and 3.2. A fan leads contaminated air
to the treatment system. After a pre-treatment step, the waste gas is fed to a
humidification chamber. The most typical unit operation introduced as a pre-
treatment step is filtration (Chapter 2). Filtration is needed to remove particles and
dust which would, otherwise, reduce the carrier's lifetime and quickly lead to
clogging problems. A humidification chamber is necessary after that pre-treatment
step in order to saturate air with water, which is needed for an optimal
biodegradation activity. If necessary, temperature and gas flow rate should also be
adjusted to their optimal values. Contaminated air is then fed either in an upf10w ar a
downflow mode through the bioreactar filled with the selected support material.
Downf1ow configuration is sometimes preferred (van Lith et al., 1990) since this will
reduce drying problems at the inlet of the reactor. Such drying problems may appear
as a result ofthe exothermic biologic al reactions (see also section 3.6). Production of
heat is higher at the inlet of the system where higher pollutant concentrations are
present. Occasionally, post-treatment units like adsorption columns, described
previously in Chapter 2, may be installed to remove residual contaminant
concentrations.
Pretreatment
Air - .
inlet
---. Treated
Dram Water . - I (elean) air
water drainage ------'
Figure 3.1. Typical schematic of a closed conventional biofilter.

Conventional Biojilters 49
The first biofilters which were built were open systems, meaning that the
upper part of the bioreactor was not c10sed (Figures 3.2 and 10.1). Although this
might present some relative advantage because of the natural supply of water in
moderately rainy regions, open bed biofilters are much more unstable than c10sed or
in-vessel systems. In closed biofilters (Figure 3.1), parameters, such as water supply,
water content, temperature, etc. are easier to control and regulate.
Treated (elean) air
Pretreatment
Air - - .
inlet
Conditioned
air inlet
Drain
water
Figure 3.2. Typical schematic of an open biofilter.
2.3 BIOFILTER ENGINEERING AND MODELLING
2.3. 1. Design and performance parameters

Several key parameters are used to characterise biofilter performance and operation.
Empty bed residence time (h or s): it represents the mean residence time a
volatile compound would theoretically spend in an unpacked biofilter.
V
EBRT = t = - (3.1)
Q
In biofiltration the volumetric gas flow rate of bulk air (Q) may reach values
of up to 200000 m3/h (Kennes and Thalasso, 1998). It would not be surprising that
higher values might be reached as a result of improved reactor design and more
efficient operation. V is the packed volume (m3) of the reactor. If the re actor is not
packed, or, in other words, if it does not contain any carrier material, the active
reactor volume would correspond to the actual reactor volume available to the
substrate or to the bulk air flow. However, in packed reactors such as conventional or
trickling biofilters, air can only occupy the void volume which depends on, among
others, the porosity ofthe carrier.
The actual residence time (e) is related to the EBRT (t):
(3.2)
where e is the ratio between the void volume available to air and the total reactor or
carrier volume.
Since it is not easy to accurately estimate the value of e, the EBRT is the
parameter most often used in biofiltration and most frequently reported in the
literature. Values of e are not necessarily identical to porosity values, for instance
when the air flow is unable to reach some pores.
Typical EBRT values are close to 1 minute, although some biofilters may
successfully be operated at residence times ranging from less than half a minute to
several minutes (Table 3.1). However, full-scale conventional biofilter operation is
often not cost effective for EBRT exceeding one minute. The influence of the EBRT
on biofilter performance is also illustrated in Figure 3.3.
-
o~
>.
(,)
t:
100 • • •
Q) 80
'0 •
!EQ)
ca 60 •
>
O
E
Q)
40
e:::
20
O 2 4 6 8 10 12 14 16 18
Gas residence time (min)
Figure 3.3. Influence ofthe EBRT on biofilter performance (Hwand and Tang, 1997),
(reproduced with permission).
Retardation factor (dimensionless): Empty bed residence time and actual

residence time represent the average time the bulk air needs for passing through the
biofilter. However, contaminants are retained for longer periods in the reactor.
Quantification ofthe extra time contaminants spend in the biofilter is expressed by a
parameter called retardationfactor (R) and defined as:
R= (3.3)
Conventional Biofilters 51
where Vs and vp are the average bulk air velocity (mlh) and the mean pollutant
velocity (mlh), respectively.
The retardation factor may be calculated by using a pulse experiment
consisting of introducing a spike of contaminant at the inlet of the biofilter and
measuring the movement of the pulse passing through the filter bed (Hodge and
Devinny, 1994). The retardation effect may be attributed to a series of factors often
grouped under the term adsorplion because adsorption is one of the main processes
involved in this phenomenon. Pollutants may indeed be adsorbed onto the carrier
material and to the biofilm, including the microorganisms and ali other material
related to the biofilm, as exopolymers produced by the biomass. However,
contaminants are also absorbed and retained in the aqueous phase, otherwise their
biodegradation would not occur, since microorganisms are active in the aqueous
biofilm or in suspension in the free water film. Under ideal conditions, at equilibrium
the amount of pollutant in the water phase is related to the contaminant concentration
in the bulk air through Henry' s law' s constant. Thus numerica! values of the
retardation factor basically depend on the time pollutants will remain adsorbed onto
solids and/or absorbed in the aqueous phase. To give an idea of orders of magnitude,
retardation factors reported for a compost biofilter and an activated carbon biofilter
reached, respectively, values of 3600 and 26000 (Hodge and Devinny, 1994).
Table 3.1. Typical EBRT used for treating different poliutants in conventional biofilters
Pollutants Mass loading Maximum EBRT (s) References

rate E.C.
(g/m 3 .h) (g/m 3 .h)
Ethylacetate 180 180 180 Deshusses el al., 1999
Hydrogen
sulphide 101' 101' 36 Yang and Allen. 1994b
Hydrogen
sulphide 138 138 38 Wani et al., 1999
Methanol 169 101 162.6 Kraislas el al., 2000
a-pinene
or methanol 20-60 Mohseni and Allen, 2000
Toluene 120 120 57 Veiga and Kennes, 2001a
Toluene 60 25 165 Shareefdeen and Baltzis,
1994
Toluene 100 70 51 Kiared el al., 1996
Toluene 230 165 78 Iorio el al.. 1998
Toluene 70 61 64 Juteau el al., 1999
TEX 72 70 57 Kennes el al., 1996
.
gsl m3 .h, E.C. = Elimination Capacity
Veiga el al., 1999
Superficial gas velocity (m/h) Of surface loading rate: It re~resents the ratio
ofthe gas flow rate (Q, m3/h) to the surface area ofthe carrier (A, m ).
Q (3.4)
v,
A
Volumetric loading rate: The volumetric loading rate is the ratio of the gas
flow rate (Q, m 3fh) to the total volume ofthe filter bed (V, m\
Q (3.5)
v =-
I V
Surface (m 3/m 2 .h) and volumetric (m 3/m 3 .h) loading rates are typically in the
range of about 60-180. However, quite lower or higher values have successfully
been used as well. Volumetric loads of 7 or 8 m 3/m 3 .h have been reported both for
the removal of odour from waste water treatment plants and for VOC biodegradation
in full-scale plants (van Groenestijn and Hesselink, 1993). Higher values reaching
426 m 3/m3 .h were applied in the pilot-scale removal of formaldehyde in plywood
production plants (van Groenestijn and Hesselink, 1993). Surface loads as high as
500 m 3/m2 .h have been applied in a compost-bark biofilter while managing
maintaining low pressure drops (Leson and Winer, 1991).
Substrate mass loading rate (g/m 3 .h): The substrate mass loading rate is the
mass of contaminant fed to the biofilter per unit time and unit volume of carrier
material,
V
Q Sin
=-- (3.6)
v V
Substrate mass loading rates that can be treated in conventional biofilters

depend on the desired removal efficiency which is. in turn. related to the
characteristics of the waste gas and the nature of the pollutants. The higher the
desired removal efficiency, the lower the treatable substrate loading rate. If the waste
gas contains only a few non-recalcitrant compounds, such as alcohols, high removal
efficiencies (> 95%) may often be reached at substrate loading rates close to or above
100 g/m3 .h (Kennes and Thalasso, 1998). For more complex substrate mixtures and
poorly biodegradable compounds with possible inhibitory effects, only lower loading
rates will be allowed unless lower removal efficiencies are acceptable.
Elimination capacity (g/m 3 .h): The eJimination capacity (EC) is the amount of
substrate ar pollutant the biofilter is able to remove per unit volume of carrier
material and unit time. Its mathematical expression is given below.
EC
Q (Sm - SOli') (3.7)
V
Elimination capacity and mass loading rate are both usually expres sed in
g/m 3 .h or mg/m 3 .h. However, pollutant concentrations are sometimes given in g/m 3
or in ppmy. Both units are related and conversion is straightfarward (Chapter 1). If
the substrate concentration is given in ppmy and if the elimination capacity under
standard conditions needs to be expressed in mg/m 3 .h, the following equation will be
used:
EC = 24 5 Q (Sin - SpUI) (3.8)
. MWV
where MW represents the molecular weight ofthe pollutant.

Elimination capacities should be as high as possible, as this will allow using
smaller biofilters, thus reducing investment costs.
Removal efficiency: The concept of removal efficiency (R.E.) in the case of

bioreactors used for air pollution control is similar to the (bio )conversion (XA) of a
given substrate in (bio )chemical reactor engineering. The conversion is the ratio of
the moles of substrate A converted into product to the moles of A fed to the reactor:
O< X = ( moles of A reacted) < 1 (3.9)

- A moles of A fed -
Similarly, for environmental engineers, the removal efficiency is the ratio of

pollutant removed to the amount of contaminant fed. It is usual to multiply this ratio
by 100 to obtain removal efficiencies expressed in percentages.
(3.10)
o$ RE (%) = 100 ( Sin ~. Sant) $ 100 (3.11)

ID
When the contaminant is fully removed, the outlet substrate concentration

(SOUI) equals zero and RE equals 100%, while RE would be null in the case Sin = SOUI.
Complete or near complete removal is often reached up to a given mass load (Figures
3.4 and 4.3). When increasing the mass loading rate, RE will gradually drop.
~80-.------------------------------.
-
..,'
E 70 -
~ 60-
:5 50
iti
g. 40
u
s:: 30
O
:;; 20 -
s::
:§ 10-
W O -~------~---------,---------,--~
O 40 80 120
Mass loading rate (g/m 3 .h)
Figure 3.4. Typical plot ofTEX (toluene, ethylbenzene, o-xylene) elimination capacity
versus mass loading rate in a perlite-biofilter (non published data).
It is worth observing that the mathematical equation representing RE only

takes into account the difference between inlet and outlet concentrations, meaning
that the removed pollutant might completely be mineralised but that it might, as well,
only be biodegraded to intermediate metabolites or adsorbed on the carrier.
RE is also the ratio ofthe elimination capacity (EC) to the mass loading rate:
(3.12)
RE
QS lI1
When plotting mass loading rate versus EC with identical scales and units on
both axes a straight line with a slope = 1 will rcpresent a 100% removal efficiency
(Figure 3.4).
2.3.2. Modelling
Reactor model equations are useful tools for calculating expected biofilter
performance as well as for scaling-up purposes. Basic concepts used in reactor
modelling are based on fundamental engineering equations valid irrespective of the
type of reactor. These equations are (bio )reaction rate equations and continuity,
energy and momentum equations (Froment and Bischoff, 1990). In the first model
applied to the treatment of polluted air in biofilters, equations used for modelling
biofilm-reactors mainly designed for waste water treatment (Harremoes, 1978;
Rittman and MacCarty, 1980) were initially adapted to biofilters for air pollution
control (Ottengraf, 1986). Such model equations have been used for many years.
They are based on simplifying assumptions presenting some limitations, although
they have successfully been applied on several occasions to lab-scale reactors and for
designing full-scale biofilters operating in different countries. Efforts have bccn
made very recently, since the early nineties, to improve biofiltcr models and several
research groups have proposed new alternatives. Although different models
described in the literature will briefly be explained, the simplest and first proposed
model (Ottengraf, 1986) will be detailed more cxtensively since it is at present sti II
the most popular choice. It is also worth noting that several of the more sophisticated
and more recently proposed model equations are, to some extent, based on that first
model or on similar equations.
2.3.2.1. Bioreaction rate equations. Although both resting cells and growing
microorganisms may be involved in the biodegradation of volatile compounds in
continuous gas phase biofilters, in aqueous batch assays mineralization of volatile
organic compounds is typically a growth Iinked biodegradation process. Under such
conditions, microbi al growth rates and removal rates of pollutants are Iinked through
the biomass yield or mass of cells produced per mass of pollutant degraded (Y XlS)
(g/g). An example of growth related biodegradation is shown in Figure 3.5 for a
Pseudomonas strain used as inoculum in biofiltration studies (Amor et al., 2001).
Some typical values ofbiomass yields for microorganisms found in gas-phase
biofilters are presented in Table 3.2. Parameters used in microbial kinetic equations
(see below) may therefore be evaluated in batch assays either by following
biodegradation patterns or growth rates.
Different equations have been proposed for describing specific microbi al

growth rates. However, two widely used mathematical expressions correspond to
Monod kinetics (Monod, 1942) and to Andrews kinetics (Andrews, 1968). The
former is used when working below inhibitory substrate concentrations while the
latter is useful when the biomass is exposed to inhibitory concentration ranges. The
term substrate usually refers to the volatile contaminant to be biodegraded, but it
could, in fact, be any other compound limiting biodegradation rates, such as a
specific nutrient or oxygen.
50 0.6
...
-
.............. __
0.5
40
1 ,
--
--
I
I
...J
0.4 I
30 E
-
Cl
E 0.3
1::
o
U)
al
1::
al
20 ~
:::l 0.2 ci
"O O
1-
10 0.1
O O
O 20 40 60 80 100 120
Time (h)
Figure 3.5. Example of growth related biodegradation of toluene by a Pseudomonas sp.
(Amor et al., 2001).
Table 3.2. Typical biomass yields (Y XIS) and Ks and K, values for Pseudomonas spp. grown
on alkylbenzenes in liquid phase batch culture
Microorganism Ks (mg/l) KI (mgll) Y XJs (g/g) Substrate References
Pseudomonas sp. 18.9 161.2 0.55 TEX (1:1:1) Veiga el al., 1999
Pseudomonas BI 1.96 1.22 Toluene Chang el al., 1993
Pseudomonas XI 0.99 1.88 Toluene Chang el al., 1993
P. putida 0.1 1.2 Toluene Pedersen el al.,
1997
P.putida 6 1980 0.6 Toluene Choi et al., 1992
TEX = Toluene + Ethylbenzene + o-Xylene
Supposing that the growth rate depends only on the concentration of one
compound (S), the Monod equation then takes the following form:
S \
f1 = f1 max
(
Ks + S) (3.13)
Ks is the substrate concentration that gives half the maximum growth rate. It
represents the affinity of a given microorganism for the substrate S. The value of Ks
must be known if one wants to use Monod' s equation for solving biofilter modelling
equations. Accurate estimation of Ks values is often not easy since biofilters are
usually colonised by mixed, often unidentified microorganisms, rather than by one
single strain. AIso, depending on the method used, quite different Ks values may be
obtained for a same microbi al strain using Monod's model (Table 3.3). Kinetic
parameters may vary in some cases by up to one order of magnitude for a given
strain, which is non negligible in designing problems and biotilter sizing.
Table 3.3. Example of variable Ks values obtained with Xanthobacter autotrophicus GJI0
grown on 1,2-dichloroethane (Ferreira Jorge and Livingston, 1999)
Reactor Ks (mg/l) Reference

Continuous membrane bioreactor 7.6 Freitas dos Santos and Livingston,
1995
CSTR 25.5 Van den Wijngaard el al., 1993
Batch 52 Hartmans el al., 1992
At high substrate concentrations, S » Ks, Monod kinetics reduces to a zero

order expression:
fi = f1 max = constant = ko (3.14)
where k o is the zero order rate constant.
At low substrate concentrations, S « Ks , Monod kinetics reduces to a tirst

order equation:
(3.15)
where k j is called the tirst order rate constant.
In such case, microbial growth rate (Il) and biodegradation rate (rs) in
biotilters are proportional and are related through a constant value according to:
r = - (dS) = (~) fi = r~) kS = k' S (3.16)

dt Yx/s \ Yx/s
if one assumes that biomass yield (YX/s) and biofilm density (B) are constant
parameters.
2.3.2.2. Simple steady state biojilm reactor model adapted to conventional biojilters
for air pollution control. The model is based on mass balance equations. As
mentioned below, several assumptions need to be made for this simple biofilter
model:
• In the biolayer, nutrients are transported by diffusion, which is described by an

effective diffusion coefficient.
• The biofilm thickness is small compared to the diameter of the carrier particles,
suggesting planner geometry.
• The microkinetics of pollutant biodegradation in the biofilm is described by the
Monod equation.
• Polluted air flows in a plug-flow mode through the biofilter.
• Interface resistance in the gas phase is negligible.
• Pollutants are removed only by aerobic biodegradation.
• The biolayer or biofilm is considered as a water phase.
• Substrate concentrations in both the gas phase and the biofilm are related
according to Henry' s law' s coefficient.
• Qualitative and quantitative biomass distribution is homogenous over the reactor
height and remains constant with time.
Substrate mass balance equations need to be written both for the biofilm and
for the bulk gas phase. Under steady state conditions the biofilm mass balance
equation is:
o Dd2SJ
(dx 2
- Reaction rate (3.17)
with
D = effective substrate diffusion coefficient in the biofilm (m 2!h)
= effective substrate diffusion coefficient in water (m2!h)
x = biofilm depth (m)
The reaction rate is calculated from the Monod equation as explained above.
In order to obtain an analytic solution, the Monod equation needs to be simplified
and reduced either to a zero order or a first order expres sion as described previously.
For a zero order reaction: O (Dd~:S) - k o (3.18)
For a first order reaction: O (DdS)

dŢ
2
- kl S (3.19)
With the following two boundary conditions:
Sg (3.20)
H
(2) dS b = O (3.21)
dx
where
Sb = substrate concentration in the biofilm (g/m3 )
Sg = substrate concentration in the bulk gas phase (g/m 3 )
H = Henry's law's coefficient (-)
The mass balance equation for the bulk gas phase along the biofilter height
under steady state conditions is:
O=-u ( -dSg) -AN (3.22)

dh S
where
u = superficial gas velocity (mlh)
= Q/A = gas flow rate (m3/h) / empty bed cross sectional area (m 2)
h = reactor height (i.e., carrier height) (m)
As = biolayer surface area per unit volume of bioreactor (m 2/m 3)
N = flux of substrate from the gas phase into the biofilm (g/m 2 .h)
= _D (dS b )
dx FO
Thus
(3.23)
with initial condition (at h = O):
(3.24)
meaning that the substrate concentration (Sg.h) at height h = O is equal to the substrate
concentration in the feed (Sg.o).
The set of equations may easily be solved for three different operating
regimes as mentioned hereafter, obtaining new equations relating the feed
concentration to the pollutant concentration at different heights in the biofilter
(Ottengraf and van den Oever, 1983).
A.l) Zero order reaction (biodegradation) rate with substrate difJusion limitation
(Figure 3.6).
(3.25)
The effective biofilm thickness or active biofilm layer corresponds in such

case to:
2dS)O.5
o= (--g (3.26)
ko H
A.2) Zero order reaction (biodegradation) rate with reaction limitation (Figure 3.6).
(3.27)
where o' is the effective biotilm thickness. The biotilm is m this case fully
penetrated.
Gas phase Biofilm Carrier

c
z
Figure 3.6. Pollutant concentration for zero order kinetics with either diffusion limitation
(Al) or reaction limitation (A2) and for tirst order reactions (8).
B) First order reaction (biodegradation) rate. For a tirst order reaction, the Thiele
modulus is used to distinguish between diffusion limitation and reaction limitation.
The Thiele modulus ($) is defined as:
~=o' --t
( k ) 05 (3.28)
Values of $ <..f2 correspond to reaction limitation, whereas values of $ > ..f2

correspond to diffusion limitation, The critical value $ = ..f2 allows ca1culation of the
biofilm thickness,
The gas phase substrate concentration profile along the height of the biofilter
is calculated from:
Sg,h _ [- (h Aş D~ tan ~)]

Sg,o - exp (O' HU) (3,29)
C) Improved models. One of the major advantages of the above mentioned model is
that it yields simple equations for which analytical solutions are quite easily
obtained, Although it is based on a series of simplifying assumptions, it often leads
to a good fit between model equations and experimental data, However, in some
cases, such as when working with mixed waste gases or in the case of uneven
biomass distribution, the model equations may not fit experimental data (Veiga and
Kennes, 2001a), Therefore efforts have been made very recently for developing
improved model equations (Devinny et al., 1999). Such models are considered to be
improved models in that they are usually a closer picture of the real situation.
However, in most cases they are more complicated to use and to apply and no
analytical solution is available,
CI) Bioreaction rate equation. Pollutant concentrations are prone to vary over time
in industrial waste gases. Contaminant concentrations do also vary with biofilter
height since the pollutant gets gradually biodegraded as the air stream passes through
the reactor. According to Monod kinetics biodegradation rates may thus switch from
zero order to first order kinetics, Such transition is not instantaneous and fractional
order kinetics may be observed over part of the total biofilter height. It is therefore
more logical to use the Monod equation (Hirai et al" 1990; Deshusses et al., 1995a)
or to switch from zero order to first order kinetic equations at a given reactor height.
Also, at high pollutant concentrations, it will be more accurate to use kinetic
equations which take into account possible inhibitory effects, such as the Haldane
equation or the Andrews equation (Andrews, 1968; Shareefdeen el al" 1993;
Shareefdeen and Baltzis, 1994). The Andrews equation has the following form:
S
(3.30)
where KI represents a concentration called inhibition constant (Table 3.2).

When several pollutants are present in a waste gas, model equations considering
cross-inhibition effects may be necessary (Baltzis and Shareefdeen, 1993; Deshusses
et al., 1995a, 1995b).
C2) Limiting substrate. Substrate concentration (S) in the kinetic equation refers to
pollutant concentration. However, in some cases oxygen might become limiting in
biofilms and control the overall removal rate. In such a case the oxygen
concentration needs to be inc1uded in the kinetic expression (Shareefdeen et al.,
1993). Possible kinetic equations are then:
(3.31 )
li = limax (K SP
or
(3.32)
where the subscripts P and O refer to pollutant and oxygen respectively.
C3) Non-steady state conditions. Biofilters are most likely to be operating under
non-steady state conditions, among others as a result ofthe variable characteristics of
industrial waste gases over time. Model equations taking into account such transient
states have been proposed and described by few authors (Shareefdeen and Baltzis,
1994; Hodge and Devinny, 1995; Deshusses et al., 1995a, 1995b; Devinny et al.,
1999).
C.4) Other parameters. Comments on other assumptions need to be made. The use of
water/air partition coefficients for calculating pollutant concentrations in the gas
phase and in the biofilm may not be valid in some cases, mainly when working with
hydrophobic pollutants (Mohseni and Allen, 2000).
Another key factor to be considered when modelling is that homogenous
quantitative and qualitative biomass distribution is generally assumed. It is quite
obvious that in natural carriers (soil, peat, etc.) this might not be true. Recent studies
have also shown that with other carriers (perlite, celite) biomass concentrations do
vary along the biofilter. Higher microbial concentrations are found near the inlet of
the reactor (Cox et al., 1997; Song and Kinney, 2000a, 2000b; Veiga and Kennes,
2001a). Up to 20% (Cox et al., 1997) or even 40% (Veiga and Kennes, 2001a) more
biomass may develop near the entrance of the system. Aiso biofilm or biomass
activity is sometimes not homogenous along the reactor height (Veiga and Kennes,
2001a) and is expected to vary with time, for instance, in the presence of variable
microbial populations. In case such variations would not be negligible, mass balance
equations should be applied to the biomass. As will be commented further in this
chapter, it was recently also shown that shifts in microbial populations are expected
to occur in conventional biofilters, even for relatively short periods of operation of a
few months (Sakano and Kerkhof, 1998; von Keitz et al., 1999).
Considering that the biofilter operates as a plug flow reactor is another

simplifying assumption that is nearly correct in many cases, though deviations from
ideal behaviours are, above ali, not unusual in the long run.
Several carrier materials used in conventional biofilters are able to adsorb
pollutants (Abumaizar et al., 1997). Such a characteristic is usually not taken into
account, which may not be a problem provided the biofilter is operating under steady
state (equilibrium) conditions.
3. Parameters affecting biofilter performance
3.1 FEED CONDITIONS AND COMPOSITION
Gas flow rate and feed composition are of prime importance in biofiltration. Besides
mass transfer rates, global removal rates or e1imination capacities are highly
dependant on biodegradation rates. Both mass transfer rate and biodegradation rate
are affected by the composition and other characteristics of the feed. At higher gas
flow rates, resulting in shorter residence times deeper biofilters will be needed. Feed
composition will affect reactor sizing as well. Usually larger biofilters will be
required when dealing with higher substrate concentrations or in the case of highly
reca1citrant compounds. One of the main advantages of biofiltration compared to
other technologies is its low cost. However, when larger biofilters are used,
investment costs do increase.
As a general rule biofiltration technology is often not the best choice for
pollutant concentrations above 4 or 5 g/m3 ; on one hand, because it would increase
investment costs and, on the other hand, because of potentially inhibitory effects on
biomass activity. Inhibitory effects observed at high pollutant concentrations or
during shock loads may partly be buffered and avoided by using carrier materials
with high adsorptive capacity, i.e., activated carbon (Devinny and Hodge, 1995;
Weber and Hartmans, 1995). Unless working at high EBRT, biofiltration at relatively
high substrate concentrations may lead to low removal efficiencies (Table 3.4) and
accumulation of toxic intermediate biodegradation products (Devinny and Hodge,
1995). Although relative1y high elimination capacities may be reached, residual
concentrations released from the bioreactor effluent might not comply with
regulatory requirements. Higher mass loading rates would also generate higher
temperature increases resulting from increased metabolic activity. When dealing with
highly loaded waste gases, biofiltration could stiU advantageously be used by
combining this technology with other non-biological processes.
Table 3.4. Influence ofthe feed concentration on the removal efficiency of some VOC s
Compound Inlet Removal EBRT References

concentration efficiency (%) (s)
(mg/m3)
Ethanol 10000 <20 186 Devinny and Hodge. 1995
Phenol 3000 24.5 Zilli et al., 1996
Toluene 6200 60 78 Jorio et al., 1998
Xylenes 8200 23 102 Jorio et al., 1998
Waste gas composition is important as well. Many lab-scale studies are

undertaken with single pollutants which allows easier and more accurate
interpretation of results and phenomena observed. At the industrial scale, one may
indeed find waste gases in which a single compound is present or, at least, in which
one compound is detected at much higher concentration than the others. However, in
many cases more than one pollutant will appear. Biofilter performance and removal
efficiencies of specific contaminants may (or not) vary when multiple pollutants are
present in the air, as compared to waste gases with single pollutants. Smet et al.
(1997) found that dimethyl sulphide removal in a wood bark-compost biofilter was
partly inhibited by the presence of isobutaraldehyde, while toluene showed no
significant effect on biodegradation of the sulphur compound. On the other hand, for
H2 S polluted air no inhibitory effect was observed when dimethyl sulphide or
dimethyl disulphide were added to the waste gas (Wani et al., 1999). In case of
mixtures of VOCs, Kennes et al. (1996) observed that a maximum TEX elimination
capacity of 70 g/m 3 .h was reached when feeding a toluene/ethylbenzene/o-xylene
mixture to a conventional biofilter which was very close to the mean maximum
elimination capacity reached when feeding toluene (72.9 g/m3 .h), ethylbenzene (85.2
g/m 3 .h) or o-xylene (63.6 g/m3 .h) individually to the biofilter, suggesting that co-
inhibition was probably basically negligible in that system.
3.2 CARRIER MATERIAL
Biofilter performance is highly dependent on the nature of the carrier material, also
called filter bed, support material or medium by different authors. The carrier
material is a solid phase on which adhesion of the biocatalyst (microorganisms) takes
place resulting in the development of a so called biojilm growing as a result of
pollutant degradation.
3. 2.1. Natural carriers

Several organic and natural carriers have been used individually or as mixtures in
biofiltration. The interest of adding inert materials to natural carriers has recently
also been tested. Among the most common natural filter beds, one should mention
compost, peat, soil and wood derivatives (Table 3.5) (Kennes and Thalasso, 1998).
Such carriers present some advantages, such as their low cost, abundant
availability and natural presence of nutrients and microorganisms. However, many of
them present a limited lifetime as a result of phenomena, such as biodegradation of
the filter bed or nutrient depletion. Compost, peat and soil are the most extensively
used natural carriers.
Compost. Although soil was probably the first natural filter medium used in
biofiltration, compost and peat gradually gained better acceptance later on and they
are nowadays probably the most popular non-synthetic carriers. Composting is a
technology used for the aerobic or, sometimes, anaerobic biologic al conversion and
stabilisation of organic matter mostly present in solid form. As a result of the
biological activity and the solid nature of the system, thermophilic conditions are
reached leading to a product free of pathogens. Aerobic composting is usually
preferred over the anaerobic process since end products of anaerobic conversion
generate odour problems which are, contrarily, basically avoided under aerobic
conditions.
64 C Kennes and M C. Veiga
Table 3.5. Examples of natural carrier materials used in conventional biofilters
Carrier material Typical volume or References

mass ratio'
Compost Ottengraf and Van den
Oever, 1983
Compost-yard trash 1/3 Yang and AlIen, 1994
Compost-wood chips Allen and Phatak, 1993
Compost-perlite 1/1 Ergas el al., 1994, 1995
Morgenroth el al., 1996
Quinlan el al., 1999
Compost-diatomaceous earth 1/2 Hodge et al., 1991
Hwang and Tang, 1997
Compost-polystyrene 1/1 Hodge and Devinny, 1994
Deshusses el al., 1995b
Compost-chaff 1/1,2/1 Hwang and Tang, 1997
Compost-GAC 1/2 Hwang and Tang, 1997
Compost-pall rings 8/2 Chou and Cheng, 1997
Krailas el al., 2000
Compost-hog fuel 1/1 Wani el al., 1998
Hog fuel-perEte 4/1 Wani el al., 1999
Peat 1 Mallakin and Ward, 1996
Arnold et al., 1997
Wu el al., 1999
Peat balls 1 Rothenbiihler et al., 1995
Peat moss 1 Acuila el al., 1999
Peat-bark-wood 711 /2 and 1/7/2 Marek el al., 1999
Peat-polyurethane foam 7/3 Shareefdeen et al., 1993
Peat-polyurethane 2/2/1 Shareefdeen et al., 1993
foam-vermiculite
Peat-polyurethane foam-perlite 2/3/5 Shareefdeen el al., 1993
Peat-perlite 2/3 Shareefdeen et al., 1993
Peat-compost 1/1 Kennes and Thalasso, 1998
Peat-glass beads 2/1 Zilli el al., 1993, 1996
Wood bark 1 Van Langenhove el al., 1986
Weckhuysen el al., 1994
W ood waste-perlite 1/1 Lee el al., 1996
Soil 1 Bohn and Bohn, 1988
Hodge el al., 1991
Soil-sand-peat-compost 20/2/3/3 Frye et al., 1992
'Orientative values, not necessarily valid for aII references cited.
Compost contains nutrients, such as nitrogen, phosphorus and trace elements

advantageously used for microbial growth and biodegradation of volatile
contaminants in biofilters. Typical characteristics and compositions compared to
other carriers are shown in Table 3.6. However, the composition of compost may
vary since it is highly dependant on the nature of the substrates originally available
for bioconversion. A problem to be taken into account when considering compost as
filter bed is that it might not be free of organic contaminants or heavy metals (Martin
and Loehr, 1996) causing inhibition of microbial activity. Many prime matters are
suitable for composting. Some examples are bark, different kinds of sludges, yard
trash, wood residues and agricultural wastes. Composts produced from such
substrates have already ali been used in biofiltration studies (Morgenroth et al., 1996;
Smet et al., 1996; Tang et al., 1996). A wide variety of microorganisms is found in
compost inc1uding bacteria, fungi (yeasts and molds), algae, protozoa and virus
partic1es. A wide microbial flora will thus naturally be present on start-up of
biofilters using such a filter bed, although they might sometimes not possess the
enzymes needed for mineralization of recalcitrant contaminants present in polluted
air streams.
Table 3.6. Typical composition and characteristics of some natural carriers
Carrier Dry bulk Specific Void C/N ratio References

density
(g/cm 3)
surface
area
fraction' or
porosity
..
(m2/g) (%)
Compost 0.272 1.4 15.2 Smet el al. 1996
Compost 0.18-0.30 84.3-89.7" 7.14-31.5 Yang and Allen,
1994
Compost 0.51 65.2*' 27 Wani et al., 1998,
1999
Hog fuel 0.25 82.4*' 325 Wani et al., 1998,
1999
Peat 0.17 1.6 49.8',82*' Zilli et al., 1996
(apparent)
0.91
(absolute)
Peat 0.11 11.5 Hirai et al., 1990
Peat 0.52-0.72 300-1300 39.5-42.9*' Wu el al., 1999
(m'l)
Compost- 0.22-0.33+ 50' Deshusses and
polystyrene Hamer. 1993
Deshusses el al.,
1995b
+Wet basis (60%)
Peal. Peat is an organic material composed of carbohydrates, minerals and a group of

substances usually called humic acids. Peat is a common carrier material in
biofiltration. It presents interesting adsorptive properties as well as a quite large
surface area to weight ratio. It has the ability of adsorbing several organic and
inorganic compounds, as well as heavy metals in aqueous, as well as in gaseous
phases (Allen et al., 1997). As an example, in the case of contaminated air, the
maximum uptake of N02 by peat was 692 mg/g, to be compared to a maximum
uptake of 1240 mg N02 /g reported for granular activated carbon.
As do most organic carriers, it contains nutrients useful for enhancing
microbial activity, though nutrient concentrations in peat are rather low. A typical
chemical composition of raw Sphagnum peat is as follows (% of dry weight) (Martir
and Manu-Tawiah, 1989): total solids, 19.50 ± 0.50, total lipids, 2.52 ± 0.17, total
nitrogen, 0.80 ± 0.08. The moisture content of wet peat was around 80.50 % with a
pH of 4.85 ± 0.11. It also contains carbohydrates and trace minerals. A typical
elemental analysis ofpeat is (%) (Hirai etal., 1990): C, 48.03; H, 5.51; N, 0.82; S.
0.21; Fe, 0.12; ash, 4.20. In this case water content was 46.7% and pH 3.75. Peat is
in most cases acidic.
Soi/o Most of the first biofilters used for air pollution control were open soil
biofilters. Polluted air was simply passed through soil either in an underground or
above ground biofilter taking advantage of the natural presence of soil
microorganisms able to degrade the contaminants. Viable heterotrophic plate counts
per gram dry weight of fertile soil have been reported as 10 8_10 10 bacteria and 10 5_
108 actinomycetes (Alexander, 1977). Dry soil is composed both of inorganic
mineral compounds and organic matter. Wide variations may be found in the ratio
between these two components and the organic maUer content may vary from less
than 1% to about 95%. Some characteristics of soils used in biofiltration are
presented in Table 3.7 (Miller and Canter, 1997). Moist bulk density does usually
range between the density of soil organic maUer (about 0.5 g/cm3) and the density of
mineral matter (about 2.7 g/cm\ Contrary to several other natural carriers, inorganic
soil is hydrophilic. It is usually less sensitive to fluctuations in moisture content but
its low permeability often leads to significant pressure drops over long term
operation periods and to the need ofusing large EBRT.
Table 3.7. Some examples of characteristics of soi! used in biofiltration
Soil type Moist bulk Permeability Organic maUer Partide size

density (glcm3) (in/h) (%) (mm)
Dougherty 0.63-2 0.79 <2.0
sand
Durant loam <0.6 0.75 <0.42
Rubicon sand 1.35-1.55 6-20 0.5-1.0 <4.7
3.2.2. Other carriers

Other inert carriers used in conventional biofilters may present some advantages
compared to the natural ones. Their composition is usually well defined and quite
stable over time. They are much more homogenous than such carriers as soil, peat or
compost. This allows minimising pressure drop and channelling problems. However,
drawbacks should also be mentioned, among which are the absence of nutrients and
microorganisms and the higher investment costs. Nevertheless, the initial absence of
nutrients may be interesting, to some extent, since their external addition allows for a
better control and regulation on the amount of nutrients available to the
microorganisms. Indeed, although being found in some natural carriers, carbon,
nitrogen, phosphorus, trace minerals and vitamins are usually not available in a
balanced ratio. Several synthetic or inert carriers have successfully been used in
biofiltration, such as GAC, perlite, glass beads, ceramic rings, polyurethane foam,
polystyrene and vermiculite, to mention only a few. The most extensively used are
perlite and aluminosilicates in general as well as activated carbon. Some researchers
also tested the possibility of using Ca-alginate supports (Chung el al., 1997) in lab-
scale biofilters, though this might not be suitable in full-scale reactors.
Activated carbon. As previously mentioned (Chapter 2), activated carbon is used in

adsorption processes. One of its major characteristics is thus its ability to adsorb
volatile pollutants and to act as a buffering agent in case of shock loads (Weber and
Hartmans, 1995). In such a case, high concentrations of the contaminants may get
adsorbed onto the carrier and gradually released later as mass load decreases. This
avoids inhibition of the microbial populations in the filter bed. Typical advantages
and disadvantages of activated carbon are respectively its very long lifetime and its
high cost.
Perlite. Perlite is an inert and inorganic material used in biofiltration, either

individually (Cox el al., 1997; Kennes et al., 1996; Veiga and Kennes, 2001a) or
mixed with organic material, such as compost, peat, etc. When mixed with organic
carriers, perlite plays the role of bulking agent and presumably reduces pressure
drop. Perlite is a volcanic mineral which has been exposed to high temperatures of
about 900 DC. As a re suit of the heating process, the volume of the mineral particles
increases considerably, forming light white granules with a density ofroughly 100 ±
20 kg/m3 for dry material. The particles are characterised by a large surface area.
Perlite granules are composed of 65-80% Si02, 12-16% Ab03, 3-5% Na20, 2-4%
K20, 1-3% Fe203 and 0-2% CaO and show a near neutral pH. Perlite has no
adsorption capacity in the presence of organic compounds, as was observed in
experiments undertaken with alkylbenzenes (toluene, ethylbenzene. xylenes)
(Kennes el al., 1996).
It has been suggested that a good filter bed should respond to the following
criteria (Bohn, 1996):
i) be chemically and physically stable,

ii) retain microbes strongly,
iii) present appropriate physical characteristics (surface area, density, etc.),
iv) produce clean drainage water,
Regarding chemical and physical stability, it was already mentioned that

natural organic carriers are characterised by the natural presence of nutrients and
microorganisms, but they do also present disadvantages, such as their limited lifetime
compared to the almost endless lifetime of synthetic carriers. Their chemi cal and
physical characteristics may thus vary over time. As a general rule, synthetic and
inert carriers may be considered much more stable than other carriers, mainly natural
organic ones. Mineralization of natural organic carriers generates a gradual bed
compaction and increasing pressure drops over time leading, in the long run, to
reduced perforrnance and the need of replacing the filter bed (Hodge and Devinny,
1994, 1995; van Lith el al., 1990). Many chemical and physical factors play a role in
microbi al mineralization of organic supports. Carbon mineralization of compost is
often faster in the earliest stages of reactor operation and vary with the C/N ratio
which, in turn, does also vary with operation time (Wani el al., 1998). It is worth
remembering that the C/N ratio will affect biofilter performance as well. According
to recent studies, higher C/N ratios seemed, to some extent, to increase removal
efficiencies (Chou and Cheng, 1997). Other factors playing a role in the rate and
extent of decomposition of organic carriers include pH, temperature, moisture and
oxygen contents, nature and concentration of pollutants present in the waste gas, and
particle size distribution ofthe bed (Tester el al., 1979).
Strong biomass adhesion on the carrier material is interesting to some extent

since, theoretically, higher biodegradative activity will be obtained at higher biomass
concentration. However, a too heavy biomass growth may not be beneficia! since it
may result in higher pressure drops and clogging problems. This is often the case
with inert or synthetic carriers to which a liquid nutrient phase is added on a regular
basis. It is therefore more typical of biotrickling filters than of conventional
biofilters. It was also shown with conventional biofilters that with thick biofilms, part
of the biomass might not be active, and when heavy growth is taking place, volatile
compounds might not reach the deeper zones of very thick biofilms.
Among relevant physical characteristics of carrier materials one should
mention the specific surface area' (m 2/m\ Higher specific surface areas should
increase mass transfer from the gas phase to the liquid phase resuIting in higher
elimination capacities (van Groenestijn and Hesselink, 1993). Nevertheless, based on
experiments undertaken with different support materials derived from peat, some
authors suggested that the opposite could be observed in some cases (Wu el al.,
1999). However, such a conclusion should be checked carefully, since in that study
other parameters as pH, which does also affect biofilter performance, varied as well,
from one carrier to another. Modifications in the characteristics of the carrier
material over long term operation periods may adversely affect overall biofilter
performance as a resuIt of increased flow resistance, presence of unused dead zones,
channelling, reduced porosity and increased energy requirements.
Since humidified air is fed to the biofilter and a water phase may occasionally
be fed as well, phenomena, such as condensation wilI result in the production of
drainage water. Such biofilter waste water should be as clean as possible. Otherwise
its treatment will be necessary, thus increasing costs.
Other parameters which are occasionally or partly related to the
characteristics of the filter bed are important as welI, although they are also strongly
dependant on other factors. Such parameters, which will be described more
extensively below, include the nature and presence of microbial populations able to
mineralise the contaminants, presence of nutrients and water needed for microbial
activity, adequate temperature, pH stability and buffering capacity, pressure drop,
etc.
3.3 MICROFLORA
3.3.1. Biofilter inoculation

On starting-up a biofilter filled with a natural carrier, microbial populations are
already naturally present in the reactor. This is, of course, not the case with synthetic
and some inert carriers. Waste water sludge is one of the most frequently used non-
defined inoculum (Kennes and Thalasso, 1998). It is worth mentioning that highly
chlorinated compounds, such as PCE, which are only biodegradable by anaerobic
bacteria (Kennes et al., 1998) could be removed from waste air in biofilters seeded
with waste water sludge (Kim, 1997). Although it was mentioned that it may be
interesting to start-up biofilters with a filter bed containing a large, diverse microbial
community, several biofiltration studies started-up with single strains or defined
consortia allowed reaching very high removal efficiencies and elimination capacities
(Cox et al., 1997; Veiga el al., 1999; Zilli el al., 1996). In some cases, although filter
bed contamination is highly probable and expected over a relatively long term
operation, the inoculated strains might remain present or even dominant for several
months (Cox et al., 1997; Veiga and Kennes, 2001a). This is particularly true when
defined inocula are used with synthetic carriers and if native microorganisms present
in such inocula were directly enriched in bioreactors fed with specific VOCs. In
other cases inoculated strains might get overgrown by other more performant
organisms (Fritsche and Lechner, 1992; Jorio et al., 1998). Biofilter inoculation may
also represent a means to speed up the start-up period (Veiga and Kennes, 2001a),
even though inoculated microorganisms might eventually get overgrown by other
strains in the long run. Examples of pure cUltures or defined mixed cultures used for
inoculation of conventional biofilters are presented in Table 3.8.
For biofilter inoculation with pure cultures, speciali sed bacterial inocula are
normally used. Recent research studies have shown that seeding eukaryotes as fungi
or algae might be useful in some cases. Studies on the biodegradation of VOCs by
fungi were reported already several decades ago (Anselmo et al., 1985; Edmonds and
Cooney, 1967; Engelhardt et al., 1977; Goldbeck el al., 1983). The removal of
sulphur compounds, such as hydrogen sulphide, methanethiol and dimethyl sulphides
by fungi has also been reported, yielding sometimes higher biodegradation rates than
with bacteria (Phae and Shoda, 1991). Studies on the use of fungal inocula in wastc
gas treatment were started more than ten years ago and were first published in the
late eighties and early nineties (Hiittermann et al., 1988; Majcherczyk et al., 1990;
Braun-Liilleman et al., 1992; Cox et al., 1994). White rot fungi are very popular
organisms in biodegradation studies. Although they have proven to be best suited for
the co-metabolic removal of semi-volatile complex aromatic substrates, such as
polycyclic aromatic hydrocarbons (Cemiglia, 1992) rather than highly volatile
substrates, some researchers suggested their use in biofiltration ofVOCs.
A potential problem with white rot fungi is the accumulation of intermediate
products (Kennes and Lema, 1994; Yadav and Reddy, 1993), although relatively low
molecular weight metabolites might get degraded by other microorganisms in
complex cultures as found in biofilters. Biofiltration studies undertaken in the
presence of strains, such as Trametes versicolur, Phanerochaete chrysusporium,
Pleurotus ostreatus and Bjerkandera adus ta using lignocellulosic materials (straw,
agricultural residues, etc.) as filter bed, allowed reaching high elimination capacities
for compounds, such as a-pinenc, styrenc and alcohols, with sometimes complete
conversion to carbon dioxide and water, which should most probably result from the
metabolic activity of other microorganisms present in filter bed next to these
basidiomycetes (Braun-Liillemann et al., 1992). Since most white rot fungi de grade
lignocellulosic material, i.e., the filter bed, studies on bed compaction and
concomitant pressure drop would be useful.
As a general rule fungi may be especially suitable for the removal of
hydrophobic pollutants, such as styrene or alkylbenzenes (van Groenestijn and
Hesselink, 1993). In biofilters the aerial hyphae of fungi represent a large surface
area in direct contact with the contaminated gas phase (Braun-Liillemann et al.,
1992). Under such conditions absorption of hydrophobic compounds is favoured
since the contaminant is directly transferred to the cell surface without phase
transition problems. Other interesting characteristics of most yeasts and filamentous
fungi are their resistance to extreme environmental conditions, such as low water
contents and low pH values although their metabolic activity under such conditions
may be reduced. Although the first reported studies were undertaken with natural
carriers and white rot fungi (Majcherczyk et al., 1990; Braun-Liilleman el al., 1992),
probably because they are well characterised and widely used strains, other fungi
~
Table 3.8. Examples of pure cultures or defined mixed cultures used in conventional biofilters
Strain Compound References

Alcaligenes xylosoxidans + a-pinene Kleinheinz et al., 1999
Pseudomonas jluorescens 0
Corynebacterium pseudodiphteriticum Ethanol, pentanol, isobutylacetate Gibson et al., 1994 ~
;::;
Hyphomocrobium MS3 Dimethylsulfide Smet et al., 1997 ;::;
Pseudomonas putida Phenol Zilli et al., 1993, 1996
I:l
'"""
Pseudomonas sp. + Bacillus sp. + Alkylbenzene mixture Veiga et al., 1999 ;::;
I:l..
Trichosporon beigelei (T, E, X) Veiga and Kennes, 2001 a ~
Rhodococcus sp. B261 n-Valeric acid Yun and Ohta, 1998 0
Thiobacillus sp. Hydrogen sulphide, methanethiol, dimethyl Hirai et al., 1990 ~
sulphide Cho el al., 1992 ciq'
I:l
Thiobacillus sp. Hydrogen sulphide Chung el al., 1997
Consortium of 8 bacteria Methanol Shareefdeen et al., 1993
Consortium of 11 bacteria and 1 fungus Dichloromethane Ergas et al., 1995
grown on inert carriers might be more performant (Cox et al., 1994, 1997).
Good reactor performance was obtained with a biofilter containing the fungus
Exophiala jeanselmei forming long filamentous mycelia. The latter strain uses the
pollutant as the sole source of carbon and energy and biodegradation does not rely on
the rresence of a co-substrate. Maximal styrene elimination capacities around 62-90
g/m .h were reached (Cox et al. , 1997). Similar results were obtained in the presence
of an alkylbenzene degrading culture composed of mixed populations of bacteria,
yeasts and filamentous fungi (Figure 3.7) (Veiga et al., 1998; Veiga and Kennes,
2001a). Some examples of yeasts and fungi inoculated in vapour phase bioreactors
appear in Table 3.9. Fungal strains inoculated in biofilters for VOC treatment should
preferably possess the following characteristics (Sanchez-Pefia el al., 2000) :
• Able to utilise pollutants as sole source of carbon and energy.

• Able to completely mineralise the pollutants.
• Non-pathogenic to mammals and other living organisms in general.
• Not easily displaced by other undesirable microorganisms.
• Minimal biomass production to prevent filter bed clogging.
• Cells stable and viable when subjected to adverse conditions, such as lack of
oxygen, pH fluctuations , periods of desiccation, and water saturation or
shutdown.
Figurc 3.7. SEM photograph of carrier material from a biofilter treating alkylbenzenes
showing the presence of numerous fungi and bacteria (Veiga and Kennes. 200 I b).
3.3.2. Microbial characterisation

In biofilters using natural organic carriers, a wide range of (micro)organisms will be
present in the system irrespective of any possible initial inoculation with pure
cultures. Conventional biofilters packed with synthetic or inert carriers do not usually
contain microorganisms at detectable levels and need to be seeded. After a few
weeks operation, although inoculated strains might still be present in the biofilter,
other new strains will most often also grow on the carrier material.
;:j
Table 3.9. Some characteristics and performance of biofilters inoculated with fungi
Pollutant Strain Max. inlet Max. E.C. Carrier material References

concentration (g/m 3) (g/ m 3 .h)
Styrene, Phanerochaete 0.1 Lignocellulosic materials Majcherczyk el al., 1990
a-pinene chrysosporium, Braun-Liillemann el al., n
Pleurotus ostreatus 1992,1997 ~
Styrene Exophiala jeanselmei 2.4 62 Perlite Cox et al., 1997 ::s
::s
91 (40% O2 )
!:l
Benzene, toluene, Phanerochaete < 10 (total) Glass beads Oh et al., 1998
'"'"
::s
xylenes chrysosporium !:l..
Toluene, ethylbenzene, Trichosporon beigelei 2 72 (total) Perlite Veiga et al., 2000 ~
o-xylene in co-culture with 2 120 (total) Veiga and Kennes, n
bacteria 2001a ~
(') ~.
Toluene Exophiala spp. 0.06 Polypropy1ene pall rings, Sanchez-Pefia et al.,
Porous silicate pellets, 2000
Compost
Toluene Scedosporium 6 80 ± 10 (*) Vermiculite/Activated Garcia et al., 2000
apiospermum carbon (85/15)
(,) Preliminary results ar short-term non steady-state experiments
A relatively wide variety of organisms is usually expected to be found in

biofilters which may include bacteria, yeasts and other fungi, protozoa, algae, etc.
Higher organisms, such as predatory mites, nematodes and collembolans have been
detected in peat biofilters (Arnold et al., 1997). Although microorganisms have been
isolated and identified by classical microbiological methods, many microorganisms
are, to some extent, reca1citrant to cultivation and it is expected that probably only
less than 20% of the strains present in bioreactors used for waste treatment can be
isolated and identified by classical microbiological methods (Wagner et al., 1993;
Wagner, 2000). Molecular tools have very recently been used for identifying hardly
culturable microorganisms in order to obtain a more reliable picture ofthe microbial
populations present in such ecosystems, although reported results are still scarce.
Among the techniques used for obtaining a deeper insight into the microbial structure
of carrier materials one should mention analysis of phospholipid fatty acids (PLFA)
(von Keitz et al., 1999) aud fluorescence in situ hybridisation (FISH) or 16 SrRNA
gene characterisation (Sakano and Kerkhof, 1998; von Keitz et al., 1999).
According to classical plating techniques, it was observed that, in a compost
biofilter treating toluene vapours, only 15% of the microorganisms growing on plates
were toluene degraders (Jutteau et al., 1999). However. it was not specified at what
stage the study was undertaken although such information is important since the
presence of more non-toluene degraders is expected on start-up than after long term
operation. Concerning stability of microbial populations over time, 16 SrRNA
studies undertaken with a biofilter treating ammonia vapours have shown that over a
102 day period, a shift in dominant populations was observed. A decrease in the
overall diversity of the heterotrophic microbial population was also detected and
about 38% of the unique clones disappeared over the 102 day period (Sakano and
Kerkhof, 1998).
3.3.3. Cel! concentration

Several studies have recently been undertaken in order to evaluate microbi al cell
concentrations in conventional biofilters.
3.3.3.1.Techniques for cel! counts. Different methods have been used for
enumeration of cells and for evaluating attached biomass concentrations in
conventional biofilters. In most cases, the first step consists of suspending a given
amount of carrier in a known volume of aqueous phase. The composition of such
aqueous phase may vary depending on the purpose and the research group. The
sample is then treated. Here again, different alternatives have been used, such as
grounding and later vortexing the sample for supernatant analysis (Wu et al., 1999)
or sonication of the sample (Cârdenas-GonzaJez et al., 1999; Veiga and Kennes,
2001a). After serially diluting the suspension, aliquots are spread on plates allowing
counting specific microbial groups on selective media. Specific inhibitors and
antibiotics may also be used if necessary (Veiga el al., 1997). If one wants to
evaluate the total amount ofbiomass attached on the support, this may be checked by
different techniques, such as estimation of dry mass content or protein concentration.
3.3.3.2. Cel! counts in biojilters. Data have been reported on cell counts in biofilters
containing either organic or inorganic carriers. Cell counts obtained in biofiltration
studies using several composts as natural carriers are reported in Table 3.10
(Cârdenas-Gonzalez et al., 1999). The biofilters were used for the removal of a
mixture of ethyl aleohol, butaraldehyde, ethyl acetate and, 1, l-diethoxybutane. Data

are reported for the start-up period and after 13 months operation. Although the
medium was only slightly acidic, the results indicate significant growth of fungi.
Three biofilters were used. The results shown in the table represent the lowest and
highest ceH counts obtained depending on the biofilter.
Table 3.10. CeH counts in a compost-biofilter
Fungi (cfulg TS ) Heterotrophs (cfulgTs )

Start-up 5 x 103 - 3 x 106 1 X 108 - 7 X lOg
13 months operation 1.7 X 106 - 4.2 X 10 7 1.6 X 108 -1.2 X 109
TS: Total Solids
Another study undertaken with a biofilter treating a mixture of toluene and

ethylacetate using carriers based on compost mixed either with wood chips or
polystyrene spheres led to similar results as in the previous case (Deshusses el al.,
1999). The total number of heterotrophic microorganisms ranged from 1.2 x 109 to
2.3 X 109 cfulg moist medium after one month of operation. In such biofilter, the
ratio ofheterotrophic toluene degrading microorganisms represented between 12.3 %
(wood chips) and 66% (polystyrene spheres) ofthe total count mentioned above. The
concentration of fungi was not estimated. Concerning heterotrophs, other authors
(Juteau el al., 1999) mention the presence of 1.0 x 10 10 cfu/g dry weight in a compost
biofilter removing toluene vapours.
In the case of inert carriers, cell enumeration was undertaken for a
conventional biofilter treating alkylbenzene vapours and fiHed with perlite as single
carrier. With such aluminosilicate carrier, after two months steady state operation,
concentrations of 5 x 10 7 - 1 X 108 cfu/g carrier and 9 x 106 - 3 X lOg cfulg carrier
were obtained, respectively for eukaryotes and prokaryotes (Veiga el al., 1997). The
re1atively high concentration of fungi resulted from, among others, the natural
acidification ofthe medium, reaching pH values around 4 (Kennes el al., 1996).
Several studies have shown that biomass concentration is often not distributed
uniformly along the biofilter and that the nature and concentration of microorganisms
often varies with reactor height (Cox el al., 1997; Song and Kinney, 2000a, 2000b;
Veiga and Kennes, 2001a). Several factors may be involved in such heterogeneity,
such as the decreasing pollutant concentration along the biofilter as air moves
through it or the non homogenous water content in the biofilter, as will be discussed
later. The sequential removal of specific pollutants in multi-compound waste gases is
another factor to be considered, since different microbial populations might degrade
different compounds.
3.4 ROLE OF NUTRIENTS AND OXYGEN
Since the majority of microorganisms found in biofilters require a range of nutrients

for growth, enzyme activity, membrane transport and, above ali, for biodegradation
of pollutants, such nutrients need to be added to the bioreactor. Nutrients may
naturally be present in organic carriers or they may be added by feeding an aqueous
phase.
Microbial cells contain carbon, nitrogen, oxygen and hydrogen as major
elements. Typical cell compositions are (Atkinson and Mavintuna, 1983) C4HgN02
or CSHs3N01.3S for bacteria or C4 H7N o.6 02 for yeasts although vanatlOns are
observed, depending on the species. Therefore such elements need to be available to
the microorganisms. Other important nutrients and trace elements include mainly S,
P, K, Mg, Ca, Fe, Mn and vitamins which may be required by some species.
Phosphorus and potassium may represent up to 5 % of the dry cell weight.
Nutrients are naturally present in organic filter beds although not necessarily
in balanced ratios. This is not the case with most inert and synthetic carriers where
addition of nutrients is always required. It is also worth mentioning that nutrients get
gradually depleted in biofilters packed with natural carriers. After a given period of
operation, addition of nitrogen has shown to improve biofilter performance in the
case of carriers, such as wood bark (Weckhuisen et al., 1993), peat (Arnold el al.,
1997), compost (Morgenroth el al., 1996) or compost-perlite media (Gribbins and
Loehr, 1998). In the latter case, elimination capacity increased from 25 to gre ater
than 100 g toluene/m 3 .h. If increasing nutrient concentration will, to some extent,
improve elimination capacities (Wu el al., 1999), it will nevertheless also enhance
microbial growth and favour pressure drop increase. Although the addition of
nutrients will increase operation costs, it should be taken into account that if such an
addition improves the removal efficiency, smaller biofilters can be used, resulting in
reduced investment costs. It is also expected that natural carriers might be used for
longer periods of time before requiring replacement as a re suit of nutrient depletion.
If the availability of nitrogen, phosphorus, etc. in the biofilter is important,
the form in which it is present is important as well. For instance, the presence or use
ofNH4 CI in non-buffered media willlead to pH drop, since microorganisms will use
nitrogen for growth, resulting in the release of a proton in the medium. A typical
example of methanol biodegradation in presence of ammonium chloride, considering
CSH9N02S as the biomass composition, is given below showing the re1ease of HCI
resulting in medium acidification (Shareefdeen el al., 1993):
Biomass yield, or the amount of cell mass produced per mass of nutrient or
substrate consumed, is also dependent on the nature of the nutrients. Here again,
ammonium will yield higher amounts of cells than, for instance, nitrate (Moo-Young,
1985). When considering toluene as substrate, theoretical biomass yields calculated
based on energetics relationships have been reported to reach 1.26 g biomass / g
toluene and 0.77 g biomass / g toluene, respectively, with ammonia or nitrate as
nitrogen source (Gribbins and Loehr, 1998). However, experimental results may lead
to quite lower yields (Choi el al., 1992; Veiga el al., 1999).
Oxygen is another important substrate naturally present in air. It is required
for the aerobic biodegradation of pollutants. Oxygen concentration needed for
pollutant removal is, in many cases, available in excess (Deshusses el al., 1996),
although it seems that it could become limiting when working at high substrate 10ads
and/or in presence of thick biofilms. Biofilters treating styrene at influent substrate
concentrations around 2.2 g/m3 and a volumetric load of 90 m3/m 3 .h proved to
perform better when working with air enriched with 40% oxygen than in presence of
only 20% O2. For otherwise identical operating conditions, e1imination capacities
were, respectively, 91 and 62 g/m3 .h (Cox el al., 1997). Other authors (Baltzis and
Shareefdeen, 1994) also showed that oxygen might get depleted faster in the biofilm
than the target pollutant, but here again this should also depend on, among others,
contaminant concentrations. It is worth mentioning that in the biofilter treating

styrene cited above (Cox et al., 1997), a mean biofilm thickness of 240-280 Ilm was
reported, with maximal values reaching 600 Ilm, while the effective biofilm
thickness was only 80 Ilm according to modelling equations, and oxygen penetration
depths in biofilms seem to be only 100-200 Ilm at most (Ottengraf and Diks, 1992).
In other cases, oxygen penetration depths of only 40 Ilm have been reported
(Shareefdeen et al., 1993). Depending on the nature of the pollutants to be treated,
for concentration ranges typical of biofilter applications, oxygen seems to be
depleted faster in the biolayer than the pollutant in the case of hydrophobic solvents
(Shareefdeen et al., 1993), while the opposite seems to occur in case of pollutants as
toluene (Shareefdeen and Baltzis, 1994). Although enrichment of air with oxygen to
avoid O2 limitation might, in some cases, improve biofilter performance, such a
procedure might not be cost-effective at industrial scale.
3.5 WATERCONTENT
3.5.1. lrifluence of water content in biofiltration

Waste air fed to biofilters needs first to be saturated with water as microbi al activity
is not possible in a dry air environment. The amount of moisture present in biofilters
is often referred to as water content. Optimising water content is a critical issue in
biofiltration.
Water content can be ca1culated and defined on a wet basis or on a dry basis.
It is not always dear which parameter is used and therefore interpretation and
comparison of published results is sometimes difficult (Kennes and Thalasso,1998).
Both definitions are given below:
mass ofwater
Water content (wet)
mass of water + mass dry carrier
mass of water
Water content (dry)
mass of dry carrier
Water activity, contrary to water content, is more representative of the actual

amount of moisture available to microorganisms. Water activity represents the
amount of water that is free. Pure distilled water has a water activity of 1.0, but if
chemicals or solids are added to the medium, part of the water may get adsorbed to a
solid surface or may be bound by a solute. Both water content and water activity
seem to be important in biofiltration (Cox el al., 1996). In a perlite biofilter used for
styrene removal, a sharp decrease in elimination capacity was observed when
decreasing water content from about 60% down to 25% while maintaining a water
activity of 1.0 (Figure 3.8) (Cox et al., 1996). Nevertheless, water content is the most
frequently used parameter since it can easily be measured and quite dear
relationships have been observed and reported between moisture content and biofilter
performance. On the contrary, high water activities typically found in biofilters are
difficult to measure. Probably many different factors related to moisture content and
other than water content or water activity should be considered to forsee biofilter
perforrnance. For example, in biofilters characterised by near optimal water contents
and high water activities, temporarily feeding dry air to the reactor would probably
damage the external layer of the biofilm and temporarily reduce overall biofilter
performance. The effect of the water content or water activity depends on the nature
of the carrier material. More water is adsorbed and bound to the filter bed in the case
of natural organic supports than with inert carriers, suggesting that less water is
available to microbes in the former than in the IaUer (Bohn and Bohn, 1999).
Most nutrients required by the micro flora found in biofilters are available in
the liquid layer present either in the biofilm itself or otherwise next or close to it. One
of the major differences between conventional biofilters and biotrickling filters is the
presence of a thicker water layer in the IaUer system meaning also an extra mass
transfer resistance (Kennes and Thalasso, 1998). In fact, a free water film may also
be present between the gas phase and the biofilm in conventional biofilters.
Neverthe1ess such film is very thin and basically negligible. Thickness of the water
film on surfaces of inorganic carriers has been estimated to be less than 10 nm (Bohn
and Bohn, 1999). This is about 50 to 5000 times less than the diameter of bacteria or
fungi. In conventional biofilters contrary to biotrickling filters the water layer is
stationary. In both reactors, flow of bulk air is turbulent and, in that region, mass
transfer of pollutants takes place by convection. Contrarily, air flow near the air-
water interface is laminar and, there, the pollutant is transferred by molecular
diffusion which is a slower mass transfer process than convection. Nevertheless,
molecular diffusion is usually not a rate limiting step as the water film is almost
negligible in conventional biofilters.
-
..c: 100
1
--
C')
r •••
E 75 0.75 ~
CD .:;
o::: 50 0.5 ~
--
fi) 10
.s10
~
o~ 25 0.25 ~
O
== O O
O 0.25 0.5 0.75 1
Distance in filter bed (m)
Figure 3.8. Water content (WC, e), water activity (.) and elimination capacity (SR,--)
versus reactor height in a biofilter treating styrene vapours (Cox el al., 1996), (reproduced
with permission).
In open biofilters, regulation of carrier humidity is difficult and highly

dependant on climatic conditions. In closed conventional biofilters, a high enough
moisture content is maintained mainly by feeding air saturated with water. This is
possible by passing the waste gas through a pre-humidification chamber prior to
reaching the inlet port of the biofilter. Saturation levels near 95% are usually
reached, though this will usually not be enough to compensate for water losses
resulting from, among others, volatilisation, enhanced by temperature increases
resulting from microbial activity. Ideally, relative humidity of influent air should
exceed 99% and be as close as possible to 100% since in a such case, theoretically,
air passing through the biofilter could not adsorb any additional water from the filter
bed unless temperature is increased. An aqueous solution often needs to be sprinkled
on top of the reactor as additional moisture regulation since 100% humidity is often
not reached by mere pre-humidification. If required such liquid solution will contain
nutrients, above all in case ofworking with non-natural carriers.
A water content of 40-60% seems to be adequate in many applications and
for most carriers, as shown in Figure 3.9 for a conventional biofilter packed with
perlite (Veiga and Kennes, 2001a). In the latter reactor, the water content remained
in the range 40--60% for at least one month simply by feeding an air stream passing
through a humidification chamber, without the external addition of any aqueous
phase through a separate circuit (Veiga and Kennes, 200Ia). Such water retention
capacity of inorganic carriers also explains why inert hydrophilic carriers, such as
perlite or ceramic can be used in conventional biofilters and not only in trickling
bioreactors. However, depending on the nature of the carrier material, slightly higher
or lower water contents seem to be suitable (Kennes and Thalasso, 1998). Some
examples are the treatment of sulphur compounds in peat biofilters operated at a
moisture content around 70--74% (Hirai el al., 1990) or the removal of propane by
soil beds at a water content of only 20% (Ebinger et al., 1987). The suitability of
such high or low water contents does not necessarily mean that these values are
optimal. Optimal water content is governed to agreat extent by the physical
characteristics of the carrier material, such as the pore size. Water in very small
micropores is hardly available to microorganisms.
--
~
o
>- 80
100
(J
r::::
CI)
'0
it:CI) 60
ca
>
O 40
E
CI)
o::
20
20 30 40 50 60 70
Water content (%)
Figure 3.9. Influence of water content on removal efficiency 111 a biofilter treating
alkylbenzene vapours (Veiga and Kennes, 2001a).
Too high a water content will lead to the formation of stagnant zones with
mass transfer limitation, possible anaerobic conditions owing to poor oxygen
transfer, and increased pressure drop. Increasing moisture content also results in
heavier drain water leaching. Too Iowa water content or water activity will reduce
microbial activity although some microorganisms, such as fungi are more tolerant
than others to such conditions. Although most often water activities above 0.9 seem
to be required for bacterial activity, fungi tolerate lower values. Many yeasts and
filamentous fungi are metabolically active at water activities around 0.80 or less. In
the case of hydrophobic compounds, such as a-pinene it was observed that a low
water content increased pollutant concentration in the wet biolayer up to toxic levels
(Apel et al., 1995).
The initial moisture content of the carrier material on starting-up the system
seems to be another relevant factor. Media which are not sufficient1y moistened
when preparing the packing might not retain enough water later on and might never
reach a high enough moisture content to ensure optimal operation. Compost-perlite
biofilters started-up at water contents of 0% and 20% and continuously fed a
humidified gas stream (95% humidity) reached, respectively, water contents of only .
4% and 9% at equilibrium which is far too low (Quinlan et al., 1999). As a general
rule, natural organic carriers are more sensitive to water content variations than inert
or inorganic carriers, among others because the structure of most natural carriers will
(sometimes irreversibly) change during drying periods impeding proper rewetting
later ono A similar problem was reported in case of a peat biofilter treating ethanol
vapours (Auri a et al., 1998). Carrier materials with hydrophobic surfaces like
compost have high air-water-solid contact angles making them difficult to rewet
when dry (Bohn and Bohn, 1999; Kraislas et al., 2000). This is also true for peat,
wood chips, activated carbon, etc. Such hydrophobic surfaces repel water spreading
and pore entry. Once dry, replacement of the filter bed is, in most cases, the best
choice in order to avoid poor performance. Hydrophilic surfaces, such as soil,
ceramic or perlite have very low contact angles and (re)wet easily.
3.5.2. Techniques for measuring and regulating water content

Several techniques are available for estimating water content in lab-scale and
industrial-scale biofilters. Two techniques are most widely used. One such method
consists in weighing the whole filter bed with a load ceH (van Lith et al., 1990). The
initial mass of dry carrier on starting-up the biofilter is a constant parameter and is
usually known. By weighing the filter bed at different time intervals the mass of
water can be estimated allowing the calculation of water content at such intervals.
Depending on the weight measured by the load cell sprinklers can automatically be
tumed on, if necessary, in order to readjust the moisture content to an appropriate
value. It should be taken into account that high pressure drops may lead to erroneous
mass estimations unless applying a correction factor for calculating the water
content. It should also be noted that this method gives a mean water content since
that parameter may vary with bed depth.
Another possibility often used at lab-scale consists in removing representative
carrier samples from the biofilter. The samples are then dried in an oven at 105 aC
until reaching constant weight. The difference between the initial weight and the
weight after drying represents the mass of water originally present in the sample
aHowing the ca1culation ofthe water content.
80 C Kennes and M C Veiga
3.6 TEMPERATURE
Most biofiltration studies have been carried out under mesophilic conditions, often at
ambient temperature, since many microorganisms are grow under such conditions.
However, waste gas temperatures are very frequently different and most often higher
than temperature ranges corresponding to the optimum activity of mesophilic
microorganisms. On the other hand, in cold regions waste gases may reach
temperature ranges more typical of psychrophilic organisms « 20 aC). It is
sometimes suggestcd that microbial growth rates double with a 10 aC temperature
rise, although this is generally true only in a limited temperature range, since for
temperature increases or decreases of about 20-30 aC above or below the optimal,
growth and substrate biodegradation are very often impossible (Figure 3.10). A
similar trcnd is often observed for biodegradation kinetics as for growth. Biofiltration
at low temperatures has been reported and may be useful in cold regions.
Conventional biofilters operating efficiently at temperatures below 10 ac have been
described, for instance, in the case of H2 S removal (Cho el al., 1992).
Microbial
activity
Psychrophilic Mesophilic Thermophilic
20 40 60
Temperature (0C)
Figure 3.10. Inf1uence oftemperature on microbial activity and biodegradation rates.
Several microorganisms are known to de grade pollutants under thermophilic

conditions. Biofiltration at temperatures above 37 uC is thereforc cxpected to be
possible, although it is only recently that such an alternative was studied. Cooling
down of hot gases to mesophilic temperatures is often not advisable since it might
significantly increase biofiltration costs. Hot gases may be diluted with ambient air in
order to lower their temperature although this would also affect flow rates, EBRT
and thus biofilter performance. The humidification process used as pre-treatment step
may be used as well for lowering inlet waste gas temperature. Similar methods can
be used for raising the temperature of cold waste gasses. Whenever possible,
thermophilic biofiltration may be a more cost effective alternative when dcaling with
hot gasses. Thermophilic biofiltration was recently successfully applied on, for
example, the removal oflow VOC concentrations at temperaturcs between 40-55 uc

in compost and wood bark biofilters (Knauf and limmer, 1994) or on the removal of
toluene at 50 °c and 60 aC in compost biofilters (Matteau and Ramsay, 1999),
Thermophilic toluene biofiltration proved even to be more efficient than under
mesophilic conditions, reaching maximal elimination capacities of 110 and 89 g/m 3 .h
under thermophilic and mesophilic conditions, respectively (Matteau and Ramsay,
1997).
Biofiltration at relatively high temperatures generates new challenges. Higher
temperatures result in higher water evaporation rates. Carrier materials arc then
prone to dry out fasteL This, in turn, may lead to channelling problems and high
pressure drops. It may also be hard to maintain a homogenous water content along
the biofilter. Another potential problem is the higher volatility or lower solubility of
pollutants and oxygen at higher temperatures. Partition coefficients and mass transfer
rates from the gas phase to the biofilm would be less favourable than at lower
temperatures. From a microbiological point of view, higher temperatures are
expected to reduce microbi al diversity.
Regarding temperature stability, as already mentioned biodegradation
reactions are exothermic. As a matter of fact temperature will increase in biofilters.
Typical temperature increases are between 2-10 aC (Morales et al., 1998). A higher
metabolic activity may often be detected near the inlet of the biofilter where higher
substrate concentrations are found. Occasionally pollutant concentrations may be
negligible before reaching the outlet of the biofilter where exothermic reactions
would be minimal. Therefore, drying out of the filter bed will appear sooner near the
inlet of the reactoL Since water is always added on top of the biofilter where the
sprinklers are located, downf1ow operation usually is more indicated in man)
applications, above ali in closed biofilters.
3.7 pH
A similar trend is observed for pH as for temperature. Biodegradation rates and

microbial growth take place over a quite limited pH range for mosI microbial
species. Microorganisms most oftcn do not toleratc pH fluctuations of more than
about 2 or 3 pH units for maintaining appreciable growth and biodegradation rates. lf
mixed microbi al communities are present as in mosI biofilters, biodegradation might
still be possible if pH variations result in a shift in dominant microbi al populations
maintaining similar biodegradation characteristics. Most bacteria grow over the pH
range 4-8 while yeasts and moulds are able to grow under more acidic conditions
(pH 2-7). This is only a generalisation since few bacteria, among which some
Thiobacil/us spp., are able to grow under highly acidic conditions, which is
advantageously used in biofiltration of hydrogen sulphide. However, one should
keep in mind that when mixed waste gases are to be treated biodegradation of
pollutants other than H2 S, even sulphur compounds, might be inhibited under acidic
conditions. Biofiltration of styrene (Cox el al., 1997) and alkylbenzenes (Kennes el
al., 1996; Veiga and Kennes, 2001a) have also proven to be possible at low pH
values (around 4) favourable to fungal growth or to the growth of mixed populations
of bacteria and fungi. In both cases elimination capacities around 70 g/m 3 h were
reached with removal efficiencies above 95%. Alkylbenzene elimination capacities
above 120 g/m 3 .h with > 99% removal efficiency were reported at such 10\\ pH as
well (Veiga and Kennes, 200Ia). Adaptation to low pH values has also been reported
for a compost-biofilter dominated by bacterial populations and treating acetone,

alkylbenzenes and hydrogen sulphide (Webster el al., 1997).
Since biofilter performance is often pH-sensitive it is useful to try and
maintain quite constant pH values. Biodegradation of compounds containing
chlorine, sulphur or nitrogen are often expected to foment pH drop through the
formation of acidifying products. Contrary to biotrickling filters where pH regulation
by means of a continuous trickling phase is quite easy, in conventional biofilters pH
regulation is most easily done through mixing specific compounds or chemi cais to
the carrier material when packing tlie reactor. Limestone, crushed oyster shells,
Ca(OH)2 or any other buffering compounds have ali occasionally been used as
additives for pH regulation. The buffering effect is limited in time requiring
replacement ofthe carrier material after exhaustion ofthe buffering capacity.
The effect of adding limestone to a compost biofilter treating dimethyl
sulphide can be given as an example. Dimethyl sulphide biodegradation leads to
medium acidification through the following reaction (Smet et al., 1996b):
Medium acidification through sulphuric acid release may partly be avoided

through addition of limestone:
The addition of 1 mole CaC03 is theoretically needed for neutralising 1 mole

H2S04 produced from the biodegradation of 1 mole dimethyl sulphide. Experimental
studies proved that such ratios were also basically adequate in practice.
In the case of pollutants containing nitrogen, as in the case of ammonia, a
drop in pH results from the following biological reaction:
Here again the addition of buffering agents will allow neutralising the
medium. The addition of chemicals, such as sodium hydroxide for neutralising pH is
not recommended as it might adversely affect the overall removal efficiency through
the accumulation of high salt concentrations and increased ionic strength (Dolfing et
al., 1993).
Contrary to synthetic carriers, some organic filter beds are naturally buffered
and pH fluctuations are reduced to a minimum. In the case of mixed waste gases, the
nature ofthe different pollutants in the air may present a neutralising effect. In a pilot
biofilter treating mixtures of sulphur compounds and ammonia, reaction between
sol- and NH3 allowed maintaining a quite constant pH (Cho et al., 1992).
3.8 PRESSURE DROP AND CLOGGING
Low pressure drop is a common feature of all filter beds during the start-up period.
Problems might show up later and are dependant on many parameters, such as the
nature and characteristics ofthe carrier and biofilm growth.
Pressure drop in packed bed reactors is a well known phenomenon in
chemical engineering. Equations such as the Ergun equation have been used to
quantify it taking into account the major parameters affecting head losses. In packed
bed bioreactors such as gas phase biofilters these parameters include: the nature of
the carrier material, particle size and shape, superficial gas velocity, biomass growth
and water content. In the section dedicated to the different types of carrier materials
used in biofiltration it was already mentioned that synthetic supports generally lead
to lower pressure drops than natural organic ones. A non-uniform flow distribution
and channelling are also more typical of organic carriers than of structured synthetic
ones. The addition ofbulking agents to natural organic carriers allows slowing down
pressure drop increase and optimising flow characteristics in conventional biofilters.
Such compounds include, among others, woods chips, bark, heather, lava particles,
perlite, glass beads or polystyrene. Their effect is shown in Figure 3.11 for compost
and for a compost-perlite mixture. Pressure drop with perlite aloţle was negligible.
Other examples can be found elsewhere (Kennes and Thalasso, 1998). These figures
also illustrate the effect of the gas phase flow rate or superficial gas velocity on
pressure drop. It can easily be observed that at higher gas velocities, head losses will
also increase. The choice of an adequate support is therefore a critical point.
2.5
c. compost
,,-
2
E
2
:;n; 1.5
Q)
I/)D-
I/)~ compost
...
Q)
D- 0.5
+ perlite
O
O 100 200 300 400
Superficial gas velocity (m/h)
Figure 3.11. Pressure drop in a compost biofilter and in a compost-perlite biofilter.
A homogenous carrier material with a regular shape, such as most inert

carriers, will, in most cases, slow down the pressure drop increase in biofilters.
Particle size is another relevant factor. Smaller particles are characterised by higher
surface areas than larger ones for a given working volume. However, small support
particles generate higher pressure drops (Figure 3.12) (Yang and Allen, 1994a).
Extensive and fast growth of microorganisms will also lead to faster pressure
drop increases and possible clogging. This is a typical situation observed with inert
carriers, when a relatively rich nutrient solution is fed on a regular or continuous
basis, which is more typical of biotrickling filters. In such cases many growing
microorganisms are present rather than resting cells, as often found in natural
carriers. Since this is more typical of trickling biofilters, the reader will find more
detailed information on this aspect in Chapter 4 dedicated to the latter reactor design.
30
-
-E 25
ca
-...
C.
.:.:.
C-
o
20
15
"C
...CI>
:::l 10
1/)
1/)
~ 5
c.
O
[A] [B] [C] [O] [E]
Parti ele size range
Figure 3.12. lnfluence of particle size of the filter bed on pressure drop (Yang and Allen,
1994a) (reproduced with permission). Particle size (mm): [A] > 12, [B] = 3.35-12. [el =
2.36-3.35, [D] = 1.18-2.36, [E] < 1.18. Gas velocities (m/h): • = 396,. = 792.
Microorganisms found in biofilters are most frequently chemotrophs. In

environments containing contaminants which may be used as carbon and energy
sources, which is often the case with VOCs, two metabolic steps will take place. The
energy generating step resulting from the presence of an energy source is called
catabolism while the biosynthetic step is called anabo!ism and results in biomass
generation from a carbon source. Theoretically one way to reduce biomass growth
and slow down clogging problems would consist in minimising anabolism. However,
the problem is quite complex since both steps are linked. In complex media
containing nutrients or growth factors which may be used as precursors, more of the
pollutant will be minera!ised with the concomitant energy generation. Nevertheless,
this will not result in less biomass growth, simply cell synthesis will take place
through the use of precursors present in the medium. Means to regulate growth
through the control of nutrient addition is not easy and still needs to he optimi sed.
Very often, limiting the presence of a given nutrient (nitrogen. phosphorus source.
etc.) will !imit biomass growth but it will also reduce removal efficiency and
elimination capacity.
4. Costs
Air treatment in biofilters is rather cheap compared to other non-biological treatment

technologies and many other, often more complex, bioreactor technologies.
Investment and operating costs are always dependant on the characteristics of the
waste stream and it is thus difficult to present one simple, general comparative table.
Costs should always be estimated through a case by case study. Technologies that
might be cheap for a given set of characteristics such as high flow rate or low
Conventional Bio(ilters 85
poIlutant concentration may become prohibitive under another set of conditions. One
should also bear in mind that the different technologies are not always suitable for
the same range of waste gas characteristics (see also Figure 2.1). Parameters
affecting operating and capital costs of a given technique might be insignificant for
another one. Recalcitrance to biological degradation which is a key parameter in the
case of biofiltration has no economic al impact on any other non-biologic al technique.
The most significant parameter linked to investment costs in biofiltration is
the price of the filter bed or carrier material. Prices may largely vary from less than
100 Euros / m3 for most natural carriers (soil, compost, wood bark, heather, etc.) to
more than 1000 Euros / m3 for some synthetic or non-natural carriers. Prices may
also slightly vary from one country to another. Some examples are listed below
(orientative prices given in Euros, year 2000):
Price (€)
Compost 40-50/ m3
Heather 100-120/m3
Wood bark 60-75/ m3
Activated carbon 6-9/ kg
For larger reactors higher volumes of filter material need to be purchased.

Hardly or slowly biodegradable poIlutants wiIl need higher retention times or larger
biofilters leading to increased capital costs. The same is true for poIlutant load. For a
given t10w rate, at higher poIlutant concentrations (i.e., higher loading rates) the
contaminated air stream must spend more time in the presence of the biocatalyst
which is most often achieved by using larger reactors. Similarly, for a given pollutant
concentration, higher t10w rates (i.e., higher loading rates) require larger reactors.
Nevertheless, it should be mentioned that although higher investment costs are
required when increasing biofilter size, the capital cost per unit filter bed volume
(Euros/m3 packing) gradually decreases while increasing reactor size.
Parameters contributing to operating costs include pre- or post-treatment,
reactor size, electricity consumption, monitoring and control, filter bed replacement,
water/nutrient consumption and, drainage water treatment. Larger biofilters usually
require the consumption of more electricity, water and nutrients.
Some data coIlected from the literature are summarised in Table 3.11. AIl
data have been converted to the same currency (Euros, €) (one Euro is equal to about
0.90-0.95 US dollar) and prices have been updated Cyear 2000), taking into account
realistic int1ation rates. Although some deviations are sometimes observed, most
reported data from different sources are in a same range of values for a given
technology. The data also agree with inforrnation obtained from some vendors of
biological waste gas treatment systems and with othcr reports available in the
literature. Costs and prices may slightly vary from one country to another. The higher
operating costs reported by Leson and Winer (1991) for biofilters built in Europe
compared to United-States reflect the higher costs of electricity in Europe.
When comparing costs appearing in Table 3.11, biofiltration stands out as onc
of the most cost effective treatment alternatives. Typical capital costs range from 10
to 25 Euros per cubic meter and hour while typical operating costs range
approximately from 0.3 to 4 € / 1000 m3 waste gas in Europe and from 0.1 to 3 € /
1000 m3 in the United States.
Table 3.11. Typical investment and operating costs for different waste gas treatment
technologies
Capital costs Operating costs Total costs

(€Im 3 .h) (€/1000m 3) (€11000m 3) References
Thermal 16.5-19.5 2.0-2.4 (1)

incineration
16.ia) (1)
5.7-7.4(b) (1)
Recuperative: Recuperati ve:

0.8-166<C) 0.25-1.5<C) (2)
Regenerative: Regenerative:
24.8-373<c) 0.35-2.5(c)
17-105(d) (3)
Catalytic 19.5-22.5 1.8-2.1 (1)

incineration
4.9-6.5(b) (1)
Fixed bed: Fixed bed:

16.9-20ic) 0.2-1.3(c) (2)
Fluidised bed: Fluidised bed:
29.5-1 83(c) 0.25-1.4(C)
22.5-67.5(d) (3)
Adsorption 7-28 0.7-1.4 (1)

2.7(e) (1)
11.4-14.6(b) (1)
1.1 (1) (4)
12.7-98.5(c) (2)
Absorption 11.2-14 l.1-1.4 (1)

7.5(g) (1)
12.7-57.7 0.4-2.0 (2)
7-140(d) (3)
Condensation 8.4-66 0.35-2.0 (2)

BiofiItration 4.2-13.9 (open) 0.4-0.7 (open) (1)
0.4-2.4 (closed) (1)
5-150 0.1-3.0 (5)
5.9-13.7 0.3-4.3 (3)
0.32-0.80 (6)
(Europe)
0.16-0.32
(USA)
(a) fuel costs only, (b) for 0.1-2.0 gvoc/m3 and 10000 m3/h with 50% energy rccovery in
case of incineration and steam regeneration in case of adsorption, (c) for 50000 m 3/h, (d)
concentration dependant, (e) including rege ner. incineration, (f) with regeneration, (g)
chlorine.
(1) Ottengraf and Diks, 1992; (2) Ruddy and Carroll, 1993; (3) van Groenestijn and
Hesselink, 1993; (4) Bohn, 1992; (5) Devinny el al., 1999; (6) Leson and Winer, 1991.
S. Applications
5.1 LAB-SCALE RESEARCH
Research devoted to optimising biofilter performance has increased exponentially

over the past two decades and more results are available at lab-scale level than at
industrial scale, despite the relatively large number of full-scale biofilters operating
successfully nowadays. Lab-scale studies are often focussing on single pollutant
waste streams, which allows easy comparison and deve\opment of engineering
principles, which can later be tested and eventually applied to mixed waste gases and
real cases.
Some examples of recent results reported in the literaturc are listed in Table
3.12. Other less recent studies can be found in some of the review papers published
over the past decade.
5.2 FULL-SCALE APPLlCATIONS
Biofilters have several applications. They may be used for the treatment of industrial
waste gases and for the removal of volatile compounds emitted by waste water
treatment plants. They may also be used in soil and groundwater remediation in
combination with technologies such as soil vapour extraction and air stripping where
contaminants present in a solid phase (soil) or in a liquid phase (aquifers) are
transferred to a gas phase (air). Air pollution generated at composting facilities or
through fugiti ve emissions from storage tanks are other examples of full-scale
biofilter applications.
Table 3.12. Examples of applications of biofiltration technology
Pollutant Carrier Inoculum or Max. Removal References

material dominant elimination efficiency
organisms capacitl) (%)
Ammonia Wood bark Mixed culture 83 Weckhuysen
et al., 1994
Ammonia Fuyolite Vi brio 22.5 gN/kg.d 85 Kim el al.,
alginolylicus 2000
BTEX Soil Mixed culture 5 8-39 Miller and
Canter,1991
Butanal Wood bark Mixed culture 90 97 Weckhuysen
el al., 1993
Cresol Compost- 70 gcf m3 .h Windsperger
bark el al., 1990
CS 2 208 Windspergcr,
1992
Dimethyl Peat Hyphomicrobium 0.59 g/kg.d(2) Zhang el al.,
sulphide sp. 1991
Dimethyl Compost Mixed 500 glm 3 d (4) > 99 Smet el al.,
sulphide enrichment 1996"
culture
Dimethyl Wood bark Hyphomicrobium 36 g/m 3 d 44 Smet el al.,
sulphide sp. 1996b
Dimethyl Compost Hyphomicrobium 362 gDMS/m 3 d 76 (DMS) Smet el al.,
sulphide + sp. 3710 100 1997
Iso- glBUT/m1.d (IBUT)
Butaraldehyde
Dimethyl Fibrous peat Mixed culture 0.68 g/kg.d(2) Cho el al.,
disulphide 1991
Ethanol Granular Mixed culture 53-219 Hodge and
activated Devinny,
carbon 1994
Ethanol Peat 30 35-40 Auria et al.,
1998
Ethene Granular Mycobaclerium 10.4 84.9 De heyder el
activated sp. al., 1994
carbon
Ethylacetate Compost- Mixed culture 175 Windsperger
bark el al., 1990
Ethylacetate + Peat-compost Mixed culture 20-40 (each) Kiared el al.,
Toluene + 1996
Butylacetate +
Butanol
(20:58: 15:7)
Ethylbenzene Perlite Mixed culture 85 > 95 Kennes el al.,
1996
Gasoline: Soil Mixed culture 50 Togna and
BTEX+ Singh,1994
Aliphatics
Hexane Compost 21 > 99 Morgenroth el
al.,1996
Hydrocarbon Activated Mixed culture 5.5 Hodge el al.,

fuel vapours carbon or 1991
diatomaceous
earth
Hydrogen Compost 130 Yang and
sulphide gsulfu/ mJ .h Allen, 1994a
Hydrogen Ceramic Mixed culture 26.5 > 99 Shinabe el al.,
sulphide mmollm'.d 1995
(in presence
ofMT)
Hydrogen Ca-alginate Thiobaci/lus sp. 47 > 95 Chung el al.,
sulphide mmol/kg.d(J) 1997
Hydrogen Compost-hog 120 30 Wani el al.,
sulphide fuel-perlite 1999
(2:2: 1)
Methanol Peat-per1ite Defined 112.8 Shareefdeen
(2:3) consortium el al., 1993
(8 bacteria)
Methanol Compost 101 60 Kraislas el al.,
2000
Methanol + Wood chips- 40 (MeOH) > 90 (both) Mohseni and
a-pinene compost- 10-12 (PIN) Allen, 1999
perlite (3:3:2)
Methanol + Compost Mixed cui ture 110 Fischer, 1992
styrene
Methyl- Activated Mixed culture 0.48 glkg.d(2) > 99 Lee and
mercaptan carbon fabric Shoda, 1989
MEK+ MIBK Compost- 50 (MEK) > 90 (both) Deshusses
PS spheres- 20 (MIBK) and Hamer,
1imestone 1993
MEK Compost- 121 > 90 Deshusses el
PS spheres- al., 1995b
limestone
MIBK Compost- 25 50-55 Deshusses el
PS spheres- al.,1995b
limestone
Pheno! Peat-glass Pseudomonas 700 24.5 Zilli el al.,
beads (2: 1) pulida 1996
a-pinene Aspen wood Pseudomonas 32 85-90 Kleinheinz el
chips fluorescens (D) + al., 1999
Alcaligenes
xylosoxydans (D)
Styrene Mixed culture 70 > 95 Togna and
Singh,1994
Styrene Perlite Exophiala 80 > 80 Cox el al.,
jeanselmei (D) 1994
Styrene Peat Mixed cuI ture 28 92 Arnold el al.,
1997
Styrene Perlite Exophiala 62 > 90 Cox el al.,
jeanselmei (O) 1997
TEX Perlite Mixed culture 70 > 95 Kennes el al.,

1996
TEX Perlite Pseudomonas sp. 120 100 Veigaand
+ Bacil/us sp. + Kennes,
Trichosporon 2001a
beigelei
Toluene Compost- 25 Weberand
activated Hartmans,
carbon 1995
Toluene Perlite Mixed culture 73 > 95 Kennes el al.,
1996
Toluene GAC- Adapted mixed 97 80-85 Hwang and
compost culture Tang, 1997
(2: 1)
Toluene GAC- Adapted mixed 58 75 Hwangand
compost culture Tang, 1997
(2:1)
Toluene Peat 2 Pseudomonas 165 71 J aria el al.,
spp. + 1998
Rhodococcus sp.
+ Arlhrobacler
sp.
Toluene Peat Defined mixed 25 31 Morales el al.,
culture 1998
Toluene Compost- 80 65 Gribbins and
perlite (7:3) Loehr, 1998
Toluene Peat beads 2 Pseudomonas 135 39 Wu el al.,
spp. + 1999
Rhodococcus sp.
+ Arthrobacter
paraffineus
TX Peat 2 Pseudomonas 115 Jorio el al.,
spp. + 1998
Rhodococcus sp.
+ Arlhrobacler
sp.
Triethylamine Compost- 140 100 Tang el al.,
chaff particles 1996
o-Xylene Perlite Mixed culture 64 > 95 Kennes el al.,
1996
Xylenes Peat 2 Pseudomonas 66 33 Jorio el al.,
spp. + 1998
Rhodococcus sp.
+ Arthrobacler
sp.
(llgpolllltan,!m3 h (unless otherwise specified), (2lgpollutan,!kgcarnet d, (3lmmols/kgdr, h,d d. (4 lshort

period or non steady-state operation, B = Benzene, MT = Methanethiol, !BUT =
Isobutaraldehyde, D = Dominant (not inoculated) strains, DMS = Dimethyl sulphide, E =
Ethylbenzene, GAC = Granular Activated Carbon, MeOH = Methanol, MEK = Methyl Ethy1
Ketone, M!BK = Methyl IsoButyl Ketone, PIN = a<-pinene, PS = Polystyrene, T = Toluene,
X=Xylene.
Conventional Bioftlters 91
Refereuces
Abumaizar, RJ., Smith, E.H. aud Kocher, W. 1997. Ana1ytical model of dual-media biofilter for
removal of organic air pollutants. J. Environ. Eng. 123: 606-614.
Acufia, M.E., Perez, F., Auria, R. And Revah, S. 1999. Microbiological and kinetic aspects of a
biofilter for the removal oftoluene from waste gases. Biotechnol. Bioeng. 63: 175-184.
Alexander, M. 1977. Introduction to soi1 microbiology, 2"d Edition. J. Wi1ey & Sons, New York.
Allen, E.R. and Phatak, S. Control of organo-sulfur compound emissions using biofiltration - methyl
mercaptan. In: Proceedings of the 86 th Annual Meeting & Exhibition of the Air & Waste Manage.
Assoc., 1993, Denver, Colorado.
Allen, SJ., Whitten, LJ., Murray, M., Duggan, O. and Brown, P. 1997. The adsorption of pollutants
by peat, lignite and activated chars. J. Chem. Technol. Biotechnol. 68: 442--452.
Amor, L., Kennes, C. and Veiga, M.C. 2001. Kinetics of inhibition in the biodegradation of
monoaromatic hydrocarbons in presence ofheavy metals. Biores. Technol. 78: 181-185.
Andrews, J.F. 1968. A mathematical model for the continuous culture of microorganisms utilizing
inhibiting substrates. Biotechnol. Bioeng. 10: 845-864.
Anselno, A.M. , Mateus, M., Cabral, J.M.S. and Novais, J.M. 1985. Degradation 01 phenol by
immobilized ce Ils of Fusariumflocciferum. Biotechnol. Le!. 7: 889-894.
Apel, W.A., Lee, B.D., Walton, M.R., Cook, L.L. and Dinerstein, K.B. Removal of cx-pinene from
off-gasses by biofiltration. In: Proceedings of the 88 1h Annual Meeting & Exhibition of the Air &
Waste Manage. Assoc., 1995, San Antonio.
Arno1d, M., Reittu, A., von Wright, A. Martikainen, PJ. and Suihko, M.-L. 1997. Bacteria1
degradation ofstyrene in waste gases using a peat filter. Appl. Microbiol. Biotechnol. 48: 738-744.
Atkinson, B. and Mavintuna, F. 1983. Biochemical engineering and biotechnology Handbook. Nature
Press, New York.
Auria, R., Aycaguer, A.-C. and Devinny, J.S. 1998. Influence of water content on degradation rates
for ethanoI in biofiltration. J. Air Waste Manage. Assoc. 48: 65-70.
Ba1tzis, B.C. and Shareefdeen, Z. Biofiltration of VOC mixtures: mode ling and pilot sca1e
experimental verification. In: Proceedings of the 87'h Annual Meeting & Exhibition of the Air &
Waste Manage. Assoc., 1994, Pittsburg, Pa.
Bohn, H.L. Biofilter media. In: Proceedings of the 891h Annual Meeting & Exhibition of the Air &
Waste Manage. Assoc., 1996, Nashville, Tennessee.
Bohn, H.L. and Bohn, R. 1988. Soil beds weed out pollutants. Chem. Eng. April, 25: 73-76.
Bohn, H. 1992. Consider biofiltration for decontaminating gases. Chem. Eng. Prog. April: 34--40.
Bohn, H.L. and Bohn, K.H. 1999. Moisture in biofilters. Environ. Prog. 18: 156-161.
Braun-Ltillemann, A., Majcherczyk, A., Tebbe, N. and Hiittermann, A. Bioluftfilter auf der Basis von
WeijJfaulepilzen. In: Biotechniques for Air Pollution Abatement and Odour Control Policies. Dragt
AJ. and van Ham J. (eds.). Elsevier, Amsterdam, The Netherlands, 1992, pp. 91-95.
Braun-Liillemann, A., Majcherczyk, A. and Hiittermann, A. 1997. Degradation of styrene by white-rot

fungi. Appl. Microbiol. Biotechnol. 47: 150-155
Cârdenas-Gonzâlez, B., Ergas, SJ. and Switzenbaum, M.S. 1999. Characterization of compost
biofiltration media. J. Air Waste Manage. Assoc. 49: 784-793.
Cemiglia, C.E. 1992. Biodegradation ofpolycyclic aromatic hydrocarbons. Biodegradation. 3: 351-

368.
Chang, M-K., Voice, T.C. and Criddle, C.S. 1993. Kinetics of competItIve inhibition and
cometabolism in the biodegradation of benzene, toluene, and p-xylene by two Pseudomonas isolales.
Biotechnol. Bioeng. 41: 1057-1065.
Cho, K., Hirai, M. and Shoda, M. 1991. Removal ofdimethyl disulfide by Ihe peal seeded with night
soi! sludge. J. Ferment. Bioeng. 71: 289-291.
Cho, K., Hirai, M. and Shoda, M. 1992. Enhanced removal efficiency of malodorous gases in a pilot-
scale peat biofilter inoculated with Thiobacil/us thioparus DW44. J. Ferment. Bioeng. 73: 46-50.
Choi, Y-B., Lee, J-Y. and Kim, H-S. 1992. A novel bioreactor for the biodegradation of inhibitory
aromatic solvents. Biotechnol. Bioeng. 40: 1403-1411.
Chou, M. and Cheng, W. 1997. Screening of biofiltering material for VOC treatment. J. Air Waste
Manage. Assoc. 47: 674-681.
Chung, Y.-c., Huang, C. and Tseng, c.-P. 1997. Removal of hydrogen sulfide by immobilized
Thiobacillus sp. strain CHil in a biofilter. J. Chem. Technol. Biotechnol. 69: 58-62.
Cox, H.H.J., Doddema, H.J. and Harder, W. 1994. Biological removal of styrene from waste gases
using fungi on inert support materials. VDl Berichte. 1104: 289-299.
Cox, H.H.J., Magielsen, F.J., Doddema, HJ. and Harder, W. 1996. Influence ofthe water content and
water activity on styrene degradation by Exophiala jeanselmei in biofilters. Appl. Microbiol.
Biotechnol. 45: 851-856.
Cox, H.H.J., Moerman, R.E., van Baalen, S., van Heiningen, W.N.M., Doddema, H.J. and Harder, W.
1997. Performance of a styrene-degrading biofilter containing the yeast Exophia/a jeanselmei.
De Heyder, B., Overmeire, A., Van Langenhove, H. and Verstraete, W. 1994. Ethene removal from a
synthetic waste gas using a dry biobed. Biotechnol. Bioeng. 44: 642-648
Deshusses, M.A. and Hamer, G. 1993. The removal of volatile ketone mixtures from air in biofilters.
Bioprocess Eng. 9: 141-146.
Deshusses, M.A., Hamer, G. and Dunn, I.J. 1995a. Behavior of biofilters for waste air treatment. 1.
Dynamic model development. Environ. Sci. Technol. 29: 1048-1058.
Deshusses, M.A., Hamer, G. and Dunn, I.J. 1995b. Behavior of biofilters for waste air treatment. 2.
Experimental evaluation ofa dynamic model. Environ. Sci. Technol. 29: 1059-1068.
Deshusses, M.A., Hamer, G. and Dunn, 1.1. 1996. Transient-state behavior of a biofilter removing
mixtures ofvapors ofMEK and MIBK from air. Biotechnol. Bioeng. 49: 587-598.
Deshusses, M., Johnson, C.T. and Leson, G. 1999. Biofiltration of high loads of ethylacetate in the
presence oftoluene. J. Air Waste Manage. Assoc. 49: 973-979.
Devinny, J.S. and Hodge, D.S. 1995. Formation of acidic and toxic intermediates in overloaded
ethanol biofilters. J. Air Waste Manage. Assoc. 45: 125-131.
Devinny, J.S., Deshusses, M.A. and Webster, T.S. 1999. Biofiltration for air pollution control, Lewis
Publisher, Boca Raton.
Diks, R.M.M. and Ottengraf, S.P.P. 1991. Verification studies of a simplified model for the removal
of dichloromethane ITom waste gases using a biological trickling filter (Part I and Part II). Bioprocess
Eng. 6: 93-99 and 131-140.
Do1fing, J., van den Wijngaard, AJ. and Janssen, D.B. 1993. Microbiological aspects ofthe removal
of chlorinated hydrocarbons from air. Biodegradation. 4: 261-282.
Ebinger, M.H., Bohn, H.L. and Puls, R.W. 1987. Propane removal from propane-air mixture, by ,ati
beds. J. Air PolI. Control Assoc. 37: 1486 -1489.
Edmonds, P. and Cooney, J.1. 1967. Identification of microorgansisms isolated from jet fuel systems.
Appl. Microbiol. 15: 411-416.
Engelhardt, G., Tillmans, G., Wallnofer, P.R. and Hutzinger, O. 1977. Biodegradation of di-idobutyl
phtalate and related dialky1 phtalates by Penicillium lilacinum. Chemosphere. 6: 347-354.
Ergas, S.1., Kinney, K., Fuller, M.E. and Seow, K.M. 1994. Characterization of a compast biofiltration
system degrading dichloromethane. Biotechnol. Bioeng. 44: 1048-1054.
Ergas, S.1., Schroeder, E.D., Chang, D.P.Y. and Morton, R.L. 1995. Control of volatile organic
compound emissions using a eompost biofilter. Water. Environ. Res. 67: 816-821.
Ferreira Jorge, R.M. and Livingston, A.G. 1999. A novel method for eharaeterisation of microbial
growth kinetics on volatile organic compounds. Appl. Mierobiol. Biotechnol. 52: 174-178.
Fischer, K. Vergleiehende Untersuehungen: Biofilter oder Biomembranverfahren zur reinigung

Liisemittelhaltiger Industrieabluft. In: Biotechniques for Air Pollution Abatement and Odour Control
Polieies. Dragt A.J. and van Ham J. (eds.). E1sevier, Amsterdam, The Netherlands, 1992, pp. 97-102.
Freitas dos Santos, L.M. and Livingston, A.G. 1995. Novel membrane bioreaetor for detoxifieation of
VOC wastewaters: biodegradation of 1,2-diehloroethane. Water Res. 29: 179-194.
Fritsche, K. and Lechner, U. Einsatz eines ehlorphcnolverwertcnden bakterienstammes zur

Abluftreinigung Biofilter. In: Biotechniques for Air Pollution Abatment and Odous Control Policies.
Dragt AJ. and van Ham J. (eds.), Elsevier, Amsterdam, The Netherlands, 1992, pp. 287-292.
Froment, G.F. and Bisehoft; K.B., 1990, Chemieal Reactor Analysis and Design, Chapter 7, 2nd
Edition,1. Wiley and Sons, New York, USA.
Frye, R.J., We1sh, D., Berry, T.M., Stevenson, B.A. and MeCallum, T. 1992. Removal of contaminant
organic gases from air in closed systems by soil. Soil Biol. Biochem. 24: 607-612.
Garcia, 1., Favela, E., Auria, R., Hemandez, S. and Revah, S. Biofiltration of toluene by the fungus
strain Scedosporium apiospermum. In: Proceedings of the 93'd Annual Meeting & Exhibition of the
Air & Waste Manage. Assoc., 2000, Salt Lake City, Utah.
Gibson, N.B., Swannell, R.P.J., Woodfield, M., Van Groenestijn, 1.W., Van Kessel, R.B.M., Wolsink,
J.H. and Hesselink, P.G.M. 1994. Biological treatment of volatile organic carbons from the faou and
drink industry. VOI Berichte. 1104: 261-272.
Goldbeck, J.H., Albaugh, SA and Radmer, R. 1983. Metabolism of biphenyl by Aspergillus

toxixarius: Inductio of hydroxylating activity and aecumulation of water-soluble conjugates. 1.
Bacteriol. 156: 49-57.
Gribbins, M.J. and Loehr, R.C. 1998. Effeet of media nitrogen eoncentration on biofilter performance.
J. Air Waste Manage. Assoc. 48: 216-226.
Hartmans, S., Kaptein, A., Tramper, J. and de Bont, J.A.M. 1992. Characterisation of a
Mycobacterium sp. and a Xanthobacter sp. For the removal of vinyl ehI oride and 1,2-diehloroethane
ftom waste gases. Appl. Microbiol. Bioteehnol. 37: 796--801.
Harremo~s, P. Biofilm kineties. In: Water Pollution Microbiology. R. Mitchell (ed.). J. WiJey and
Sons, New York, 1978, pp. 82-109.
Hirai, M., Ohtake, M. and Shoda, M. 1990. Removal kineties of hydrogen sulfide, methanethiol and
dimethylsulfide by peat biofilters. J. Ferment. Bioeng. 70: 334-339.
Hodge, D.S., Medina, V.F., Islander, R.L. and Devinny, J.S. 1991. Treatment of hydroearbon fuel
vapors in biofilters. Env. Teehnol. 12: 656-662.
Hodge, El.S. and Devinny, J. 1994. Biofilter treatment of ethanol vapors. Environ. Prog. 13: 167-173.
Hodge, D.S. and Devinny, J. 1995. Modeling removal ofair eontaminants by biofiltration. J. Environ.
Eng. 121: 21-44.
Hiittermann, A., Loske, D., Braun-Liillemann, A. and Majeherezyk, A. 1988. Der Einsatz von
Wei[3faulepilzen bei der Sanierung kontaminierte BMen und als Biofilter. Bioengineering. 3: 156-
160.
Hwang, S-J. and Tang, H-M. 1997. Kinetie behavior ofthe toluene biofiltration process. J. Air Waste
Manage. Assoe. 47: 664-673.
Jorio, H., Kiared, K., Brzezinski, R., Leroux, A., VieI, G. And Heitz, M. 1998. Treatment of air
polluted with high eoneentrations of toluene and xylene in a pilot-seale biofilter. J. Chem. Teehnol.
Bioteehnol. 73: 183-196.
Juteau, P., Laroeque, R., Rho, D. And LeDuy, A. 1999. Ana1ysis ofthe relative abundanee of different
types of bacteria capable of toluene degradation in a compost biofilter. Appl. Microbiol. Biotechnol.
52: 863-868.
Kennes, C. and Lema, J.M. 1994. Degradation of major compounds of creosotes (PAH and phenols)
by Phanerochaete chrysosporium. Bioteehnol. Let. 16: 759-764.
Kennes, C., Cox, H.H.J., Doddema, H.J. and Harder, W. 1996. Design and performanee ofbiofilters
for the removal ofalkylbenzene vapors. J. Chem. Technol. Biotechnol. 66: 300-304.
Kennes, C., Veiga, M.C. and Bhatnagar, L. 1998. Methanogenie and perehloroethylene-dechlorinating
activity ofanaerobie granular sludge. Appl. Microbiol. Biotechnol. 50: 484-488
Kennes, C. and Thalasso, F. 1998. Waste gas biotreatment technology. J. Chem. Technol. Bioteehnol.
72: 303-319.
Kiared, K., Bibeau, L., Brzezinski, R., VieI, G. and Heitz, M. 1996. Biological elimination ofVOCs in
biofilters. Environ. Prog. 15: 148-152.
Kim, J.O. 1997. Gaseous TCE and PCE removal by an activated carbon biofilter. Bioprocess Eng. 16:
331-337.
Kim, N.J., Sugano, Y., Hirai, M. and Shoda, M. 2000. Removal eharacteristics ofhigh load ammonia
gas by a biofilter seeded with a marine bacterium, Vibrio alginolyticus. Bioteehnol. Le!. 22: 1295-
1299.
Kleinheinz, G.T., Bagley, S.T., St. John, W.P., Rughani, J.R. and McGinnis, G.D. 1999.
Characterization of alpha-pinene-degrading mieroorganisms and application to a bench-scale
biofiltration system for VOC degradation. Areh. Environ. Contam. Toxieol. 37: 151-157.
Knauf, S. and Zimmer, H. 1994. Biofiltration bei Ablufttemperaturen liber 40 ac. Staub Rein. Luft.
54: 41-44.
Kraislas, S., Tuan Pham, Q., Amal, R., Jiang, J.K. and Heitz, M. 2000. Effect of inlet mass loading,
water and total bacteria count on methanol elimination using upward flow and downward flow
biofilters. J. Chem. Technol. Biotechnol. 75: 299-305.
Lee, S.-K. and Shoda, M. 1989. Biological deodorization using activated carbon fabric as a carrier of
microorganisms. J. Ferment. Bioeng. 68: 437-442.
Lee, B.O., Apel, W.A., Walton, M.R. and Cook, L.L. Treatment ofmethanol contaminated air streams
using biofiltration. In: Proceedings of the 89" h Annual Meeting & Exhibition of the Air & Waste
Manage. Assoc., 1996, Nashville, Tennessee.
Leson, G. and Winer, A.M. 1991. Biofiltration: an innovative air pollution control technology for
VOC emissions. J. Air Waste Manage. Assoc. 41: 1045-1054.
Majcherczyk, A., Braun-Ltillemann, A. and Htitterman, A. Biofiltration of polluted air by a complex

filter based on white-rot fungi growing on lignocellulosic substrates. In: Advances in Biological
Treatment of Lignocellulosic Materials. Coughlan, M.P. and Amaral Coli aco M.T. (eds), Eisevier
Applied Sciences, London, 1990, pp. 323-329.
Mallakin, A. and Ward, O.P. 1996. Oegradation of BTEX compounds in liquid media and in peat
biofilters. J. Indust. Microbiol. 16: 309-318.
Marek, J., Paca, J., Koutsky, B. and Gerrard, A.M. 1999. Oetermination of local elimination capac it ies
and moisture contents in different biofilters treating toluene and xylene. Biodegradation. 10: 397-313.
Martin, A.M. and Manu-Tawiah, W. 1989. Study on the acid hydrolysis of pe a!: composition of the
extracts from Sphagnum peat. J. Chem. Technol. Biotechnol. 45: 171-179.
Martin, F.-J. and Loehr, R.C. 1996. Effect of periods of non-use on biofilter performance. J. Air
Waste Manage. Assoc. 46: 539-546.
Matteau, Y. and Ramsay, B. 1997. Active compost biofiltration of toluene. Biodegradation. 8: 135-
141.
Matteau, Y. and Ramsay, B. 1999. Thermophilic toluene biofiltration. J. Air Waste manage. Assoc.
49: 350-354.
Miller, O.E. and Canter, L.W. 1991. Control of aromatic waste air streams by soil bioreactors.
Environ. Prog. 10: 300--306.
Mohseni and allen. 2000. Biofiltration of mixtures of hydrophilic and hydrophobic volatile organic
compounds. Chem. Engin. Sci. 55:1545-1558.
Monod, J. 1942. Recherches sur la croissance des cultures bacteriennes. Hermann, Paris.
Moo-Young, M. 1985. Comprehensive Biotechnology, voI. 1, Pergamon Press, Oxford, UK, p. 227.
Morales, M., Revah, S. and Auria, R. 1998. Start-up and the effect of gaseous ammonia additions on a
biofilter for the elimination oftoluene vapors. Biotechnol. Bioeng. 60: 483-491.
Morgenroth, E., Schroeder, E.O., Chang, O.P.Y. and Scow, K.W. 1996. Nutrient limitation in a
compost biofilter degrading hexane. J. Air Waste Manag. Asoc. 46: 300--308.
Oh, Y.S., Choi, S.c. and Kim, Y.K. 1998. Oegradation of gaseous BTX by biofiltration with
Phanerocahete chrysosporium. J. Microbiol. 36: 34-38.
96 C. Kennes and MC Veiga
Ottengraf, S.P.P. and Van den Oever, A.H.C. 1983. Kinetics of organic compound removal from
waste gases with a biological filter. Biotechnol. Bioeng. 25:3089-3102.
Ottengraf, S.P.P. Exhaust gas purification. In: Biotechnology: A comprehensive treatise. Rhem, H.J.
and Reed G. (eds.), VCH, Weinheim, Germany, 1986, VoI. 8, pp. 425--452.
Ottengraf, S.P.P. and Diks, R. Review paper: process technology of biotechniques. In: Biotechniques
for Air Pollution Abatement and Odour Control Policics. Dragt A.J. and van Ham 1. (eds.). Elsevier,
Amsterdam, The Netherlands, 1992, pp. 17 -32.
Pedersen, A.R., Moller, S., Molin, S. and Arvin, E. 1997. Activity oftoluene-degrading Pseudomonas
putida in the early growth phase of a biofilm for waste gas treatment. Biotechnol. Bioeng. 54: 131-
141.
Phae, C.-G. and Shoda, M. 1991. A new fungus which degrades hydrogen sulfide, methanethiol,
dimethyl sulfide and dimethyl disulfide. Biotecnol. Let. 13: 375-380.
Pomeroy, R.D. 1957. Deodorizing gas streams by the use of microbiological growths. US Patent
2.793.096.
Quin1an, c., Strevett, K., Ketcham, M and Grego, 1. 1999. VOC elimination in a compost biofilter
using a previously acclimated bacterial inoculum. J. Air Waste Manage. Assoc. 49: 544-553.
Rittman, B.E. and McCarty, P.L. 1978. Variable-order model of bacterial-film kinetics. J. Environ.
Eng.104:889-899.
Rittman, B.E. and McCarty, P.L. 1980. Model for steady-state biofilm kinetics. Biotechnol. Bioeng.
22: 2343-2357.
Rothenbtihler, M., Heitz, M., Beerli, M. and Marcos, B. 1995. Biofiltration of volatile organic
emissions in reference to flexographic printing processes. Water Air Soil. PolI. 83: 37-50.
Ruddy, E.N. and Carroll, L.A. 1993. Select the best VOC control strategy. Chem. Eng. Prog. July: 28·
35.
Sakano, Y. and Kerkhof, L. 1998. Assessment of changes in microbial community structure during
operation of an ammonia biofilter with molecular tools. Appl. Environ. Microbiol. 64: 4877--4882.
Sanchez-Pefia, S.R., Kazenski, S.L., Woertz, 1.L., Kinney, K.A. and Szanislo, P.J. An assessment of
fungal biofiltration of volatile organic compounds. In: Proceedings of the 93'd Annual Meeting &
Exhibition ofthe Air & Waste Manage. Assoc., 2000, Salt Lake City, Utah.
Shareefdeen, Z., Baltzis, B.C., Oh, Y.-S. and Bartha, R. 1993. Biofiltration of methanol vapor.
Shareefdeen, Z. and Baltzis, B.C. 1994. Biofiltration of toluene vapor under steady-state and transient
conditions: theory and experimental results. Chem. Eng. Sci. 49: 4347--4360.
Shinabe, K., Oketani, S., Ochi, T. And Matsumura, M. 1995. Characteristics of hydrogen sulfide
removal by Thiobaci/lus thiooxidans KS I isolated from a carrier-packed biological deodorization
system. 1. Ferment. Bioeng. 80: 592-598.
Smet, E., Chasaya, G., Van Langenhove, H. and Verstraete. W. 1996a. The effect of inoculation and
the type of carrier material used on the biofiltration ofmethyl sulphides. Appl. Microbiol. Biotechnol.
45: 293-298.
Smet, E., Van Langenhove, H. and Verstraete, W. 1996b. Long-term stability of a biofilter treating
dimethyl sulphide. Appl. Microbiol. Biotechnol. 46: 191-196.
Smet, E., Van Langenhove, H. and Verstraete, W. 1997. Isobutaraldehyde as a competitor of the
dimethyl sulfide degrading activity in biofilters. Biodegradation. 8: 53-59.
Song, J.H. and Kinney, K.A. 2000a. Biodegradation capacity and enzymatic activity in a vapor-phase
biofilm reactor. In: 4'h International Symposium on Environmental Biotechnology, Noordwijkerhout,
The Netherlands, ApriIIO-I2, pp. 113-116.
Song, J.H. and Kinney, K.A. 2000b. Effect of vapor-phase bioreactor operation on biomass
accumu1ation, distribution, and activity: linking biofilm properties to bioreactor performance.
Tang, H.-M., Hwang, S.-J. and Hwang, S.-C. 1996. Waste gas treatment in biofilters. J. Air Waste
Manag. Assoc. 46: 349-354.
Tester, C.F., Sikora, L.J., Taylor, J.M. and Parr, J.F. 1979. Decomposition ofsewage sludge compost
in soil: III. Carbon, nitrogen and phosphorus transformations in deifferent sized fractions. J. Environ.
Qual. 16: 25-28.
Togna, A.P. and Singh, M.A. 1994. Biological vapor-phase treatment using biofilter and biotrickling
filter reactors: practical operating regimes. Environ. Prog. 13: 94-97.
Van Langenhove, H., Wuyts, E. And Schamp, N. 1986. E1imination ofhydrogen sulfide from odorous
air by a wood bark biofilter. Water Res. 20: 1471-1476.
van Lith, C., David, S.L. and Marsh, R. 1990. Design criteria for biofilters. TransIChemE. 68: 127-
132.
van Groenestijn, J.W. and Hesselink, P.G.M. 1993. Biotechniques for air pollution control.
Biodegradation. 4: 283-301.
Veiga, M.C., Fraga, M., Amor, L. and Kennes, C. Microbial and kinetic characterization of biofilters
degrading alkylbenzene vapors under acidic conditions. In: 3rd International Symposium
Environmental Biotechnology, Oostende, Belgium, ApriI21-23, 1997, pp. 185-188.
Veiga, M.C., Amor, L., Fraga, M., Văzquez, A. and Kennes, C. 1998. Microbial alkylbenzene
degradation under extremophilic conditions. Meded. Fac. Landbouww. Gen!. 63: 1919-1922.
Veiga, M.C., Fraga, M., Amor, L. and Kennes, C. 1999. Biofilter performance and characterization of
a biocatalyst degrading alkylbenzene gases. Biodegradation. 10: 169-176.
Veiga, M.C., Mondelo, S., Amor, L. and Kennes, C. Long-term performance of a biofilter removing
alkylbenzene vapors. In: 4'h International Symposium Environmental Biotechnology,
Noordwijkerhout, The Netherlands, April 10-12,2000, pp. 456-459.
Veiga, M.C. and Kennes, C. 2001 a. Parameters affecting performance and modeling of biofilters
treating alkylbenzene-polluted air. Appl. Microbiol. Biotechnol. 55: 254-258.
Veiga, M.C. and Kennes, C. 2001b. Biomass characteristics in biofilters and their effect on
performance and model ing. In: Proceedings of the 94'h Annual Meeting & Exhibition of the Air &
Waste Manage. Assoc., 2001, Orlando, Florida.
Von Keitz, Schramm, A., Altendorf, K. and Lipski, A. 1999. Characterization of microbial
communities of biofilters by phospholipid fatty acid analysis and rRNA targeted oligonucleotide
probes. System. Appl. Microbiol. 22: 626--634.
Wagner, M., Amann, R., Lemmer, H. and Schleifer, K.H. 1993. Probing activated sludge with
oligonucleotides specific for proteobacteria: inadequacy of culture-dependent methods for de scrib ing
microbial community structure. Appl. Environ. Microbiol. 59: 1520-1525.
Wagner, M. The Iink between microbial community structure and function. In: 4 th International
Symposium Environmental Biotechnology, Noordwijkerhout, The Netherlands, ApriI10-12, 2000, pp.
92-95.
Wani, A.H., Branion, R.M.R. and Lau, A.K. 1998. Degradation kinetics of biofilter media treating
reduced sulfur odors and VOCs. J. Air Waste Manage. Assoc. 48: 1183-1190.
Wani, A.H., Lau, A.K. and Branion, R.M.R. 1999. Biofiltration control of pulping odors - hydrogen
sulfide: performance, macrokinetics and coexistence effects of organo-sulfur species. J. Chem.
Technol. Biotechnol. 74: 9-16.
Weber, F. and Hartmans, S. 1995. Use ofactivated carbon as a buffer in biofiltration ofwaste gases
with fluctuating concentrations oftoluene. Appl. Microbiol. Biotechnol. 43: 365-369.
Webster, T.S., Devinny, J.S., Torres, E.M. and Basrai, S.S. 1997. Microbial ecosystems in compost
and granular activated carbon biofilters. Biotechnol. Bioeng. 53: 296-303.
Weckhuysen, B., Vriens, L. and Verachtert, H. 1993. The effect of nutrient supplementation on the
biofiltration removal ofbutanal in contaminated air. Appl. Microbiol. Biotechnol. 39: 395-399.
Weckhuysen, B., Vriens, L. and Verachtert, H. 1994. Biotreatment of ammonia- and butanal-
containing waste gases. Appl. Microbiol. Biotechnol. 42: 147-152.
Wijngaard, A. van den, Wind, R.D. and Janssen, D.B. 1993. Kinetics of bacterial growth on
chlorinated aliphatic compounds. Appl. Environ. Microbiol. 59: 2041-2048.
Windsperger, A., Buchner, R. and Stefan, K. 1990. Reinigung losungmittelhaltiger Abluft mit
Biofiltern. Staub Rein. Luft. 50: 465-470.
Windsperger, A. Optimierung von biologischen Abluftreinigungsanlage an praktischen Beispielen. In:

Biotechniques for Air PolIution Abatement and Odour Control Policies. Dragt AJ. and van Ham J.
(eds.). Elsevier, Amsterdam, The Netherlands, 1992, pp. 107-117.
Wu, G., Conti, B., Leroux, A., Brzezinski, R., VieI, G. And Heitz, M. 1999. A high performance
biofilter for VOC emission control. J. Air Waste Manage. Assoc. 49: 185-192.
Yadav, J.S. and Reddy, C.A. 1993. Degradation of benzene, toluene, ethylbenzene and xylenes
(BTEX) by the Iignin-degrading basidiomycete Phanerochaete chrysosporium. Appl. Environ.
Microbiol. 59: 756-762.
Yang, Y. and AlIen, E.R. 1994a. Biofiltration control of hydrogen sulfide. 1. Design and operational
parameters. J. Air Waste Manage. Assoc. 44: 863-868.
Yang, Y. and AlIen, E.R. 1994b. Biofiltration control of hydrogen sulfide. 2. Kinetics, biofilter
performance, and maintenance. J. Air Waste Manage. Assoc. 44: 1315-1321.
Yun, S.-1. and Ohta, Y. 1998. Removal of gaseous n-valeric acid in the air by Rhodococcus sp. B261
immobilized onto ceramic beads. World J. Microbiol. Biotechnol. 14: 343-348.
Zhang, L., Hirai, M. and Shoda, M. 1991. Removal characteristics of dimethyl sulfide, methanethiol
and hydrogen sulfide by Hyphomicrobium sp. Y55 isolated from peat biofilter. J. Ferment. Bioeng.
72: 392-396.
ZilIi, M., Converti, A., Lodi, A., Del Borghi, M. and Ferraiolo, G. 1993. Phenol removal from waste
gases with a biological filter by Pseudomonas putida. Biotechnol. Bioeng. 41: 693--699.
ZilIi, M., Fabiano, S., Ferraiolo, G. and Converti, A. 1996. Macro-kinetic investigation on phenol
uptake from air by biofiltration: influence of superficial gas flow rate and inlet pollutant
concentration. Biotechnol. Bioeng. 49: 391-398.
CHAPTER 4 BIOTRICKLING FILTERS
Huub H.J. COX and Marc A. DESHUSSES
1. Introduction
As mentioned in Chapter 3, biotreatment for air pollution control started in the 1950s
using soi! beds. The technology has since matured and industrial applications of
biofilters are numerous. More recent1y research has focused on a variation of
biofilters called biotricklingjilter, or trickling biojilter. Biotrickling filters work in a
similar manner to biofilters, except that an aqueous phase is trickled over the
packing, and that the packing is usually made of some synthetic or inert material,
such as, among others, plastic rings, open pore foam, lava rock. The trickling
solution contains essential inorganic nutrients such as nitrogen, phosphorous,
potassium, etc. and is usually recycled.
Biotrickling filters are more complex than biofilters but are usually more
effective, especially for the treatment of compounds difficult to degrade or
compounds that generate acidic by-products, such as H2S (Oh and Bartha, 1997;
Mpanias and Baltzis, 1998; Cox and Deshusses, 2000b). Biotrickling filters can also
be built taller than biofilters, which reduces their footprint. It is also easier to control
the conditions because ofthe free liquid phase, so that difficult applications are better
handled by a biotrickling filter than by a biofilter. These include but are not limited
to the treatment of high concentrations of contaminants, treatment of hot gases and/or
of acid producing contaminants, treatment of air streams containing grease or
particles, etc. Biotrickling filters are more recent than biofilters. and have not yet
been fully deployed in industrial applications, but the prospective future applications
are promising. The advantages and disadvantages of biofiltration and biotrickling
filtration are compared to those of conventional air pollution control techniques in
Table 4.1.
2. Biotrickling filtration fundamentals
2.1 BIOTRICKLING FILTRATION PRINCIPLE
The basic mechanisms of biotrickling filtration are shown in Figures 4.1 and 4.2.
Contaminated air is passed co- or counter-current1y through a packed bed of inert
materials on which a pollutant-degrading biofilm has established. At this time, there
is no clear agreement as to whether co-current or counter-current is better.
Mathematical modelling of kinetically limited systems (Diks and Ottengraf, 1991 b,
Zuber, 1995) suggests that co-current is better, since it minimises stripping near the
air exit, but experiments have failed to demonstrate that this was indeed true. For
mass transfer limited systems, the flow direction may not mat1er (Lobo el al., 1999).
Constraints specific to the equipment construction might prevail until further
demonstration that one or the other operating mode is preferred.
99
C. Kennes and M. C. Veiga (eds.), Bioreactorsfor Waste Gas Treatment, 99-131.
100 HHJ Cox and MA. Deshusses
Table 4.1. Comparison of biofiltration, biotrickling filtration and conventional air pollution
control techniques
Control Advantages Disadvantages

Technology
Biotrickling - Simple and low cost technology - More complex to construct and
filtration - Medium capitallow operating costs operate than biofilters
- Effective removal - Clogging by growing biomass iftoo
- Treats effectively acid produc ing much nutrient is added and high
contaminants concentrations of VOCs are treated
- Low pressure drop
Biofiltration - Simple and low cost technology - Relatively large footprint
- Low operating and capital costs requirement
- Effective removal of low - Medium replacement every 2-5 years
concentrations - Less suitable for high concentrations
- Low pressure drop than biotrickling filters
- No further waste streams produced - Moisture and pH sometimes difficult
to control
- - - - - ._---- - Particulate matter m<lY_c:logJh(:.jJ~. __
Wet - Medium capital costs - Very high operating costs
scrubbing - Can operate with particulate in gas - Reduced performance by scale
stream deposit
- Relatively smail footprint - Need for complex chemical feed
- Ability to handle variable loads systems
- Weil proven technology - Does not remove most VOCs
- Requires toxic and dangerous
--=:---:--- - - - - _ . _ _.. _ - chemicals
Carbon - Short retention time/smail unit - High to extremely operating costs
adsorption - Consistent, reliable operat ion - Carbon life reduced by moist gas
- Moderate capital costs - Creates secondary waste streams
(spent carbon)
- Medium pressure drop ______._
Incineration - Effective removal of compounds - High operating and capital costs
irrespective of nature and - High flow/ low concentrations not
concentration cost- effective
- Suitable for very high loads - UsuaIly requires additional fuel
Performance is uniform and reliable - Creates a sccondary waste stream
- Smail footprint (NO x)
- Scrutinised by the public
Contrary to biofilters, biotrickling filters do not require the air to be pre-

humidified or pre-conditioncd prior to treatment. In most cases the air will rapidly
reach water saturation. If the air contains small amounts of dust or particulate matter,
possibly grease, they will be leached via the medium purge. The medium feed
consists of mineral nutrients, and can vary from a full basal salt medium with trace
elements and pH buffer, as is usual in laboratory studies and in selected small field
applications, to a very crude mix of fertilisers containing the necessary nitrogen,
phosphorous and potassium. In waste water treatment plants, industrial water (e.g.,
chlorinated secondary effluent), containing small amounts of dissolved organic
materials and some residual nitrogen, has been used successfully to feed H2S
degrading biotrickling filters (Morton and Caballero, 1996). It should be noted
though that H2 S degrading biotrickling filters host autotrophic organisms and have
usually a low nutrient requirement.
Biotrickling Fi/ters 101
Blowcr ,...--_ _ _ _--,
Illlct air
(contamillatcd)
Figure 4.1. Schematic ofbiotrickling filter set up (co-current and forced draft shown).
Biofilm ConlamÎnalrd air
il
Pollutanl End Product
H~drocarbon CO, - H,O
H,S --. '0, '
Reduced ""Ifur 0;- + CO.
NO." I-CO.)
Pollul anl (. ,1 •
Figure 4.2. Basic principles of biotrickling filtration.
The feed rate and nutrient concentration is based on the expected nutrient
requirements of the process culture. Based on typical growth media recipes,
biotrick!ing filters often receive less nutrients than stoichiometrically required. This
is because the nutrients are internally recycled by cell death, endogenous biomass
digestion, and predators such as protozoa and rotifers. Nutrient !imitation can also be
a deliberate choice of the operator, in an attempt to limit bacterial growth and slow
down biomass accumulation, as discussed in Section 5.5. Interestingly, the nutrient
deficiency can be as high as a factor 20 of the theoretical amount required without
severe consequences as far as pol\utant removal rates are concerned. The
determination ofthe feed rate and feed composition should also consider keeping the
concentration of any salt or metabolite below any possible inhibitory concentration,
although as wil\ be discussed in Section 5.5, some investigators have on purpose
al\owed the salt concentrat ion to rise, again, in an attempt to !imit biomass growth.
Biotrickling filters differ from biofilters in that there is a free liquid phase. In
most cases the trickling solution is recyc1ed, although in a few application with very
low trickling, a one pass of the liquid has been used. Overall, a wide variety of
conditions have been experienced for trickling rates usually ranging from 0.05 to 20
m/h. The trickling rate influences the wetting of the packing and therefore the mass
transfer ofthe pollutants to the biofilm. The trickling rate also influences the pressure
drop over the packed bed. The fact that biotrickling filters work with only partial
wetting of the biofilm, suggests that pollutants can transfer from the gas to the
biofilm directly, or they can transfer from the gas to the trickling liquid, and then to
the biofilm. While the latter path is generally assumed, mechanisms and rates for the
direct transfer from the gas to the biofilm has been vastly ignored (Cox and
Deshusses, 1998).
As far a pollutant degradation is concemed, most ofthe degradation occurs in
the biofilm, although in selected cases, significant activity can occur in the recyc1e
liquid, as reported by Cox el al. (2000). Examination of Figure 4.2 reveals that
pollutant elimination is the result of a complex combination of mass transfer
(diffusion, convection) and biological processes, such as growth, death and lysis,
predation. These are complicated by the fact that the biofilm houses multiple species
in a complex architecture. All ofthese are affected by the operating conditions ofthe
system and also affect the pollutant removal and the general dynamic of the
biotrickling filter. Under these circumstances a complete description of ali the
phenomena involved in a biotrickling filter is not yet available. While an improved
fundamental understanding would be desirable for re actor design and operation, the
lack of fundamental understanding does not prevent application ofthe technology.
2.2 DEFINITIONS, PERFORMANCE REPORTING
As in the case of conventional biofilters, operation and performance of biotrickling

filters for air pollution control is generally reported in terms of removal efficiency, or
pollutant elimination capacity as a function of the pollutant loading, or the gas empty
bed retention time (EBRT). These terms extensively defined in Chapter 3 are briefly
recalled below (Equations 4.1-4.4):
.
Remova l EffilClency = RE = C m - C OU!
100 (%) (4.1)
C in
Pollutant Elimination Capacity = EC = (4.2)
V (4.3)
Empty Bed Retention Time = EBRT = Q (s or min)
Pollutant Loading = L = ~ Q (g / m 3 . h) (4.4)

where Cn and Cout are the inlet and outlet pollutant concentrations (usually in g/m3 or
in dilution to threshold for odour removal), respectively, V is the volume of the
packed bed (m3) and Q is the air flow rate (m 3/h). Pollutant concentrations are
usually reported as mass per volume; conversion of volumetric to mass
concentrations (C) is done using the ideal gas law which reduces to Equations 1.1 or
1.2 recalled below for room temperature conditions (see Chapter 1):
/ 3 _ Molecular Weight of pollutant (g / moI) C (ppmv) (4.5)

C (g m) - 24776
It should be stressed that the elimination capacity and the loading are
calculated using the volume of the packed bed and not to the total volume of the
reactor. Depending on the reactor design, the volume of the packed bed volume will
be about 40-90% of the total reactor volume. Aiso the EBRT is calculated on the
basis of the total volume of packed bed (Equation 4.3) and not the void space in the
packing. The actual gas residence time wi\l be lower depending on the porosity of the
packing, the dynamic liquid hold-up and the amount of biomass attached to the
packing. The porosity of packing ranges from about 50% (lava rock) to 95% (ali
random or structured packings), the liquid hold-up is usually less than 5% ofthe bed
volume, and biomass may occupy 5% to 30% of the bed volume. Hence, the actual
gas residence can be less than halfthe EBRT.
A typical elimination capacity vs. pollutant loading curve is shown in Figure
4.3. Note that it is similar to curves obtained for conventional biofilters. It is usual to
report the performance as a function of the load, i.e., inlet concentration x air flow.
rather than the concentration. This enables comparison of systems of different sizes
operated under different conditions. One underlying assumption is that the
performance depends only on the pollutant load, hence, that low concentrations-high
flowrates conditions lead to similar elimination capacities than high concentrations-
low flowrates. This assumption is generally valid because the pollutant
concentrations commonly encountered in biotrickling filters are often high enough
for the micro-kinetic to be of zero order. As schematically shown in Figure 4.3, this
is no longer true at very low pollutant concentrations (typically below 0.05-0.1
g/m 3), in particular for pollutants with high Henry's law coefficients, because first
order kinetics will prevail in the biofilm resulting in a reduction of the maximum
elimination capacity.
190% removalline
Pollutanlload (g m-1 bol)
Figure 4.3. Schematic of a typical elimination capacity VS. (oad curve for a biotrickling
fiIter.
Examination of Figure 4.3 reveals that there are essentially three operating
regimes, as for the case of conventional biofilters:
1. Low loading, also called first order regime. The elimination capacity and the
loading are identical and the pollutant is completely removed. The biotrickling
filter is operated well be10w its maximum e1imination capacity. The performance
increases proportionally with the loading.
2. Intermediate range. Breakthrough of the pollutant occurs. With higher inlet
concentrations or higher air flow rates, the e1imination capacity increases, but to
a lesser extent than the loading.
3. High loading, also called zero order regime. The biotrickling filter is operated at
its maximum elimination capacity. Increases in loading do not result in further
increases in elimination capacity, the removal efficiency decreases. As indicated
on the figure, if the concentration is very low, the maximum elimination capacity
is reduced accordingly.
For the evaluation ofbiotrickling filter performance, one should consider both
the maximum elimination capacity and the removal efficiency. For practical reasons
academic research is mainly concemed with the maximum elimination capacity OI
with high performance, which occur at relatively high pollutant concentration and
often less than -90% removal efficiency. On the other hand, reactor design for
industrial application often needs to meet a certain discharge requirement, or achieve
a high removal percentage. Thus there might be some chaUenges in extrapolating
research data for reactor design.
2.3 BIOTRICKLING FILTER CONSTRUCTION
To be cost effective, biotrickling filters need to be both inexpensive and reliable.

This is achieved by careful material selection, reactor design, and reactor operation
and monitoring. Basic construction considerations are discussed in this section.
The biotrickling filters are always in-vesse1 types of reactors, although
conceptually, one could consider open bed biotrickling filters, in a similar manner to
open bed biofilters (see Chapter 3). The vessel needs to provide the structural
strength to support the bed. It can be constructed out of fibreglass OI fibreglass
reinforced plastic, plastic polymers (PE, PP, less frequently PVC), metal, or concrete.
Fibreglass or fibreglass reinfOIced plastic are commonly used in air scrubbers and
provide a good structural strength and excellent resistance to various chemicals.
Reactors can be build relatively taU (l0-15 m). Fibreglass or fibreglass reinforced
plastic reactors are relatively heavy and can be expensive. Polyethylene (PE) or
polypropylene (PP) vessels offer the convenience of lightweight materials, low costs
and good chemical resistance. The tensile strength of PE or PP is somewhat limited,
so that special considerations are required for tall reactors. Biotrickling filters can
also be constructed out of carbon stee1, stainless steel or other metals. Usually this is
a costly option. Special care should be paid to cOIrosion resistance, in which case an
epoxy coating, or a similar corrosion preventing coating, is warranted. Biotrickling
filters can possibly be made out of stackable containers to save space, as are some
biofilters commerciaUy available (Sabo el al., 1995). Finally, biotrickling filters
could be constructed in concrete (custom or precast cinder blocks), although to our
knowledge this has not yet been attempted. Here again, special consideration for
corrosion resistance is warranted, especially for the treatment of hydrogen sulphide

or if low pH is expected. In all cases, in a similar manner to biofilters, thermal
insulation may be required around the vesse1 for reactors installed in cold climate.
In most cases the bottom of the reactor is used as a reservoir for the trickling
liquid, which eliminates the need for a separate tank. The bottom of the reactor is
also the preferred location for any liquid heating, which is be best done using a
submerged electric heater. In line heat exchangers are not recommended. In addition
to being more costly, heat exchangers would be subject to rapid fouling by biomass,
which lowers heat transfer and increases liquid pressure drop. The biotrickling filter
vesse1 includes various duct work and liquid piping. For air ducting, fibreglass, PE,
galvanised or stainless stee1 is common. In most cases inexpensive PVC piping is
adequate for the liquid. The concentrations ofpollutants are generally too low to alter
the PVC. However, special care should be placed on the liquid distribution system on
top of the bed, as improper distribution will affect the reactor's performance. A
convenient way to distribute the liquid onto the packing is via a series of spray
nozzles placed so that a homogeneous distribution is achieved. Non-clogging nozzles
are recommended, and monthly inspection of the nozzles should be performed. The
recycle flow is ensured by a pump, usually a high efficiency centrifugal pump. In
most cases the pump will be relative1y small « 3 HP). If trickling is intermittent a
timer is installed, or the pump is remotely operated by the central process control
software. In key applications two pumps should be installed, one as a backup in case
the other fails. Having a re1iable liquid distribution and pumping is key, since the
biotrickling filter will rapidly fail in absence of liquid trickling.
As in the case of biofilters, careful selection of the blower is a key to cost
effective operation. In large applications a variable speed drive is recommended.
While these are much more expensive than single speed drive, they allow for
adjusting the flow and the head to compensate for the building up of biomass and the
slow increase of pressure drop over time. This saves on energy costs otherwise
wasted in dampers or other flow regulating devices. The blower can be installed
upstream (forced draft) or downstream (induced draft) of the biotrickling filteI. In
any cases, sufficient space should be allocated above and be10w the bed, for the air
distribution and the air exit, and for the liquid distribution. This is particularly true
for large reactors, where poor air distribution can cause significant short-circuiting
and decrease pollutant removal (Chitwood et al., 2000). Depending on the reactor
configuration (co- or counter-current) and on the air velocity, installation of a
demister will be required to prevent carry over of water droplets. It can consist in a
simple water knock out drum or a section of bare packing, or a specially designed
non-clogging demisteI.
The biotrickling filter will usually include a tank for the nutrient solution,
which is best stored as a concentrated solution to avoid bacterial growth in the stock
solution, and a tank for caustic if the reactor includes a pH control. These are usually
made of PE or PP and their volume should be such as to ensure at least a two week
supply. Nutrients or caustic are fed manually or by pump based on a timer, or
pumped as triggered by a process controIler.
The degree of sophistication of the controls will usually determine the nature
ofthe ancillary equipment as well as reactor operation and maintenance. A minimum
of control is desired to ensure safe operation. This inc\udes installation of a low level
switch at the bottom ofthe reactor to prevent the recycle pump to run dry. If deemed
necessary, the biotrickling filter may be equipped with sophisticated sensors and
control routines using logic statements to detect abnormal operation, and propose or
implement remedial action via feedback controllers (Webster et al., 1999). For
instance, the medium feed and purge might be controlled using the input of
conductivity probe (as an indicator of the overall salt concentration) installed in the
recycle liquid, and the input of a level switch. However, it should be stressed
sophisticated controls are not always necessary. The low tech alternative for the
above will consider an operator removing a given volume of the recycle liquid on a
given schedule (daily or weekly) and replace it with water and a mineral nutrient
mix, often fertiliser grade. The choice of sophisticated controlor mostly manual
operation depends on the application, the size of the reactor, the regulatory agency
and the available budget.
Monitoring the re actor performance is often defined in the permitting of the
biotrickling filter. It can be on-line integrated with the process control, or off-line by
grab samples. Monitoring is an important part of operation and maintenance. Proper
monitoring helps in troubleshooting and in optimising the operation of the reactor.
Most important is the monitoring of pollutant removal. For VOCs this is best done
using a flame ionisation detector, while for H2 S, and ammonia, special meters are
available. Monitoring tubes can also be conveniently utilised. For odour treatment,
odour panels are still the best measurement although they are expensive. The
measure of the odour removal efficiency is usually based on the odour dilution-to-
threshold, although the character of the odour should also be considered. Inlet and
outlet should always be measured. The frequency of the measurement depends on the
regulatory requirement, the equipment used and the operating conditions of the
biotrickling filter (steady conditions or rapid transients).
Other parameters to consider in a monitoring plan include airflow rate,
reactor temperature, pressure drop, biomass content, and parameters pertaining to the
recycle liquid, such as flow rate, pH, ionic strength, biological oxygen demand, and
possibly dissolved oxygen in the case of the treatment of high pollutant
concentrations. Structural inspection and proper functioning of the biotrickling filter
components (e.g., spray nozzles, feed lines, etc.) are also important. The challenge of
defining a good monitoring plan will be to schedule the necessary sampling that
allows for a good control of the biotrickling filter and an early detection of possible
problems, while minimising maintenance and analytical costs.
3. Biotrickling fiUer performance
3.1 EXAMPLES OF POLLUTANTS TREATED IN BIOTRICKLING FILTERS
Over the past decade, many pollutants have been tested for treatment in biotrickling
filters, mostly in the laboratory. Effective removal has been reported for aliphatic and
aromatic hydrocarbons, oxygenated hydrocarbons, chlorinated hydrocarbons and
inorganic compounds. Typical examples are presented in Table 4.2. Most pollutants
contain carbon and they serve as a carbon and energy source for aerobic,
heterotrophic microorganisms. Other biological processes observed Of applied in
biotrickling filters include:
• Oxidation of inorganics as a source of energy. An example is hydrogen sulphide,

which is oxidised to sulphuric acid by Thiobacillus species (Section 3.3). These
species are in general autotrophic, using carbon dioxide from air as the carbon
source for growth. Another example is the oxidation of ammonia to nitrate.
Effective removal of ammonia in a nitrifying biotrickling filter has been
demonstrated by van Groenestijn and Lake (1999). In a second denitrification
reactor, the produced nitrate was reduced to nitrogen gas as the final product of
the combined system.
• Co-metabolism. An example is the removal of trichloroethylene in biotrickling
filters (Sun et al., 1998). Cometabolism requires the presence of an inducing
growth substrate, which may be already present in the waste gas or has to be
added artificially.
• Anaerobic removal of compounds that are persistent under aerobic conditions
(Section 4.2).
Biotrickling filtration, like any other biological technique, is successful only

if the pollutant is biodegradable and if it is available in concentrations sufficiently
high to support a thriving process culture. Quantitative criteria for pollutant
biodegradability have not been established, hence, the pollutants in Table 4.2 are
classified in three groups of increasing biodegradability. The pollutant availability to
the process culture is related to the pollutant mass transfer into and in the biofilm.
Table 4.2 lists the dimensionless Henry coefficient as an approximation of the
pollutant availability. The following general trends can be discemed:
• The start-up or lag phase increases with decreasing biodegradability. Whereas

biotrickling filters treating easily biodegradable compounds may operate at their
maximum elimination capacity within a few days, effective removal of
compounds such as MTBE may only be observed after several months of
operation. As discussed in Section 4.4, the start-up phase may be shortened by
inoculation of the biotrickling filter with microorganisms with a high specific
activity towards degradation of the pollutant of interest.
• If the gas/liquid partition is favourable (low H), the elimination capacity
increases with higher biodegradability of the pollutant. This is illustrated in
Figure 4.4. For biotrickling filters that are kinetically limited. maximum
elimination capacities of 200-300 g/m 3 .h can be obtained for easily
biodegradable pollutants. Performance at higher pollutant loads is more likely to
be limited by factors other than the biodegradability, for instance by oxygen
limitation (Section 4.2) or by rapid accumulation ofbiomass (Section 5).
• Irrespective of the biodegradability, removal of hydrophobic compounds such as
alkanes may be limited by the mass transfer rate. As a rule of thumb, mass
transfer limitation can be expected to play an increasingly important role for
compounds with a Henry coefficient gre ater than 0.5-1.
3.2 TOLUENE AS MODEL POLLUTANT: COMPARISON OF BIOTRICKLING

FILTER PERFORMANCES
Toluene can be considered as moderately water-soluble and easily biodegradable

with a maximum elimination capacity in biotrickling filters of about 80 g/m 3 .h (Table
4.2). Over the past ten years, many researchers have used toluene as their model
pollutant. This enables a direct comparison of how reactor design and operation
affect the overall performance. It is worthwhile to mention that biotrickling filter
108 HHJ Cox and M.A. Deshusses
performance is very reproducible provided that the reactors are operated under
identical conditions. In a comparative study with twenty identical, toluene-degrading
biotrickling filters operated in parallel, the average maximum elimination capacity
was 79 g/m 3 .h with a standard deviation of 5.7 g/m 3 .h (Cox and Deshusses, 2000a).
~
u 200
ro
o..
ro
u
C 150
.Q ~
....,
ro . ~
C(Y)
E __
.- E 100
=<1l~
OJ
E 50
:::J
E
'xro
O
~
+ ++ +++
Biodegradability of pollutant
Figure 4.4. Maximum elimination capacities averaged for pollutants grouped by increasing
biodegradability.
Table 4.3 summarises the design, operation and performance of toluenc-

degrading biotrickling filters in five independent studies conducted by different
laboratories. The maximum elimination capacity varied from 18 to 120 g/m 3 .h,
which greatly exceeds the variability observed for the elimination capacity of
identical biotrickling filters as discussed above. Not unexpectedly, parameters related
to the mass transfer of the pollutant (e.g., specific surface area of packing, superficial
liquid velocity), the microbiology (e.g .. type of inoculum. supply of nutrients). the
design (e.g., reactor dimension and type of packing) and the operation of biotrickling
filters (e.g., gas residence time, inlet concentration, co-current versus counter-current
operation) affected the performance. Currently. research is attending to identify and
quantify parameters crucial for the optimum biotrickling filter performance. Some is
discussed in next sections.
Owing to the complexity of the process, the design engineer has at present
only a few objective criteria for developing full-scale biotrickling filters that go
beyond the thumb-of-rule principle. Despite efforts to model biotrickling filter
performance (e.g., Diks and Ottengraf, 1991a; Pcdersen and Arvin, 1995; Mpanias
and Baltzis, 1998; Okkerse el al., 1999; Lobo el al., 1999), these have not yet proven
to be very helpful for design. Therefore, it is stiH common practice to perform pilot
plant studies before designing and constructing full-scale reactors. This pilot tests are
important because conditions at field locations are usually not constant, but subject to
changes of the temperature, waste gas composition and/or concentration of
pollutants. This may have a marked effect on the performance of biotrickling filters.
which is often not considered in the constant environment of the laboratory (Webster
el al., 1999).
Table 4.2. Examples of pollutant removal in biotrickling filters.
Pollutant Biodegradability' Henry coefficient' Lag phase EC (glml.h) References

Aliphalie hydrocarbons
Hexane ++ 74 50 days 7.5 Plaggemeier et al., 1997
Heptane ++ 83.2 NA 24 Schindler and Friedl, 1995
Aromatie hydrocarbons
Styrene ++ 0.11 15 days 32 Pol el al, 1998
Styrene' ++ 0.11 NA 35 Webster et al, 1999
Toluene +++ 0.28 10 days 80 Cox and Deshusses, 1999'
Oxygenated hydrocarbons
Propionaldehyde +++ 0.0024 NA 300 Kirchner et al., 1992
Acetone +++ ~
0.0016 NA 500 Kirchner et al., 1992
Methyl ethylketone +++ 0.0024 NA 40 Chou and Huang, 1997
o'
'::;'
Methanol +++ 0.00019 5 days 100 Allen el al., 2000 ;::;.
n-butanol +++ 0.00035 NA 100 Heinze and Friedrich, 1997 i:t
S·
Diethyl ether + 0.028 2 weeks 60 Zhu el al., 1996 (Jq
Methyl-tert-butyl ether + 0.023 7 months 45 Fortin and Deshusses, 1999 ~
::;:-
Chlorinated hydrocarbons ;:;'"
Dichloromethane +++ 0.093 1 week 200 Hartmans and Tramper, 1991
+++ 0.093 1 week 157 Diks and Ottengraf, 1991 b
Chlorobenzenes ++ 0.18 NA 300 Ob and Bartha, 1994
++ 0.18 NA 60 Mpanias and Baltzis, 1998
Nitrogen and sulphur compounds

Nitrobenzene + 0.00098 4 weeks 50 Oh and Bartha, 1997
Carbon disulphide" ++ 0.39 NA 220 Hugler el al., 1999
Nitric oxide + NA 6 weeks 25 Chon and Lin, 2000
Hydrogen sulphide' +++ 0.94 2 weeks 100 Kraakman et al., 1998
" Pilot-plant study .....
b From poorly (+) to highly biodegradable (+++) (Devinny el al., 1999). a\Q
, Dimensionless Henry coefficient. Major sources: Nirmalakhandan and Speece, 1988, and Lide, 1997.
......
......
a
Table 4.3. Selected laboratory studies on toluene removal in biotrickling filters.
Parameter Pedersen and Arvin, Cox and Deshusses, Smith et al., Schiinduve et al., Weber and Hartmans,
1995 1999a 1996 1996 1996
Reactor design
Reactor, lD x H (cm) 9x70 15xl30 15x114 15x25 30xlO0
Packing Steel Pall rings PP Pall rings Celite pellets PP Pali rings Pall rings ;::r::
Specific surface area (m 2/m 3 ) 317 220 1190 NA 110 ;::r::
Inoculation Adapted consortium Pseudomonas Sample of reactor Adapted consortium Bacterial Fungal '-..
corrugata consortium consortium
Q
Reactor operation ~
"
;:s
Gas flow Upflow Downflow Downflow Upflow Upflow \:)..
Empty bed gas residence time (5) 32-160 56 60-120 17-132 36
~
Toluene inlet concentration (glm 3 ) 0.19-1.61 0.5-3.5 0.93 0.2-3.8 0.7 ~
Superficial liquid velocity (m/h) 3.3-4.7 7.9 0.05 22.6 1.6
Continuous Semi-continuous' Batch Batch
O
ti)
Nutrient feed Batch
Nutrient composition Semi-full medium Full medium Full medium Full medium Full medium Complete, limiting ;::-
'"
;:::
N-NH/ N-N0 3' concentrations
ti)
'"'"
Liquid residence time (day) NA 0.5 NA NA
'"
Reactor performance
Load (g/m 3 h) 6-150 35-225 60 -50-1300 70
Elimination capacity (g/m 3 .hl 35 71-83 60 120 33 18 34
'No liquid recycling, but once-through pass of fresh medium.
Biotrickling filters 111
3.3 PILOT-PLANT STUDIES AT US WASTE WATER TREATMENT PLANTS
Although biotrickling filter research in the US has trailed Europe over the past two
decades, recently rapid progress has been made especially in the area of odour
control at waste water treatment facilities. Waste gases at these facilities contain
hydrogen sulphide (H2S) as the principal odour causing agent in concentrations up to
100 ppmv as well as lower concentrations (0-100 ppbv) of various VOCs and
chlorinated hydrocarbons. Chemical scrubbers are currently employed to reduce the
odour problem. Although they are effective in removing H2S, the high consumption
of chemicals (caustic soda, hypochlorite) and their ineffectiveness in controlling
VOCs are drawbacks of increasing concern. Demonstration projects, many of them
located in Southern California and conducted by universities, waste water treatment
facilities and industry, focus on using biofilters and biotrickling filters as alternatives
for chemical scrubbers.
In 1993, a biotrickling filter study was done at the Los Angeles County
Sanitation Districts (Morton and Caballero, 1996). Greater than 98% and 99%
removal of respectively H2S and odour was demonstrated in a pilot unit of 0.3 m3
with lava rock as the packing material. The gas residence time was varied between
12 and 30 seconds, which is considerably more than the typical detention time in
chemical scrubbers (2-5 s). A constant pH of 2.0-3.0 was maintained by
continuously supplying secondary effluent water and purging the produced sulphuric
acid. The secondary effluent water also proved to be good source of nutrients for
growth and activity of H2S oxidising bacteria. Because of concerns of clogging of the
lava rock, a second experiment was set up using a porous plastic p.acking. The H2S
removal efficiency was lower, presumably because of mass transfer limitation due to
a lower specific surface area of the packing. The study of Morton and Caballero
(1996) illustrates the importance of packing material selection.
Co-treatment ofH2S and VOCs has been studied in a number ofprojects. For
optimisation of both removals, there may be a conflict of pH optima. H2S is in
general oxidised by Thiobacillus species with an optimum pH of about 2. VOCs are
degraded by heterotrophs that generally prefer a neutral pH, although VOC removal
in low pH, H2S oxidizing biofilters and biotrickling filters has been observed (Torres
and Basrai, 1998; Chitwood el al., 1999; Cox and Deshusses, 2000b). Chitwood el
al., 1999 investigated a two-stage process at the Ojai Valley Sanitary District. This
process consisted of an acidic reactor with lava rocks for H2S removal, followed by a
pH-neutral biofilter for VOC removal. The acidic reactor had a very low intermittent
trickling rate. Overall performance and removal efficiency of the two stage process
seemed slightly better than combined removal in a single stage biofilter (Chitwood el
al., 1999). Other studies have focused on biotrickling filters operated at a neutral pH
by automated caustic soda addition, for simultaneous removal of H2S and VOCs. A
pilot-scale biotrickling filter at the County Sanitation Districs of Orange County
removed greater than 87% of H2S but removal of VOCs was disappointingly low at
11 % (Torres and Basrai, 1998). Low VOC loadings and frequent systems upsets
were the presumed causes for poor VOC removal. Nevertheless, this study shows the
potential of pH-neutral biotrickling filters for the simultaneous removal of H2S and
VOCs.
4. Factors influencing biotrickling fllter performance
4.1 TEMPERATURE
Most biotrickling filters in industry are subject to changing temperatures throughout

the day as well as seasonal variations. which both are not experienced in the
1aboratory. Remarkably few studies have however investigated the influence of the
temperature on the performance of biotrickling filters. In general, biotrickling filters
are operated at temperatures between 10 and 40 DC, which typically is the
temperature range for growth of mesophilic microorganisms.
As discussed earlier, biotrickling filter performance is in general limited by
the biological reaction rate or by the mass transfer rate. The temperature has a
different effect on either limitation, consequently temperature effects on the
performance vary depending on the particular application. For mass transfer limited
reactors the effect of the temperature should be discussed with respect to the
parameters that influence mass transfer. The driving force for mass transfer depends
on the Henry coefficient, which will decrease with increasing temperature. On the
other hand, the diffusion coefficient wilI increase, facilitating mass transfer inside the
biofilm. These two effects may counterbalance each other. Thus Diks and Ottengraf
(1991 b) observed no effect when the temperature was increased from 20 to 30 °e in a
biotrickling filter degrading dichloromethane. With the same polIutant Hartmans and
Tramper (1991) also observed no changes between 18-30 DC when their reactor was
operated at a high gas flow rate and low pollutant concentrations (mass transfer
limitation). Operation of the same reactor at a low gas flow rate and high pollutant
concentration (biological reaction limitation), however, showed increasing
performance at higher temperature. This was correlated to increasing activity of the
microorganisms (Hartmans and Tramper, 1991). The influence ofthe temperature on
microbial activity has been well studied and, in the sub-optimum range (~1 0-
35 0c), can in general be described by the Arrhenius equation. Obviously,
temperatures beyond the optimum of the process culture will cause a decline
of the performance and even cause cell death if the maximum temperature is
exceeded.
Many waste gases in industry have temperatures beyond the mesophilic
range, but treatment of those gases has received only little attention (Van Lith et al.,
1997). Cooling of hot waste gases would be a solution, but is expensive especially
when the gas is saturated with water. The use of thermophilic microorganisms
adapted to high temperatures is an interesting area, as it alIows for treatment without
prior cool ing. As an example, effective removal of a-pinene and methanol at
biotrickling filter temperatures of respectively up to 60 and 70 °e has been
demonstrated (Allen et al., 2000). In our own laboratory, we observed that overall
performance of a thermophilic biotrickling filter treating ethanol vapours was
comparable to that of a mesophilic control (Cox and Deshusses, 2000c). Operation at
53 °e resulted in the selective enrichment of a thermophilic population. The ethanol
removal rate was not affected by the temperature in the range of 50 to 62 DC.
Thermophilic treatment is expected to be a promising, new area of application of
biotrickling filters.
4.20XYGEN
The oxygen concentration in most waste gases is in general several orders of a

magnitude higher than the pollutant concentration. However, because of the low
oxygen water solubility, biotrickling filter performance may be limited by mass
transfer of oxygen into the biofilm and/or by diffusion in the biofilm. Kirchner et al.
(1992, 1996) observed significant increases of the elimination capacity when the
oxygen partial pressure in the gas was increased, thus demonstrating that at normal
oxygen content, VOC removal was limited by oxygen availability. Oxygen limitation
occurs when the penetration depth of oxygen is less than that of the pollutant,
causing anaerobic layers in deeper parts of the biofilm close to the substratum
(Figure 4.5). This has experimentally been confirmed using an oxygen
microelectrode technique, showing depletion of oxygen at a depth of about 400 Ilm
into the biofilm of a biotrickling filter treating diethyl ether (Alonso et al., 1998a).
- Gas -_-14--- Biofilm -------...t-- Packing ....
Oxygen ---------- A) Low pollutant

concentration: there is an
"
\~, excess oxygen
Pollutant ~""'~~~- ... _---------------------------
- Gas --~-- Biofilm -------...t-- Packing ....

Po\lutant
Oxygen ----------
concentration: oxygen
limitation occurs
Figure 4.5. Schematic representation of oxygen and pollutant concentration profiles in the
biofilm.
Occurrence of oxygen limitation depends on the utilisation rates of oxygen

and the pollutant on one hand, and the respective diffusion rates in the biofilm on the
other hand. This can, in a first approximation, be calculated using Equation 4.6 (Diks
and Ottengraf, 199Ia):
(4.6)
where D02 and Dpoll are the diffusion coefficients of oxygen and of the pollutant in
the biofilm, C02 ,i and Cpoll,i are the concentrations of respectively oxygen and the
pollutant at the biofilm interface, while the ratio Vpo ll/V02 (mg pollutant/mg oxygen)
reflects the reaction stoichiometry. Oxygen limitation will occur when A < 1,
whereas A > 1 corresponds to the situation of pollutant mass transfer limitation.

Equation 4.6 predicts that oxygen limitation occurs when the pollutant at the
gas/biofilm interface exceeds a certain concentration, as illustrated in Figure 4.5.
Indced, pollutant removal in biotrickling filters becomes more sensitive to oxygen-
related parameters at high inlet concentrations of the pollutant (Mirpuri el al.,
1997a). The ratio of pollutant degradation to oxygen consumption (VpolI/V02) simply
follows from reaction stoichiometrics. However, a factor often neglected is the
oxygen consumption by the secondary population in the biofilm. It has been
estimated these secondary processes may significantly contribute to the overall
oxygen consumption rate in biofilms (Villaverde el al., 1997), but it is not known to
what extend this may limit the removal rate ofthe primary pollutant.
A final note on the ratio of the interfacial concentrations C02 ,/C polI ,i' In the
past, it has been convenient to as sume that a) no resistance occurred at the gas-liquid
or gas-biofilm interface, and b) that the partition coefficient of the pollutant or
oxygen was this of the partition with water. While the former assumption can easily
be corrected using the two film theory and calculation on individual mass transfer
coefficients, the latter assumption may be more questionable. Recent work by
Davison el al. (2000) suggests that the partition coefficient of hydrophobic
compounds like propane, but possibly oxygen, to biofilm differ considerably from
the partition to water. While the reasons for the deviation are sti Il unclear, this may
explain why some hydrophobic compounds are better degraded than predicted and/or
why oxygen limitation may not be as frequent as one would predict from Equation
4.6.
Anaerobic zones in the biofilm are unwanted for aerobic pollutant
degradation, however, they may facilitate the biodegradation of compounds that
require anaerobic conditions for biodegradation. For instance, the removal of PCE
from aerobic waste gases in biofilters has been reported (Devinny el al., 1995). To
date, it is unclear how to best exploit anaerobic processes in mostly aerobic
biotrickling filters.
4.3 PACKING MATERIAL
Many different sorts of packing materials have been tested in biotrickling filters.
Important requirements to the packing include a large specific area, high porosity,
high chemi cal stability and structural strength, low weight, suitable surface for
bacterial attachment and growth, and low cost.
A commonly used packing material is lava rock (e.g., Morton and Caballero,
1996; Fortin and Deshusses, 1999). This material has the advantage of providing a
large specific surface, a porous structure that facilitatcs colonisation by
microorganisms (Fortin and Deshusses, 1999). and a low price. Disadvantages are
the low porosity (-50%) and the heavy weight which requires special construction of
the reactor. Another disadvantage we recently observed in the laboratory is that lava
rock is not chemically inert. A substantial weight loss was observed over several
months at a low pH. This might be of importance for biotrickling filter applications
with a low pH, e.g., odour treatment and H2S removal at waste water purification
plants as discussed in Section 3.3.
Random dump plastic packings such as Pall rings have bccn used in many
laboratory studies and large scale reactors. These packings are casy to handle.
however, experiments indicate that start-up is relatively long presumably because
Biolricklingfilters 115
poor biofilm establishment on the surface (Fortin and Deshusses, 1999; Kazenski and
Kinney, 2000). AIso, the relatively low specific surface area is a disadvantage for
achieving a high elimination capacity (Webster el al., 1999). On the other hand, the
use of plastic packing may be beneficial because of its stability, low cost and high
porosity (Kazenski and Kinney, 2000). Structured packings made from stainless steel
or plastic combine a high porosity and large specific surface. Good performance has
been observed with these type of packings (Zuber, 1995; Webster et al., 1998), but
they are more expensive.
Activated carbon-based packings have adsorbing properties that are not
encountered in other materials. It is often assumed that adsorption is beneficial
especially when the pollutant concentration in the waste gas is fluctuating ('peak
dampening'). However, activated carbon particles in biotrickling filters will bc
covered by a biofilm, which will decrease the absorptive capacity of the carbon. A
better option would be using a separate activated carbon unit preceding the
biotrickling filter (Weber and Hartmans, 1995).
A relatively new packing used in biotrickling filters is open-pore
polyurethane foam products. Published data on the performance are scarce, but some
major European vendors offer these products, often for biotrickling filters designed
for H2S removal. Our own experiences in the laboratory (unpublished) indicated
improved performance with polyurethane foam cubes over other types of packing,
especially at high gas flow rates with low H2S concentrations. Owing to the open
structure and high porosity, the pressure drop over the packing remained low at a
relatively high gas velocity. The large specific surface area proved beneficial with
respect to mass transfer limitation observed at low H2S concentration.
4.4 INOCULATION AND MICROBIAL ECOLOGY
Unlike biofilters with an indigenous microbial population, biotrickling filters need to

be inoculated with microorganisms. The following sources are commonly used:
• Activated sludge from waste water treatment plants;

• Soil or water samples from sites or plants contaminated with the pollutant of
interest;
• Consortia that are enriched in the laboratory on the pollutant of interest;
• Pure cultures, degrading the pollutant of interest and obtained either from culture
collections or isolated from mixed consortia;
• Samples of biotrickling filters treating the same or a comparable waste gas
stream.
Selection of the inoculum source becomes incrcasingly important when the

pollutant is more difficult to de grade (Figure 4.6).
For easily biodegradable pollutants or for complex waste gases such as
odours, a general source such as activated sludge is often sufficient. Activated sludge
contains a wide spectrum of bacteria, capable of degrading many different
compounds, and is thus a good choice for odour treating biotrickling filters. Use of
adapted consortia or pure cultures with high biodegradation potential is favoured
when treating gases containing poorly biodegradable pollutants. An example is the
treatment of MTBE (Figure 4.7). Effective removal of MTBE after inoculation with
exposed soil and water samples was observed only after seven months of operation.
The start-up phase could be significantly reduced by inoculation with an adapted

consortiurn. Start-up was the fastest when using a concentrated pure culture isolated
in the laboratory and capable of degrading MTBE at a high rate.
Biodegradability of Specificity of the Example

the pollutant inoculum
Activated sludge
High Low
Soillwater sample from
contaminated site
Mixed consortium, adapted to

the pollutant in the lab
Low High Pure culture of pollutant-

degrading microorganisms
Figure 4.6. Inoculation guide for biotrickling filters.
1.2
1.0
0.8 - I - - -.....--9Il.......I • pure culture

iii o adapted consorlium
c
1;::: + contaminated soil/water
u
w 0.6
u
w 0.4
0.2
0.0
O 50 100 150 200 250 300
Time (days)
Figure 4.7. Start-up ofMTBE degrading biotrickling filters inoculated with different sources
of microorganisms (Cox and Deshusses, unpublished data for the pure culture and adapted
consortium, Fortin and Deshusses, 1999 for the contaminated soil and water inoculum data).
As discussed in Section 2.1, a complex ecological community exists in

biotrickling filters. A broad variety of bacteria, yeasts, fungi, protozoa, nematodes,
algae and other higher organisms is generally observed (e.g., Hugler et al., 1996;
Sch5nduve et al., 1996; Cox and Deshusses, 1999a). This can be attributed to various
factors:
Biotrickling filters Jl7
• Biotrickling filters are open systems, hence, invasion by and establishment of

microorganisms and other organisms from the environment is unavoidable.
• Biofilm growth is coupled to the production of polymers, cell lysis products and
mctabolites from the primary pollutant degrading population. These products
serve as sources of carbon and energy for the secondary population. Analysis of
the biofilm indicate that these alternative carbon sources may constitute a
significant part of the biofilm (Arcangeli and Arvin, 1992; Pedersen et al., 1997),
which may explain the existence of a thriving secondary population in
biotrickling filters. Apart from this, the cells in the biofilm (the primary and
secondary population) itself may serve as food source for predating organisms
such as protozoa (Cox and Deshusses, 1999a).
• Although biotrickling filters in mathematical modelling are considered to be
homogenous reactors, the reality is quite different. Heterogeneity with respect to
parameters affecting microbi al activity (e.g., pollutant, oxygen and nutrients
concentrations, pH, metabolic products) may exist over the height of the reactor
as well as locally insi de the biofilm. Consequently, different microbi al
populations may develop depending on the local conditions. Biofilm
heterogeneity in biotrickling filters has been demonstrated by various
microscopical tcchniques (Hugler et al., 1996; Moller et al., 1996; Schonduve et
al., 1996; Pedersen et al., 1997).
The use of molecular techniques has improved the understanding of the

dynamics of microbi al populations in biotrickling filters. The structure of microbial
communities as determined by DNA finger printing appears to be more complex than
found with c1assical techniques such as plating of biofilm samples on
microbiological media (Allen et al., 2000). DNA finger-printing also allows for a
comparison of microbial communities in different biotrickling filters and adaptations
of communities to changing conditions in the reactor. Of particular interest are the
studies ofPedersen et al. (1997) and Moller et al. (1996). They used epi-fluorescence
microscopy and a 16S rRNA targeting probe to study the establishment and in situ
activity of a Pseudomonas putida species in the biofilm of a toluene degrading
biotrickling filteL One important outcome of these studies is that the specific activity
of Pseudomonas putida in the biofilm was lower than as found in batch cultures with
the pure species. This demonstrates the potential risk of overestimating the biological
degradation rate in biotrickling filters if biokinetic parameters derived from batch or
continuous cultures are used.
4.5 NUTRIENTS
In order to maximise the pollutant elimination capacity, nutrients should be present at

concentrations high enough to maintain an active, growing culture. In general, the
nutrient requirement is qualitatively the same as the elemental composition of
biomass. With the waste gas pollutant scrving as a source of carbon and energy (i.e.,
VOCs), essential elements other than carbon should be supplied to the biotrickling
filter (Figure 4.8). Biotrickling filter performance is maximised when these are added
in excess, i.e., the pollutant in the waste gas is the limiting substrate.
The amount of nutrients to be added can be estimated from biomass yield
coefficients an each of the essential elements (Pirt, 1975). In the following
Source Element Bacterial cell
Pollutant - C
Air/water - - O, H - - - 1 - +
Medium - N, S, P, K, Mg
~ Trace elements
Fe, Cu, Mo, etc.
Figure 4.8. Utilisation of nutrients by heterotrophic bacteria with the pollutant as the carbon
source for growth.
ca1culations it is assumed that the waste gas pollutant is the carbon source for
microbial growth. Since the pollutant is the limiting substrate, the amount of biomass
formed is proportional to the amount of pollutant degraded. The first step is to
calculate the expected biomass accumulation rate (X,ee in g dry biomass/m3 reactor.
h) from the desired or design value for the e1imination capacity (g C-pollutant/m3
reactor.h) and the biomass to carbon yield coefficient (g dry biomass/g carbon):
X,ee = EC Y XlC (4.7)
For a growing cultures, YXlC is approximately 1.0--1.15. As discussed below,

the overall biomass yie1d of the mixed process culture in a biotrickling filter is lower
because of predation by higher organisms and other secondary processes. To prevent
limitation of growth by element E, the minimal supply rate (SR E in g E/m 3 reactor.h)
can be ca1culated according to:
(4.8)
where YXIE is the biomass yie1d coefficient on element E (g dry biomass/g element
E). Typical examples of biomass yield coefficients are listed in Table 4.4. Note that
these are for pure cultures growing in bioreactors, recommended values for
biotrickling filtration are probably close to those listed as the upper limit.
The above calculation assumes that nutrients supplied are irreversibly
incorporated into a continuously growing primary culture. Cell death and Iysis, and
the presence of a secondary population and predators in the biotrickling filter is thus
neglected. These processes result in nutrient-recycling, causing a larger availability
of nutrients. The supply rate of nutrients can be decreased accordingly, but this is
very difficult to quantify. Experiments with toluene degrading biotrickling filters
have indicated that nutrient recycling by secondary processes may reduce the
required nutrient supply rate by a factor two (Cox and Deshusses, 2000c). Not
unexpectedly, the recycle of nutrients by predation and secondary processes is
Table 4.4. Examples of biomass yield coefficients on some essential nutrients
Element Y XlE (g dry biomass/g element)

Lower !imit Upper !imit
Nitrogen 83 20
Phosphorous 23 285
Potassium 25 111
Sulphur 38 3300
Calcium 111 28000
Iron 200 50000
Copper 5000 100000
irrelevant when excess nutrient is provided. Under these circumstances rapid growth
of biomass and re actor clogging is often experienced. This is further discussed in
Section 5.
For full-scale biotrickling filters it is often not practical or too expensive to
feed a chemically defined medium. Alternative nutrient sources include agricultural
fertilisers and secondary effluent water from waste water treatment plants. Although
these sources may not have the most ideal composition, they have shown to be
satisfactory in supporting sustained pollutant elimination in pilot-scale biotrickling
filters (Webster et al., 1999).
4.6 LIQUID RECYCLING
A distinctive feature of biotrickling filters is the recycling of a liquid phase over the
packing. The recycle liquid supplies water and nutrients to the biofilm and it removes
metabolic products that would otherwise accumulate in the biofilm in possibly toxic
concentrations. Continuous monitoring and adjustment of chemi cal parameters in the
recycle liquid allow for better control of microbial activity in the biofilm, which is an
important advantage ofbiotrickling filters over biofilters.
Biotrickling filters can be operated with the gas and liquid flowing co-
currently or counter-currently. Co-current operation is often advocated, as counter-
current operation would allow stripping of the pollutant from the liquid phase at the
gas outlet side of the re actor, thereby causing a lower elimination capacity. This is
merely a matter of theoretical concern. Experimental data indicate no major
difference between either mode of operation (Diks and Ottengraf, 1991b). Moreover,
stripping effects are likely to occur only in biotrickling filters overloaded with the
pollutant, causing a high pollutant concentration in the liquid phase. In most field
applications a near complete removal of the pollutant is usually the objective, in
which case the pollutant concentration in the liquid phase will be negligible as well
as possible stripping effects.
The rate of liquid recycling is often expressed as the superficial liquid
ve1ocity. Ve10cities typically range from 0.05 up to 20 m/h. Usually an increase of
the elimination capacity is observed with a higher superficial liquid velocity. This
has been attributed to an increase of the wetted biofilm surface area (Diks and
Ottengraf, 1991 b), or to a reduction of mass transfer resistance in the liquid phase
(Hartmans and Tramper, 1991). Treatment of pollutants that release acid end-
products is a case that requires special attention. At too Iowa trickle rate, the liquid
may not be able to remove the produced acids fast enough so that the pH could be
reduced to the extent of inhibiting microbi al activity (Diks and Ottengraf, 1991 b; Oh
and Bartha, 1994). This situation can be assessed by caiculating the expected pH at
the bottom of the column using the trickle rate, the acid production rate and the
trickling liquid buffer capacity. Altematively, one could measure the pH in the liquid
trickling from the bottom of the packed bed.
The upper !imit of the superficial liquid velocity is determined by the
flooding point of the reactor, which will de crease over time as biomass accumulates.
The superficial velocity will also have an effect on the removal of hydrophobic
pollutants. Thick layers of recycle liquid covering the biofilm may act as a barrier,
causing a reduced mass transfer rate of the pollutant from the gas phase into the
biofilm (Zhu el al., 1998; Nascimento el al., 2000). This type of mass transfer
resistance can be reduced by lowering the superficial liquid velocity. Also
intermittent instead of continuous liquid recycling may improve the mass transfer of
hydrophobic pollutants and improve their removal in biotrickling filters (Wolff,
1992; Pol el al., 1998).
5. Biomass growth and long term performance of biotrickling fiIters
With continuous supply of nutrients, the biotrickling filter will accumulate large
amounts of biomass. At present, clogging of thc re actor over the long run is one of
the major obstacles at the industrial scale. This section describes the various stages of
biomass accumulation, the consequences for the performance and long-term stabi!ity,
and methods to control biomass growth.
5.1 BIOMASS GROWTH KINETICS AND POLLUTANT ELIMINATION
In general, pollutants are used by the primary culture to produce new biomass and to
generate energy for maintenance activities of the existing biomass. The biotrickling
filter process can at first instance be considered as a continuous culture (e.g.,
neglecting secondary processes, local heterogeneities, and mass transfer effects):
EC = (1l1Y + m) X (4.9)
).Ima, S (4.10)
).1
Ks + S
Equation 4.9 describes the elimination capacity as a function ofthe amount of

active biomass of the primary culture per volume of re actor (X), the specific growth
rate (Il), the maintenance energy requirement coefficient (m) and the biomass yield
on the limiting nutrient (Y). Equation 4.10 is the Monod equation for growth on
pollutant S as defined in Chapter 3, which can be extended to include oxygen
limitation (i.e., the factor OI {Ks o + O}), nutrient limitations (element E) other than
the pollutant (EI {KS,E + E}) or the presence of inhibitors (II {l + I/KI}).
The population in biotrickling filters consists of different sub-populations,
e.g., the primary culture, the secondary population, predators, and other higher
organisms. Biomass accumulation in the biotrickling filter is the addition of growth,
cell death and lysis, predation by higher organisms and washout of detached biomass
Biolricklingjilters 121
via the liquid purge each sub-population present in the system. This is shown in
Equation 4.11, which expresses the overall biomass accumulation rate in the
biotrickling filter as the sum of accumulation rates of sub-populations (e.g., primary
culture, secondary population, predators) present in the biotrickling filter.
Overall rate of biomass accumulation = I {(Il-d) X - predation - washout} (4.11)
where d is the specific death rate of the sub-population. Predation and wash-out of
biomass have received little attention in biotrickling filter research and these
processes have not been quantitatively described as yet. Despite the limitation of
being virtually impossible to solve, Equation 4.11 demonstrates that the overall
biomass accumulation rate in biotrickling filters depends on the intrinsic microbial
growth rate on the one hand, and various mechanisms of biomass reduction on the
other hand. This allows for development of various strategies to control the
accumulation ofbiomass, as discussed in Section 5.5.
5.2 STAGES OF BIOMASS FORMATION
Biotrickling filters are often operated with sufficient supply of nutrients to maximise
the pollutant removal rate. Various stages of biomass accumulation can be
distinguished after inoculation of the biotrickling filter with an appropriate source of
microorganisms. At first, a lag phase will occur during which the microorganisms
adapt to environmental conditions in the reactor and to the pollutant (Figure 4.9). The
duration of the lag phase mainly depends on the biodegradability of the pollutant.
Lag phases as short as a few hours to a few days have been reported for e.g. ethanol
and toluene. On the other hand, removal of methyl-terl-butyl ether (MTBE), a
gasoline additive with low biodegradability, may only be observed after several
months of operation (see Section 4.4).
The exponential growth phase follows the lag phase (Figure 4.9), with rapid
biomass accumulation and an exponentially increasing biofilm thickness (Pedersen el
al., 1997). Pollutant degradation takes place throughout the entire biofilm and the
overall pollutant removal rate increases with the thickness. During this phase a short
term peak in the elimination capacity can sometimes be observed. This has been
linked to a rapid proliferation of suspended biomass in the recyc1e liquid (Cox et al.,
2000). Generally, the. pollutant removal rate soon reaches a constant value (Figure
4.9, pseudo steady state). At this point, diffusion limitation ofthe pollutant (Okkerse
et al., 1999), oxygen (Kirchner el al., 1996; Mpanias and Baltzis, 1998) and/or
nutrients (Zhu el al., 1996; Rihn el al., 1997) in the biofilm causes biodegradation to
take place only in the upper layer at the gas/biofilm interface. Similar to fixed film
waste water treatment processes, the effective biofilm thickness in biotrickling filters
is as a rule of thumb in the order of 25-300 Ilm (e.g., Kirchner el al., 1996;
Schonduve el al., 1996; Mirpuri et al., 1997a).
Prolonged operation with nutrient supply causes further growth of the
biofilm, however, the effective biofilm thickness and the elimination capacity remain
constant (Zuber, 1995; Cox and Deshusses, 1999a; Okkerse el al., 1999). Prolonged
operation therefore primarily results in formation of biomass not actively involved in
biodegradation. Thick biofilms in biotrickling filters contain large fractions of
inactive or dead cells and inert material (Schonduve el al., 1996; Hugler et al., 1996;
Lag Pseudo steady Clogging

phase state
Effective biofilm
Diffusion limited
Inactive
Time
Figure 4.9. Schematic of the long-term performance of biotrickling filters with continuous
supply of nutrients. The upper section shows the development of the elimination capacity
over time, whereas biofilm grown is shown in the lower section. (Note that the schematic is
not to scale).
Okkerse et al., 1999; Pedersen et al., 1997). Finally, biotrickling filters subject to
excessive biomass formation (Figure 4.9, clogging) face c10gging with increasing
pressure drops across the reactor and a decreasing pollutant removal rate (Cox et al.,
1998; Cox and Deshusses, 1999a; Okkerse el al., 1999).
5.3 BIOFILM ARCHITECTURE AND MASS TRANSFER
The gas/biofilm interfacial area is an important parameter in assessing the mass

transfer in biotrickling filters. Usually a planar geometry is assumed, with aflat and
homogeneous biofilm and a specific surface are a the same as that of the supporting
packing. Recently a few innovative techniques have been developed that demonstrate
the heterogeneous structure of the biofilm. Scanning confocal laser microscopy
(SCLM) showed a porous biofilm with cell-free channels in a toluene-degrading
biotrickling filter (Moller el al., 1996). One-dimensional scanning of the biofilm
demonstrated a rough surface with significant variation in the local biofilm thickness
(Okkerse el al., 1998). Computed axial tomography was used to visualise in situ
biofilms in toluene degrading biotrickling filters (Deshusses el al., 1998). This
technique allows for quantification of the distribution and sizes of air and water
channels and the determination of the surface area without disturbing the biofilm
structure as is the case with most ex situ techniques. A large variation in pore
diameter of air and water channels was found, and the surface area of the biofilm was
larger than expected due to a rough surface.
Biofilm heterogeneity resuIts in a greater area for mass transfer -and hence,
in a faster mass transfer rate- than from planar biofilm geometry would be expected.
This has been experimentally confirmed by Pedersen and Arvin (1997a), who
showed that the mass transfer coefficient (KLa) in biotrickling filters was 25-140 %
larger when the packing was covered by a biofilm.
5.4 STEADY STATE VERSUS NON-STEADY STATE
Although many mathematical models are based on growing cultures according to a

Monod type of equation, they frequently assume a steady state with no accumulation
of biomass over time. This obviously is a weakness; steady state models may
accurately predict the performance at a certain point in time, but they fail to predict
the long-term performance. The calculations shown in Table 4.5 demonstrate the
importance of c1ogging. For a toluene degrading biotrickling filter with an
elimination capacity of 40 g/m3 .h it may take only II weeks to produce an amount of
biomass equal to the reactor volume. However, flooding may already be observed at
biomass contents of about 50% by volume of the reactor (Cox et al., 1998). Aiso as
indicated by the calculation in Table 4.5, the pollutant load and elimination capacity
as well as the biomass yield are the principal parameters in determining the biomass
accumulation rate (Okkerse et al., 1999; Cox et al., 1998). The critical parameter is
the biomass yield coefficient, which e.g. for pure cultures of Pseudomonas species
with excess of nutrients has been found to range from 0.3 to 1.2 g biomass/g toluene
(Mirpuri et al., 1997b). Secondary processes in biotrickling filters -e.g., cell death
and lysis, cryptic growth and predation by higher organisms- enhance nutrient
recycling, thus decreasing the actual growth yield and extending the operating life
time (Diks et al., 1994a; Cox and Deshusses, 1999a).
Table 4.5. Example ofthe estimation ofthe biomass accumulation rate in biotrickling filters
treating toluene vapours. The amount of biomass produced in 77 days is equal to the total
reactor volume
Toluene elimination capacitya 40 g toluene/m l reactor.h

= 36.5 g C-toluene/ml reactor.h
Degree of carbon mineralizationb 69%
C-toluene to C-biomass conversion ratio 31%
C-biomass accumulation rate 11.3 g C-biomass/ml reactor .h
Dry biomass accumulation rateC 25.7 g dry biomass/ml reactor .h
Wet biomass accumulation rated 559 g wet biomass/ml reactor .h
-0.013 ml biomass/m 3 reactor .day
a Average value from Pedersen and Arvin, 1997b. b Cox el al., 1998. C Carbon content in dry biomass
of 44%, and d water content in wet biomass of 4.6% (Cox and Deshusses, 1999a).
Recently, dynamic mathematical models have been developed that account

for biomass accumulation over time (Alonso et al., 1997, 1998b). These models are
useful in identifying the underlying cause of decreasing pollutant efficiency in
c10gged biotrickling filters. Although one would expect higher removal rates with
increasing amounts of biomass (Equation 4.9), overall performance decreases
because of diminishing accessibility of the pollutant to the biomass. With increasing
thickness, biofilms on adjacent partic1es will overlap each other, thus decreasing the
actual biofilm specific surface available for mass transfer into the biofilm (Alonso et
al., 1998).
124 H.H.J Cox and MA Deshusses
5.5 PREVENTION OF CLOGGING
The first option to prevent clogging is reduction of the biomass accumulation rate
(Equation 4.11). According to Equations 4.9 and 4.11, this can be achieved by
reducing thc biomass yield coefficient (Y XlS), by increasing the maintenance
requirements (m s) or by stimulating predation or the death rate. The challenge is to
maintain a high pollutant removal rate (Equation 4.9), however, growth, yield,
activity and maintenance are interrelated parameters reflecting general cell
metabolism and, as such, they are often difficult to influence independently.
Various strategies to reduce the specific growth rate in biotrickling filters
have been investigated in the laboratory (Table 4.6). These include limiting the
supply of nutrients essential for growth, the use of nitrate as a nitrogen source instead
of ammonium and the addition of compounds such as NaCI in concentrations that
partiaIly inhibit microbial growth. In general, these strategies also result in reduction
of microbial activity. Hence, a relatively large re actor volume will be required to
treat the same volume of waste gas at the same efficiency. An interesting option is
the use of fungi, as these types of microorganisms show a higher removal rate than
bacteria under nutrient-limiting conditions in toluene degrading biotrickling filters
(Weber and Hartmans, 1996). Predation of biomass by higher organisms such as
protozoa is a relatively unexplored area in biotrickling filter research, although the
phenomenon is quite known in waste water treatment systems. Protozoa are natural
inhabitants of bioirickling filters, and stimulation of their activity was showed ta
reduce the biomass accumulation rate (Cox and Deshusses, 1999a). The advantage of
predation over options mentioned earlier is that the process culture is allowed to
grow at the maximum rate, and hence with maximum microbi al activity. In fact,
stimulation of pratozoan predation caused a slight increase of the maximum
elimination capacity, which was tentatively attributed ta an increased rate of nutrient
recycling (Cox and Deshusses, 1999a). The predation rate was, however, not high
enough to fully counterbalance micrabial growth, and only delaying clogging of the
reactor could be achieved.
The second option to prevent clogging is periodical removal of the produced
biomass. In this scenario the biotrickling filter is operated at a high climination
capacity allowing rapid accumulation of biomass. Removal of biomass can be done
physically or chemically (Table 4.6). Physical removal of biomass relies on biofilm
detachment by high shear forces. This can be do ne by backwashing of the reactor, or
by periodical stirring of the packed bed. Although these techniques result in
prolonged, stable biotrickling filter operation, certain drawbacks exist (Table 4.6). It
should be noted that shear stress of the trickling liquid during normal operation of the
biotrickling filter is not sufficient to remove substantial amounts of attached biomass
(Pedersen et al., 1997; Cox and Deshusses, 1999b). Chcmical removal of biomass is
a simple operation as no major changes of the reactor configuration are required. A
stable toluene degrading biotrickling filter has been obtained by periodical washing
of the packing with a NaOH solution for 3 hours (Weber and Hartmans, 1996). A
post-treatment with HCI was needed to restore the pH to a neutral value. Other
chemicals such as sodium hypochlorite and hydrogen peroxide may be more
effective in removing biomass, but they are also more toxic to the micrabial
population (Cox and Deshusses, 1999b). This could potentiaIly slow down the restart
ofthe reactor.
Table 4.6. Control ofbiomass accumulation in biotrickling filters
Biomass control option Principle Disadvantage References

Reduction ofthe biomass
accumulation rate
N utrient limitation Reduction ofthe biomass Reduction of the specific Holubar et al., 1999; Weber and
yield coefficient microbia1 activity Hartmans, 1996; Cox et al., 1998
Wilbker and Friedrich, 1996
N-N0 3" instead ofN-NH/ Reduction ofthe biomass Reduction of the specific Schonduve et al., 1996; Smith et
yield coefficient microbial activity al., 1996
Addition of growth inhibitors Reduction ofthe biomass Reduction of the specific Schonduve et al., 1996; Diks et b::i
yield coefficient microbial activity al., 1994b o·
'?
L se of specific microbial species Selection of species with low Selective advantagc for species Weber and Hartmans, 1996 Ci·
biomass yield and high with high biomass yield ~
S·
activity liQ
Protozoan predation Reduction ofthe biomass Protozoan predation does not fully Cox and Deshusses, 1999a ~
yield coefficient balance microbi al growth ~
..,
'"
P"riodical removal of excess biomass
Backwashing Biomass removal 40% larger reactor volume needed Sori al et al., 1995; Smith et al.,
for full packing fluidisation. 1996; Smith et al., 1998
Requirement for packing media
(hat can be fluidised
Periodical stirring Biomass removal Complicated reactor design and Laurenzis et al., 1998; Wilbker et
construction. al., 1997
Chemical washing Biomass removal Toxicity to microorganisrns. Weber and Hartmans, 1996; Cox
Secondary wastes. and Deshusses, 1999b
.......
N
V,
126 fIfJJ Cox and MA. Deshusses
It is worth to stress that the biomass control strategies discusscd in this

Section have only been investigated in the laboratory. Very little experience is
available at industrial scale, because the few existing full-scale biotrickling filters
have ali been designed for applications with a low potential for clogging. Cost-
effective, long term operation of biotrickling filters requires finding the optimum
between two extreme cases (Deshusses and Cox, 1999):
1. Operation of large volume, low performance biotrickling filters that do not or

only very occasionally require biomass removal.
2. Operation of small volume, high performance biotrickling filters with frequent
biomass removal.
It may be expected that with further understanding of microbiological

processes in biotrickling filters solutions will be found that combine high microbial
activity and a low biomass accumulation rate.
6. ConcIusions
Biological waste water treatment has become part of our everyday life. Although, the
process is not fully understood, waste water treatment is a major improvement for
water quality. More than 100 years after the introduction of the first waste water
treatment plants, bioreactors for the treatment of gaseous effluents have the potential
to play a similar role for air quality. In the past decades major progress has been
accomplished in the development of vapour-phase bioreactors, in particular in
biotrickling filters. While the level of understanding of the biotrickling filtration
process sti Il remains limited, the evident success of waste water biotreatment should
be a motivation to pursue active research. Clearly, thc full potential of biotrickling
filtration for air pollution control has not yet been explored. There is great promise
for both effective and environmentally friendly biotreatment of contaminated gases.
References
AlIen, D.G., Kong, Z., Fulthorpe, R.R. and Farhana, L. Thermophilic biofiltration of volatile organic
compounds. In: Proceedings of the 93,d Annual Meeting & Exhibition of the Air & Waste Manage.
Assoc., 2000, Pittsburgh, PA.
Alonso, c., Suidan, M.T., Sorial, G.A., Smith. F.L., Biswas, P., Smith, PJ. and Brenner. R.C. 1997.
Gas treatment in trickle-bed biofilters: biomass, how much is enough? Biotechnol. Bioeng. 54: 583-
594.
Alonso, C., Zhu, X., Suidan, M.T., Kim, B.R. and Kim, BJ. 1998a. Modelling ofthe biodegradation
process in a gas phase bioreactor - estimation of intrinsic parameters. In: Proceedings of the USC-
TRG Conference on Biofiltration, F.E. Reynolds Jr. (ed.). The Reynolds Group, Tustin, CA, pp. 211-
218.
Alonso, c., Suidan, M.T., Kim, B.R. and Kim, B.1. 1998b. Dynamic mathematical model for the
biodegradation ofVOCs in a biofilter: biomass accumulation study. Environ. Sci. Technol. 32: 3118-
3123.
Arcangeli, 1.P. and Arvin, E. 1992. Toluene biodegradation and biofilm growth in an aerobic fixed-
film reactor. Appl. Microbiol. Biotechnol. 37: 510-517.
Chitwood, D.E., Devinny, J.S. and Reynolds, F.E. 1999. Evaluation of a two-stage biofilter for
treatment ofPOTW waste air. Environ. Progr. 18: 212-221.
Chitwood, D.E., Devinny, J.S. and Maiburg, E. 2000. Heterogeneous flow through biofilter media. In:
Proceedings of the 93'd Annual Meeting & Exhibition of the Air & Waste Manage. Assoc., 2000,
Pittsburgh, PA.
Chou, M.-S. and Huang, J.-J. 1997. Treatment of methyl ethyl ketone in air stream by biotrickling
filters. J. Environ. Eng. 123: 569-576.
Chou, M.-S. and Liu, J.-H. 2000. Biotrickling filtration of nitric oxide. J. Air Waste Manage. Assoc.
50: 502-508.
Cox, H.H.J. and Deshusses, M.A. 1998. Pollutant biodegradation in biotrickling filters revisited. In:
Proceedings of the USC-TRG Conference on Biofiltration, F.E. Reynolds Jr. (ed.). The Reynolds
Group, Tustin, CA, pp. 219-226.
Cox, H.H.J., Nguyen, T.T. and Deshusses, M.A. 1998. Elimination oftoluene vapors in biotrickling
filters: performance and carbon balances. In: Proceedings of the 91 SI Annual Meeting & Exhibition of
the Air & Waste Manage. Assoc., 1998, Pittsburgh, PA.
Cox, H.H.J. and Deshusses, M.A. 1999a. Biomass control in waste air biotrickling filters by protozoan
predation. Biotechnol. Bioeng. 62: 216-224.
Cox, H.H.J. and Deshusses, M.A. 1999b. Chemi cal removal of biomass from waste air biotrickling
filters: screening ofchemicals of potential interest. Water Res. 33: 2383-2391.
Cox, H.H.J. and Deshusses, M.A. 2000a. Innovative experimental setup for the parallel operation of
multiple bench scale biotrickling filters for waste air treatment. Environ. Technol. 21: 427--435.
Cox, H.H.J. and Deshusses, M.A. 2000b. Combined removal of H2S and to1uene in a single-stage
biotrickling filter. In: Proceedings of the 93'd Annual Meeting & Exhibition of the Air & Waste
Manage. Assoc., 2000, Pittsburgh, PA.
Cox, H.H.J. and Deshusses, M.A. 2000c. Thermophilic biotrickling filtration of ethanol vapors. In:
Proceedings of the USC-TRG Conference on Biofiltration, F.E. Reynolds Jr. (ed.). The Reynolds
Group, Tustin, CA, pp. 159-166.
Cox, H.H.J., Nguyen, T.T. and Deshusses, M.A. 2000. Toluene degradation in the recycle liquid of
biotrickling filters for air pollution control. Appl. Microbiol. Biotechnol. 54: 133-137.
Davison, B.H., Barton, J.W., Klasson, K.T. and Francisco, A.B. 2000. Influence of high biomass
concentrations on alkane solubilities. Biotechnol. Bioeng. 68: 279-284.
Deshusses, M.A., Cox, H.H.J. and Miller, D.W. The use of CAT scanning to characterize bioreactors
for waste gas treatment. In: Proceedings of the 91 SI Annual Meeting & Exhibition of the Air & Waste
Manage. Assoc., 1998, Pittsburgh, PA.
Deshusses, M.A. and Cox, H.H.J. A cost benefit approach to reactor sizing and nutrient supply for
biotrickling filters for air pollution control. In: Proceedings of the 92 nd Annual Meeting & Exhibition
ofthe Air & Waste Manage. Assoc., 1999, Pittsburgh, PA.
Devinny, J.S., Webster, T.S., Torres, E. and Basrai, S. 1995. Biofiltration for removal of PCE and
TCE vapors from contaminated air. J. Hazard. Waste Hazard. Mat. 12: 283-293.
Devinny, J.S., Deshusses, M.A. and Webster, T.S. 1999. Biofiltration for air pollution control. Lewis
Publishers, Boca Raton.
128 HHJ. Cox and M.A. Deshusses
Diks, R.M.M. and Ottengraf, S.P.P. 1991 a. Verification studies of a simp1ified model for the removal
of dichloromethane from waste gases using a biological trickling filter (Par! 1). Bioproc. Eng. 6: 93-
99.
Diks, R.M.M. and Ottengraf, S.P.P. 1991b. Verification studies ofa simplified model for the removal
of dichloromethane from waste gases using a biologic al trickling filter (par! II). Bioproc. Eng. 6: 131-
140.
Diks, R.M.M., Ottengraf, S.P.P. and Vrijland, S. I 994a. The existence of a biological equilibrium in a
trickling filter for waste gas treatment. Biotechnol. Bioeng. 44: 1279-1287.
Diks, R.M.M., Ottengraf, S.P.P. and Van den Oever, A.H.C. 1994b. The influence of NaCI on the
degradation ofdichloromethane by Hyphomicrobium sp. Biodegradation. 5: 129-141.
Fortin, N.Y. and Deshusses, M.A. 1999. Treatment of methyl tert-butyl ether vapors in biotrickling
filters. 1. Reactor startup, stcady-state performancc, and culture characteristics. Environ. Sci. Technol.
33: 2980-2986.
Hartmans, S. and Tramper, 1. 1991. Dichloromethane removal from waste gases with a biotrickling
filter. Bioproc. Eng. 6: 83-92.
Heinze, U. and Friedrich, c.G. 1997. Respiratory activity of biofilms: measurement and its
significance for the elimination of n-butanol from waste gas. Appl. Microbiol. Biotechnol. 48: 411-
416.
Holubar, P., Andorfer, C. and Braun, R. 1999. Effects of nitrogen limitation on biofilm formation in a
hydrocarbon-degrading trickle-bed filter. Appl. Microbiol. Biotechnol. 51: 536-540.
Hugler, W.c., Cantu-De la Garza, 1. and Villa-Garcia, M. Biofilm analysis from an odor-removing
trickling filter. In: Proceedings ofthe 89'" Annual Meeting & Exhibition ofthe Air & Waste Manage.
Assoc., 1996, Pittsburgh, PA.
Hugler, W., Acosta, C. and Revah, S. 1999. Biological removal of carbon disulfide fram waste aiI'
streams. Environ. PragI'. 18: 173-177.
Kazenski, S.L. and Kinney, K.A. Biofiltration of paint spray booth emissions: packing and media
considerations and VOC interactions. In: Proceedings of the 93'd Annual Meeting & Exhibition of the
Air & Waste Manage. Assoc., 2000, Pittsburgh, PA.
Kirchner, K., Wagner, S. and Rehm, H.-1. 1992. Exhaust gas purification using biocatalysts (fixed
bacteria monocuhures) - the influence of biofilm diffusion rate (0 2) on the overall reaction rate. Appl.
Microbiol. Biotechnol. 37: 277-279.
Kirchner, K., Wagner, S. and Rehm, H.-1. 1996. Removal of organic aiI' pollutants from exhaust gases
in the trickle-bed bioreactor. Effect of oxygen. Appl. Microbiol. Biotechnol. 45: 415-419.
Kraakman N.J.R., Melse, R.W., Koers, B. and Van Dijk, 1. 1998. Biological treatment ofwaste gases
containing H 2S in combination with odor and or CS1 . In: Praceedings ofthe USC-TRG Conference on
Biofiltration, F.E. Reynolds lI'. (ed.). The Reynolds Group, Tustin. CA, pp. 91-98.
Laurenzis, A., Heits,H., Wiibker, S.-M., Heinze, U., Friedrich, C. and Werner, U. 1998. Continuous
biological waste gas treatment in stirred trickle-bed reaclor with discontinuous removal of biomass.
Leson, G. and Winer, A.M. 1991. Biofiltration - An innovative air pollution control technology for
VOC emissiollS. l. Air Waste Manage. Assoc. 41: 1045-1054.
Lide, D.R. 1997. Handbook ofChemistry and Physics. 77'" edition, CRC Press, Boca Raton.
Lobo, R., Revah, S. and Viveros-Garcia, T. 1999. An analysis of a trickle-bed bioreactor: carbon
disulfide removal. Biotechnol. Bioeng. 63: 98-109.
Mirpuri, R., Sharp, W., Villaverde, S., Jones, W., Lewandowski, Z. and Cunningham, A. 1997a.
Predictive model for toluene degradation and microbial phenotypic profiles in flat plate vapor phase
bioreactor. J. Environ. Eng. 123: 586-592.
Mirpuri, R., Jones, W. and Bryers, J.D. 1997b. Toluene degradation kinetics for planktonic and
biofilm-grown cells of Pseudomonas putida 54G. Biotechnol. Bioeng. 53: 535-546.
Morton, R. and Caballero, R. 1996. The biotrickling story. Water Environ. Technol. June: 39--45.
Moller, S., Pedersen, A.R., Poulsen, L.K., Arvin, E. and Molin, S. 1996. Activity and three-
dimensional distribution of toluene-degrading Pseudomonas putida in a multispecies biofilm assessed
by quantitative in situ hybridization and scanning confoeal laser microscopy. Appl. Environ.
Microbiol. 62: 4632--4640.
Mpanias, e.J. and Baltzis, B.e. 1998. An experimental and model ing study an the removal of mono-
chlorobenzene vapor in biotrickling filters. Biotechnol. Bioeng. 59: 328-343.
Nascimento, D.M., Hudepohl, N.J., Schroeder, E.D. and Chang, D.P.Y. Bio-oxidation of nitric oxide
in a nitrifying, aerobic filter. In: Proceedings of the 93'd Annual Meeting & Exhibition of the Air &
Wastc Manage. Assoc., 2000, Pittsburgh, PA.
Nirmalakhandan, N.N. and Speece, R.E. 1988. QSAR model for prcdicting Henry's constant. Environ.
Sci. Technol. 22: 1349-1357.
Oh, Y.-S. and Bartha, R. 1994. Design and performance of a trickling air biofilter for chlorobenzene
and o-dichlorobenzene vapors. Appl. Environ. Microbiol. 60: 2717-2722.
Oh, Y.-S. and Bartha, R. 1997. Removal of nitrobenzene vapors by a trickling air biofilter. J. Ind.
Microbiol. Biotechnol. 18: 293-296.
Okkerse, WJ.H., Ottengraf, S.P.P., Osinga-Kuipers, B. and Okkerse, M. Dynamic modelling of

biotriekling filters for waste gas purification: possibilities and limitations. In: Proceedings of the 91'1
Annual Meeting & Exhibition ofthe Air & Waste Manage. Assoc., 1998, Pittsburgh, PA.
Okkerse, WJ.H., Ottengraf, S.P.P., Osinga-Kuipers, B. and Okkerse, M. 1999. Biomass accumulation
and clogging in biotrickling filters for waste gas treatment. Evaluation of a dynamic model using
dichloromethane as a model pollutant. Biotechnol. Bioeng. 63: 418--430.
Pedersen, A.R. and Arvin, E. 1995. Removal of toluene in waste gases using a biological trickling
filter. Biodegradation. 6: 109-118.
Pedersen, A.R. and Arvin, E. 1997a. Effect of biofilm growth on the gas-liquid mass transfer in a
biotrickling filter for waste gas treatment. Water Res. 31: 1963-1968.
Pedersen, A.R. and Arvin, E. 1997b. Toluene removal in a biofilm reactor for waste gas treatment.
Water Sci. Technol. 36: 69-76.
Pedersen, A.R., Moller, S., Molin, S., Arvin, E. (1997). Activity of toluene-degrading Pseudomonas
putida in the early growth phase of a biofilm for waste gas treatment. Biotechnol. Bioeng. 54: 131-
141.
Pirt, SJ. 1975. Principles ofmicrobe and cell cultivation, J. Wiley & Sons, New York.
Plaggemeier, T., Lămmerzahl, O. and Engesser, K.-H. Puri ficat ion of n-hexane polluted air by use of
a biological trickling filter: monitoring the primary-degrading microflora. In: Proceedings of the
130 H.H.J Cox and MA. Deshusses
International Symposium on Biological Waste Gas Cleaning, April 28-29, 1997. Prins W.L. and van
Ham J. (eds.). VDI Verlag GmbH, DUsseldorf, Germany, pp. 257-269.
Pol, A., Van Haren, F.J.J., Op den Camp, H.J.M. and Van der Drift, C. 1998. Styrene removal from
waste gas with a bacterial biotrickling filter. Biotechnol. Let. 20: 407-410.
Rihn, M.J., Zhu, X., Suidan, M.T., Kim, B.J. and Kim, B.R. 1997. The effect of nitrate on VOC
removal in trickle-bed biofilters. Water Res. 31: 2997-3008.
Sabo, F., Schneider, T. and Fischer, K. 1995. lmprovement and application of high-etliciency
biofilters. In: Proceedings of the USC-TRG Conference on Biofiltration, D.S. Hodge and F.E.
Reynolds Jr. (eds.). The Reynolds Group, Tustin, CA, pp. 217-224.
Schindler, l. and Friedl, A. 1995. Degradation of toluenelheptane mixtures in a trickling-bed

bioreactor. Appl. Microbiol. Biotechnol. 44: 230-233.
Schonduve, P., Săra, M. and Friedl, A. 1996. Influence of physiologically relevant parameters on
biomass formation in a trickle-bed bioreactor used for waste gas cleaning. AppI. Microbiol.
Smith, F.L., Sorial, G.A., Suidan, M.T., Breen, A.W., Biswas, P. and Brenner, R.C. 1996.
Development of two biomass control strateg ies for extended, stable operation of highly efficient
biofilters with high toluene loadings. Environ. Sci. Technol. 30: 1744-1751.
Smith, F.L., Sorial, G.A., Suidan, M.T., Pandit, A. and Biswas, P. 1998. Evaluation oftrickle bed air
biofilter performance as a function of inlet VOC concentration and loading, and biomass control. J.
Air Waste Manage. Assoc. 48: 627-ti36.
Sorial, G.A., Smith, F.L., Suidan, M.T., Biswas, P. and Brenner, R.C. 1995. Evaluation oftrickle bed
biofilter media for toluene removal. J. Air Waste Manage. Assoc. 45: 801-8\0.
Sun, A.K., Hong, J. and Wood, T.K. 1998. Modelling trichloroethylene degradation by a recombinant
Pseudomonas expressing toluene ortho-monooxygenase in a fixed-film bioreactor. BiotechnoI.
Bioeng. 59: 40-51.
Torres, E.M. and Basrai, S.S. Evaluation of two biotechnologies - Biofiltration and biotrickling
filtration for controlling air emissions from POTWs. In: Proceedings of the 91" Annual Meeting &
Exhibition ofthe Air & Waste Manage. Assoc., 1998, Pittsburgh, PA.
Van Groenestijn, J. W. and Lake, M.E. 1999. Elimination of alkanes from off-gases using biotrickling
filters containing two liquid phases. Environ. Progr. 18: 151-155.
Van Lith, c., Leson, G. and Michelsen, R. 1997. Evaluating design options for biofilters. J. Air Waste
Villaverde, S., Mirpuri, R.G., Lewandowski, Z. and Jones, W.L. 1997. Physiological and chemical
gradients in a Pseudomonas putida 54G biofilm degrading toluene in a flat plate vapor phase
bioreactor. Biotechnol. Bioeng. 56: 361-371.
Weber, F.J. and Hartmans, S. 1995. Use ofactivated carbon as a buffer in biofiltration ofwaste gases
with fluctuating concentrations oftoluene. Appl. Microbiol. Biotechnol. 43: 365-369.
Weber, F.J. and Hartmans, S. 1996. Prevention of clogging in a biological trickle-bed reactor
removing toluene from contaminated air. Biotechnol. Bioeng. 50: 91-97.
Webster, T.S., Togna, A.P., Yang, Y. and Guarini, W.J. From bench- to pilot-scale experiments: the
treatment of volatile organic compound emissions from spray paint booth applications using a
biological trickling filtration reactor. In: Proceedings of the 91 SI Annual Meeting & Exhibition of the
Air & Waste Manage. Assoc, 1998, Pittsburgh, PA.
Webster, T,S" Cox, H.H,J, and Deshusses, M,A, 1999, Resolving operational and performance
problems encountered in the use of a pilotlfull-scale biotrickling filter reactor, Environ, Progr. 18:
162-172.
Wolff, p, Biologische Abluftreinigung mit einem intermittierend befeuchteten Tropfk6rper, In:

Biotechniques for Air Pollution Abatement and Odour Control Policies, Dragt A.J, and van Ham j,
(eds,), Eisevicr, Amsterdam, The Netherlands, 1992, pp, 49--62,
Wiibker, S,-M, and Priedrich, C. 1996, Reduction of biomass in a bioscrubber for waste gas treatment
by limited supply ofphosphate and potassium ions, Appl. Microbiol. Biotechnol. 46: 475-480,
Wiibker, S,-M" Laurenzis, A" Werner, U, and Friedrich, C. 1997, Controlled biomass formation and
kinetics of toluene degradation in a bioscrubber and in a reactor with a periodically moved trickle-bed,
Biotechnol. Bioeng, 55: 686--692,
Zhu, X" Rihn, M,J" Suidan, MT, Kim, B,J, and Kim, B.R, 1996, The effect of nitrate on VOC
removal in trickle bed biofilters, Water Sci, Technol. 34: 573-581.
Zhu, X" Cao, H, and Suidan, MT The effect of moisture content on VOC removal in trickle-bed
biofilters, In: Proceedings of the 91" Annual Meeting & Exhibition of the Air & Waste Manage,
Assoc" 1998, Pittsburgh, PA.
Zuber, L. 1995, Trickling filter and three-phase airlift bioreactor for the removal of dichloromethane
from air. Ph,D Thesis, Swiss Federal Institute of Technology Zurich, Switzerland,
CHAPTER 5 BIOSCRUBBERS
Johan W. VAN GROENESTIJN
1. Introduction
Bioscrubbers for waste gas treatInent are characterised by a physica1 separation of

absorption of volatile compounds into water and subsequent biological treatment of
the water in two unit operations. The waste gas is c1eaned in an absorber, a gas-liquid
contactor, in which the pollutants are transferred from the gas phase to the aqueous
phase. The gas-liquid contactor preferred for this purpose is an absorption column
with a packed bed and counter-current flow of gas and water. The water leaving the
absorber, loaded with dissolved pollutants, is regenerated by biological treatment in a
bioreactor. Generally this bioreactor is a tank, aerated with air bubbles and containing
suspended activated sludge, with a much larger volume than the absorber. The
c1eaned water is recyc1ed to the top of the absorber again. In Figure 5.1 a drawing of
a bioscrubber can be found with the unit operations involved. To guarantee microbial
biodegradation activity, a nutrient solution should be added to the water phase. In
addition, titrants (acid or alkali) may be added to control the pH of the aqueous
phase, which can be required for an optimum biological activity and/or an optimum
absorption process. Since suspended biomass and dissolved compounds may
accumulate in the water phase, a stream of wastewater should be produced. To
replenish wasted and evaporated water, fresh water is added regularly or
continuously. Besides volatile compounds also dust and aerosols are absorbed in the
water phase. If too high concentrations of partic1es are expected in the water phase.
because of this solids absorption process, particles in the gas phase may be removed
by a filter or separate scrubber prior to the bioscrubber.
The advantages ofbioscrubbers compared to biofilters are:

• Smaller volume of the equipment;
• Better pH control;
• More reliable and predictable;
• No clogging problems ofpacking materials;
• Lower occurrence of toxic concentrations in the water phase.
The last two points are also advantages compared to bio-trickling filters.
Disadvantages:
• Only cost-effective for pollutants with a dimensionless Henry's law's coefficient
be10w 0.01;
• More difficult to attain elimination efficiencies higher than 98%;
• Chance ofthe slowest-growing micro-organisms being washed out;
• Stagnation periods of some days are not permitted;
• Disposal of sludge;
• More complicated start-up procedure;
• Higher operational costs.
133
C. Keflfles afld M. C. Veiga (eds.), Bioreactorsfor Waste Gas Treatmeflt, 133-162.
134 J W van Groenestijn
Demister Absorber liquid
Packed bed
Water
level
sensor
0---:
o - o
o o o
Absorber o o o o o o o
o o
o o o o
Fresh waler o
Waste
Tilranl -------(:t----t.,-~~__:::_:-___:-~~J------'--~-- waler
Nutrienl
solulion
Blower
Figure 5.1. Bioscrubbcr (the waste gas is introduced in the absorber in which the pollutants
are transferred to the water phase, subsequently the water is transferred to the bioreactor in
which the pollutants are biodegraded).
As a result, bioscrubbers are interesting if pH control is important, e.g., for

elimination of H 2S and S02, and for removal of highly water soluble compounds
such as alcohols, aldehydes and fatty acids. The smaller volumes are appreciated by
industries, even if slightly higher investment and operating costs are involved.
Bioscrubbers are newer than biofilters. Presumably the first bioscrubbers were
installed in the nineteen seventies in German foundries for the removal of amines,
phenol, formaldehyde and odour from waste gases (Fischer. 1990; Schippert, 1994).
In a part of the literature the word 'bioscrubber' is also used to describe bio-
trickling filters. Some authors distinguish suspended-growth bioscrubbers and fixed
film bioscrubbers (Overcamp et al., 1993). In this chapter 'bioscrubber' means
'suspended-growth bioscrubber'.
In the sections below, the construction, design and the mechanism of
operation of bioscrubbers are explained and examples of their application are given.
The chapter concludes with a cost analysis.
2. Absorbers
2.1 CONSTRUCTION
The primary function of the gas absorption contactor is to provide an extensive area
of liquid surface in contact with the gas phase under conditions favouring mass
Bioscrubbers 135
transfer (see also Chapter 2). Contactors normally employ at least one of the
foIlowing mechanisms: (1) dividing the gas into small bubbles in a continuous liquid
phase (e.g., bubble cap trays) , (2) spreading the liquid into thin films that flow
through a continuous gas phase (e.g., packed columns), and (3) forming the liquid
into smalI drops in a continuous gas phase (e.g., spray chambers) (Kohl and Nielsen,
1997).
Important types of absorbers are:
• Counter-current packed tower;

• Co-current packed tower;
• Cross-flow packed tower;
• Wet cyelone;
• Venturi scrubber;
• Spray tower;
• Eductor venturi (jet).
For bioscrubbing, the packed towers are recommended, as the other types
show poor elimination efficiencies for compounds relatively poorly soluble in water,
i.e. with dimensionless Henry's law coefficients near 0.01. Venturi scrubbers are best
in entrainment of small mist drops « 10 ~m), fumes « 1 ~m) and fine dust (l-S
~m), but weak in transfer of volatile compounds. In Table 5.1 typical gas velocities
and pressure drops for scrubbers can be found. The low pressure drops in counter-
current packed contactors are appreciated in industry, because of the lower energy
costs connected, despite of the larger volumes compared with co-current packed
contactors. Co-current operation aIlows the use of very high velocities without
concern of flooding. However, the gas velocities which are possible in co-current
packed absorbers are seldom required in bioscrubbing. Moreover, the absorption
efficiencies of co-current absorbers are lower than that of counter-current absorbers,
as in the latter the elean gas is contacted with the elean (also called 'Iean') water.
Cross flow operation of packed absorption towers is characterised by a horizontal
flow of the gas and a vertical flow of the water. However, the water wiIl not move
perfectly verticaIly, but with a slope, because of the horizontal forces of the flowing
gas. Therefore, it is recommended to use packed beds shaped with a slope
(parallelogram).
For bioscrubbers counter-current packed towers are recommended. The
towers comprise the folIowing elements (from bottom to top): a liquid reservoir or
sump, a gas inlet (generalIy one large opening), a packed bed, a space with spray
nozzles or other liquid distribution devices, a demister and a gas oudet. The packed
bed is mostly filIed with special low pressure, high efficiency plastic packing.
Examples are Jaeger Tri-packs (Jaeger Products Inc.), Intalox Snowflake packing
(Norton Chemical Process Products Corp.), TeUerettes (Ceilcote Air PoUution
Control) and Maspac (Clarkson Controls and Equipment Co.) (Kohl and Nielsen,
1997). These are randomly placed materials. Conventional random packing materials
are Raschig and PaU rings and saddles. Recently, structured packing materials gain
more popularity. The packing materials mostly have specific surface areas between
136 1. W van Groenestijn
100-300 m 2/m 3 (Fischer,1990), but not al! parts of the surface may be wetted aII the
time (2/3 of the total surface seems realistic). The sizes of packing elements range
from 2 to 8 cm. Kohl and Nielsen (1997) recommend a ratio of the diameter of the
column and the packing elements of at least 15: 1. Because of the deformability,
plastic packing is limited to an unsupported height of 3 to 5 m.
Table 5.1. Typical gas velocities and pressure drops for wet scrubbers (adapted from Kohl
and N ielsen, 1997)
Scrubber type Typical gas velocities (m/s) Pressure drop in em water

Packed contactors
Counter-current 1.5-3 1.5-3 (per m)
Co-current 3-20 8-25 (per m)
Cross flow 1.5-3 10 (total)
Wet eyclone
Entrance 10-25
Superficial 3-5 10 (total)
Spray tower
Counter-current 1-2.5 8-25 (total)
Venturi scrubber
Throat 40-50 20-30 Ctotal}
It is normally considered good practice to design for a gas flow rate that gives
a pressure drop of less than 3 cm water per meter of packing. In the case in which
the liquid tends to foam, less than 2 cm waterlm is recommended (Kohl and Nielsen,
1997). In bioscrubbing the gas loading rate ranges between 4000 and 10000 m 3/m 2 .h
for volatile compounds readily soluble in water (Kok, 1991; VDI/DIN, 1996), which
corresponds with gas veloeities in the column of 1-3 mls (Fiseher, 1990). The height
of the absorbers used in bioscrubbing range from 2 to 8 m, with 3 m as a good
average (VDI/DIN, 1996). As a consequence the gas retention time in the absorber is
a few seconds.
The water flow in the absorption tower should wet the packing surface as
much as possible. This implies a good distribution over the packing and a sufficiently
large flow rate. The number of liquid streams provided by the feed distributor should
be 10-20 per m2 and liquid redistributors are required every 5-10 tower diameters for
rings, and at least every 7 m for all types of dumped packing. The latter may be good
reason to limit the height of bioscrubber absorbers to 7 m. For a good wetting of the
surface packing about 10 m3 water/m 2 column horizontal section/h is required.
However, to prevent attachment of biomass on the packing, and eventually tower
clogging, the shear ofthe water flow should be sufficiently high. This can be reached
by applying a surface loading rate of at least 20 m3 waterlm 2.h (Kok, 1991; Schippert,
1994). VDI/DIN (1996) recommends a trickling density of 20 to 60 m 3/m 2.h and a
ratio ofliquid/air flow of 0.002-0.003.
The materials used for construction of the absorption tower cylinder should be
resistant to corrosion caused by liquids and gases and thermal!y and mechanical
resistant for the temperatures and loading rates applied. The material of out-door
absorbers should also be UV resistant. Plastics such as polyvinyl chloride,
Bioscrubbers 137
polyethylene, polypropylene and glass fibre reinforced plastics are suitable

(Anonymous, 1993; VDI/DIN, 1996).
2.2 INTERPHASE MASS TRANSFER AND KINETICS
For the design of the absorber column (height, diameter) and water flow rate for a
given gas flow, it is important to know the Henry's law's coefficient in order to
calculate how much water at least is required for a certain elimination efficiency.
Subsequently, mass transfer rate considerations give the required surface area for
interphase mass transfer. From the required surface area, the packing volume can be
calculated for each type of packing material.
When polluted gas is contacted in a counter-current absorption column, the
water leaving the column is brought into contact with the influent gas. The
concentration of a compound in this so called 'rich' water can not be higher than the
concentration in equilibrium with the concentration in the gas phase, according to
Henry's law:
(5.1)
in which
H = dimensionless Henry's law coefficient (g.m- 3 gasi g.m-3 water),
Cg = concentration of a compound in the gas phase (g/m 3 ),
CI = concentration of a compound in the liquid phase (g/m 3).
This implies that the ratio between the water flow and gas flow should be at
least H, otherwise high elimination efficiencies (> 95%) cannot be attained. For this
purpose, the absorption factor is created (Overcamp, 1993):
(5.2)
in which
A = absorption factor (dimensionless),
FI = liquid flow rate (m 3 /h),
Fg = gas flow rate (m 3 /h).
An absorption factor of at least 1 is a prerequisite for a high efficiency. A

second prerequisite for a high elimination efficiency is a sufficiently low
concentration of the 'lean' absorption liquid. According to Henry's law, in counter-
current absorbers, the concentration of the target compounds in the gas phase cannot
be lower than the concentration in the water phase multiplied by H. Low water phase
concentrations have to be created by regeneration of the absorption liquid in the
bioreactor.
If the concentration in the gas phase is higher than the concentration in the
water phase multiplied by H, in any part of the column, a driving force exists for
transfer ofthe compound from the gas phase to the liquid phase.
The mass transfer in the absorber can be described by the two film theory
developed by Lewis and Whitman (1924). In this model two bulk phases (gas and
liquid) have different concentrations and are not in equilibrium according to Henry's
law. Only at the interphase such equilibrium exists. In a very small film of liquid near
the interface, a concentration profile exists connecting liquid bulk concentration and
interphase concentration. The same is true for the gas phase. In these films the target
molecules move with a certain velocity, depending on the properties of the
compound and the solvent. These velocities are defined in mass transfer rate
coefficients (m/s): both films have different mass transfer rate coefficients, which can
be combined to one overall mass transfer coefficient:
(5.3)
in which
KI = overall mass transfer coefficient (m/s),
kl = partial mass transfer rate coefficient in the liquid phase (m/s),
kg = partial mass transfer rate coefficient in the gas phase (m/s),
mi = distribution coefficient (gas/liquid) of the target compound (-),
Under conditions in which Henry's law can be used mi can be replaced by H.

KI approaches k l for compounds with high Henry's law's coefficients. For these
compounds the diffusion through water is the rate-limiting step in interphase transfer.
KI is frequently multiplied with a second rate-determining factor in interphase
transfer rate, which is the specific interphase area 'a' (m 2/m3 column or reactor). The
term Kla comprises alI concentration independent factors that determine the
interphase transfer rate. Alternatively, Kg is used by absorption column designers,
which is also an overall mass transfer coefficient and related to KI by:
(5.4)
in which H is the Henry's law's coefficient expressed in (atm.m3/mol) and Kg is

expressed in (mol/m2 .s.atm).
The mass transfer from the gas phase to the water phase occurs at a rate F:
(5.5)
where 'a' is the specific interface area (m 2/m 3 ) and F is expressed in (kg/m3 .s).
However, in absorber column design a tradition exists to work with HTU

(height ofa transfer unit) and NTU (number of transfer units) (see Chapter 2). HTU
Bioscrubbers 139
is the height over which the gas phase concentration target compound is reduced with
a factor 'e', which has a value ofabout 2.7. In bioscrubbers HTU ranges between 0.2
and 2 m (Fischer, 1990). NTU is the number of sequential HTUs required to attain
the desired elimination efficiency. One HTU gives an efficiency of 57%, while 95%
requires three and 99% five HTUs.
1
NTU=ln-- (5.6)
E
1--
100
and
h = (NTU) (HTU) (5.7)
in which E is the elimination efficiency in percentage, and h is the height (m) of the
packing in the absorber.
Equation 5.6 is only true if the absorption water fed to the top of the column
is absolutely free of the compound to be absorbed, which is seldom found in
bioscrubbing. To account for the presence of the target compound in the lean water,
the following formula should be used:
(5.8)
in which
GM = superficial molar mass velocity of gas stream (moles/m2 .h),
LM = superficial molar mass velocity ofliquid stream (moles/m2.h),
X2 = mole fraction solute in liquid fed to top ofthe column,
Yl = mole fraction solute in gas fed to bottom ofthe column,
Y2 = mole fraction solute in gas leaving top ofthe column,
H = dimensionless Henry's law's coefficient (g m- 3 gas / g m-3 water).
Note that the group HGM/L M appears several times in Equation 5.8: it is the
stripping factor, the reciprocal absorption factor (Equation 5.2). Equation 5.8 is only
true if Henry's coefficient is constant over the whole concentration range involved
and requires that temperature does not change in the column (as it changes H).
An estimation ofthe value ofHTU can be made using formula 5.9:
HTU=~ (5.9)
KgaP
in which P is the total gas pressure ofthe system (atm).
140 J. W van Groenestijn
The most difficult point in practice is the value of Kg or of Kga. The valuc
depends on the nature of the compound, the properties of the gas and the liquid
(viscosity, composition, mixing, turbulence) and temperature. In addition, mixtures
of pollutants in gas make the design even more complex. Therefore pilot plant tests
cannot be avoided. They should give the basis for further scale up calculations using
the formulas described above. The given formulas can also be used to compare
absorbers, to estimate if existing absorbers can be approved and how they react on
changes in, among others, gas flow rate, liquid flow rate, gas concentrations, liquid
concentrations, height ofthe packing.
Formulas 5.1 to 5.9 were obtained from various literature sources (Ottengraf,
1986; Fischer, 1990; Riet and Tramper, 1991; Overcamp, 1993; Schippert, 1994;
VDIIDIN, 1996; Kohl and Nielsen, 1997) and adapted for this chapter.
3. Bioreactors
The aqueous effluent from absorbers is loadcd with absorbed pollutants and can be
regenerated in a bioreactor. In this regeneration process the pollutants are
biodegraded and converted into CO 2 , H20 and mineral products. As a resul! the water
can be reused in the absorber. Most bioreactors in bioscrubbers arc tanks with a
suspension of activated sludge and spargers for bubble aeration. The properties and
design of such bioreactors show great similarity with activated sludge tanks for
wastewater treatment. The most important difference is that the hydraulic retention
time is much longer, generally the same as the sludge retention time. The
consequence is that generally no measures are required for sludge retention, such as
sedimentation and sludge retum and addition of support material for biomass
attachment. Because of the low hydraulic loading rate, the concentration of the
activated sludge can be higher than that in activated sludge tanks in wastewater
treatment systems. Nevertheless, enhanced sludge retention in bioscrubber
bioreactors can be an option in case micro-organisms have to be retained with
extreme low growth rates (e.g. nitrifiers at low temperatures) or in case the water has
to be refreshed at a high rate because of accumulation of degradation products.
The following steps should be followed in the design ofbioreactors:
• Decide which concentration of the target compound in the lean solvent is desired
(connection to the design ofthe absorber);
• Estimate the growth rate ofthe micro-organisms involved;
• Estimate the specific substrate conversion rate;
• Calculate the required amount of biomass in the reactor;
• Calculate the required addition of nutrients;
• Calculate the required addition oftitrants;
• Calculate the rate of accumulation of reaction products, nutrients and titrants;
• Calculate the dilution rate ofthe reactor (water refreshment);
• Calculate the amount of oxygen required and design the aeration system;
• Calculate the volume ofthe bioreactoL
Bioscrubbers 141
Knowledge on the growth rate is required for the ca1culation of the minimum
sludge retention time and may also be used in ca1culation of the specific substrate
conversion rate. The growth rate is dependent on environmental factors such as
temperature, pH, ionic strength, presence of toxic compounds and the substrate
concentration. The dependence on the substrate concentration is expressed by the
Monod equation as also explained in previous chapters:
(5.10)
in which
).l = specific growth rate (g biomass/g biomass .h),
Ilmax = maximum specific growth rate (g biomass/g biomass .h),
CI = substrate concentration (g/m3),
Ks = Monod constant or substrate affinity constant (g/m3).
In a completely mixed tank reactor the substrate concentration CI in the

reactor is the same as that in the lean solvent sent to the top of the absorber.
The Ilmax of many micro-organisms growing on readily biodegradable organic
compounds at 20°C-30°C and optimum chemical medium composition, generally is
higher than 0.1 h· l . The growth rate on chlorinated hydrocarbons can be lower
(Dolfing el al., 1993), while nitrifiers have maximum specific growth rates near 0.01
h· l . The Ks is mostly around 1 mgl!. To prevent wash out of the active biomass, the
dilution rate (volume fresh water added/hour/total volume water in reactor) should
not be higher than the growth rate. The sludge production is lower than experienced
in biological waste water treatment plants (for the same organic loading rate), as the
inert fraction is missing. In addition, the sludge is generally less toxic than that from
domestic waste water treatment in which heavy metals are accumulated. If the ratio
of suspended organic matter and suspended dry matter is between 70% and 75%, the
sludge is in a good condition (VDI/DIN, 1996).
The calculation of the required amount of biomass in the reactor needs an
estimation of the specific substrate conversion rate: the amount of substrate
converted by 1 g of biomass per hour. Two ways for such an estimation can be
followed:
1. If Il is divided by the yield factor Y (g biomass dry weight produced/g substrate

consumed), the amount of substrate used for growth can be ca1culated. In addition
the amount of substrate per hour for the maintenance of 1 g dry biomass should be
added (this is the maintenance coefficient 'm'). According to Pirt (1985):
Total rate of rate of consumption + rate of consumption

consumption for growth for maintenance
The total rate of consumption, expressed per gram ofbiomass dry weight, is:
1!..+m (5.11)
y
142 J W. van Groenestijn
2. Use ofthe Michaelis-Menten equation, originally developed for enzymes, but also
true for whole cells:
v (5.12)
in which
V = specific substrate conversion rate (g substrate/g dry biomass .h),
V max = maximum specific substrate conversion rate (g substrate/g dry biomass .h),
Km = Michaelis-Menten constant (g/m\
Using the substrate conversion rate and the expected loading rate of the
pollutant(s), the required amount of biomass can be calculated. Note that the water
from the absorber sometimes can contain COD concentrations as high as 5000 mgl!.
In bioscrubber bioreactor design the kinetic parameters as used in Equations
5.10 and 5.11 are not always known and if the waste gas contains mixtures of
biodegradable compounds, calculation becomes too complex. Therefore design rules
of thumb are used, while the equations above can be used to estimate the effect of
changes in operation and design. Mostly the bioreactor is oversized to guarantee
stability and to cope with fluctuating loading rates. According to the VDI/DIN
guidelines (1996), the design should be based on the maximum half hours average of
the absorbed load. The concentration of biomass in bioscrubber bioreactors ranges
between 1 and 15 g dry weight (Schippert, 1994). Sludge concentrations higher than
15 g/Ilead to too much clogging ofthe absorber packing (VDIIDIN, 1996). Ottengraf
(1986) recommends evcn lower sludge conccntrations: not more than 5-8 g dry
weightl!. The volumetric conversion rate of bioscrubber bioreactors is 20-200 g/m 3.h
(Schippert, 1994) and maximum 400 g/m 3 .h for readily biodegradable compounds
(Fischer, 1990; VDI/DIN, 1996), while the sludge loading rate is 0.005-0.006 g
substrate/g biomass dry weight . h. The yield factor observed in practice ranges
between 0.1 and 0.4 g dry sludge/g substrate. The bioreactor can be started up by
inoculation with activated sludge from waste water treatment plants.
Because of the completely mixed character of the liquid in the bioreactor, the
concentrations ofthe biodegradable compounds are low in the entire system, which is
an advantage in case these compounds are toxic (e.g., aromatic compounds and
formaldehyde). In biofilters, however, high (toxic) concentrations may exist at the
biofilm surface.
The growth of biomass requires nutrients. The amount can be estimated from
the sludge production rate (10) and element composition of biomass. According to
Stanier et al. (1978) the average microbial cell contains (based on dry biomass) 50%
C, 20% 0,14% N, 8% H, 3% P, 1% S, 1% K, 1% Na, 0.5% Ca, 0.5% Mg, 0.5% CI,
0.2% Fe, othcr elements 0.3%. Tracc clemcnts generally are present in the waste gas,
the used water and in the other raw materials, a part of the macro elements may also
be present in the gas (N and S compounds, dust) or in the used tap water (Ca and
Mg). Often, P is thc limiting factor and N and Parc added as the only nutrients, but
care should be taken not to create element limitation for growth.
Bioscrubbers 143
The type of suspended biomass in the bioreactor is more sensitive to periods

without substrate than the biomass in biofilters. If the duration of the intermission is
more than two days, artificial addition of substrate is recommended to maintain a
high microbial activity (VDVDIN, 1996). It is advisable to continue aeration during
intermissions to keep the aerobic microorganisms alive.
The addition of nutrients, titrants (to compensate for inorganic acid
production from N, S and CI compounds), compounds from fresh water and
inorganic reaction products will increase the ionic strength of the water phase. To
prevent inhibition ofmicrobial growth and/or activity, the water is slowly wasted and
replaced by fresh water. Measurement of the electric conductivity gives a good
indication ofthe ionic strength. A maximum of 5 mS/cm is recommended (VDI/DIN,
1996) or lOg dissolved inorganics/l. The rate of accumulation (weight/h) can be
estimated from the biological reaction equations and the planned rate of nutrient
addition.
A second reason to waste water is the accumulation of sludge. The sludge
production (g/h) is the sludge concentration (gll) multiplied by the waste water flow
(1/h). In rare cases produced or absorbed non-biodegradable organic compounds can
also accumulate, even up to toxic levels, thus creating another reason for water
refreshment. The residence time of the water mostly is not longer than 20 to 40 days
(VDVDIN, 1996). Schippert (1994) recommends a waste water discharge of 0.2% of
the tank volume per hour. Fresh water addition compensates for waste water
discharge and evaporation losses.
In contrast with aeration tanks in waste water treatment, the sludge in
bioscrubbers forms much smaller flocs. The shear forces in the pump and the
absorber do not allow a large floc size. As a consequence, separation of sludge and
water by natural sedimentation or by flotation is difficult. As a result still 5000 mg
sludge COD/l can be found in the effluent. For sludge retention, if required, and
further sludge processing filtration and centrifugation are used. Mostly, the sludge
can be discharged in a sewer system together with the waste water.
The addition of oxygen in bioscrubber bioreactors generally proceeds via
bubble aeration. From the waste gas flow rate, gas composition and the biological
reaction equations or waste gas COD, the required amount of oxygen can be
calculated. This amount generally is much higher than the amount absorbed in the
absorber. A minimum concentration of 1 mg dissolved oxygen is advised for
sufficient biological activity (Fischer, 1990). The required airflow rates and aeration
equipment can be derived from experiences in biological waste water design
(Tchobanoglous and Burton, 1991; Koot, 1980). Note that in bioreactors with a water
height of 3 m only about 10% of the oxygen in the used air is absorbed in the water
phase (Koot, 1980). The efficiency of aeration is generally more than 1 kg O2
transferredJkWh air blower energy consumed. The aeration tank should have a free
board of 0.5-1 m to anticipate on foaming, scrubber drainage and turbulence. To
prevent sludge sedimentation in the aeration tank, the liquid flow rate in the reactor
should be at least 0.15 mls and preferably 0.3 mls (Ottengraf, 1986), which can be
attained by bubble aeration. In a large part of the existing bioscrubbers the absorber is
placed on top ofthe bioreactor, allowing the water flowing to this reactor by gravity.
The volume of the aeration tank is normally larger than that of the absorber.
In VDVDIN (1996) a list of design values of 14 German bioscrubbers can be found.
The volumes of the bioreactors range from 7 to 1000 m3 and the ratio between gas
flow rate (m3/h) aud bioreactor volume (m3) ranges between 100 and 3000, with
1000 as a balauced average. Compared with biofilters, which are mostly loaded with
100 m3 gas/m3 filter bed .h, the volume of bioscrubbers is almost an order of
magnitude lower thau the volume ofbiofilters.
The pump between the bioreactor and the absorber should be a positive
displacement pump or a slow running centrifugal pump, to minimise destruction of
sludge flocs. The pumps should be cOITosion resistant aud their shaft seal should be
able to withstaud solids. A dry running protection is recommended.
4. VOC removal: cases
4.1 EXPERlENCES IN THE NINETEEN EIGHTIES
In the early nineteen eighties bioscrubbers were installed to treat gases from
enamelling ovens (alcohols, glycols, ketones, glycolether, aromatic compounds aud
resins), incinerators, foundries (amines, phenol, formaldehyde aud ammonia) aud fat
smelteries (Ottengraf, 1987).
One of the ear1y comprehensive reports on a successfu1 bioscrubber
application was that from Schippert (1986), who described how waste gases from a
paint booth in Ma1mo were treated. The factory emitted 35000 to 60000 Bm3/h with
30-45 kg VOC aud 2 kg resin aerosol per hour at a temperature of 160 aC. The gas
first pas sed aventuri scrubber for cooling and remova1 ofthe aerosol. In this scrubber
1500 kg water evaporated aud the heat (200 kW) was recovered. Subsequently two
scrubbers with a diameter of 4 m aud a height of 6 m were used aud two activated
sludge tanks with a total added volume of 280 m3 for the regeneration of the
absorption 1iquid. Sludge retum was mauaged via flotation aud waste water was
produced at a rate of 0.5 m3/h. Nutrients were added to maintain a high microbial
activity: 0.15 kg Na2HP04, 0.6 kg (NH4)2S04 aud 0.1 autifoam agent per hour. 180
kW energy was required at full capacity aud minor amounts of NaOH as a titraut.
99% remova1 ofVOCs was observed from a mixture ofbuty1 glycol, n-butanol, ethyl
glycol, isobutanol, xylene aud MIBK (methyl isobutyl ketone) in air. Xylene aud
MIBK showed the poorest elimination efficiencies (70%) but their concentration in
the influent gas was only 10w.
After three years the author reported again about this bioscrubber. The
e1imination efficiency still was more than 99%, of which 99% was attained in the
first stage. Therefore, it was conc1uded that the bioscrubber was over-dimensioned.
Another problem was the attachment of biomass on the scrubber packing. It was
observed that this effect coincided with the addition of excess amounts of nitrogen in
the second stage. The mechanism behind this effect was not studied further
(Schippert, 1989a).
Another early report of a bioscrubber treating gases from a paint booth was
from Beyreitz (1989). A biofilter aud a bioscrubber were 10aded simultaueously with
a gas containing acetone, methy1 ethyl ketone, toluene, butyl acetate, 2-propauol aud
di-acetone alcoho1, to compare their performance. The biofilter had a loading rate of
70 m 3 /m 2/h, the bioscrubber had a higher surface loading rate because of its smaller
size. The biofilter removed 90% ofthe VOCs present, ofwhich toluene was the most
problematic with only 51 % elimination. The bioscrubber removed 80% VOCs of
Bioscrubbers 145
which 0% toluene. The conclusion was that bioscrubbers are attractive for removal of
hydrophilic compounds, because of its reduced size, but biofilters still are needed for
removal of hydrophobic compounds. The author suggested that a combination of a
bioscrubber and a biofilter might save space, which could be an important
consideration in an industrial environment.
An example of a bioscrubber at a foundry can be found in a paper written by
Buren (1989). The bioscrubber treated 36000 Bm3/h waste gas containing phenol,
formaldehyde and ammonia. The bioscrubber comprised an absorber column,
followed by a clarifier (a plate settler for removal of dust) and a bioreactor. The
bioreactor was a hydrolift reactor in which air and water were introduced at the
bottom section and activated carbon particIes with aUached biofilm were fluidised.
The advantage of the use of activated carbon as a support material was the buffering
of concentration fluctuations. The bioscrubber reduced the phenol concentration from
50 to less than 3 ppmv, the formaldehyde concentration from 5 to less than 0.5 ppmv,
while ammonia (30 ppmv) was removed with an efficiency of at least 50%.
More bioscrubbers for the removal of phenol and formaldehyde are described
by Huckschlag (1992). Pilot plant investigations have been carried out at the sites of
wood factories, textile plants and glass wool manufacturers between 1985 and 1991.
Phenol always could be removed at least 97% and formaldehyde at least 85%.
4.2 EXPERIENCES IN THE NINETEEN NINETIES
Interesting details on the design and operation of a state of the art bioscrubber can be
found in Kellner and Flauger (1998). The bioscrubber cleans waste gases from a
coating process that contains a mixture of hydrophilic VOCs of which 84% is
acetone. The air flow rate is 20000 m3/h, the acetone concentration 800-2000 mg/m3
and the temperature 22°C. The bioscrubber comprises a packed bed co-current
scrubber and an activated sludge tank. The scrubber has a column diameter of 2 m
and a packing height of 2 m, while the volume of the activated sludge tank is 140 m 3 .
The activated sludge system also comprises a sludge sedimentation tank and sludge
recycling. Fertilizers are added to support the growth of micro-organisms: 24 kg/day
during starting up and 50 kg/week during steady state operation. The biomass
concentration in the activated sludge tank is 2-3 g dry weightll and 5 m3 of the
recirculation liquid is discharged in the sewer each day.
At the start of the operation in 1995 a liquid loading rate of 10 m3/m2 .h was
used. This appeared to be too low: the packing was clogged with biomass and the gas
pressure drop increased up to 1800 Pa. Pilot plant studies carried out by Loy et al.
(1998) revealed that a lower packing height and a higher liquid loading rate (28
m3/m2.h) prevented clogging of the absorber packing. These insights were
implemented on full scale: the packing height was reduced from 2 m to 1.5 m and the
liquid loading rate was increased to 24 m3/m2 .h. From that moment a high cleaning
efficiency could be obtained without any problems of clogging.
Bioscrubbers are very suitable for treating gases in which organic nitrogen
compounds are present. The production of ammonia as a result of biodegradation and
subsequent nitrate accumulation by nitrification may yield problems in biofilters. In
bioscrubbers, however, control of the pH and concentration of inorganics is very
easy. A good example is the bioscrubber described by Hansen and Rindel (1992) for
the treatment ofwaste gases from a fish feed production factory. 25000 m3 gas/h was
treated using a packed absorption column of 14 m3 and an aeration tank with a

volume of 28 m3 . Titrants were added for pH control, water drainage took place to
remove accumulated matter, nutrients were added to support bacterial growth and
fresh water to compensate for evaporation. The water recirculation between the two
process units was 250 m3/h. The bioscrubber removed 95 % ofthe odour and 80 % of
the TOC. For this type of off-gas it is interesting to have information on
accumulation of biodegradation products in the scrubber liquid. Maximum 20 g
salts/I were alIowed to accumulate, of which the largest part was nitrate. Not more
than 0.1 g ammonia/I was accumulated, indicating a biological nitrification process.
In addition, 100-1000 mg COD and 3 g biomass dry weight were present in a litre
scrubbing liquid. The authors have calculated a yie1d factor of 0.07 g biomass dry
weight per g of COD removed. The data indicate that the concentration inorganics
can be controlled by dilution, while still maintaining sufficient microbial activity.
Bioscrubbers can be combined with other treatment technologies. Already
Beyreitz (1989) proposed a sequence of a bioscrubber and a biofilter for gases
containing mixtures of hydrophilic and hydrophobic compounds. Hansen (1998) used
a sequence of a bioscrubber and an adsorption filter for the same reasons. Frohlich
(1994) combined chemical scrubbers and bioscrubbers for off-gases containing
mixtures of ammonia and VOCs. 41000 m3fh gas from a cigarette factory was treated
in a chemi cal scrubber with liquid using a pH of 3 and a bioscrubber in which the
liquid had a pH of 6. The odour units decreased from 2000 to less than 400 and the
ammonia concentration from 35 to less than 1 mg/m3 According to Frohlich
chemi cal pre-treatment yields a higher reliability.
The suggestion of Beyreitz was followed up by Standefer and WilIingham
(1998) with a realisation in practice. A combination of a bioscrubber and a biofilter
was used to treat 60000 m3 gasfh containing organic nitrogen compounds, glycols,
alcohols and aldehydes. The absorber of the bioscrubber had a gas residence time of
less than 1 second, but was able to remove 92% of the VOCs loaded. This was
possible because of the readily solubility of the compounds in water. Nevertheless,
the efficiency of 92% was not sufficient and one of the compounds showed
fluctuating absorption efficiencies (70-95%). The biofilter polishing step was
required to reach a guaranteed elimination efficiency of more than 95%. In fact, the
bioscrubber reduced the load and size of the biofilter. The total gas residence time
was not more than only 20 seconds.
The presence of a biological waste water treatment plant according to the
activated sludge process on the same site as the bioscrubber can yie1d interesting
advantages. The bioreactor of the bioscrubber can be connected to the activated
sludge tank and liquid can be exchanged. The advantages are:
• Generally, no additional nutrients required in the bioscrubbers;

• Better coping with waste gas loading intermissions;
• Fast restart after bioreactor deactivation.
A prerequisite is that the compounds to be treated in the waste gas are the
same as or show sufficient similarity with compounds introduced in the waste water
treatment plant.
Bioscrubbers 147
Diehl and Schiifer-Treffenfeldt (1997) described two applications of this

principle. 120000 m 3 gas/h with 300 volatile compounds, under which organo-
sulphur compounds, was treated in a bioscrubber comprising an absorber with a 200
m 3 liquid/h recirculation rate. 25% ofthe flow rate mentioned was recirculated over a
bioreactor and 5% was exchanged with an activated sludge tank. Almost no chemical
additions were required, apart from very low amounts of titrants to maintain the pH
between 6.5 and 7.5. The influent gas contained 4000-22000 odour units, of which
80% was removed in the bioscrubber. The residual odour had an acceptable earth-
like character.
A second application can be found at a plastics factory at which 10000 m 3!h
with volatile plastic monomers (Henry's law's coefficient = 0.015) was treated in a
bioscrubber with a bioreactor of 100 m 3 volume. 10-20 m3 liquid/h was exchanged
with an activated sludge tank at the same factory site. More than 80% VOC removal
was obtained (Diehl and Schiifer, 1997).
Although bioscrubbers are supposed to loose their cost-efficiency for
compounds with a dimensionless Henry's law's coefficient higher than 0.01,
exceptions have been reported. Parkinson (1996) described a bioscrubber for the
elimination of styrene from waste gases from a factory for automobile parts. The
bioscrubber was able to reduce the styrene concentration from 400 to 5 ppmv, while
the Henry's law' s coefficient is 0.1. According to the author the operational costs are
20% and the capital costs are 40% lower than biofilters for styrene removal. A reason
may be that removal of high concentrations of styrene by biofilters needs very large
filter bed volumes.
Bioscrubbers are applied to treat odour from waste water treatment plants. In
sewage treatment plants odour is produced everywhere at which anaerobic conditions
appear in combination with open water surfaces, e.g., in the influent pit, primary
sedimentation tank, sludge thickener and sludge dewatering equipment. Nurul Islam
el al. (1998) determined the 5 most important odour compounds at sewage treatment
plants: H2S, dimethyl sulphide, methyl mercaptan, NH3 and trimethyl amine, of
which the latter contributes most to the odour.
The treatment of these gases is difficult. Treatment using biofilters yields
problems with acidification resulting from bioconversion of nitrogen and sulphur
compounds, while bioscrubbers show limited absorption efficiencies for the
relatively hydrophobic organo-sulphur compounds. Hansen (1998) described a
bioscrubber with polishing filter with adsorbent for application at a waste water
treatment plant. As most of the pollutants were removed in the bioscrubber, the
adsorbent had a long life expectancy, thereby ensuring economic al operation. The
bioscrubber c1eaned 6000 m 3!h gas from the inlet structure of a waste water treatment
plant (primary pumping station, screen and aerated grit chamber). Only phosphate
was used as a nutrient and caustic soda was used as a titrant to keep the pH at 8.5-9.
At this pH H2S is absorbed well, while the absorption of C02 is limited. The inlet
concentration H2S was average 10 mg!m 3 and maximum 75 mg!m 3 and the
elimination efficiency more than 99%. The concentration of organo-sulphur
compounds in the outlet gas was lower than 0.1 mg!m3 .
Many other examples of bioscrubber applications are given in VDVUIN
(1996). A summary is given in Table 5.2.
Table 5.2. Applications of bioscrubbers at various industries (adapted from VDI/D1N, 1996)
Industry Waste gas compounds Waste

gas flow
rate
(m%)
Mineral fibre/metal gaskets Odour, aerosols 9000
Orinks can painting (steel) Alcohols, glycols, toluene 57000
Orinks can painting (aluminium) Alcohols, glycols 48000
Pasteboard packages for drinks Ethanol, ethyl acetate, I-methoxy propane-2-ol 150000
Oetergents and cosmetics Alcohols, acrylate monomers, odorants,
toluene 30000
Tobacco industry Odour, dust 45000
Tobacco industry Ammonia, nicotine, odour 35000
Animal rendering Ammonia, amines, sulphide 30000
Animal rendering Odour 50000
Grinding wheel production Phenol, amlllonia, amines, forlllaldehyde 28500
Conditioning of paint residues Ethyl acetate, acetone, MEK, aromatics,
alkanes 500
Industrial waste water treatlllent Odour 500
Foodstuffs production Allllllonia, sulphide, odour 750
Sewage treatlllent Organo-sulphur compounds, pyrazines 670
Aluminium foundry Phenol 120000
Grease melting Odour 15000
Grey cast iron foundry Phenol, ammonia 13000
Chipboard production Forlllaldehyde 500
Glass wool production Phenol, forlllaldehyde 400
Molding pl. for lining elements Phenol, forlllaldehyde, ammonia 350
Printing Alcohols, esters 60000
5. VOC removal: developments
5.1 THERMOPHlLIC BIOSCRUBBING
Most probably the largest bioscrubber installed up to now is a thermophilic one

(Anonymous, 1999). In 1999, in Austria, a bioscrubber was installed to treat 400000
m3 gas/h from a wood plate factory. The hot gas contained formaldehyde, organic
acids and wood particles and was contacted in a scrubber in which a mist of water
was sprayed in upward direction. No packing was present. The mist drops with
absorbed pollutants subsequently entered a large vessel with a demister in the top
section and an aerated sump in the bottom section. The sump was a suspension of
thermophilic micro-organisms and the place where the pollutants were biodegraded.
5.2 BroSCRUBBERS WITH TWO LIQUID PHASES
Conventional bioscrubbers can only be used for compounds which are readily soluble
in water. Thc !imit is set by a Henry's law's coefficient of 0.01. The addition of an
organic solvent to the water phase, however, can make bioscrubbers fit for
Bioscrubbers 149
elimination of more hydrophobic compounds, even alkanes. The addition of 10-30%

water immiscible solvent with a high boiling point to the liquid phase enables the
absorption of hydrophobic compounds from the gas phase in the absorber (Schippert,
1989). The waterlsolvent emulsion with the absorbed pollutants mainly in the organic
phase are subsequent1y transported to a bioreactor in which microorganisms first
degrades the pollutant dissolved in the aqueous phase. The lower pollutant
concentration in aqueous phase resulting from this process forms a driving force for
mass transfer from the organic phase to the aqueous phase. Thus, complete biological
regeneration of the liquid is possible. It is hypothesised that microorganisms may
accumulate at the aqueous/organic interphase, which may enhance the regeneration
process.
The organic solvent should be selected carefully. According to Schippert
(l989b) and Cesario et al. (1992) the organic solvent has to meet the following
conditions:
• High solvent capacity;

• Immiscible with water;
• Low solubility in water;
• Inert to biodegradation;
• Not toxic for biocatalysts;
• Low vapour prcssure;
• Relatively low viscosity;
• Density different from the density of water;
• Odourless;
• Favourable price.
Silicon oii (poly-dimethyl siloxane), di-n-octylphtalate and di-n-nonylphtalate

are good candidates. Alkanes and certain aromatic compounds show a 100 to 1000
better solubility in these solvents compared with water.
In 1989 Schippert did pilot plant experiments with the so-called 'Biosolv
Verfahren'. The bioscrubber absorber was loaded with 6000 m3 gas/m 2 h containing
styrene. Thc liquid contained 13% solvent and was loaded in the absorber at a rate of
25 m3/m 2 .h. Under the absorber a bioreactor was placed in form of an aerated tank.
90% styrene could be removed from the gas. Poppe and Schippert (1992)
demonstrated the new technique by treatment of a mixture of 13 volatile compounds
in air by a two stage bioscrubber. The first bioscrubber was conventional, while the
second contained the organic solvent. Each stage included an aerobic biological
regeneration reactor and had a volume of 0.8 m3 . The gas flow rate was 200-600
m3/h. In the first stage mainly hydrophilic compounds such as ethyl acetate, butyl
acetate and acetone were removed, while hydrophobic compounds pas sed
completely. They were removed in the second stage only. Positive results were
obtained with toluene, ethyl benzene and xylene. By this treatment, the xylene
concentration decreased from 16 mg to 2 mg/m 3 gas. Based on these results a few
full-scale bioscrubbers were installed in Germany.
Deziel et al. (1999) reviewed the state of the art in the use of two-liquid-phase
media for degradation of hydrophobic or toxic compounds by microorganisms. A
tradition of more than 35 year exists. Substrates such as alkanes, benzene, styrene,
phenol, naphthalene and pentachlorophenol have been tested and solvents as pristane,
dibutyl phtalate, hexadecane, dodecane, heptamethyl nonane, silicon oiI and paraffin
oiI were used. Not all solvents meet the demands given above. In continuous
degradation experiments with long durations, ali kinds of negative effects can appear.
Gardin et al. (1999) further selected suitable solvents. Based on known physical
properties pristane, isopar V, isopar L, perfluoro hydrocarbons and silicon oiI were
preselected. However, biological experiments revealed that pristane, isopar V and L
were not inert: they were degraded by microorganisms, while perfluoro hydrocarbons
were toxic. Silicon oii seems the best choice up to now. It was proven that the
presence of silicon oii in a aqueous medium could greatly enhance the growth of
micro-organisms on xylene as a substrate. Besides the improvement of
bioavailability, solvents can reduce the toxicity of compounds (by decreasing the
concentration in the water phase) and can act as a buffer system for fluctuating loads
of hydrophobic or toxic substrates (Deziel et al., 1999; Gardin et al., 1999).
According to Deziel et al. (1999) it is important to create an aqueous/solvent
interphase area as large as possible. An optimal phase ratio exists that will generate
the highest interphase area. This ratio depends on the physical conformation of the
bioreactor, properties of the solvent and the mixing rate. 2~O% solvent in water
gives the highest areas. A higher oii phase ratio may lead to larger drop sizes (Deziel
et al., 1999) or even phase inversion (Groenestijn and Lake, 1999). A small drop size
may be created by emulsifiers, inc1uding the natural emulsifiers excreted by the
microorganisms in the system.
As with aqueous bioscrubbers, water and sludge in two-liquid-phase
bioscrubbers have to be discharged after some time of operation. In running full-scale
systems this is achieved by separation ofthe two phases ofthe waste liquid stream by
centrifugation, followed by recyc1ing of the expensive organic phase and discharge of
the aqueous sludge suspension. Not more than 3% of the organic solvent is lost
annually (Keramchemie, personal communication), which in addition limits
environmental damage.
Separation of the aqueous and organic phase plays an important role in the
bioscrubber deve10ped by Cesario and co-workers (1992). They propose a
combination of a spray tower and a liquid-impelled loop reactor. In this system a
water-immiscible organic liquid with a high solvent capacity for the pollutant is
recyc1ed between the absorber and the bioreactor. The latter is designed as a liquid-
impelled loop reactor, a liquid/liquid contactor which contains the culture medium
and the cells (Tramper et al., 1987). This reactor comprises a riser and a down-
comer. The organic solvent is introduced at the bottom ofthe riser and causes the low
density water/solvent mixture to rise. At the top of the reactor the two phases are
separated. The organic phase is recirculated via the spray scrubber to the bottom of
the riser, while the aqueous phase is recirculated via the downcomer. Without
bacteria the system works perfectly, however, phase separation becomes difficult
when free suspended bacteria are present (Cesărio and Tramper, personal
communication).
5.3 ANAEROBIC BIOSCRUBBERS
In conventional bioscrubbers aerated bioreactors are used to biodegrade the

pollutants aerobically. However, certain pollutants can only be degraded under
Bioscrubbers 151
anaerobic conditions. Parker et al. (1998) used scrubbers and anaerobic reactors of
the UASB type to degrade perchloroethylene. As chlorinated hydrocarbons are
relatively insoluble in water, an organic solvent was added to the scrubbing liquid.
One of the organic solvents tested was vegetable oiI. In addition, anaerobic
bioscrubbers are used for elimination ofNOx and SOx (see NOx and SOx sections).
5.4 COMETABOLIC BIOSCRUBBING
Some compounds are aerobically biodegradable only as a result of cometabolism

when growth takes place on a different substrate. Hecht et al. (1995) developed a
cometabolic bioscrubber for the eIimination of trichloroethylene from gases. The
gases were lead through a bioreactor containing Pseudomonas cepacia in form of
bubbles. Phenol was added as the growth substrate. Although this reactor was not a
cIassical bioscrubber with separated absorber and bioreactor, the idea may be
valuable for consideration in conventional configurations.
5.5 FOAMS
Foam is an alternative way to contact gas and a suspension of microorganisms. In

Germany laboratory experiments were carried out with biological foams and waste
gases (Anonymous, 1998). Toluene was used as a modei compound. The researchers
observed that the bacteria sometimes accumulated at the gas-liquid surface, which is
expected to enhance the absorption process. In the USA, experiments were carried
out with a laboratory scale biologically activated foam reactor (Phipps and Ridgway,
1995). In this reactor a nutrient solution containing a surfactant and microorganisms
was mixed with gas in a static mixer. Foam was produced with gas bubble diameters
of 0.1-1 mm, having a specific surface area of 8000-80000 m 2/m 3 . In a column, the
foam is transported to a foam breaker (spray nozzle) and the liquid is recycIed.
Experiments were carried out using benzene and toluene at loading rates of 0.3
g/m3 .h. About half of the load was eliminated, which implies a very low volumetric
elimination capacity compared with conventional bioscrubbers and biofilters. The
technology needs further optimisation.
6. H 2 S and SOx removal
6.1 H2 S REMOVAL FROM AEROBIC GASES
Bioscrubbers as developed by Hansen (1998) can remove more than 99% of H2S
from aerobic gases. In the bioreactors the sulphide is converted to sulphate, a reaction
that needs alkali addition for pH control and frequent water refreshment to control
salt concentration. Alternatively, the absorbed sulphide can be biologically converted
to eIemental sulphur. This incomplete oxidation can be carried out by bacteria from
the genus Thiobaci/lus if subjected to oxygen limited, but aerated, conditions. Such
system is described by Janssen et al. (1997). A pilot scale bioscrubber was used for
treatment of 160 m 3 sour gas per hour from a refinery. The influent concentration
contained 70-500 ppmv H2S and the effluent concentration O ppmv. This way less
caustic soda is required for pH control as the alkalinity can be recycled over the
bioreactor and the absorber.
6.2 HzS REMOV AL FROM ANAEROBIC GASES
Natural gas, biogas, synthesis gas and Claus tail gas mostly contain HzS. In large
scale (> 15 tons S/day) gas treatment, amine absorbers and Claus plants can be used
to remove HzS. For smaUer quantities, liquid redox systems, based on reaction with
iron chelates, are used. Bioscrubbers may be an alternative for these redox systems. A
system similar to HzS removal from aerobic gases can be applied to clean anaerobic
gases. According to Janssen et al. (l999a; 2000) the technology is competitive in the
0.1-15 tons S/day range. The reactions in the absorber and the bioreactor are
respectively:
An early report is given by Dijkman (1995) in which a bioscrubber for 400 m3

biogas/h is described. The used absorber was a packed spray tower and the bioreactor
was a 72 m3 submerged fixed film reactor. The bioscrubber removed more than 99%
of the HzS introduced (10000-15000 ppmv in influent gas and 20-120 ppmv in
effluent gas). According to the author, compared with a caustic soda scrubber, 90%
ofthe amount ofNaOH required can be saved ifa bioscrubber is used. According to
Janssen el al. (2000) 5 fuU scale bioscrubbers for biogas treatment were installed
worldwide.
Besides biogas, high-pressure natural gas can be treated in a similar way. Pilot
plant experiments have been carried out with this type of gas. The pilot plant
comprised an absorption column, operating at pressures between 5 and 53 Bar, a
flash vessel, a 0.4 m3 aerated bioreactor, operating under atmospheric prcs:-,urc and CI
plate settler for the separation of sulphur and water. In the bioreactor the pH, redox
potential, and solution conductivity were controlled. The control of the redox
potential is required, since the bacteria involved have the tendency to form sulphate
at too high redox potentials. In this way, less than 3.5% of the sulphide is converted
to sulphate in these types of bioreactors. Since the bacteria are recycled over the high
pressure absorber, they experience high pressure differences. Nevertheless, no
negative effects of these pressure variations on the biological activity could be
observed. The bacteria were able to oxidise almost ali sulphide loaded: the rich
solvent (from the absorber) contained 100-1000 mg sulphide/l and the lean solvent
less than 20 mg/l. Only small amounts of C02 (10%) were absorbed in the absorber
to compensate for stripping of COz in the aerobic reactor and loss of NaHC0 3 in the
bleed stream. The sulphur slurry can be concentrated in a decanter centrifuge,
yielding a sulphur cake with 40% water and dry matter with 95-99% SO (Janssen el
al., 1999b). Various options exist to make sulphur of different purities and reuse it in
agriculture or sulphuric acid manufacturing. The demonstration plant was able to
c1ean natural gas containing up to 8% H2 S with a near 100% efficiency in a 100 days
experiment (Janssen et al., 2000). For high sulphur loading rates gas lift bioreactors
may be used.
Bioscrubbers 153
A different approach for the treatment of biogas was followed by Nishimura

and Yoda (1997). Biogas is mostly produced on a site on which both anaerobic and
aerobic biological reactors for waste water treatment are present. The aerobic reactors
serve as post-treatment and remove ammonia, residual COD and sulphide from the
anaerobic effluent. These aerobic reactors that convert sulphide into sulphate, can
also be used as bioreactors in a bioscrubber. This was demonstrated by the authors:
40 m 3 biogas/h was treated in a 3 m 3 absorber with 13 bubble trays over which 20 m 3
activated sludge sus pension was recirculated. The HzS concentration in the gas
decreased from 2000 to 20 ppmv.
A third system was developed by Pagella el al. (1996), using Fe(III) solutions
to absorb HzS and convert it to elemental sulphur, and bioreactors that convert Fe(II)
into Fe(III). The absorber liquid conta ins Fe(III) at a low pH. A pH lower than 3 is
required to prevent precipitation of ferric hydrates. At this low pH, no chelating
agents are required. In the absorber the following immediate reaction takes place:
To regenerate the absorber liquid, the suspended e1emental sulphur should be

removed and the produced Fe 2+ should be oxidised to Fe 3+ again. However, this
oxidation cannot take place by aeration or by clectrochemical reactions, only by
biological oxidation by bacteria as Thiobacillusferrooxidans. The biological reaction
is:
2 Fe z+ + HzO + 1/2 Oz ~ 2 Fe 3+ + 2 OR
Pagella and co-workers revealed that a pH as low as 1.5 was acceptable for
this reaction. A pH of I appeared lethal. The system was further worked out by
Pagella and De Faveri (2000) in a laboratory scale test using a bubble column
absorber with a: gas residence time of 1.5-2.3 s and a tixed bed bioreactor. The
optimum pH for growth of the bacteria used to oxidise the ferrous ions appeared to
be 2.2.
6.3 SOx REMOV AL FROM FLUE GASES
SOx can be removed from flue gases by scrubbing with dilute solutions of caustic
soda or limestone. However, in these processes the costs for chemi caIs are
experienced as high and products as disodium sulphate can hardly be disposed
(Cetinkaya el al., 2000). A biological alternative was developed (Buisman el al.,
1994; Cetinkaya el al., 2000; Janssen el al., 2000). In this bioscrubber the hot gases
pass an absorber in form of a reverse jet wet scrubber in which particulates and SOx
are absorbed. The most important reactions are:
S02 + NaHC03 ~ NaHS0 3 + COz
The latter reaction concerns only a part of the NaHS0 3 formed in the tirst
reaction. The liquid from the absorber is transferred to an anaerobic reactor in which
the sulphites and sulphates are biologically converted to sulphide. For this process an
electron donor is required. For large scale applications H2 gas is preferred, while in
small-scale plants also methanol and ethanol may be used (H2 generation on small
scale is relatively costly). The reactions involved are:
Subsequently two possibilities exist to further process the sulphides produced:

(1) formed H 2S gas can be treated in existing amine absorbers, or (2) aerobic
biological treatment of the effluent containing sulphide. The conversion of sulphide
to sulphur is described in the section above. A simplified flow sheet of the process
can be found in Figure 5.2.
Feed Treated Gas Spent

gas gas bleed air
Makeup
NaOH Anaerobic Aerobic
reactor reactor
Particulates H2 Air Sulphur
Figure 5.2. Flow sheet of a bioscrubber for SOx removal from flue gases (adapted from
Cetinkaya el al., 2000).
Pilot plant experiments have been carried out at a 600 MW power plant that
produced 2 mi/lion m 3 flue gas per hour. 6000 m3/h gas with a temperature of 120 °C
was used for the experiments. The plant comprised a 6 m high absorber tower and a
10 m high anaerobic 'Internal Circulation' reactor, in which gas circulated by an
external compressor for a good mixing. The aerobic reactor was an airlift loop
reactor. The water recycle flow was varied from O to 6 m 3/h and the water bleed from
O to 500 IIh. The water from the absorber had a pH of 7-8. Not higher, because that
led to pH problems in the aerobic bioreactor in which an alkaline reaction took place.
The first experiments were carried out using a thermophilic anaerobic reactor (50 0c)
and ethanol as electron donor. The start-up took six weeks. The second reactor
started with the production mainly of sulphate, but after increasing sulphide loading
Bioscrubbers 155
rates, mainly sulphur was produced. The plant was loaded with 6 kg S02Jh and the
SOx removal efficiency was 98% (Buisman et al., 1994; Janssen et al., 2000).
7. NO x removal
A bioscrubber for the removal ofNO x from flue gases is described by Cetinkaya and
co-workers (2000). The problem connected to NO x removal from flue gases is that
95% of the NO x is NO, a compound poor1y soluble in water. To overcome this
solubility problem Fe(II)EDTA is used in the scrubber liquid to react with NO. As a
result, Fe(II)[EDTA]N0 2-, a nitrosyl complex, is forrned and NO is absorbed from
the gas at a high rate, at any temperature. The scrubber liquid containing the absorbed
N0 2 and the nitrosyl complex is subsequently transferred to an anoxic bioreactor in
which biological denitrification takes place. The nitrogen compounds mentioned are
reduced to dinitrogen gas using an electron donor, e.g., ethanol. The Fe(II)EDT A in
the liquid can be reused again in the absorber. Removal efficiencies of more than
80% can be achieved.
8. NH3 removal
The reactions involved in NH 3 eIimination in bioscrubbers are:
NH 3 + H20 -7 NH/ + OR (absorption and dissociation),
NH/ + 2 O2 -7 N03- + 2 H+ + H20 (biological nitrification),
N0 3- + 5 (H) -7 Y2 N 2 + OR + 2 H20 (biological denitrification; anoxic

conditions required).
In the latter reaction (H) represents an electron donor (reductor). This can be
an organic compound, sulphide or H2. Note that the overall reaction is not acidic nor
alkaline. Therefore, a denitrification step in the bioscrubber is attractive because of
the better pH control and because the water can be used a longer time, since this way
no inorganic compounds are accumulated.
Nitrifying micro-organisms have low growth rates. The generation times
range from 3 to 10 days depending on environmental conditions. Since the sludge age
of bioscrubber bioreactors can be very high, nitrification is possible. Ammonia
removal by bioscrubbers has been reported manifold. An example can be found in
Hansen and Rindel (1992): a bioscrubber for the treatment of gases from a fish feed
factory. Nitrate accumulated in the bioreactor liquid, because no denitrification step
was incorporated.
Hvidtfeldt Rasmussen et al. (1994) developed a bioscrubber in which both
nitrification and denitrification took place. The bioscrubber was installed to treat
40000 m3Jh waste gas from a fish feed factory. During production hours organic
compounds were absorbed in the absorber at which also some nitrification took
place, while denitrification took place in the anoxic sump under the column. In the
production intermission periods, further nitrification took place in the same sump,
156 J W van Groeneslijn
but undemeath aerobic conditions. As organic compounds absorbed from the gas
served as electron donor for the denitrification process, addition of chemicals, such
as methanol, were not required.
Groenestijn el al. (1997) propose the use of separate absorbers, nitrification
reactors and denitrification reactors for the removal of ammonia from waste gases
from intensive livestock breeding. Many decentral absorbers can be connected to one
central biological treatment unit. This way ammonia is transported through small
water pipes rather than large gas pipes. Nitrification reactors could be constructed as
trickling filters with natural air draft. The denitrification process can be carried out
with methanol as electron donor. Based on laboratory experiments a 2 m3 absorber, a
4 m3 trickling filter, and a 1 m3 denitrification reactor should be used to treat 8000 m3
waste gasJh containing 15 mg NH 3/m3 .
9. Costs
9.1 BIOSCRUBBERS FOR VOC AND (AEROBIC) H2S
First statements about costs in review papers are discussed and subsequently these
are checked with data from individual reports. For ca1culation of amounts in Euros
(€), the exchange rates of September 2000 are used.
According to Kok (1991) the investment costs of bioscrubbers range from
NLG 20 (€ 9) to NLG 60 (€ 27) per m3Jh gas flow rate. The bioreactor accounts for
the largest part of these costs. According to Menig el al. (1997) a bioscrubber for
10000 m3Jh costs DM 500000 (€ 256000) and one for 60000 m3Jh DM 1200000 (€
613000), which means DM 20-50 (E 10-51) per m3Jh, depending on the size of the
bioscrubber.
In one of the earliest Dutch review papers on bioscrubbing (Joziasse and
Wiering, 1992) operation costs inc1uding capital depreciation range from NLG 1.20
to 4.50 (€ 0.54-2.04) per 1000 m3 treated gas in the 30000-130000 m 3Jh flow rate
range. Kok (1991) gives a wider range: NLG 1-5 (E 0.45-2.27) /1000 m3 . According
to Menig el al. (1997) the exact operation costs, inc1uding the capital costs, depend
on the time the bioscrubber is operational (day/night, weekends, etc.). In a certain
bioscrubber that operated 8000 h per year treatment of each 1000 m3 gas cost DM 1.3
(E 0.66), while DM 2.3 (€ 1.18) costs would be made ifthe same bioscrubber would
operate for only 3000 hours per year. According to the authors catalytic oxidation is
50% more expensive and adsorption even 100%. According to Kok (1991) 60% of
the operating costs of bioscrubbers are capital costs. Energy can be the next
important cost factor. Bioscrubbers that treat gases containing less than 0.5 g
VOC/m3 consume 1.5-3 kWhl1000 m3 treated gas. If the VOC concentration is
higher, the energy consumption increases.
The operation costs also can be expressed per kg VOC removed. Kok (1992)
ranges the costs for bioscrubbing between NLG 2 (€ 0.91) and NLG 5 (€ 2.27) per kg
VOC removed, with the lowest values for off-gases with high concentrations and
compounds readily soluble in water.
Specific cost analyses can be found in Kellner and Flauger (1998) in which a
bioscrubber for treatment of 20000 m3/h costs DM 450000 (€ 230000). According to
the authors this is less than many other bioscrubbers which would have cost DM
Bioscrubbers 157
700000 (€ 358000). Neverthe1ess, their investment costs are 40% higher than that for
a biofilter, but 35% lower than required for adsorption. The costs given are low, but
within the ranges given above. The bioscrubber described by Biiren (1989) had
operation costs exc1uding capital costs of DM 0.5 (E 0.26) per 1000 m3 gas. A
bioscrubber for treatment of 120000 m3 gas/h described by Fischer (1990) costs DM
1100000 (E 562000), which can be regarded as lower than the ranges given above.
The operation costs exc1uding capital costs were DM 0.33-0.45 (E 0.17-0.23) per
1000 m3 gas, of which 80-90% were electricity costs. Standefer and Willingham
(1998) compared their combination of a bioscrubber and a biofilter with sole
biofilters. In their specific case, the investment costs connected to the combination
were 20% lower than those of a biofilter, and the operation costs were 40% lower
(US$ 0.09 or € 0.10411000 m3 exc1uding capital costs). Incineration would have cost
nine times more.
Although Frohlich (1994) did not give exact cost figures for the 41000 m 3/h
capacity bioscrubber described, the electricity use was 60-90 kW and the water use
100-500 l/h, while the chemi cal costs were 17 times smaller than the e1ectricity
costs. According to Hansen (1998) bioscrubbing of gases containing 10 mg H 2S/m3 is
2 times more cost-effective compared with chemical scrubbing. This factor grows to
3 ifthe H2 S concentration is 50 mg/m 3 .
Janssen et al. (1997) arrived at operation costs for H2S removal from aerobic
gases, comparable with those described above. US$ 400000 (€ 464000) inc1uding
capital costs is required annually for treatment of 50000 m3 gas/h. In case the
operation was continuous, this would mean US$ 0.91 or € 1.0611 000 m 3 .
In summary it can be stated that the investment costs for bioscrubbers range
from € 10 to € 25 per m3 gas/h, mainly depending on the size. These costs may be
about 40% higher than those of biofilters. The operation costs range from € 1 to € 3
per 1000 m3 gas treated, mainly depending on the concentration, the solubility of the
compounds in water and the number of operating hours per year. High
concentrations, high solubility and a high number of operation hours lead to lower
operation costs. These costs can be higher or lower than those of biofilters. For gases
containing VOCs with Henry's law's coefficients lower than 0.001 and in
concentrations higher than 0.5-1 g VOC/m 3 bioscrubbers may be more cost-
effective.
9.2 OTHER BIOSCRUBBERS
According to Cetinkaya et al. (2000) bioscrubbing ofNO x from flue gases costs US$
0.81 or € 0.94/lb N and bioscrubbing of SOx costs US$ 333 (€ 386) per ton S, which
is 2.5 lower than SOx using caustic soda scrubbers. The operational costs of the
biological process for SOx removal are lower than those of the caustic soda process,
while the investment costs are higher. The pay back of the higher capital investment
will be site specific and based on the sulphur loading and flue gas flow rate.
Paybacks within two years can easily be achieved at higher sulphur loads. According
to Grootaerd et al. (1997) the operational costs for biological SOx are slightly lower
than that of the limestone forced oxidation process. Only if the gypsum market
becomes saturated the biological process becomes competitive. In addition, the
growing concern about heavy metals in gypsum from the limestone process may
influence decisions on flue gas treatment.
A bioscrubber for the removal of 5000 ppmv H2S from 42000 Nm3 natural
gas per hour costs US$ 5500000 (€ 6380000) (Janssen et al., 1999a). The operation
costs excluding depreciation of the capital costs amount US$ 430000 (€ 500000)
annually (about US$ 1.17 or € 1.36 per 1000 m\ which is lower than the operation
costs ofliquid redox technology.
10. ConcIusions and future opportunities
Comparable with biofilters and biotrickling filters, bioscrubbers can be cost effective
for treating waste gases containing VOC concentrations lower than a few g/m3. The
application of bioscrubbers can be more attractive than the use of biofilters if the gas
contains relatively high concentrations ( > 0.5 g/m 3) of hydrophilic compounds such
as alcohols, aldehydes, fatty acids and glycols. In addition, bioscrubbers can be cost-
effective for the treatment of gases containing N, S and halogenated compounds, and
attractive in case land area is limited. Combinations of bioscrubbers and polishing
steps can be interesting for treatment of gases with mixtures of hydrophobic and
hydrophilic compounds.
Considerable progress is made in the development of bioscrubbers for the
treatment of flue gases and anaerobic gases. The development of a bioscrubber for
NOx is still in its infancy, but promising. More attention still should be given to the
development of bioscrubbers containing alternative solvents. The addition of organic
absorbents, solid absorbents, adsorbents and reagents to the aqueous phase can
facilitate the absorption ofhydrophobic compounds such as alkanes and NO.
At present, biological waste gas treatment accounts for only about 2% of the
total world waste gas treatment financial turn over. By developments as described
above, physico-chemical processes can be substituted by biological ones and this
substitution process may be the most important opportunity for growth of the turn
over in biological waste gas treatment.
Abbreviations
COD chemical oxygen demand

DM Deutsch Mark
HTU height of a transfer unit
MEK methyl ethyl ketone
MIBK methyl isobutyl ketone
NLG Netherlands Guilders
NTU number of transfer units
ppm part per miII ion
TOC total organic carbon
UASB upf10w anaerobic sludge blanket
VOC volatile organic compound
€ Euro
$ Dollar
Bioscrubbers 159
Acknowledgement
The author gratefully acknowledges the support given by TNO Environment, Energy and Process
Innovation, Apeldoorn, the Netherlands, to write this chapter.
References
Anonymous, 1993. Composite flooring chosen for bioscrubber, Advanced Composites Bulletin
(March): 6-7.
Anonymous, 1998. Einsatz bio10gischer Schaume zur Ab1uftreinigung, Brochure from Umsicht,
Aachen, F.D. Germany.
Anonymous, 1999. Holzfaserplattenproduktion: Biowaschersystem nutzt thermophile Bakterien,

Wasser, Luft und Boden (9/1999): 62.
Beyreitz, G., Hilbner, R. and Saake, M. 1989. Biotechnologische Behandlung losemittelhaltiger

Abluft, Wasser Luft und Boden (9/1989): 53-57.
Buisman, C., Dijkman, H., Prins, W., Verbraak, P. and Hartog, H. den 1994. Biological (flue) gas
desulphurization, Lucht. 11(4): 135-137.
Bilren, E. 1989. Bio-catalytische Abgasreinigung in einer Kernmacherei. VDI Berichte. 735: 89-98.
Cesârio, M.T., Beeftink, H.H. and Tramper, 1. Biological treatment ofwaste gases containing poor1y-
water-soluble pollutants. In: Biotechniques for Air Pollution Abatement and Odour Control Policies.
Dragt AJ.and van Ham J. (eds.). Elsevier, Amsterdam, The Netherlands, 1992, pp. 135-140.
Cetinkaya, B., Sahlin, R.K., Abma, W.R., Dijkman, H., Meyer, S.F. and Kampeter, S.M. 2000.
Control FCC flue-gas emission. Hydrocarbon Processing. 79: 55-62.
Deziel, E., Comeau, Y. and Villemur, R. 1999. Two-liquid-phase bioreactors for enhanced
degradation ofhydrophobic/toxic compounds. Biodegradation. 10: 219-233.
Diehl, B. and Schllfer-Treffenfeldt, W. Biow!ischer - Aufbau, Verfahrensvarianten und

Betriebserfahrungen mit einem neuen Biowăscherkonzepl. In: Proceedings of the International
Symposium on Biological Waste Gas Cleaning, Maastricht, The Netherlands, April28-29, 1997. Prins
W.L. and van Ham J. (eds.). VDI Verlag GmbH (Publisher), Diisseldorf, F.R. Germany, 1997, pp.
345-352.
Dijkman, H. 1995. Biological gas desulphurization, Med. Fac. Landbouww. Univ. Gent 60/4b: 2677-
2684.
Dolfing, J., Wijngaard, AJ. van den and Janssen, D.B. 1993. Microbiologica1 aspects ofthe remova1
Fischer, K. 1990. Biologische Abluftreinigung, Expert Verlag (Publisher), Ehningen, F.R. Germany.
Fr1ih1ich, S. 1994. Entscheidungshilfen und Praxisbeispiele fur den Einsatz von Biowăschern zur
Abluftreinigung und Geruchsminderung. VDI Berichte. 1104: 81-94.
Gardin, H., Lebeault, J.M. and Pauss, A. 1999. Biodegradation of xylene and butyl acetate using an
aqueous-silicon oii two-phase system. Biodegradation. 10: 193-200.
Groenestijn, J.W. van, Harkes, M.P. and Baartmans, R.F.W. A novel bioscrubber for the removal of
ammonia from off-gases. In: proceedings of the International Symposium on Biological Waste Gas
Cleaning, Maastricht, The Netherlands, Apri! 28-29, 1997. Prins W.L. and van Ham J. (eds.). VDI
Verlag GmbH (Publisher), Diisseldorf, F.R. Germany. 1997, pp. 305-312.
Groenestijn, J.W. van and Lake, M.E. 1999. Elimination ofalkanes from off-gases using biotrickling
filters containing two liquid phases. Environ. Prog. 18: 151-155.
Grootaerd, H., De Smul, A. and Verstraete, W. 1997. Biological flue gas desulphurisation, Med. Fac.
Landbouww. Univ. Gent 62/4b: 1529-1534.
Hansen, N. 1998. Odour's biological option, Water Quality Int. (July/Aug): 12-14.
Hansen, N.G. and Rindel, K. Recent experiences with biological scrubbers for air pollution control in
Denmark. In: Biotechniques for Air Pollution Abatement and Odour Control Policies. Dragt A.J. and
van Ham J. (eds.). Eisevier Science Publishers B.V., Amsterdam, The Netherlands, 1992, pp. 143-
154.
Hecht, V., Brebbermann, D., Bremer, P. and Deckwer, W.-D. 1995. Cometabolic degradation of
trichloroethylene in a bubble column bioscrubber. Biotechnol. Bioeng. 47: 461-469.
Hvidtfeldt Rasmussen, H., Hansen, N.G. and Rindel, K. 1994. Treatment of odorous nitrogen
compounds in a bioscrubber comprising simultaneous nitrification and denitrification. VDI Berichte.
1104: 491-497.
Huckschlag, W. Biotechnologische Behandlung Phenol- und Formaldehydhaltiger Abluft. In:

Biotechniques for Air Pollution Abatement and Odour Control Policies. Drag! A.J. and van Ham J.
(eds.). Eisevier Science Publishers B.V., Amsterdam, The Netherlands, 1992, pp. 279-286.
Janssen, A.J.H. and Buisman, C.J. 1999a. The Shell-Paques desulfurisation process for H 2S removal
from high pressure natural gas, synthesis gas and Claus tail gas. In: Proceedings Ninth Gas Research
Institute Sulfur Recovery Conference. October 24-27.1999. San Antonio. Texas.
Janssen, A.J.H., Dijkman, H. and Janssen, G. 2000. Novel biological processes for the removal of H 2S
and S02 from gas streams, In: Environmental Technologies to Treat Sulfur Pollution. Lens P. and
HulshoffPol L. (eds.), IWA (Publisher), London.
Janssen, A.J.H., Hoop, K. de and Buisman, C.J.N. The removal of H2S from air at a petrochemical
plant. In: Proceedings ofthe International Symposium on Biological Waste Gas Cleaning, Maastricht,
The Netherlands, April 28-29, 1997 (Eds. W.L. Prins and J. van Ham), VDI Verlag GmbH
(Publisher), Diisseldorf, F.R. Germany, 1997,359-364.
Janssen, A., Marcelis, C. and Buisman, C. 1999b. Industrial applications of new sulphur
biotechnology. Water. 21 (Nov/Dec): 55-57.
Joziasse, J. and Wiering A.C.F. 1992 Monografieen Informatiesysteem Milieutechnieken -

Compartiment Lucht, RIVM report nr 736101007, Distribution Centre Ministry ofVROM (Publisher),
Zoetermeer, the Netherlands.
Kellner, C. and Flauger, M. 1998. Reduction of VOCs in exhaust gas of coating machines with a
bioscrubber. In: Proceedings 9Ist Annual Meeting & Exhibition ofthe Air & Waste Manage. Assoc.,
June 14-18, 1998, San Diego, California.
Kohl, A.L. and Nielsen R.B. 1997. Gas Purification, 5th edition, GulfPublishing Company, Houston,
Texas.
Kok, H.J.G. 1991. Biowassysteem voor de behandeling van koolwaterstofhoudende afgassen - Fase 1:
Haalbaarheidsstudie, TNO-IMET report 91-151, Apeldoorn, the Netherlands.
Bioscrubbers 161
Kok, H.J.G. Bioscrubbing of air contaminated with high concentrations of hydrocarbons. In:
Biotechniques for Air Pollution Abatement and Odour Control Policies. Drag!, AJ and van Ham, 1.
(eds.), Eisevier Science Publishers B.V., Amsterdam, The Netherlands, 1992, pp. 77-82.
Koot, AJ.C. 1980. Behandeling van afvalwater, Waltman (Publisher), Delft, The Netherlands.
Lewis, W.K. and Whitman, W.G. 1924. Principles ofgas absorption. Ind. Eng. Chem. 16: 1215-1220.
Loy, J., Hofmann, A. and Egerer, B. 1998. Troubleshooting in industrial bioscrubbers and biotrickling
plants in the elimination ofodors and VOCs. In: Proceedings 91" Annual Meeting & Exhibition ofthe
Air & Waste Manage. Assoc., June 14-18, 1998, San Diego, California.
Menig, H., Krill, H. and Jăschke, T. Kosten und Effizienz biologische Verfahren im Vergleich zu
anderen Abluftreinigungssystemen. In: Proceedings of the International Symposium an Biological
Waste Gas Cleaning, Maastricht, The Netherlands, April 28-29, 1997. Prins W.L. and van Ham J.
(eds). VOI Verlag GmbH (Publisher), Dilsseldorf, F.R. Germany, 1997, pp. 27-58.
Nishimura, S. and Yoda, M. 1997. Removal of hydrogen sulfide from an anaerobic biogas using a bio-
scrubber. Water Sci. Technol. 36: 349-356.
Nurul Islam, A.K.M., Hanaki, K. and Matsuo, T. 1998. Fate of dissolved odorous compounds 111
sewage treatment plants. Water Sci. Technol. 38: 337-344.
Ottengraf, S.P.P. Exhaust Gas Purification. In: Biotechnology: A comprehensive treatise. Rehm H.-J.
and Reed G. (eds.), VCH, Weinheim, Germany. 1986, VoI. 8, pp. 425-452.
Ottengraf, S.P.P. 1987. Biological systems for wastegas elimination. Tibtech. 5: 132-136.
Overcamp, TJ., Chang, H.-C. and Grady Jr, C.P.L. 1993. An integrated theory for sllspended-growth
bioscrubbers. Air & Waste. 43: 753-759.
Pagella, C. and De Faveri, D.M. 2000. H2S gas trea!men! by iron hioprocess. Chem. Engin. Sci. 55:
2185-2194.
Pagella, c., Silvestri, P. and De Faveri, D.M. 1996. H2 S gas treatment with Thiobacillus ferrooxidans:
overall process peformance and the chemical step. Trans IChemE. 74(A): 123-132.
Parker, WJ., Collins, J., Wells, J. and Kennedy, K. Biological treatment of air streams containing
chlorinated organic compounds. In: Proceedings of the 91" Annual Meeting & Exhibition of the Air &
Waste Managc. Assoc., June 14-18, 1998, San Diego, California.
Parkinson, G. 1996. Bioscrubber cleans styrene from waste gases. Chem. Engin. April: 15.
Phipps Jr, D.W. and Ridgway, H.F. Using biologically activated foam for the degradation of vapor
phase volatile organic contaminan!s. In: Proceedings 1995 Conference on Biofiltration. Hodge D.S.
and Reynolds F.E. (eds.). October 5-6, 1995, Los Angeles, California, The Reynolds Group
(Publisher), Tustin, California.
Pirt, SJ. 1985. Principles of Microbe and Cell Cultivation. Blackwell Scientific Publications
(Publisher), Oxford, UK.
Poppe, W. and Schippert, E. Das KCH-Biosolv-Verfahren in Kombination mit einem Biowăscher

herkiimlichcr Art - eine Verfahrensentwicklung zur Abluftreinigung fur wasserliisliche und schwer
wasserliisliche Schadstoffe. In: Biotechniques for Air Pollution Abatement and Odour Control
Policies. Dragt AJ.and van Ham J. (eds.). Elsevier, Amsterdam, The Netherlands, 1992, pp. 71-76.
Riet, K. van 't, and Tramper, J. 1991. Basic Bioreactor Design. Marcel Dekker, New York.
Schippert, E. 1986. Biologische Abluftreinigung. Umwelt. 8: 527-529.
Schippert, E. 1989a. Biowăscher nach einer Dosenlackieranlage. VOI Berichte. 735: 77-88.
Schippert, E. 1989b. Das Biosolv-Verfahren von Keramchemie zur Absorption von schwer
wasserloslichen Losemitteln. VOI Berichte. 735: 161-178.
Schippert, E. 1994. Biowăschertechnologie. VOI Berichte. 1104: 39-56.
Stand efer, S. and Willingham, R. 1998. Industrial application ofbiofilter technologies. In: Proceedings
USC-TRG Conference on Biofiltration, October 22-23, 1998, Los Angeles, California. ReynoldsJr,
F.E. (ed.). The Reynolds Group (Publisher), Tustin, California: 1-8.
Stanier, R.Y., Adelberg, E.A. and Ingraham, J.L. 1978. General Microbiology. Macmillan (Publisher),
London.
Tchobanoglous, G. and Burton, F.L. 1991. Wastewater Engineering. Metcalf & Eddv. Inc. (ed.).
McGraw-Hill. New York.
Tramper, J., Wolters, 1. and Verlaan, P. The liquid-impelled loop reactor: a type of density-difference-
mixed bioreactor. In: Biocatalysis in Organic Media. Laane c., Tramper J. and Lilly M.D. (eds.).
Elsevier , Amsterdam, The Netherlands, 1987, pp. 31 1-316.
VOI/OIN-Handbuch Reinhaltung der Luft, Volume 6: Biological Waste Gas Purification (VOI 3478).
1996. VOI Verlag (Publisher), Dtisseldorf, Germany.
CHAPTER 6 MEMBRANE BIOREACTORS
Sarina J. ERGAS
1. Introduction
Successful biofiltration applications have been Iimited to control of relatively soluble

VOCs at low loading rates. Biofiltration is also of limited use for compounds that
produce acidic or toxic metabolites or are degraded via cometabolism. To overcome
the limitations described above, gas phase biological treatment systems must be
developed which:
1) incorporate mass transfer of gas phase pollutants across a media with high
specific surface area and low diffusion length;
2) incorporate high biomass concentrations to maintain high biodegradation rates;
3) provide a method for wasting biomass to prevent clogging;
4) provide a method for addition of pH buffers, nutrients, co-metabolites, and/or
other amendments to support the microbial population and neutralise acidic
metabolites.
Hollow fibre membrane bioreactor (HFMB) systems have been under
investigation which meet the above requirements and have been shown to achieve
high VOC removal efficiencies in small reactor volumes. A schematic of a typical
HFMB system is shown in Figure 6.1. The hollow fibre membranes serve as a
support for the microbi al population and provide a large surface area for VOC and
oxygen mass transfer. Waste gases containing VOCs are passed through the lumen of
the hollow fibres. Soluble compounds in the gas phase are transferred through the
membrane pores and partition into a VOC-degrading biofilm surrounded by a
circulating nutrient media. Compounds in the biofilm are available for
biodegradation. Potential advantages of hollow fibre membrane bioreactors for waste
gas treatment include:
• Ability to continuously remove biomass to prevent clogging at high VOC loading

rates;
• Ability to remove degradation by-products and add pH buffers, nutrients and/or
co-substrates to support the microbial population and neutralise acidic
metabolites;
• Ability to treat low solubility compounds such as nitric oxide (NO) and
methane;
• Separation of the microbial process from the gas being treated. This may be
useful in indoor air applications where carryover of biomass into the ambient air
is a concern;
• Optimal humidification ofthe gas stream;
• Low pressure drop;
• Gas and Iiquid flow rates can be varied independently without flooding or
foaming;
• Modular design, no mov ing parts;
163
C. Keflfles afld M. C. Veiga (eds.), Bioreactorsfor Waste Gas Treatmeflt, 163-177.
164 S. J Ergas
Disadvantages of HFMBs for waste gas treatment include high capital costs
and that the technology has not been demonstrated at full-scale .
2. Membrane bioreactor applications for polIution control
Hollow fibre membrane bioreactors have been used in a number of pollution control
applications inc1uding:
• Separation and retention ofbiosolids;

• Bubble free aeration ofbioreactors;
• Providing hydrogen and other low solubility gases to bioreactors;
•
Contaminated
InJet Air
MOOngI
Recirculation
Flow
HolJow
Fiber
Membranes
Figure 6.1. Schematic of a hollow fibre membrane bioreactor for biological air polllltion
control (after Shllmway, 1997).
Membrane bioreactors 165
• Extractive membrane bioreactors for controlled transfer of pollutants from

industrial waste waters;
• Biological air pollution control.
Membrane bioreactors are most often used for separation of biomass. The
ultrafiltration of microfiltration membranes are used as a substitute for
sedimentat ion. These systems are used to produce high quality effluent and/or
provide high solids retention times without washout of biomass. These applications
are well documented in the literature and are outside the scope of this discussion
(Stephenson el al., 2000).
Gas to liquid transfer membrane systems were first introduced for use in
blood oxygenation and have been studied extensively for this application (Lund el
al., 1996). Pollution control research into membrane aeration HFMBs has been
focused on the enhancement of oxygen transfer in high oxygen demanding
applications such as the treatment of high strength industrial waste waters
(Yamagiwa el al., 1998) and nitrification (Brindle and Stephenson, 1996b; Brindle el
al., 1998). These systems have also been used for bioreactors degrading VOCs, since
the membranes allow for reactor aeration without stripping volatile compounds to the
atmosphere (Pressman el al., 1999). Owing to the low solubility of oxygen, oxygen
transfer often limits aerobic degradation of high oxygen demanding waste waters.
Many facilities switch from atmospheric air to pure oxygen to enhance oxygen mass
transfer; however, conventional oxygenation systems have high power requirements
and are incompatible with attached growth reactors (Stephenson el al., 2000).
Hollow fibre membranes provide a very high surface area for transfer of oxygen
directly to the biofilm. Hydrophobic hollow fibres have been developed with sealed
hydrophilic ends that enable 100% utilisation efficiency in gas to liquid mass transfer
applications (Ahmed and Semmens, 1992a, I 992b).
A novel application of HFMB gas transfer technology, hydrogenotrophic
denitrification of drinking water, has been investigated by several researchers
(GantzerI995; Lee and Rittmann, 2000; Ergas and Reuss; 2001). A number of
common genera of bacteria can use hydrogen as an electron donor and nitrate as an
electron acceptor under anoxic conditions. These organisms have been shown to
denitrify nitrate contaminated drinking water to acceptable levels. Disadvantages of
hydrogenotrophic denitrification include lower denitrification rates and the difficulty
in dissolving sufficient quantities of hydrogen into the water owing to its low
solubility. Use of HFMBs has been shown to support high biomass densities and
improve hydrogen mass transfer rates while preventing the waste of excess hydrogen
and avoiding accumulation of explosive gasses in a confined space.
Extractive membrane bioreactors have been used to biologically trea!
industrial waste waters in the presence of high concentrations of acids, bases and
salts that can inhibit degradation (Livingston, 1994; Livingston el al., 1998).
Extractive membrane bioreactors utilise dense silicone membranes that selectively
extract organic pollutants such as cholorethanes, chlorobenzenes and toluene, from
polar and ionic compounds. A VOC degrading biofilm grows on the surface of the
membranes creating a driving force for mass transfer. Owing to the selectivity of the
membranes, the biofilm is isolated from the harsh conditions in the waste water to
ensure high biodegradation rates.
166 S. J Ergas
3. Membrane fundamentals
A membrane separates two distinct bulk phases of a system while allowing the
transport of compounds from one phase to the other. In waste gas treatment
applications, gases are most often blown through the lumen (inside) of tubes made
from membrane materials. Pollutants from the gas phase diffuse through the
membranes to a liquid phase on the shell si de (outside) of the membranes. The
membranes also serve as a support for the microbi al population. Once in the Iiquid,
compounds are biodegraded, creating a concentration gradient which serves as a
driving force for mass transfer.
Membranes are available in a wide variety of materials, porosities, and pore
sizes. For successful application, membrane materials must strike a balance between
reasonable mechanical strength, high permeability and high selectivity (Stephenson
el al., 2000). For microporous membranes, high selectivity requires a membrane
material with a narrow range of pore sizes and high permeability requires a
membrane material with a high porosity.
Dense and porous membranes are fundamentaIly different from each other.
Dense membranes rely on physical-chemical interactions between the permeating
compounds and the membrane materials. The mass transfer rate through a dense
membrane depends on the solubility and the diffusivity of the permeating compound
in the dense matrix (Reij el al., 1998). Dense membranes are limited to polymeric
materiaIs, such as latex and silicone, they can be operated at high gas pressures, are
resistant to chemical and mechanical abrasion (Fitch el al., 2000; Stephenson el al.,
2000). Dense membranes have also been shown to be more resistant to biofouling
than porous membranes (Cote el al., 1988; 1989), possibly because the hydrophobic
nature of silicone resists attachment of microorganisms.
Microporous hydrophobic membranes are most often used in gas transfer
applications because they provide a high gas permeability, while not aIlowing
transport of water across the membrane. The membrane pores remain gas filled and
compounds transfer from the gas stream through the membrane pores by gaseous
diffusion. At excess liquid side pressures above the critical pressure, M> cr. water
enters the pores of the membranes, dramatically decreasing mass transfer rates
(Ergas and Reuss, 2001). Gas side pressures greater than the bubble point result in
bubble formation in the liquid phase (Semmens et al., 1999). Thus over the excess
pressure range of O to ~Pcr the gas-liquid interface is immobilised at the mouth ofthe
membrane pore on the liquid side (Sikar, 1992).
Microporous hydrophobic membrane materials include polytetrafluorethylene
(PTFE), polypropylene, Teflon™, Gortex™ (PTFE/nylon) and other composites.
Hydrophobic microporous membranes coated with an extremely thin layer of
silicone have aIso been investigated (Sikar, 1992). The thin silicone layer increases
mass transfer resistance but also decrease biofouling. Microporous hydrophobic
membranes are available with pore diameters between 0.1 and 1.0 f.U11 (Stephenson el
al., 2000). The membranes are manufactured as small diameter (200-400 )lm ID)
hoIlow fibre bundles that provide surface area to volume ratios as high as 30 to 100
cm'! (Sirkar, 1992). This is an order of magnitude greater than equivaIent sized
packed towers.
Gas to liquid transfer can also be carried out using microporous hydrophilic
membranes such as polysulfone and cellulose membranes. In these applications, gas
Membrane bioreaclors 167
pressure must be higher than liquid side pressures. The gas-liquid interface is thus
immobilised at the pore mouth on the gas side ofthe membrane.
Control of biomass thickness has been shown to be a key operational
consideration in continuously operated HFMBs for aeration of waste waters
(Stephenson el al., 2000) and biological waste gas treatment (Ergas and McGrath,
1997). Decreased HFMB performance has been observed in bioreactors afier the
development of a thick biofilm owing to substrate mass transfer limitations,
membrane fibre plugging, decreased biomass activity, and/or metabolite
accumulation (Brindle and Stevenson 1996a; Freitas dos Santos el al. 1997). Hollow
fibre bundles also tend to clump together when biofilm growth is high. resulting in
fibre tangling and reduction of available membrane surface area. Several operational
strategies have been used to maintain film thickness at an optimum level including
the use of cross-flow membrane configurations (Ahmed and Semmens, 1996) and
periodic shearing of biomass from the membranes using high liquid velocities
combined with scouring with gas bubbles (Pankhania el al., 1994; Dolasa and Ergas,
2000).
4. Research overview
A summary of the laboratory scale investigations of HFMBs for air pollution control
is shown in Table 6.1. In an early study by Hartmans et al. (1992) a HFMB was used
to control air emissions of toluene and dichloromethane. Mass transfer coefficients
were determined for a number of different membrane materials. Using the
experimentally determined mass transfer coefficients, dichloromethane removal was
simulated. Results of simulations suggested that a significantly lower reactor volume
would be required for a HFMB than for a biotrickling filter. Greater than 95%
removal of toluene and dichloromethane were observed in experiments with aflat
sheet membrane bioreactor.
Parvatiyar el al. (1996a) used a two module-in-series polysulfone HFMB to
investigate toluene removal from a contaminated airstream. Toluene removal reached
84% with a 16 second gas residence time and an inlet concentration of 600 ppmv. A
similar experimental system was used by the authors to study degradation of
trichloroethylene (TCE) (Parvatiyar et al., 1996b). The biofilm was initially
acclimated to toluene, then gradually weaned from a toluenerrCE mixture to 100%
TCE. A 30% TCE removal efficiency was achieved with a 36 second gas residence
time.
Fitch et al. (2000) compared mass transfer and biodegradation rates for
benzene and butanol contaminated gases in HFMBs that utilised dense (latex and
silicone), microporous hydrophobic (polypropylene) and microporous hydrophilic
(polysulfone) membranes. The highest overall pressure drops were observed with the
polypropylene membranes due to the smaller diameter of these fibres compared with
the other fibres. Significant sorption of benzene was observed in initial tests with the
polysulfone membranes; therefore, butanol was used in subsequent experiments with
these membranes. Removal efficiencies for butanol of up to 99% were obtained at an
inlet concentration of 200 ppmv. Dense high permeability latex and silicone
membranes were found to have high benzene mass flux rates, possibly because the
greater solubility of benzene in the polymers than in air creates a greater effective
concentration gradient than observed in the air phase. The low total surface area of
the dense membrane tubes limited overall removal; however, and the polypropylene
168 S. J Ergas
Table 6.1. Summary ofHFMB for waste gas treatment research
Reactor Compound Membrane Fibre Pore Membrane References

Type materials ID size area (m z)
(mm) (Il m)
Toluene, Polypropylene 0.l0 0.004 Hartmans

dichloromethane et al., 1992
Flat Propene Polypropylene NA 0.1 0.0040 Reij etal.,
sheet 1995
Hollow Propene Polypropylene 1.8 0.2 0.10 Reij and
fibre Hartmans,
1996
Hollow Toluene Polysulfone 1.1 0.028 Parvatiyar
fibre et al .. 1996"
Hollow Trichloroethene Polysulfone 1.1 0.028 Parvatiyar
fibre el al.,
1996b
Flat Propene Polypropylene NA 0.1 0.0040 Reij el al.•
sheet 1997
Hollow Toluene Polyethylene 0.28 0.23 Ergas and
fibre McGrath.
1997
Hollow Toluene Polypropylene 0.20 0.05 0.37 Ergas el al.,
fibre 1999
Hollow ToluenelTCE Polypropylene 0.20 0.05 0.37 Dolasa and
fibre Ergas,2000
Hollow Ammonia Polyolefin 0.20 0.063 Keskiner
fibre multilayer and Ergas,
2001
Hollow Benzene Latex rubber 9.5 NA 0.012 Fitch et al.,
fibre 2000
Hollow Benzene Silicone 9.5 NA 0.012 Fitch el al.,
fibre rubber 2000
Hollow Benzene Polypropylene 0.20 0.2 0.30 Fitch el al.,
fibre 2000
Hollow Butanol Polysulfone 1.1, 0.05 0.030, Fitch el al.,
fibre 1.9, 0.022, 2000
2.7 0.013
Hollow Trichloroethene Polypropylene 0.24 0.03 0.70 Pressman el
fibre al., 2000
membrane unit was the most effective on the basis of removal per total unit volume
ofreactor.
In my own laboratory we have conducted a number of studies using toluene
as a model VOC. In our first set of experiments (Ergas and McGrath, 1997) a
laboratory scale HFMB was constructed and operated with toluene at varying loading
rates. The gasses pas sed from an inlet manifold to a membrane distributor made from
336 polypropylene hollow fibres (280 flm ID, 63% porosity and active fibre length of
1.1 m). A plot of removal efficiency vs. gas flow rate over the experimental period is
shown in Figure 6.2. Toluene removal efficiencies of gre ater than 97% were
achieved with an inlet toluene concentration of 100 ppmv and gas flow rates less
than 1.0 l/min (l.4 s residence time). When the gas flow rate was increased above 1.2
l/min (1.1 s residence time) a significant decrease in removal effieieney was
observed. Removal efficieney was found to deerease over the four month operational
period due to c10gging of the bioreaetor with microbi al biomass, possibly due to the
growth of nitrifying bacteria.
In subsequent experiments (Ergas el al., 1999) we investigated the effeets of
toluene loading rate, gas residenee time, and liquid phase turbulenee on toluene
removal in a laboratory-scale HFMB. Nitrate was used as a nitrogen souree to
diseourage the growth of nitrifying baeteria. Initial aec1imation of the mierobial
eulture to toluene oeeurred over a period of nine days, after which a 70% removal
efficiency was achieved at an inlet toluene concentration of 200 ppmv and a gas
residence time of l.8 s (elimination capacity of 20 g/m 3 .min). At higher toluene
loading rates a maximum elimination capaeity of 42 g/m 3 min was observed. Liquid
phase reeirculation rate had no effect on toluene removal in the HFMB.
100,--,~,-,--=~~=-------------------,
90
~
o
>-
()
c
Q) 80
·13
it:
w
ro> 70
o
E
Q)
n::
60
0.2 0.4 0.6 0.8 1.2 1.4 1.6

flowrate (L/min)
Figure 6.2. Removal efficiency in a HFMB vs. gas flow rate (after Ergas and McGrath,
1997). Removal efficiency was determined after an acclimation period at each flow rate.
The inlet toluene concentration was maintained at 100 ppmv, and the liquid recirculation rate
was41/min.
170 S. J Ergas
We have also investigated the use of a HFMB to control ammonia using

nitrifying bacteria (Keskiner and Ergas, 2001). The reactor utilised polyolefin
multilayer membrane bundles consisting of 200 fibres, with a length of 50 cm, an
inner diameter of 200 Ilm and porosity of 42%. Greater than 92% removal efficiency
was obtained at an inlet ammonia concentration of 60 ppmv and a gas residence time
of less than OA second. Ammonia mass transfer rates were found to increase at
higher recirculation rates. Biomass adhesion and accumulation was not a problem in
this system, possibly owing to the slow growth rates of nitrifying bacteria.
Experiments are currently being conducted with this reactor using nitrifying bacteria
ta remove nitric oxide from combustion gas streams.
4.1 TREATMENT OF LOW SOLUBILITY COMPOUNDS
In biological air pollution control systems, compounds must partition from the gas
phase into the moist biofilm before they can be degraded. Pollutant concentrations at
the gas/biofilm interface can be described by Henry's law:
(6.1)
where H is the Henry's law' s coefficient, Cg is the gas phase concentration, and SL is
the liquid phase concentration. Conventional biofilters have therefore been limited to
the control of relatively soluble compounds. A number of environmentally relevant
compounds such as nitric oxide (NO), hexane and methane are biodegradable but
have high Henry's law constants. For these compounds mass transfer from the gas
phase to the biofilm limits removal unless very large reactor volumes are used.
Owing to their higher mass transfer rates, HFMBs may make biological treatment of
these compounds more economically feasible.
A number of studies have been carried out using propene as a model VOC
because its low solubility makes it difficult to remove in conventional biofilters (air-
water partition coefficient at 2S o C of 8.6). Aflat sheet microporous polypropylene
HFMB inoculated with Xanthobacter Py2 was investigated by Reij et al. (1995).
After five days with an inlet propene concentration of 2300 ppmv, the biofilm
acclimated and 58% propene removal was maintained for the duration of the thirty-
day test. Because ofpropene's poor solubility, all mass transfer resistance was found
to be in the liquid phase. For more soluble compounds, the authors determined that
membrane phase resistance could approach the same order of magnitude as liquid
phase resistance. A similar reactor system and bacterial culture was investigated by
Reij et al. (1997) to degrade propene at concentrations varying from 10 to 1000
ppmv. Once a biofilm was established, propene flux to the membranes was stable,
even at low concentrations (9-30 ppmv) when mass transfer limitations should be
greatest.
Reij and Hartmans (1996) investigated a HFMB that utilised 40
polypropylene hollow fibres with a length of 500 mm, an inner diameter of 1.8 mm
and a pore size of 0.2 Ilm. Propene was again used as a model compound due to its
low solubility. Maximum propene removal rates were 70-110 g/m 3 .h. A gas
residence time of 80 s was required for 95% removal at an inlet propene
concentration of 480 ppmv. The reactor was changed from ammonia as a nitrogen
source to nitrate to discourage the growth of nitrifying bacteria. Increasing the shell
side velocity was found to alleviate clogging of the fibres with biomass. A gradual
decrease in propene degradation was observed, possibly owing to ageing of the

biofilm.
4.2 COMETABOLISM
A number of chlorinated organic compounds, such as TCE, can only be degraded

aerobically under co-metabolic conditions. Cometabolism is defined as the
transformation of a compound by a microorganism that is unable to use the substrate
as a source of energy or as an essential nutrient element. A second substrate (primary
substrate) is used to support growth and induce the enzymes necessary to co-
metabolise the target compound. The primary substrates often inhibit metabolism of
the target compound due to enzyme competition (Alvarez-Cohen and McCarty,
1991). In plug flow reactors, such as biofilters, the induced bacteria are located near
the inlet of the bioreactor where the primary substrate is readily degraded. In these
areas, competitive inhibition occurs between the two compounds. In the remaining
sections, the bacteria do not get sufficiently induced with the low primary substrate
concentrations remaining to co-metabolise the target compound. Microbial activity
can also be inhibited by the toxicity of the target compound or its metabolites. The
result is low removal efficiencies for TCE in conventional biofilters (Speitel and
McLay,1993).
Several researchers have studied HFMBs for cometabolism in TCE in waste
waters. Aziz et al. (1995) investigated a HFMB with an externa) semi-batch reactor
for TCE cometabolism. Waste water contaminated with TCE flowed insi de the
hollow fibres, with a liquid media containing an active methanotrophic culture
circulating around the fibres. At residence times of 5 to 9 minutes in the fibre lumen,
TCE conversions of 80% to 95% were observed. Pressman et al. (1999) presented a
follow up study with a similar system and methanol as the primary substrate. Over
93% ofthe transferred TCE was biodegraded.
Pressman et al. (2000) used a HFMB for cometabolism of TCE contaminated
gases. A pure culture of Methylosinus trichosporium OB3b PP358 was grown in a
continuous flow chemo stat and circulated through the fibre lumen of a HFMB while
TCE contaminated air was circulated on the shell side of the reactor. Methylosinus
trichosporium OB3b PP358 is a methanotrophic bacterium that has the ability to
rapidly co-metabolise chlorinated solvents when grown on either methane or
methanol. Between 54% and 84% TCE transfer was observed and 92% to 96% ofthe
transferred TCE was co-metabolised at gas residence times of 1.6-5.0 minutes.
Biomass clogging did not occur in this system, possibly because the biomass was
pumped through the lumen rather than the shell side of the fibres.
In my own laboratory we have investigated TCE cometabolism in HFMBs
using toluene as a primary substrate (Doi asa and Ergas, 2000). A TCE co-
metabolising culture enriched from a waste water seed was inoculated into the
HFMB. Initially toluene was supplied to the reactor to build a sufficient biomass
density on the fibres. After steady state toluene removal was achieved, TCE was
added to gas phase of the reactor. Toluene was added in three different
configurations: (1) as a mixture with TCE in the gas phase; (2) by pulsing into the
gas phase; or (3) to the liquid phase. Addition of a toluenerrCE mixture through the
fibres resulted in an initial decrease in toluene removal followed by complete
recovery within 5 days and a maximum TCE removal efficiency of approximate1y
30%. Pulsing oftoluene and TCE through the membranes did not result in significant
172 S. J Ergas
TCE removal. Adding TCE in the gas phase and toluene in the liquid phase resulted
in a maximum TCE removal efficiency of 23%, however, results were highly
variable and appeared to be related to liquid phase toluene utilisation rates and
biofilm thickness.
5. Theoretical models
A number of researchers have presented mathematical models of mass transfer and

biodegradation of substrates by biofilms growing on the surfaces of gas-transfer
membranes (Livingston 1993; Essila et al. 2000). The general biofilm model
reported here was developed in collaboration with Dr. Mark W. Fitch of the
University of Missouri, Rolla, and tested using experimental data from HFMB
studies conducted using toluene as a model VOC (Ergas et al. 1999). The model uses
an inert surface to establish a boundary condition, with substrate entering the biofilm
from the gas-liquid interface. The model was derived for a single lumen and related
to the total removal by the number of fibres. A conceptual model of this system is
shown in Figure 6.3. Model assumptions include: steady state operation; Monod
biodegradation kinetics; and constant biomass density, Pb. Because concentration
varies both axially and radially, no analytical solution exists for a single lumen.
Therefore the lumen is divided along the axis into n sections, each with an axial
length, I'1z. The influent gas concentration to the ntil section is the concentration
exiting the previous section, Cn- l • The concentration exiting the nth section, Cn , is
equal to the influent less the removal in the section due to the mass transfer and
biodegradation.
Memhrane Fiber ReacrOf

Wal!
~:s~~:o~-]
Figure 6.3. Conceptual model ofthe phases ofthe hollow fibre membrane reactor (after
Ergas el al., 1999).
5.1 MEMBRANE MASS TRANSFER
The flux, Jn, of substrate through the membrane can be expressed as:
(6.2)
where km is the membrane mass transfer coefficient, Am is the area of the membrane
in the section, and C n.m is the gas phase concentration on the outcr face of the
membrane.
A number of authors have assumed that gas and membrane resistances are
negligible compared to liquid phase resistance (Ergas and McGrath, 1997; Yang and
Cussler, 1986; Cote et al., 1989) and therefore C n•m is approximately equal to the
concentration in the gas stream, C n-1• Since the membrane is surrounded by biofilm,
the liquid phase concentration at the biofilm interface, Sn,O, can be related to Cn-I
using Henry's law:
Sn,O = Cn _ 1 / H (6.3)
Equation 6.3 sets the inner surface (left-hand) boundary condition for the biofilm
model.
5.2 SUSPENSION MASS TRANSFER AND DEGRADATION
The suspension (liquid volume) was treated as a continuous flow stirred tank reactor
(CFSTR). The mass flux of the substrate, h, from the biofilm to the liquid can be
described by:
(6.4)
where kL is the liquid mass transfer coefficient, Ab is the outer surface area of the
biofilm, SIl.i is the VOC concentration at the outer surface of the biofilm, and SL is
the bulk liquid VOC concentration. Assuming Monod biodegradation kinetics, a
mass balance on the liquid volume yields:
PL J.lmaxSL V=QS -QS +J (6.5)

Y K +S o L h
S 1.
where V is the liquid volume, PL is the biomass density in the liquid, flmax is the
maximum specific growth rate, Y is the yield coefficient, Ks is the half saturation
coefficient, Q is the liquid flow rate, and So is the inf1uent VOC concentrat ion to the
suspension. Substituting Equation 6.4 into 6.5 and solving for Sn,i yields:
(6.6)
This relationship sets the exterior surface (right hand) boundary condition for the
biofilm model.
174 S. J Ergas
5.3 BIOFILM MASS TRANSFER AND DEGRADATION
Assuming no advection in the biofilm, steady state (aS I at = O), no concentration

gradient in the z direction, and Monod substrate utilisation kinetics, the continuity
equation for the biofilm in cylindrical coordinates (r, z, and 8) is
(6.7)
where S is the substrate concentration in the biofilm and Ds is the VOC diffusion
coefficient in the biofilm. There is no analytical solution for Equation 6.7 therefore a
numerical solution was generated.
To fit the models, the biofilm density was varied until the predicted removal
matched the observed removal (Ergas el al., 1999). The biofilm biomass density was
calibrated at 29000 mgll, slightly higher than that reported by Characklis and
MarshaU (1990). The numerical model predicted the observed trend that shell side
liquid flow rate, had little effect on observed removal once a biofilm was established.
The model slightly underpredicted the effect that substrate loading rate had on
removal in the system. The only case of poor prediction by the model was the effect
of gas residence time on removal. Observed removals were higher than predicted
removals for these mns. If the pore space of the membranes became water-filled
during the course of experimentation, as discussed above, an added resistance to
mass transfer would be expected, and mass transfer to the biofilm could become
dependent upon the gas flow rate (gas to liquid transfer). Sensitivity analysis
indicated that removal was a strong function of the biofilm phase biomass density
and also of the biofilm diffusion coefficient, with diffusion rates below (lOr9 m2/s
resulting in decreased removal rates.
6. Conclusions
Hollow fibre membrane bioreactors are a promising technology for the treatment of
biodegradable gas phase poUutants. It has been shown to be effective for aerobic
degradation of a range of compounds including ammonia, benzene, butanol,
dicholoromethane, propene, TCE and toluene. Advantages of HFMBs include high
mass transfer rates, low pressure drops and small reactor volume requirements. In
addition, the ability to separate the microbial population from the gases being treated
allows for independent optimisation of each phase of the system. In the opinion of
this author, the most promising areas of air pollution research for HFMBs include:
• Low solubility compounds;

• Cometabolism of chlorinated organic compounds;
• Compounds requiring specialised microbial populations or conditions;
• lndoor air applications.
The greatest disadvantages for the technology are high capital costs and that
it has yet to be demonstrated at fuU scale. Owing to their modular design: however.
HFMBs should be relatively easy to scale up. EnviroGen Inc. and the Medical
University of South Caro lina have developed a prototype with fund ing from the
Membrane hioreactors 175
Department of Energy that they hope to field test in 2001 (Togna, 2000). Long term
studies have been conducted with membrane aeration bioreactors (Stephenson et al.,
2000), which are similar in concept and operation. Membrane reactors have also
been in use for several decades for blood oxygenation (Sikar, 1992) and separation
and retention of biosolids (8rindle and Stephenson, 1996a). In common with
conventional biofilters, HFM8s have problems with excess biofilm growth and long
term stability of VOC degrading biofilms. It is also unknown what effect long
periods of association of membrane materials and biomass will have on mass transfer
rates and mechanical strength in these systems.
Acknowledgements
1 would like to thank my students Ayesha Dolasa, Carolyn Gendron, Yenner Keskiner,
Michael McGrath, Fereshteh Mehmandoust, Andreas Reuss and Leslee Shumway for their
work on these projects. This material is bascd on work supported by the National Sciencc
Foundation. Any opinions, findings, conclusions or recommendations expres sed in this
material are those ofthe authors and do not necessarily reflect the views ofNSF.
References
Ahmed, T., Semmens, M.J. 1992a. Use ofsealed end hollow fibers for bubbleless membrane aeration:
Experimental studies. J. Membr. Sci. 69: 1-10.
Ahmed, T., Semmens, M.J. 1992b. The use of independently sealed microporous hollow fiber
membranes for oxygenation ofwater: Model development. J. Membr. Sci. 69: 11-20.
Ahmed, T., Semmens, M.J. 1996. The use of transverse hollow fibers for bubbleless membrane
aeration. Water Res. 30: 440-446.
Alvarez-Cohen, L., McCarty, P. 1991. Effeets of toxicity, aeration, and reduetant supply on
triehloroethylene transformation by a mixed methanotrophie culture. Appl. Environ. Microbiol. 57:
228-235.
Aziz, CE., Fiteh, M.W., Linquist, L.K., Pressman, JG., Georgiou, G., Spcitel, GT. 199')
Methanotrophie biodegradation of triehloroethylene in a hollow fiber membrane bioreactor. Environ.
Sei. Teehnol. 47: 2574-2583.
Beyenal, H" Seker, S" Tanyo1ae, A" Salih, B, 1997. A mathematieal model for hollow fiber biofilm
reaetors. AIChE J, 43: 243-250,
Brindle, K., Stephenson, P, 1996a. Mini review: The applieation of membrane biological reactors for
the treatment ofwastewaters. Bioteehnol. Bioeng. 49: 601-610.
Brindle, K., Stevenson, P. 1996b. Nitrification in a bubbleless oxygen mass transfer membrane
bioreactor. Water Sci, Technol. 34: 261-267.
Brindle, K., Stevenson, p" Semmens, M.J. 1998. Nitrification and oxygen utilization in a membrane
aeration bioreactor. J, Membr. Sci. 144: 197-209.
Characklis, W,G" Marshall, K,C 1990, Biofilms. Wiley, New York,
Cote, p" Bersillion, J.L., Huyard, A. 1989, Bubble-free aeration using membranes: mass transfer
ana1ysis.1. Membr. Sci. 47: 91-106.
Cote, P., Bersillon, J.L., Huyard, A., Faup, G. 1988, Bubble-free aeration using membranes: process
analysis. J. Water Pollution Control Fed. 60: 1986-1992,
Dolasa, A,R., Ergas, S.J. 2000, Membrane biorcactor for cometabolism of trichloroethcne air
emissions. J, Environ, Engin, 126: 969-973.
176 S J Ergas
Ergas, S.J., Reuss, A. 2001. Hydrogenotrophic denitrification of drinking water using a hollow fiber
membrane bioreactor. Journal of Water Supply: Research and Technology-Aqua. 50: 161-171
Ergas, S.1., McGrath, M.S. 1997. Membrane bioreactor for control of volatile organic compound
emissions. J. Environ. Engin. 123: 593-598.
Ergas, S.1., Shumway, L., Fitch, M.W., Neeman 1.1. 1999. Membrane processes for biological
treatment ofcontaminated airstreams. Biotechnol. Bioeng. 63: 431-441.
Essila, N.1., Semmens, M.1., Voller, V.R. 2000. Modeling biofilms on gas permeable supports:
Concentration and activity profiles. J. Environ. Engin. 126: 250-257.
Fitch, M., Sauer, S., Zhang, B. 2000. Membrane biofilters: Material choices and diurnal loading
effects. Proceedings USC- TRG Conference on Biofiltration, University of Southern California,
October 19-202000, 83-90.
Freitas dos Santos, L.M., Pavasant, P., Strachan, L.F., Pistikopoulos, E.N., Livingston, A.G. 1997.
Membrane attached biofilms for waste treatment - Fundamentals and applications Pure AppL Chem
69: 2459-2469.
Gantzer, C. J. 1995. Membrane dissolution ofhydrogen for biological nitrate removal. In: Proceedings
ofthe 68th Annual Conference and Exhibition, Water Environment Federation, Research Symposium,
Miami Beach, FL, 49--{50.
Hartmans, S., Leenen, E.J.T.M., Voskuilen, G.T.H. Membrane bioreactor with porous hydrophobic
membranes for waste-gas treatment. In: Biotechniques for Air Pollution Abatement and Odour
Control Policies. Oragt A.1. and van Ham J. (eds.). Eisevier, Amsterdam, The Netherlands, 1992, pp.
103-106.
Keskiner, Y., Ergas, S.1. Control of ammonia and NO x emissions using a nitrifying membrane
bioreactor. In: Proceedings of the 94 th Annual Meeting & Exhibition of the Air & Waste Manage.
Assoc., 2001, Orlando FL.
Lee, K.C., Rittmann, B.E. 2000. A novel hollow-fibre membrane biofilm reactor for
autohydrogenotrophic denitrification of drinking water. Water Sci. Technol. 41: 219-226.
Livingston, A.G. 1993. A novel membrane bioreactor for detoxirying industrial wastewater: 1.
Biodegradation ofphenol in a synthetically concocted wastewater. Biotechnol. Bioeng. 41: 915-926.
Livingston, A.G. 1994. Extractive membrane bioreactors - a new process technology for detoxifying
chemical-industry wastewaters. J. Chem. Technol. Biotechnol. 60: 117-124.
Livingston, A.G., Arcangeli, J.P., Boam, A.T., Zhang, S.F., Marangon, M., dos Santos, L.M.F. 1998.
Extractive membrane bioreactors for detoxification of chemical industry wastes: Process
development. J. Membr. Sci. 151: 29-44.
Lund, L.W., Federspiel, W.J., Hattler, B.G. 1996. Gas penneability of hollow fiber membranes in a
gas-liquid system. J. Membr. Sci. 117: 207-219.
Parrkhania, M., Stephenson, T, Semmens, M.J. 1994. Hollow fibre bioreactor for wastewater treatment
using bubbleless membrane aeration. Water Res. 28: 2233-2236.
Parvatiyar, M.G., Govind, R., Bishop, 0.1'. 1996a. Biodegradation oftoluene in a membrane biofilter.
J. Membr. Sci. 119: 17-24.
Parvatiyar, M.G., Govind, R., Bishop, D.F. 1996b. Treatment of trichloroethylene (TCE) in a
membrane biofilter. Biotechnol. Bioeng. 50: 57-64.
Pressman, J.G., Georgiou, G, Speitel, G.E. 1999. Demonstration of efficient trichloroethylene
biodegradation in a hollow-fiber membrane bioreactor. Biotechnol. Bioeng. 62: 681--{592.
Pressman, J.G., Georgiou, G., Speitel, G.E. 2000. A hollow-fiber membrane bioreactor for the
removal oftrichloroethylene from the vapor phase. Biotechnol. Bioeng. 68: 548-556.
Reij, M. W., de Gooijer, K.O., de Bont, lA.M., Hartmans, S. 1995. Membrane bioreactor with a
porous hydrophobic membrane as a gas-liquid contactor for waste gas treatment. Biotechnol. Bioeng.
45: 107-115.
Reij, M.W. and Hartsmans, S. 1996. Propene remova1 from synthetic waste gas using a hollow fiber
membrane bioreactor. J. Appl. Microbiol. Biotechnol. 45: 730-736.
Reij, M.W., Hamann, E.K., Hartmans, S. 1997. Biofiltration ofair containing low concentrations of
propene using a membrane bioreactor. Biotechnol. Prog. 13: 380-386.
Reij, W.M., Keurentjes, J.T.F., Hartmans, S. 1998. Membrane bioreactors for waste gas treatment. J.
Rittmann, B.E., McCarty, P.L. 1980. Model of steady state biofilm kinetics. Biotechnol. Bioeng. 22:
2343-2357.
Semmens, M.J., Gulliver, J.S., Anderson, A. 1999. An analysis of bubble formation using
microporous hollow fiber membranes. Water Environ. Res. 71: 307-315.
Shumway, L. 1997. A membrane process for biological treatment of contaminated airstreams. M.S.
Thesis, Department of Civil and Environmental Engineering, University of Massachusetts, Amherst.
Sirkar, K.K. 1992. Other new membrane processes. In: Membrane Handbook. Ho, W.S.W. and Sirkar,
K.K. (eds). Chapman & Hali, New York.
Speitel, G.E.; McLay, D.S. 1993. Biofilm reactors for treatment of gas streams containing chlorinated
solvents. J. Environ. Engin. 119: 658--{)78.
Stephenson, T., Judd. S., Jefferson, 8., Brindle, K. 2000. Membrane Bioreactors for Wastewater
Treatment. IW A Publishing, London.
Togna, P. 2000. personal communication. EnviroGen Inc. Lawrenceville NJ.
Yamagiwa, K., Yoshida, M., Ilo, A., Ohkawa, A. 1998. A new oxygen supply method for
simultaneous organic carbon removal and nitrification by a one-stage biofi!m process. Water Sci.
Technol., 37: 117-124.
Yang, M-C, Cussler, E.L. 1986. Designing hoIlow fiber contactors. AIChE J. 32: 1910.
CHAPTER 7 COMBINED ADV ANCED OXIDATION
AND BIODEGRADATION
lohan W. VAN GROENESTIJN
1. Introduction
In existing biological technologies for waste gas treatment pollutants are transferred
to a water phase and subsequently biodegraded in a biofilm or suspension of micro-
organisms. These two processes impose two important prerequisites on biological gas
treatment: the pollutants should be readily soluble in water and readily
biodegradable. Bioscrubbers, biofilters and biotrickling filters encounter problems
with compounds with dimensionless Henry's Law's coefficients higher than 1 (e.g.,
alkenes, alkanes). In addition, the biological treatment of gases with low
concentrations of compounds with Henry's Law' s coefficients higher than 0.1 (e.g ..
many aromatic hydrocarbons) may be limited by low interphase mass transfer rates
(Groenestijn and Hesselink, 1993). Problems with biodegradability can be caused by
the fact the compounds need anaerobic conditions for their degradation (e,g.,
perchloroethylene) or are only aerobically degraded by co-metabolism (e.g.,
trichloroethylene). Some compounds are in principle biodegradable, but their
elimination in biofilters needs long start up periods. It has been reported that
elimination of certain organo-sulphur compounds only started 5 months after start up
of a biofilter (Cho el al., 1992).
These problems can be solved by a pre-treatment of the waste gas using
advanced oxidation technologies, such as photochemical, photocatalytical, and non-
thermal plasma technologies. This pre-treatment may be carried out in the gas phase
or after absorption, in a water phase. By generating hydroxyl radicals and oxygen
atoms in the gas, hydrophobic compounds can be partly oxidised, which mostly yield
compounds that are more hydrophilic and bettcr biodegradable. Gases with these
partly oxidised compounds can be biologically treated in biofilters or related
equipment at higher volumetric elimination capacities and higher stability than
untreated gases. The combination of a chemical and a biological system is expected
to be more cost efficient than chemi cal oxidation solely, since for a complete
chemi cal oxidation of compounds large amounts of energy are required.
Combined biotreatment and advanced oxidation in water is better known than
that in gases. Complex and recalcitrant compounds in water, such as dyes,
chlorinated hydrocarbons and pesticides can be degraded by photochemical processes
in combination with ozone or hydrogen peroxide, photocatalytical treatment, and
electron beam treatment. It depends on the composition of the water whether
chemical oxidation is used as a pre- or post-treatment in combination with biological
process. If a waste water contains readily biodegradable compounds at higher
concentrations than the recalcitrant compound, it is recommended to start the
treatment process with a biological process, to save energy or chemicals in the
chemi cal process. If the water mainly contains recalcitrant pollutants. chemi cal
oxidation can be the initial operation. A subsequent biological treatment step may be
required because of (1) effluent standards, (2) water reuse and (3) toxicity of products
of advanced oxidation.
179
C. Kennes and M. C. Veiga (eds.), Bioreac/orsfor Was/e Gas Trea/ment, 179-200.
Such reasoning can also be made for waste gas treatment. If the gas mainly
contains hydrophobic or poorly biodegradable target compounds, a partial oxidation
by a chemical process should precede a biological process. This sequence can lead to
the lowest costs for near complete mineralisation. If the mentioned compounds
represent only a small part of the target compounds, advanced oxidation may be used
as a post-treatment step. In this post-treatment process complete mineralisation can
be aimed at or only chemical modification (to change the odour properties of
molecules), depending on the specific emission requirements. Even a sequence of
biological treatment, advanced oxidation and a second biological treatment may be
advantageous.
In the sections below important advanced chemical oxidation technologies for
gas treatment are discussed. Almost all studies in this field aim at complete
mineralisation, as the technology is considered to be the sole unit operation to
de grade pollutants in a gas. There are only a few examples of studies which aim at
combinations ofbiological processes with advanced oxidation.
2. Photochemical treatment of gases
2.1 TECHNOLOGY AVAILABLE
Two different photochemical treatment processes can be distinguished: homogenous

photolysis and heterogenous photolysis (or photocatalysis).
In homogenous photolysis no catalysts are used. In this technology, ultra-
violet radiation with wavelengths lower than 190 nm is absorbed by 02 and volatile
organic compounds and breaks the atomic bonds, resulting in oxygen and
hydrocarbon radicals (direct photolysis). Direct photolysis of certain volatile organic
compounds can also take place with ultra-violet radiation of higher wavelengths.
Subsequently, the oxygen radicals can react with O2, yielding ozone, or with H20,
yielding hydroxyl radicals and hydrogen peroxi-radicals. The organic radicals can
react with O2 yielding RCOO radicals. The produced radicals and ozone are very
reactive and can oxidise volatile organic or inorganic compounds (indirect
photolysis). Volatile compounds are first partially oxidised and eventually, afier a
series of subsequent oxidation steps, to carbon dioxide and water. Typically, low-
pressure mercury lamps are used which emit UV radiation with wavelengths of 185
nm and 254 nm (Sint Annaland, 1996).
In photocatalysis, photons are absorbed by the catalyst and produce radicals,
which can react with organic compounds that have adsorbed onto the same catalyst
surface. If H20 and O2 are present hydroxyl, perhydroxy and superoxide radicals are
produced. The hydroxyl radical and the electron gap in the catalyst surface are
assumed to be the primary oxidants (see Figure 7.1). The volatile compounds are
subjected to a series of oxidations, leading to complete mineralisation. As the
produced organic intermediates easily stay adsorbed onto the catalyst, they are mostly
completely converted into carbon dioxide and water. The wavelength of the UV
radiation required for this process can be higher than that in the homogenous
photolysis process: 300-380 nm is suggested. Suitable catalysts are Ti0 2 , ZnO and
CdS, but it is generally accepted that Ti0 2 shows better activities in the destruction of
a variety of compounds (Nogueira el al., 1997). Among the different commercial
Combined oxidation-biodegradation 181
forms of Ti0 2, Degussa P25 has become a research standard due to its wel!-defined
nature and a high photocatalytic activity. Further improvements have been suggested
by using Ti0 2!V205/Pt mixtures (Sanchez et al., 1995).
In most of the gas/solid reactors Ti O2 is present as a coating. Cylindrical flow
reactors are used with Ti0 2 coated on the insi de wal!, while in the middle of the
cylinder a long cylindrical lamp is placed (Sakamoto et al., 1997). They are
alternatively cal!ed annular thin film reactors. The placement of coated bafflcs in the
cylinder may improve the performance (Raissi el al., 1998). Without the catalyst,
these reactors can also be used for photolysis.
Red"ctU
Oxidiser'
e-
k
~~
(9/-
~(I
;:.,
Conductor band 0..,
Photons < 380 nm
E(gap)
Valence bond
Figure 7.1. At the surface of Ti0 2 an electron is liberated by the action of photons. The free
electron and the electron hole both take part in reduction and oxidation reactions.
2.2 PHOTOCATALYTIC OXIDATION
Most of the studies on photochemical gas treatment deal with photocatalytical

processes, using Ti0 2 as the catalyst. The reason may be that the subject involves
relatively simple, safe and new technology. Almost ali studies described in journals
and conference proceedings were carried out on laboratory scale Of in pilot plants.
Only a few fuI! scale experiences are reported. Examples of successful experiments
are given below.
2.2.1. Examples ofsuccessful studies

Hal! and col!aborators (1997) used a reactor with Ti0 2 coated honeycomb monoliths
il!uminated by UV light to eliminate mixtures of formaldehyde, toluene and ammonia
in air in the 100-1000 ppbv range. Chapman and Mook (1997) were able to
photocatalytically eliminate 99% of the trichloroethylene (TCE) present in gas. In

experiments carried out by Hoang-Van et al. (1997) ali odour was removed from
gases containing pyridine, 2-furfural and butadione and it was found that mainly C02
was produced and no intermediate organic compounds. The experiments were
interesting as 2-furfural and butadione are emitted during cooling of foodstuffs
containing cereals or milk respectively, and pyridine is present in tobacco smoke.
The treatment of VOCs from tobacco smoke can be an interesting indoor application
of photocatalysis. A commercial scale photocatalytic reactor was successfully tested
for ethylene removal from indoor air in a concentration range of 80 to 1500 ppbv at
room temperature (Tompkins et al., 1997). Besides VOCs also inorganic pollutants
can be removed from indoor air, using photocatalytic reactors. S02, H2S, NOx , NH3
and CO were eliminated with efficiencies between 75% and 99% (Qinglin and
Chanjuan, 1995). According to Graf et al. (1997) the photocatalytic reactors can also
be used in plant growth facilities in spacecrafts. For this application the reactor
should be able to reduce the ethylene concentration to levels lower than 50 ppbv,
which was proved to be possible. A greater challenge is to remove methane from air
in space applications. It was studied by Selzer et al. (1997) using photocatalytic
reactors.
Many studies on photocatalytic elimination deal with NO x in gases. Ibusuki
and collaborators (1995) have evaluated that photocatalytic oxidation of several'
ppmv of nitric oxide (NO) and nitrogen dioxide (N02) in air to N02 and HN03 by
semiconductor particles like Ti0 2 proceeds very rapidly. They found that mixtures of
Ti02 and particles of activated carbon with high adsorptivity (affinity) for N02 can
oxidise most of NO and N02 completely to HN03, which is captured on the surface
of the particles. Addition of Fe203 or MgO could increase the catalytic activity. In
Japan there is a lot of interest to use Ti0 2 coated walls to clean ambient air in city
streets and buildings (e.g., parkings and toilets). An example is given by Murata et al.
(1997) who did tests with air purifying pavement blocks in form of photocatalytic
concrete blocks, made from cement that contains Ti0 2 in the upper layer. The blocks
can be applied to outdoor surfaces where they are exposed to sunlight for activation
and to rain for regeneration (wash away the accumulated products). In a simulation
experiment 80% of the offered NO was eliminated at concentration ranges of 50-
1000 ppbv.
Most VOCs can be oxidised in photochemical reactors, however, a few
exceptions exist. Tetrachloro carbon (Sănchez et al., 1997) and freons (chlorofluoro
carbons) (Weaver et al., 1997) are inert to chemical oxidation, as they lack
unsaturated bonds and a hydrogen atom that can be split off (removal of an hydrogen
atom is the first step in oxidation). Addition of an electron donor may lead to their
reduction in photochemical reactors. This was tested with trichlorofluoro methane
and the addition ofNaHC02.
2.2.2. Intermediate products

In photocatalytic oxidation the presence of intermediate products in the gas phase has
been reported frequently. During the photocatalytic oxidation of perchloroethylene
(PCE), phosgene can be detected. This product only accumulates at short gas
residence time in the photo-reactor. Other intermediates detected from PCE oxidation
are dichloro acetyl chloride and tetrachloro carbon (Murabayashi et al., 1997; Li et
al., 1997). Benoit-Marquie and coworkers (1997) detected six intermediate products
in the gas phase when photocatalytically oxidising butanol, but at the end, afier
prolonged reaction, only carbon dioxide and water. Oxidation of butylamine yielded
3 intermediates. Vorontsov et al. (1995) found acetaldehyde in the gas phase as an
intermediate product ofthe photocatalytic oxidation of ethanol, and acetaldehyde and
ethylacetate as a result of photocatalytic oxidation of diethyl ether.
Usually the intermediates stay adsorbed onto the catalyst and are further
oxidised. Ameen el al. (1997) revealed that isopropyl a1cohol tirst is adsorbed,
converted into acetone, which may desorbe or may be oxidised to acetaldehyde.
Adsorbed acetaldehyde was rapidly oxidised to acetic acid and further rapidly to
formaldehyde and formic acid, then slowly to carbon dioxide. The presence of
acetone in the gas phase during the photocatalytic oxidation of isopropyl alcohol was
earlier demonstrated by Ameen and collaborators (1995).
The higher solubility of intermediates was positive1y used in a process
developed by De1prat et al. (1997) in which an air flow containing 2J-dimethyl
pyrazine passed a Ti0 2 suspension in water, irradiated with 340 nm UV waves. 2,3-
dimethyl pyrazine, an odorous compound produced by food processing, was
eliminated from the gas phase, while (intermediate) products were retained in the
water phase. Fourteen products were identitied. Organic nitrogen was transformed
into ammonium ions, which were very slowly oxidised into nitrate ions.
2.2.3. Catalyst deaclivation

Although the photocatalytic process has shown high efticiency in the destruction of
VOCs in air streams, some researchers have observed catalytic deactivation, with
consequent changing of catalyst surface colour afier extensive continuous operation,
especially when working with aromatic compounds. This deactivation has been
associated to both absence of water vapour and adsorption of by-products onto the
Ti02 surface (Alberici and Jardim, 1998). Deactivation also has been reported in
photocatalytic treatment of gas containing triethylamine (Chen et al., 1997b).
Alberici and Jardim (1998) found that the toluene conversion in an annular thin film
reactor and a black light lamp with a radiation maximum at 365 nm decreased from
87% to 21% afier 150 minutes use. The authors assumed that this deactivation
process is due to aromatic intermediates formed during the photocatalytic process,
which remain tirmly adsorbed onto the catalyst surface. It was proved that addition of
ozone (concentration 1 mgll gas) prevents this deactivation .
2.2.4. Kinetics
The economy and space requirement of photocatalytic reactors in industrial
applications depend on the required gas residence time, which depends in turn on the
degradation rate of the target compounds. The rate of degradation of volatile
compounds in photocatalytic reactors with Ti0 2 as the catalyst depends on the light
intensity, the available surface area of Ti0 2, the concentration of the target
compound, the water vapour concentration, the oxygen concentration and the nature
ofthe compound.
Many researchers have found satisfactory removal efticiencies in photo-
reactors using gas residence times of a few seconds. Graf el al. (1997) were able to
eliminate ethylene down to a level of 50 ppbv within 1 to 3 seconds. Another study
indicated that more than 99% photocatalytic conversion of PCE could take place in
0.5-2 seconds. The degradation rate appeared to be directly proportional to the light
184 J. W van Groeneslijn
intensity to the power 1.2 and water vapour inhibited degradation (Hung and Yuan,
1998). Humidity appears to be important: Li et al. (1997) demonstrated that the effect
of water vapour depends on the nature of the compound. The conversion of PCE
might be inhibited by water vapour, but the oxidation of toluene was stimulated by
increasing relative humidities in the range 0-10%. Higher humidities did not have an
additional stimulating effect. According to the authors these different effects of
humidity on the conversion of toluene and PCE imply that thcir photocatalytic
oxidation mechanisms are different. The photocatalytic oxidation of toluene may be
through hydroxyl free radicals, which needs water molecules to initiate, while for
PCE the oxidation mechanism may be through chlorine free radicals, which does not
need water molecules. The water molecules can be adsorbed onto the Ti0 2 surface
and therefore block the oxidation reaction.
Lichtin and Sadeghi (1997) studied the effect of VOC concentration on the
degradation rate, using benzene as a model compound in the range of 90-12000
ppmv. First order kinetics were found: the rate of initial oxidation of benzene was
directly proportional with its concentration. In their test equipment the initial
oxidation took minutes, while the CO 2 production approached completeness in more
than 1 hour. This indicates the accumulation of intermediate compounds and
demonstrates that partial oxidation may be economically more attractive than
complcte oxidation. The same study revealed the independence of initial benzene
oxidation on O2 concentrations between 0% and 100%, and the increase of CO 2
production rate with increasing 02 concentrations. Sakamoto el al. (1997) a1so found
first order kinetics for toluene in the 200-1200 ppbv range. The removal of toluene
was always 75%, independently on the toluene concentration. As expected, the
removal efficiency was dependent on the catalyst surface area and the gas flow rate
through the reactor. Although first order kinetics is common in gas phase
photocatalysis, exceptions exist. Photocatalytic oxidation of isopropy1 alcohol shows
first order kinetics up to concentrations of 40 ppmv, but near independence was
found at concentrations above 60 ppmv (Ameen et al., 1995). This behaviour
supports Langmuir-Hinse1wood kinetics.
The effect of light intensity was studied in a detai1ed way using TCE as a
model compound (Upadhya and Ollis, 1998). At high concentration ranges (1000
ppmv) the amount of VOC converted per g Ti0 2 per minute appeared to be directly
proportional to light intensity in a linear dependency, but in such way that low light
intensities were more efficient. At low concentrations (60 ppmv) such linear relation
could not be found.
2.2.5. Energy use

Besides elimination rate, the energy required for photocatalytic oxidation determines
the costs of industrial scale photo-reactors. In scientific studies the quantum yield is
used, which is defined as the moles of pollutant destroyed per Einstein (= 1 mole of
photons). However, figures on energy consumption are seldom reported, as
measurement of quantum efficiency is complex. With regard to the energetics of the
photocatalytic process, the concept of electrical energy per order of reduction of thc
concentration of target compounds (EE/O), Bolton Criteria, is often used (Bolton el
al., 1995; Raissi el al., 1998). EE/O is defined as the input of electrical energy
required by a waste treatment process in order to reduce the concentration of a target
compound in the waste stream by one order of magnitudc (factor 10). In
photocatalysis, the amount of energy used seems to be independent of the

concentration of the compound: a reduction from 100% to 10% requires the same
amount of energy as a further reduction from 10% to 1%, which is the result of first
order kinetics. The EE/O was deve10ped for water and can be expressed as kWh per
1000 US gallons of water (Bolton et al., 1995). The EE/O assumes first order kinetic
behaviour and provides a practical overall energy efficiency figure of merit to allow
comparison of different processes for degradation of a given pollutant. The EE/O
times the electricity price gives the energy costs connected to the UV lamps of
photoreactors. Raissi and collaborators (1998) alternatively proposed the use of
EP/O, based on electric power, for gas treatment.
One of the few studies on energy use indicated that TCE can be converted
with a quantum efficiency of 0.1-0.27 moI/Einstein (Cabrera et al., 1997), while
Hung and Yuan (1998) reported a quantum efficiency of 0.052 moi/Einstein in the
photocatalytic mineralisation ofPCE. This implies that a very large part ofthe energy
generated, is not used for the conversion process. Moreover, PCE and TCE are
relatively easily photo-degradable because of the reactive double carbon bond. Many
other organic compounds will demand even more energy for degradation.
2.3 PHOTOL YSIS
Although the number of studies on photolytic conversion of gaseous compounds in

reactors is much lower than that on photocatalytic oxidation, the process may be
more interesting for combining it with biofiltration. In photolysis the partial oxidised
intermediates cannot adsorb onto a catalyst surface and stay in the gas phase.
Sint Annaland (1996) compared photolysis and photocatalytic oxidation of 1-
butene. Irradiation of air containing 100 ppmv l-butene with 185 nm UV waves
(photolysis) revealed that more than 90% of l-butene was photolysed into at least 9
organic intermediates and only 5% into CO 2. The quantum yie1d was 0.5 (molecules
l-butene converted/produced photons) (Kok, 1996; Sint Annaland, 1996).
Photocatalytic oxidation of l-butene, however, yielded 85-95% CO 2 , but also some
CO. The initial oxidation of l-butene by photolysis was 2-5 times faster than the
photocatalytic oxidation. According to Sint Annaland a higher photolysis rate
compared to the photocatalysis rate is exceptional: l-butene absorbs 185 nm UV
radiation very well and is very reactive with ozone. In an economic analysis it was
calculated that the quantum yield in photocatalysis and photolysis of l-butene is 50-
100 times too low for a cost effective gas treatment technology. However, the range
mentioned is strongly dependent on reaction conditions. The energy requirement for
the treatment of gas containing 100 ppmv l-butene is estimated 300 kWhlI000 m 3 .
This requirement will be lower at lower concentrations, which means that photo-
conversion may be cost efficient for ppbv range pollutants, which is relevant for
indoor applications and in the case of strict emis sion standards.
Besides its application in reactors photochemical conversion of volatile
organic compounds (VOCs) is studied well in atmospheric chemistry. In a review on
gas-phase tropospheric chemistry, Atkinson (1990) gives the reaction rates of many
VOCs with hydroxyl radicals at 298 K. It appears that the reactivity of alkanes
correlates very well with their molecular weight: higher alkanes react faster. In
particular, a striking difference exists between methane and the other alkanes.
Methane reacts orders of magnitude slower. Alcohols react faster than alkanes and
alkenes react faster than homologue alcohols. The same posltlve correlation of
reaction rate and molecular weight was found for these chemical classes. A same
effect was found for non-oxygenated (real hydrocarbon) aromatic compounds. The
addition of one or more hydroxyl groups to a molecule makes aromatic compounds
more reactive with hydroxyl radicals. For instance, phenol is 20 times more reactive
than benzene. This means that the accumulation of such oxidation intermediates is
unlikely. Neverthe1ess, phenol can be detected during photochemical oxidation of
benzene, while benzaldehyde, o-cresol, m-cresol and p-cresol can be measured as
products from toluene oxidation and photochemical conversion of o-xylene produces
o-tolualdehyde, 2,3-dimethylphenol and 3,4-dimethylphenol.
A few researchers combine photocatalysis and photolysis. They use UV
radiation with a relative high wavelength, Ti02 as catalyst and addition of ozone to
promote indirect photolysis and to prevent catalyst deactivation. Ozone can be
produced in situ or by remote ozone generators. Alberici and Jardim (1998)
investigated gas-phase photocatalytic oxidation of different classes of VOCs,
including alkanes, ketones, a1cohols, chlorinated compounds and aromatic
compounds, using an annular thin film reactor with a black light lamp (maximum
radiation at 365 nm) and Ti0 2 as catalyst. For all organic compounds tested, catalytic
deactivation was not observed, except for toluene. Co-injection of ozone could
prevent this deactivation. The addition of ozone had only clear advantages for
toluene (to prevent deactivation) and for pyridine. The destruction of 620 ppmv
pyridine was increased from 16 to 68% by addition of 1 mg 0 3/1 air, produced by a
high voltage ozone generator. Similar experiences of increased elimination rates by
addition of ozone in photocatalytic reactors can be found in Obee and Hay (1997)
and Ollis (1995).
Other researchers combine photolysis and high temperatures: the photo-
thermal process. Chen et al. (1997a) proved that at 600 °C BTEX mixtures could be
treated within 10 seconds in a reactor with UV radiation with a wave1ength of 185
and 254 nm. 62%, 87%,94% and 96% destruction was observed ofbenzene, toluene,
ethylbenzene and xylene respectively.
3. Treatment of gases using a non-thermal plasma
A plasma is a mixture of free moving electrons and positively charged ions. It can be
generated in a medium between two electrodes. It already can be produced if a high
voltage AC (10-30 kV) is applied for a very short time (1-10 milliseconds). During
these short pulses the medium will not be heated, but the plasma is nevertheless
produced. It is called a non-thermal plasma. If the electricity pulses are too long, the
medium acts as a conductor and is substantially heated. As a result high energy costs
are involved. Non-thermal plasmas are an excellent source of gas-phase free radica1s
(O, OH and H) and other active species useful for destroying pollutants (Cal and
Schluep, 2000). Therefore non-thermal p1asmas are the logic option for gas
treatment. They can be generated by die1ectric barrier discharge (DBD), corona
discharge or an e1ectron beam.
According to Cal and Schluep (2000) using non-thermal plasmas for gas-
phase pollution control shows much promise, but is still in its early stages of research
and development.
3.1 DIELECTRIC BARRIER DISCHARGE AND CORONA DISCHARGE
Dielectric barrier discharge (DBD) or corona discharge utilises a dielectric material

between the discharge gap and one of the two discharge electrodes. Typically, a
material with a high dielectric strength (V /mm) and a high dielectric constant, such as
glass or quartz, is used as the dielectric. When the potential across the gap reaches
breakdown voltage, the dielectric acts as a stabilising material leading to the
formation of a large number of micro-discharges of short pulses that are statistically
spread over the discharge gap (Lee and Chang, 1998). In the gas phase electrical
discharges are produced, known as streamers. The high energy electrons present at
the leading edge of these propagating streamers can generate the radical species
necessary to interact with the pollutant molecules. A possible reactor design is a
cylindrical reactor consisting of a centre discharge electrode surrounded by a
dielectric barrier (e.g., a quartz tube) and an outer electrode (Cal and Schluep, 2000).
Altematively, the space between the inner and outer electrode can be filled with a
packed bed of pellets made of a dielectric material.
A summary of research on VOC destruction by DBD or corona discharge
carried out by various researchers can be found in Cal and Schluep (2000). Studies
are available on methylene chloride, hexane, methyl-ethyl ketone, cyclohexane,
methane, aromatic hydrocarbons, TCE, PCE, trichloroethane, carbon tetra chloride
and carbon tetrachloride. Efficiencies and reaction rates differed as a result of
different reaction conditions and nature of the compound. Interestingly, methane
oxidation proceeded much slower than benzene and toluene oxidation, a similarity
with photolysis.
In addition, a wide range of volatile compounds that can be oxidised in a
pul sed corona discharge reactor was studied by Grothaus et al. (1995). It was proved
that C7Hg, CChF 2, CH2Ch, CH3CCh, Nh SF6, C2F6, CF 4 and NO x can be oxidised
successfully. At least 95% conversion ofNO x , S02, CO and benzpyrene was attained
by Abolentsev et al. (1997). Odorous VOCs and aerosols from a fish smoking plant
were removed in a two-stage gas treatment plant comprising of an electrostatic
precipitator and a pulse corona discharge reactor. 90% of the VOCs were removed
(Chiraevski et al., 1995). The same authors tried pulse corona discharge to remove
benzpyrene from gases from aluminium plants. 100 W power was required to remove
95-99.8% benzpyrene from 50 m3 gas/h, which can be regarded as cost effective.
The energy used in oxidation with non-thermal plasma is mainly put into
generating highly energetic electrons that attack the target pollutants and matrix gas
components (which products can also react with the pollutants). This leads to low
energy consumption for purification of low concentrated waste gas streams (Sj5berg
et al., 1997b). DBD has been studied as a possible technology to remove S02, NO,
NH3 and VOCs from gases. Examples are given below.
Sj5berg et al. (1997a) studied the oxidation of toluene in gases using a
commercially available dielectric barrier discharge unit. They found that the removal
efficiency of toluene in the plasma was higher at higher energy input, i.e., higher
applied voltage and frequency, low gas flow rates and higher water vapour content.
Identified oxidation products at low plasma energy inputs were benzaldehyde,
benzyla1cohol, cresol as well as ring-opening products as acetaldehyde, formic and
acetic acid and carbon dioxide. The identified oxidation products were better water
soluble and better biodegradable than toluene. In addition, solid products were
observed, probably as a result of polymerisation reactions. Sjoberg et al. (1997b)

suggest that the major oxidation mechanism in these systems is the attack of OH
radicals on the toluene molecule. The OH radicals are produced by electron impact
dissociation of water. Electron impact dissociation of toluene is, however, unlikely
because the probability of toluene-electron collision is low in diluted gas streams
« 1000 ppmv).
Since energy consumption of the plasma increases exponentially when
approaching complete removal of VOCs, it was suggested to use a biological post
treatment step, rather than put more energy in the system to reach complete chemical
oxidation (Sjoberg et al., 1997a). The authors estimated the use of energy by
different types of non-thermal plasma technologies and compared the data with those
found in waste gas incineration. For this purpose they combined literature data with
own calculations. The results are summarised in Table 7.1.
Table 7.1. Comparison of energy use by different processes for 90 % removal of VOCs
from dilute waste gases (1 g VOC/m 3)
Process Energy consumption (kW/lOOO m3 .h)

Non-thermal plasma
- Barrier discharge 1-1300
- Corona discharge 4-154
- Electron beam 1-24
Incineration
- Regenerative 2-30
- Catalytic 2-30
The wide range of energy consumption given for non-thermal plasma

processes reflects the lack of experience with this technology. Nevertheless, the
authors expect that the working range of large-scale non-thermal plasma plants is less
than 1 g VOC/m 3 . Above this concentration regenerative and catalytic incineration
work more efficient. This suggested working range combines well with the optimal
concentration range of biologic al treatment systems (lower than 1-5 g VOC/m 3,
dependent on the nature of the compound) (Groenestijn and Hesselink, 1993). This
implies that combinations of non-thermal plasma and bio1ogical processes for waste
gas treatment may be cost effective at low concentration ranges.
In the same period a similar study was carried out by Lee and Chang (1998),
but using p-xylene as a model compound. By DBD with a voltage of 20 kV they
found C02, H20, CO, C2H2 and C2H4 as end products. The inlet concentration of
xylene was 500 ppmv. Their study was not focussed on the production ofhydrophilic
VOCs from xylene, which appearance may be assumed at lower energy use.
Yang and collaborators (1998) compared the destruction efficiencies of
toluene, ethyl acetate and 2-butanone under similar conditions. A rotating spark gap
pulsed power supply was used to produce pulse plasma for elimination of pollutants
in air at a concentration of 50 ppmv. The plasma reactor destroyed more than 96% of
the toluene in the air stream at an electric field strength of24 kV/cm and a residence
time of 3.5 seconds. Under the same conditions 85% of ethyl acetate and 2-butanone
were destroyed. It indicates that, like photo(cata)lysis, each compound reacts at a
different specific rate. This implies that for mixtures of pollutants in air a high energy
use and long gas residence time may be required ta treat the gas according ta
emis sion standards. The design of the system will be most dependent an the
compound with lowest reaction rate. In this respect it may bc interesting again ta
study a possible complimentarity between oxidation in non-thermal plasmas and
biological processes with respect to the ease ta oxidise different classes of VOCs.
Inorganic compounds can also be oxidised in non-thermal plasmas produced
by DBD. Double dielectric barrier discharge has been demonstrated effective at
oxidising NO into N02 and HN0 3 (Federle et al .. 1997). The much higher solubility
of the products in water made it possible to treat the gas with a more conventional
gas treatment system: a horizontal wet scrubber utilising hydrogen peroxide and
sodium hydroxide to absorb the products and convert them to nitrate salts. The
authors did not consider a biological post treatment system, but it is expected that an
absorber and denitrification re actor may be an interesting option to treat these gases
with NO (e.g. fiue gases). The authors developed the system for diesel-powered
generators and found a NO x elimination efficiency greater than 70%. The NO
conversion efficiency was 17 eV per NO molecule oxidised and the overall system
efficiency was 40 eV per NO x molecule removed. The power consumption of the
N Ox removal process was estimated to be less than 10% of the total power produced
by the generator.
Slightly different results were found by Yang el al. (1997): 130 ppmv NO in
air was converted into N2, O2 and N0 2 (not HN0 3 !) in packed beds with dielectric
materials (barium titanate) in which non-thermal plasmas were produced by electric
discharges. The 40 ppmv N0 2 produced from 130 ppmv NO was furthcr removed
from the gas phase in a post treatment step Ca caustic wet scrubber).
A humid waste gas containing S02 and NH 3 can be cleaned within 2 - 6
seconds in a re actor with pulsed corona discharge. In such reactor particles of
(NH4)2S04 were produced (Chang and Choi, 1998). Another study describes the
conversion of S02 in fiue gas into S03, which precipitate onto the fiy ash (Amirov et
al., 1995). The authors found that the presence of particles in the gas phase
stimulated the reaction efficiency and that the addition of water vapour reduced the
energy costs. One Wh is required ta remove 12 ppmv from one Nm 3 flue gas. A
small pul sed corona streamer pilot plant for 200 Nm 1 flue gas per hour was tested by
Mattachini et al. (1995). A gas residence time of 0.6-0.7 s seemed to be sufficient.
3.2 ELECTRON BEAM DISCHARGE
The e1ectron beam technology is based on an electron gun that shoots high energetic
electrons to a target object. The electron gun comprises an electron source, i.e .. a
glow cathode, and an acceleration part that gives the electron a high speed
(comparable with a TV or an electron microscope). The electron gun, or EB (electron
beam) generator, is made vacuum to avoid hindrance of electron movement. The
electrons leave the EB generator via a window, made of thin but solid material, e.g.,
aluminium or titanium foiI. By interaction and collision with matter on their pathway,
the high-energetic electrons generate a multitude of exited molecules. X-rays are a
secondary effect of electron beam discharge. This secondary radiation creates
re active species as well, but also requires specific safety measures connected to this
technology.
VOCs and inorganic volatile compounds can be oxidised by electron beam

discharge. An example of VOC removal from gases is described by Prager et al.
(1995). Chlorinated hydrocarbons (C2 compounds) were stripped from polluted
groundwater and 1000 Nm3 gas/h was fed into a self-shielded electron beam unit and
irradiated with a beam of 190 keV electrons with a power up to 11 kW. 90% of
organic chlorine was removed. The out coming gas was filtered with lime and
residual ozone was removed by an additional charcoal filter. However, produced
chloro-acetic acids in a concentration of 0.1 mg/m 3 were stiU present as traces in the
effluent gas. The total treatment costs were estimated 0.33 DM/m3 treated
groundwater. According to Penetrante et al. (1995) for a variety of VOCs
(chlorinated hydrocarbons, toluene, benzene and methanol) electron beam processing
is more energy efficient than either pul sed corona or dielectric barrier discharge
processing.
Vitale el al. (1996) made a comparison of EB elimination of toluene,
chlorinated compounds and freons from gases in a concentration range of 1-3000
ppmv. The authors concluded that many chlorinated compounds were more easily
decomposed than toluene. For instance, for 99% removal ofTCE 10 eV per molecule
were required, while 99% elimination of toluene required an amount of energy equal
to 170 eV per molecule. Freon 113 was even more stable: 680 eV/molecule were
required for 99% removal. The analyses of decomposition products revealed that at
least eight different VOCs were produced from 1,1,I-trichloroethene, such as mono,
di and tri chloro acetyl chloride. The latter two products were also detected in EB
degradation of TCE.
EB units up to 1 MW are commercially available (Vitale et al., 1996). A full-
scale combination of a photochemical process and electron beam collision was
applied to treat odorous gases from a food factory. 75% reduction of odour was
found, which was satisfactory in this specific case (Anonymous, 2000).
The use of EB for flue gas treatment already has been studied since 1972
(Sato et al., 1992). Both S02 and NO x can be removed, yielding H2S0 4 and HN03 as
products. An example of flue gas treatment can be found in Deminsky et al. (1995).
According to the authors NO can be converted to N0 2 and HN0 2 or HN03 using
electron beam discharge. The power requirement seemed to be moderate.
Chmielewski et al. (1993) reported the use of 100 kW for treatment of 20000 m3/h
flue gas from a power plant and Doi et al. (1993) required maximum 15 kW for
treatment of 1000 m3/h flue gas from municipal waste incinerators.
4. The use of advanced oxidation technologies in water treatment
The number of studies on advanced oxidation technologies in water is larger than that
in gases. AlI technologies described above can be applied in a water phase. In water
treatment, the combination with biological technologies is considered more
frequent!y.
In 1994 the Dutch Government has made an overview on the possibilities of
chemi cal oxidation technologies for water treatment. A broad view on the effects of
such treatment was presented, e.g., the changed eco-toxicology of water and the
changed BOD/COD ratio, anticipating on biological post-treatment. In this study the
advanced oxidation technologies were only represented by photochemical processes
Combined oxidalion-biodegradation 191
(Bijsterbosch, 1994). Experiences and future opportullltJes of three specific

photochemical processes were discussed: (1) Ozone/UV oxidation (photolysis in
combination with ozone addition); (2) photolysis in combination with hydrogen
peroxide addition; and (3) photocatalysis. Ozone/UV oxidation is used at full scale
for the elimination of aromatic compounds, chlorinated aromatic compounds and
chlorinated aliphatic compounds. It is used to remove dyes from textile waste water,
to treat paper mill waste water, to remove cyanides from water and to treat spent
cooling water containing biocides and AOX (absorbable halogenated organic
compounds). Combined hydrogen peroxide/UV oxidation can be used to eliminate
cyanides, phenol, other aromatic compounds and chlorinated aliphatic compounds. It
can be used for the treatment of cooling water, polluted ground water, landfill
drainage water, drinking water (at present mainly sole ozone or UV treatment) and
industrial waste water (paper mill water, water containing phenol, organo-CI, dyes,
cyanides and pesticides). This technology is applied at full scale, but in less extent
than ozone/UV oxidation. Photocatalytic oxidation of water hardly knows any full-
scale applications. It can be used to eliminate aromatic compounds, pesticides,
organo-CI compounds and dyes from industrial waste water and contaminated
groundwater.
Ti0 2 is the most popular catalyst again and it can be used in form of a coating
on a reactor wall, similar to gas phase treatment or in sus pension in the water phasc.
Instead of artificial UV radiation, sunlight can be used to irradiate plates coated with
Ti0 2 on which a thin layer of water is flowing (Wang, 2000). Nevertheless, the low
quantum yield of photocatalytic processes did not yet stimulate its practic al
application. Cost effective application is waiting for improvement of the catalysts.
The number of studies/applications of photocatalytical treatment of water is still
increasing. Examples of recent studies are given below.
In particular, the decolourisation of textile waste water gets much attention. Li
and Zhao (1999) used biologically pre-treated textile waste water in laboratory scale
experiments with photocatalysis in Ti0 2 suspensions. To retain the expensive catalyst
partic1es, a membrane filtered the suspension and the slurry was recycled to the
photo-reactor. Complete decolourisation was obtained and 90% removal of
nonbiodegradable residual COD in 10-20 hours, which can be regarded as rather
long. Their method is fit for waste water containing reactive dyes. In a review paper
on treatment of textile waste water, Vandevivre el al. (I998) state that UV light has
been tested in combination with hydrogen peroxide or solid catalysts for the
decolourisation of dye solutions. While the UV/hydrogen peroxide process appeared
too slow, costly and little effective for potential full-scale application, the
combination UV ITi0 2 seems more promlS1l1g. The authors recommend
photocatalytic oxidation as a post-treatment process, after a biological process,
chemical precipitation and ideally after ozonation. Commercial mobile pilot plants
comprising ozonation, UV treatment and hydrogen peroxide addition are available.
Hung and Yuan (2000) tried a new combination of oxidation technologies:
hydrogen peroxide assisted photocatalysis, which was successful to remove sodium
dodecyl sulphate from groundwater. 95% of this surfactant was removed and it was
proved that the addition ofhydrogen peroxide stimulates the degradation.
Another novel combination of advanced oxidation technologies was recently
developed by Stock and coworkers (2000): the combination of sonolysis and
photocatalysis for textile dye degradation. The azo dyc naphtol blue black was
degraded using a high-frequency ultrasonic generator and UV photocatalysis. An

additive effect on degradation rate was observed when the sonolysis and
photocatalysis experiments were carried out in simultaneous or sequential manner.
Sonolysis is effective for inducing faster degradation of the parent dye. while Ti0 2
photocatalysis is effective for promoting mineralisation.
The use of electron beam technology for water treatment is described a.o. by
Sato el al. (1992), Kuruck el al. (1995), Cleland el al. (1996) and Getoff (1996).
Interestingly. Getoff (1996) observed the production of aldehydes and carboxylic
acids from TCE. These products were further decomposed at higher doses of
radiation. The author did not consider biodegradation as a way to further convert the
products mentioned.
EB decolourisation of dyes is more energy efficient than oxidation using
ozone and EB is more effective than UV radiation in case the water is highly turbid.
However, the use of EB for disinfection of cIear water is less energy efficient than
photocatalysis.
S. Combined advanced oxidation and biological treatment of gases
Several authors have suggested or even proved that combination of different

oxidation technologies can be beneficial in specific cases. The combination should
use complimentary properties of the different technologies, e.g., initial oxidation
versus mineralisation of intermediates, hydrophobic compounds versus hydrophilic
compounds, oxidation of compounds of different chemi cal cIasses, and
biodegradable compounds versus persistent ones.
As exhibited above, many of the studies on gas phase advanccd oxidation are
focussed on complete mineralisation, and many of the studies mention the high costs
involved for this complete oxidation. Therefore a partial advanced oxidation of
pollutants in gases followed by biological treatment may be interesting. By partial
oxidation many hydrophobic hydrocarbons and NO can be converted inta hydrophilic
and readily biodegradable compounds, which reduces the costs for biological waste
gas treatment or makes biological treatment possible. In addition. partial oxidation
can reduce the costs of advanced oxidation.
As far as known by the author of this chapter only two studies deal with this
combination. The work of Sjoberg el al. (l997a, 1997b) on treatment of waste gases
in non-thermal plasmas is specifically focussed on production of intermediates and
minimisation of energy consumption. with the intention to apply biological post-
treatment. According to the authors energy consumption of the plasma increases
exponentially when approaching complete removal of VOCs. At low plasma energy
inputs, however, intermediate oxidation products dominate which efficiently can be
mineralised in a biological treatment unit. As described in earlier sections they used
toluene as a model compound and found aromatic and aliphatic alcohols. aldehydes
and acids as oxidation products. These oxidation products are more hydrophilic and
are expected to be transferred from gas to water faster than toluene, which may have
positive consequences for biological post-treatment systems.
A second study was carried out by Groenestijn el al. (1994) who used styrene
as model compound and combined photochemical reactors and biofilters. First the
unit operations were studied separately. Photolysis of styrene in air yielded
benzaldehyde and traces of benzoic acid. Interestingly, the reaction continued in the
effluent gas collection vessel outside the photochemical reactor. Probably radicals
and instable intermediates produced in the UV reactor stiU reacted with styrene after
leaving the reactor. This not -optimi sed reactor system consumed 15 k Wh per 1000
m 3 gas containing 100 mg styrene/m 3 and oxidised styrene for 80%. This high energy
consumption may be lower after optimisation ofthe photochemical reactor.
Besides photolysis also photocatalalysis was studied. For this purpose a
closed system was used with a gas recycling loop and a reactor with a Ti0 2 coating
and a UV lamp. Styrene was only eliminated if both the UV light and Ti0 2 were
present in the reactor. Immediately after the UV irradiation was started styrene
disappeared, while CO 2 was only slowly released over a period of 135 minutes. As
expected, eventually 8 moi CO 2 were produced per moi eliminated styrene. This
means that complete oxidation takes a lot oftime and energy.
To prove that products from partial oxidation of styrene can be converted in
biofilters faster than styrene, laboratory scale biofilters filled with compost were
tested with different gases containing one organic poUutant: styrene, benzaldehyde,
acetophenone or I-phenylethanol. The elimination capacities found after sufficient
adaptation time can be found in Table 9.2.
Table 7.2. Elimination capacities of different aromatic compounds in a compost biofilter
Compound Elimination capacity (g VOC/m 3 filterbed .h)

Styrene 40
Acetophenone 60
Phenylethanol 100
Benzaldehyde 100-120
The results confirm that styrene oxidation products can be e1iminated at a

higher rate in biofilters compared with styrene.
The combination of photocatalysis and biofiltration was tested by adding
humidified air containing 500 mg styrene/m3 at a rate of 60 l/h to a photoreactor and
a biofilter in series. Expressed as carbon, the influent gas contained 38 mmol styrene-
C/m3, the gas after the photoreactor contained 15 mmol styrene-C, 15 mmol
(produced) CO 2-C and 5.3 mmol benzaldehyde-C per m3 and after the biofilter the
gas contained only produced CO 2-C (33 mmol/m\ The experiment, which ran 17
days, shows that the mechanism of partial oxidation by photocatalysis and
subsequent complete oxidation of the products works. The energy consumption of
the non-optimised photoreactor was 800 kWh per 1000 m3 treated gas. Further
research should be directed to maximise the production of benzaldehyde and
minimise energy use and gas residence time in the photoreactor.
6. Conclusion and future opportunities
It can be concluded that the technology of advanced oxidation of poHutants in gases

in stiH in deve1opment. The technology seems promising with respect to variety of
194 J. W van Groenestijn
compounds that can be eliminated, neverthe1ess, a very limited number of full-scale

applications can be found. This may be the result of uncertainties about the costs
involved. In particular, the power demand can be high. Estimations range from
optimistic to very pessimistic. In photolysis and photocatalysis the quantum
efficiency is still too low. The latter technology is waiting for better catalysts. Only
few researchers realise that partial oxidation of VOCs and combination of other
technologies can be a way to minimise the energy costs and size of advanced
oxidation equipment. Post-treatment may be chemi cal scrubbing, adsorption or
biodegradation. In this view more research can be recommended on how to use
energy and reactor space efficiently to oxidise compounds just sufficiently to make
them fit for further biological degradation. From this future work we may learn
which of the advanced oxidation technologies introduced in this chapter is most
efficient with respect to costs and ability to convert a variety of VOCs. The first
impression is that photolysis may stay too energy inefficient, that photocatalysis
shows problems with desorption of intermediate compounds and that non-thermal
plasmas may be most promising.
The application of such combined systems may be particularly sought in
waste gases containing very hydrophobic compounds such as alkenes. For waste
gases with aromatic compounds it depends on the specific situation. At present many
alternative biological systems to cope with aromatic compounds are in development.
The combined system may also be applied for gases containing highly chlorinated
compounds, which cannot easily biodegraded under aerobic conditions. Conversion
of NO into N02 may be interesting for the development of a biological DeNOx
process.
Combined advanced oxidation and biological treatment of gases seems
promising, but the deve10pment is stil! in its infancy.
Abbreviations
AOX Absorbable halogenated organic compounds

BOD Biochemical oxygen demand
BTEX Benzene, toluene, ethyl benzene and xylene
COD Chemical oxygen demand
DBD Dielectric barrier discharge
DeNOx RemovalofNO.
EB Electron beam
EE/O Electrical energy per order of reduction
EP/O Electrical power per order of reduction
PCE Perchloroethylene
TCE Trichloroethylene
UV Ultra violet
VOC Volatile organic compound
Acknowledgements
The author gratefully acknowledges Jan WilIem Assink and Hennan Kok for advices and information
on advanced oxidation technologies. This chapter was written with support from TNO Environment,
Energy and Process Innovation, Apeldoorn, the Netherlands.
References
Abolentsev, v., Korobtsev, S., Medvedev, D. and Shiriaevski V. 1997. Chemical process in
heterogeneous pulsed corona discharge for pollution control. In: Proceedings Fourth International
Conference on Advanced Oxidation Technologies for Water and Air Remediation. Orlando, Florida,
Naval Surface Warfare Center (Publisher), Maryland, USA: 63.
Alberici, R.M. and Jardim, W.F. 1998. Gas-phase destruction of VOCs using TiO~/UV and
Ti02/O)UV. J. Adv. Oxid. Technol. 3: 182-187.
Ameen, M., Annapragada, R.V. and Raupp, G.B. 1995. Photocatalytic oxidation kinetics of dilute
isopropyl alcohol in air. In: Proceedings World Environmental Congress. Science & Technology
Integration Inc. (Publisher), London, Ontario, Canada: 353.
Ameen, M.M., Burrows, V.A. and Raupp, G.B. 1997. In situ infrared spectroscopy oftitania thin films
during photocatalytic oxidation of dilute isopropyl alcohol in air. In: Proceedings Third International
Conference on Ti02 Photocatalytic Purification and Treatment of Water and Air. Orlando, Florida,
Naval Surface Warfare Center (Publisher), Maryland, USA: 29.
Amirov, R.H., Asinovsky, E.I., Desyaterik, Y.N., Samoilov, I.S., Belyaev, V.V., Kropp, L.1. and
Zykov, A.M. 1995. Flue gas e1ean up utilizing neutralization property of fly ash and and nanosecond
corona discharge. In: Proceedings World Environmental Congress. Science & Technology Integration
Inc. (Publisher), London, Ontario, Canada: 346.
Anonymous. 2000. Aerox-injector beperkt geuremissie. NieuwsbriefMilieutechnologie. 12: 12.
Atkinson, R. 1990. Gas-phase tropospheric chemistry of organic compounds: a review. Atmospheric

Environ. 24A(l): 1-41.
Benoit-Marquie, F., Wilkenhoner, U., Andre, L., Puech-Costes, E., Braun, A.M., Oliveros, E. and
Maurette, M.T. 1997. Photocatalytic degradation of organic substances in the gas phase: intermediates
and mechanisms of reaction using Ti02 supported. In: Proceedings Third International Conference on
Ti0 2 Photocatalytic Purification and Treatment of Water and Air. Orlando, Florida, Naval Surface
Warfare Center (PubIisher), Maryland, USA: 148-149.
Bijsterbosch, J. W. 1994. Chemische oxydatietechnieken: een overzicht van de stand der techniek,
RIZA (Publisher), Lelystad, The Netherlands. Report 94079.X.
Bolton, J.R., Stefan, M.1. and Hislop, K.A. 1995. Mechanisms and efficiencies in the
photodegradation of organic pollutants in aqueous solution. In: Proceedings World Environmental
Congress. Science & Technology Integration Inc. (Publisher). London, Ontario, Canada: 229.
Cabrera, M.I, Alfano, O.M. and Cassano, A.E. 1997. Quantum efficiencies of the photocatalytic
decomposition of trichloroethylene in water, a comparative study for different varieties of titatnium
dioxide catalysts. In: Proceedings Third International Conference on TiO~ Photocatalytic Purification
and Treatment of Water and Air. Orlando, Florida, Naval Surface Warfare Center (Publisher),
Maryland, USA: 37-38.
Cal, M.P. and Schluep, M. Destruction of benzene in a dielectric barrier discharge plasma reactor. In:
Proceedings of the 93'd Annual Conference & Exhibition of the Air & Waste Manage. Assoc., 2000,
Salt Lake City, USA.
Chang, H. and Choi, Y. The scavenging of S02 in flue gas by injection of NH3 and corona plasma
reaction. In: Proceedings of the 91st Annual Meeting & Exhibition of the Air & Waste Manage.
Assoc., 1998, San Diego, USA
Chapman, T.E. and Mook, P.H. 1997. Experience and limitations oftitanium dioxide photocatalytic
treatment for vapor phase contaminants at McClellan air force base. In: Proceedings Third
International Conferenee on Ti0 2 Photoeatalytie Purifieation and Treatment of Water and Air,
Orlando, Florida, Naval Surfaee Warfare Center (Publisher), Maryland, USA: \06.
Chen, C.T., Tafuri, A.N., Dellinger, B. and Graham, J.L. 1997a. Photothermal enhaneed destruction of
aromatic eompounds. In: Proeeedings Fourth International Conferenee on Advaneed Oxidation
Teehnologies for Water and Air Remediation. Orlando. Florida. Naval Surfaee Warfare Center
(Publisher), Maryland, USA:114-l 15.
Chen, J., Huang, A., Spiess, F., Cao, L.X. and Suib, S.L. 1997b. Deactivation oftitania photocatalysts
in the oxidation of triethylamine in air. In: Proeeedings Third International Conference on Ti02
Photocatalytic Purification and Treatment of Water and Air. Orlando, Florida, Naval Surfaee Warfare
Center (Publisher), Maryland, USA: 183.
Chiraevski, V., Korobstsev, S., Medvedev, D. and Tulupov. M. 1995. Cleaning off-gas rrom mixture
complex VOCs by pulse corona diseharge. In: Proceedings World Environmental Congress. Science &
Teehnology Integration Ine. (Publisher). London, Ontario, Canada: 292.
Chmielewski, A.G., Tyminski, B., Licki, J., IIler, E., Zimek, Z. and Dobrowolski, A. 1993. Pilot plant
for f1ue gas treatment with electron beam - start up and tow stage irradiation tests. Radiat. Phys.
Chem. 42: 663-668.
Cho, K.S., Hirai, M. and Shoda, M. 1992. Enhanced removal efficieney of malodorous gases in a
pilot-scale peat biofilter inoculated with Thiobacil/us thioparus DW44. J. Ferment. Bioeng. 73: 46-
50.
Cleland, M.R., Lawrence, C.B. and McKeown. J. 1996. Technical and economical aspects of electron
beam processing for environmental applications. In: Proceedings Third International Conference on
Advanced Oxidation Technologies for Water and Air Remediation. Cincinnati, Ohio, Naval Surface
Warfare Center (Publisher), Maryland, USA.
Delprat, H., Pichat, P., Guillard, C. and Courbon, H. 1997. Intermediate produets rrom the degradat ion
of 2,3-dimethylpyrazine, a food-processing odorous compound, in UV -irradiated Ti02 aqueous
suspension. In: Proeeedings Third International Conference on Ti0 2 Photocatalytie Purifieation and
Treatment of Water and Air. Orlando, Florida. Naval Surface Warfare Center (Publisher). Maryland,
USA: 125-126.
Deminsky, M.A., Fridman, A.A. and Potapkin, B.V. 1995. Physical-chemieal mechanism of NOx
oxidation stimulated by eleetron beam. In: proceedings World Environmental Congress. Scienee &
Teehnology Integration Ine. (Publisher). London, Ontario, Canada: 339.
Doi, T., Osada, V., Morishige, A., Tokunaga, O., Miyata, T., Hirota, K., Nakajima, M. Komiya, M.,
Miyajima, K. and Baba, S. 1993. Pilot-plant for NO" S02 and HCI removal rrom f1ue-gas of
municipal waste incineration by electron beam irradiation. Radiat. Phys. Chem. 42: 679-682.
Federle, S.P., Littrell, D.M., Rogers, J.W. and Rolader, G.E. 1997. A pilot scale NO x removal system
for diesel engine exhaust. In: Proeeedings Fourth International Conference on Advanced Oxidation
Technologies for Water and Air Remediation. Orlando, Florida, Naval Surface Warfare Center
(Publisher), Maryland, USA: 25.
Getoff, N. 1996. Radiation-indueed degradation of water pollutants - state of the art. Radiat. Phys.
Chem. 47: 581-593.
Graf, J., Kittrell, J.R. and Chintawar, P. 1997. Photocatalytic destruction of ethylene in air. In:
Proceedings Third International Conference on Ti0 2 Photocatalytic Purification and Treatment of
Water and Air. Or\ando, Florida, Naval Surface Warfare Center (Publisher), Maryland, USA: \05.
Groenestijn, J.W. van and Hesselink, P.G.M. 1993. Bioteehniques for air pollution control,
Groenestijn, J.W. van, Doddema, H., Kok, H. and Koster, T.P.M. 1994. Combined photochemical and
biological treatment of off-gases. VOI Berichte. 1104: 313-324.
Grothaus, M.G., Hutcherson, R.K., Brown, R, Roush, R, Korzekwa, R.A, Ingram, M.W . Beck, se.
George, M., Ridgeway, R.G. and Pearce R.V. 1995. Process off-gas abatement using a pulsed eorona
discharge. In: Proceedings World Environmental Congress. Seience & Techno[ogy [ntegration [ne.
(Publisher), London, Ontario, Canada: 213.
Hali, RJ., Bendfeldt, P., Obee, T.N, and Sangiovanni, J.J. 1997. Computational and experimental
studies of UV/titania photocatalytic oxidation of VOCs in honeycomb monoliths. In: Proceedings
Third International Conference on TiO, Photocatalytic Purification and Treatment of Water and Air.
Orlando, Florida, Naval Surface Warfare Center (Publisher), Maryland, USA: 78.
Huang-Van, c., Pichat, P., Courbon, H., Disdier, J. and Guillard, C. 1997. PhotocatalySlS over IlO, as
a deodorization technique: elimination in air of pyridine, 2-furfural and butanedione. In: Proceedings
Third International Conferenee on TiO, Photocatalytic Purification and Treatment of Water and Air.
Orlando, Florida, Naval Surface Warfare Center (Publisher), Maryland, USA: 123-124.
Hung, C. and Yuan, C. Gas-phase photocatalytic degradation of perchloroethylenc on glass pellets

irnrnobilized with TiO,. In: Proceedings of the 91 st Annual Meeting & Exhibition ofthe Air & Waste
manage. Assoc., 1998, San Diego, USA.
Hung, C. and Yuan, C. H,Orassisted photocatalysis of sodium dodecylsulfate surfactant for

groundwater remediation. In: Proceedings of the 93 ,d Annual Conference & Exhibition of the Air &
Waste Manage. Assoc., 2000, Salt Lake City, USA.
Ibusuki, T., Takeuchi, K., Nishikata, S., Shin-kai, K., Miyamoto, M. and Takahashi, T. 1995. Removal
of aiI' pollutants from ambient air by using a mixture of photocatalyst and adsorbent powders. In:
Proceedings World Environmental Congress. Science & Technology [ntegration Inc. (Publisher).
London, Ontario, Canada: 249.
Kok, HJ.G. 1996. Conversie van micro-verontreinigingen bij procesgasstromen door middcl van UV-
straling in combinatie met ozon, TNO-MEP (Publisher), The Netherlands. Report R 96'374
Kurucz, C.N., Waite, T.D., Cooper, W.J. and Nickelsen, M.G. 1995. Empirical models for estimating
the desctruction of toxic organic compounds utilizing electron beam irradiation at full scale. Radiat.
Phys, Chem. 45: 805-816.
Lee, H,M. and Chang, M.B. Non-thermal plasma: an emerging technology for VOCs control. In:
Proceedings ofthe 91st Annual Meeting & Exhibition of the Air & Waste manage. Assoe., 1998, San
Diego, USA.
Li, K., Liu, S,Y.c., Khetarpal, S. and Chen, D.H. 1997. Ti02 photocatalyuc oxidation oftoluene and
PCE vapor in the air. In: Proceedings Third [nternational Conference on TiO, Photocatalytic
Purification and Treatment of Water and Air. Orlando, Florida, Naval Surface Warfare Center
(Publisher), Maryland, USA: 172-173.
Li, X.Z, and Zhao, Y.G. 1999. Advanced treatment of dyeing wastewater for reuse. Water Sci.
Technol. 39: 249-255.
Lichtin, N,N. and Sadeghi, M. 1997. PCO of bcnzene vapors over TiO,. [n: Proeecdmgs Third
International Conference on TiO, Photocatalytic Purification and Treatment of Water and Air.
Orlando, Florida, Naval Surface Warfare Center (Publisher), Maryland, USA: 170-171.
Mattachini, F., Sani, E., Yan, K., Gallimberti, 1., Rudolf, R. and Marode, E. 1995. Flue gas cleaning
by pulsed corona streamer fundamental study and pilot plant design. In: Proceedings World
Environmental Congress. Science & Teehnology lntegration [ne. (Publisher). London, Ontario,
Canada: 215.
Murabayashi, M., !toh, K. and Togashi, K. 1997. Gas-phase photocatalytic degradation of organic
chlorine compounds in water. In: Proceedings Third International Conference on TiO, Photocatalytic
Purification and Treatment of Water and Air. Or1ando, Florida, Naval Surface. Warfare Center
Murata, Y., Tawara, H., Obata, H. and Takeuchi, K. 1997. Air purifying pavement: development of
photocatalytic concrete blocks. In: Proceedings Third International Conference on TiO, Photocatalytic
Nogueira, R.F.P., Alberici, R.M., Morais, J.F. de and Jardim, W.F. 1997. An alternative Ti0 2 variety
for aqueous and gas-phase decontamination. In: Proccedings Third International Conference on Ti0 2
Photocatalytic Purification and Treatment of Water and Air. Orlando. Florida. Naval Surface Warfare
Center (Publisher), Maryland, USA: 83-84.
Obee, T.N. and Hay, S.O. 1997. The augmentation of UV photocatalytic oxidation with trace
quantities of ozone. In: Proceedings Third International Conference on Ti0 2 Photocatalytic
(Publisher), Maryland, USA: 80.
Ollis, D.F. 1995. Heterogeneous photocatalysis for purification and treatment of air: progress,
problems, prospects. In: Proceedings Wor1d Environmental Congress. Science & Technology
Integration Inc. (Publisher). London, Ontario, Canada: 230.
Penetrante, B.M., Hsiao, M.C., Bardsley, J.N., Merritt, B.T., Vogtlin, G.E., Wallman, P.lI., Kuthi, A.,
Burkhart, c.P. and Bayless, l.R. 1995. Comparison of electron beam and electrical discharge
processing of volatile organic compounds in air streams. In: Proceedings World Environmental
Congress. Science & Technology Integration Inc. (Publisher). London, Ontario, Canada: 182.
Prager, L., Mehnert, R., Sobottka, A., Paur, H.-R., Baumann, W. and Mătzing, H. 1995. Electron
beam degradation of chlorinated hydrocarbons in air: results of a field study In: Proceedings World
Environmental Congress. Science & Technology Integration Inc. (Publisher). London, Ontario,
Canada: 280.
Qinglin, L. and Chanjuan, X. 1995. Application of TiO, based photocatalyst for purification of
contaminated indoor air. In: Proceedings Wor1d Environmental Congress. Science & Technology
Integration Inc. (Publisher). London, Ontario, Canada: 351.
Raissi, A.T., Martin, E.D., Muradov, N., laganathan, S., Painter, C.R. and Kemme, M.R. 1998. Mass
transfer considerations in the design ofvapor-phase photocatalytic reactors. J. Adv. Oxid. Technol. 3:
188-198.
Sakamoto, K, Tonegawa, Y. and Ishitani, O. 1997. Destruction of indoor air pollutants in TiOrwall
coated cylindrical flow reactor under 254 nm UV irradiation. In: Proceedings Third International
Conference on Ti0 2 Photocatalytic Purification and Treatment of Water and Air. Orlando, Florida,
Naval Surface Warfare Center (Publisher), Maryland, USA: 94-95.
Sănchez, B., Romero, M., Fabrellas, B., Blanco, l. and Avilla, P. 1995. Comparative assessment of
TiOrbased monoliths for photocatalytic destruction of organic pollutants in gaseous emissions. In:
Proceedings World Environmental Congress. Science & Technology Integration Inc. (Publisher).
London, Ontario, Canada: 355.
Sănchez, B., Romero, M., Cardona, A., Fabrellas, B., Avila, P. and Bahamonde, A. 1997.
Photocatalytic treatment of VOCs in gaseous emissions, success and limitations. In: Proceedings Third
International Conference on Ti0 2 Photocatalytic Purification and Treatment of Water and Air.
Sato, S., Tokunaga, O., Arai, H. and Hashimoto, S. 1992. Electron beam irradiation technology for
environmental conservation. In: Proceedings International Symposium on Applicatoons of Isotopes
and Radiation in Conservation ofthe Environment: 3-15.
Selzer, V.H., Graf, J.e., Crittenden, le. and Hand. D. W. 1997. Gas-phase photocatalytic destruction
of methane in air for space applications. In: Proceedings Third International Conference on TiO,
Photocatalytic Purification and Treatment of Water and Air. OrI an do, Florida, Naval Surface Warfare
Sint Annaland, M. van, 1996. Ontwikkeling van een heterogeen gas-vast TiO,-fotokatalytisch
reactorsysteem voor de verwijdering van microverontreinigingen uit procesgasstromen. Report from
CHERA, The Netherlands.
Sjăberg, A., Heinzle, E., Kut, O.M. and Hungerblihlcr, K. 1997a. Waste gas treatment using advanced
oxidation processes: OJ/H,O, and non-thermal plasma for combination with biotreatment. In:
Proceedings of the International Symposium on Biological Waste Gas Cleaning, Maastricht, the
Netherlands, April 28-29, 1997. Prins, W.L. and van Ham, l (eds.). VOI Verlag GmbH, Dlisseldort;
Germany: 373-376.
Sjăberg, A. Teich, T.H., Heinzle, E. and Hungerblihler. K. 1997b. VOC treatment with a non-thermal
plasma. In: Proceedings Third International Conference on TiO, Photocatalytic Purification and
Treatment of Water and Air. Orlando, Florida, Naval Surface Warfare Center (Publisher), Maryland,
USA: 60-62.
Stock, N.L., Peller, J., Vinodgopal, K. and Kamat, P.V. 2000. Combinative sonolysis and
photocatalysis for textile dye degradation. Environ. Sci. Technol. 34: 1747-1750.
Tompkins, D.T., Bula, R. and Anderson, M.A. 1997. Characteristics of a commercial-scale

photocatalytic oxidation device for removing ethylene from indoor air. In: Proceedings Third
International Conference on TiO, Photocatalytic Purification and Treatment of Water and Air.
Upadhya, S. and Ollis, D.F. 1998. A simple kinetic model for the simultaneous concentration and
intensity dependencies ofTCE photocatalyzed destruction. J. Adv. Oxid. Technol. 3: 199-202.
Vandevivere, P.e., Bianchi, R. and Verstraete, W. 1998. Treatment and reuse of wastewater from the
textile wet-processing industry: review of emerging technologies. J. Chem. Technol. Biotechnol. 72:
289-302.
Vitale, S.A., Hadidi, K., Cohn, D.R., Bromberg, and L., Falkos, P. 1996. Decomposing VOCs with an
electron-beam plasma reactor. Chemtech. 26: 58-63.
Vorontsov, A.V., Barannic, G.B., Savinov, E.N. and Parmon, V.N. 1995. Gas-phase photocatalytic
oxidation of volatile organic compounds by air on TiO, deposited on a honeycomb support. In:
Proceedings World Environmental Congress. Science & Technology Integration Inc. (Publisher).
London, Ontario, Canada: 37!.
Wang, Y. 2000. Solar photocatalytic degradation of eight commercial dyes in TiO, suspension. Water
Res. 34: 990-994.
Weaver, S., Mills, G. and Slaten, L. 1997. Gas-phase transformations of freons photoinitiated by TiO,
partic1es. In: Proceedings Third International Conference an TiO, Photacatalytic Purification and
Treatment of Water and Air. Orlando, Florida, Naval Surface Warfare Center (Publisher), Maryland,
USA:3!.
Yang, C., Yamamota, T., Beltran, M.R. and Kravets, Z. 1997. Corona induced chemical scrubber for
the control of nitrogen oxides. In: Proceedings Fourth International Conference an Advanced
Oxidation Technologies for Water and Air Remediation. Orlando, Florida, Naval Surface Warfare
Yang, c., Beltran, M.R., Kravets, Z. and Yamamoto, T. Development ofpulsed plasma technology for
VOC destruction. In: Proceedings of the 91st Annual Meeting & Exhibition of the Air & Waste
manage. Assoc., San Diego, USA.
CHAPTER 8 ROTATING BIOLOGICAL CONTACTORS
Philipp RUDOLF VON ROHR and Patrik RUEDIGER
1. Introduction
Waste gases are found in alI branches of industries. As long as the contents of
organic or inorganic substances are be10w local admissible values the waste gas may
be released to the environment without any additional treatment. If these specific
values are higher the waste air shall be treated before leaving the plant site. Clearly
gaseous wastes are difficult to store; therefore most of the waste must be cleaned in
place where it is generated mostly by fast chemical absorption, adsorption or
incineration (see Chapter 2). Industries like to be flexible in producing products.
Therefore variable waste production may also be observed. AlI biological waste gas
systems are extremely well suited for constant gas loads and constant waste
composition which is not the standard case. Normally the loads and contents of
organic by-product in waste gases is changing within short periods of time. This
explains why only few systems are incorporated in industrial applications. However,
many different approaches have been attempted to overcome the load fluctuations.
The RBC was originally deve10ped for waste water streams as an alternative
to standard sewage systems. Several discs -acting as biofilm carrier- are mounted on
a shaft at defined distances from each other. The discs are rotating at a speed of about
two revolutions per minute in order to increase mass transfer and to control the
biofilm thickness by the shear force induced between the biofilm and the waste
water. Charaklis (1990) has divided the growth of the biomass on inert carriers (here
discs) in eight steps:
• Accumulation of organic molecules on the carrier;

• Cells reach the carrier surface from the bulk;
• Cells adsorb on the carrier surface;
• Part ofthe reversibly adsorbed cells desorb;
• Part ofthe celIs adsorb irreversibly and immobilise;
• Immobilised celIs grow with the available nutrients and metabolic products leave
the system;
• Further cells adsorb at the surface and increase the biomass film;
• Parts ofthe biomass film detach and are collected in the liquid sump.
Charaklis and Marshall (1990) experimentally investigated biofilm growth in

a tube. It was shown that at increasing substrate concentrations biomass
accumulation increased, ending up with a bigger stationary film thickness. Different
authors (Peyton and Charaklis, 1993; Stewart, 1993) described the influence of the
content of organic material in the feed on biofilm detachment. Such influence was
dominant whereas the shear stress influence was negligible in the investigated range
of operation. The detachement of the biofilm is originated by different mechanisms,
among which erosion, sloughing and abrasion.
From a process standpoint several conditions should be fulfilled in order to
implement those systems. One main goal is the controlled limitation of biofilm
growth which could be reached by finding a way to run a system in which the
201
C. Kennes and M. C. Veiga (eds.), Bioreactors for Waste Gas Treatmells, 201-214.
202 Ph. Rudolf von Rohr and P. Ruediger
biofilm detachment rate is equal to its growth rate. This may ideally be realised in a
rotating system with fixed carriers.
1.1 RBC FOR WASTE WATER TREATMENT
The idea of RBCs in waste water treatment goes back to the beginning of the
tewntieth century. Tests were performed in Germany and in the USA between 1920
and 1930, but the idea did not spread out. In 1950 the RBC was further developed at
the University of Stuttgart in Germany. The first industrial system was built in 1960
(Antoine, 1976). Because of a very poor construction the shaft of this first RBC
broke down. Furthermore the discs on that shaft were unsufficiently fixed. The
rumors about these problems were not very favorable for further introduction of the
system (Andowski, 1956).
The carrier discs were fixed vertically to the horizontaJly mounted and
rotating shaft. The waste water entered on one side of the mostly covered part of the
reactor and flew from one cell to the next one. Each cell contained two discs. The
mean residence time was very short compared to conventional sewage plants. In
other constructions curved PVC sheets have been used, winding them around the
shaft like a big cylinder. The flow of the waste water is in such case paralle1 to the
shaft. Depending on the design, between 40 and 60% of the volume of the carrier is
immersed in the liquid. During the rotation the biofilm ahsorbs the oxygen in the gas
phase (upper part) and the organic material in the liquid phase. The rotation leads to
a good mixing and holds detached biofilm particles in a floating condition (Antoine,
1976). In the fie1d of waste water treatment RBCs are considered to present the
foJlowing advantages (Bitton, 1994):
• Low investment costs;

• Low maintenance;
• Good buffering capacity;
• High reliability: low chance to plug.
Compared to conventional systems the energy consumption is reduced by a

factor two and the necessary space is reduced by 20 to 60%. The structure of the
biofilm in a stirred, rotating system is more filamentous compared to smooth closed
pores in calm systems. This difference allows ---even for thicker films- transport of
the substrate and oxygen to places close to the carrier and therefore guarantees an
efficient bioreaction.
1.2 HISTORY OF RBC FOR WASTE GAS TREATMENT
A similar problem occurs in biological waste gas treatment systems as in waste water
systems, Le., the accumulation of biomass leading to clogging phenomena. A
Canadian GrOUpl (Kahler and McKim, 1989) reported on the use of a rotating system
for waste gas treatment. Our group at ETH (Ruediger, 1999) concluded experimental
tests and has adapted RBC systems since 1997. Sabo and Fischer (1996) reported
about a rotating biofilter. The goal of their investigations was to improve the wetting
of the biomaterial and to decrease the tendency of compression of the filter material.
I CMS Group Inc., Concord, Ontario, Canada

Rotating Biological Contactor 203
2. Experience with the RBC for waste gas treatment
There is only limited experience with this kind of reactor. The above mentioned
group performed tests in a pilot plant with phenol during 373 days. For gas feed
concentrations between 1 and 7 g/m3 the removal efficiency remained above 99.9 %.
Our group wanted to investigate this novel system and carried out tests under defined
conditions. The Canadian group2 kindly provided parts of the pilot facility. The pilot
facility (Figure 8.1) consisted ofthree parts:
• Gas loading unit;

• Reactor with differently shaped discs made from different materials (metals,
polyethylene );
• Conditioning unit for the liquid part.
Dichloromethane (DCM) was chosen as model pollutant. Pressurised dry air

was mixed with a fully loaded (through a bubble column) flow of DCM. Flow rate
and inlet concentrations were continuously controlled and adjusted, working in the
range of 27 to 300 mg/m3. The inoculum used for start-up consisted of a mixture of
two types ofmicroorganisms, called DMI and DM2.
Figure 8.1. Process diagram of an RBC pilot facility (RBC with four chambers) (Ruediger,
1999).
2 ibid.
The reactor consisted of a cylindrical outer shell with several windows and an
inner shaft with several carrier discs fixed on it. The diameter of each disc was 28
cm. The distance between each disc was 10 mm. The shaft could rotate at variable
speeds between 5.8 and 8.8 rpm. The liquid container was conditioned with sodium
hydroxide (NaOH) and specific nutrients. Furthermore the temperature was
maintained at 30° C. The main parameters are summarised below:
Totallength (m) 1.112

Diameter (m) 0.312
Total volume (1) 85.7
Liquid volume (1) 26.3 at V G = 60.7 l/min
Gas volume (without discs) (%) 59.4
Total carrier surface (m2) 7.44
Specific surface (m2/m3 ) 87
2.1 RESULTS WITH THE RBC
The results are presented using the performance parameters defined below. Some of
these parameters are similar as for biofilters and have been extensively defined
previously (Chapter 3):
Elimination capacity:
(8.1)
Removal efficiency:
RE 100 (8.2)
Surface-specific elimination rate:
VG (c GO - c G•ou' ) EC (8.3)
SEC =
a VR a
Surface-specific substrate load:
SLO = (8.4)
Data on elimination capacity (EC) and removal efficiency (RE) are shown in
Figure 8.2.
140 I I
120 f- t, O
I O
O -
100 f- t, O -
~ t,
-
1
O
80 l-
60 l- t, <> <> OO
~t,<>t, <El
~
O
-
w -
40 1-
20 i/fS)<> O
O
-
O
100
80
~
~ 60 O
<> o O
w 40 $> O O O
D:: O
~~t,oMO °t,O O
O
20 <> <> t, t, O
V G
<>
O
O 5 10 15 20
eoo [g/m'1
OV 0= 23.6 IImin oV 0= 41.81/min 1':. V 0= 60.7 I/min <> V o =72.2 I/min

Figure 8.2. Elimination capacity (EC) and removal efficiency (RE) of dichloromethane from
air using a rotating biological contactor (as described in this Chapter).
The elimination capacity tends to assymptotically approach a saturation

value. This behaviour led to the conclusion that in this concentration and operating
range the limiting factor appeared to be the mass transfer rate rather than the
bioreaction. The elimination capacity reached 120 g/ m~ h and seemed to be
independent ofthe residence time (VRIV G) or the gas flow rate. The residence time at
the maximum measured Ee was 151 s. Of significant interest is the dependency of
the removal efficiency on the rotational speed. Figure 8.3 shows the results for two
different inlet concentrations.
Figure 8.3 shows that for the highest feed concentration the removal
efficiency approached a maximum at 3 rpm. This means that there was no mass
transfer limitation. At lower inlet concentrations a different behaviour was observed
and the RE increased even for numbers ofrevolution per minute higher than 8. Mass
transfer limitation was detected, owing to the rotational speed. The data represented
in Figure 8.2 indicate that there is no reaction limitation due to the rotation of the
discs within the investigated range. Therefore it was conc1uded that mass transfer
limitation took place at the level of the gas-biofilm transfer or the gas-liquid
transfer, which are not influenced by the rotational speed.
Summarising the effects, it can be conc1uded that:
(i) when rising the rotational speed:
• Mass transfer of oxygen and dichloromethane into the liquid phase is increased;
• The amount of liquid increases on the discs whereby the mass transfer from the
gas phase into the liquid decreases. The rate of absorption increases;
• The liquid is better mixed. Oxygen and the pollutant are better distributed in the
reactor;
• Energy consumption increases;
(ii) when decreasing rotational speed:
• Bad mixing is observed. Detached parts of the biofilm sediment to the reactor
floor and anaerobic zones appear;
• The uptake of oxygen is decreased;
• The shear forces on the biofilm decrease;
• The release of the non-volatile components from the biofilm is hindered. They
may accumulate eventually leading to inhibition ofthe activity ofthe biofilm.
An improved removal efficiency at increasing rotational speeds has also been

observed with RBCs used for waste water treatment (Lu el al., 1997) as long as a
specific value of the rotation speed is not exceeded. Tests proved that even after five
months of continuous operation the RBC was not plugged; and its elimination
capacity did not decrease (Ruediger, 1999).
2.2 COMP ARISON OF THE RBC AND A BIOTRICKLING FIL TER
In the following section, the degradation capacity of the rotating biofilm reactor will
be compared to that of biological trickling filters. The DCM biodegradation data
from Diks (1995) and from Zuber (1995) will be used for comparison. In order to
make a valid comparison, geometrical and operational values must be considered.
These data are presented in Tables 8.1 and 8.2.
35
o
30
o
o
25 v
o O c Go =42 g/m 3
~ 20 o o c Go =16.1 g/m 3
UJ r.
c:: 15
D D D o D
10 D
D
5
o
O 2 4 6 8 10
n [rpm]
Figure 8.3. Influence ofthe rotational speed on the removal efficiency (RE) of the described
rotating biological contactor. Substrate: dichloromethane in air; gas flow rate: V G = 60.7
I/min.
Table 8.1. Geometrical variables ofbiotrickling filters and RBC for the removal of DCM
A VR Atot, dK AK L Type of References
m2/m l 1 m- m 10·3 m' m packing
750 339 254 0.4 125.6 2.7 y," ceramic Diks,1995
Novalox-
saddle
350 7.6 2.66 0.096 7.28 1.05 Mellapak Zuber, 1995
350.Y
87 85.74 7.44 0.312 76.45 1.112 Discs RBC for waste
air
Table 8.2. Operational data ofbiotrickling filters and RBC for the removal ofDCM
vG tG b T cGO,max LO ma• EC max References
m/h s m3/m 2 .h ec glm 3 glmR3 .h glm/.h
160 61 3.6 20 Il 650 157 Diks,1995
160 23.6 II 30 20 3051 420 Zuber, 1995
18.4- 71.3- n = 5.8 rpm 30 17 858 120 RBC for
65.1 218 waste air
For comparison with biotrickling filters the surface-specific elimination

capacity (SEC) is plotted against the surface-specific dichloromethane load (SLO).
With such method of comparison neither the residence time nor the specific surface
affect the results.
Figure 8.4 compares degradation measurements in the rotating biological
contactor and in the two biological trickling filters. The data from Zuber (1995)
consisted of three sets of measurements obtained at various biomass concentrations
in the biotrickling filter.
0.5 2
"
6 6 1-
0.4
1,5
:c- :c- 6
6
E NE
~.~
,
0.3 6
:2! :2!
()
w
cn 0.2 In
• D ()
w
cn
if!J [ [] •
O 0,5
6Î
0.1 w O ~
~DD
o o
O 0.2 0.4 0.6 0.8 O 4 8 12 16
SLO [g/m'.h] SLO [g/m'.h]
[1 Diks. 1995; t.. Zuber. 1995; • rotating biofilm
Figure 8.4. Comparison of dichloromethane removal from the gas phase in biotrickling
filters and in a rotating biological contactoT. (Diks, 1995; Zuber, 1995; Ruediger, 1999).
208 Ph. Rudalf van Rahr and P Ruediger
With a surface-specific load of less than 1 g/m 2 h, the performance values for
the rotating biofilm reactor were lower than those obtained by Zuber (1995) and
equivalent to values in the upper range of those measured by Diks (1995). This
characteristic is found even though the part of the rotating biofilm re actor submerged
in the liquid phase, which would be expected to have a greater resistance to diffusion
than the unsubmerged part, was inc1uded in the calculations of the surface-specific
elimination capacity. Comparison with the data from the literature showed that the
part ofthe biofilm covered with the liquid phase as well as the biomass suspended in
the liquid phase can contribute to biodegradation. An increased biomass activity in
the rotating biological contactor is less probable because all reactors were inoculated
with microorganisms especially adapted to DCM, namely DMI and DM2.
At higher loads, those gre ater than approximately 4 g/m 2 h, the values from
Zubcr (1995) level off, indicating an increasing limitation of the reaction. In contrast,
the values for the RBC continue to increase. Therefore the degradation capacity of
the rotating biological contactor was limited by diffusion even at high loads.
2.3 PARAMETERS AND MODELS
In the present section differences between short term and long term influences on the
elimination capacity will be considered. Long term influenccs are paramcters
influencing the structure and the activity of the biofilm. These influences determine
the maximum performance of the RBC plant. Long term influences are assumed to
take place on a time scale of several days or weeks. Short term influences affect only
the combined diffusion and degradation. In the lattcr case the characteristics of the
biofilm are considered to remain unchanged. Short term influences have a time scale
of minutes. Parameters such as the number of revolutions per minute and substrate
concentrations can present both short term and long term effects on the degradation
capacity of an RBC. Parameters as pH, temperature, and nutrient or salt
concentrations, which have both short and long term influences, are assumed to
remain constant.
2.3.1. Lang term parameters and madels

In the long term, both the rotational speed and the substrate load influence biomass
growth and biofilm structure. Therefore a stationary degradation characteristic,
determined by the thickness, the structure, and the activity of the biofilm develops
over a longer period of time.
Using residence time analysis Sassi el al. (1996) examined the parameters
influencing an RBC reactor. Depending on the surface-specific substrate load and on
the mixing time owed to disc rotation, operating conditions as shown in Figure 8.5
were suggested.
With a longer mixing time, i.e., a slower disc rotation, diffusion is limiting.
At shorter mixing times resulting from higher rotational speeds of the discs, a higher
shear rate develops and the shear forccs limit substratc degradation. Substrate
limitation and limitation of the bioreaction are dependent on the substrate load.
Optimal operation is found in the range in which the bioreaction rate is limiting.
According to Sassi el al. (1996), the most important operating parameters are
substrate load, volumetric flow rate ofthe liquid phase, and Reynolds numbcr, which
inc1udes parameters as: shaft speed, liquid density, diameter of the discs, and the
dynamic viscosity.
surface-specffic Ioa<!
Figure 8.5. Working fields of a RBC (Sassi et al., 1996).
2.3.2. Short term injluences

In order to create a model for the short term effects of the rotational speed and the
substrate concentration on the elimination capacity of a Rotating Biological
Contactor, it will be assumed that the biofilm does not change. Figure 8.6 shows
substrate movement in a rotating biofilm reactor used for the treatment of waste gas.
Four coexisting compartments can be seen.
gas phase biofilm
liquid bulk
Figure 8.6. Waste gas flows in a rotating biofilm reactor.

210 Ph. Ruda/f van Rahr and P. Ruediger
• Waste gas flows through the gas filled space in the re actor;
• The largest part ofthe liquid phase (the bulk fluid) is collected in the lower part of
the reactor;
• The rotation ofthe discs lifts liquid from the bulk fluid up into the gas-filled area.
This liquid interacts with the biofilm. To simplify the model one could regard this
liquid as a film adhering to the biofilm;
• The biofilm.
In the short term the rotational speed and the substrate load int1uence
diffusion in the reactor and the kinetics of biological degradation. The measurements
presented above examine these short term phenomena.
The substances can be absorbed from the gas phase either by the liquid film
adhering to the biofilm or by the bulk liquid. Because of the rotation of the discs a
constant transfer between the liquid adhering to the biofilm and the bulk liquid
occurs.
In their model of the dependence of the distribution of a liquid on a disc in a
RBC Vaidya and Pangarkar (1987) used the following parameters: disc rotational
speed; distance from the centre of the disc; viscosity and surface tension of the
liquid, and the extent to which the discs were immersed. In the model it was assumed
that the liquid remained attached to the disc. The movement of the liquid relative to
the disc while it is in the gas phase was ignored. Using the time periods during which
the individual points on the disc remained in the gas phase, the concentration profile
could be determined. Using such a profile, absorption rates and, in the end, substance
transfer coefficients could be calculated. According to these calculations. the mass
transfer coefficient for a disc half immersed in the liquid phase depends solely on the
rotational speed. The interaction between the biofilm and the liquid was not
discussed.
In practice it must be assumed that the t1uid moves relatively to the disc. The
extent to which this t10w contributes to the total absorption rate should be further
investigated.
Owing to the rotation of the disc the oxygen concentration in the biofilm
varies. In the gas filled space the biofilm adsorbs oxygen from the gas phase. Thus
oxygen concentration in the biofilm increases. On the submerged part of the disc thc
oxygen concentration in the biofilm decreases. Gujer and Boller (1990) created a
model that also considered anaerobic biodegradation of organic substances, flake
formation, and shear-rate-dependent detachment ofthe biofilm. Modelling of oxygen
diffusion into the liquid was of great interest. In this model the fluid taken from the
bulk liquid within the biofilm was ignored. Oxygen from the air was assumed to be
transfered directly from the gas phase to the biofilm. Factorial analysis was applied
to the variations in oxygen concentrations within the biofilm resulting from disc
rotation. The period of exposure of the biofilm in the gas filled area was too short to
seriously influence the oxygen concentration profile caused by the biological
degradation. Only at small angular velocities did the increased exposure time re suit
in an increased penetration depth. Gujer and Boller (1990) concluded that the
influence of disc rotation on the degradation capacity of the biofilm must be taken
into account only at very small rotational speeds. In their model they did not
differentiate between diffusion into the liquid film and diffusion into the bulk liquid,
but used instead an average kLa value.
Rotating Biological Contactor 2]]
Boumansour and Vase! (1998) examined the relationship between diffusion

from the gas phase to the liquid phase and rotational speed in a disc equipped RBC.
They used propane as a tracer as it is neither biologically degraded nor biologically
produced. Measurements were made with clean and with biofilm covered discs. In
this way, the influence of thc biofilm on diffusion could be determined. The results
showed a decrease in the ha value for the biofilm covered discs compared to the
clean ones. The ha values increased with rising speed of rotation. Boumansour and
Vasel (1998) correlated the measurements by means ofa Sherwood re!ationship. The
Sherwood number for diffusion in a RBC with clean discs was calculated using the
Reynolds and Froude numbers, both being dependent on the rotational speed and on
disc diameter, as well as a dimensionless number that depended on the extent to
which the discs were immersed.
Studies on the dependence of the removal efficiency on the speed of rotation
showed that diffusion caused by disc rotation was a limiting factor at low pollutant
concentrations. Disc rotation had a visibly smaller inf1uence at higher concentrations.
However, plotting the surfice-specific elimination capacity against the surface-
specific substrate load showed that up to a substrate load of 18 g/m 2 .h. the reaction
rate was not the limiting step. These conclusions are summarised in Figure 8.7.
limited by
mass transfer.
independent of
rotation speed
limited by reaction
Iimited by
mass transfer,
dependent of
rotation speed
substrate load
Figure 8.7. Regimes dependent on the rotationa! speed and the substrate load for constant
biofilm conditions.
3. Design improvements
In the experimental RBC the gas phase f1ew tangentially along the discs (Figure 8.8).
For such construction, different residence times re suit for the gas molecules in the
reactive part of the RBC. Therefore it was suggested feeding the gas into the re actor
through slots in a hollow shaft. Such construction forced the waste gas to f10w in a
radial direction with respect to the discs.
212 Ph. Rudal( van Rohr and P. Ruediger
tangential air path radial air path

out of hollow shaft
Figure 8.8. Gas flow paths through a rotating biofilm reactor.
Central introduction of the gas phase would also make it possible to use a
packing instead of discs as support for the biofilm. For this purpose the packing
could be held in the shape of a drum by a wide meshed screen. The extent to which
the packing would hinder removal of detached biofilm pieces must still be
investigated. Preliminary results obtained with a similar construction as the RBC
(hollow shaft) showed quite high removal efficiencies of toluene (Vinage el al ..
2001). Fora flow rate of 1 m 3/h with a high concentration of3.1 g/m 3 , the removal in
the RBC was 1.9 g/h. Referring to the active biofilm in the used re actor, the surface-
specific elimination capacity was 0.37 g/m 2.h.
4. Conclusions
The RBC is an alternative technology to other biological waste gas trcatment systcms
for low loaded waste gas flows. Simple design, flexibility, and good elimination
capacity are some ofthe highlights.
Further theorctical and experimental work is needed to check long term
stability and to define scaling up criteria. Industrial experience will be necessary to
set up a simple automation system and to test the constructional parts, corrosion
problems, its adaptation to different pollutants in a practical environment and to
confirm scale-up criteria and theoretical predictions.
The combination with other physical or chemical waste gas treatments (like
adsorption or chemi cal oxidation) may be necessary to overcome the two major
drawbacks Iinked to biological waste gas treatment systems, namely the long
adaptation time of the microorganisms and the limited degradation. However,
biological waste gas treatment systems are easy to handle and seem to require only
minimalmaintenance. Therefore they are interesting for further investigations.
Symbols
a m 2/m 1 specific diffusion surface
Atot m2 total diffusion surface
A m2 diffusion surface
AK m2 cross-sectional area ofthe column
b m1/m 2 .h sprinkling rate
CGO,max kglm 1 maximum concentration at gas entry
dK m column diameter
dSoh m disc diameter
EC g/m1.h eliminatiOl1 capacity
EC max glm1.h maximum eliminatiOl1 capacity
kLa Ils mass transfer coefficient
L m reactor length
LO glm'.h load
LOmax glm 3 h maximum load
n rpm shaft speed
RE % removal efficiency
SEC g/m 2 h surface-specific eliminatiOl1 rate
SLO g/m 2.h surface-specific load
T ec temperature
tG s mean hydraulic holdup time for the gas phase
VG m1/s volumetric gas flow
VI ml/s volumetric liquid flow
VR 1 reactor volume
m
VG m/s superficial gas velocity
Refcrcnces
Andowski, L. 1996. Past, present and future ofthe RBC industry. Water Eng. Manage. 143: 31-34.
Antoine, R.L. 1976. Fixed biological surfaces-wastewater treatment; the rotating biological contactoL
CRC Press Inc., Cleveland, Ohio.
Bitton, G. 1994. Wastewater microbiology. Wiley Series in Ecological and Applied Microbiology,
Wiley-Liss., New York.
Boumansour, B.E. and Vasel, J.L. 1998. New tracer gas method to measure oxygen transfer and
enhancement factor on RBC. Water Res. 32: 1049-1058.
Characklis, W.G. and Marshall, K.c. 1990. Biofilms. J. Wiley & Sons, Inc., New York.
Diks, R.M.M. 1995. Ph.D. thesis, Technische Universiteit Eindhoven, The Netherlands.
Gujer, W. and Boller, M. 1990. A mathematical model for rotating biological contactors. Wat. Sci
Technol. 22: 53-73.
Heits, H., Laurenzis, A. and Werner, U. 1997. Biologische Abgasreinigung mit kontrolliertem
Biomasseaustrag im periodisch rlickgesplilten Rieselbettreaktor. Gefahrstoffe Reinhaltung der Luft.
57: 153-158.
Kahler, B.D. and McKim, M.P. 1989. Biological contact gas scrubber for waste gas purification.
Ontario Research Foundation, Canada.
214 Ph. Rudalf van Rahr and P. Ruediger
Lu, c., Li Hung, c., Lee Lian, Y. and Lin Min, R. 1997. Effects of disc rotational speed and
submergence on the perfonnance on an anaerobic rotating biological contactor. Environment Internat.
23: 253-263.
Peyton, B.M., Charaklis, W.G. 1993. A statistical analysis of the effect of substrate utilization and
shear stress on the kinetics ofbiofilm detachement. Biotechnol. Bioeng. 41: 728-735.
Ruediger, P. 1999. Ph.D. thesis, ETH Zurich, Switzerland.
Sabo, F. and Fischer, K. 1996. Entwicklung und Erprobung eines neuartigen Rotor-Biofilters. In:
Fortschritte bei der thermischen, katalytischen, sorptiven und biologischen Abgasreinigung.
Kommission Reinhaltung der Luft im VDI und DIN (ed.). VDI Verlag, Diisseldorf, pp. 559-565.
Sassi, G., Ruggeri, B., Bosco, F. and Specchia, V. 1996. Relaxation time analysis of a rotating
biological contactor. Chem. Engin. Sci. 51: 2853-2858.
Stewart, P.S. 1993. Model ofbiofilm detachement. Biotechnol. Bioeng. 41: 111-117.
Vaidya, R.N. and Pangarkar, V.G. 1987. Convective diffusion model for mass transfer in a rotating
biological contactor: disc submergence less than 50%. Water. Res. 21: 1499-1503.
Vinage, 1., Wegmann, A., and Rudolf von Rohr, Ph. 2001. Novel reactor for waste gas purification:
Modified Rotating Biological Contactor. In: Proceedings ofthe 94'1> Annual Meeting & Exhibition of
the Air & Waste Manage. Assoc., 2001, Orlando, FL.
Zuber, L. 1995. Ph.D. thesis, ETH Zurich, Switzerland.

CHAPTER 9 ACTIVATED SLUDGE AND SUSPENDED GROWTH
BIOREACTORS
Ange1a R. BIELEFELDT
1. Overview
Of the number of alternative designs for biological treatment of contaminated air

streams, activated sludge reactors are often overlooked. The concept behind activated
sludge treatment is simple: in a single reactor the contaminant transfers from gas into
bulk liquid where it is degraded by a suspension of bacteria. The gas is generally
introduced into the bottom of the liquid in the reactor in the form of dispersed
bubbles. Using this most simplistic definition, a number ofterms have been used to
describe these systems with subtle differences implied. In true activated sludge
treatment, the primary function is to biologically treat dissolved contaminants such
as the organic carbon in conventional municipal waste water. The air contaminants,
such as odours or volatile organic compounds, are co-degraded with the
contaminants dissolved in the inlet waste water. In general these reactors are not
specifically designed for optimal gas treatment. Slight variations where the reactor is
specifically designed for gas treatment and is not co-degrading waste water
contaminants have been termed sparged suspended growth bioreactors (Bie1efeldt
and Stense1, 1998), suspended growth reactors (Neal and Loehr, 2000), and bubble
columns (Andrews and Noah, 1995). More speciali sed designs of these simple
systems are airlift bioreactors which are most widely used in chemi cal engineering
applications (Chisti, 1989; Cesario et al., 1995).
The mass transfer characteristics of gas to liquid have been poor1y defined in
many reactors where contaminated gases were treated in a suspended culture of
bacteria. It appears to have been assumed that mass transfer would not limit the
operating efficiency of the reactor. This is the case in many studies with fermenters
and similar systems. This chapter will illustrate that optimal designs explicitly
account for the physical/chemical characteristics of the contaminants to be treated
and the physical elements ofthe reactor that impact mass transfer.
The second critical element in the design of suspended growth gas treatment
reactors is the biodegradation of the contaminants. The consortia of bacteria present
and their growth conditions control biodegradation. Compared to biofilters, where a
variety of conditions can exist within the structure of the biofilm, it is easier to
control the growth conditions in a suspended culture reactor where complete mixing
of the liquid phase can be assumed. Electron acceptors, pH, and metabolite build up
can be readily controlled in the suspended growth system. With optimal control of
conditions in the reactor. biodegradation can maintain minimal concentrations of the
contaminants in the liquid thereby optimising the mass transfer efficiency.
Compared to alternative biological gas treatment methods, activated sludge
reactors have some advantages. Media plugging and bed drying concerns that are
common in biofilters are not an issue for suspended growth systems. Anaerobic
micro-zones that can develop in biofilms and change biodegradation pathways,
potentially creating odours or other undesirable by-products, are not a problem in
activated sludge systems. Bioscrubbers are traditionally two-stage. The gas phase
contaminants transfer into a liquid spray in the first reactor and are then biodegraded
215
C. Kenlles and M. C. Veiga (eds.), BioreactorsJor Waste Gas Treatme/lt, 215-254.
216 A.R. Biele{eldt
in the liquid suspension primarily in the second re actor. This two-reactor system is
more expensive because of having multiple reactors. In addition, biomass plugging
of spray units and attached biogrowth in the first stage are difficult to control.
This chapter summarises the pertinent design factors for the general category
of activated sludge reactors. Differences in the various sub-types of reactors are
described. Fundamentals of mass transfer and biodegradation are incorporated into a
model to aid the design of suspended growth gas treatment reactors. Published
applications of suspended growth reactors to treat various contaminants are
summarised. Finally, the costs of using activatcd sludge systems to treat
contaminated gases are discussed.
2. Process description
There are a variety of sub-classifications of processes that are variations on activated

sludge treatment. These processes include: traditional activated sludge reactors,
sparged suspended growth reactors (SGRs) or bubble columns, and airlift reactors.
Each is described in more detail below.
2.1 ACTIV ATED SLUDGE
In true activated sludge treatment the primary function of the re actor is to treat
dissolved contaminants in waste water. The re actor has generally been optimised for
biodegradation of the contaminants in the liquid. Gas is introduced into the reactor to
provide oxygen as an electron acceptor allowing suspended aerobic bacteria to
degrade dissolved organic contaminants. Mass transfer considerations in thc design
of these reactors are generally focused on oxygen delivery aud mixing. Current
application of activated sludge reactors to treating contaminated gases has required
little, if any, modification of the waste water treatment basins. The contaminated air
is simply used in addition to, or in place of, the clean air commonly used to supply
oxygen. A diagram showing the major components of activated sludge treatment
systems is shown in Figure 9.1.
Activated sludge is a very common process used to treat municipal wastc
water and industrial process wastewaters. The organic contaminants in municipal
waste water are generally readily biodegradable and have a high oxygen demand.
The waste water flows into a tank where suspended bacteria de grade the organic
compounds and/or transform inorganic compounds (i.e., ammonia to nitrate). To
supply sufficient oxygen to the bacteria, aerators with good mass transfer
characteristics are installed in the tank.
Most activated sludge systems supply the oxygen that is required via diffused
air. Altematives to air diffusion are pure oxygen diffusion or l11echanical surface
aeration. The diffuser placel11ent and gas tlow rates used for aeration are generally
sufficient to provide l11ixing of the liquid and to l11aintain the bacteria in suspension
(greater than 3 std Cl11 3 airlcl112 basin floor area.min, recol11l11ended by Metcalf &
Eddy, 1991). There are a limited number of cases where both submerged diffusers
and mechanical mixers are used to achieve desired levels of mixing and oxygen.
The two most commonly used types of diffused aerators are: (1) porous or
fine bubble diffusers; and (2) non-porous or coarse bubble diffusers. The porous
diffusers are generally made of ceramics in the form of plates or individual domes;
Activated Sludge 217
Irealed gas
o o 4-Bm
o o o Iiquid
risinggas o
depth
o bubbles o
o o o o Irealed
waslewaler
t
b '!>C t r i ~ I o waslewaler
suspenslo9J o
filter o o o
II...!:C:2.0n!!!l:!!!am~in~al~e.9.d+lCJa~ir~dlliiff~U~S@io~n:gd~ev~ice~==:Jl
tr_ o-o
o
\
gas
- - - - ----- i
"
moislure blower I activaled slud~)
+ - - - - seltled
- - -biomass
- - -(relurn
---
~ ~,
biomass ~
Figure 9.1. Conventional activated sludge reactor.
they are altematively made of membrane material on a plate, disc, or tube. As air is
forced through the porous diffuser material it enters the liquid in the form of tiny gas
bubbles typically ranging in diameter from 1 to 4 mm. Non-porous diffusers
generally consist of small holes in pipes or plates produc ing large gas bubbles of 6 to
10 mm diameter in the liquid. Smaller gas bubbles result in higher mass transfer
efficiency because greater contact area between the gas and liquid allows diffusion of
the gas molecules into the liquid. Therefore porous diffusers are generally superior to
coarse bubble aerators. Despite gre ater headloss of air inside fine bubble diffusers
(greater than 5 cm of water head) versus coarse bubble aerators, fine bubble diffusers
have still been shown to cost up to 50% less to aerate an activated sludge system
treating municipal waste water (US EPA, 1989).
Clogging of the porous diffusers can occur due to particulates in the inlet gas
and bacterial growth on the surface of the diffuser. Filtration of the gas prior to the
inlet manifold of the activated sludge tank is frequently used to prevent c10gging of
porous diffusers. Gas filtration is not required for coarse bubble aeration. Excessive
bacterial growth on the surface of the diffusers in contact with the water can occur
with either type of aeration device. The bacterial growth increases the headloss
across the diffuser and may change the aeration characteristics in the basin. This
fouling effect will decrease the aeration efficiency over time. Fouling can be detected
operationally by an increased blower discharge pressure and/or coarser bubbles. To
alleviate the effects of bacterial fouling, it is recommended that porous diffusers be
routinely c1eaned. Ceramic diffusers can be c1eaned by draining the basin and using
an aggressive water spray (fire hose) or by immersion of the diffuser heads in an
acidic solution. For membrane systems, a burst of higher airflow may dislodge
attached bacteria.
Manufacturers traditionally report the characteristics of their diffuser devices
as a specific oxygen transfer efficiency (SOTE) at a given submersion depth of the
diffuser and gas flow rate. The number of diffusers located in the tank is selected to
provide the desired mass loading of oxygen to the system and typically ranges from 1
to 7/m2 (US EPA, 1989; Metcalf & Eddy, 1991).
Activated sludge tanks are generally concrete basins that are 4.5 to 7.6 m
deep with a width to depth ratio of 1 to 2.2 : 1 to allow good mixing characteristics of
218 A.R. Bielefeldt
the water (Metcalf & Eddy, 1991). The total biomass concentration in the reactor
generally ranges from 2000 to 5000 mg/L These high biomass concentrations are
maintained in conventional activated sludge systems by operating with a longer
biomass retention time (SRT; 5-10 days) than hydraulic residence time (HRT; 5 to
10 h). A elarifier generally follows the aeration basin to settle out the biomass in the
waste water. The concentrated biomass settled at the bottom of the elarifier is
recyeled to the aeration bas in (return activated sludge). A selected portion of this
biomass is also wasted as sludge or untreated biosolids rather than being returned to
the aeration basin. If hydrophobic contaminants present in the gas are treated in an
aeration basin, adsorption to biosolids may aid gas treatment. Note that in studies
where contaminated gas has been treated in full-scale activated sludge tanks, the tank
dimensions and mass transfer characteristics for the gas injection have generally not
been reported in detaiL
Although there is a high concentration of bacteria in activated sludge tanks, a
significant portion of the suspended biomass may not be actively degrading the gas-
phase contaminants. Therefore, care must be taken if the total biomass is used to
predict reactor performance. A study with synthetic organic compounds in activated
sludge treatment found that the active fraction of biomass in a mixed system treating
a mixture of organic compounds may be estimated by the fraction of the organic
loading attributable to the individual compound of interest (Magbanua et al.. 1998).
If this is true the active biomass toward the gaseous contaminants may be small since
waste water carbon loading is often on the order of 120 to 200 mg/l (300 to 800
mg/Ld) biological oxygen demand (BOD). The concentration and type of bacteria
that are actively degrading the gas contaminants determines the biodegradation
kinetics and therefore the liquid concentrations of the contaminants. The system
should maintain low dissolved contaminant concentrations to optimise mass transfer
efficiency and prevent discharge in the eft1uent water.
The significant loading of carbon from the contaminated waste water means
that oxygen transfer will generally control the design of the gas delivery system. To
prevent the slowing of aerobic bacterial processes owing to low dissolved oxygen
(DO) concentrations, DO greater than 1 mg/l is generally desired. The relatively high
DO coupled with the low solubility of oxygen means that the mass transfer of
oxygen is inefficient. At inlet gas concentrations of 20% oxygen, eft1uent gas from
activated sludge basins generally contains 13-18% oxygen. Since most volatile
organic contaminants have a much higher affinity for the liquid phase versus gas
phase compared to oxygen (as evident by lower Henry's coefficients). the system is
generally over designed for mass transfer ofthe gas contaminants.
One critical element that should not be overlooked in the design of gas
treatment in activated sludge tanks is the influence of the liquid characteristics on
mass transfer from the gas. For example, in treating waste water it has been observed
that the oxygen mass transfer coefficient (K1a) is generally lower in the waste water
plus biomass suspension (activated sludge) than in clean water. This effect has been
characterised by 'alpha' (a), where:
K, a in wastewater
a (9.1)
KI a in elean water
Tewari and Bewtra (1982) measured alpha values for oxygen in waste water
of 0.67 to 0.925. A variety of compounds in the waste water could cause this effect
ineluding the biomass itself, surfactants (from soap and biological surfactants), ions,
etc .. Bielefeldt (1996) found that 200 mgll of methanol in tap water resulted in an
oxygen alpha of 0.63 compared to methanol-free tap water. However, the ratio of the
Kla of BTEX (benzene, toluene, ethylbenzene, xylenes) compounds to the Kla of
oxygen was the same with or without methanol. This indicates that the alpha values
for mass transfer of oxygen and BTEX compounds were the same. Roberts el al.
(1984) tested the stripping of five chlorinated organic compounds via bubble aeration
in various types of liquid. Using oxygen transfer as a metric for comparison they
found an alpha of 0.8 for oxygen transfer into filtered secondary effluent. The ratios
of the K1a of the VOCs to the Kla of oxygen were not significantly different in
deionized water and filtered secondary effluent. These relationships are shown
mathematically in Equation 9.2:
VOC Kja in elean water VOC Kja in wastewater (9.2)

/fi =
oxygen Kja in elean water oxygen Kja in wastewater
Rearranging Equation 9.2 indicates that the VOC u is equal to the oxygen u.
In any case, the mass transfer characteristics of a selected gas sparging device may
differ between clean water tests (as generally reported by manufacturers or measured
in laboratory tests) and full-scale application in systems treating contaminated water.
The characteristics of the liquid in the activated sludge system can also
influence both the Henry's coefficient (H) and the diffusivity in water (Dw) of
oxygen and contaminants. Changes in the Henry's coefficient generally relate to
changes in the solubility of the compounds due to co-solvent effects. The Henry's
coefficient is critical because it reflects the compound's affinity for the air versus
liquid phase. Diffusivity in water controls the resistance to mass transfer from the gas
to liquid.
The characteristics of the liquid in the reactor are important and should be
taken into account during design. Owing to the variability of these conditions and
difficulty quantitatively predicting their effects, bench scale tests may be required
prior to full-scale design.
2.2 SPARGED SUSPENDED GROWTH REACTORS AND BUBBLE COLUMNS
Sparged, suspended growth systems utilise a reactor design similar to activated

sludge trea1ment except that the sole function is to treat contaminated gases. Figure
9.2 diagrams the major components of a sparged suspended growth biotreatment
reactor. The contaminants in the gas serve as the growth substrate for the bacteria
unless aerobic cometabolic treatment is desired which requires the addition of a
specifically selected growth substrate. This results in nearly 100% of the biomass in
the reactor directly contributing to biodegradation of the gaseous contaminants. The
reactors are designed explicitly for gas treatment, so mass transfer can be optimised
for specific contaminants. This can allow for significantly more shallow tanks to be
used « 2 m) than is common for waste water activated sludge tanks designed for
oxygen supply to degrade organics in the waste water (4.5 to 7.6 m deep). For
example, in a model scenario outlined in Bielefeldt (1996) it was estimated that a
reactor with a liquid depth of 1 m could achieve greater than 95% removal efficiency
220 A.R. Bielefeldt
for compounds with a Henry's coefficient (H) of less than 0.4 (given an oxygen K,a
of 0.24 min-' for a QglA of 100 l/min.m2). Estimating the power required to operate
the gas blower, this shallow reactor depth requires approximately 76% less energy to
operate the blowers than a traditional activated sludge reactor. To achieve a maximal
compact design air diffusers should be located at a somewhat closer spacing with
gre ater coverage of the reactor floor area than is typical for activated sludge reactors
treating waste water. To date, full-scale suspended growth bioreactors designed
explicitly for gas treatment have not been reported in the literature.
treated gas
1-3m
O O o liquid
rising gas O depth
o bubbles o o
o O o
water with nutrients, o
pH buffer, etc. b 'be te r ia I o
,f----- suspeo"sioR
o o
o
o
I contaminated
air diffusion device
I gas
~ _ _ _ _ _ s~ttI~dl:~i0!!la~s i!:e~rn_a<:!!v~e~ sl~d9!l) _ _ _ _ _ _ _ _ j
Figure 9,2. Sparged suspended growth bioreactor.
Most suspended growth gas treatment systems are designed for aerobic
biodegradation of the contaminants with the contaminated gas supplying oxygen to
the bacteria. In these cases, it is possible that oxygen requirements rather than
contaminant removal will control the reactor design. Since the contaminants being
treated in the gas phase are frequently at dilute concentrations, the mass transfer of
the contaminants rather than oxygen may still control the design.
A variety of liquid pumping methods used in bench-scale gas treatment
systems could be applied to full-scale reactors. A mechanical mixer can be used in
addition to the air flow to achieve good mixing of the liquid and suspension of the
bacteria. Desired bacterial nutrients and buffers should be added into the in1et water
to the reactor. To treat organic contaminants, nitrogen and phosphorus should be
added at stoichiometric requirements of about 14 g N and 3 g P per 100 g of C. The
inlet water should also contain sufficient alkalinity to maintain the pH of the system
in a range that is optimal for the bacteria that degrade the contaminants. A
continuous flow ofwater or batch feeding at selected intervals can be used to operate
the system at a selected hydraulic residence time (HRT). This water flow will also
serve to remove metabolic by-products of the bioreactions that might otherwise
accumulate to inhibitory levels. Inhibitory by-product accumulation was noted by
Neal and Loehr (2000) and Bielefeldt (1996) in SGRs treating toluene and BTEX,
respectively.
Another key element in achieving an efficient system is to maintain a high
concentration of biomass in the reactor. This biomass can then rapidly remove the
contaminants from the liquid, maintaining optimal mass transfer. In a completely

mixed system the average HRT will be equivalent to the biomass retention time
(SRT). However, it may be advantageous to maintain a longer SRT than HRT; this
would result in less waste water generation and generally less waste biomass
production. The most common strategy to achieve a longer SRT than HRT is to
recirculate biomass back to the gas treatment reactor. This recyde could be similar to
conventional activated sludge systems where the liquid effluent from the SGR is held
in a quiescent basin to settle the biomass and retum it to the aerated reactor.
Altematively, the liquid in the bioreactor could be operated similar to a sequencing
batch mode where gas feed to the reactor is stopped, the biomass is allowed to settle
to the bottom of the reactor, and liquid is removed from the top of the reactor. This
configuration would work best if two gas treatment reactors were operated in
parallel, and during the settling phase of one reactor, the other handled the entire
flow of contaminated gas.
A reactor design similar to sparged suspended growth systems but using
terminology more familiar to chemical engineers is a bubble column. However, the
concept for a full-scale bubble column is dimensionally a reactor that is taller than it
is wide, rather than a 'tank' where the heightwidth ratio is less than 1. These systems
are generally operated at significantly higher gas flow rates per unit area (1.5 to 28
cm 3lcm2 -s) than traditional activated sludge systems. Chisti (1989) reported pilot-
scale studies with bubble columns having rectangular and circular cross-sections.
The rectangular column had a volume of 981,2.146 m tall, with a perforated plate
(50 holes of 1 mm diameter) sparger giving an oxygen KJa of 1.2 to 3.6 min· J with
power input of 150 to 600 W/m 3. The circular bubble column was operated with
depths of 1.5 to 5.9 m giving a liquid volume of 70 to 273 1. The column also
contained a perforated plate sparger with 106 holes of 1 mm diameter. The measured
oxygen KJa in clean water ranged from 1.5 to 5 min,J with power input increasing
from 250 to 1000 W/m3 . Generally, given the same power input per unit volume, the
oxygen mass transfer coefficient was not significantly different in the bubble column
with a rectangular versus circular cross-section. In these bubble columns, most of the
energy input resulted from gas expansion rather than the kinetic energy of the air
entering the water. This is because ofthe pressure difference at the bottom and top of
the reactor that results from the liquid depth.
A non-engineered small-scale suspended growth bioreactor that has been
used to test biological gas treatment is a traditional fermenter. Fermenters have been
primarily used to test pure cultures of bacteria for treatment of gas phase
contaminants. In most fermenters the liquid heightdiameter ratio ranges from 1 to
2: 1. Most are aerated through a perforated ring of tubing with additional mechanical
mixing. The mass transfer characteristics of the system have not been measured
and/or reported in many studies with fermenters. The results from studies in
fermenters indicate the potential to biologically treat contaminated gas using a
suspension of bacteria. However, reported poor system performance may be due to
inadequate mass transfer and do not represent the performance that could be
achieved in a properly designed suspended growth bioreactor. Mechanical mixing
should increase the aeration efficiency compared to the sparger alone due to a higher
overall energy input into the reactor. Many authors have noted increases in the mass
transfer coefficient with increases in power density (power input per unit liquid
volume) in the reactor.
222 A.R. Bielefeldt
2.3 AIRLIFT REACTOR
An airlift reactor is a type of bioreactor designed by chemi cal engineers to produce

pharmaceuticals, polymers, chemi cal feed stocks, etc. As in activated sludge reactors,
traditionally gas is introduced into an airlift reactor for the purpose of supplying
oxygen to the bacteria and to provide mixing. Airlift reactors are still not widely used
in full-scale industrial applications (-7% of products produced in airlift reactors;
Chisti, 1989) despite offering advantages over traditional stirred tank bioreactors.
However, numerous full-scale processes using airlift reactors to produce desired
aqueous products are in operation world wide, with up to 1500 m3 working volume
ofliquid reported (Chisti, 1989). Chisti (1989) reported results with pilot-scale airlift
reactors of 50 to 1000 I volume. The use of airlift reactors for biological gas
treatment has been reported in lab-scale studies (Ensley and Kurisko, 1994; Wei et
al., 1999; Ritchie and Hill, 1995) but no reports of large-scale airlift reactors for
contaminated gas treatment were found in the literature.
Airlift reactors are different from the other suspended growth systems due to
the presence of two distinct zones in the reactor liquid. In one zone the air is
introduced and liquid ris ing is induced (the riser); in the other 'ungassed' zone the
liquid and residual gas bubbles recirculate down to the bottom of the reactor (the
downcomer). Baffles in the reactor or an externalloop cause the distinct pattern of
liquid circulation. There are a variety of configurations for airlift reactors that
incorporate the general principle of a riser and a downcomer. The main types of
airlift reactors are concentric internal loop, split cylinder internal loop, and external
loop. Sketches of these reactor types showing plan and profile views are provided in
Figure 9.3. While most airlift reactors have a circular cross-section, rectangular or
square cross-sections may also be used (Chisti, 1989).
concentric·loep
split·cylinder externalloop
Figure 9.3. Airlift biorcactors.

The hydrodynamics in airlift reactors are significautly different from

activated sludge tanks aud bubble columns. In airlift systems, the liquid circulation
rate is primarily dependent on the gas flow rate (with some influence of reactor
geometry aud sparger placement). With the higher liquid veJocities, higher gasflow
rates are possible without so called 'slugging' of the gas (where the gas bubbles
coalesce into larger bubbles as they rise through the liquid). An excellent review of
airlift reactor design aud resulting mass transfer characteristics is provided in Chisti
(1989).
3. Process variables
To fully and accurately define the operation of suspended growth gas treatment
reactors, a number of parameters are typically used. They are sometimes slightly
different from those defined for other reactors used for air pollution control and
described in Chapters 3-8. These parameters can be used to estimate reactor
performance and are used in the reactor modelling equations presented in Section 4.
Since different terms are traditionally used by civil engineers for activated sludge
reactor design, environmental and mechanical engineers for gas treatment reactor
design, and chemical engineers for airlift design, a set of parameters has been
selected by the author from these various pools for use in this chapter. Definitions for
the key reactor design and operating parameters and their importance are provided
below, with correJations to terms common to the various literature sources provided.
The reported ranges of selected parameters for the different sub-classes of suspended
growth gas treatment bioreactors are summarised in Table 9.1.
Mean Gas Residence Time, tR = V g/Qg. Mean gas residence time is the
average time that an individual gas bubble is in contact with the liquid in the reactor.
This parameter is frequently not measured but commonly ranges from 1-5 seconds
depending on the bubble size and reactor depth.
Gas Hold-Up, E = V g I (V g + VI). Gas hold-up is equivalent to the gas void
fraction in the reactor which is the volume fraction of the gas phase dispersed in the
total gas-liquid phase. E has a theoretical maximum of 0.74 aud an operational
maximum in the riser of an airlift reactor of 0.3. The gas hold-up in typical activated
sludge tanks aud sparged suspended growth reactors, bubble columns, and the riser
of external loop airlift reactors, concentric internal loop airlifts, aud split cylinder
airlift reactors are approximately 0.005, 0.05-0.2, 0.02-0.2, 0.05-0.2, aud O.OS-O.l3,
respectively (Chisti, 1989).
Empty Bed Retention Time, EBRT = VI/Qg. EBRT is equal to the liquid
volume in the reactor divided by the gas flow rate. Although this parameter is not
truly useful for modelling reactor performance it is easy to calculate and frequently
reported. It also allows a basis of comparison to biofilters where the EBRT is
commonly reported.
Superficial Gas Velocity, Ug = Qg/A. Ug is the volumetric gas flow rate per
unit area of reactor. This is important when there are land or space limitations for a
given reactor. The information can also help to predict the potential for coalesced gas
bubbles versus maintaining discrete fine bubbles of gas in the reactor. Approximate
values ofUg range from 0.001 to 0.2 mls for bubble columns and 0.06 to 0.6 m/s in
the riser of airlift reactors (Chisti, 1989).
224 A.R. Biele{eldt
Mass Loading Rate, MLR = Cgin Qg ! VI. The mass loading rate is equal to the
mass of contaminants entering the reactor per volume of liquid in the reactor per
time. The higher the mass loading ratc that can be treated to the required level, the
smaller the volume of the reactor. This parameter is again useful for comparing
between different types of gas treatment bioreactors.
Specific Loading Rate, SLR = Cgin Qg ! X. Specific loading rate is the mass
loading rate normalised to the concentration of biomass in the reactor liquid or the
mass of contaminant per mass of bacteria per time. If the specific loading rate is too
high the biodegradation capacity of the bacteria in the reactor can be overwhelmed.
This will result in thc contaminant accumulating in the reactor liquid thereby
decreasing mass transfer. In addition, for some organic and inorganic contaminants
high liquid concentrations may be inhibitory to the bacteria. '1'0 prevent accumulation
of the contaminant, the specific loading to a suspended growth bioreactor should be
maintained below the specific degradation rate of the bacteria for the contaminant.
The specific degradation rate of the contaminant varies as a function of temperature,
oxygen concentration, pH, and the Iiquid concentration of the contaminant. The
specific degradation rate is discussed in more detail in Section 4 for modelling.
Overall mass transfer coefficient for the reactor, Kla (units: time- I ). Kla
describes the rate of mass transfer from the gas into the liquid. It is generally
predominated by liquid film resistance to mass transfer (gas phase resistance is
assumed negligible). The Kla includes both the mass transfer resistance (k l) and the
gas:liquid contact area for mass transfer (a). Because it is difficult to measure or
compute the total surface area of the numerous gas bubbles of varying size that are
present in the re actor (a), Kla is measured as a lump parameter. Despite the
importance of Kla for defining the gas:liquid mass transfer characteristics of a
system, it has not been reported for many activated sludge systems or fermenters. Kla
is influenced by the reactor geometry, aeration device, gas flow rate, liquid
turbulence, temperature, and the chemical composition ofthe Iiquid. For example, in
airlift reactors the height of liquid affected the oxygen Kla (Chisti. 1989). In both
bubble columns and airlift reactors, increased solids concentration decreased the
oxygen Kla (Chisti, 1989). The mass transfer coefficient is also frequently correlated
to the power input per liquid volume. More discussion on Kla is given in Section 4.
Table 9.1. Range of aeration and oxygen mass transfer coefficients for different reactors
Reactor Qg/A, Liq depth Temp Oxygen Air addition and References
type (cm/min) (m) (OC) Kla, (min-') mixing
Activated 3-30 4.5-7.6 Env. 0.17-D.38 Ceramic or US EPA, 1989
sludge membrane diffusers Metcalf & Eddy,
at standard flaor 1991
coverage
Sparged 95 0.6-1.22 22-24 1.48-1.8 Ceramic fine Romain, 1996
suspended bubblc diffuser
growth 100% flaor
(SGR) coverage
1.5-16 0.40-D.42 19-26 0.25-D.73 sparge stone Bielefeldt,1996
NR ('.'- stirring) Neal and Loehr,
2000
Bubble 50-1700 1.4-6.0 20-22 O. 18-4.2 non-porous plate Chisti. 1989
column
Airlift 120-1800 1.4-5.2 20-22 0.12-6 non-porous plate Chis ti, 1989
4. Modelling
4.1 GAS-LIQUlD MASS TRANSFER
Equations describing gas-liquid mass transfer for fine bubble aeration have been
developed based on the two-film theory (Whitman, 1923). In any reactor with a
se1ected aeration device and 1iquid volume, and operated at a given temperature and
gas flow rate, there is a definitive gas-liquid volumetric mass transfer coefficient for
each compound in the system called K,a. According to the two-film theory. in a
system with turbulent mixing, the total resistance to mass transfer across the
interfacial films is the sum of the diffusional resistance due to the liquid film and gas
film. A higher resistance is represented by a smaller mass transfer coefficient (k).
(9.3)
For many compounds, the diffusivity in gas (Da) is much larger than the
diffusivity in liquid (Dw) resulting in less mass transfer resistance in the gas film. In
addition, for compounds with a much greater affinity for the gas phase than the liquid
phase, the H is large. These two factors contribute to the second term in Equation 9.3
being negligible compared to the liquid film resistance. For example, with oxygen,
the gas film resistance is considered negligible and the overall mass transfer rate can
be related simply to the liquid film mass transfer in which the oxygen diffusion rate
affects the transfer rate to the liquid. If liquid film resistance controls the mass
transfer of both a contaminant (such as a volatile organic compound, VOC) and
oxygen, then the K,avoc should be re\ated to the K,ao2 and the ratio of the molecular
diffusivities ofthe compounds in water. This is shown by Equation 9.4,
(9.4)
where Dwvoc is the diffusion coefficient of the contaminant in water (cm 2/s) and
DW02 is the diffusion coefficient of oxygen in water (cm2/s), and n is a constant.
Depending on specific conditions in the re actor, n ranges from 1.0 to 0.5.

Under strict two-film theory, the mass transfer coefficient is a first order function of
the diffusion coefficient. Under highly turbulent conditions, the surface-renewal
model and penetration theory predict that the mass transfer coefficient will be a
square-root function of the diffusivity (n = 0.5 in Equation 9.4). Under less turbulent
conditions, n approaches 1. The most conservative approach to estimate the mass
transfer coefficient of a compound based on that of oxygen is to assume that n = 1
(this applies for ali compounds with a smaller diffusion coefficient in water than
oxygen; as a rule of thumb this includes compounds with a molecular weight greater
than 32).
The oxygen gas-liquid mass transfer capability of aerators used in suspended
growth reactors for waste water treatment has been widely studied. The oxygen K,a
226 A.R. Bielefeldt
(K\ao2) can be experimentally measured by a well established and re1ative1y simple

AS CE Standard Method for oxygen transfer in dean water; correlations to correct for
temperature effects on the K\ao2 have also been deve10ped (ASCE, 1984):
(9.5)
where T = temperature in cc, and e is a coefficient that varies with test conditions,
averaging 1.024.
As a result of chemical handling and analytic difficulties, measurement of the

K\a for contaminants (K\avoc) is significantly more challenging than measuring the
K\ao2' Therefore, the ability to estimate K\a values for any contaminant from K\a02
values is desired.
Matter-Muller et al. (1981) studied fine bubble aeration stripping of six
VOCs with Henry's coefficients (H) of 0.14 to 15.7. The experimentally determined
VOC to oxygen K\a ratio (1jI) was independent ofthe aeration system and turbulence
conditions in tests with tetrachloroethene. When ljI values were predicted using
Equation 9.4 with n = 1, these predicted ratios were within the 95% confidence
interval ofthe experimentally measured values. For example. the measured \jf values
for toluene were 0.51 and 0.54 as compared to a predicted value of 0.53. This
research indicated the ability to estimate the VOC mass transfer rates based on
measured oxygen mass transfer rates.
In work by Roberts and Dandliker (1983) studying the mass transfer of six
organic compounds (H = 0.22 to 11) and oxygen during surface aeration, the
experimental results indicated that n in Equation 9.4 was equal to 0.66. In later work
with a packed column air stripper and the same six VOCs, n was 0.5 when the gas
film resistance was negligible compared to the liquid film resistance (Roberts el al.,
1985). For liquid film resistance to be gre ater than 10 times the gas film resistance,
they reported that H must exceed 0.3 when the gas:liquid volume ratio in the reactor
was between 1 and 10. If this criteria was not met, a correction to the liquid diffusion
ratio was needed to account for the significant influence of gas film resistance on the
overall gas:liquid mass transfer rate.
In diffused aeration stripping experiments with 20 VOCs (H 0.04 to 1.3) and
oxygen, Hsieh et al. (1993) found that the K\a ratios predicted using the diffusivity
ratio (Equation 4 with either n = 0.5 or n = 1) did not agree with experimentally
measured K\a ratios for any of the compounds. For example, the measured ljI values
for toluene ranged from 0.17 to 0.37 under the different test conditions, which are
significantly lower than the values of 0.51 to 0.54 measured by Matter-Muller et al.
(1981). Hsieh attributed the variance in K\a ratios to significant gas film resistance to
mass transfer. This differs from the H values daimed for liquid film mass transfer
control of> 0.14 (Matter-Muller et al., 1981) and > 0.3 (Roberts et al., 1985). Libra
(1993) studied gas-liquid mass transfer of toluene, dichloromethane, and 1,2-
dichlorobenzene (H = 0.095 to 0.24) in a sparged, stirred tank and reported that the ljI
values for each compound were a function of both the diffusivity and the energy
input. For example, the measured ljI value for toluene increased from 0.13 to 0.66
with increasing energy input. In stripping experiments with individual and mixtures
of BTEX compounds (H = 0.21 to 0.29), the measured ljI values were within the
range ofvalues estimated on the basis ofthe liquid diffusivity ofthe compounds with
n = 1 (Bie1efe1dt, 1996). These Kla ratios did not vary significantly at different liquid
depths or gas flow rates that encompassed PIV I from 3 to 46 W/m 3 .
The fundamental differential equation that describes the change in the
contaminant (such as a VOC) concentration in the bubbles of contaminated gas
sparged into the reactor as they rise through the liquid is:
dCg(t) (9.6)
dx
where Cg is the VOC concentration in the gas (mg/l), VI is the liquid volume in the
reactor (m\ Vg is the volume of gas bubbles in the re actor liquid (m\ Klavoc is the
overall mass transfer rate of the VOC in clean water (min·\ a = constant to correct
for mass transfer in activated sludge water versus clean water, CI is the VOC
concentration in the liquid (mg/l), and t is the time the gas bubble has risen in the
column (min).
At t = O, Cg(O) = Cgm, which is the VOC concentration in the influent gas
(mg/l). When t = the total bubble retention time in the re actor, Cg (tR) = CgOll!, which
is the VOC concentration in the effluent gas (mg/l). Although the VOC concentration
in the rising gas bubbles are changing with depth in the reactor, the liquid
concentration in the reactor can be considered constant in a well mixed activated
sludge system at steady state. Under steady state conditions, Equation 9.6 can be
solved for the VOC concentration in the effluent gas, yielding Equation 9.7:
C gin exp (9.7)
where Qg = gas flow rate (m 3/min), D = the liquid depth in the activated sludge tank
(m), and A = the floor area of the tank (m2 ). This equation is equivalent to equations
used to describe bubble columns by Andrews and Noah (1995; Equation 9.15) and
airlift reactors by Cesario et al. (1995; Equation 9.1).
If biodegradation maintains extremely low contaminant concentrations in the
reactor liquid, the second terrn on the right hand side of Equation 9.7 becomes
negligible; gas treatment efficiency is then controlled by mass transfer and can be
estimated using the first term in Equation 9.7.
4.2 BIODEGRADATrON KINETrCS
The biodegradation kinetics of a compound serving as a sole growth substrate for a

bacterial culture are frequently described by Monod kinetics (see also more detailed
explanation ofthe equation in Chapter 3):
(9.8)
228 A.R. Bielefeldt
where C, = concentration of compound in the liquid (mg/l), k = maximum specific

substrate degradation rate (glg.d), Ks = the half-saturation concentration (mg/l), and
X is the biomass concentration in the liquid (mg/l). The kinetic coefficients k and Ks
are constant for a given bacterium degrading a given compound. In this case, the
compound in the liquid being biodegraded is the gas phase contaminant that has
transferred into the liquid. The specific loading rate, SLR, of the contaminant to the
reactor must be below the specific degradation rate:
Specific degradation rate (gl g.d) = (9.9)

Ks + CI
Important modifications to the Monod equation may be required depending

on the specific situation. For example, some compounds are toxic to bacteria. In this
case Equation 9.8 can be modified to include substrate toxicity by the Andrew's
model as:
dS
(9.10)
dt
where Ki is an inhibition parameter for the toxic substrate. If oxygen (or the required
electron acceptor) is limiting, a second term can be added to the Monod expression in
Equation 9.8 to reflect that removal rates become slower when oxygen is limiting
(generaIly below 2 mgll DO).
Since an activated sludge reactor is a continuously stirred tank with SRT
control, the steady state liquid substrate concentration can be described based on
Monod kinetics by rearranging Equation 9.8 (Metcalf & Eddy, 1991):
Ks (1 + b SRT) (9.11 )
CI = [SRT(Yk _ b)] - 1
where Y = the cell yield on the substrate (g biomass/g substrate), b = the biomass
endogenous decay (day"), and SRT = the operational solids retention time of the
reactor (dai'). This allows an estimation of the liquid concentration of the
contaminant under steady state conditions given the biokinetics and operating SRT.
If a group of contaminants is being treated simultaneously, there are two
scenarios that may apply. First, there may be different groups of bacteria within a
mixed culture that degrade various contaminants without direct interaction (such as
an activated sludge system). In this case no changes in biokinetics are needed to
account for treating the mixture of contaminants.
For the case where multiple contaminants present in the gas, such as BTEX,
can be degraded by the same bacteria, competitive inhibition kinetics may be used to
predict the biodegradation rates in the reactor by adding an inhibition (I) term to the
standard Monod equation. 'I' is based on the steady state concentrations of
competing compounds in the reactor Iiquid (mg/l) and the Ks values ofthe competing
compounds (the vaIidity ofthis model was shown in Bielefeldt and Stensel, 1999b).
For two compounds that competitively inhibit one another and only one serves as a
growth substrate (such as TCE and toluene):
Ks 1 + b SRT (1 + ~ITCE)
STCE
(9.12)
SRT (kY - b) - 1
Under steady state conditions, the mass transfer rate ofthe contaminant out of
the gas into the liquid must equal the biodegradation rate and washout rate:
(9.13)
The biomass concentration at steady state is a function of the amount of

growth substrate removed, the synthesis yield, endogenous decay rate coefficient,
and the re actor SRT.
Qg Y (C g 111 - C go",) (9.14)

X=
VI (S~T + b)
In addition, at steady state the dissolved oxygen concentration (DO in mg/l)
in the reactor is constant so the oxygen transfer rate into the liquid is equivalent to
the oxygen consumption rate. The oxygen consumption rate is re1ated to the
contaminant biodegradation rate, Y, and b:
CgO,ill(l- expA) - [(Cgin - Cguut)(R - 1.42Y)] - (1.42bX ~J

DO = (9.15)
Ha, (1 - exp A)
where A = - aK,ao2 D AI Qg H02 ' and R = the stoichiometric ratio of mg oxygen

required for 1 mg of contaminant oxidised.
4.3 MODEL PREDICTED GAS TREATMENT
The model presented above was validated in a number of experiments in a 2 1 lab-

scale SGR treating BTEX compounds (Bielefeldt, 1996; Bielefeldt and Stensel,
1999). The model predicted treatment of individual BTEX compounds and mixtures
of BTEX compounds matched well with experimentally measured values.
The model can be used to aid in reactor design, specifically when selecting
the operating retention time and reactor depth required based on selected aeration
equipment. Using literature reported values for diffusivity in water, Henry's
coefficient, and biokinetics, estimates of reactor design can be made.
230 A.R. Bielefeldt
A test scenario was developed to illustrate the use of the steady-statc

suspended growth reactor design. In general, the test scenario illustrates the reactor
depth and SRT required to achieve greater than 98% treatment efficiency of various
inlet concentrations of toluene and naphthalene at 25°C. Representative biokinetic
coefficients from the Monod or Michaelis-Menten model were selected from the
literature (Bielefeldt, 1996). The gas flow rate per unit area (QgI A) and resulting
oxygen K1a are based on fine bubble aeration tests in lab and pilot scale columns by
Romain (1996) and Bielefeldt (1996). Results are summarised in Table 9.2.
First, the required gas treatment efficiency was used to calculate the
alJowable effluent gas concentration. Next, the Henry's coefficient was used to find
the maximum alJowable Jiquid concentration (C1m<lX = CgOll! IH). Using this maximum
!imit as a guide, a lower desired liquid concentration of the contaminant was
selected, such as 10% Cl max ' Then Equation 9.11 was used to compute the SRT
needed to maintain the selected liquid concentration. Next, Equation 9.4 was used to
estimate the Kla of the contaminant from the oxygen K1a of the selected aeration
device at a given gas flow rate per unit are a (a for the organic compound and oxygen
were both assumed to equal 1, and n = 1). Then Equation 9.7 was used to calculate
the depth of the re actor required to achieve the selected treatment efficiency. The
biomass concentration was estimated using Equation 9.14, which allows an estimate
of biomass wasted that must be handled in some way. This also allows a feasibility
check with typical biomass concentrations. Finally, thc oxygen requirement was
checked (Equation 9.15) to ensure that oxygen transfer would not control the design
depth of the reactor rather than the mass transfer of the contaminant. If the DO
dropped below 0.1 mg/l it is assumed that the reactor depth should be re-designed to
alJow for sufficient oxygcn availability.
The results of the model shown in Table 9.2 illustrate important aspects of
design and operation of sparged suspended growth gas treatment bioreactors. Very
high gas treatment efficiency cannot be easily obtained with very low inlet gas
concentrations due to a minimum liquid concentration needed to maintain biomass in
the reactor. For example, the 321 day SRT for 0.1 mg/l inlet toluene ilJustrates that
this condition is near 'washout' where there is not enough biogrowth at the low
liquid concentrations to maintain an active bacterial culture in the reactor. Depths
required to achieve 98% removal oftoluene are 1 to 1.2 m using the ceramic sparger,
reactor design, and gasflow rate of the Romain (1996) re actor, compared to 0.55 to
0.6 m depths using the sparger and reactor dimensions of the lab-scale reactor
reported by Bielefeldt (1996). To achieve 99.8% removal with similar reactors. 1.8 to
1.9 m and 1 to 1.7 m reactor depths are needed, respectively.
Owing to the high affinity of naphthalene for water (Iow H) very shallow
liquid depths of 0.1 m can be used to achieve adequate mass transfer of naphthalene.
However, insufficient oxygen transfers into thc liquid for the aerobic bacteria except
at the lowest inlet naphthalene concentration modeJled of 0.1 mg/l. This is evident by
steady-state dissolved oxygen concentrations in the Jiquid of less than 0.1 mg/l.
Depths of 0.15 m and 1.3 mare needed to transfer enough oxygen to biotreat inlet
naphthalene concentrations of 1 and 10 mg/L respectively.
In general, the biomass concentrations required to biotreat 0.1 to 10 mg/l
toluene or naphthalene in the inlet gas are significantly lower than is common in
municipal waste water activated sludge reactors. This indicates that the amount of
biomass wasted from the reactors would be minimal, resulting in a small amount of
additional costs to ireat wasie liquid from the bioreactor.
Table 9.2. Model estimated gas treatment
Compound Qg/A VOCKla Biokinetics Cgin, eg out, CI, SRT, D, X, (mg/I) DO,
(crnlmin) (min") (mgll) (mgll) (mgll) (d) (m) (mg/I)
Toluene K = 1.0 glg.d 0.1 0.002 0.003 321 12 20 8
H = 0.25 K,=0.18mgll 0.02 0.008 41.3 l.l 30 7
Dw= 1.09 100 0.93 Y = 0.8 g/g 0.02 0.02 14.3 1.15 10 8
b = 0.01 dai' 10 0.2 0.08 4.2 1.1 30 <0.1
10 0.02 0.02 14.3 1.76 65 I
k = 1.0 g/g.d 0.1 0.002 0.008 ** ** ** ** ::....
K, = 0.18 mgll 0.02 0.03 43 4 C)
I 70 1.2
100 0.93 Y = 0.8 g/g 0.02 0.06 1.45 5 ~.
10 7 t:)
b = 0.1 dai' 10 0.2 0.08 6.8 LI 49 <0.1 ~
10 0.02 0.04 22 1.9 91 <0.1 t:l...
100 1.29* k = 1.0 g/g.d 0.1 0.002 0.003 32\ 0.85 29 8 S:2
::::
18.6 0.36 K, = 0.18 mg/l 0.1 0.002 0.003 32\ 0.6 7 8 ~
18.6 0.36 Y = 0.8 g/g la 0.02 0.06 5.3 1.0 8 3 '"
b = 0.01 dai'
0.\ 0.002 0.\ \0.\ 0.16 146 7
0.02 0.1 10.\ 0.10 83 <0.1
Naphtha1ene 100 0.73 k = 0.45 glg.d 0.02 0.2 6.3 0.11 52 <0.1
H = 0.02 K, = 0.25 mgll 10 0.2 0.\ 10.\ 0.10 281 <0.1
Dw= 0.86 Y = 0.85 g/g la 0.2 3.4 0.10 833 <0.1
b = 0.01 dai'
I 0.004 0.1 la 0.15 56 1
10 0.014 0.9 3.6 1.3 22 0.9
* estimated from K1a oxygen ifn = 0.5 in Equation 9.4
** washout condition: cannot achieve such a low substrate concentration; b must be less than 0.03 dai l
tv
'-
"'"
232 A.R. Bielefeldt
4.4 POWER
Since one of the primary operating costs associated with suspended growth
bioreactors is the energy to operate the blower, it is important to estimate the power
requirements ofthe system. From Metcalf & Eddy (1991):
Pw= Qgpa RT[( L2J 0.283 -11 (9.16)

0.283 e pl
where Pw = power, pa = density of air, e = efficiency of blower, p2 = outlet pressure,

pl = inlet pressure, R = universal gas constant, and T = absolute temperature. The
efficiency of most blowers is about 0.7 to 0.9. In general the inlet pressure of air into
the blower is near atmospheric conditions and the effluent pressure is the amount
needed to overcome the headloss in the aeration piping, across the diffuser, and head
of water in the tank. This equation makes it easy to estimate the amount of energy
that will be used.
5. Experience treating contaminated gases
The first reported full-scale systematic treatment of contaminated gases in an

activated sludge system began in 1959 at the Hyperion waste water treatment plant in
Los Angeles, CA, USA (pomeroy, 1982, 1963). Odour gases recovered from the
headworks and sludge handling facilities were pumped into existing activated sludge
tanks where municipal waste water was being aerobically treated. The contaminated
air served as both the oxygen source for the bacteria degrading waste water organics
and removed the odorous compounds. Since then, numerous municipal waste water
plants in the U.S., Canada, and Japan (over 28 reported in published literature) have
treated odorous gases by bubbling into activated sludge tanks. In general the systems
were not designed for optimal odour treatment, but rather for fortuitous odour
removal in a system optimised for oxygen transfer and biodegradation of the waste
water organics. Full-scale treatment of other types of contaminated gases in similar
systems using suspended bacteria (but not co-treating contaminated water) are not
reported in the literature. However, there are numerous reports of pilot-scale and
bench-scale reactors that have successfully treated a range of inorganic and organic
contaminants. Table 9.3 summarises the different gas contaminants that have been
treated in activated sludge or suspended growth systems. The compounds are
grouped into inorganics and organics (including aromatic and chlorinated). The
results ofthese studies are covered in more detail in the subsequent sections.
5.10DOURS
5.1.1. Full-scale treatment in activated sludge

Odorous gases can be re1eased from a variety of sources at municipal waste water
treatment plants including the headworks, open basins, and sludge processing
facilities. The odorous gases generally contain a variety of organic and inorganic
compounds including hydrogen sulphide (rotten egg smell; HzS), ammonia (NH4),
and sulphur-containing organics (such as mercaptans that smell like skunk). This
complex mixture of gases is frequent1y captured to prevent uncontrolled release.

Treatment efficiencies of various compounds in sewer gases by the City of Los
Angeles in activated sludge basins were 90-98% for H2S, > 99% for methyl
mercaptan, > 98% for dimethyl sulphide, 98% for dimethyl disulphide, and > 94%
ED50 for odour (Haug, 1993). Adsorption of the odour compounds onto the biomass
in the activated sludge tank was cited as one of the removal mechanisms at work, in
addition to biodegradation (Haug, 1993).
The use of odorous gases to aerate activated sludge tanks is a simple and low
cost treatment method. The primary advantage of activated sludge treatment
compared to alternative gas treatment methods is the effectiveness achieved using
existing facilities, which is especially important in locations where land area is
limited. To date, there have been 29 waste water treatment plants reported in the
literature that use or have used activated sludge for odour control in the U.S., Canada
and Japan (Bowker, 1999; Bowker, 1996; Ando, 1980; Ryckman-Siegwarth and
Pincince, 1992). In general, few details on these systems have been provided. A
sampling of the locations that treat odorous air in activated sludge basins is provided
in Table 9.4. Note that in many cases, very large gas flow rates are being treated.
These odorous gases are either used as the only source of air to the aeration basins or
are mixed with fresh air prior to entry into the activated sludge basins.
Table 9.3. Summary of compounds treated in suspended growth gas treatment bioreactors
Compound(s) H (25°C) Mixture Reactor type References
Odour, including: YES

H 2S 0.385 Municipal ww 1,2,3
Ammonia 0.0025 activated sludge;
Amines 0.2-7.10.4 SSGR 4
Mercaptans O. 13 2
H2S 0.385 8 alone Fermenter 6,7, 19
S02 0.03 8 alone Fermenter 6
NO 21 8 alone Fermenter 6
Methanol (MeOH) 3.6.10'6 DCM & propanol SSGR 9
Ethanol (EtOH) 1.2.10.3 5 alone Airlift 10
2-Propanol 5.8.10.4 DCM & MeOH SSGR 9
Phenol 1.6.10.5 alone Airlift 11
p-cresol 2.9.10. 5 5 alone Airlift 10
Benzene 0.22 alone and BTEX SSGR 12
Toluene 0.28 alone and BTEX SSGR; fermenter 12, 13; 14
o-Xylene 0.22 alone and BTEX SSGR 12
Ethylbenzene 0.32-0.35 alo ne and BTEX SSGR 12
p-Xylene 0.26 BTEX, TX SSGR 15
Trichloroethene 0.37 toluene (gas) fermenter 14
(TCE) phenol or tol (Iiq) airlift 16, 17
Dichloromethane 0.10-0.13 alone airlift 18
(DCM) MeOH+propanol SSGR 9
References: IOstojic el al., 1992; 2Bowker, 1996; 3Ryckman-Siegwarth el al., 1992; 4Romain,
1996; sCard, 1998; 6Dasu el al., 1993; 7Basu el al., 1996; 8CRC, 1998; 9Vanderberg-Twary el
al., 1997; \OWei el al., 1999; IIRitchie and Hill, 1995; 12Bielefeldt and Stensel, 1999; 13Neal
and Loehr, 2000; 14Landa el al., 1994; 15Bielefeldt, 1996; 16Ensley and Kurisko, 1994;
17Hecht el al., 1995; 18 Zuber et al., 1997; 19Sublette and Sylvester, 1987.
234 A.R. Bielefeldt
Table 9.4. Full-scale waste water treatment facilities treating odorous air in activated sludge
basins.
Location WWTP Source of odorous Q,Odour Activated Removal

liquid air a1r sludge efficiency
tmt treated parameters
capacity (ml/min)
(m'/d) (cfm)
Toyohiragawa 118200 sludge heat NR NoI reported 99% odour
WWTP; Sapporo trealmenloff-gas
Cit):', Ja12an(l)
Concord WWTP; 39000 sludge holding 19.8 Coarse diffusers 95% odour;
New Hampshire, tank (700) 3-m deep 92% H 2 S
USA(2) Fine diffusers > 99.5%
3-m dee12 odour & HzS
Valley Forge; 30000 influent chamber, 62.3 4.3 m deep > 99% odour
Pheonixville, PA, 10 c1arifiers (2200) membrane 99.9% H 2S
USA(2) diffusers
Springfield, MA, 254000 compost facility 127 8 fi deep coarse > 99% odour
USA(3) for 10 and 2° (4500) bubble diffusers
sludge
Reedy Creek 34000 in vessel 159 coarse diffusers > 98 10
WWTP Orlando, composting (5600) 100% odour
FL, USA(5) re actor
Hyperion; 1.59.106 Headworks, 5098 19% odorous air > 83% H2 S
Los Angeles, CA, influent channels, (0.18M) + fresh air; fine 96-99%
USA(4) primary c1arifier bubble diffusers odour
80% VOCs
Long Beach; CA, 95000 primary c1arifier 227 coarse bubble 100% H 2S
USA(4) (8000) diffusers removal
Glendale, Los 76000 bar screens, 566 originally coarse N ot reported
Angeles, CA, influent pump (20000) bubble,
USA(2) station, primary switched to fine;
c1arifiers 100% odour air
References: (l)Ando, 1980; (2)Bowker, 1999; (3)Ostojic el al., 1992; (4)Ryckman-Siegwarth
and Pincince, 1992; (5)Stillwell el al., 1994.
NR = not reported
5.1.2. Potential problems in full-scale treatment

In general, few problems have been reported with treating odorous gases in activated
sludge basins. Many plants have operated for over 20 years without significant
problems (Bowker, 1996; Bhattarai, 2000). The primary concern is the corrosion of
blowers and piping materials owed to H2S. In treating odorous gases (with the
primary chemi cal of concern being H2 S), a few cases of steel corrosion of inlet filters
and air piping have been reported (Bowker, 1999; Ryckman-Siegwarth and Pincince,
1992). However, there are almost no reports of actual internal aeration blower
corrosion (Bowker, 1999). At one location (not pumping into activated sludge), the
aluminium lobes pitted and steel guide vanes corroded but this was attributed to the
lack of a moisture trap. New activated sludge processes being designed for co-
treatment of contaminated odour gases have selected corrosion-resistant materials,
the most common being stainless steel and plastic (high density polyethylene [HDP]
or polyvinyl chloride [PVC]). These materials have been used at Terminal Island Los
Angeles CA, Bondi Island Springfield MA, Lowell MA, Los Coyote CA. (Ryckman-
Siegwarth and Pincince, 1992).
Diffuser plugging is also a concern but has rarely been reported. Because of
this concern, Ostoj ic et al. (1992) recommends that odorous gases be treated via
coarse bubble diffusers to prevent c10gging or fouling of fine diffusers. However, it
is not apparent that this is necessary. Of the information available 16 plants used
coarse bubble aeration and 14 used fine bubble aeration (note that some plants have
used both, in general switching from coarse to fine when facilities were upgraded).
No problems have been reported more frequently using either diffuser type. It is
important to note that diffuser fouling can also occur in activated sludge tanks
sparged with c1ean air (instead of contaminated air).
It appears that fairly typical blower and sparger maintenance can minimise
any potential problems of treating odour gases. The accumulation of sludgy material
in air pumps has been reported in some cases (Lexington KY, Long Beach CA, Los
Coyote CA, LowelI MA, Ventura CA; Ryckman-Siegwarth and Pincince, 1992). It is
possible that smalI quantities of this material actualIy he1p to protect the blower
materials from corrosion (Bhattarai, 2000).
5.1.3. Pilot-scale treatment in a sparged suspended growth reactor

A pilot-scale study for odour treatment in a sparged suspended growth reactor was
conducted at the Renton, Washington, waste water treatment plant by Romain
(1996). The reactor was constructed of a 19 cm diarneter x 1.4 m talI Plexiglas
column with a ceramic fine bubble diffuser covering the entire floor of the re actor.
The measured c1ean water oxygen mass transfer coefficient in the reactor averaged
l.6 min- 1. Contarninated gas treated in the reactor originated from the headspace over
a dissolved air flotation waste sludge thickening system. A 0.09 kW (1/8 hp)
diaphragm pump supplied air to the gas treatment reactor at approximately 1 m 3/min-
m 2 . During long-term gas treatment the reactor contained 35 1 liquid at a 1.27 m
depth. Biomass in the reactor was initialIy seeded with activated sludge, and twice
per week I I of the re actor content was removed and replaced with secondary
effluent from the waste water treatment plant (nitrogen, phosphorus, and bicarbonate
buffer were added as needed). After 1 month the reactor appeared to have reached
approximate1y steady conditions at a biomass (as VSS) concentration of 250 mg/!.
The approximate SRT in the reactor was 123 days. Hydrogen sulphide, ammonia,
mercaptans, and amines were measured in the influent and effluent gas from the
reactor over a 1.5 month period. The operating temperature in the reactor ranged
from 11 to 22 ac, with most temperatures between 13 and 17 ac. The pH in the
reactor was maintained between 7.8 and 8.4. Influent concentrations averaged 2.7,
0.6, 0.2, and 0.3 ppmv for HzS, amines, ammonia, and mercaptans, respectively
(approximate1y 0.004 mg/l HzS and 0.0001 mg/l ammonia). In alI cases, the
concentrations of these compounds in the effluent gas were below the method
detection limit of 0.1 ppmv.
At the end ofthe 2.5 month period of continuous operation the liquid depth in
the reactor was decreased by settling the biomass in the reactor and removing the top
portion, thereby maintaining nearly equal total biomass quantity in the reactor during
the tests. Short-term tests were conducted where the influent and effluent gas
concentrations were measured once at each depth. At inlet concentrations ranging
from 1.8 to 5, 0.1 to 0.25, 0.1 to 0.3, and 0.25 to 0.5 ppmv for H2S, amines,
ammonia, and mercaptans, respectively, effluent concentrations of alI the compounds
were below the detection limit of 0.1 ppmv at liquid depths between 1.3 and 0.6 m.
Although computed removal efficiencies were low due to the smalI differences
236 A.R. Bielefeldt
between inlet concentrations and the detection limits, successful treatment of odour
compounds was achieved in the specifically designed SGR at liquid depths much
shallower (down to 0.6 m) than traditional activated sludge tanks used for full-scale
treatment (4.5 to 7.6 m).
When tests were conducted with unacclimated biomass (35 I of activated
sludge; depth 1.27 m) in the same reactor, concentrations of H2S were reduced from
4 to 5 ppmv to < 0.1 ppmv, but detectable concentrations of amines, ammonia, and
mercaptans were present in the effluent gas, representing 20 to 43%, O to 44%, and
47 to 90% removal, respectively. This indicated that a bioconsortia acclimated for
degradation of odour compounds was critical to achieve good gas treatment
efficiency in the reactor. With the unacclimated biomass, biodegradation, rather than
mass transfer, limited the gas treatment efficiency.
5.1.4. Comparison to other reactor types

Despite the 10ng-term success in many locations for treating odour gases in activated
sludge this method is still not widely used. Biofilters are commonly used to treat
odour gases at municipal waste water treatment facilities (Williams and Miller, 1992;
Torres et al., 1997). However, reliable performance of biofilters requires careful
control of moisture and pH. Tests were conducted by Torres et al. (1997) treating
headworks off-gas from a municipal waste water treatment plant in Fountain Valley,
California, in bench scale biofilters containing yard waste compost, sewage sludge
compost, granular activated carbon, or zeolite as support media. All biofilters
removed 100% H2S (from inlet concentrations averaging 3.1 ppmv) but the removal
of organics (measured as total non-methane organics; average inlet 26 ppmv) was
only 79-83% for ali media types. The numbers of other studies with biofilters are too
numerous to summarise here, but good references include Sklandany et al. (1998),
Torres et al. (1997), and other studies cited in Chapter 3.
Bioscrubbers have also been used for odour treatment (Joyce and Sorensen,
1999; Joyce and Sorensen, 1998) (see also Chapter 5). The processes described were
a combination of biomass attached to solid media and suspended in liquid trickled
through the reactor. Joyce and Sorensen reported that these systems have
successfully removed H2 S from odour gas reducing concentrations from 1000-2000
ppmv to zero under autotrophic conditions (no supplemental carbon added). Removal
of organic odour compounds ranged from 70-85%. The use of existing activated
sludge treatment facilities on site is an equally attractive alternative for gas treatment.
If odour gas flow rates collected from the facilities exceed the aeration demand of the
activated sludge tanks, design of a sparged SGR specifically for gas treatment may
provide reliable and inexpensive treatment in a shallow depth system.
5.2 INORGANICS
The primary use of biologic al gas treatment for inorganic compounds has been
proposed for treating sulphur based contaminants associated with fuels. This includes
both desulphurisation of natural gas (Basu et al., 1996) and treatment of S02 in
combustion off-gases (Dasu et al., 1993). The presence of NO in combustion off-
gases may also be treatable in a suspended growth bioreactor (Dasu et al., 1993).
Treatment of these three inorganics has been studied in lab-scale fermenters. Reactor
operating conditions are summarised in Table 9.5 and the results are discussed in
more detail in the following subsections.
5.2.1. Hydrogen sulphide (H2S)

Low quality natural gas (LQNG) makes up a large fraction of known reserves. This
LQNG requires treatment to remove N 2, HzS, and CO 2 prior to use. In general, the
H2S and CO 2 are removed by amine scrubbers and are then stripped out to regenerate
the scrubbing solvent. The remaining gas stream can be expensive to treat. Basu et
al. (1996) proposed suspended growth biological treatment to remove HzS and caz
from gas streams. In general these studies were conducted in batch fermenters with
poorly characterised mass transfer efficiency.
Table 9.5. Treatment of inorganics in lab-scale suspended growth bioreactors
Cmpd Liquid Qg/A MLR SLR EBRT K,a Inlet Removal Temp
depth, (cmlmin) (mg/ (g/g VS.d) (min) min- I conc, % (DC)
(m) l.h) (t R, s) VOC (mg/I)
(0,2
H2S(1) 0.2 0.3 9.8-20 0.59-1.2 47 NR 7-15 >99 30
(1-2)
H2S(2) 0.15 0.3-1 56-74 NR > 30.8 NR 38 100 30
56-95 > 24' 38 100
NO(3) 0.2 0.6 8 1.5-8.5 50 NR 6.4-6.7 > 96,10030
S02(3) 0.2 0.6 8-33 0.3-0.6 6.8 NR 0.9-2.7 > 95 30
References: (I)Sublette and Sylvester, 1987; (2)Basu el al., 1996; (3)Oasu el al., 1993; '" with
su!phur removal from the \iquid.
Numbers in italics indicate values not directly reported, that were estimated.
NR = not reported.
A variety of bacteria have the ability to transform hydrogen sulphide. These

bacteria have a range of growth requirements and reaction by-products. Basu et al.
(1996) used Chlorobium thiosu/fatophilum to transform H2 S to elemental sulphur
(So). Under anaerobic conditions, this bacteria uses CO 2 for carbon (autotrophic
metabolism) and light and H2S oxidation for energy (phototrophic). The relevant
reaction is:
2 H2S + CO 2 -light~ 2 S + CH 20 + H20

biomass
The sulphur formed by C thiosu/fatophilum is deposited outside the ceH

membrane which makes it easy to recover from the reactor. It was also found that N2
and CH4, common co-contaminants with H2S, do not interfere in the reaction. The
need for light to drive the reaction dictates reactor design constraints and an
additional energy requirement for 24 hour operation.
Studies for continuous H2S treatment by C thiosu/fatophilum were conducted
in a fermenter containing 1.25 1 of culture at 30°C and operated with a 20-hour liquid
retention time. Two, 200 Watt tungsten bulbs provided light to the reactor.
Contaminated gas contained 2.5% (25000 ppmv) H2S, 10% CO 2, 15% CH4, and
72.5% He. The gas flow rate was varied to give EBRT values of 14 to 40 minutes. At
EBRTs of 30.8 min and above, 100% H2 S removal was achieved, decreasing to 58%
removal at a 13.7 min retention time. When the reactor was operated with side
treatment to remove elemental sulphur from the reactor Iiquid, 100% H2S removal
238 A.R. Bielefeldl
was achieved at EBRTs of 23.9 min and above; 77% removal was achieved at an
EBRT of 14 min. This study demonstrated successful treatment ofH 2S.
An alternative sulphur-oxidising bacterium, potentially useful for natural gas
desulphurisation, was studied by Sublette and Sylvester (1987) and Dasu el al.
(1993). Thiobacillus denitrificans is a strict autotroph and facultative anaerobe.
Under anaerobic conditions, nitrate will be used as an e1ectron acceptor and reduced
to nitrogen (N 2 gas). If sulphide is limiting, the bacteria will also grow anaerobically
on sulphur compounds, converting S-2 or S03-2 to So or S04-2. Soluble sulphide is
toxic to the bacteria so careful balancing in the reactor must maintain sulphide levels
below the toxicity threshold (~ 150 mgll).
Gas containing 5000 to 10000 ppmv H2S and 50000 ppmv CO 2 in nitrogen
(N2) was added at approximately 30 ml/min via a ring sparger into a fermenter
(Sublette and Sylvester, 1987). The fermenter contained 1.4 l of a T denitrificans
suspension in nutrient media (which initially contained 5000 mg/l KN0 3). Batch
tests (~36 h) were conducted at 30°C with biomass concentrations ranging from 200
to 450 mg protein/l. In tests with an H2S SLR of 3.3 to 4.1 g/g.d, effluent
concentrations were < 0.07 mg/l H2S when the reactor was stirred at 200 rpm (> 99.1
to 99.55% removal). These effluent H2S concentrations dropped below the method
detection limit when the fermenter was stirred at 300 rpm (>99.95-99.96% removal).
In the batch liquid, H2S was converted into S04, biomass growth occurred, and both
NH4+ and nitrate were consumed. The strong growth of biomass in the system
implies that long-term treatment with T denitirificans would be possible if liquid
flow through the system maintained adequate nitrate and ammonia availability.
Somewhat large quantities of nitrate are needed, averaging 1.6 moi N0 3· per moi
H2S. (Sublette and Sylvester, 1987).
Under higher H2S loadings of 4.5 to 8.9 g/g.d, reactor upsets occurred as
evidenced by some H2S conversion to So, sulphide accumulation in the reactor
liquid, and N20 detected in effluent gas. Under these conditions, H2S removal
efficiency dropped to 96% to 87%. Reactor recovery was possible if the cui ture had
been exposed to the high soluble sulphide concentrations for less than 2 to 3 hours.
(Sublette and Sylvester, 1987).
Other bacteria have reportedly been used for treatment of H2S contaminated
gases, but no direct information on these systems could be obtained. Sublette and
Sylvester (1987) cite a study where Thiobacillus lhioparus was used to remove H2S
from gas that was bubbled through the culture (from G. Yu. Ass and B. Sh. Shpiner,
USSR Patent No. 986,469, 1983). Sublette and Sylvester (1987) also cite the use of
mixed cultures of Beggialoa and Thiolhrix to treat H2S gases in a similar manner
(from Nishihara Environmental Sanitation Research Corporation, Japanese Patent
No. 57,170,181,1982).
5.2.2. Sulphur dioxide (SO.)

Sulphur dioxide (S02) is a common contaminant in flue gas that results from burning
low quality coal contaminated with sulphur. It is also present in other combustion
gases including solid waste incinerator gases. Concerns about acid rain led to strict
regulations of S02 in the U.S. under the Clean Air Act. These regulations require that
sulphur compounds must be removed from the off-gas from combustion processes.
In many cases, NOx may be a co-contaminant in these gas streams.
In general, bacteria capable of sulphur-reduction require a carbon source and
strict anaerobic conditions. In biological gas treatment studies by Dasu el al. (1993),
a mixed culture of Desulfovibrio desulfuricans (A TCC 13541) and approximately

50% heterotrophs was utilised. The heterotrophs maintain the absence of dissolved
oxygen in the reactor liquid, which provides the low redox that al!ows sulphur-
reducing bacteria to thrive. This is required since most combustion off-gases contain
residual levels of oxygen that would otherwise prevent sulphur reduction. It is also
unique because most biological gas treatment reactors are designed for aerobic
treatment. Glucose or pre-treated sewage sludge (pre-treated with alkali and heat to
release organics into soluble form) was used to supply carbon to the bioreactor. The
study was conducted in a fermenter containing 1.5 L of the suspended culture. The
inlet gas contained 5 moles of CO, per moi of SO" the balance being N, (about
99.8% ofthe gas flow was N,; used to strip out the H,S formed). At a loading rate of
10 mg SO,/l.h this system achieved nearly 100% conversion of the inlet SO, to H,S.
The average total biomass in the reactor was 5000 mgll MLSS. Given the MLSS of
the inlet sewage, this represented an increase of about 500 mgll with the
consumption of 1500 mg/I carbon (soluble COD) from the sewage. Although the
system worked well, the specific SO, reduc ing capability of the culture was higher
when fed glucose as a carbon source (which was converted into ethanol used by the
D. desulfuricans) rather than treated sludge.
In the same study (Dasu et al., 1993), the fermenter was operated with a
culture of Desulfotomaculum orientis (ATCC 19365) and H, utilising bacteria.
Initial!y, lactate was fed as a carbon source. The inlet gas contained SO, (to serve as
a terminal electron acceptor; 1 mole equivalent), CO, (used as a carbon source; 77.9-
98.3 mole equivalents), and H, (to provide energy; 933-729 mole equivalents). In
addition, N, gas was fed at 1479 to 1867 times the SO, moi feed rate to strip out H,S
generated in the reaction. At an SO, feed rate of 8 to 10 mg/l.h the system achieved
greater than 98% SO, reduction to H,S.
5.2.3. Nitric oxide (NO)

Dasu et al. (1993) reported studies in fermenters with numerous bacterial types to
remove nitric oxide (NO) from contaminated gases. Nitric oxide can be converted to
nitrogen gas (N,) by many types of denitrifying bacteria.
Thiobacillus denitrificans (ATCC 23642) can use N0 3 or NO as a terminal
electron acceptor, thiosulphate (S,03 -2) as an energy source, and uses CO, for cel!
carbon. The stoichiometry measured for these reactions was I moI S,03-2 for I moi
NO reduction. The suspended growth gas treatment reactor achieved 96-98% NO
removal of an inlet concentration of 4800 ppm NO loaded at a rate of 42 mg NO/l.h.
Increased loading up to 65 mg NO/l.h reduced the treatment efficiency below 79%.
T denitrificans used ammonium for biomass formation and about 380 mg biomass
grew per 5700 mg NO reduced.
Further NO treatment tests were conducted with Parococcus denitrificans by
feeding succinate or pre-treated municipal sewage sludge (alkali and heat treated) as
a carbon source (Dasu el al., 1993). In short term tests when inlet gas containing
0.5% NO, 5% CO" and 94.5% N, was fed at 30 ml/min into a 2 I fermenter,
complete removal of NO was achieved. The specific NO loading in these tests
ranged from 1.6 to 8.5 g NO/g.d. The treatment efficiency was 12 mg NO/h with 6
moi NO reduced per moI succinate utilised. Although identical treatment efficiency
was obtained with either carbon source, NO breakthrough did occur with some
amount of organics (measured as soluble COD) from the sludge stil! present but
apparently non-bioavailable.
240 A.R. Bielefeldt
In similar tests by Dasu et al. (1993) with Pseudomonas denitrificans, growth

on yeast extract or pre-treated municipal sewage sludge under the same NO loading
conditions resulted in similar NO treatment efficiency. The mass transfer
characteristics in the fermenters operated by Dasu el al. (1993) to study S02 and NO
treatment were not reported; since nearly 100% removal was achieved, the systems
were apparently not mass transfer limited.
5.3 AROMA TIC ORGANICS
Aromatic organic compounds are frequently toxic to humans and therefore many are
c1assified as Hazardous Air Pollutants (HAPs) by the U.S. Environmental Protection
Agency. Due to health concems these compounds are regulated for rele ase into the
environment. The removal of a wide variety of aromatic organic compounds has
been studied in lab-scale suspended growth bioreactors. The BTEX compounds,
which are present in gasoline and solvent vapours, have been the most wide1y
studied. The studies achieving successful treatment of aromatic organics are
summarised in Table 9.6.
Some ofthe results shown in Table 9.6 were generated during short-term tests
where the biomass was not allowed to accumulate and the system was not operating
at steady state. The specific loading in these tests was generally high, such that the
contaminants accumulated in the liquid. Using the second term on the right hand side
of Equation 9.7, it was estimated that 83, 28, and 63 percent of the benzene,
ethylbenzene, and o-xylene in the effluent gas was owed to contaminant
accumulation in the reactor liquid, respectively. This indicates that longer tests
aUowing biomass selection and accumulation could achieve significantly improved
removal efficiency for benzene and o-xylene (Bie1efeldt, 1996).
The types of bacteria used to treat the aromatic organic compounds varied. In
general the aromatic organics studied are all readily biodegradable by a wide variety
of aerobic bacteria which mineralise the compounds to carbon dioxide and water. For
the Landa el al. (1994), Ritchie and Hill (1995) and Wei el al. (1995) studies, pure
cultures of Burkholderia cepacia G4 (formerly Pseudomonas cepacia) and
Pseudomonas pUlida strain IlO (ATCC 17484) were used. Although no mention of
changes in the culture over time was made, it seems like1y that long-term operation
under full-scale conditions would not maintain a pure culture in the bioreactor. Using
a defined mixed culture, Kennes el al. (1993) found that most wild type strains
inoculated into a non-sterile fermenter treating gas contaminated with chloroalkanes
and chlorobenzene were washed out after several weeks. Mixed toluene degrading
bacteria were used in the Neal and Loehr (2000) studies. A culture predominated by
a filamentous bacterium (A-l, ATCC No. 55581; but not pure) and grown on phenol
prior to introduction into the gas treatment reactor was used in studies by Bie1efe1dt
(1996) and Bie1efeldt and Stensel (1998).
In two studies the build up of metabolites during BTEX treatment was noted.
Bielefeldt (1996) and Bielefeldt and Stense1 (1998) found that a brown metabolite
formed from long term BTEX gas treatment by a mixed consortium of bacteria in a
sparged suspended growth reactor. The bacteria were enriched from soil from a
manufactured gas plant and maintained in CSTR degrading BTEX for over 1 year
prior to use in the gas treatment reactor. The soluble metabolites were not
specifically identified but were found to accumulate at high loading (> 20 mg
BTEXll.h) and interfere with BTEX biodegradation. Neal and Loehr (2000) observed
a yellow metabolite, also unidentified, during toluene degradation. The metabolite

was only observed at loadings above 17 mg/l.h, although in some cases loadings
above this level did not result in metabolite accumulation. The Microtox toxicity of
the liquid increased in the presence of the yellow metabolite. This indicates that
reactor operation must be controlled to prevent the build up of inhibitory substances.
As an alternative to maintaining low loading to the reactors, the soIids retention time
(SRT) of the system could be uncoupled from the hydraulic residence time (HRT).
This could dilute the soluble metabolite away while maintaining a high biomass
concentration in the system. A simple cIarifier (as is commonly used in activated
sludge waste water treatment systems) could be used to accomplish this goal.
BTEX compound treatment in alternative types of gas treatment reactors have
been widely studied. For example, Neal and Loehr (2000) studied toluene treatment
Table 9.6. Treatment of aromatic organic compounds in suspended growth bioreactors
Compound Reactor Q't!A MLR SLR EBRT K,a !nlet Removal Temp.
depth, (eml (mg/l.h) (glg (min) VOC conc., % (0C)
(m) min) VS.d) (tR,S) (0 2) (mg/l)
(min")
Toluene(l) 0.415 2.71 5 0.88 15.4 NR 1.28 99.4 22-25
1.79 14.2 0.98 23.3 5.52 99.9 20-25
1.48 \7.3 1.37 28.2 8.56 98.\ 20-25
2.\2 30.2 1.02 \9.6 9.83 96.9 22-25
Toluene(2) 0.40 3.8 11.6 0.47 10.5 (0.21 ) 2.1 99.4 22-25
6.4 27 0.43 6.7 (0.34) 3.0 98.6
(1.2)
Toluene(3) 0.16- 0.5-1 59.8 3.88 16 1.67 15.95 99 28
0.08 (2.61)
Benz·(4) 0.40 4.0 58.9 4.02 10 0.\3 9.7 96" 20-2\
(1.2) (0.27)
Ethyl- 0.4 8.0 37.2 2.38 5 0.16 3.1 91" 21
benz.*(4) ( 1.2) (0.39)
o-Xylene(4) 0.4 8.0 34.9 1.05 5 0.12 2.9 89" 25-27
(1.2) (0.39)
T_OX(2) 0.4 5.8 21.2 0.37 6.9 (0.31) 1.5/1 98.3 23-25
(1.2)
BT(2) 0.4 6.4 15.6 0.32 6.25 (0.34) 0.8/1 98.8 21-23
( 1.2)
BTE(21 0.4 5.2 \6.0 0.32 7.7 (0.28) 0.7/0.8. 98.2 22-24
( 1.2) /0.7
BTEX(21 0.4 4.6 \4.9 0.27 8.7 (0.25) 0.6 ea. 98.2 22-24
( 1.2}
Phenol(51 0.6-0.7 4-32 1.8- 1.3 " 3-22 r 0.65- > 99.7 20-25
r 16.2 (0.83) 0.85
p-Cresol(6) 1.5 r 21 r 1.03- 1.2 * 7r 1.2 0.43 > 99.5 30
2.4 (5) (1.66)
References: (I~eal and Loehr, 2000; (2)Bielefeldt and Stensel, 1998; (3)Landa el al., 1994;
Bielefeldt, 1996; (5)Ritchie and HiII, 1995; (61 Wei el al., 1999; r = riser of airIift; /\ dry
(4l
weight; NR = data not reported; itai ies = values not directly reported that were estimated;
'short term tests (therefore, specific loading rate likely higher than would occur under
steady-state conditions).
242 A.R. Bielefeldt
in a biofilter column packed with compost aud perlite. At mass loadings of 3.9 to
28.1 mgll.h, 96.8 to 99.7% ofthe toluene was removed. The inlet air was humidified
to maintain moisture content in the reactor, but drying aud cracking of the bottom
layer of the reactor was visible on Day 32 in the high loading experiment, which
resulted in a 20% drop in toluene removal efficiency. The treatment performauce
achieved in the biofilter was comparable to the suspended growth system except that
the biofilter was loaded with much higher gas flow rates (1100-2340 ml/min) aud
much lower inlet concentrations (0.09-0.8 mgll) to give the same mass loading as the
suspended growth system.
5.4 CHLORINATED ORGANICS
Mauy chlorinated organic compounds are c1assified as Hazardous Air Pollutants

(HAPs) by the U.S. EPA and therefore require treatment. However, only a limited
number of gaseous chlorinated organics have been treated in biologic al treatment
systems. Since most of the compounds occur at low concentrations in air, the
presence of oxygen dictates that the compounds must be degradable by aerobic
bacteria. This eliminates biological treatment for fully chlorinated compounds (such
as carbon tetrachloride and tetrachloroethene) that are not aerobically biodegradable.
In addition, the ability to aerobically degrade chlorinated organic compounds is not
as widespread as non-chlorinated aromatics treatment. Therefore, more care is
needed when selecting a source to seed the bacteria in the reactor. Biological
treatment of chlorinated organics in suspended growth systems has only been
reported in lab-scale reactors. A summary of the relevant operating conditions which
achieved high treatment efficiency (> 80% removal) are summarised in Table 9.7.
5.4.1. Dichloromethane (DCM)

Dichloromethaue (DCM, also known as methylene chloride [MeCI]) can be used as a
carbon source aud growth substrate by some types of aerobic bacteria. A key
consideration for biologic al treatment of dichloromethane is maintaining a neutral
pH, since HCI is generated as a degradation by-product. Most biological reactions
operate best at near neutral pH; to maintain operational stability the water should
contain alkalinity or pH buffering capacity. As a result of complete liquid mixing in
suspended growth systems, pH control can be easily maintained by adding a buffer
into the inlet water to the reactor.
Biotreatment of DCM in a bubble column was studied by Vanderberg-Twary
et al. (1997) in mixed gas streams containing methanol and 2-propauol. The mixtures
were formulated to simulate paint stripping wastes. The reactor contained 1.2 lliquid
aud was mixed with a magnetic stir bar. Contaminated gases were introduced into the
reactor via a non-porous diffuser. The reported mass transfer coefficients (Kla) for
DCM, methanol, and 2-propanol were 0.44, 3.7 x 10-5, and 4.1 x 10-3 min- 1 (however,
looking at data shown, the K1a for methauol appears closer to 0.23 min· 1; also,
relative Kla values between the organics should not be as variable as reported based
on their relative diffusivity in water). The liquid was buffered to maintain a neutral
pH aud inoculated with Hyphomicrobium sp. strain DM-2 (ATCC 43129) and
Rhodococcus rhodochrous strain OFS (A TCC 29672). The authors reported that
optimal treatment was achieved at a gas flow rate of 2.36 lImin which provided
sufficient oxygen without loss of biomass. No inlet or effluent concentrations
of contaminants were reported, but slug masses of contaminants (associated with
Table 9.7. Reactor operation and performance for treating chlorinated organics
Compounds Reactor Liquid Qg/A MLR SLR EBRT Kla Inlet Removal Temp. References
type depth, (cm/min) (mg/l.h) (g/gVS.d) (min) (min· I ) concentration (%) (0C)
(m) (t R , s) VOC (mg/l)
(0 2 )
TCE F 0.16- 0.5-1 0.92 0.06 16 1.7 0.24 90 28
+ toluene 0.08 55.2 3.53 (2.6) 14.7 99
TCE A 0.30 •• 2.0 6.86 1.5 35 NR 4.0 94 20-25 2 ;,.,
+ phenolliquid 0.51 0.11 35 0.3 97 ~
~.
TCE • B 0.4 4.7 0.5-8.2 0.01--D.16 8.5 0.09 0.07-1.2 77-82 22-24 3 I:l
(1.2) (0.2) ~
I:l..
TCE B 1.5 25 0.7-2.7 0.01--D.05 6 0.13 0.07--D.27 75-80 25 4 ~
+ phenolliquid :::
TCE B 1.5 144 8 0.15 1.04 0.55 0.14 50 25 4
~
+ phenolliquid '"
TCE A 1.5 48-168 15-19 0.25--D.57 1·3 0.18-0.64 0.31-1 50 25 4
+ phenol liquid
DCM A-S 0.83 305 441 NR 0.27r NR 2 85 30 5
A = airlift; A-S = airlift with sand; B = bubble column; F = fermenter;
References: lLanda et al., 1994; 2Ensley and Kurisko, 1994; 3Bielefeldt 1996; 4 Hecht el al., 1995; 5Z uber el al., 1997.
r = riser basis; NR = Not reported; italics = value not directly reported, so estimated;
*short-term test with endogenous toluene-degrading bacteria;
*. reported reactor dimensiol1s nut consistent - 3.5 L Ţ 30 cm depth and 8 cm diameter.
N
....
Uv
244 A.R. Bielefeldl
contaminated rags) were treated via gas recirculation until complete destruction was
achieved.
Zuber el al. (1997) used the results of three years of bench-scale treatment of
DCM in a 3-phase airlift bioreactor (sand:water:air) to estimate scale-up to a full-
scale treatment system. Owing to the presence of the sand the system contained both
suspended biomass and attached biomass. These results are discussed in more detail
in Section 7.1.
5.4.2. Trichloroelhene (TCE)

Bench-scale studies have been conducted for trichloroethene (TCE) biotreatment in
suspended re actor systems. Under aerobic conditions, TCE can only be biodegraded
cometabolically by specific bacteria. Aerobic cometabolism of TCE requircs that a
co-substrate compound induce the necessary oxygenase enzymes required for
biotransformation of the TCE. Because the bacterial cells do not grow or obtain
energy from the TCE degradation, the necessary co-substrate must be added. In
addition, transformation by-products of TCE have bcen shown to be inhibitory to
some bacteria via so called intermediate toxicity. Further, since the same enzymes in
the bacteria are used to degrade both growth substrates (such as phenol or toluene)
and TCE, competitive inhibition will occur. lf the contaminated gas contains both
TCE and rcquired co-substrates (such as methane, propane, and toluene), no
additional substrate feeding into the reactor will be needed. However, in most cases a
co-substrate will be specifically selected and fed into the bioreactor to grow enough
active bacteria to maintain low concentrations of TCE in the reactor liquid (thereby
maintaining optimal mass transfer ofTCE out ofthe contaminated gas). Owing to the
complexity of co-substrate feeding and the potential for inactivation of the bacteria
due to TCE intermediate toxicity, a suspended growth system may have significant
advantages over attached growth systems since better control of the growth
conditions can be maintained.
Three studies have been conducted with pure cultures of Burkholderia
(formerly Pseudomonas) cepacia G4 (Landa el al., 1994; Ensely and Kurisko, 1994;
Hecht el al., 1995). A toluene mono-oxygcnase enzyme, which can be induced by
toluene or phenol, degrades the TCE. Results are summarised in Table 9.7. Landa el
al. (1994) fed toluene (14.7 mgll) along with TCE (0.24 mg/l) in the gas phase to a
0.8 I fermenter, achieving 99% toluene removal and 90% TCE removal. They tested
higher inlet concentrations of both TCE and toluene, but the other conditions
achieved poorer removal of both toluene (97 to 71 %) and TCE (70-7%).
In studies by Ensley and Kurisko (1994) conducted in a 3.5 I volume
concentric loop airlift reactor, 0.1 to I g phenoIll.d fed directly into the reactor liquid
supported biodegradation of TCE (90-95% removal of 0.3 to 4 mg/l) by B. cepacia
G4. The TCE removal efficiency decreased over time and was attributed to decreased
bacterial activity. The time interval over which high TCE removal efficiency could
be maintained varied depending on the phenol and TCE loading, but ranged from just
over 24 hours to 30 days. Ensley and Kurisko (1994) also stated that tests were
conducted with Pseudomonas mendocina KR-I with toluene fed as a co-substrate,
but quantitative results from these studies were not reported. The mass transfer
characteristics in the airlift reactor were not reported for these studies.
Hecht el al. (1995) also fed phenol into the liquid to cometabolically treat
TCE by B. cepacia G4. Their 30 l working volume column treated TCE at varying
concentrations and gas flow rates. The presence or absence of a baffle was used to
compare a bubble column to a concentric loop airlift reactoL Under high TCE
loading conditions where 50% removal was achieved, the airlift system could operate
at a higher MLR and SLR than the bubble column (see Table 9.7). In general, the
treatment efficiency declined with increasing loading (from higher TCE inlet
concentrations and/or higher gas flow rates). The biomass concentrations in the
reactor were high, ranging from 900 to 1650 mg dry weightll.
Bielefeldt (1996) attempted co-degradation of TCE and toluene from
contaminated gas in a 2 1 bubble column containing a filamentous bacterial culture,
but poor treatment efficiency of TCE and toluene were achieved, probably owing to
competitive inhibition effects. Subsequent tests used endogenous cells that were pre-
induced by growth on toluene, achieving 77-82% TCE removal. For long term
operation, this system would require two reactors to continuously treat TCE
contaminated air. While one reactor was treating TCE, the other reactor would be fed
the appropriate co-substrate (in this case, toluene). Then the reactors would switch.
This would have the advantage that co-substrate would not be present during TCE
degradation to compete. In addition, during co-substrate feeding the desired amount
of biomass in the system could easily be controlled without worrying that the
competition would decrease the TCE removal efficiency. Previous studies with the
filamentous culture showed that endogenous TCE degradation could be maintained
over 24 hours after growth on phenol (Bielefeldt et al., 1995). Therefore, the
proposed treatment strategy shows promise for suspended biological reactors treating
TCE contaminated gas on alternating cycles with substrate feeding.
5.4.3. Comparison ta other biological gas treatment methods

In a study with a celite packed biofilter seeded with Pseudomonas putida FI, Cox el
al. (1998) reported that 30 to 60% TCE was removed when toluene was present as a
co-contaminant to induce cometabolism of TCE. The biofilters were operated with
an EBRT of 2.6 minutes, inlet TCE concentrations of 0.065 to 0.18 mgll, and inlet
toluene concentrations of O (intermittent endogenous tests) to 0.1 O mgl!. Toluene
removal was 100% and TCE removal was 20-30% when fed 0.10 mg/l toluene and
0.065 mgll TCE. Maximum TCE removal of 60% was achieved in endogenous tests
over periods ofup to 4 h but dropped to negligible TCE removal within 24 hours.
In a lab-scale soil biofiiter, Kampbell el al. (1987) reported > 90% removal of
TCE during co-degradation of propane, isobutane. and n-butane, operating at a 15
minute EBRT. Shields et al. (1993) achieved an average of 90% TCE removal over 4
days when TCE was mass loaded at 0.03 mgll.min (0.13 mgll inlet) to a 26 1 oyster-
shell biofilter seeded with Pseudomonas cepacia G4 PRl (a mutant bacteria that
constitutively expresses the mono-oxygenase necessary for cometabolism of TCE).
In a similar experiment in a 3 I reactor, greater than 96% TCE removal was achieved
when the reactor was loaded at 0.06 to 0.09 mg/l.h (inlet 0.8 to 1 mgll) for 3 days.
TCE removal in suspended growth systems appears to be comparable to these
biofiiter systems.
TCE contaminated gas treatment was studied by Speitel and McLay (1993) in
a biotrickling filter packed with celite that was operated with continuous liquid and
gas flow. Natural gas was fed into the reactor as a source of methane to support
cometabolism of TCE by Methylosinus trichosporium OB3b. The average methane
concentration was reduced from 23 mgll in the inlet air to 18 mg/l in the effluent air.
At a TCE mass loading rate of 0.57 mgll.h (0.33 mg/l inlet), TCE removal averaged
28% over 20 days. In a second experiment with a TCE mass loading of 0.66 mg/l.h
246 A.R. Bielefeldt
(0.40 mg/l inlet), TCE removal averaged 62% over 25 days; however, TCE removal
was decreasing over time probably owing to toxicity. In a final experiment at a TCE
loading rate of 1.34 mgll.h (0.72 mgll inlet), the removal rate averaged 21% and
remained fairly constant over 25 days of operation. These results are less optimal
than gas treatment of TCE using toluene or phenol as co-substrates.
Dichloromethane (DCM) is somewhat easier to treat since it can be degraded
as a bacterial growth substrate. Ergas et al. (1994) achieved greater than 98%
removal ofDCM in a 211 compost biofilter loaded at 0.57 to 9.5 mg/l.h (inlet 0.01 to
0.19 mg/l). Diks and Ottengraf(l99l) achieved greater than 90% removal ofDCM at
mass loadings less than 50 mg/l.h (inlet 1 mg/l) in a 377 1 lab-scale trickling filter
packed with ceramic saddles and seeded with Hyphomicrobium GJ21. The results
from the attached growth systems cannot be directly compared to the results reported
in SGRs due to the significantly different test conditions.
5.5 OTHER ORGANICS
Wei et al. (1999) investigated the treatment of ethanol contaminated gas in a bench-
scale extemal-loop airlift reactor. The 12 1 reactor was inoculated with Acetobacter
aceti (ATCC 15973) and sparged via a spinning flat disc (giving an oxygen K,a of
approximately 1.67 min"). Ethanol is not fully mineralised by A. aceti with acetic
acid a major by-product. Therefore, growth on ethanol was slow (Y = 0.064). In a
short-term test (80 h), an inlet concentration of 70 mg/l ethanol (mass loading 220
mg/l.h) was treated to > 99% removal from 0.62 l/min gas flow by 800-1700 mg/l
biomass. Ethanol concentrations in the liquid ranged from 150 to 50 mg/l but no
breakthrough of ethanol occurred owing to its high affinity for the liquid phase. It
was predicted that at steady-state conditions with a liquid retention time of 4.2 days
that the steady state biomass in the reactor would be approximately 1500 mg/l.
Vanderberg-Twary et al. (1997) treated contaminated gases from a waste
barrel using a gas loop that was passed through a bubble column bioreactor. The gas
contained methanol, 2-propanol, aud dichloromethane, representing paint stripping
wastes; concentrations were not reported. Biomass concentrations of 75 to 270 mg/l
were able to de grade contaminants from the gas but insufficient data were reported to
determine the mass loading or specific loading to the reactor during the experiments.
6. Costs to construct and operate suspended growth gas treatment systems
The costs for an activated sludge gas treatment system will include both capital and
operating (power, aeration maintenance) costs. Table 9.8 lists the primary costs to
build and operate a suspended growth gas treatment bioreactor. The main costs to
construct a new suspended growth gas treatment basin will be: the basin, the air
diffuser system including piping and diffuser heads, airflow and pressure monitoring
equipment, an air blower, blower controls, liquid feed (and waste) pump, and
biomass separation basin aud retum pump (if needed). Optimal design for gas
treatment resulting in more shallow tanks thau are common for municipal waste
water activated sludge tanks will save money on the bas in and blower (owing to less
head needed). If an existing activated sludge system is present for waste water
treatment little additional capital costs may be required.
In addition to the initial expenditures, operating the system will require
energy, periodic labour for maintenance, and materials. The primary operating cost
Activated Siudge 247
Table 9.8. Costs associated with suspended growth gas treatment bioreactors
Cost category Specific item Notes

Capital Tank Liquid depth sufficient for mass transfer
Aerators Fine bubble generally optimal
Liquid pump For continuous or intermittent addition of
nutrients, buffer, etc.
B10wer Sized for air flow rate and head
lnlet filter To remove particles !Tom the gas
Moisture trap To extend life ofblower
Piping (liquid and air) Compatible material
(optional) (Bio settling tank) If SRT> HRT is desired
(Bio recirculation pump) lf SRT> HRT is desired
(Oxygen sensor) If oxygen critical, may use to trigger
additional oxygen addition
(pH probe and adjustment) May be for chlorinated organics treatment
Operating Blower energy Minimised in shallow reactor
Liquid pump energy Minimised with longer HRT
Nutrient and pH chemicals N, P, trace nutrients, buffers
(optional) (Bio recirculation pump energy) If SRT> HRT is desired
Maintenance diffuser cleaning Routine periodic maintenance to ensure
pumps and blower upkeep optimal reliability and performance
for the activated sludge system is generally the power for the blower. This power
requirement is a function of the pressure at which the gas must be injected into the
reactor. The pressure needed is significantly higher than for the other types of gas
treatment systems because it must overcome the head of water (water depth in the
tank) and headloss in the aerator. For example, the headloss across biofilters has been
reported to range from 0.003 to 0.17 m of water head per m of bed depth (Govind et
al., 1993; Kiared et al., 1997; Torres et al., 1997). Equation 9.16 can be used to
estimate the power requirements for aeration (Metcalf & Eddy, 1991). Given nearly
equal head loss across the aerators, regardless of the tank depth, a general
relationship between tank depth (giving the water head of the air to overcome) and
the power required can be generated. Assuming an air flow rate of 100 m3/min (at
standard temperature and pressure) at 0.5 m and 5 m liquid depths, the blower power
required is approximately 10.4 and 91.7 kW, respectively. This indicates an energy
savings of about 88% in the shallow reactor. The nutrients and buffer chemi cais
added to support biogrowth (such as nitrogen, phosphorus) may also be a significant
operating cost. Maintenance required for the suspended growth system will include
periodic cleaning andlor replacing diffusers, and routine maintenance for the
blowers, liquid pumps, and instrumentation.
Zuber el al. (1997) estimated that a basic investment cost of $174300 would
be needed to construct an airlift re actor with sand media to treat 100 m3/h of air
containing 2 mg/l dichloromethane to 99.5% removal. Additional operating costs for
energy, chemicals, water, and personnel time averaged to $79720/yr. Bowker (1996)
estimated the cost to treat 3738 m 3/h of odour gases, including 0.18 mg/l H2 S, in a
dedicated SGR at $230000 capital and $95000/yr operation and maintenance.
However, the analysis assumed that a 4.27 m deep tank would be needed (a
traditional tank depth for a waste water treatment activated sludge basin), which is
significantly higher than the depth that would actually be needed. The model
248 A.R. Biele{eldt
indicates that a significantly lower depth can be used (about 1 m), saving both capital
costs (smaller basin, smaller blower) and operating costs (energy for blower).
For specific cost estimates for blowers, fine bubble diffusers, and
maintenance costs for activated sludge aeration, the US EP A Design Manual "Fine
Pore Aeration Systems" (1989) offers a good, though old, summary of data. Using
this information the cost of treating gases in an optimally designed SGR has been
estimated. Costs are adjusted to 1996 values using ENR cost indexes in order to
compare to the Zuber and Bowker values. The test case was treating 2 mg/l toluene
in 100 m3/h contaminated gas to 99.5% removal. Using the modelling method
described in Section 4 it was determined that a tank with 0.75 m liquid depth and 9
m2 area could be used (operating at a Qg/A of 18.6 cm/min). This system operated at
a 31 d SRT to give a toluene concentration in the liquid of 0.01 mg/l, biomass
concentration of 12 mgll, and adequate oxygen was available ( > 2 mg/l DO if the
inlet air contains 20% oxygen). The long SRT and small volume ofthe tank require
only a very small liquid pump, perhaps operated once every three days to feed
nutrients to the reactor and waste liquid contents. Capital costs for purchase,
construction, and installation of the tank, diffusers, blower, liquid pump, nutrient
feed tank, electrical, and plumbing would total $7000 to $9000. Energy costs for the
blower, liquid pump, chemical costs, and labour cost for maintenance would total
about $14000/yr. To treat 4000 m 3/h ofthe same gas and treatment efficiency, a 360
m2 x 0.75 m deep tank would be needed, at a capital cost of about $190000 to
$250000 and yearly cost of $90000 (realising economy of scale). These costs
compare very favourably with the previous estimates of Zuber et al. (1997) for the
airlift reactor and are similar to those from Bowker (1996) for odour treatment.
7. Variations
Some innovative changes to standard suspended growth biotreatment have been

proposed including (1) addition of a solid carrier (activated carbon, sand) and (2)
addition of co-solvent. These modifications are discussed briefly in the following
sections.
7.1 SOLID CARRIER ADDITION TO LIQUID
Solid carriers such as activated carbon, plastic media, or sand can be added into the
sparged system. Depending on the amount of solid carrier added, the re actor may end
up more similar to a fluidised bed reactor than a suspended growth system. A
significant amount of the total biomass in the reactor may be attached to the solid
carriers in a biofilm rather than being suspended in the liquid. This should result in a
longer mean retention time of bacteria in the system since they will not be washed
out with the liquid. The addition of activated carbon into the SGR may enhance
reactor operational stability and performance due to its sorption capacity. Ye et al.
(1994) modelled the addition of activated carbon into the riser of a concentric-loop
airlift reactor. In the 5.5 m deep liquid, the model predicted > 90% removal of
benzene or toluene, and greater than 97% DCM removal. No information on the
mass transfer characteristics of the airlift reactor and activated carbon concentration
were provided, and the modei was not experimentally validated. Zuber et al. (1997)
studied treatment of DCM contaminated gas in a concentric loop airlift reactor
containing 195 g sand/1 reactor volume. In a bench-scale reactor of 0.83 m liquid

height, DCM concentration was reduced from 2 mgll to 0.3 mg/!. Modelling was
used to estimate the design and performance in a full-scale re actor.
7.2 CO-SOLVENT ADDITION TO LIQUID
Cesario et al. (1997) proposed the use of a co-solvent to improve mass transfer into
water. The water-immiscible solvent is dispersed in the liquid and the contaminant of
interest (and/or oxygen) has a higher affinity for the solvent than water. The solvent
FC40 was chosen to study toluene removal (conc. gas/conc. solvent = 0.012
compared to H of 0.21 for toluene). FC40 is a perfluorocarbon that is non-
biodegradable, non-toxic, and has a low vapour pressure (so it will not be lost to
volatilisation). The FC40 was maintained dispersed by mechanical mixing at 800
rpm. The K1a oftoluene (0.017 S-I) was not influenced by volumetric concentrations
of 0-10% FC40. Similarly, the oxygen K1a was also insensitive to FC40 conc (0.017
S-I). In the presence of 10% of FC40, the toluene mass transfer was enhanced by a
factor of 1.1 (mass transfer with versus without solvent) and oxygen by a factor of 2.
This indicated a negligible benefit for contaminant treatment unless oxygen supply
controlled the reactor design. However, different co-solvents might prove more
successfu!.
8. Conclusions
Biological treatment of contaminated gases can be achieved using suspended bacteria

in an aerated tank or column. The technology has been demonstrated at numerous
full-scale plants treating odour contaminated gases in activated sludge basins with
municipal waste water. Numerous inorganic and organic contaminants have also
been successfully treated in lab-scale reactors with column, fermenter, or airlift
configurations. Basic modelling of the mass transfer of contaminants from the gas
phase into the bulk liquid has demonstrated that shallow liquid depths (less than 3 m)
can generally be used. Shallow depths are advantageous due to cost savings in the
tank, blower, and aeration power; designs should optimise mass transfer through the
selection of efficient aeration diffusers and good liquid mixing. Bacteria should be
selected to maintain low concentrations of the contaminants in the liquid in order to
maximise the driving force for mass transfer. Owing to the nature of the suspension it
is easy to model contaminant biodegradation. The reactor should be designed and
operated to ensure that adequate nutrients and oxygen (for aerobic biodegradation
reactions when desired) are present. The prevention of toxic metabolite build up in
the reactor liquid is a concern; strategies to avoid this problem include controlling
loading rates or maintaining a lower HRT than SRT. Operational ease and reliability
are primary advantages of suspended growth treatment reactors compared to
alternative bioreactor designs.
Important Symbols and Acronyms
A floor area ofthe reactor, m2

B endogenous biomass decay rate, day'!
BTEX benzene, toluene, ethylbenzene, xylenes
250 A.R. Bielefeldt
Cg concentration in the gas (generally ofthe contaminant), mg/l

CI concentration in the liquid (generally ofthe contaminant), mgll
D liquid depth in the reactor, m
Da diffusivity in air
DCM dichloromethane
Dw diffusivity in water
E gas hold -up; unitless volume fraction
EBRT empty bed retention time, s.
H unitless Henry's coefficient; equilibrium cone. in air/equilibrium conc. in water;
m3 liq/m 3 air
HRT hydraulic residenee time
k maximum specific substrate degradation rate, glg cells.d
Kla overall mass transfer coefficient, time'l
Ks half saturation concentration biokinetic parameter, mgll
MLR mass loading rate, mgll.h
Qg gas flow rate
SGR suspended growth reactor
SLR specific loading rate, mg eontaminant/mg VSS.h
SRT solids residence time, represents the average time a bacterium is in the system
TCE trichloroethene
VI liquid volume in the reactor
VOC volatile organic compound
VSS volatile suspended solids; a measure of biomass concentration
Y ceH yield on growth substrate, g cells/g substrate
a waste water effect on mass transfer = Kla waste water I Kla elean water
Kla voc I Kla02
References
Ando, S. 1980. Odor control ofwastewater treatment plants. J. WPCF. 52: 906-914.
Andrews, G.F. and Noah, K.S. 1995. Design of gas-treatment bioreaetors. Biotechnol. Prag. II: 498-
509.
ASCE. 1984. ASCE standard measurement of oxygen transfer in elean water. Am. Soc. Civ. Eng.,
New York, USA.
Basu, R., Clausen, E.C. and Gaddy, J.L. 1996. Biological conversion of hydragen sulfide into
elemental sulfur. Environ. Prog. 15: 234-238.
Bhattarai, R. 2000. Wastewater Regulatory Manager, City of Austin, USA. Personal Communication.
Bielefeldt, A.R. 1996. Ph.D. Dissertation. University of Washington. Seattle, WA, USA.
Bielefeldt, A.R., Stensel, H.D. and Strand. SE 1995. Cometabolic degradation of TCE and DCE
without intermediate toxicity. J. Environ. Eng. 121: 791-797.
Bielefeldt, A.R. and Stensel, H.D. 1998. BTEX-contaminated gas treatment in a shallow, sparged,
suspended-grawth bioreactor. Bioremed. J. 1: 241-254.
Bielefeldt, A.R. and Stensel, H.D. 1999. Treating VOC-contaminated gases in activated sludge:
meehanistic model to evaluate design and performance. Environ. Sci. Technol. 33: 3234-3240.
Bielefeldt, A.R.; Stensel, H.D. 1999b. Modeling competitive inhibition effects during biodegradation
ofBTEX mixtures. Water Res. 33: 707-714.
Bowker & Assoc. 1996. Biological odor control by diffusion into activated sludge basins. NEWEA
Journal. 30: 137-146.
Bowker, RP.G. 1998. Biological systems: Aeration basins. In: Odor and VOC control handbook. H.J.
Rafson (Ed.). McGraw-Hill, New York, USA.
Bowker, R.P.G. 1999. Activated sludge diffusion. Water Environ. Technol. II: 30-35.
Card, T.R. 1998. Fundamentals: chemistry and characteristics of odors and VOCs. In: Odor and VOC
control handbook. H.1. Rafson (Ed.). McGraw-Hill, New York, USA.
Cesario, M.T., Beeftink, H.H. and Tramper, J. 1995. Feasibility of using water-immiscible organic
solvents in biological waste-gas treatment. Bioproc. Eng. 12: 55-63.
Cesario, M.T., Beverloo, W.A., Tramper, J. and Beeftink, H.H. 1997. Enhancement of gas-liquid mass
transfer rate of apolar pollutants in the biological waste gas treatment by a dispersed organic solvent.
Enzyme Microb. Technol. 21: 578-588.
Chisti, M.Y. 1989. Airlift bioreactors. Eisevier, London, U.K.
Cox, C.D., Woo, H-J. and Robinson, K.G. 1998. Cometabolic biodegradation of trichloroethylene
(TCE) in the gas phase. Water Sci. Technol. 37: 97-104.
CRC. 1998. Handbook of chemistry and physics, 79th Edition. D.R Lide (Ed.). CRC Press, Boca
Raton, USA.
Dasu, B.N., Deshmane, V., Shanmugasundram, R., Lee, C-M. and Sublette, K.L. 1993. Microbial
reduction ofsulfur dioxide and nitric oxide. Fuel. 72: 1705-1714.
DeHoliander, G.R., Overcamp, TJ. and Grady, c.P.L. 1998. Performance of a suspended-growth
bioscrubber for the control ofmethanol. J. Air Waste Manage. Assoc. 48: 872-876.
Diks, R.M.M, Ottengraf, S.P.P. and Vrijland, S. 1994. The existence ofa biological equilibrium in a
trickling filter for waste gas purification. Biotechnol. Bioeng. 44: 1279-1287.
Dolfing, J., Wijngaard, A.J. van den and Janssen, D.B. 1993. Microbiological aspects ofthe removal
Ensely, B.D. and Kurisko, P.R 1994. A gas lift bioreactor for removal of contaminants from the vapor
phase. Appl. Environ. Microbiol. 60: 285-290.
Ergas, S.1., Kinney, K., Fuller, M.E. and Scow, K.M. 1994. Characterization ofa compost biofiltration
system degrading dichloromethane. Biotechnol. Bioeng. 44: 1048-1054.
Govind, R., Utgikar, V., Zhao, W., Shan, Y., Parvatiyar, M. and Bishop, D.F. 1993. Development of
novel biofilters for treatment of volatile organic compounds. In: IGT Symposium on Gas, Oii and
Environmental Biotechnology. Colorado Springs, USA.
Haug, R.T. 1993. Odor management I - QuantifYing and treating. In: The practical handbook of
compost engineering. Lewis Publishers, Boca Raton, USA.
Hecht, V., Brebbermann, D., Bremer, P. and Deckwer, W.-D. 1995. Cometabolic degradation of
trichloroethylene in a bubble column bioscrubber. Biotechnol. Bioeng. 47: 461-469.
Hsieh, C., Babcock, R.W. and Stenstrom, M.K. 1993. Estimating emissions of20 VOCs. II: Diffused
aeration. J. Environ. Eng. 119: 1099-1118.
252 A.R. Bielefeldt
Joyce, J. and Sorensen, H. 1999. Bioscrubber design: How to improve odor-control flexibility and
operational effectiveness. Water Environ. Technol. II: 37-42.
Joyce, J. and Sorensen, H. 1998. The design ofbioscrubbers for improved odor control flexibility and
operational effectiveness. In: Proceedings of the Water Environment Federation, 71 SI Annual
Conference and Exposition. 7: 249-255.
Kampbell, D.H., Wilson, J.T., Reed, H.W. and Stocksdale, T.T. 1987. Removal of aliphatic
hydrocarbons in a soil bioreactor. J. Air PolI. Control Assoc. 37: 1236-1242.
Kennes, c., Huistra, A., and Janssen. D.B. 1993. Continuous biodegradation of toxic chlorinated
compounds by a defined bacterial consortium. Med. Fac. Landbouww. Univ. Gent. 58: 1749-1753.
Kiared, K., Wu, G., Beerli, M., Rothenbuhler, M. and Heitz, M. 1997. Application ofbiofiltration to
the control ofVOC emissions. Environ. Technol. 18: 55--{)3.
Landa, AS., Sipkema, E.M., Weijma, J., Beenackers, AA.C.M., Dolfing, J. and Janssen, D.B. 1994.
cometabolic degradation of trichloroethylene by Pseudomonas cepacia G4 in a chemostat with
toluene as the primary substrate. Appl. Environ. Microbiol. 60: 3368-3374.
Libra, J.A. 1993. Stripping of organic compounds in an aerated stirred tank reactor. Fortschr.-Ber.
VDI Reihe 15 Nr. 102. Dusseldorf. VDI-Verlag.
Magbanua, B. S. Jr., Poole, L. J. and Grady, C. P. L. Jr. 1998. Estimation ofthe competent biomass
concentration for the degradation of synthetic organic compounds in an activated sludge culture
receiving a multicomponent feed. Water Sci. Technol. 38: 55-62.
Matter-Muller, c., W. Gujer, and W. Giger. 1981. Transfer of volatile substances from water to the
atmosphere. WaterRes. 15: 1271-1279.
Metcalf & Eddy. 1991. Wastewater Engineering, Third Edition. McGraw-Hill, New York, USA.
Neal, A.B. and Loehr, R.C. 2000. Use of biofilters and suspended-growth reactors to treat VOCs.
Waste Mgmt. 20: 59--{)8.
Ostojic, N., Les, A.P., and Forbes, R. 1992. Activated sludge treatment for odor control. Biocycle. 33:
74-78.
Pomeroy, R.D. 1982. Biological treatment ofodorous air. J. WPCF. 54: 1541-1545.
Pomeroy, R.D. 1963. Controlling sewage plant odors. Consult. Eng. 20: 101-104.
Ritchie, BJ. and Hill, G.A 1995. Biodegradation of phenol-polluted air using an external loop airlift
bioreactor. J. Chem. Technol. Biotechnol. 62: 339-344.
Roberts, P.V. and Dandliker, P.G. 1983. Mass transfer of volatile organic contaminants from aqueous
solution to the atmosphere during surface aeration. Environ. Sci. Technol. 17: 484-489.
Roberts, P.V., Munz, C. and Dandliker, P. 1984. Modeling volatile organic solute removal by surface
and bubble aeration. J. WPCF. 56: 157-164.
Roberts, P.V., Hopkins, G.D., Munz, C. and Rojas, A.H. 1985. Evaluating two-resistance models for
air stripping of volatile organic contaminants in a countercurrent, packed column. Environ. Sci.
Technol. 19: 164-172.
Romain, M. 1996. M.S. Thesis. University of Washington. Seattle, WA, USA

Ryckman-Siegwarth, J. and Pincince, A.B. 1992. Use of aeration tanks to control emissians from
wastewater treatment plants. In: Praccedings afthe 65 1h Annual Conference afthe Water Enviranment
Federation, 10: 11~22.
Shields, M.S., Reagin, M.J., Gerger, R.R., Somerville, c., Schaubhut, R., Campbell, R. and Hu-
Primmer, J. 1993. Constitutive degradation ofTCE by an altered bacterium in a gas phase reactof. In:
Bioremediation of chlorinated and polycyclic aromatic hydrocarbon compounds. R.E. Hinchee, A.
Leeson, L. Semprini and S.K. Ong (Eds.). Lewis Publishers, Boca Raton, USA.
Sklandany, G.J., Oeshusses, M.A., Oevinny, J.S., Togna, A.P., and Webster, T.S. 1998. Biological
systems: Biofilters. In: Odor and VOC control handbook. H.J. Rafson (Ed.). McGraw-HiIl, New
York, USA.
Speitel, G.E. and McLay, O.S. 1993. Biofilm reactors for treatment of gas streams containing
chlorinated solvents. J. Environ. Eng. 119: 658~678.
Stensel, H.O. and Bielefeldt, A.R. 1999. A device and method for removal of gas contaminants
through a shaIlow sparged bioreactor. U.S. Patent No. 5985649, lssued Nov. 16, 1999.
Stillwell, S.A., Hans, O.E., and Katen, P.c. 1994. Biological scrubbing of foul air in activated sludge
treatment reduces odors and ROGs from headworks and primary claritiers. In: Proceedings - Odor and
VOC emission control for municipal and industrial wastewater treatment facilitics. Alexandria, VA.
Water Environment Federation.
Sublette, K.L. and Sylvester, N.O. 1987. Oxidation ofhydrogen sulfide by Thiobacillus denilrificans:
desulfurization of natural gas. Biotechnol. Bioeng. 19: 249~257.
Tewari, P.K. and Bewtra, J.K. 1982. Alpha and beta factors for domestic wastewater. J. WPCF. 54:
1281-1287.
Torres, E.M., Oevinny, J., Basrai, S.S, Carlson, L.J., Gossett, R., Kogan, V" Ahn, T., Kardos, O"
Shao, J., Webster, T. and Stolin, B. 1997. Biofiltration: controlling air emissions through innovative
techno10gy. Water Environment Research Foundation, USA.
U.S. EPA. 1989. Design manual: fine pore aeration systems. Oftice of Research and Oevelopment.
Cincinnati, OH. 625/1 ~89!O23.
Vaderberg-Twary, L., Steenhoudt, K., Travis, B.J., Hanners, J.L., Foreman, T.M. and Brainard, J.R.
1997. Biodegradation of paint stripper solvcnts in a modified gas lift loop bioreactor. BiotechnoI.
Bioeng. 55: 163~169.
Wei, V.Q., Hill, G.A. and Macdonald, O.G. 1999. Bioremediation of contaminated air using an
external-Ioop airlift bioreactof. Can. J. Chem. Eng. 77: 955~962.
Whitman, W.G. 1923. The two-tilm theory ofabsorption. Chem. MeI. Eng. 29: 147.
WilIiams, T.O. and Miller, F.C. 1992. Odor control using biofilters. BiocycIe. 33: 72~77.
WilIiams, T.O. and Miller, F.C. 1992. Biotilters and facility operations. Biocycle. 33: 75~79.
Wu, G., Chabot, lC., Caron, J.J. and Heitz, M. 1998. Bio10gica1 eliminatioll of volatile organic
compounds from wastc gases in a biofilter. Waler Air Soil PolI. 101: 69~78.
Ye, L., Khandan, N.N., and Edwards, F.G. 1994. Biological treatment ofairstreams contaminated with
organic vapors. Water Sci. TechnoI. 30: 71~74.
Zhu, H., Keener, T.C., Bishop, P.L., Orton, T.L., Wang, M. and Siddiqui, K.F. 1999. Aeration
recirculation in air and high purity oxygen systems for control of VOC emissions from wastewater
aeration basins. Environ. Prog. 18: 101~106.
254 A.R. Bielefeldt
Zhu, H., Keener, T.C., Orton, T.L., Bishop, P.L., Khang, SJ. and Siddiqui, K. 1997. The effectiveness
of aeration recirculation in controlling VOC emissions from publicly owned treatment works. 1. Air
Waste Manage. Assoc. 47: 1259-1267.
Zuber, L., Dunn, I.J. and Deshusses, M.A. 1997. Comparative scale-up and cost estimation of a
biological trickling filter and three-phase airlift bioreactor for the removal of methylene chloride from
polluted air. J. Air Waste Manage. Assoc. 47: 969-975.
PART 2. APPLICA TIONS
CHAPTER 10 BIOFILTRATION OF WASTE GASES FROM
A DAIRY INDUSTRY
Christian KENNES
1. Background
Biological waste gas treatment in above ground conventional biofilters using natural
filter beds is a simple and efficient technology (Chapter 3). It is suitable for the
treatment of polluted air originating in production plants and waste water treatment
plants from food and drinking industries (Passant et al.. 1992). The technology has
recently been app1ied by Casals Cardona Ind. for the treatment of waste gases
released from a waste water treatment plant and containing mainly su1phur
compounds, ammonia and some VOCs.
2. Technology selection
Preliminary studies showed that the mean waste gas flow rate reached 4000 m 3/h and
was mainly composed of a mixture of low concentration organic and inorganic
volatile pollutants (Table 10.1) released at ambient temperature. The VOC mixture
was mainly composed of lactic acid and low concentrations of several non-identified
volatile fatty acids.
Table 10.1. Composition and mean pollutant concentrations
Identified pollutants Mean inlet Description of odour Odour threshold'

concentrations (mg/m 3)
(mg/m 3)
H2S 5 Rotten eggs 0.014
CH3SH Sulphide 0.082
(CH3)2 S + (CH3)2 S2 Rotten cabbage, 0.0508 and 0.346
sulphide
NH 3 3 Pungent 39.600
VOC 5
'100% panel recognition threshold
The main reasons for the biofilter vendor (Casals Cardona Ind.) and for
members of the dairy industry to select a biological process for solving the air
pollution problem were the low investment and operation costs compared to other
technologies. high efficiency and minimal supervision and monitoring requirements.
Biodegradability of the pollutants is a prerequisite for applying biofiltration.
Several studies have shown that the compounds listed in Table 10.1 can relatively
easily and completely be mineralised under aerobic conditions (Weckhuysen et al.,
1994, Smet and Van Langenhove, 1998). The above mentioned gas flow rates and
temperature are also suitable for biofiltration. A conventional biofilter was selected
257
C. Kelllles and M. C. Veiga (eds.). Bioreactors for Waste Gas Treatment. 257-260.
258 C. Kennes
as it represents the simplest and cheapest design and because ofthe recent experience
of Casals Cardona Ind. in operating such systems. Although biodegradation of
sulphur compounds might lead to pH drop, many of the microorganisms degrading
such compounds are known to be quite resistant to acidic conditions, although VOC
degraders might be more sensitive. The presence of ammonia in the waste gas might
also partly buffer pH fluctuations in the presence of metabolites formed during
biodegradation of sulphur compounds (Chapter 3) (Kennes and Thalasso, 1998). It is
also worth observing the quite low concentration of pollutants representing in fact
conditions typical of odour problems to which biofiltration technology was applied
originally. It can easily be calculated that mass loading rates of only a few g/m 3 .h
were applied. Such low concentrations will also contribute to maintain limited pH
variations. An advantage of using a natural carrier is that it can be used as organic
fertiliser when it gets exhausted and once it needs to be replaced for maintaining
proper operation ofthe bioreactor.
3. Biofilter design
A 4 kW centrifugal fan made out of non-corrosive material was used for polluted air
ducting from the emission source to the treatment system at a flow rate of 4000 m 3/h,
trying as far as possible to avoid mixing c1ean air with the polluted air stream which
would otherwise generate higher flow rates and increased energy costs. The parts in
contact with the fluid were made of a biphenolic resin/fibre glass mixture.
3.1 PRE-TREATMENT
A pre-treatment step is often required for regulating moisture content, pH,

temperature and for removal of partic1es (Chapters 2 and 3). Although partic1es were
not detected in the present case, the waste air stream needs to be humidified to near
saturation. Air conditioning was performed by passing the polluted stream through a
3 m 3 counter-current vertical scrubber (Tecnium) made of synthetic resins/glass fibre
mixtures and operating with a LlG factor of2.361/m3 . pH ofthe scrubbing liquid was
checked regularly. The scrubber was equipped with a mist eliminator to avoid
reentrainment of liquid droplets which would otherwise result in the emission of
droplets to the atmosphere and partial 105S of the scrubbing liquid. The washing
liquid is recirculated over the column by means of a 2.2 kW centrifugal pump
(Tecnium). The waste gas reaching humidity levels of about 98% was then fed to the
bioreactor.
3.2 BIOFILTER
Treatment of preconditioned polluted air was performed in a 75 m 3 open biofilter

constructed in concrete and filled with 45 m 3 of a 1.8 m high filter bed, leading to an
empty bed residence time around 40 s (Figure I 0.1). Sizing of the full-scale biofilter
and evaluation of optimal running conditions, such as residence time were based on
preliminary small scale experiments. The filter bed consisted of a mixture of coconut
fibre/fibrous peat (1: 1) previously cut and prepared by the vendor (HOBIFILTER,
Switzerland) with dry and wet densities of 0.35 and 0.55 glcm 3, respective1y. The
biofilter was provided with four sprayers for water addition twice a day. The total
Case sludy - Open biofiller 259
volume of water added daily to the system was approximate1y 200 litres, value that
could either be increased or decreased depending on weather conditions. The bottom
and walls of the biofilter were recovered with a impermeable layer made of
thermoplastic materials.
4. Operation of the treatment system
The biofilter has been operating for approximately two years so far. It was not
inoculated and the start-up period lasted about one month. Although inoculation with
speciali sed cultures has proven to be sometimes necessary when dealing with methyl
sulphide pollution (Smet el al., 1996), this appeared not to be a prerequisite in the
present case, although such a procedure would have been interesting if a shorter
start-up stage was required. Although sampling effluents of such open biofilters is
difficult, monitoring data allowed showing that the start-up period was characterised
by overall removal efficiencies of about 40-60%. Sampling was undertaken by
placing a hood provided of a stack above the biofilter bed. After the first month
operation, removal efficiencies could be maintained above 60% most often reaching
above 99% removal, and eliminating complaints conceming odorous emissions.
After almost two years operation, pressure drop was still minimal. Previous
experience of Casals Cardona Ind. with such filter beds has shown that after two to
three years continuous operation, the addition of 30% fresh carrier material would be
required. The total lifetime of the carrier under the present running conditions is
expected to reach five years. After such a period, complete renewal of the filter bed
is necessary as a result of one or several ofthe following potential problems (Chapter
3):
Figure 10.1. General view ofthe biofilter.

260 C. Kennes
• Bed compaction;
• Excess dead biomass;
• Exhaustion of relevant nutrients of the filter bed;
• Presence of high salt concentrations;
• Channelling effects;
• Excessive accumulation of dust particles.
The old filter bed is advantageously used as fertiliser. Experience has shown that
covering the filter-bed with a special grass layer may be useful for improving and
maintaining optimal moisture conditions and to avoid unwanted growth of weed or
the presence of large roots eventually altering homogeneity and composition of the
filter-bed. Such a procedure also improves the general aspect ofthe biofilter.
References
Kennes, C. and Thalasso, F. 1998. Waste gas biotreatment technology. l. Chem. Technol. Biotechnol.
72: 303-319.
Passant, N.R., Richardson, S.J., Swannell, R.P.J., Woodfield, M., van der Lugt, l.P., Wolsink, l.H.,
Hesselink, P.G.M., Hecht, V., Brebbermann, D. and Bischoff, H. Biodegradability of the volatile
organic compound (VOC) emissions from the food, drink and metal degreasing industries. In:
Biotechniques for Air Pollution Abatement and Odour Control Policies. Dragt A.J. and van Ham J.
(eds.). Eisevier, Amsterdam, The Netherlands, 1992, pp. 315-320.
Smet, E., Chassaya, G., Van Langenhove, H. and Verstraete, W. 1996. The effect ofinoculation and
the type of carrier material used on the biofiltration of methyl sulphides. Appl. Microbiol. Biotechnol.
45: 293-298.
Smet, E. and Van Langenhove, H. 1998. Abatement of volatile organic sul fur compounds in odorous
emissions from the bio-industry. Biodegradation. 9: 273-284.
Weckhuysen, B., Vriens, L. and Verachtert, H. 1994. Biotreatment of ammonia- and butanal-
containing waste gases. Appl. Microbiol. Biotechnol. 42: 147-152.
CHAPTER 11 TREATMENT OF HIGH VOC LEVELS IN A CLOSED
BIOFILTER
Jan-Carel NIEUWLAND and Michael S. McGRA TH
1. Introduction
Traditionally, biofilters have been used to solve odour problems throughout Europe
and more recently in North America (Leson and Winer, 1991; van Groenestijn and
Hesselink, 1993; Kennes and Thalasso. 1998). Although just as important as the
odour industry, less attention has been paid to biofilters treating volatile organic
compounds (VOC) in the chemical, pharmaceutical, printing and automotive
industries. Poorly designed biofilters have caused a great deal of negative publicity.
Consequently, biofilters in the USA and to a lesser extent in Europe are typically not
considered a reliable and cost efficient solution to VOC emission problems. In
Europe however, partly because of regional strict emission regulations, biofilters
have established themselves as cost effective and reliable VOC abatement solutions
in some industries (Chapter 1). A number of companies have built several full scale
biofilter installations in various different fields including the food, food flavouring
and fragrance, municipal, pharmaceutical and chemi cal industry.
A company in the Netherlands which produces industrial separation
membranes was contracted by the Dutch government to solve their methanol and
trichloroethene (TCE) emis sion problem. Their day time emis sion Ievels were 1800
mg/m 3 ofmethanol and 200 mg/m 3 ofTCE, while a maximum of 100 mg/m 3 ofTCE
was allowed as well as a total VOC maximum of 150 mg/m 3 . Therefore the methanol
concentration had to be decreased to a maximum of 50 mg/m 3 . The TCE levels were
to be decreascd to 100 mg/m3 by alteration of the production process and a
previously installed activated carbon system.
Typically it is assumed that biofiltration is not technically or economically
feasible when treating high levels ofVOC is required. The most common choices for
treating these streams are thermal and catalytic oxidation and adsorption. However,
if an accurate description of the off-gas is obtained a biofilter can be properly sized.
In addition, experience must play a large role in overcoming the practical hurdles in
designing for highly concentrated off-gases. If a biofilter has the proper residence
time, proper moisture addition system and nearly continuous loading, biofiltration
can achieve high removal efficiencies (gre ater than 90%) for methanol in a very cost
effective manner.
2. Technology selection
Many treatment technologies were considered viable options. The options

investigated by the gas separation company were regenerative adsorption of TCE and
methanol, non-regenerative adsorption of methanol, regenerative adsorption of TCE,
absorption of methanol, biofiltration of methanol and bioscrubbing of methanol.
These options were evaluated on an economic basis and were ali assumed to be
capable of solving the VOC abatement problem. Figure 1l.1 shows the results of the
261
C. Kelliles and M. C. Veiga (ee/s.J. BioreactorsJor Waste Gas Treatment, 261-284.
262 J-C Nieuwland and MS McGrath
economic analysis performed by the membrane separation company. Catalytic

oxidation was not considered a viable opt ion due to the potential poisoning effects of
TCE on most catalyst and potential hazard of the formation of toxic reaction by-
products.
Owing to the relatively low life cycle cost of biofiltration, the company
decided that they would investigate suppliers of biofilters. After some time the
company purchased a BIOTON biofiltration system from MEC.
800000
700000
600000
~
'"
lfJ 500000
2- 400000
t1o 300000
u
''"O
1-<
200000
100000
o
o 3 4
Years of operation
(1) ~ Non regenerative adsorption ofmethanol

(2) ~ Regenerative adsorption of TCE and methanol
(3) ~ Biowashing ofmethanol
(4) ~ Absorption ofmethanol
(5) ~ Regenerative adsorption ofTCE (no solution)
(6) ~ Biofiltration of methanol
Figure 11.1. Cost comparison of potential technologies.
3. Biofilter design
The system consisted of a packed column humidifier, biological oxidation chamber

(l4m x 3m x 3m), BIOTON medium (63 m\ control system (including load cell
based fully automated moisture addition system), process fan and exhaust stack.
Figure 11.2 shows a typical flow diagram of a BIOTON biofiltration system.
The inlet airflow is first led through a pre-humidification system, which
nearly saturates the airflow, preventing the desiccation of the filter medium under
non loading conditions. Secondly the air is led through a bed of porous, biologically
active filter material. Finally, the air is sucked out of the system by means of a fan
after which it is emitted to the air through the means of a small stack. The process air
left the manufacturing process and entered the biofilter at 22 °c and 50% relative
humidity and was virtually free of particulate matter.
Heat Solenoid To Exhaust i
-------------su-ppie"mentăr------------:
Addition
Mist Valve Watering System Stack :
ElimÎnator ţ t t
I
Water 6) • I
,, ~
Oistributor I~ ,
I
'"i2'
Packing
iiii· ,
I
I
~
I
I
I
I ~
I
I '"t:>..
Proces. I ~ I CJ-
Gas : I
I ~
41 I ~
c........... , Bio Media I ....
I
IITo Drain ,1
~~.----~
Load CeH .• -------- .-------
Process
Water
Figure 11.2. Schematic ofthe BIOTON biofiltration system.

1"
~
264 J-C Nieuw/and and MS. McGrath
The concentration of methanol was expected to be reduced from 1800 mg/m3

to 50 mg/m3 in the biofilter system, an overall removal of 97%. Overall off-gas flow
rate was determined to be 2700 m3/hr. TCE, present in the off-gas, was assumed not
to be degraded by the microorganisms and to have no effect on the performance of
the biofilter.
Ali experiments were conducted at a full-scale BIOTON biofilter.
Measurements were taken by visual inspection or physical sampling. Dry mass values
were determined by weighing samples of the filter medium before and after drying in
a stove. Pollutant concentrations were determined by on-line sampling using FTIR
(Fourier Transform Infrared Spectrometry) and heated ducting.
4. Biofilter performance
4.1 START-UP AND OPERATION
Soon after start-up of the biofilter system, it was dear that the installation did not
perform as specified. Fourier Transform Infrared Spectrometry (FTIR) analysis ofthe
process gas is shown in Figure 11.3, and indicates an average inlet loading of 2500
mg/m3 and peaks of 3000 mg/m3 . The peak durations were too short for the
microorganisms to adapt, so peaks left the system virtually untreated. Under the
above conditions, the outlet concentration fluctuated between 300 and 500 mg/m3
(approximately an overall removal efficiency of 85%) much more than allowed by
the regulatory body. Measurements also showed that inlet concentrations of TCE
were not reduced during passage of the filter medium. Even on a longer timescale
the inlet and outlet concentrations were the same, indicating that no buffering within
the biological system occurred. The assumption that TCE would have no effect on
the biology was confirmed by the fact that the biofilter performed within the design
limits, taking the desiccation the filter medium and the spiked loading into account.
A detailed investigation of the system revealed that it was not possible to
remove the higher leve1s of methanol with the amount of medium in the system.
Even if the system was fully operating according to the design parameters the
capacity of the system was not sufficient. In addition, owing to the high inlet
concentrations high levels of heat were formed during oxidation resulting in
desiccation of the filter material. Specifically, when heat is generated from bio-
oxidation, moisture is transferred from the filter medium to the air stream. Normally
the automatic moisture addition system would compensate for the moisture lost
through evaporation. However, the pressure of the water in the piping system feeding
the overhead spray nozzles was less than design pressure. This resulted in a
deficiency in the coverage zone of the nozzles, therefore the moisture of the medium
could not be adequate1y maintained. At lower inlet concentrations of polluting
compounds these problems do not result in a critical situation. In this particular case
the rapid desiccation and the limited nozzle coverage created dry areas in the filter
bed (as shown in Figure 11.4). Dry filter material has a lower pressure drop than wet
filter material, so the inlet air tended to flow trough dry filter material. As a result,
not the whole volume of the filter material is active in the degradation of the
methanol. This does also mean that more methanol, and therefore more substrate for
heat production, is preferentially going through the dry areas. A low moisture content
Case study - Closed biofilter 265
Weekly offgas profile
3500 , - - - - - - - - - - - - - - - - - - - - - . ,
~3000
~2500
~ 2000 - - Maximum inlet
o
~ 1500 ....... Maximum Qutlet
'E
l'.l1000
1:
O
o 500
W
C O~~~~~~~~~~~~====~--~
o 2 4 6 8 10
Time (days)
Figure 11.3. Initial performance of the biofilter system.
has a negative influence on mass transfer from the air to the biofilm, as well as an
inhibiting effect on the activity of the microorganisms. Escape from this vicious
circle is not possible in this normally self-regulating system and appropriate measures
are required to correct the situation and to prevent the eventual failure of the biofilter.
4.2 OPTIMIZA nON OF BIOFILTER PERFORMANCE
Three major hurdles required attention to obtain a successfully operating biofilter.

These were higher than expected inlet concentrations, spikes of inlet concentrations
of methanol, aud improper off-gas flow distribution through the biofilter medium.
Firstly, the higher than expected inlet concentrations of methanol required more
retention time and therefore more filter medium to be adequately removed. It was
decided to install a smaller second biofilter vessel in series with the first one, thereby
increasing the overall biofilter medium volume from 63 m3 to 90 m 3 and increasing
the retention time from 84 seconds to 120 seconds. The overall gas surface velocity
of 64 m/h was maintained by adding the second vessel in series and not in paralleL If
the second vessel were added in parallel, the surface velocity would have fallen
below au acceptable level resulting in an uneven distribution of the airflow. Also,
owing to the low gas velocities, adding the second vessel in series did not
significantly increase the overall system pressure drop.
The peaks of methanol were another problem. Microorgauisms require a
certain time to adapt to changing circumstances, but the peak duration did not allow
for this adaptation. However, the excellent solubility of methanol in water together
with the already installed humidifier offered a solution. The sump of the humidifier
could be used as a buffering vessel for methanol. During day time peaks of methanol
were collected in the sump, smoothing the pollution profilc of the inlet gas stream.
During night time, when there was no production and the inlet methanol
concentrations were much lower, the methanol in the water phase was released to the
off-gas stream through equilibrium partitioning and treated in the biofilter.
266 J-C Nieuwland and MS McGrath
Humiclffieatlon~
$Y$tem
. . -'~Filter meclium
Figure 11.4. Nozzle coverage at the reduced nozzle pressure.
The third problem was the low water pressure causing inadequate water
coverage from the overhead nozzles in the bio-oxidation chamber. Since the water
pressure could not be increased, additional nozzles were installed and the nozzle
layout was improved. The nozzles were rearranged to give a more even distribution
of water of the filter surface. The automatic spraying system was now capable of
humidifying the whole of the filter bed, providing the optimal conditions for the
microorganisms. The even distribution of water also allowed an even distribution of
the airflow through the filter bed, resulting in optimal biological activity throughout
the entire filter medium. Furthermore the even distribution of water created an
increased total specific surface area by improving the exchange surface between
water and air. This allowed the biofilter to take up more methanol out ofthe airflow.
S. Conclusions
After corrective measures were taken, additional off-gas sampling was performed
throughout the system. The results of this sampling are shown in Figure 11.5.
Sampling confirmed the high inlet loading and the day time peak loadings of
methanol. In addition, it was confirmed that the packed bed humidification system
allowed the peaks to be dampened. The more even distribution of water over the
filter surface, resulting in a better use of the total filter medium, led to an improved
performance of the first biofilter vessel. Although the performance was not good
enough to comply with govemmental regulations, the outlet concentrations of 100 -
Case study - Closed biofiller 267
Weekly offgas profile
3500
~ 3000
C,
§. 2500
CI)
c::
- - - Maximum inle/
:8 2000
....... Maximum oullel
~
'E<Il 1500 - . - . - 'n'el Me/hano' afler humidifier
(.l
c::
o 1000
a;"
:s 500
Ou!l~t m~!/lanol
o
o 2 4 6 8 10
Time (days)
Figure 11.5. Final performance ofbiofi/ter system.
150 mg/m 3 carne near the required value of 50 mg/m 3 , Finally, it was obvious that by
adding the additional bio-oxidation vessel complete removal of methanol could be
achieved.
Detailed models do exist to determine the volume of medium required for
applications and to predict biofilter performance. However, biofilter performance is
extremely dependent on this sizing tool. If adequate data of off-gas composition do
not exist, it is highly recommended that VOC specification and quantitication
sampling is performed prior ta design. In addition, experience must always be
combined with modelling results when sizing biotilters. Understanding the effects of
heat generated from bio-oxidation, peak loading effects, the importance of
humidification and an even water distribution and other complications can only be
accounted for through empirical data. The present case study is an example of how
combining experience with proven design too1s can provide reliab1e remova1 of high
levels of methanol, a common VOC.
References
Kennes, C. and Thalasso, F. 1998. Waste gas biotreatment technology. J. Chem. Technol. Biotechnol.
72: 303-319.
Leson, G. and Winer, A.M. 1991. Biofiltration: an innovative air pollution control technology. For
VOC emissions. J. Air Waste Manage. Assoc. 41: 1045-1054.
Van Groenestijn, J.W. and Hesselink, P.G.M. 1993. Biotechniques for air poIIution control.
CHAPTER 12 NEW BIOREACTOR SYSTEM FOR TREATING
SULPHUR- OR NITROGEN- COMPOUNDS
N.J.R. KRAAKMAN
1. Introduction
During recent years, biological techniques have been applied in an increasing

nurnber of situations to solve industrial polluted air emission problems. More and
more biologic al techniques overcome a lot of drawbacks and disadvantages
associated with classical, physical--chemical air purification techniques like
absorption, adsorption and incineration.
The well known disadvantages of the traditionally used air treatment
techniques are high energy costs (incinerators), the use of chemicals (chemical
scrubbers) and the production of waste products (incinerators, scrubbers, activated
carbon).
Conventional organic media biofilters have often been applied to treat very
different types of air streams. Although some failures have sometimes been detected,
many ofthese installations have been successful in treating polluted air (Chapter 3).
Much effort has been made over the past ten years aimed at improving
biological air purification technologies. Progress has been made in research fields as
microbiology, for example by using thermophilic microorganisms (Heslinga and
Groenestijn. 1997), in the biodegradation of chlorinated compounds (Seignez et al ..
2000), in developing applications with fungi (Cox, 1995; Kraakman el al., 1997), in
the use of predators to prevent clogging (Cox el al., 1999; van Heiningen el al.,
2000), in the biodegradation ofNOx (Woertz et al., in press). Progress has also been
made in research fields like biofilm and biofilter modelling (Picioreanu el al., 1999;
Alonso et al. 1998), reactor design and reactor operation with or without using
additional techniques (Song and Kinney, 2000; Cox and Deshusses, 1999; Melse and
Kraakman, 1998).
This chapter describes a bioreactor technology that has been developed by the
company Bioway b.v. and that has been applied to solve several industrial emis sion
problems discussed below.
2. Why develop a new bioreactor?
Polluted air streams containing biodegradable compounds, readily soluble in water,

can be effectively treated in a biofilter. In general the most cost efficient way to treat
polluted air streams when the concentrations of pollutants in the air stream are below
5 g/m3 is by implementing biological techniques (Groenestijn and Hesselink, 1993).
Unfortunately biologica! treatment is not possible for ali types of air streams yet.
Each air stream is different in terms of composition of pollutants and in terms of air
stream conditions like temperature, relative humidity and emission frequency.
Because each air stream is different, each one needs to be treated in a different way.
Conventional biofilters have been applied successfully to air streams
containing VOC and odour compounds, readily biodegradable and with relatively
270 NJR. Kraakman
high water solubilities (Chapter 3). However, conventional biofilters present some
limitations in treating polluted air streams that contain:
• Compounds which are difficult to degrade biologically (e.g., large molecules as

PAH, chlorinated compounds, etc.);
• Compounds that are very poorly soluble in water (e.g., hexane, pentane, 1,3-
butadiene);
• Compounds released during the biodegradation process and resulting in a build-
up ofintermediates like catechols, acetic acid, etc.;
• Compounds biodegraded into acidic products (e.g., sulphur and nitrogen
compounds).
A disadvantage of some conventional biofilters is that organic media need to

be replaced rather frequently. These media need to be replaced even more often if
they have not been stabilised before use. The costs of media replacement can
represent up to 40% ofthe operational cost (Bioway, data not shown).
Conventional biofiltration systems need a re1atively large amount of space
when compared to other air treatment techniques. They are restricted by the media
volume and by the media height. The media volume necessary to treat the air
effective1y is defined by the limitation of the diffusion of the components in the air
stream to the biomass layer or the limitation of the biological reaction rate in the
biomass layer. The height of the medium in a conventional biofiltration system does
usually not exceed 1 to 1.5 meter in order to prevent high energy costs owed to high
pressure drops over the media.
When biological air treatment technologies are applied in industries where air
pollution control is very strictly regulated, a good control of the biological systems is
a prerequisite. In many industries the production processes are changing over time
due to variations in the production process itself. In order to be able to operate a
biofiltration system continuously under optimal conditions a good understanding and
a good control of important process conditions are necessary. Not a black box, but a
system where important conditions for optimal microbial activity can be measured
and controlled.
3. The different types of Bioway's bioreactor system
Below we will describe a bioreactor system capable oftreating air streams containing
mixtures of compounds like hydrogen sulphide, organic sulphur compounds,
ammonia and organic nitrogen compounds.
The bioreactor can be characterised by:
• The inorganic media;

• The multi-stage layers of media (up to three layers);
• A water distribution system designed in such a way that water can be added to
each layer separately;
• The ability to recirculate the process water within the bioreactor;
• An automatic pH/EC-control;
• A controllable interface system for the operator.
Case study - New bioreactor 271
The bioreactor is built to be flexible for different situations. A modular

standardised construction is used with materials resistant to low pH values. The
bioreactor installation is not a biotrickling filter because it is more flexible than a
biotrickling filter is known for. The general definition of a biotrickling filter is the
following: a biofilter containing media with microorganisms attached to it and with
the process water continuously recirculating over the media (Groenestijn and
Hesselink, 1993) (see also Chapter 4). The new bioreactor can be operated as a
biotrickling filter, but it is more than that. One important difference is that the
bioreactor always contains two or more layers of media that can be irrigated
separately. Besides, nearly none of these bioreactors is operated with a constant
recirculation over each layer. The layers are often not continuously irrigated with
process water, but only discontinuously without any recirculation. Recirculation is
applied only when this appears to be really necessary. However, like biotrickling
filters as well as some conventional biofilters the new bioreactors contain inorganic
media and allow optimal control of the biological process via the composition of the
liquid medium and through an easy wash out ofbiodegradation products.
For different applications different bioreactors have been developed. Up to

now three different types of systems have been deve1oped:
• For municipal waste water treatment plants and lift stations the ZEROCHEM
unit has been deve1oped.
Air streams emitted from municipal waste water treatment plants and lift
stations in sewer lines contain odorous gases. The volume of the stream is often
small (mostly up to 2000 m 3/h) and its composition is relatively constant. High
removals of odours, among which hydrogen sulphide, are required in these
situations. The bioreactor should be able to be operated either with or without a fan.
When functioning without a fan the air comes out a dwell only when waste water
enters the dwell. Low pressure drop over the bioreactor is therefore required. Low
maintenance and simplicity of operation are important factors for such application.
• In food industries like breweries, vegetable oii production plants or industries

with industrial waste water treatment installations the PURSPRING bioreactor
has been developed (Figures 12.1 and 12.2).
The composition of air streams emitted from food industries is highly
variable but it is often characterised by the presence of high H2S concentrations (up
to 2500 ppmv) in combination with other compounds like odorous VOC compounds
or ammonia (see also Chapter 10). The volume ofthe air stream can be large (up to
10000 m3/h or higher) and the characteristics ofthe air stream are directly related to
the production process in the plant. Large variations of the concentration of the
compounds are sometimes observed. Stable operation of the bioreactor with high
removal efficiencies is required. Operators are often at the site to take care of the
bioreactor, therefore a good interface between operator and installation is desirable.
• For the Viscose industry the V-SPRING bioreactor has been deve1oped.
Air streams that contain H2S in combination with CS 2 are emitted by Viscose
industries. The volume ofthe streams is re1ative1y large (more than 50000 m 3/h) and
CS2 emissions forrn the driving force to treat them. The emissions of CS2 are
regulated in many countries and conventional air treatment techniques are often not
272 N..J. R. Kraakman
Figure 12.1. A PURSPRING bioreactor treating off-gases from an oii seed extraction plant.
cost-effective. It is important that the air is treated at any time and special
requirements for the bioreactor are necessary because of the risk of explosion linked
to the presence of CS2. The process control is much more complicated and special
training of the operators is required. The degradation of the compounds in
bioreactors is performed by special\y cultivated microorganisms and the degradation
process is control\ed by the salt content of the process water. The pH of the water is
extremely low. In some applications it is desirable that the effluent water coming
from the bioreactor can be reu sed in the production process in the plant.
4. Operating strategies
4. 1 INTRODUCTION
Pol\uted air streams nearly always contain a mixture of very different compounds.
Each application is different and the new bioreactor provides the required flexibility
needed to treat such pol\uted air streams. Examples of different operating strategies
that can be used are mentioned below.
Figure 12.2. A PURSPRlNG bioreactor treating off-gases from an anaerobic waste water
treatment plant of a brewery.
4.2 CONTROL PER LA YER OF MEDIA
Polluted air streams often contain a mixture of different compounds with different
chemical properties. These compounds differ from each other in their ability to be
biodegraded and their solubility in water. The compounds need to be transported to
the biofilm and degraded in it. To degrade aII these compounds a specialised mixture
of microorganisms is required. But not aII these microorganisms will present optimal
activity under the same process conditions. Microorganisms differ in their ability to
form a good biomass structure, in their growth rate, in their biodegradation capacity
as weB as optimal pH and nutrient conditions. Besides microorganisms differ from
each other with respect to their capacity to obtain their energy and nutrient sources.
As a result of the presence of different compounds in the air stream different
microorganisms are necessary and therefore different environmental conditions are
required.
One example in which different environmental conditions are preferred is
when a consortium of autotrophic and heterotrophic organisms represents the best
combination to treat an air stream containing a mixture of compounds. Many odorous
274 N.JR. Kraakman
air streams often contain a mixture of compounds that can best be treated by a
combination of autotrophic and heterotrophic organisms. The most optimal way to
treat such polluted air stream is by dividing the bioreactor into different levels
allowing optimal activity of different organisms under different environmental
conditions. Autotrophic organisms can biodegrade H2 S whilc hcterotrophic
organisms will remove odorous compounds separately.
Autotrophic organisms use carbon dioxide as carbon source in the air stream.
In a biofiltration system treating odour compounds these autotrophic organisms often
get the cellular energy for growth and respiration from thc oxidation of H2 S (often by
growing first Thiobacillus spp.). These aulotrophic organisms grow relatively slowly
and show high uptake rates even when sulphuric acid is present at such a
concentration that the pH becomes rather acidic. Heterotrophic organisms use
organic compounds (VOC) to obtain their cell carbon. Their growth rate is often
higher than for autotrophic organisms but they often require near neutral pH
environments.
In the bioreactor autotrophic organisms degrading HzS and heterotrophic
organisms degrading organic compounds are separated in different layers of media.
Because of the low Henry's constant of HzS compared to most other organic
compounds it is most common to degrade HzS in the first layer of thc bioreactor
while the organic compounds can be removed in the second layer. Removal of HzS in
the first layer prevents acidification of the media in the second layer and provides
optimal conditions for growth and survival of the heterotrophic organisms. So by
using different layers with different microorganisms a broader range of compounds
can be biodegraded.
4.3 TRRTGATION
Water can be added on top of each layer of media separately. Not only the amount of
water but also the frequency of irrigation can be regulated separately for each layer.
Water can also be added either continuously or discontinuously to a layer; in such a
way that the thickness of the water boundary can be regulated on the biofilm layer.
The water boundary is necessary for an efficient mass transfer and substrate
utilisation. Too much water leads to less void volume in the media, which is negative
for the removal of poorly soluble compounds. Compounds that are poorly soluble in
water need a long retention time in the media in order to reach efficient mass
transfer. Depending on the differences in compound solubilitics and mass transfer,
the irrigation program per layer must be adapted to each specific situation.
Manipulating the irrigation method per layer can increase the flexibility and air
purification efficiency.
Some additional considerations regarding operating strategy arc mentioned
hereafter. The process water in the bioreactor can either be recirculated over the
different layers or it can be added to one layer and then simply removed from the
bioreactor. When low pollutant concentrations are required in the out1et air stream of
the bioreactor it is better not to recirculate the water over the upper layer. Cleaned air
can otherwise come into contact with a concentrated water phase leading to stripping
of certain compounds from the water phase to the c1eaned air. On the other hand
recirculation can be interesting in some other situations. Microorganisms may get
slowly washed out from the media or they may be partly removed when
environmental conditions are not favourable for some of them in the biofilm.
Recirculation may be interesting when it allows maintaining high concentrations of

active biomass in the water phase. Research from Cox et al. (1998) is an example
that showed the benefits of such operational strategy in biotrickling filters. For a
stable operation of the installation in some situations reseeding microorganisms from
the bottom part to the upper part of a layer is required (Bioway, results not shown).
4.4 MEDIA COMPOSITION
The operating strategy can be changed by modifying media composition. In the new
bioreactor an inorganic medium is used. For different situations a different medium
composition is considered reaching either a more open or a less open medium
structure.
By changing media composition the surface area and the void volume can be
made different for each layer. A high surface area is necessary for a proper mass
transfer of the compounds from the gas phase to the water phase. It is also necessary
to be able to maintain as many active microorganisms as possible. It is known that
nearly only the outside of the biolayer is responsible for the biodegradation of
pollutants. This is owing to the restriction of oxygen and compounds diffusion
through the biolayer. The void volume is necessary in order to maintain as long as
possible the waste gas in contact with the medium. The removal of compounds with
high Henry's constants from an air stream is highly dependent on the mass transfer
and needs therefore a long retention time in a bioreactor.
Besides surface area and void volume the gas velocity can be of importance.
The gas velocity through the media affects the resistance of the layers of media. A
higher gas velocity leads to a higher resistance (i.e., pressure drop). Low resistance is
required because otherwise energy costs for the blower can reach up to 40% of the
total operational costs of a biological air treatment installation (Bioway, data not
shown). Often high gas velocities through a bioreactor are possible when air streams
contain low pollutant concentrations. When the overall removal of pollutants is not
being affected by creating a more or less open structure in the medium the
operational costs can be influenced significantly. So, media compositions can be
adjusted to the different types of air streams for different reasons. The most
important reason for changing medium composition is that it allows modifying the
surface area, the void volume and the resistance ofthe medium.
4.5 pH CONTROL
Many parameters may influence the actJvlty and growth of microorganisms.

Temperature, water activity, pH, nutrient conditions, oxygen concentration, substrate
toxicity are some examples of such parameters determining which microorganisms
will be active and to what degree. The pH is an important parameter in biological air
treatment because it has a great impact on microbial populations. H+ and OK ions
are the most mobile of all ions. Therefore small changes in their concentrations have
large effects. Most microorganisms prefer a pH corresponding to approximately
equal concentrations ofH+ and OK ions (pH = 7). Microorganisms are characterised
by different optimal pH ranges. Most bacteria prefer near neutral pH values while
fungi prefer lower pH ranges. Although most microorganisms may be active in a pH
range between 6 and 8, some of them may also be active near an extreme pH of 1 or
276 N.JR. Kraakrnan
near pH 13. In bioreactors pH needs therefore to be optimised and controlled

carefully.
In applications where biodegradation of pollutants leads to the formation of
acidic metabolites the pH of the medium wilI decrease. Air streams containing high
concentrations of compounds degraded to acid products can lead very fast to a
decrease in biological activity. For example, an air stream containing 1000 ppmv
H2S may lead within a couple of hours to a more than 50% reduction of biological
activity when the pH is not being controlled (Bioway, data not shown).
There are two ways to control the pH in a biological air treatment installation:
1) By the addition of caustic soda;

2) By the use of microorganisms being active at low pH.
In bioreactors sodium hydroxide is not often used to control the pH. When
using caustic soda, recirculation of the process water in the biological air treatment
installation is preferable. However, as discussed earlier recirculation is not always of
interest to reach high purification efficiencies. Another disadvantage of using caustic
soda is that it may significantly affect total operational costs. Moreover, the use of
chemicals requires the application of specific safety rules.
Microorganisms active at a low pH may be used in bioreactors. Then, the
acids formed during the degradation process can be concentrated and removed with
water. Microorganisms active at a pH below 2 have been used in several bioreactor
applications. With such tolerant strains, relatively high concentrations of acids may
be produced, meaning that water consumption during bioreactor operation can be
reduced. In some industrial applications produced acids can be reused in the
production plant. If not, the running off water should be mixed with other water
streams. When other water streams are not available this running off water can be
neutralised with, for example, lime if necessary. So, using microorganisms adapted
to an extreme pH allows overcoming the disadvantages of using caustic soda for pH
regulation.
In some bioreactor applications the presence of ammonia in the air stream
alIows natural pH regulation. Ammonia dissolves in the water phase and forms NH4+
ions, capable of neutralising acid degradation products (sulphuric and nitric acids in
our applications).
When the process water is at a very low pH the measurements are often not
very ac curate as the pH scale is logarithmic. Therefore in some applications not only
the pH but also the salt content is measured. To determine the salt content water
conductivity is measured.
5. Start-up and process control
5.1 START-UP
For efficiently removing pollutants from a waste gas, the best available
microorganisms should be used. Because most of the waste gasses contain a mixture
of compounds a variety of microorganisms is mostly necessary to be able to degrade
aII the compounds. It is often thought that when the environmental conditions in a
biotreatment system are optimal microbial populations present in the waste gas or
present on the (organic) medium wiII rapidly show up. Nevertheless, different
situations have shown that the growth of the optimal populations can take weeks or
even months (Chapters 3 and 4). It is important to ensure a good build up of the
biofilm on the media especially when the media are inorganic like in our bioreactors.
In our bioreactors inocula are always seeded on start-up. Biocultures capable
of efficiently degrading the compounds are used as inocula. Important parameters are
for example the ability to be active in a certain temperature range, at an extremely
low pH or, with a minimal use of nutrients. Start-up takes normally one to about four
weeks. During start-up, the process water is recirculated over the bioreactor.
Depending on the water source for the installation extra specific nutrients are used
during this growing phase of the microorganisms. Some factors like pH and
temperature can be of a large influence during the start-up. It has for example been
observed in one application that with only a few degrees (3 to 4 DC) temperature
increase the start-up period could be reduced from three weeks to one week.
5.2 PROCESS CONTROL AND OVERCOMING OPERATING PROBLEMS
It is important to be able to carefully measure and control process parameters.

Depending on the application transmitters for temperature, pH, conductivity, airflow,
pressure drop, water flow, water pressure, nutrient flow and compound
concentrations allow on-line measurements in the bioreactor. The standard control
system is provided with an automatic pH controlling device and contains an interface
inc1uding remote control. Alarms can be generated and manuals are provided with
standard operational procedures. Maintenance is very important for every biological
installations, also for the present type of bioreactor. Ali the bioreactors are visited
two to four times a year for a check up of the bioreactor and if necessary reports with
advices are provided to the operators.
When interruptions are necessary during holidays or maintenance periods
special shut down procedures have been developed so that a bioreactor can come
back to full service as fast as possible after shut down. An example of this is the
procedure and assistance for the use of thiosulphate every year during a three week
summer shut down of a specific bioreactor application (data not shown).
6. Examples
Several applications ofthe different types of bioreactors are shown below.
6.1 CASE 1: POTATO PROCESSING PLANT
Application
Settling tank of anaerobic water purification at a potato processing plant.
Airflow characteristics
• Flow: 250 m31h, temperature: 10-35 ec, humidity: 70-90%
• H2 S concentration raging from 400 up to 1000 ppmv
• CH4 concentration « 5% LEL)
• Odour compounds (fatty acids, organic sulphur compounds, ketones, aldehydes,
etc. )
278 N.JR. Kraakman
Type oJbioreactor
• PURSPRING system PS2150
• Footprint bioreactor: 2 m 2
• Built in 1997
Removal efficiencies
• H 2S removal > 98% (average over three years, measured daily first and measured
ones a week after the first three months)
• Odour removal > 97% (measured twice)
6.2 CASE 2: BREWERY
Application
Air from an anaerobic water purification plant at a brewery.
Airjlow characteristics
• Flow: 700 m 3/h, temperature: 15-25 DC, humidity: 60-80%
• H 2 S concentration ranging from 700 up to 1400 ppmv
• Odour compounds (organic sulphur compounds, fatty acids, mainly acetic acid,
ketones, aldehydes, aromatic compounds, terpenes, alkanes, mainly methane)
Type oJbioreactor
• PURSPRING system PS3250 (see Figure 12.2)
• Built in 1998
H2 S removal > 97% (average over 1 year. measured daily).
6.3 CASE 3: SPONGE MANUFACTURING PLANT
Application
Sponge manufacturing process air.
Airjlow characteristics
• flow: 8500 m 3/h, temperatures 25-35 DC, humidity: 80-90%
• H 2S concentrations ranging from 200 up to 400 ppmv
• CS 2 concentrations ranging from 200 up to 400 ppmv
Type oJbioreactor
• V-SPRING system PS3400 (Figure 12.3)
• Built in 1999
• H2S removal > 90% (average over 1 year, measured twice a day)
• CS2 removal > 75% (average over 1 year, measured twice a day)
Case sludy - New hioreaclor 279
Figure 12.3. V-SPRING bioreactor treating off-gases from a sponge mallufacturing plan!.
6.4 CASE 4: MUNICIPAL WASTE W ATER TREATMENT PLANT
Application
Municipal waste water treatment plant.
Airflow Characteristics
• flow: 500 m 3/h, temperature: 5- 35 ec, humidity: 50- 90%
• H2S concentrations up to 300 ppmv
• Odour compounds
Type oJbioreactor
• ZEROCHEM-unit ZC1850 (Figure 12.4)
• Built in 1999
Removal ejjiciencies
H2S removal > 99% (average over three months during the summer period, on-line
measurement). See Figure 12.5 for H2S removal during day and night.
280 N.J R. Kraakman
Figure 12.4. ZEROCHEM bioreactor operating at a waste water treatment plan!.
A list of aII bioreactors that have been built and operated successfully is
presented in Table 12.1. It can be seen that some of the bioreactors have been
operating in the field for many years now. The number of applications is growing
and it is expected that many more bioreactors will be built in the near future.
Information on the investment and the operating costs of a bioreactor system
is presented Table 12.2. The costs are compared with other conventional techniques
like chemical scrubbing, incineration, adsorption and conventional biofilters.
7. Future developments
Each kind of industrial appIication needs a tailor made solution. As describe above
three types of bioreactors are standardised for specific applications. Future
developments will focus on many other applications. One of the developments is the
ability to increase removal efficiencies of very poor soluble compounds. Because of
the abiIity to operate the bioreactors in different ways to obtain optimal process
control this is one ofthe future development that is necessary.
Table 12.1. List ofbioreactors that have been built and operated successfully
AIRFLOW
RAlE IMPORTANT YEAROF
PROJECTS (m 3fh ) COMPONENTS INSTALATION
Waste water treatment plant 500 H2S,odour in progress
Liftstation off-gas 500 H2S,odour in progress
POTW off-gas 2000 H2S,odour In progress
Sponge manufacturing process air 50000 H2S, CS 2 in progress
Waste water treatment plant 700 H2S,odour 2000
Rayan manufacturing process air 12500 H2S, CS 2 in progress
Brewery ground water processing off-gas 34000 H2 S 2000
POTW off-gas 1700 H2S,odour 2000
Oii production plant 2500 H2S (up to 1500 ppmv) 2000
Sewerage 500 H2S, adour 1999

Sponge manufacturing process air 8000 H2S,CS 2 1999
H 2S (up to 2000 ppmv),
Brewery waste water off-gas 700 odour 1999
H 2S (up to 2000 ppmv),
Brewery waste water off-gas 700 odour 1998

Refuse dump waste water holding tank 350 H2S, NH3, odour 1998
Pesticides manufacturing process air 9000 H2S,CS 2 1998
Rayon manufacturing process air 500 H2S, CS 2 1998
H2S (up to 1000 ppmv),
Oii production plant 1000 adour 1997
Extract deposit tank of anaerobic water H 2S (up to 1000 ppmv),
purification 250 adour 1997
Liftstation off-gas 1200 H2S, adour 1996
POTW off-gas Methylsulphides, H2S,
600 adour 1996
Chicken muck processing, Ysselstein DMDSIDMTS, adour,

1400 H2 S,NH3 1994
N
f:3
Table 12.2. The investments and the operating costs of a bioreactor system compared with other air purification techniques.
Based on an air stream of 8500 m 3/h with 50 ppmv H2S
Single stage Multistage Engineered Carbon adsorption Purspring

wet scrubber wet scrubber biofilter bioreactor
Capital cost $75000' $110000 $180000 $90000 $130000
Annual power $5150 $7800 $4600 $4600 $500
cost (1)
Annual sodium hydroxide cost (2) $4600 $4600 $0 $0 $0
(28 gpd) (28 gpd)
Annual sodium $47600 $11000 $0 $0 $0
hypochlorite (180 gpd) (41 gpd)
:<:
<....,
cost (3) ~
Annual carbon $0 $0 $0 $67000 $0
cost (4) (42000Ib/yr) ~
I:l
Annual media $0 $0 $34000 $0 $0 I:l
cost (5)
Annual maintenance labour cost (6) $3900 $3900 $2600 $2600 $2600
tI
I:l
;:;
(3 h1week) (3 h/week) (2 h/week) (2 h/week) (2 h/week)
Total annual $63900 $24650 $41200 $74200 $3100
operating cost
Total present value cost, 10-year life $467000 $261000 $433000 $546000 $149000
(1) Bascd an $0.07 per kWh. Source: Industrial Wastewater, May/June 1999
(2) Based on 25% solution, $0.45/gal.
(3) Based an 12.5% solution, $0.73/gal.
(4) Based an $I.OO/lb.
(5) Annualised, based an complete media replacement in five years.
(6) Based an $25/h.
0.95 $ (US Dollar) ~ aprox. 1 € (Euro)
H2S REMOVAL
ZEROCHEM-BIOREACTOR
300,---______________________________________,30
250~-----------------------------------------t25
UJ •••• • • (1)
~ 200fWl~._~------~~~~----_,~~~~~·~--~20 ~;
~~
~ ~
~
~ Q.
a
UI..... 150 15 B5"""::
~~ ~~
u~ 8~
UlM 100 10 o
(l)N
~ ~ ~
50~--------~~------------~~------------~-t,
....")~~~IO'?~o)'Y~--;~°!j'Y<;)rţ'?°?j'Yc'\.'?~c:5'Y~~~~'Y~.'?~<S'Y~rţ'?O...;~'!i'?(')l6ftO~~...;'Y~'?'?~(cj'Y~qj'?~...;'Y~'"'J'?0oţ'YCl~'?()'\.'Y<:;:).
T1ME
Figure 12.5. Biological air purification in a ZEROCHEM bioreactor. Fluctuating H2 S

concentrations due to changing conditions during day and night.
A future development is also to increase the degradation capacltles of the

bioreactors. Higher degradation capacities will lead to a reduction of the investment
costs. Different microorganisms will be investigated in different applications. One of
the risks when the degradation capacities are increased is that biomass growth might
lead to c10gging ofthe media. More strategies will be developed to prevent excessive
biomass growth and to ensure stable operation in the long run.
The drainage water of the bioreactor is often acidic and can be reused in some
specific applications. Research is at this moment focussed on treating the effluent
water with physical processes for these situations where recycling of the water is
desirable.
References
Alonso, C., Zhu, X., Suidan, M.I., Kim, B.R. and Kim, B.R. 1998. Modeling of the biodegradation
process in a gas phase bioreactor - estimation of intrinsic parameters. In: Proc. Conference on
Biofiltration. Reynolds F.E. (ed.), Tustin, California.
Cox, H.H.J. 1995. Ph.D. Dissertation, State University of Groningen, Groningen, The Netherlands.
Cox, H.H.J. and Deshusses, M.A.. 1999. Chemical removal of biomass !Tom waste air biotrickling
filters: screening ofchemicals of potential interes!. Water Res. 33: 2383-2391.
Cox, H.H.J., Nguyen, T.T. and Deshusses, MA 1998. Pollutant biodegradation in biotrickling filtes
revisited. In: Proc. Conference on BiofiItration. Reynolds F.E. (ed.), Tustin, California.
Cox, H.H.J., Nguyen, 1.1. and Deshusses, M.A. 1999. Predation of bacteria by the protozoa
Tetrahymena Pyriformis in toluene-degrading cultures. Biotechnol. Le!. 21: 235-239.
Groenestijn, J.W. van and Hesselink, P.G.M. 1993. Biotechniques tor air pollution control.
284 NJR. Kraakman
Heiningen, W.N.M. van, Kraakman, N.J.R., Gerbcns, S. and Groenestijn, lW. van. 2000. The
development of a biofilter system using fungi on inert carner material in combination with mites. In:
Proc. Conference on 13iofiltration Reynolds VE (ed ), Tustin, California
Heslinga, D.e. and Groenestijn, J.W van. 1997. "Jbennophilic biological wastc gas cleaning. In: Proc.
Intem. Symp. Biological Waste Gas Cleaning. Prins W.L and van Ham 1 (eds.). VDI Verlag GmbH,
DUsseldorf, Germany.
Kraakman, N.J.R., Groenenstijn, J. W. van, Koers, B. and Heslinga, D.C. 1997. Styrene removal using
a new type of bioreactor with fungi. In: Proc. Intem. Symp. Biological Waste Gas Cleaning. Prins
W.L and van Ham] (eds.). VOI Verlag GmbH, Diisseldorf, C;cnnany
Melse, R. W. and Kraakman, N.J.R.. 1998. Biological treatment of waste gases containing !I,S and CS 2
combined with the production of concentrated NaOH and H,S04' Meded. Fac. Landbouww. Gen!. 63:
1841-1847.
Picioreanu, C., Loosdrecht, M.e.M. van and Heijnen, 11 1999. Discretc-differential modelling of
biofilm slructure. Water Sci. Technol. 39: 115-122.
Seignez, e., 13odart, V., Lombardot, T., Stettler, M., Thoeni, C., Peringer, P and I1011iger, C 2000. A
biotrickling filter to treat chlorobcnzene-contaminated waste gas: Inoculum production, performance
and bacteria involved. In: Proc. 4lh International Symposium Environmental Biotechnology. IIartmans
S. and Lens P. (OOs.), Noordwijkerhout, The Netherlands, pp. 105-108.
Song, J. and Kinney, K.A. 2000. Effect of vapor-phase bioreactor operation on biomass accumulation,
distribution, and activity: linking biofilm properties to bioreactor perfonnance. Biotechnol. Bioeng. 68
50S-S I Ii.
Woertz, J.R., Kinney, K.A., Mclntosh, N.D.P. and Szaniszlo, P.J. 2001 1 Air Wastc Manage. Assoc.
(in press).
CHAPTER 13 BIOSCRUBBER FOR TREATING WASTE GASES
FROM WASTE WATER TREATMENT PLANTS
Niels G. HANSEN and Kim RINDEL
1. Introduction
The need to prevent public nuisance, particularly for installations near residential or
tourist areas, means that there is an increased need for the control of odour emissions
from waste water treatment plants (WWTP). Consequently odour emissions must be
reduced through the use of a better technology and cleaning of the resulting
emissions.
In waste water treatment, odour nuisances typically arise in the sewerage
system and at the waste water treatment plant, particularly at the inlet works and in
sludge treatment facilities. Problems are primarily caused by volatile sulphurous
compounds such as hydrogen sulphide and mercaptans. By covering and ventilating
the critical sections of the plant it is possible to create a satisfactory working
environment. However, it is often necessary to remove the odorous substances from
the ventilation air to avoid nuisances in the neighbourhood.
The waste water treatment plants Lynetten and Damhusaaen are owned and
operated by the partnership Lynettefaellesskabet IlS. The plants treat waste water
from eight owner municipalities. In total the plants treat approximately 275000 m3/d,
which is a population equivalent of 1.1 million. During the period 1991-97, both
treatment plants were extended to include nutrient removal. The costs of the
extension amount to f 160 millions.
1.1 DAMHUSAAEN WASTE W ATER TREATMENT PLANT
The Damhusaaen WWTP treats sewage from a population of 226000 (350000 PE).
Primary treatment comprises 6 mechanical screens, 2 aerated grit and grease
chambers and 7 circular primary sedimentation tanks with a hydraulic capacity of
28000 m3/h. The screens and the grit and grease chambers are covered and
ventilated. The combined biological and chemi cal waste water treatment (Bio-
Denipho with simultaneous precipitation) takes place in four lines comprising
anaerobic tanks and aeration tanks having a total volume of 80000 m 3 . Final
sedimentation takes place in twenty-four circular tanks with a hydraulic capacity of
10000 m 3/h. The waste water is discharged into the Sound via a 1.5 km long pipe.
Primary sludge is digested in four digesters totalling 7600 m 3 and then dewatered
together with the secondary sludge in four lines comprising drum screens and
centrifuges. The produced gas is used to generate electricity and heat. The dewatered
sludge is transported by lorry to the Lynetten WWTP for drying and incineration
together with sludge from Lynetten. The ash is deposited in an environmentally
approved landfill at Lynetten.
The works dates back to 1930 and has later been extended by stages. As the
built up area becomes larger, the residential area has come c1oser, which means there
are residences very c10se to the plant.
As a result of the odour nuisances from the treatment plant, in 1992-93 a
survey was carried out to reveal the odour sources. The inlet structure, particularly
285
C. Kelllles alld M. C. Veiga (eds.), Bioreactorsfor Waste Gas Treatment, 285-298.
286 N. G. Hansen and K. Rindel
Figure 13.1. Aerial view ofthe Damhusaaen WWTP.
the screen ing and the grit and grease chamber, tumed out to be the most important
odour sources at the treatment plant. It was therefore decided that measures should
be taken to reduce the odour emission from this area.
An aerial view of the Damhusaaen WWTP is shown in Figure 13.1. The inlet
works is seen in the middle, left hand section of the picture ne ar the circular primary
clarifiers.
1.2 ASSESSMENT OF METHODS TO REDUCE ODOUR EMIS SION
Several methods of odour reduction were available. In addition to covering/enclosing

the process equipment, combined with ventilation and cleaning of the ventilation air,
the possibility of dosing one or more suitable chemicals into the sewer system was
considered. A well prepared dosage of oxygen Of nitrate into the waste water could
probably prevent the formation of odorous substances to a high degree. However,
based on a general assessment it was decided not to carry through this option, mainly
due to the relatively high operating costs and the necessity of a comprehensive
preliminary study because of the extensive and complex sewer system in the
catchment area. It was consequent!y decided to cover/enclose the most critical
sections together with ventilation and cleaning of ventilation air.
1.3 AIR CLEANING METHODS
The choice of the air cleaning method was based on considerations in relation to
operating costs, operational reliability, environmental friendliness, and space
requirement.
Case study - Bioscrubber 287
Conventional air c1eaning methods at waste water treatment facilities are

chemical scrubbing and activated carbon filtration. However, these methods are
characterised by relatively high operating costs. Furthermore, chemical scrubbing
involves oxidising agents, whose use may be environmentally questionable.
Biological methods, especially the so called biofilters, are becoming
increasingly popular, mainly owing to their environmental friendliness and low
operating costs. In the biofilter, biological degradation of odorous substances is
achieved by means of naturally occurring materials, such as bark, compost and peat,
on which a biofilm develops. Air flows through the biofilter and the pollutants are
adsorbed onto the biofilm and subsequently degraded.
Experience in recent years has, however, shown that certain basic
requirements must be met to ensure an efficient and reliable operation of a biofilter
(Williams, 1994). Important parameters are the moisture content, pH, accumulation
of reaction products and the mechanical stability of the filter material. Most
frequently the reason for poor functioning of biofilters is inadequate moisture
control, but inhibition owing, for example, to acidification and accumulation of
inorganic saIts may also reduce the effectiveness of a biofilter. Modern biofilter
constructions can to a certain degree combat these problems, but at considerably
higher construction costs than the more primitive constructions (van Lith et al.,
1997).
Another approach to biological air c1eaning is the use of a bioscrubber, in
which biological degradation takes place in a liquid phase containing suspended
microorganisms (Chapter 5). In this way the above mentioned requirements are
satisfied. Based on positive experience over a number of years in animal feed
production applications (Hansen and Rindel, 1992; Rasmussen et al., 1994) and
extensive pilot testing at the Lynetten WWTP, the bioscrubber has been further
developed for c1eaning air in applications such as waste water treatment plants.
A bioscrubber can be operated with much higher air loads than a biofilter
(3000-4000 m 3jm2 .h compared to 100-150 m3 jm2 .h). The higher air load reduces
space requirements and thereby construction costs. Furthermore, operating costs for a
bioscrubber are, for example, lower than for a chemical scrubber because the
microorganisms are working for free - a situation which becomes more evident the
greater the pollutant content of the inlet air.
The above considerations thus conc1uded in a solution which comprises cover
of the inlet pumping station and the grit and grease chamber, enc10sure of screens
and ventilation from containers for grit and screenings. The total ventilation air, 6000
m 3jh, is cleaned in a bioscrubber. Furthermore, a not insignificant part ofthe decision
basis was that a process based on biological activated sludge treatment scemed to be
the most 'natural' choice in a waste water treatment plant.
2. Plant description
2.1 COVERS AND ENCLOSURES
2.1.1 Grit and grease chamber

The aerated grit and grease chamber is covered with aluminium profiles. The cover is
provided with a track, through which the scraper arm for the grit chamber bridge can
pass. The track is covered with rubber plates which deflect when the scraper arm
288 N G. Hansen and K. Rindel
passes. The amount of exhaust air is sufficient for produc ing a weak negative
pressure under the cover, thus taking care of both the process air used for aeration of
the chamber and atmospheric air taken in through openings in the cover.
2.1.2 Screens
The screens are covered with aluminium plates, allowance being made for access by
the staff in connection with service inspection and repair works (Figure 13 .2).
Ventilation is not taken directly from the screen building but from the inlet
chamber below the screen building, where a weak negative pressure is maintained.
By also establishing a weak positive pressure in the screen building, a reduction of
the exhaust air flow is obtained and with it a less expensive air c1eaning plant and a
better working environment for the staff.
2.2 AIR CLEANING PROCESS
The bioscrubber method combines two well-known processes: (i) an absorption unit
with inert media (a packed column) in which the pollutants are transferred to a liquid
phase; and (ii) a bioreactor based on the activated sludge principle in which the
pollutants are oxidised by means of microorganisms. A flow diagram of the
bioscrubber process is shown in Figure 13.3. The activated sludge from the
bioreactor is used as the absorption liquid in the column. Thus the two sub-processes,
mass transfer and biological oxidation, take place in separate but interconnected
units, simplifying the dimensioning and operat ion ofthe plant.
The absorption column contains an inert plastic or metal packing which
provides maximum contact between the air and the liquid phases and well defined
and predictable flow conditions. Before the c1ean air is released to the atmosphere. it
passes through a demister, in which liquid droplets are separated from the air.
Figure 13.2. Screen covers.

filler
P1 NaOH
Figure 13.3. Flow diagralll ofthe bioscrubber.
The bioreactor is designed and constructed on the basis of the degradation

rates for the specific poJlutants involved. The end products from the biological
degradation of the odorous sulphur compounds are low load waste water containing
sulphate and smaJl amounts of surplus sludge.
Effective process control is ensured through application of relevant
instrumentation to control the liquid level, pH and salt concentration. Control of the
dissolved oxygen content in the sludge suspension is normaJly not necessary because
sufficient oxygen is transferred in the absorption column. The bioscrubber is
therefore based on relatively simple systems and facilities, aJlowing a high degree of
automation and reliability.
The design of the absorber section is based on common principles of
absorption of gases into a liquid in which a chemi cal or biological reaction enhances
the mass transfer. Figures 13.4 and 13.5 show the results of measurements of
c1eaning efficiency in relation to hydrogen sulphide (Heist et al. , 1995).
Figure 13.4 shows the influence of liquid load on the absorption of H2S, and
Figure 13.5 iJlustrates the effect of varying the pH. The full drawn Iines represent
model calculations based on the proposed absorption theory. In both cases the
measured values adapted weJl to the model calculations. The model was developed at
air velocities of about 1 m/s, which is approximately 30 times the air velocity of
biofilters.
95
90
x
L= 32 m 3j(m2 .h)
85 +.--------.-------~--------~------~
0.50 0.75 1.00 1.25 1.50
Air velocity (m/s)
Figure 13.4. Influence of liquid load L on the cleaning efficiency for hydrogen sulphide.
Packed height = 6 m, pH = 8.5.
100
~
90
tf<
'--'
>-. 80
g 70
v
'u 60
}5
v
gp 50
'S
CIj
40
v
O 30 .
20 -
0.50 0.75 1.00 1.25 1.50
Air velocity (m/s)
Figure 13.5. Influence of pH on the cleaning efficiency for hydrogen sulphidc.

Packed height = 6 m, liquid load = 32 m 3jm 2·h.
2.3 FULL-SCALE PLANT
The bioscrubber at Damhusaaen WWTP (Figure 13.6) was commissioned during the
latter part of 1996. The bioscrubber is made of stainless steel and operates
automatically as an integral part of the SCADA system (Supervisory Control And
Data Acquisition) ofthe waste water treatment plant.
It was required that the air leaving this particular plant should be as odourless
as possible. The air from the bioscrubber is therefore treated in a polishing filter to
remove any remaining odorous substances of low water solubility, which are not
absorbed in the scrubber section. The polishing filter conta ins an adsorbent. As most
of the polIutants are removed in the bioscrubber, the adsorbent has a long life
expectancy, thereby ensuring economical operation.
Figure 13.6. Bioscrubber at the Damhusaaen WWTP. In front the gri! chamber.
The construction is adapted to the existing screen building. The polishing

filters, for instance, are located on the roof of the building, while machinery and
measuring equipment are located in a free corner inside the screen structure.
Scrubber column and bioreactor are placed close to the building so that the entire
construction takes up a minimum of additional space.
Main design data for the bioscrubber are as folIows :
Variation
Air flow, (m 3/h) 6000 0- 6000

Air temperature, (0C) 15 5- 30
Moisture content, (% r.h.) 80 40- 100
Inert dust, (mg/m 3 of dry air) <I 0-5
Hydrogen sulphide, (mg/m 3 of dry air) IO 0-50
Methyl mercaptan, (mg/m 3 of dry air) 2.5 0-5
Dimethyl sulphide, (mg/m 3 of dry air) 2.5 0- 5
Dimethyl disulphide, (mg/m 3 of dry air) 0.25 0-1
The dimensions ofthe bioscrubber plant are as follows:
Diameter of scrubber, (m) 1.6

Height of scrubber, (m) 12
Packed volume, (m3) 12
Sump volume, (m 3) 7
Liquid circulation rate, (m 3/h) 70
Diameter ofbioreactor, (m) 1.9
Volume ofbioreactor, (m 3) 11
Volume of polishing filter, (m 3) 2 x 0.68
Start-up of the biological process was performed by addition of activated

sludge from the aeration tanks of the waste water treatment plant. After a few days
the original microbial culture, which was dominated by heterotrophic bacteria that
convert organic matter, adapted to the new conditions. Subsequent1y, autotrophic
sulphide oxidisers, presumably of the Thiobacillus species, dominated. Figure 13.7
illustrates this with results from pilot tests (Heist el al., 1995).
80 , - - - - - - - -
70
60
~ 50
~J
<;\; 40
Ei
~ 30
o
20
la
o~~~~~~~~~~
o 100 200 300 400
:-.Jumber Dt days 1'rom start
Figure 13.7. Oxygen Uptake Rate of activated sludge.
The Oxygen Uptake Rate (OUR), i.e., the oxygen consumption during the
biological oxidation (and endogenous rcspiration) in terms of the suspended solids
(SS) concentration was dctermined both with and without addition of a carbon source
(acetate) and with addition of sulphide. A fast adaptation of the sludge to the
sulphide was seen which, after a period of time, was oxidised 5-10 times the original
rate in unadapted activated sludge.
The tests with and without acetate showed that readily biodegradable organic
substances were not present in the samples and that the activated sludge only slowly
degraded acetate after a lag time. This indicated that the activated sludge contained
fewer heterotrophic microorganisms than normally, prcsumably owing to a low
content of organic matter in the air.
Therefore, it was not necessary to seed with speciali sed bacterial strains for
this application. This was also the case with regard to other applications of biological
air cleaning plants where native compounds were to be treated. In most cases, it will
be an advantage to start with a mixed culture as it is more robust than a specially
prepared culture. Furthermore, in many applications a specially prepared culture will
be difficult to maintain, as the introduced air is not sterile. Only in cases where non-
native compounds such as chlorinated solvents are treated, it seems an advantage to
use specially prepared microbial cultures.
Certain precautions should be taken to maintain biological activity:
• Salt concentration should be as constant as possible, approx. 2% w/w, as

otherwise inhibition of bacteria may occur. In practice, this was done by
discontinuous discharge of waste water, based on measurement of conductivity of
the liquid, as this is directly proportional to the salt concentration. Predominantly,
the dissolved salts consisted of sodium sulphate formed by the biological
oxidation of sulphur compounds in the air. The sulphuric acid thereby formed
was neutralised by sodium hydroxide. During the automatic discharge procedure,
the sludge flocs settled down in the bioreactor, and the supernatant was
discharged.
• Likewise, pH should be kept rather constant to avoid inhibitory effects. Sodium
hydroxide was dosed on the basis of continuous pH measurements. The pH
optimum was near 8.5-9.0, where a high biological activity could be maintained,
as well as an effective absorption of hydrogen sulphide obtained at the same time
(Figure 13.5).
• Normally, nutrient salts and micronutrients were not added with the polluted air.
Therefore, continuous dosing of small amounts of phosphorus, potassium and
nitrogen was necessary together with different trace metals and vitamins. This
dos ing could be avoided or reduced by using treated waste water as feed water
for the scrubber. However, in the current case the relatively small water
consumption did not justify execution of the necessary installations.
3. Process experience
After more than four years' operation of the bioscrubber, the following important
experience has been gained.
3.1 START-UP
Start-up ofthe plant was rapid and straightforward because the activated sludge from
the aeration tanks was used as seeding material. This means that seeding with special
cultures was not necessary.
3.2 CLEANING EFFICIENCY
The cleaning efficiency is high, > 99 % for hydrogen sulphide in the scrubber section
alone with inlet concentrations of up to 75 mg H2S/m 3 . Outlet concentrations lower
than 0.1 mg/m 3 are easily obtained with the polishing filter in operation. Organic
294 N. G. Hansen and K Rindel
sulphur compounds only appear at low concentrations and are reduced to less than
0.1 mg/m 3 through the plant. The ave rage hydrogen sulphide load is about 10 mg/m 3
on an annual basis and approximately 25-35 mg/m 3 during the summer and autumn.
3.3 PRESSURE DROP
The pressure drop of air in the scrubber section is stable, which means that there is
no uncontroUed biomass growth on the plastic packing and therefore no risk of
clogging. This is a result of the low growth rate of the sulphide oxidising bacteria
combined with the continuous recirculation of scrubbing liquid.
3.4 pH MEASUREMENT
The pH measuring and control equipment proved to be very reliable. pH is an

essential parameter for hydrogen sulphide removal (as illustrated in Figure 13.5) and
must be controUed if a high cleaning efficiency is to be maintained. The pH is
measured by two separate units. Any deviation between the two measurements is
monitored by the PLC and causes an alarm to be activated.
3.5 MAINTENANCE
Maintenance is reduced to control and servicing of instruments and machines and

regular checks of the operation by means of the SCADA system. It is necessary to
add smaU amounts of nutrients, mainly phosphate at intervals. Contrary to certain
misconceptions, maintenance of the bioscrubber is thus limited and does not include
operations unknown to the operating staff.
3.6 CONSUMPTION OF CHEMICALS
The consumption of sodium hydroxide for neutralising the produced sulphuric acid is
stoichiometric (2.4 kg NaOH/kg H2 S) in relation to the amount ofhydrogen sulphide
absorbed at a pH level of 8.5-9.0 in the scrubber liquid. This means that no carbon
dioxide is absorbed from the air, which would have meant a higher consumption of
chemi caIs - as is the case in chemical scrubbers which are operated at a higher pH.
The consumption of adsorbing agent in the polishing filters depends upon the
content of substances with low solubility in the ventilation air, because, in practice,
aU the hydrogen sulphide is removed in the scrubber section. Fresh adsorbing agent
added in March 1998 was not exhausted until October 1999. Therefore, the
consumption can be expected to be approximately 400-500 kg/year. As the pollutant
load during winter is rather low, fresh adsorbing agent was not added until April
2000 and it stiU did not show any sign of exhaustion in November 2000.
It is to be expected that foaming might occasionaUy occur in the sludge
suspension. To prevent foaming, an antifoam agent is dosed into the bioreactor.
Dosing takes place discontinuously in predetermined amounts controlled by a timer
function. The average consumption is about 1 l/d. Furthermore, a foam sensor placed
above the liquid surface in the scrubber section activates the dosing pump in order to
suppress any uncontrolled foaming.
3.7 OPERATING COSTS
The operating costs for the bioscrubber are only about half those for a chemical
scrubber at inlet concentrations of about 10 mg H2S/m3• This is mainly owed to the
lower consumption of chemicals, because the oxidising agent of a chemical scrubber
is replaced by microorganisms in the biological process, and the near-stoichiometric
consumption of sodium hydroxide in the bioscrubber. The higher the inlet
concentration, the more advantageous the biological process. Therefore, at an inlet
concentration of 50 mg H2S/m3, the operating costs for the bioscrubber are as low as
one fourth of the operating costs of a chemical scrubber. Figure 13.8 illustrates this
for two different inlet concentrations.
The comparison is based on a one-stage bioscrubber with a polishing filter
placed downstream and a traditional two-stage chemical scrubber with an acid first
stage used for removal of nitrogen compounds (which may otherwise cause
undesirable side reactions in the next stage), and a basic and oxidising scrubber
section utilising sodium hypochlorite.
1.4
scrubber
1.2
.~
ao 0.8
o
8
4l 0.6 Chemical
.... scrubber
'"
o 0.4
U
0.2
o
Low inlet concentration High inlet concentration
~ Power ~ Chemicals o Water lllllI Maintenance
Figure 13.8. Operating costs of a bioscrubber and a chemical scrubber. Low inlet
concentrations correspond to 10 mg H2S/m1 and 3 mg org.S/m1, high inlet concentrations
correspond to 50 mg H2S/m1 and 3 mg org.S/m1 •
The calculations are based on current Danish prices of chemicals, electricity,

water and manpower. The calculations on the chemical scrubber are based on
inforrnation from an experienced supplier, who has delivered and operated numerous
plants in connection with waste water treatment.
Under the above conditions the operating costs of a bioscrubber and a modern
biofilter will be almost identical, provided that a reasonably long life is obtained for
the filter medium of the latter. The operating costs for a biofilter are rather sensitive
to the life span ofthe filter media.
The construction costs for a bioscrubber are at the same level as for a two-
stage chemical scrubber. However, the construction of modern biofilters has proved
296 N.G. Hansen and K. Rindel
to be costly; therefore, in comparison, a bioscrubber is an advantageous solution

even at a relatively low air flow.
3.8 EFFECT OF ODOUR EMISSIONS ON THE NEIGHBOURHOOD
In the summer period of 1997, complaints were received of odour being a nuisance to
the neighbourhood of the waste water treatment plant. The Lynettefaellesskabet
contacted the Environmental Control Body (Miljokontrollen) in Copenhagen, which
is the supervising authority for the plant. It was agreed that the Body should make a
number of odour observations in the are a in the periods October 1997 to February
1998 and May 1998 to September 1998. The observations, which were made in the
plant lee side in radii of 200 m and 400 m respectively from the primary treatment
(potentially the largest odour source), where also the bioscrubber is located, were
characterised as eno odour', 'noticeable odour which, however, does not justify
complaints' and 'excessive odour'. The total number of observations amounted to
138 in the first and 149 in the second period, each period reaching a twenty days
duration.
A survey of the vicinity of the plant revealed a number of potential odour
sources, which were not generated by the Lynettefaellesskabet. The sources in
question were three pumping stations, a sludge receiving plant and a combined sewer
overflow. There are other potential odour sources in the area, for example, a
composting plant is located nearby. The results of the odour observations are given
in Table 13.1 (Miljokontrollen, Copenhagen, 1998).
Table 13.1. ResuJts of odour observations in the neighbourhood of the Damhusaaen WWTP
Period Observations (numbers) Observations (%)

No Noticeable Excessive No odour Noticeable Excessive
odour odour odour odour odour
Oct. 1997-
Feb. 1998 113 24 82 17
May 1998-
Sept. 1998 136 13 O 91 9 O
Based on the survey, it was concluded that Damhusaaen WWTP did not emit
any noticeable odour during the test periods.
4. Future measures
Similar covering of the grit and grease chamber at Lynetten WWTP (750000 PE) is
intended. In addition, the overflow channels in the primary sedimentation tanks will
be covered. The covering is made partly for the sake ofthe working environment and
partly to avoid odour being a nuisance to the adjoining site, which no longer houses a
ship building industry but other activities such as a drive in cinema. To begin with,
the discharges from the screens and the grit and grease chamber are collected, and
whether cleaning in a bioscrubber is necessary will be decided based on

measurements of the discharge.
In general, the bioscrubber might be used for cleaning ventilation and process
air from, for example, waste water treatment plants, waste treatment plants,
composting plants, fish industry, animal feedstuff industry and rendering plants.
However, other types of odour emitting industries might also benefit from using the
bioscrubber for waste air cleaning.
When high concentrations of acidifying substances such as hydrogen sulphide
and ammonia (acidification through biological nitrification) are to be treated and
neutralised, the bioscrubber offers an advantage over the biofilter because the latter
will quickly be acidified and the life ofthe filter media will be reduced. In a chemical
scrubber, the consumption of oxidising agent will considerably increase in case of
high influent concentrations. Under such circumstances carbon consumption in an
activated carbon plant will also be high.
Many years of experience with the bioscrubber has been gained in the fish
industry and animal feedstuff industry, and it has proved possible to remove almost
ali nitrogenous compounds from the air and to reduce odour emissions by at least
95%.
Possible improvements of the bioscrubber for waste water treatment plant
applications could be:
• Improved absorption efficiency for organic sulphur compounds in the scrubber

section, for instance by means of addition of pul veri sed activated carbon or other
absorption-enhancing agent to the scrubber liquid;
• Improved capacity of the absorption section for handling inlet concentrations of
hydrogen sulphide higher than approximately 75 mg/m3 . Possible means could be
increasing the buffer capacity of the scrubbing liquid, as substantial oxidation
and sulphuric acid production occur in the liquid hold-up of the packing
elements. This causes the pH to drop decreasing the mass-transfer efficiency;
• Examination of causes and cures of foaming in the scrubber liquid. Foaming
seems to occur mainly during start-up and during periods with low inlet
concentrations.
5. Conclusions
1. Covering and enclosing the critical sections such as grit and grease chamber and
screens at the inlet to the Damhusaaen WWTP, together with cleaning of the
ventilation air, have reduced the odour emissions to a level of practically no
noticeable odour nuisance to the treatment plant neighbourhood.
2. Cleaning of the ventilation air takes place in a bioscrubber, which is a process

that combines absorption of odorous substances in a packed column with
degradation in an activated sludge plant.
3. Experience from more than two years of operation has shown that the
bioscrubber can be characterised as follows:
(a) Environmentally friendly process without the use of oxidising chemicals;
(b) High cleaning efficiency with respect to hydrogen sulphide and organic
sulphur compounds;
(c) Low operating costs owed to low consumption of chemicals. Operating costs
are halfto one quarter ofthose of a chemical scrubber at inlet concentrations
in the range of 10-50 mg H2S/m3;
(d) Small space requirement at larger air flows owed to the relatively high
velocity ofthe air through the scrubber;
(e) High reliability as drying out, acidification, clogging and accumulation of
reaction products do not occur;
(i) Maintenance requirements are relatively low, because the process can be
widely automated;
(g) Quick start-up with ordinary activated sludge from a waste water treatment
plant;
(h) On the basis of simple design rules, the process can be designed in an
optimum way for the given amount of air and concentration of pollutants; and
(i) Suited to removal of acidifying substances without any negative effect on the
life of the plant.
4. Therefore the bioscrubber, possibly in combination with a polishing filter, is a

competitive and environmentally friendly alternative to conventional chemi cal or
adsorptive processes. Where reliability, process control, high loading rates and
space requirements are of importance, the bioscrubber should be preferred to the
biofilter.
Acknowledgement
The authors acknowledge the CIWEM organisation.
References
Hansen, N.G. and Rindel, K. Recent experience with biological scrubbers for air pollution control in
Denmark. In: Biotechniques for Air Pollution Abatement and Odour Control Policies. Dragt AJ. and
van Ham J. (eds.). Eisevier, Amsterdam, The Netherlands, 1992, pp. 143-154.
Heist, J.H., Hansen, N.G. and Rasmussen, H.H. Control of odor emissions from wastewater treatment
plants in a bioscrubber. In: Proc. Water Environment Federation 68 1h Annual Conf. & Exposition,
Water Environment Federation, 1995, Alexandria, VA.
Milj0kontrollen. 1998. Odour survey in the surroundings of the Damhusaaen wastewater treatment
plan!. The Environmental Control Body in Copenhagen, (in Danish).
Rasmussen, H.H., Hansen, N.G. and Rindel, K. 1994. Treatment of odorous nitrogen compounds in a
bioscrubber comprising simultaneous nitrification and denitrification. VDI Berichte. 1104: 491-497.
Van Lith, c., Leson, G. and Michelsen, R. 1997. Evaluating design options for biofilters. J. Air Waste
Williams, T.O. Biofiltration for control of odorous emissions and volatile organic compounds from
wastewater and sludge processing facilities. In: Proc. WEF Specialty Conf. on Odor and Volatile
Organic Compound Emission Control for Municipal and Industrial Wastewater Treatment Facilities,
Water Environment Federation, 1994, Alexandria, V A, pp 4.1-4.13.
CHAPTER 14 ODOUR CONTROL AT WASTE WATER
TREATMENT PLANTS BY DIFFUSION
INTO ACTIVATED SLUDGE BASINS
Robert P.G. BOWKER
1. Introduction
Diffusion of odorous air into activated sludge bas ins has been practised for over 30
years as a technique for controlling odours at waste water treatment plants. At least
25 facilities in North America currently use or have used this approach. Table 14.1 is
a list of these plants with information on sources of odorous air, year of installation,
foul air flowrate, and problems experienced.
2. Summary of reported problems
2.1 CORROSION
Concern for cOITosion of activated sludge blowers has generally impeded more
widespread use of the odour diffusion technique. While some plants have reported
cOITosion of blower inlet components such as filters or silencers, documentation of
cOITosion of the internal workings of the blower is very limited. Two plants reported
cOITosion of stee1 inlet guide vanes. In one of those cases, the carbon steel vanes
were replaced with nickel plated vanes with no further problems. At the other
facility, introduction of acidic condensate was suspected as the cause of corrosion as
there was no moisture trap upstream ofthe blower. At a minimum, a demister should
be inc1uded before the blower to remove acidic droplets or condensate formed
upstream in the ductwork. All inlet components should be of corrosion-resistant
material such as 316 SS. Some new facilities have specified nicke1 plated internal
blower components or a corrosion-resistant synthetic resin coating.
2.2 ORGANIC FILM ACCUMULA nON
Many facilities practising odorous air diffusion have reported the accumulation of a
'tar-like substance' or 'greasy film' on the internal components of the blowers. No
information is available on the chemical make up of this substance. It is likely from
grease aerosols that become 'baked' onto the blower components at the e1evated
internal temperatures. Most plants remove this material by steam c1eaning, and one
facility installed steam injection ports on the blower housing. An industrial waste
water treatment plant uses household over c1eaner to remove the film. One plant in
Los Angeles County, California reportedly solved the problem with installation of a
two stage air filtration system.
299
C. Kelliles alld M. C. Veiga (eds.), Bioreactors for Waste Gas Treatment, 299-305.
© 2001 Kluwer Academic Publislzers.
Table 14.1. Summary ofselected facilities in North America using activated sludge diffusion for odour control
"'2"
Treatment plant Plant Source of Yeac Odorous
and location capacity adour controlled installed airflow Comments
ml/d m'jmin
(mgd) (cfm)
Banny Brook WWTP 500 Fennenters for BNR 1994 40 Corrosion cfinlet filters occurred, caus ing damage to blower. New guide vanes are nickel-plated.
CaIgary, AB (130) process (1400)
Kyrene Water Reclamation Aerated equalisation tank 1993 150 FRP fan transfers foul air ta centrifugal blowers for diffusion via jet aerators High H1S ( > 50
Facility (3)
" (5400) ppmv). No corrosion problems reported
Tempe, AZ
Las Coyote WWTP 140 Primary clarifiers 1970 280+ Filters cleaned every 6 months. Blo\\'ers rebalanced and c1eaned every year. No corrosion reported.
Cerritos, CA (37) Influent wet \vell (10000+) Coarse bubble ditTusers. Steel blowers w/coal tar epoxy. SS ducting
~
Long Beach WWTP 95 Primary c1anticr 1973 170+ No filters 011 compressor suction. Have to steam clerul suction and compressor once a ycar Coarse
~
Long Beach. CA (25) (6000+) bubble diffusers: no c1oggmg. Concrete corrosion. 100% removal ofH 2S
0
Hyperion WWTP 1140 Headworks 1959 5100 Have had to c1ean blowers twice since start-up. Fine bubble ditfuscrs. No corroslon rcported. lJ:j
a
Las Angeles, CA (300) Primary c1arificrs ( 180000) Clean plus raul air pass through ane systcm Estimate adaur removal efficiencies 96-99%, VO(· ~
;.;-
DAF thickencrs removal 80%. No Ion ger practlsed after con vers ion ta pure oxygen activated sludge ...,
'"
Emuent pumr statian
LA Glendale WWTP 76 Bar screens 1976 570 Originally coarse bubble dltfusers. From 1979-1991 transitioned to fine bubblc No corrosion or
Las Angeles, CA (20) Influent pumr station (20000) clogging reported. Foul air IS only souree of air ta aeration tanks.
Primaries
Tennmal Island WWTP 80 Headworks 1977 850 Fine bubble difrusers; steel blo\l,'er; PVC piping. No plugging. COlTosion only of concrete in
Las Angeles, CA (21) Primary dariliers (30000) covered tanks.
Centrifuge bldg
Sludge blendl!lg tank
TillmanWWTP 150 Primary c1arilicrs 1984 Fine bubb1c ditfusers. no clogging. No cOlTosion reported
Las Angeles. CA (40) Sase of scrc\\ pumps
Bar screens
PO!TIonaWWTP 49 Primary c1arifiers 1965 170+ Fine bubble diffusers. Chatlge filters quanerly.
Pomona CA (13) (6000+)
Treatment plant Plant Source of Year Odoralls

and location capacity odour control led installed airflow Comrnents
mVd m3/min
(mgd) (cfm)
Annapolis WRF 38 Primary clarifiers, 2001 1850

Annapolis, MD (10) gravlty thickeners, (7000)
belt press
Ventura WWTP 15 Primary clarifier 1972 400 Fine bubble diffusers. c1eaned every 2 years as regular maÎntenance. No corrosion reported
San Buena, CA (4) launders (14000) Blowers sandblasted cvery 2-3 years.
Gravlty thlckener
~
~
DAF thlckener S:?

;::
Whittier Narraws WWTP 57 Primary clarifiers 1962 140+ Clean or replace filters an blower suction quarterly ta annually No corrosÎon reported. Fme bubble ~
Sa EI Monte, CA (15) (5000+) dlffusers, no clogging I
~
San lase Crcck WWTP 240 PrimaT)' daritiers 1971 570+ Recently sWltched from coarse ta fine bubble diffusers. New filter system an blower suctlOn Steel ~
~.
Whittier, CA (62) (20000+) blower. Odours from ~eams of aluminium CQvers ;::,
<ii'
Reedy Creck WWTP 34 Compostmg facility 1988 160 Odaur control blower,> plpmg, and dlfllisers separate rrom mam system aeratlon. Coarsc bubble i::>...
Orlanda. FL (9) (5600) dlffuscrs. No COITOS](lll reported ""
:2"
Visla Royale WWTP 0.8 LiftstatlOl1 1993 No corrosion apparelll after three ycars Positive dlsplacement blower, coarse bubble dlfluscrs ~
~
Vero Beach. FI. (0.2)
§;
Town Branch WWTP 68 Zimpro sy"<tem 1980 42 Odaur control plping ,Uld diffusers separate from aeratian. Coarse bubble diffusers. No filler'i~ ~
Lexington, KY (18) Vacuum lilter bldg (1500) blowers c1eaned ever. 3-6 months. No cOITosion ar pluggmg. Very effective system Zimpro "";::sO·
system no Ion ger In service.
Lawell WWTP 120 Gravity thlckeners 1984 260 Switched from caarsc 10 tine bubble diffusers In 1991. AII SS piping. ~o fillcrs. Same parllculate
Lowell,MA (32) Siudge dc\\atering (9000) build-up al dlffuser hrads
Springfield Regional WWTP 250 In-vesselwmpostlllg NA 130 Interim measure beforc construction ofthennal oxidiser SS coarse bubble diffusers, Impellcrs
Springfield, MA (67) (4500) coated w/s)'nthetic re"ln SS and HDP piping: ilO corrosion or plugging
Scarborough WWTP 6.8 Aerated sludge holdmg 1983 2.8 Foul air is separate s~ . . tem from clean air supply; coarse bubblc diffusers, steci blower Conosion
Scarborough, ME (1.8) tank (100) only an DI intake plpC) and steel inlet silencer v..,
<::>
......
""~
Treatment plant Plant Souree of Year Odorous
and locatien capacity odour control led Înstalled airtlow Comments
ml/d m3/min
(mgd) (cfrn)
Concord WWTP 39 Siudge holding tanks 1994 20 Interim solutian until scrubbers built in 1996. High adour and H2S levels to aeration. Fine bubblc
Concord, NH (102) (700) diffusers performed better than coarse. No corrosion after 2 yr.
Concord WWTP 39 Intluent channels, grit 2000 71 Perfonnance data to be collected in 2001
Concord, NH (102) chamber, primary clarlfier (2500)
launder
Durham WWTP 9.S Siudge holding tanks 1993 37 COITosion of steel înlet filter; replaced with SS. Fine bubble diffusers; no plugging. Foul air is 100%
Durham, NH (2.S) (1300) of aeration air
Roche Chemi cais II Primary clarifier 4S0 Industrial WWTP, Coarse and fine bubbIc diffusers; condensate traps; tar-Iike substance 011 blowcr
Belvedere, NJ (3) (16000)
~
impellers "'"
o
Musonetcong WWTP 9.5 In-vesse! composting 1992 140 Coarse bubble diffusers. 2 m (6 ft) deep strictly for odour control (basins are mechanically aerated)
Stanhope, NJ (2.5) (SOOO) B10wers have syllthetic resin coatmg on impeJlers. heads, and sections
~
~
Newberg,OR 15 In-vesscl composting 1988 68 Oedicated system for diffusion of odours into oxidation ditch; coarse bubbIc ditfusers; 110 corraSlon ~
....
(4) (2400)
Valley Forge WWTP 30 Influent structure 1996 62 Dcdicated fine bubble diffuser system for adour control, Aeration tanks mechanically aerated High
Phoenixville, PA (8) Primary clarifiers (2200) inlet I-{zS. 'Tar' build up caused conventional PO blowers to shut down Intake duct and tilter
modifications soIved problems.
Walnut Creek WWTP 230 Headworks 1977 1060 Fine bubble diffusers Filters on blower intake. No corrosion reported. Fibreglass fans transICr air ta
Austin, TX (60) Primary clarifiers (37500) blower plenum. 'Greasy' film on bIower Iabes removed with steam
Flow equallsation
Sewer tUHnei
Trinity River Central WWTP 615 Raw sewage PS 1997 510 Fine bubble dîffusers Filters on intakes of centnfugal blowers. Now used as back up to biofilter
Headworb. Pri, clar systern
Dallas, TX (162) (18100)
efflllent channel
Gravity th ickeners
Sludge holding tanks
Arlington Ca. WPCP 110 Headworks 1995 570 SteeI blowers, coarse bubble diffllsers, No corrosion problems. Regular inspection program jor
Alexandria, VA (30) (20000) blower alignment, oii change.
Case Study ~ Activated sludKe diffusion 303
3. Con cord, New Hampshire, case study
The waste water treatment plant in Concord, New Hampshire used activated sludge
diffusion as an interim measure to control odours from a sludge holding tank. The air
from the headspaee of the holding tank was pulled through an existing positive
displacement blower and diffused into a meehanieally aerated activated sludge tank
at approximately 37 m 3/min. Coarse bubble diffusers were first tested. As shown in
Table 14.2, removal of odour (as measured in dilutions to threshold or odour units)
was over 96 pereent, but odour conccntrations at the surface of the aeration tank were
almost 50 times background levels. Hydrogen sulphide concentrations were also
elevated well above background concentrations. The same test was conducted using
tubular, flexible membrane, fine bubble diffusers. Odour and H2 S removals improved
dramatically sueh odour and H2 S concentrations at the surface were reduced to the
background leve1s measured when no foul air was introduced.
Table 14.2. Performance of activated sludge diffusion for odour control at Concord, New
Hampshire, USA
Odour coneentration
HzS, llllmv
O/T,O.U.
Odour Souree
Back- Back-
In Out In Out
Ground Ground
Sludge holding tank
Coarse bubble 24150 850 10 ~100 7.8 0.2
Fine bubble 39000 18 14 ~100 0.3 0.2
In the year 2000 the Concord waste water treatment plant replaeed the
existing mechanieal aerators with fine bubble diffusers and used the new blower to
diffuse foul air. The blower receives 100 pereent foul air from the influent channels.
aerated grit chamber, and primary clari fier effluent launders. The system is etIective
and no problems have been reported. Since the plant uses biotowers upstream of the
aeration tanks for BOD reduction, the demand for oxygen in the aeration tanks is
relatively low. As a result blower air flow rates have only been 260 to 530 m 3/min,
less than the design foul air flow of 660 m 3/min based on a ventilation rate of 6 air
changes per hour from the covered odour sourees. Performanee testing of the odour
diffusion system is scheduled for the summer of 200 1.
4. Valley Forge, Pennsylvania, case study
The 30 MLld Valley Forge plant had experienced high odour emissions from the
turbulent influent chamber, primary clarifier feed wells and effluent launders, and
primary effluent distribution ehamber. Hydrogen sulphide concentrations at turbulent
surfaees were 50 to 500 ppmv in the air. The sewer authority chose activated sludge
diffusion over chemi cal scrubbers or other odour treatment alternatives. even though
the activated sludge tanks were meehanically aerated. A dedicated system of blowers
and diffusers was installed in 1996 to treat 62 m 3/min of odoraus air. Design criteria
are summarised in Table 14.3.
304 R.P.G. Bowker
Table 14.3. Key design criteria for odour control system at Valley Forge WWTP
Parameter Value
Odour sources Influent chamber
Primary claritier feed wells
Primary claritier effluent launders
Primary effluent splitter box
Air flow 62 m3/min (2200 cfm)
Air exchange rate below covers 12AC/hr
Inlet H2S (estimated) 120 ppmv (summer)
Blowers 2--45 kW (60 hp) positive displacement
Diffusers 394 tubular, flexible membrane type at
4.3 m (14 fi) depth
Materials of construction Covers FRP
Ductwork PVC, 316 SS
Blower tilters, silencers 316 SS
Blowers steel
Discharge piping 316 SS
Concrete protection Epoxy coating above water line
Performance data are shown in Table 14.4. The system was effective in
reducing odours and H2S levels at the surface of the aeration tanks to background
levels. During the tirst year, problems were experienced with accumulation of
organic matter in the positive displacement blowers that caused the blowers to shut
down. This was attributed to suction piping being too close to the surface of the
covered primary claritier feed well, causing grease to be pulled into the blower, and
to an inefticient tilter system. Afier these problems were corrected, the system has
provided reliable service. However, the epoxy coating used to protect the concrete in
the covered tanks failed in the presence of sulphuric acid.
Table 14.4. Performance of activated sludge diffusion for odour control at Valley Forge
WWTP, Pennsylvania, USA
Odour concentration
HzS, ppmv
D/T.O.U.
Odour Sources
Back- Back-
In Out In Out
Ground Ground
Influent chamber, primary 19000 5 5 77 0.1 0.05
claritier feed wells and 7 30 0.1 0.1
effluent launders,
distribution box
Case study - Activated sludge difJusion 305
Power costs for a dedicated odour diffusion system are high, and may make
use of a dedicated blower and diffuser system uneconomical compared to other
alternatives. According to Valley Forge staff, however, mechanical aerators that are
elosest to the foul air diffusers are now operated at low speed, and no net increase in
power consumption has been experienced.
5. Annapolis, Maryland, case study
Prior to implementing a full-scale odour control system using activated sludge

diffusion, the Annapolis Water Reelamation Facility conducted pilot testing using a
small blower and diffuser system to treat 8 m3/min of air collected from a primary
sludge distribution chamber. A 0.9 m x 2.1 m floating polycarbonate chamber was
used to isolate the surface above the diffusers to facilitate collection of samples and
monitoring of hydrogen sulphide levels. Air flowrates through each diffuser were
maintained at the maximum rate recommended by the manufacturer. Results of the
testing are shown in Table 14.5.
Table 14.5. Removal of reduced sulphur compoullds alld odour by activated sludge diffusion,
Ailnapolis WRF pilot study
Avg. inlet Avg.outlet Removal Avg.

Compound
(ppbv) (ppbv) (%) background
Hydrogen sulphide 68.2 <4 > 94 <4
(no supplemental H2S)
Hydrogen sulphide 14350 19.8 99.9 <4
(with supplementa1 H2 S)
Methyl mercaptan 144 5.9 95.9 <4
Dimethyl sulphide 82.1 <5 > 94 <4
Isopropyl mercaptan 44.3 <4 > 91 <4
Odour concentration 1350o.u. 1500.u 133 o.U.
Removal of reduced sulphur compounds and odour was excellent at the

loadings experienced. Odour concentrations at the surface of the aeration tank were
not significantly different from background concentrations when elean air was
introduced.
A full-scale system to treat 1850 m3/min of foul air from the primary
elarifiers, gravity thickeners, and belt press room will likely go into operation in late
2001. As a precautionary measure, the existing centrifugal blowers will be shipped
back to the manufacturer for application of a corrosion-resistant synthetic resin
coating on the internal components exposed to H2 S.
INDEX
A c1osed, 49,261-267
Absorption, 17,24-27,86, 134, 152 rotating, 202
Acetobacter sp., 246 Biomass retention time, see Siudge
Actinomycetes, 66 retention time
Activated carbon, 32-35, 62,67,88- Bioreactor,4, 17, 140-144, 152
90,248 Bioscrubber,5-7, 133-162,215,236
Activated sludge, 9, 115, 140,215-254 Biotrickling filter, 5-7, 99-131, 206-
Adsorption, 17,32-37,51,86,100, 207,246
115,218,233,261-263,282 Board industry, 7
Column,25 BoIton criteria, 184
Isotherm,34-37 Breakthrough curve, 36
Aeration, 140, 143, 153 Breweries, 278
Aerosol, 7,9, 133 BTEX, 4,88,186,219-220,228-229,
Air 233,240-241
Composition, 3 Bubble column, 219-221, 223-224,
Velocity, 50-51,136, see a/sa gas 242,245-246
velocity Burkholderia sp., see Pseudomonas
Airlift, 154,222-224,233,243-249 sp.
Alcaligenes sp., 70, 89 Butanol, Il, 88
A1gae, 73, 116 Butene, 185
A1kylbenzenes: 55,69-71, 78, 81-82
A1pha(a),218-219 C
A1pha-pinene, see Pinene Carrier
Ammonia, 3, 7-8, 73,75,82,88, 107, organic, 63
144-148,155,168,170,181,233, inorganic, 67, 73, 77-79
257-258,271,276,282 synthetic, 114
Anaerobic, 6, 79, 114, 147, 150-154, Catalyst, 39, 180
179,237-238,273,277-278 Celite, 61, Il O, 245
Andrews equation, 55, 60, 228 Cell
Animal rendering, 7, 148 composition, 74-75,142
Arrhenius equation, 112 counts, 66, 73-74
Arthrabacter sp., 90 Cellulose, 166
Centrifugal force, 19-20
B Ceramic, 26, 66, 78-79,89,224
Bacillus sp., 70,90 Channelling, 66, 68, 81, 83, 260
Backwashing, 124-125 Chemical industry, 7
Baghouse, 20 Chemical Oxygen Demand, 142, 143,
Bark, 64, 75, 83, 85, 88 146,191,239
Basidyomycetes, 69 Chlorinated compounds, 44, 186, 190,
BET equation, 36 194
Biofilm Chlarobium sp., 237
model, 57, 172-173 Clogging, 41, 82-84,100, 124-126,
thickness, 76, 121, 172,201 142, 145,202,217,235,283
Biofilter, 2, 5-7, 47-98,100,223,236 Coconut fibre, 258
open,49, 77,104,257-260 COD, see Chemical Oxygen Demand
308 Index
Collection efficiency, 18-19,23-34 Einstein, 184-185

Colony fonning unit (CFU), 74 Electron beam discharge, 189-190
Combustion processes, 4-5 Electrostatic precipitator, 22, 187
Cometabolism, 107, 17L 244-245 Elimination capacity, 52-54, 102-
Compaction, 67-69, 260 103, 204
Compost, 63-67, 73-75, 79, 81-85, Elimination efficiency, 139
88-90 Empty Bed Retention Time, 49-50,
Composting, 6, 7, 234 102,223
Concrete, 104,217,258 Equilibrium curve, 30
Condensation, 17,42-44, 87 Ergun equation, 82
Condenser Ethanol, 62,70,88,112,121,148,
direct-contact, 42 154, 183,233,239,246
surface,42 Ethylacetate, 51, 88, 183
Corona, 23,186-190 Ethylene, 182
Corynebacterium sp., 70 Ethyl-tert-butyl-ether (ETBE), 5
Co-solvent, 219, 248-249 European Reference Odour Mass
Costs, 84-87,100,156-158,217,232, (EROM), II
246-248,262,283 Exophiala sp., 71-72, 89
Creso!, 88, 186-187,233,241 Exopolymers, 51
Cyclone,19
-scrubber, 24 F
Fabric filter, 20
D Fatty acids, 6, 134, 158,257,277-278
Dairy industry, 257-260 Feed composition, 62-63
Darcy equation, 20 Fertiliser, 100, 260
Demister, 105, 135, 148,258,263, Fibreglass, 7, 104-105,258
288 Filter bed, see Carrier
Design, 26-27,42,47,140,258 Filtration, 20
Desorption, 33 Fish feed production, 7
Desulfotomaculum sp., 239 Flow rate
Desulfovibrio sp., 239 gas,28,49, 62,136-137,230
Deutsch equation, 23 liquid, 28,137,143,145
Diatomaceous earth, 89 water, see liquid
Dichloromethane, 70, 109, 167-168, Fluorescence in situ hybridization
203-208,233,242-244,246,248-249 (FISH),73
Diffusion limitation, 59 Foams, 151
Diffusivity, 225,232 Food factory, see Food industry
Dissolved oxygen (00),218,229,289 Food industry, 7, 190,257-260,271
DNA finger printing, 117 Formaldehyde, 52,134,144-145,
Downtlow, 48, 81 148,181,183
Drag coefficient (C o), 18 Foundry, 7,145
Drainage water, see Drain water Fourier Transform Infrared
Drain water, 67, 79, 85,143,146,191 Spectrometry (FTIR), 264
Dust, 9, 48,100,133,135,145,148, Freundlich isotherm, 35
260 Fugitive emissions, 5, 87
Full-scale, 87, 150,222,232-235,264
E Fume, 9,135
EBRT, see Empty Bed Retention Time Fungi, 65, 69-74, 79, 81, 124
Index 309
G Liquid,
Gas hold-up, 223 hold-up, 103
Gravity settling chamber, 18 velocity, 119-120, 137
Groundwater poliution, 6 Lithographic operations, 8
Growthrate(~),55-56, 120, 141, 173 Livestock, 156
Load celi, 79,263
H Loading rate, 51-53, 136,224
Haldane equation, 60
Hardboard manufacturer, 7 M
Heather,85 Maintenance, 120, 124, 141
Henry's law coefficient, 51,57-58, Mass transfer, 28-32, 122, 137-138,
107,109, 112, 133, 137-139, 170, 173-175,205-206,225-227
219-220,226,230 zone(~Z),36-37
H2S, 63, 80-81, 89, 99-100, III, 115, Membranes, 44, 166-168
134,147,151-154,157-158,232- Membrane reactor, 163-177
234,237-239,247,257,271,274, Mercury lamp, 180
271-279,281 Methanol, 51, 70, 75, 89,109, 112,
HTU, 31, 138-139 154,171,190,219,233,242,261
Humic acid, 65 Methyl-ethyl-ketone (MEK), 89, 148,
Humidification, 48, 77-78, 80 187
Hydrogen sulfide, see H2S Methyl-iso-butyl-ketone (MIBK): 89,
Hyphomicrobium sp., 88,242,246 144
Methyl-tert-butyl-ether (MTBE), 5,
1 107, 1I5-116, 163, 170
Incineration, 17,37-41,86-87,100, Methylosinus sp., 245
157, 188,261-262 Michaelis-Menten,
Inhibition constant, see K1 constant (K.o), 142
Inoculation, 68-71, 115, 142,259 equation, see Monod equation
Mist eliminator, see Demister
K Mobile sources, 2-3
K1, 55, 60, 120,228 Modeliing, 49, 54-62, 172-174,225-
Kinetics, 55-60, 120, 137, 183,227- 232
229 Moisture, see Water content
Kt, see kL Monod equation, 55-57, 60,120,141,
Kt,29-30, 138, 173,225 173,227-230
KL~ 122, 130,210-211,218-221, Mould,81
224-227 Mycohacterium sp., 88
Ks, 55-56,141,173,228,231
N
L Naphthalene, 230-231
Lacquering industry, 7, 9 Nitric oxides (NOx), 2, 8, 39, 65, 15 L
Lactic acid, 257 155, 157-158, 182,238
Langmuir isotherm, 36 Nitrogen compounds, 269
Lava rock, 99, 103, 111, 114 Non-steady state, 61, 123
Leather industry, 7 NTU, 31,138-139
Lift stations, 271 Nutrients, 63-68, 74-76,117-119,
Lignoceliulosic material, 69,72 142-144, 146,260
310 Index
o 203, 234
Odour,7-9,232-236,257,282 Polypropylene (PP), 104-105, 137,
definition, 10 166--170
Olfactometer, 12 Polystyrene, 64-66, 74, 83,90
Operating line, 28 Polysulfone, 167-168
Overall heat transfer coefficient (U), Polytetrafluoroethylene (PTFE), 166
43 Polyurethane foam, 64, 66, 115
Oxidation, see Incineration Polyvinyl chloride (PVC), 104-105,
Oxygen limitation, 76, 113, 128 136,202,234
Oyster shells, 82, 245 Porosity, 27, 32-33, 49-50, 65, 114-
115,169-170
p Power, 154, 187, 189-190, 22()""'221,
Packed tower, 26-28, 135 232,247,282
Packing, see Carrier plant, 154
Paint booth, 144 Predation, 102, 1l8, 121, 124
PalI rings, 26, 64, 72, 110, 114, 135 Pressure drop, 18-20, 26-27, 82-84,
Paracoccus sp., 239 100, 136, 163
Parts per million Pre-treatment, 10, 17,48,80,146,
definition, Il 179,258
ParticIes, 2-3, 7-8, 17-25,48,99, Process control, 105-106,276--277
133,189,217,247,258,262 Propene, 168, 170, 174
P articulates, see P articI es Protozoa, 65, 73,101,116--117,124
Peat; 64--68, 78-79, 88-90, 258 Pseudomonas sp., 55, 70, 89-90, 110,
Perchloroethylene (PCE), 69, 151, 1l7, 123, 151,240,244-245
182-185,187 Pu1p and paper industry, 8
Perlite, 53, 61, 64, 67, 72, 78-79, 83, Pulse Jet collector, 22
88-90 PVC, see Polyvinyl chloride
Permeabi\ity constant (K), 21
Phanerochaete sp., 69, 72 R
pH, 81-82, 105, III, 119, 133,258, Raschig rings, 26, 135
275-276 Rate equation, 54, 60
Phosgene, 38, 182 Rayon manufactures, 9, 281
Phospholipid Fatty Acid Analysis RBC, see Rotating Biological
(PLFA),73 Contactor
Photocatalytic oxidation, 181-184 Reaction
Photochemical, 10, 44 \imitation, 59-60
Photolysis, 10, 185-186 rate, 57-60
Pinene, 51,69-70, 72, 79, 89, 112 Recirculation, see Recycling
Plasma, 179, 186, 188 Recycling, 119-120, 146, 169,263
Plate column, 31 Refinery, 151
Plate counts, see Cell counts Refuse dump, 282
Plug flow, 57, 62 Removal efficiency, 53, 102, 204
Pollutants Rendering industry, 8
anthropogenic, 4 Resin production, 8
natural,4 Retardation factor, 50..... 51
Pollution Reverse f10w cIeanmg, 21 ..... 22
sources, 4-6 Reynolds number (Re), 18,208,211
Polyethylene (PE), 104-105, 137, 168. Rhodococcus sp., 70. 90, 242
Index 311
RNA, T
16S-rRNA, 73,117 Tabaco industry, 8
Rotating Biological Contactor, 5-7, Teflon, 166
201-214 Temperature, 32-33, 38-45, 80- 81,
Rotarional speed, 205-207, 210-211 112
Terminal settling velocity, 18
S Tetrachloroethylene, see
Sand, 64, 66 Perchloroethylene
Scanning confocallaser microscopy Thermophilic, 80-81, 112, 148, 154
(SCLM),122 Thiele modulus, 59-60
Scrubber, 23-25,111,133,135-136, Thiobacillus sp., 70,81,89, 106, III,
258,282,291 151,153,238-239,274,292
SEC, see Surface-specific elimination Ti0 2, 180-182, 184, 186, 191
capacity Toluene, 30-40, 51, 55, 63, 67, 72,
Shaker,21 75-76,81,88,90,107-110,121-124,
Silica gel, 33 144,148-149,151,167-169,190,
Site remediation, 4 212,220,226,231,233,241,243
SLO, see Surface-specific substrate Toxic, 5, 62,124, 141-142
load Trametes versicolor, 69
Sludge retention rime (SRT), 218, 221, Trich1oroethylene (TCE), 38, 107,
230,235,241 151,167,182,187,190,229,233,
Soil, 66, 79, 87-88 243-246,261
Soil Vapor Extracrion (SVE), 6 Trickling
Sonolysis, 191-192 density, see rate
Sparged suspended growth bioreactor, filter,47
219-222,224 rate, 102, 120, 136
Specific oxygen transfer efficiency Triethylamine, 90
(SOTE),217
Specific surface area, 68,135-136, U
163 UltraViolet (UV), 44, 136, 180, 185-
Sponge manufactures, 6, 278 186,190
Spray tower, 24, 135 Uplow Anaerobic Sludge Bed reactor
Start-up, 68-69,115-116,259,292- (UASB),151
293
Stationnary sources, 3 v
Storage tanks, 8 Vapour
Stripping factor, 139 concentration, II
Styrene, 69, 71-72, 77,89,109, 147, definition, 10
149,192-193 Vegetable oii production, 8
Sulphur compounds, 38, 69, 109, 147- Vehicles,4-5
148,257-258,269 Venturi scrubber, 24, 135
Superficial Vermiculite, 64, 66, 72
gas velocity, 51, 58, 83, 223 Virus, 65
liquid velocity, 119-120 Viscose industry, 271
Supply rate, 118 Void fraction, 26, 65
S urface loading rate, 51
Surface-specific elimination capacitv, W
204,207 Waste water, 26, 42, 68, 100, 165,281
312 Index
Waste water treatment, 6, 9, 54, 87, y

lil, 121, 140-144, 146-148,202,279 Yeast, 65, 69, 71, 73,75,79,81,116
Water activity, 76-77, 79 Yield
Water content, 76--79 biomass (Y), 54-55, 75, 118-120,
Wood chips, 64, 74, 79, 89 123-124, 141,228
Woodindustries, 7,9,145,148
Z
X Zeolite,33
X-rays, 189

Christian Kennes, MarÃ A C. Veiga (Auth.), C. Kennes, M. C. Veiga (Eds.) Bioreactors For Waste Gas Treatment 2001 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Christian Kennes, MarÃ A C. Veiga (Auth.), C. Kennes, M. C. Veiga (Eds.) Bioreactors For Waste Gas Treatment 2001 PDF

Uploaded by

Copyright:

Available Formats

Bioreactors for Waste Gas Treatment

SPRINGER-SCIENCE+BUSINESS MEDIA, B.V.

ISBN 978-90-481-5772-3 ISBN 978-94-017-0930-9 (eBook)

Printed on acid-free paper

AII Rights Reserved

Chapter 1 Fundamentals of air pollution 3

3.5. Water content 76

5.3. Anaerobic bioscrubbers 150

2. Experience with the RBC for waste gas treatment 203

PART2 APPLICA TIONS 255

Chapter 10 Biofiltration ofwaste gases from a dairy industry 257

3. Biofilter design 262

and post-graduate courses related to environrnental engineering and technology, and

Angela R. Bielefeldt, University of Colorado, Department of Civil, Environmental

Huub H.J. Cox, University of California, Department of Chemical and

Mare A. Deshusses, University of California, Departrnent of Chemi cal and

Sarina J. Ergas, University of Massachusetts, Department of Civil and

Johan W. van Groenestijn, Institute of Environmental Sciences, Energy Research

Christian Kennes, University of La Corufia, Department of Chemical Engineering,

MiehaeI S. MeGrath, Monsanto Enviro-Chem Systems Inc., S1. Louis, MO.

Jan-Carel Nieuwland, Monsanto Europe S.A./N.V., Avenue de Tervuren 270-272,

Osear Prado, University of La Coruna, Department of Chemical Engineering,

Kim Rindel, Lynettefaellesskabet IlS, 250 Refshalevej, DK-1432 Copenhagen.

Patrik Ruediger, ETH Centre, Institute of Process Engineering, CH-8092 Zurich.

Maria C. Veiga, University of La Corufia, Department of Chemica1 Engineering,

Christian KENNES and Maria C VEIGA

Table 1.1. Composition of elean dry air

Compound % volume (or 10 4 ppmv)

A compound is considered to be a contaminant when it is present in the air at

C. Kennes and M. C. Veiga (eds.), BioreactorsJor Waste Cas Treatmel1t, 3-15.

2. Sources of air pollution

2.1 MOBILE SOURCES

2.2 STATIONARY SOURCES

2.2.1. Combustion processes

2.2.2. Industrial waste gases

2.2.3. Fugitive emissions

2.2.4. Site remedia/ion

2.2.5. Waste water treatment

Sources Typical pollutants Treatment References

Table 1.2. (Continued)

Landfill gas Hydrogen sulphide Biofilter Sabo,1989

Table 1.2. (Continued)

Viscose rayon H2S, CS 2 Biotrickling filter Chapter 12

3. Types of air pollutants

3.2 V APOURS AND ODOURS

4.1 PARTICLES AND VAPOUR CONCENTRATIONS

While the concentration of particulate matter in air is generally expressed in g/m3 ,

At25°C: glm3 = (MW/24.5) ppmv (1.1)

At OaC: glm3 = (MW/22.4) ppmv (1.2)

where P is the pressure in atmospheres and T is the temperature in Kelvin.

Dynamic olfactometry is a normali sed method allowing measurment of odour

Demiriz, A.M. Neue Einsatzgebiete biologischer Filteranlagen: Giei3erei-Bereich und

Eitner, D. Emissionsminderung in Olmiihlen durch Biofilter - Erfahrungsbericht. In: Biotechniques

Hofmann, W. 1989. Biofilter nach Kakaorăstereien. VOI Berichte. 735: 233~242.

Huber, J. Planung, Durchftihrung und erste Erfahrung zum Biofilter Tierkorperbeseitigungsanlage

Kersting, U. Behandlung gror3volumiger Abluftstrome durch Biofilter, vorgestellt an Beispielen der

Lehtomăki, J., Torronen, M. and Laukkarinen, A. 1\ feasibility study of biological waste-air

Maier, G. 1989. Biofiltration von Gier3ereiabgassen. VOI Berichte. 735: 285~292.

Poulopoulos, S. and Philippopoulos, C. 2000. lnfluence of MTBE addition into gasoline on

Schmidt, M. 1993. Abluftreinigungsanlage mit Biofiltern. Chemie-Umwelt-Tech. 22: 74-75.

Christian KENNES, Maria C. VEIGA and Oscar PRADO

In order to make an adequate choice of the most suitable treatment technology, it is

2. Removal of particulate matter

N A = KG (PG - p) = K L (C - C L ) (2.16)

or NA = Ky (ye; - y) = Kx (x - XL) (2.17)