Elefteria Psillakis, Vassilis Gekas (Auth.), Xavier Nicolay (Eds.) - Odors in The Food Industry-Springer US (2006)

Integrating Safety and Environment Knowledge
Into Food Studies

Xavier Nicolay towards European Sustainable Development
Odors in
the Food
Industry
Odors in the Food Industry
ISEKI-FOOD SERIES
Series Editor: Kristberg Kristbergsson, University of Iceland
Reykjavík, Iceland
Volume 1 FOOD SAFETY: A Practical and Case Study Approach

Edited by Anna McElhatton and Richard Marshall
Volume 2 ODORS IN THE FOOD INDUSTRY

Edited by Xavier Nicolay
Volume 3 UTILIZATION OF BY-PRODUCTS AND TREATMENT OF

WASTE IN THE FOOD INDUSTRY
Edited by Vasso Oreopoulou and Winfried Russ
Volume 4 PREDICTIVE MODELING AND RISK ASSESSMENT

Edited by Rui Costa and Kristberg Kristbergsson
Volume 5 EXPERIMENTS IN UNIT OPERATIONS AND PROCESSING

OF FOODS
Edited by Maria Margarida Cortez Vieira and Peter Ho
Volume 6 CASE STUDIES IN FOOD SAFETY AND ENVIRONMENTAL

HEALTH
Edited by Maria Margarida Cortez Vieira and Peter Ho
Odors in the Food
Industry
Edited by
Xavier Nicolay
Institut Meurice
Brussels, Belgium
Xavier Nicolay
Ingénieur
Service de Génie Chimique & Biochimique
Instiut Meurice – HELdB
1, avenue Emile Gryzon, bât. 2
1070 Bruxelles
Belgium
xnicolay@ceria.be
Series Editor
Kristberg Kristbergsson
Professor of Food Science
Dept. Food Science and Human Nutrition
University of Iceland
Hjardarhaga 2-6
107, Reykjavik
Iceland
Library of Congress Control Number: 2006926455
ISBN-10: 0-387-33510-2
ISBN-13: 978-0387-33510-0
e-ISBN-10: 0-387-34124-2
e-ISBN-13: 978-0387-34124-8
©2006 Springer Science+Business Media, LLC

All rights reserved. This work may not be translated or copied in whole or in part without the written
permission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York,
NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in
connection with any form of information storage and retrieval, electronic adaptation, computer
software, or by similar or dissimilar methodology now known or hereafter developed is forbidden.
The use in this publication of trade names, trademarks, service marks, and similar terms, even if they
are not identified as such, is not to be taken as an expression of opinion as to whether or not they
are subject to proprietary rights.
9 8 7 6 5 4 3 2 1
springer.com
SERIES ACKNOWLEDGMENTS
ISEKI Food is a thematic network on food studies, funded by the European Union as
project N° 55792-CP-3-00-1-FR-ERASMUS-ETN. It is a part of the EU program in
the field of higher education called ERASMUS, which is the higher education action
of SOCRATES II program of the EU.
v
SERIES PREFACE
The single most important task of food scientists and the food industry as a whole
is to ensure the safety of foods supplied to consumers. Recent trends in global food
production, distribution, and preparation call for increased emphasis on hygienic
practices at all levels and for increased research in food safety in order to ensure a
safer global food supply. The ISEKI-Food Series is a collection of books where vari-
ous aspects of food safety and environmental issues are introduced and reviewed by
scientists specializing in the field. In all of the books a special emphasis was placed
on including case studies applicable to each specific topic. The books are intended
for graduate students and senior level undergraduate students as well as profession-
als and researchers interested in food safety and environmental issues applicable to
food safety.
The idea and planning of the books originates from two working groups in the
European thematic network “ISEKI-Food,” an acronym for “Integrating Safety and
Environmental Knowledge In Food Studies.” Participants in the ISEKI-Food net-
work come from 29 countries in Europe and most of the institutes and universities
involved with food science education at the university level are represented. Some
international companies and nonteaching institutions have also participated in the
program. The ISEKI-Food network is coordinated by Professor Cristina Silva at
The Catholic University of Portugal, College of Biotechnology (Escola) in Porto.
The program has a website at: http://www. esb.ucp.pt/iseki/. The main objectives of
ISEKI-Food have been to improve the harmonization of studies in food science and
engineering in Europe and to develop and adapt food science curricula, emphasizing
the inclusion of safety and environmental topics. The ISEKI-Food network started
on October 1st in 2002, and recently has been approved for funding by the European
Union for renewal as ISEKI-Food 2 for another three years. ISEKI has its roots in
an EU-funded network formed in 1998 called Food Net where the emphasis was on
casting a light on the different food science programs available at the various univer-
sities and technical institutions throughout Europe. The work of the ISEKI-Food
network was organized into five different working groups with specific task, all aiming
to fulfill the main objectives of the network.
The first four volumes in the ISEKI-Food book series come from WG2 coordi-
nated by Gerhard Schleining at Boku University in Austria and the undersigned.
The main task of the WG2 was to develop and collect materials and methods for
teaching safety and environmental topics in the food science and engineering cur-
ricula. The first volume is devoted to food safety in general with a practical and
case study approach. The book is composed of 14 chapters that were organized
into three sections on preservation and protection, benefits and risk of microorgan-
isms, and process safety. All these issues have received high public interest in recent
years and will continue to be in the focus of consumers and regulatory personnel
for years to come. The second volume in the series is devoted to the control of air
vii
viii Series Preface
pollution and treatment of odors in the food industry. The book is divided into eight
chapters devoted to defining the problem, recent advances in analysis, and methods
for prevention and treatment of odors. The topic should be of special interest to
industry personnel and researchers due to recent and upcoming regulations by the
EU on air pollution from food processes. Other countries will likely follow suit with
stricter regulations on the level of odors permitted to enter the environment from
food-processing operations. The third volume in the series is devoted to utilization
and treatment of waste in the food industry. Emphasis is placed on sustainability of
food sources and how waste can be turned into by-products rather than pollution
or landfills. The book is composed of 15 chapters starting off with an introduction
of problems related to the treatment of waste and an introduction to the ISO 14001
standard used for improving and maintaining environmental management systems.
The book then continues to describe the treatment and utilization of both liquid
and solid waste with case studies from many different food processes. The last book
from WG2 is on predictive modeling and risk assessment in food products and pro-
cesses. Mathematical modeling of heat and mass transfer as well as reaction kinetics
is introduced. This is followed by a discussion of the stoichiometry of migration in
food packaging, as well as the fate of antibiotics and environmental pollutants in the
food chain using mathematical modeling and case study samples for clarification.
Volumes five and six come from work in WG5 coordinated by Margarida Vieira at
the University of Algarve in Portugal and Roland Verhé at Gent University in Bel-
gium. The main objective of the group was to collect and develop materials for teach-
ing food safety-related topics at the laboratory and pilot plant level using practical
experimentation. Volume five is a practical guide to experiments in unit operations
and processing of foods. It is composed of 20 concise chapters each describing differ-
ent food-processing experiments outlining theory, equipment, procedures, applicable
calculations, and questions for the students or trainees followed by references. The
book is intended to be a practical guide for the teaching of food-processing and engi-
neering principles. The final volume in the ISEKI-Food book series is a collection of
case studies in food safety and environmental health. It is intended to be a reference
for introducing case studies into traditional lecture-based safety courses as well as
being a basis for problem-based learning. The book consists of 13 chapters contain-
ing case studies that may be used, individually or in a series, to discuss a range of
food safety issues. For convenience the book was divided into three main sections
on microbial food safety, chemical residues and contaminants, and a final section on
risk assessment and food legislation.
The ISEKI-Food book series draws on expertise from close to a hundred universi-
ties and research institutions all over Europe. It is the hope of the authors, editors,
coordinators, and participants in the ISEKI network that the books will be useful to
students and colleagues to further their understanding of food safety and environ-
mental issues.
March 2006 Kristberg Kristbergsson

PREFACE
Olfactory nuisances, to the same extent as noise pollution, appear to be a modern

social problem that is drawing increasingly more attention and is closely associated
with urbanization, industrialization, and overall population density.
This issue represents a major concern for decisionmakers in industry as well as for
regional or national authorities. Air pollution and odor problems involve the respon-
sibility of industrial plant managers not only in terms of brand image, but also in
terms of possible economic consequences due to the implementation of treatment
processes, or in the absence of adapted precautions, legal actions from neighboring
industries or persons.
Authorities also are concerned by this public health problem as they are in charge
of implementing regulations and planning the national and regional development
which will directly affect the socio economic situation of the region or country.
This book was edited within the framework of the ISEKI Thematic Network
and was originally intended for students and teachers in food engineering. In this
context, its primary aim is to raise future industrial decisionmakers’ awareness in
odors and air pollution problems. However, this book also offers to representatives
of industry the appropriate tools to better apprehend and manage odors problems,
should they ever be confronted with this type of situation whether as a victim or as
a liable actor.
In particular, this book deals with the diverse questions raised by odors in the food
industry and the closely related volatile organic compounds: ranging from percep-
tion of the issue to implementation of regulations, from prevention of the problems
to their possible treatment, through specific case studies and analysis methods illus-
trating the different measurement technologies.
This exhaustive approach was made possible thanks to the extraordinary extent
of the ISEKI Thematic Network and its privileged relationships among multidisci-
plinary experts from universities in 29 countries across Europe.
At this stage I would like to thank all contributors for their efficient and enthusi-
astic collaboration.
January 2006 Xavier Nicolay
ix
CONTENTS
CONTRIBUTORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
1. ODOR PROBLEMS IN THE FOOD INDUSTRY . . . . . . . . . . . . . . . . . . . 1

Elefteria Psillakis and Vassilis Gekas
2. ODOR MEASUREMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Elefteria Psillakis
3. PRECONCENTRATION PRIOR TO GAS CHROMATOGRAPHY . . . . 41

Elefteria Psillakis
4. THE APPLICATION OF INTELLIGENT SENSOR ARRAY FOR AIR

POLLUTION CONTROL IN THE FOOD INDUSTRY . . . . . . . . . . . . . . . 47
Saverio Mannino, Simona Benedetti, Susanna Buratti,
and Maria Stella Cosio
5. ELECTRONIC-NOSE TECHNOLOGY: APPLICATION FOR

QUALITY EVALUATION IN THE FISH INDUSTRY . . . . . . . . . . . . . . . . 57
Guðrún Ólafsdóttir and Kristberg Kristbergsson
6. ODORS PREVENTION IN THE FOOD INDUSTRY . . . . . . . . . . . . . . . . 75

Regina Nabais
7. ODORS TREATMENT: PHYSICOCHEMICAL TECHNOLOGIES. . . . 105

Regina Nabais
8. ODORS TREATMENT: BIOLOGICAL TECHNOLOGIES . . . . . . . . . . . 125

Bram Sercu, João Peixoto, Kristof Demeestere, Toon van Elst,
and Herman Van Langenhove
INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
xi
CONTRIBUTORS
Simona Benedetti
Department of Food Science, Technology and Microbiology, University of
Milan, Via Celoria 2, 20133 Milan, Italy
Susanna Buratti
Maria Stella Cosio

Kristof Demeestere
EnVOC Research Group, Department of Organic Chemistry, Faculty of
Bioscience Engineering, Ghent University, 9000 Gent, Belgium
Toon van Elst

Project Research Gent nv, 9030 Gent, Belgium
Vassilis Gekas
Department of Environmental Engineering, Technical University of Crete,
Polytechnioupolis, GR-73100 Chania-Crete, Greece
Kristberg Kristbergsson
Department of Food Science and Human Nutrition, University of Iceland,
Hjardarhaga 2-6, Reykjavík, Iceland
Herman Van Langenhove

Saverio Mannino
Department of Food Science Technology and Microbiology, University of
Regina Nabais
C.E.R.N.A.S. -Centro de Recursos Naturais, Ambiente e Sociedade, Escola
Superior Agrária de Coimbra, Instituto Politécnico de Coimbra. Bencanta,
3040-316 Coimbra, Portugal
xiii
xiv Contributors
Gudrun Olafsdottir
Icelandic Fisheries Laboratories, Skulagata 4, 101 Reykjavík, Iceland
João Peixoto
Department of Biological Engineering, CEB, University of Minho, 4710-057
Braga, Portugal
Elefteria Psillakis
Department of Environmental Engineering, Technical University of Crete,
Polytechnioupolis, GR-73100 Chania-Crete, Greece
Bram Sercu
1
Odor Problems in the Food
Industry
Elefteria Psillakis and Vassilis Gekas
1. INTRODUCTION
With growing population, industrialization, and urbanization, the odor problem

has been assuming objectionable proportion. Rapidly growing industrialization
has aggravated the problem through odorous industrial operations. Undesirable
odors contribute to air quality concerns and affect human lifestyles. On the eco-
nomic front, loss of property value near odor-causing operations/industries and
odorous environment is partly a consequence of offensive odor. Odor is undoubt-
edly the most complex of all the air pollution problems. Malodors generated by
the food industry vary enormously since they can be generated during the produc-
tion, processing, or even in the waste water treatment areas of the plant.
In general, odor can be defined as the “perception of smell” or in scientific
terms as the “organoleptic attribute perceptible by the olfactory organ on sniff-
ing certain volatile substances” (ISO 5492:1992). Whether pleasant or unpleas-
ant, odors are induced by inhaling airborne volatile organics or inorganics.
Unlike conventional air pollutants, odor has distinctly different characteristics,
which, to an extent, can be comparable with noise pollution given that similar to
noise, nuisance is the primary effect of odor on people.
2. HUMAN RESPONSE TO ODORS
Odor is complex both because of the large number of compounds that contrib-
ute to it and because it involves a subjective human response. In general, different
people find different odors offensive at different concentrations and this is fre-
quently related to the way different people perceive odors. According to a simple
ELEFTERIA PSILLAKIS AND VASSILIS GEKAS Department of Environmental Engi-

neering, Technical University of Crete, Polytechnioupolis, GR-73100 Chania-Crete, Greece. e-mail:
epsilaki@mred.tuc.gr, vgekas@enveng.tuc.gr
1
2 E. Psillakis and V. Gekas
model (Frechen, 1994) describing human odor perception it is the physiological

reception and the psychological interpretation that results in a mental impres-
sion of a specific odor. Consequently, although the human olfactory organ is
quite sensitive, the response to odor is more related to past memories or cultural
experiences. It appears that odors have the ability to evoke memories that have
been suppressed for many years and some of them may be associated with sad-
ness or unpleasant situations. Our response to an odor may not be caused by
high concentration or intensity but rather to a strong memory or an impres-
sion instilled many years ago. Discrimination between “good” and “bad” smells
is important since pleasant and unpleasant smells require different behavioral
responses (Jacob et al., 2003). Bad smells warn us of danger, poor air quality,
“off ” food, poisons, even illness—each of which requires some immediate deci-
sion to be made and action to be taken, for example, avoidance or withdrawal.
Pleasant smells, on the other hand, do not necessitate immediate actions or deci-
sions. In fact, the biological significance of pleasant smells is not immediately
obvious. Recent evidence suggests that malodors activate different areas of the
brain from “pleasant” odors (Zald and Pardo, 2000). The human olfactory sys-
tem consists of the olfactory epithelium, the olfactory bulb, and the olfactory
cortex. The olfactory epithelium is an area of approximately 5 cm2 located in
the upper nasal cavity, containing between 107 and 108 receptor cells (Lancet,
1991). The receptor cells connect via olfactory neurons to the olfactory bulb,
where preprocessing of the electrical outputs from the receptor cells takes place
before passing to the olfactory cortex where further processing takes place in the
higher-order olfactory structures of the central nervous system (Pearce, 1997).
It is said that humans can differentiate about 10000 odors with differing qual-
ities. To date it is not possible to predict an odor sensation due to the chemical
structure of an odorant with a view to establishing an odorant classification sys-
tem (Gostelow et al., 2001). This is due to the fact that substances of similar or dis-
similar chemical constitution may have similar odors. Furthermore, the nature and
strength of odor may change on dilution and weak odors are not perceived in the
presence of strong odors. When odorous compounds with the same strength blend
to produce a combination, it is possible that one of them may be unrecognizable.
As mentioned earlier, memories are strongly correlated with odors. How-
ever, there are several other parameters affecting human perception of odors
and these include offensiveness, duration of exposure to odor, frequency of
odor occurrence, and tolerance/expectation of the receptor. For example, fatigue
from continued exposure to an odor may affect the human sense of smell. This
phenomenon is called adaptation. Adaptation may reduce both perceived odor
intensity and perceived odor quality (Stuetz et al., 2001). The degree of adapta-
tion resulting from exposure to an odorous air will depend on the odor con-
centration experienced. The weaker the odor concentration of an air sample,
the more adaptation affects perceived strength. The phenomenon of adaptation
frequently reveals itself in industrial situations, with workers reporting that an
initially repulsive odor eventually seemed less repulsive.
Odor Problems in the Food Industry 3
Other parameters such as age and gender also may contribute to the ability
to perceive an odor (Bliss et al., 1996) and to a lesser extent health (e.g., cold,
nasal allergy), personality, educational background, training, or hereditary defi-
ciencies in odor sensitivity.
3. ODORS AND HUMAN HEALTH
An important area of research is whether odors are simply a nuisance or a legiti-

mate health threat. In fact, very little information is available on the impact of
odor on human health. Most studies address the impact of individual chemicals
on human health (e.g., the smell of hydrogen sulfide and its lethality to humans).
There are several problems to that approach: the toxicity of most chemicals, even
those with high production volumes, is not known, and frequently the chemical
composition of materials resulting from complex industrial processes is unidenti-
fied (Rosenkranz and Cunningham, 2003). It should be mentioned, however, that
much of the research to date on the relationships between individual chemicals
and human health has been conducted indoors, where gas concentrations are
higher than in open conditions.
Although there is limited evidence that serious risks to physical health
occur downwind, for example, of livestock confinement facilities, some research
suggests that odor-causing substances can cause health effects such as eye, nose
and throat irritation, headache, and drowsiness, and possibly aggravate allergies,
asthma, and bronchitis (Sohn et al., 2003). More frequently, though, strong,
unpleasant, or offensive smells can interfere with a person’s enjoyment of life
especially if these odors are frequent and/or persistent. Studies indicate that
odors may alter a person’s attitude (Schiffman et al., 1995; Radon et al., 2004).
However, it is unclear if the attitude altering is a psychological or physiological
response to odor. Some people may feel angry and frustrated because of the
distasteful smell of the gases rather than being physically affected by the gases.
The argument may still be made, however, that in either case the person’s attitude
was altered, making the presence of odors a valid health concern. More research
is underway, but it is important to recognize that while odors from food industry
may not affect public health in general, odors can affect a person’s lifestyle, and
in such cases a person’s psychological well-being may be just as important as his
or her physical health.
4. ODOR GUIDELINES AND REGULATIONS
The regulations and guidelines vary from country to country. It is neither pos-
sible nor practical to provide all guidance or regulations of different countries.
As such, the aim of this section is to provide some useful information relevant
to the idea behind odor nuisance. It also should be mentioned here that the term
“guidelines” in the context of this chapter implies not only numerical values but
also any kind of guidance given.
4.1. Germany
After World War II, the rapid economic development and the densely populated
areas were responsible for the early regulatory legislation regarding air pollution
in Germany. The legal basis for any requirement with respect to air quality is the
German Federal Protection Act for Ambient Air (1974/1990) and the Technical
Instruction on Air Quality Control (1986). According to the Federal Protection
Act for Ambient Air, odors caused by plants and are treated as an annoyance,
although the problem is to find out whether an annoyance has to be considered
as significant (relevance of the annoyance).
The state of Northrhine-Westfalia is undoubtedly the most densely popu-
lated and industrialized area in Germany and for decades now the authorities of
this state have been developing and testing a new regulation/directive, according
to which a complete system has been designed, beginning with measurement
methods of the existing odor load and calculation of the expected odor load,
and concluding with ambient air quality requirements expressed as limit values
in terms of maximum permitted odor frequency in ambient air in certain areas
(Frechen, 2000; Both, 2001).
These limit values were developed on the basis of investigations in which the
existing odor load measured as odor frequency (Guideline VDI 3940, 1997) and
the degree of odor annoyance of residents assessed by questionnaires according
to guideline VDI 3883 Part 1 (1997) were correlated. The result was to set a limit
impact concentration at 1 odor unit /m3 (see definition in Chapter 2, paragraph
3.2) and then to limit the time percentage during which a higher impact con-
centration is tolerable (Both, 2001). These time percentages (odor frequencies)
were between 10% for residential areas and 15% for industrial areas. In order to
assess more correctly the extent of odorous emission to people it is stipulated
that an hour may be recognized as exceeding the limit value if the limit value is
exceeded during 10% of 1 hour, thus during 6 minutes. In practice, it is assumed
to be sufficient to multiply the hourly mean by a factor of 10 (the so-called factor
10 model). Thus an impact concentration of 0.1 odor unit /m3 (hourly average)
equals to 1 odor unit /m3 in the sense of the directive.
4.2. The Netherlands
The general policy in Netherlands is to keep the population as free as possible from
annoyance. The term “annoyance” is translated as the percentage of population
perceiving sometimes or even often an annoying odor. In this context in the year
2000, 12% of the population was annoyed by industrial odors (Frechen, 2001).
The pig production sector in the Netherlands is considerable in size, relative
to both the size of the population and the surface area of the country. Annual
production is approximately 30 million pigs, which amounts to 2 pigs per head
of the population. It is therefore not surprising that odor impact of pig produc-
tion is a major environmental issue, given the high density and proximity of both
residents and pigs.
The first guideline on how to take account of environmental odor aspects
for licensing as a result of application of the existing the Nuisance Law was first
issued in 1971 and revised several times in later years.
4.3. United Kingdom
The Environmental Protection Act of 1990 provides the legal framework for
avoiding and controlling odor nuisance in the United Kingdom. A comprehensive
overview over “The legal context of odor annoyance” in the United Kingdom can
be found elsewhere (Salter, 2000). The regulations, however, do not set any general
valid emission standards concerning odor and retracts to more general statements
concerning the odor nuisance. This is expected to change in the near future after
publication of Technical Guidance Note H4 (2002), Integrated Pollution Preven-
tion and Control (IPPC), Horizontal Guidance for Odor by the Environment
Agency.
4.4. United States of America
The regulations for odor in the United States of America vary from state to
state. Depending on the state regulation may concern hydrogen sulfide limits,
detection to threshold limits, or use as the main legal basis the nuisance law. An
overview on the regulation of odors in the United States is given by Thomas
Mahin (2001) and is summarized in Table 1.
Table 1. Examples of Ambient Standards for Odor-Causing Compounds [all agencies

listed are state agencies unless otherwise noted; from Mahin (2001)]
Location Compound Ambient Odor Standard
California Hydrogen sulfide 30 ppbva (1-hr average)

Idaho Hydrogen sulfide 10 ppbv (24 hr average)
30 ppbv (30 min average)
Minnesota Hydrogen sulfide 30 ppbv (30 min average)b
50 ppbv (30 min average)c
Nebraska Total reduced sulfur 100 ppb (30 min average)
New York City Hydrogen sulfide 1 ppbv (for wastewater plants)
North Dakota Hydrogen sulfide 50 ppbv (instantaneous, two readings 15 min apart)
Pennsylvania Hydrogen sulfide 100 ppbv (1 hr average)
5 ppbv (24 hr average)
Texas Hydrogen sulfide 80 ppbv (30 min average) residential/commercial
120 ppbv industrial, vacant or range lands
a
Parts per billion by volume.
b
Not to be exceeded more than two days in a 5-day period.
c
Not to be exceeded more than two times per year.
4.5. World Health Organization Guideline Values
For substances with malodorous properties at concentrations below those where

toxic effects occur, guideline values likely to protect the public from odor nui-
sance were established by the World Health Organization (1987, 2000). These
were based on data provided by expert panels and field studies for a limited
range of substances as 24-hour average concentrations. They were derived with
the aim of providing a basis for protecting the public from the adverse effects of
air pollution.
For a few of these substances that exhibit malodorous properties at con-
centrations below that at which toxic effects occur, guideline values have been
established for avoidance of substantial annoyance. For example, in the case of
hydrogen sulfide, guideline values based on sensory effects or annoyance reac-
tions, for an average time of 30 minutes, were set at 7 μg m−3 in order to protect
against substantial annoyance. It is important to note that these guidelines have
been established for single chemicals. Mixtures of chemicals can have additive,
synergistic, or antagonistic effects.
5. ODORS CHEMISTRY
A number of factors affect the emission of compounds from food industry opera-
tions. Most of the substances emitted are the products of microbial processes and
in most cases it is the microbial environment that will determine which substances
are generated and at what rate. This section intends to describe the chemical and
biological mechanisms that affect the formation and release of emissions. Most of
the information used in this section was taken by the draft report of the US Envi-
ronmental Protection Agency (2001).
5.1. Ammonia
The microbial decomposition of organic nitrogen compounds under both aerobic

and anaerobic conditions leads to the production of ammonia as a by-product. The
amount of ammonia volatilized from any operation depends on total ammo-
nia concentration, temperature, pH, and storage time (US Environmental Pro-
tection Agency, 2001). Furthermore, in solution, the partitioning of ammonia
between the ionized (NH4+) and unionized (NH3) species is controlled by pH
and temperature. As expected, at lower pH, ammonium is the predominant
species, and ammonia volatilization occurs at a lower rate than at higher pH
values where the rate of ammonia volatilization is increased. Given that the pH of
manures handled as solids ranges between 7.5 and 8.5, ammonia volatilization
is enhanced. In the case of manure, other aminocompounds including aliphatic
amines (methyl- and ethylamine) also may present but at lower concentrations. It
should be mentioned here that during the storage of fresh manure, amino acids
are most likely to undergo decarboxylation producing putrescine, cadaverine,
and ammonia (Zhu, 2000).
5.2. Sulfur Compounds
Hydrogen sulfide and other reduced sulfur compounds are produced during
anaerobic decomposition and hydrogen sulfide is the predominant reduced sul-
fur compound emitted from animal feeding operations. In the case of animal
manures, there are two primary sources of sulfur (US Environmental Protection
Agency, 2001): (1) sulfur amino acids contained in the feed and (2) inorganic
sulfur compounds, such as copper sulfate and zinc sulfate, serving as growth
stimulants and used as feed additives to supply trace minerals.
In general, the magnitude of hydrogen sulfide emissions depends on the
liquid phase concentration, temperature, and pH. It is well known that the solu-
bility of hydrogen sulfide in water increases at pH values above 7. Therefore, as
pH shifts from alkaline to acidic, the potential for hydrogen sulfide emissions
increases (Stumn and Morgan, 1996).
As already stated, under anaerobic conditions, any excreted sulfur will be
microbially reduced to hydrogen sulfide. Thus, all manures managed as liquids
or slurries are potential sources of hydrogen sulfide emissions; as such emissions
from confinement facilities with dry manure handling systems and dry manure
stockpiles should be negligible if there is adequate exposure to atmospheric oxy-
gen to maintain aerobic conditions.
Other sulfur compounds emitted from animal feeding operations include
methyl mercaptan, dimethyl sulfide, dimethyl disulfide, and carbonyl sulfide. In
general, the very offensive smelling compound methyl mercaptan is a product
of sulfur-containing amino acid decomposition, and it can be oxidized to form
the unpleasant-smelling compounds dimethyl disulfide or dimethyl sulfide (Zhu,
2000).
5.3. Volatile Organic Compounds
During the degradation of organic matter, volatile organic compounds (VOC)

are formed as intermediate metabolites. Under aerobic conditions, microbial
degradation will lead to the formation of VOC, which are rapidly oxidized to
carbon dioxide and water. Under anaerobic conditions, complex organic com-
pounds are degraded microbially to volatile organic acids and other volatile
organic compounds, which, however, in turn may be converted to methane
and carbon dioxide by methanogenic bacteria (US Environmental Protection
Agency, 2001).
It should be mentioned here that volatile fatty acids (VFA) represent a
large portion of VOC and are responsible for a significant proportion of odor
in emission plumes from swine production facilities. Together with CO2, H2, as
well as ammonia, they can be produced from the deamination of amino acids
that are produced during the process of protein degradation and breakdown of
carbohydrates. Typical acids in this group consist of acetic, propionic, butyric,
iso-butyric, valeric, iso-valeric, caproic, and capric acids.
5.4. Phenolic Compounds
Phenolic compounds such as phenols and p-cresols are produced from the micro-
bial degradation of tyrosine and phenylalanine in the intestinal tract of animals
(Ishaque et al., 1985). Indole, skatole, cresol, and 4-ethylphenol appear to be
the major components included in this group of compounds. In addition, the
metabolism of tryptophan can result in the production of indoleacetate, which
is subsequently converted into skatole (3-methylindole) and indole by a different
group of bacteria (Zhu, 2000).
6. FOOD INDUSTRY COMPONENTS
The major components in the food industry are depicted on Figure 1. As can
been seen edible products go through a number of stages before reaching the
consumer. It is generally accepted that odor problems in the food industry are
more likely to occur during the first two stages. Food service and food retail
operations are themselves generally not significant sources of air or odor pollu-
tion. No data are available on the portion of total emissions of specific pollut-
ants that can be ascribed directly to the food service and food retail industries.
The one exception is emissions from vent hood systems of food service equip-
ment, where in some cases users of certain food service cooking equipment are
requested to install pollution control measures.
A selection of the food industry activities of the first two stages are given in
Table 2. In the food industry large livestock operations, poultry farms, slaugh-
terhouses, food and meat processing industries, and bone mills are among major
contributors to odor pollution.
Grower/ Processing Wholesale Service/ Consumption

Fisherman Retail
Figure 1. Major components in the food industry. Adapted from Davies and Konisky (2000).
Table 2. Selected Food Industry Activities
Industry Activities Description
Livestock & poultry feed operations • Cattle feedlots

• Swine feedlots
• Poultry houses
• Dairy farms
Animal & meat products preparation • Meat packing plants
• Meat smokehouses
• Meat rendering plants
• Manure processing
• Poultry slaughtering
Fish processing • Fish canning
Dairy products • Natural and processed cheese
Preserved fruits and vegetables • Canned fruits and vegetables
• Dehydrated fruits and vegetables
• Pickles, sauces, and salad dressings
Grain processing • Grain elevators & processes
• Cereal breakfast food
• Pet food
• Alfalfa dehydrating
• Pasta manufacturing
• Bread baking
• Corn wet milling
Confectionary products • Sugar processing
• Cane sugar processing
• Sugar beet processing
• Salted & roasted nuts & seeds
• Almond processing
• Peanut processing
• Vegetable oil processing
7. SELECTED CASE STUDIES OF ODOR PROBLEMS IN

THE FOOD INDUSTRY
7.1. Livestock Enterprises (including animal feeding operations and manure

processing)
Odors and gases emitted are a by-product of the microbial decomposition

of manure and other organic matter. The amount and type of emissions are
dependent on the amount and type of microbial activity and consequently to
the moisture content, temperature, pH, oxygen concentration, and other envi-
ronmental parameters. Any changes in these parameters will alter odor and
gas emissions. For example, as temperature decreases, microbial activity slows
down, and as such during the winter months few odors are generated. Emis-
sion sources are distributed among livestock buildings, manure storage units,
land application of manure, and pastures (Skinner et al., 1997). Gas emissions
from buildings and storages are relatively constant and vary with seasonal tem-
peratures. In contrast, land application of manure has the potential to emit
large amounts of gases periodically throughout the year. Gas emissions vary
significantly with management practices and manure system design.
The main odorous compound groups associated with animal feeding
operation and wastes are fatty acids, amines, ammonia, aromatics, as well as
inorganic and organic sulfur (US Environmental Protection Agency, 2001).
The most frequently detected odorous compound from livestock operations
is undoubtedly hydrogen sulfide. Another group of malodorous compounds
released during animal feeding operations are the volatile organic compounds.
Next to the odor problem created by these compounds, the majority of them
have an environmental impact, given that some of them participate in atmo-
spheric photochemical reactions, while others play an important role as heat-
trapping gases.
7.2. Meat Processing
Meat smokehouses are used to add flavor, color, and aroma to various meats,
including pork, beef, poultry, and fish. The operations typically involved in the
production of smoked meat are tempering or drying, smoking, cooking, and
finally chilling. In general, the pollutants associated with meat smokehouses
include particulate matter (PM), carbon monoxide (CO), volatile organic com-
pounds (VOC), polycyclic aromatic hydrocarbons (PAH), organic acids, acro-
lein, acetaldehyde, formaldehyde, and nitrogen oxides. Acetic acid (followed
by formic, propionic, butyric, and other acids) has been identified as the most
prevalent organic acid present in smoke that is the primary source of emissions
in the smokehouses. Furthermore, heating zones in continuous smokehouses
(and the cooking cycle in batch smokehouses) are not a source of combustion
compounds [e.g., polycyclic aromatic hydrocarbons (PAH)]; they are, however, a
source of odor including small amounts of VOC as a result of the volatilization
of organic compounds contained in the meat or the smoke previously applied to
the meat (AP-42, 2004).
7.3. Meat-Rendering Plants
Meat-rendering plants process animal by-product materials for the production

of tallow, greas, and high-protein meat and bone meal. There are two types of
animal-rendering processes and these are (1) the edible rendering plants that
process fatty animal tissue into edible fats and proteins and inedible rendering
and (2) the inedible rendering plants produce inedible tallow and grease, which
are used in livestock and poultry feed, soap, and production of fatty acids. Vola-
tile organic compounds are the primary air pollutants emitted from these
rendering operations (AP-42, 2004) and are considered to be an odor nuisance in
residential areas in close proximity to rendering plants. The major constituents
that have been qualitatively identified as potential emissions include organic sul-
phides, disulfides, C-4 to C-7 aldehydes, trimethylamine, C-4 amines, quinoline,
dimethyl pyrazine, other pyrazines, and C-3 to C-6 organic acids. In addition,
lesser amounts of C-4 to C-7 alcohols, ketones, aliphatic hydrocarbons, and aro-
matic compounds are potentially emitted.
7.4. Fish Canning
Fish processing includes both the canning of fish for human consumption and
the production of fish by-products such as meal and oil. Although smoke and
particulates may be a problem, odors are the most objectionable emissions from
fish processing plants. The fish in the by-products segment is often in a further
state of decomposition, thus causing more odor problems than canning itself. The
fish meal driers form part of the fish by-products segment and are the largest odor
source. In addition, the odorous gases from reduction cookers consist primarily
of hydrogen sulfide (H2S) and trimethylamine [(CH3)3N] but are emitted from this
stage in appreciably smaller volumes than from fish meal driers (AP-42, 2004).
7.5. Cane Sugar Processing
One of the main environmental problems associated with sugar processing is

the management of effluent. All stages in sugar processing and refining produce
wastes that collectively are very rich in organic content, high in suspended solids,
and also may be colored (AP-42, 2004). The release of untreated wastewaters to
surface waters will cause extensive enrichment, stripping oxygen from the water
and killing aquatic life. As such, wastewater will need to be treated on-site before
released to sewer or surface water. Odor from the waste lagoons and from pro-
cessing may result in local nuisance complaints. If deemed a nuisance by the
authorities, mitigation measures may need to be implemented, though these usu-
ally relate to housekeeping measures.
7.6. Baking
The oven exhaust from baking industries releases water vapor, CO2, VOC, and
various combustion products. The VOCs are primarily ethanol produced by the
yeast during the fermentation process. It should be noted, however, that many
individuals find the odor of fresh baked bread very desirable.
8. CONCLUSIONS
How much odor should a community or individual have to tolerate? This is the
primary question that must be answered before any good odor policy can be
developed. Everyone would probably agree that if one person smells a facility
one day out of the year, that facility should not be declared a nuisance. But
several hundred people smelling a facility nearly every day throughout the year
would be considered a nuisance. The reality at most sites lies somewhere in the
middle. A definition of nuisance must be established that takes into account
odor intensity and frequency, meaning how bad and how often. An odor nui-
sance also may have to consider a relationship between the number of people
annoyed. Included in this decision also might be the economic impact of odor
control on the food industry.
9. ACRONYMS AND ABBREVIATIONS
mve Mestvarkeneenheden or fattener units

PM Particulate matter
PAH Polycyclic aromatic hydrocarbons
VFA Volatile fatty acids
VOC Volatile organic compounds
10. REFERENCES
AP-42 (2004) “Compilation of Air Pollutant Emission Factors, AP-42, Fifth Edition, Volume I:
Stationary Point and Area Sources.” U.S. Environmental Protection Agency. http://www.epa.
gov/ttn/ chief/ap42/index.html
Both, R. (2001) “Directive on odor in ambient air: an established system of odor measurement and
odor regulation in Germany.” Water Science & Technology 44 (9), 119–126.
Bliss, P.J., Schulz, T.J., Senger, T., Kaye, R.B. (1996) “Odor measurement—Factors affecting olfac-
tometry panel performance.” Water Science and Technology, 34 (3–4), 549–556.
Davies, T., Konisky, D.M. (2000) “Environmental Implications of the Foodservice and Food Retail
Industries.” Discussion Paper 00–11, Resources for the Future, Washington DC. http://www.rff.
org [2004. Nov 23]
Environmental Protection Agency, Ireland (2001) “Odor Impacts and Odor Emission Control Mea-
sures for Intensive Agriculture.” Environmental Research R&D REPORT SERIES No. 14.
http://www.epa.ie/pubs/
Federal Protection Act for Ambient Air (1974/1990) Act on the Prevention of Harmful Effects on
the Environment Caused by Air Pollution, Noise, Vibration and Similar Phenomena (Federal
Emission Control Act, Bundes-Immissionschutzgesetz—BImSchG) Federal Ministry for Envi-
ronment, Nature Conservation and Reactor Safety, Bonn (BGBl. I p.880). (available in English).
Frechen, F.-B. (1994) “Odor emissions from wastewater treatment plants—recent German experi-
ences.” Water Science and Technology, 30 (4), 35–46.
Frechen, F.-B. (2000) “Odor measurement policy in Germany.” Water Science and Technology, 41
(6), 17–24.
Frechen, F.-B. (2001) “Regulation and policies.” In Stuetz, R.M., Frechen, F.-B. (eds), Odours in
Wastewater, Treatment: Measurement, Modeling and Control. London: IWA publishing.
Gostelow, P., Parsons, S.A., Stuetz, R.M. (2001) “Odor measurements for sewage treatment works.”
Water Research, 35 (3), 579–597.
Guideline VDI 3883 Part 1 (1997). Effects and Assessment of Odors. Psychometric Assessment of
Odor Annoyance Questionnaires. Dósseldorf (German/English).
Guideline VDI 3940 (1993). Determination of Odorants in Ambient Air by Field Inspections.
Düsseldorf. (German/English).
Ishaque, M., Bisaillon, J.G., Beaudet, R., Sylvestre, M. (1985) “Degradation of phenolic compounds
by microorganisms indigenous to swine waste.” Agricultural Wastes, 13 (3), 229–235.
ISO 5492:1992 (1992) “Sensory analysis—vocabulary” International Organization for Standardization.
Jacob, T.J.C., Fraser, C., Wang, L., Walker, V., O’Connor, S. (2003) “Psychological evaluation
responses to pleasant and maldor stimulation in human subjects; adaptation, dose response
and gender differences.” International Journal of Psychophysiology, 48 (1), 67–80.
Lancet, D. “Olfaction—The strong scent to success.” Nature 351 (6324), 275–276.
Mahin, T.D. (2001) “Comparison of different approaches used to regulate odors around the world.”
Water Science and Technology, 44 (9), 87–102.
Pearce T.C. (1997) “Computational parallels between the biological olfactory pathway and its analogue.
The Electronic Nose .1. Biological olfaction.” Biosystems, 41 (1), 43–67.
Radon K., Peters A., Praml G., Ehrenstein V., Schulze A., Hehl O., Nowak D. (2004) “Livestock
odors and quality of life of neighboring residents.” Annals of Agricultural and Environmental
Medicine 11 (1), 59–62.
Rosenkranz, H.S., Cunningham, A.R. (2003) “Environmental odors and health hazards.” The
Science of the Total Environment, 313 (1–3), 15–24.
Salter, J. (2000) “The legal context of odor annoyance.” Proceedings of the International Meeting on
Odor Measurement and Modeling, Odor 1, Cranfield University.
Schiffman, S.S., Sattely Miller, E.A., Suggs, M.S., Graham, B.G. (1995) “The effect of environmental
odors, emanating from commercial swine operations on the mood of nearby residents.” Brain
Research Bulletin 37 (4), 369–375.
Skinner, J.A., Lewis, K.A., Bardon, K.S., Tucker, P., Catt, J.A., Chamber, (1997) “An overview of
the environmental impact of agriculture in the U.K.” Journal of Environmental Management,
50 (2), 111–128.
Sohn, J.H., Smith, R., Yoong, E., Leis, J., Galvin G. (2003) “Quantification of odors from piggery
effluent ponds using an electronic nose and an artificial neural network.” Biosystems Engineering
86 (4), 399–410
Stuetz, R.M., Gostelow, P., Burgess, J.E. (2001) “Odor perception.” In: Stuetz, R.M., Frechen, F.-B. (eds),
Odours in Wastewater Treatment: Measurement, Modeling and Control. London, IWA publishing.
Stumn, W., Morgan J.J. (1996) Aquatic Chemistry: Chemical Equilibria and Rates in Natural Waters.
Third Edition. Environmental Science and Technology, New York, John Wiley & Sons, Inc.
Technical Guidance Note H4, Integrated Pollution Prevention and Control (IPPC), Horizontal Guid-
ance for Odor, Part 2 Assessment and Control, Environment Agency, UK, 2002. Available on
www.environment-agency.gov.uk
Technical Instruction on Air Quality Control (1986) (Erste Allgemeine Verwaltungsvorschrift zum
Bundes-Immissionsschutzgesetz – Technische Anleitung zur Reinhaltung der Luft – TA Luft)
Federal Ministry for Environment, Nature Conservation and Reactor Safety, Bonn (GMBl. p. 95).
(available in English).
US Environmental Protection Agency (2001) “Emissions from animal feeding operations, draft.”
EPA Contract No. 68-D6-0011, Emission Standards Division, Office of Air Quality Planning
and Standards, Research Triangle Park, NC.
World Health Organization (1987) “Air quality guidelines for europe.” World Health Organization
Regional Publications, European Series No 23.
World Health Organization (2000) “Air quality guidelines for Europe. Second edition”. World Health
Organization Regional Publications, European Series No 91.
Zald, D.H., Pardo, J.V. (2000) “Functional neuroimaging of the olfactory system in humans.” Inter-
national Journal of Psychophysiology, 36 (2), 1165–1181.
Zhu, J. (2000) “A review of microbiology in swine manure odor control.” Agriculture, Ecosystems &
Environment, 78 (2), 93–106.
2
Odor Measurement
Elefteria Psillakis
1. INTRODUCTION
It generally is recognized that for effective odor control measures to be implemented

the problem must first be quantified. Odor measurement data can be used to:
Predict odor impact in the vicinity of an operation for odor impact assess-
ment purposes.
Provide information on the strength and intensity of odors.
Identify the causes of an odor problem and quantify the scale of odor
emission from a particular source.
Measure/evaluate the performance of an odor control technology imple-
mented by a company.
However, odors are temporal and spatially dimensioned and can be considered
to be one of the most difficult challenges for scientists to investigate. A person’s
response to an odor is highly subjective: different people find different odors
offensive and at different concentrations given that physiologically odor recogni-
tion is associated with the emotional center of brain. Furthermore, some of the
odorous compounds can be detected by the human nose in very low concentra-
tions (e.g., hydrogen sulfide) while others cannot be detected even at very high
concentrations (e.g., methane). This is further complicated by the fact that some
combinations of compounds may be more odorous than the sum of the indi-
vidual gases. Determining the impact area of odorous gases also is very difficult.
As wind direction and speed change, the odor impact area and intensity change.
Depending on specific conditions, odorous gases can travel several meters or
several kilometers. Gas transmission and impact area also depend on the specific
gas. For these reasons, there is no universally accepted method for the quanti-
fication of odors, and odor measurement often has been regarded as an art as
opposed to a science (Gostelow et al., 2001).
ELEFTERIA PSILLAKIS Department of Environmental Engineering, Technical University

of Crete, Polytechnioupolis, GR-73100 Chania-Crete, Greece e-mail: epsilaki@mred.tuc.gr.
15
16 E. Psillakis
2. SAMPLING
The sampling of an environmental parameter, such as odor, is an extremely com-

plex task for environmental engineers and scientists. Prior to preparing a sam-
pling program it is necessary for a field inspection to be carried out in order to
determine the potential odor sources at the site to be investigated. Factors to be
noted include location and conditions affecting odor emissions (controlled or
uncontrolled such as weather), accessibility of odor sampling points, and char-
acteristics of odor emissions (e.g., temperature and humidity) as well as safety
requirements due to toxicity. Sampling points as well as frequency, duration, and
averaging time of sampling should reflect both spatial and temporal pattern of
the facility to be investigated. Increasing the number of samples to be taken will
increase the cost of the study. In some cases composite samples across several
sampling points are used instead of a large number of discrete samples (Jiang
and Kaye, 2001). It therefore is essential to carefully design the sampling pro-
gram and to coordinate it in advance with the testing program, thus minimizing
the time elapsed between sampling and analysis.
2.1. General Considerations for Sampling Odorous Compound
In those parts of the sampling equipment coming into contact with the odor-
ant sample, appropriate materials such as PTFE, PET, FEP, Tedlar™, glass, or
stainless steel must be used. Materials such as brass and silicone or natural rub-
ber must not be used in sampling lines and fittings. In general, representative
odorous samples are collected in the field using special purpose atmospheric
sampling bags, whose quality must conform to criteria such as being odor free,
nonadsoprtive, leak-free, reasonably robust, and having sufficient volumetric
capacity. Commonly, sampling bags are made of Tedlar™, FEP, or the low-cost/
single-use Naoplhan NA™. Every new sampling bag should be tested for their
background odor concentration as well as for leakage. Because of their high cost
it is a common practice to reuse these bags. Nonetheless, where materials are
reused, procedures for cleaning and conditioning must be applied as defined in
certified methods. In addition, the interval between sampling and measurement
should not exceed 30 hr, during which time samples must be kept in the dark and
at a temperature below 25°C (EN 13725,2003).
Sampling bags may be filled using either the direct or indirect sampling
techniques. According to the direct sampling, the bag is filled under pressure
by pumping the air sample into the bag. In indirect sampling, the odor sample
is collected using an odor sampling system, having a vacuum pump and a 12-
volt battery built into a sampling drum as shown in Figure 1. After placing a
sampling bag into the sealed sampling vessel, air is pumped out of the sampling
drum, creating a vacuum inside the drum. Sample air is drawn into the bag by
the pressure difference between the inside and outside of the bag. Because of the
Odor Measurement 17
Odour Sample Intake
Tedlar Bag
Vacuum
Switch
Exhaust
Pump and battery
Figure 1. Arrangement of apparatus for indirect odor sampling system.
risk of contamination the direct method is seldom used and the indirect sam-
pling approach is generally recommended.
2.2. Emission Source Types
Typically emissions sources are characterized as point, area, and volume.

Another type of emission source is the fugitive one (e.g., leaking valves). How-
ever, according to the EN 13725:2003, quantification of such emissions incurs
large errors and consequently no known technique can be recommended.
2.2.1. Sampling of Point Sources
Typically a point source will be a stack with a known flow rate such as a dis-
charge stack from abattoir or even a vent from a pig shed. Gaseous samples from
these sources can be collected using a sampling train consisting of a probe, a
delivery pipe, and an optional particulate filter before the collection system (EN
13725:2003) such as the one described in the US EPA Method 0030 (1986). It is
important that air velocity, dimensions of the vent, temperature, and humidity are
measured before a sample is taken. Appropriate guidance regarding the selection
of sampling points and velocity measurement and sampling point location can be
found in ISO Method 9096: 1992(E) (Stationary source emissions—Determina-
tion of concentration and mass flow rate of particulate material in gas-carrying
ducts—Manual gravimetric method), ISO 10780: 1994(E) (Stationary source
emissions—Measurement of velocity and volume flow rate of gas streams in
ducts), and Australian Standard AS 4323.1 1995 (Stationary Source Emissions,
Method 1—Selection of Sampling Positions).
Predilution of the sample should be undertaken for samples taken directly
from combustion processes where the air temperature and relative humidity
18 E. Psillakis
exceed 50 °C and 90%, respectively, thus preventing condensation in the sample

bag as well as reducing the concentration to levels suitable for analysis (Jiang
and Kaye, 2001).
2.2.2. Sampling of Area Sources
Typically an area source will be a water or solid surface such as the water surface
of a slurry storage tank or a cattle feedlot and can be sampled using different
methods such as the portable wind tunnels. Area sources can be distinguished
in sources with a measurable outward airflow and sources that do emit odor but
have no measurable outward flow.
2.2.2a. Sources Without Outward Flow
In the early 1970s, Lindvall (1970) introduced an odor emission hood used in the
comparison study of odor strengths from different sources. Later, Lockyer (1984)
designed a wind tunnel system to measure ammonia loses from pastures. Recently,
an improved portable wind tunnel system was developed at the University of New
South Wales (Jiang et al., 1995; Bliss et al., 1995) to measure odors emission rates
from liquid and solid area sources (sewage and industrial wastewater treatment
plants, cattle feedlots, mushroom composting, piggeries, etc). The principle of
the wind tunnel system is that controlled air (recently filtered by activated carbon
through a series of devices) forms a consistent flow over a defined liquid or solid
surface. Convective mass transfer takes place above the surface as odor emission
happens in the natural atmosphere. The odor emissions are then mixed with clean
air and vented out of the hood. A proportion of the mixture is drawn into a sam-
pling bag via Teflon tubing using the sampling vessel. The air velocity used inside
the wind tunnel is 0.3 m/sec, which is the lowest reliable measurable air velocity
directly inside the main section of the wind tunnel. The aerodynamic performance
of the wind tunnel has been validated based on the wind speeds and height where
most complaints occurred. An isometric sketch of a portable wind tunnel sys-
tem is shown at Figure 2. It should be mentioned here that according to the EN
13725:2003 this method is of limited application when the tunnels use unfiltered
atmospheric air as intake, which is not always odor free.
An alternative to wind tunnel systems is the isolation chambers, also called
flux hoods (Klenbusch, 1986; Gholson et al., 1991). According to this method,
during sampling a representative area of the source surface is enclosed by the
isolation flux chamber. Then a controlled flow of clean sweep gas (normally
nitrogen or odor- and hydrocarbon-free air) is released into the chamber and is
used to transport the emission from the surface to be sampled. The performance
of isolation chambers largely depends on the configuration of the enclosure and
operating procedures (Hwang, 1985).
A comparative study on portable wind tunnel system and isolation cham-
ber for the determination of VOCs from real sources (Jiang and Kaye, 1996)
Odor Measurement 19
Mixing Chamber
Main Section
Floating Tubes
Figure 2. Isometric sketch of portable wind tunnel system.
(Table 1) reported emission rates that were not in accordance and emphasized
the implications of the Henry’s Law constant for mass transfer processes in the
case of the isolation chamber. The same report concluded that this was due to
the fact that the use of isolation chambers may result in significant underestima-
tions of emission rates for real situations and suggested the use of wind tunnel
systems. Nonetheless, the EN 13725:2003 clearly indicates the need for further
research and standardization.
Table 1. Comparison of Wind Tunnel and Isolation Chamber (Jiang, and Kaye, 1996)
Wind tunnel Isolation chamber
Aerodynamics Parallel and even air flow inside Conditions at the emission
the wind tunnel surface not known
Emissions and air well mixed at the exit Emissions and sweep gas
may not be well mixed
Operating Temperature and relative humidity are Temperature and relative
parameters close to ambient condition during humidity are influenced
sampling by solar heating during
long sampling duration
Air velocity = 0.3 m/s Sweep gas = 5L/min
Stabilization time = 5 min Equilibrium time = 24 min
Sample collection rate = 20 L/min Sample collection rate = 2 L/min
Limitations Sample contamination due to Not for chemicals that have
atmospheric air as intake low Henry’s Law constants
20 E. Psillakis
2.2.2b. Sources with Outward Flow
Although, the use of wind tunnels has been demonstrated for all types of area
sources (Jiang and Kaye, 1996), significant limitations may exist in the cases of
area sources with outward flow. For example, in some instances the placement of
the wind tunnel system may create back pressure, limiting the flow of outward
moving air into the wind tunnel, and leading to an underestimation of odor
emission rates (Jiang and Kaye, 2001). The ideal situation would be to cover the
whole total source (or a large part) with foil. The cover must be left open at some
point in order to allow the air to escape as well as to allow sampling using the
point source apparatus (EN 13725:2003).
2.2.3. Sampling of Building Sources
Typically building sources, such as chicken and pig sheds, have a number of
openings. Prior to about 10 years ago, little research was undertaken on the
determination of odor emissions from buildings. For building sources, measure-
ments of both odor concentration and air ventilation rate are required. The air
ventilation rate from animal housing is dependent on operational conditions
(e.g., opening or closure of side flaps or shutters) and ambient wind speed and
direction (Jiang and Kaye, 2001).
For animal sheds, odor samples are normally taken from several points
within a shed. Experience indicates that composite sample may be sufficient to
represent a single shed at a particular time. Additional samples can be taken at
different times of the day or week or to understand the fluctuation of the odor
concentration levels within a day or a week. Similarly, sampling may be carried
out for different weeks during the grow-out cycle or for different seasons during
a year or longer.
2.2.4. Calculation of Emission Rates
Emission rates, required for odor impact assessment, are calculates using the
odor concentration measured by olfactometry together with other measured
properties of the emission source and the sampling apparatus. For point sources
the odor emission rate (OER) is calculated using the odor concentration (OC:
ou/m3; see definition in paragraph 3.2) measured by olfactometer and the mea-
sured gas flow rate (Q,m3/s) according to the following equation:
OER = Q × OC
The specific odor emission rate (SOER) may be defined as the quantity (mass)
of odor emitted per unit time from a unit surface area. The quantity of odor
emitted is not determined directly by olfactometry but is calculated from the
concentration of odor (OC: ou/m3) as measured by olfactometry, which is then
multiplied by the volume of air passing through the hood per unit time. The
volume per unit time is calculated from the measured velocity through the wind
Odor Measurement 21
tunnel, which is then multiplied by the known cross-sectional area of the wind
tunnel. Given that A is the area covered by the wind tunnel (m2), then SOER is
calculated by the expression:
Q × OC
SOER =
A
Finally, for building sources the odor emission rate (OER) is calculated
using the odor concentration measured by olfactometer through the door and
window openings. The following equation can be applied and the measured gas
flow rate (Q, m3/s) according to the following equation:
OER = Q × OC
where Q is the gas ventilation rate (m3/s) and OC is the odor concentration mea-
sured by olfactometer (ou/m3).
3. POINTS TO CONSIDER
3.1. Terms Associated with an Odor Measurement
In considering odor measurement, it is important to distinguish between odor-

ants and odors. An odorant is a substance that stimulates a human olfactory
system so that an odor is perceived, whereas an odor is an organoleptic attribute
perceptible by the olfactory organ on sniffing certain volatile substances (EN
13725:2003). The linkage between odorant properties and odor perception is
not clear due to the lack of a comprehensive theory of olfaction. Two broad
classes of odor measurement exist as a result: analytical measurements, referring
to odorants, and sensory measurements, employing human subjects relating to
odors (Gostelow et al., 2001). The terms (also called dimensions) associated with
an odor measurement refer to the parameters of an odor that can be measured.
There are four generally accepted terms: concentration, intensity, character, and
hedonic tone.
Odor concentration is the most frequently measured parameter and can
be measured analytically or by sensory means. Analytical measurements give
the physical concentration for specific odorants, whereas sensory concentration
measurements determine the number of dilutions required to reduce an odor
to its threshold concentration, which is the lowest concentration at which an
odor either can be detected or recognized (Gostelow et al., 2001). Recognition
thresholds are typically higher than detection thresholds by a factor of 1.5–10
(Dravnieks and Jarke, 1980). It should be mentioned here that odorant concen-
tration is the only odor dimension that can be measured analytically. The rest of
the parameters can be measured only using sensory methods.
Odor intensity is the strength of the perceived odor sensation. It is related
to the odorant concentration. The odor intensity usually is stated according to a
predetermined subjective rating system (i.e., faint, moderate, strong) by subjective
22 E. Psillakis
magnitude estimates, or (i.e., odor A is twice as strong as odor B) or by reference

to a specific odorant, whose concentration is adjusted until the reference and test
odorant have the same perceived intensity (Gostelow et al., 2001).
It is important to recognize the distinction between odor intensity and gas
concentration. Odor intensity is a measure of detection sensed by the nose. Odor
concentration is the actual concentration of the gas in the air. Nonetheless, the
two terms are related and the perceived intensity increases with increasing odor
concentration, although the relationship is not linear. The laws that have been
proposed to explain intensity–concentration relationships are the Weber–Fechner
law and Steven’s law (Sarkar and Hobbs, 2002).
Odor character or quality is the property to identify an odor and to differ-
entiate it from another odor of equal intensity. Examples of odor characters for
specific odorants are given in Table 2.
The hedonic tone is a property of an odor relating to its pleasantness or
unpleasantness. When an odor is evaluated for its hedonic tone in the neutral
context of an olfactometric presentation, the panelist is exposed to a controlled
stimulus in terms of intensity and duration. The degree of pleasantness or unpleas-
antness is determined by each panelist’s experience and emotional associations.
Often, negative values are used to represent unpleasant odors and positive values
represent pleasant odors.
3.2. The Unit of Odor Measurement
The odor unit is a difficult unit to define because it relates a physiological effect
to the stimulus that caused it. Odor concentrations derived by threshold olfac-
tometry are dimensionless. The concentrations may be termed threshold odor
numbers (TON) or dilution to threshold (D/T) ratios. Nonetheless, it is becom-
ing increasingly common to envisage odor concentrations as physical concentra-
tions and to express them as odor units per cubic meter (ou/m3). In the United
States, the same measure has been expressed as odor units per cubic foot (ou/ft3).
Numerically, TON, ou/m3, and ou/ft3 are identical, so care must be taken when
the concentration ratios are expressed as a physical concentration.
According to the EN 13725:2003, the European odor unit (ouE) is that
amount of odorant(s) that when evaporated into 1 m3 of neutral gas at standard
conditions elicits a physiological response from a panel (detection threshold)
equivalent to that elicited by one european reference odor mass (EROM) evapo-
rated in 1 m3 of neutral gas at standard conditions. One EROM evaporated into
1 m3 of neutral gas at standard conditions is the mass of substance that will elicit
the D50, which is 50% incidence of effect in physiological response (detection
threshold) assessed by an odor panel in conformity with this standard and has
by definition a concentration of 1 ouE/m3. For n-butanol one EROM is 123 mg.
Evaporated in 1 m3 of neutral gas, at standard conditions, this produces a con-
centration of 0,040 mmole/mole (which is equal to a volume fraction of 40 parts
per billion). There is one relationship between the ouE for the reference odorant
Odor Measurement 23
Table 2. Odorants Associated with Sewage Treatment Works (Abbott, 1993; Bonnin
et al., 1990; Brennan, 1993; Gostelow et al., 2001; Young, 1984)
Class Compound Formula Character
Sulfurous Hydrogen sulfide H2S Rotten eggs

Dimethyl sulfide (CH3)2S Decayed vegetables, garlic
Diethyl sulfide (C2H5)2S Nauseating, ether
Diphenyl sulfide (C6H5)2S Unpleasant, burnt rubber
Diallyl sulfide (CH2CHCH2)2S Garlic
Carbon disulfide CS2 Decayed vegetables
Dimethyl disulfide (CH3)2S2 Putrification
Methyl mercaptan CH3SH Decayed cabbage, garlic
Ethyl mercaptan C2H5SH Decayed cabbage
Propyl mercaptan C3H7SH Unpleasant
Butyl mercaptan C4H9SH Unpleasant
tButyl mercaptan (CH3)3CSH Unpleasant
Allyl mercaptan CH2CHCH2SH Garlic
Crotyl mercaptan CH3CHCHCH2SH Skunk, rancid
Benzyl mercaptan C6H5CH2SH Unpleasant
Thiocresol CH3C6H4SH Skunk, rancid
Thiophenol C6H5SH Putrid, nauseating, decay
Sulfur dioxide SO2 Sharp, pungent, irritating
Nitrogenous Ammonia NH3 Sharp, pungent
Methylamine CH3NH2 Fishy
Dimethylamine (CH3)2NH Fishy
Trimethylamine (CH3)3N Fishy, ammoniacal
Ethylamine C2H5NH2 Ammoniacal
Diamines NH2(CH2)5NH2 Decomposing meat
Pyridine C6H5N Disagreeable, irritating
Indole C8H6NH Fecal, nauseating
Scatole or Skatole C9H8NH Fecal, nauseating
Acids Acetic (ethanoic) CH3COOH Vinegar
Butyric (butanoic) C3H7COOH Rancid, sweaty
Valeric (pentanoic) C4H9COOH Sweaty
Aldehydes / Formaldehyde HCHO Acrid, suffocating
ketones Acetaldehyde CH3CHO Fruit, apple
Butyraldehyde C3H7CHO Rancid, sweaty
Isobutyraldehyde (CH3)2CHCHO Fruit
Isovaleraldehyde (CH3)2CHCH2CHO Fruit, apple
Acetone CH3COCH3 Fruit, sweet
Butanone C2H5COCH3 Green apple
and that for any mixture of odorants. This relationship is defined only at the D50
physiological response level (detection threshold), where:
1 EROM ≡ 123 mg n-butanol ≡ 1 ouE for the mixture of odorants
This linkage is the basis of traceability of odor units for any odorant to that of
the reference odorant. It effectively expresses odor concentrations in terms of
“n-butanol mass equivalents”. The odor concentration can be assessed only at a
24 E. Psillakis
presented concentration of 1 ouE/m3. As a consequence the odor concentration

is expressed as a multiple of one ouEin a cubic meter of neutral gas. The odor
concentration, in ouE/m3, can be used in the same manner as mass concentra-
tions (kg/m3) (EN 13725:2003).
The stimulus in this case can be a multitude of substances. In that sense the
odor unit is very similar to the LD50, as used in toxicology assessments, indicat-
ing the dose that causes a lethal effect in 50% of a well-defined test population.
The physiological reaction is the unifying reaction that can be caused by a wide
range of substances, at an equally wide range of dosages. In odor research the
D50 could be described as the dose that 50% of a population that can detect as a
sensory stimulus (EN 13725:2003).
4. OVERVIEW OF METHODS
There are a number of different methodologies in use for odor analysis. Selection
of a particular method will depend upon the purpose of the measurement, the
frequency of monitoring, sampling location, type of source emission (e.g., point
source), as well as the nature and complexity of the emission.
In general, odor can be “measured” using chemical (analytical) or sensory
methods. It is generally accepted that these categories do not have clear cutoff
points and some assessment methodologies could be considered to fall into more
than one. The analytical techniques include chemical analysis and direct reading
instrumental analysis. Chemical analysis is the indirect assessment involving the
collection of a sample which, when analyzed, will give the concentration of the
various chemical species present. This includes substance-specific wet chemistry
methods, as well as sample collection followed by analysis by means of instruments
such as gas chromatography (GC). Direct reading instrumental analysis provides
information on the concentration of specific chemical species or their concentra-
tions relative to each other. This includes among others portable analyzers (such as
portable GC), the “electronic nose,” as well as colorimetric tubes.
The sensory methods on the other hand relate to human response giving
an assessment of the physiological response to a particular mixture—strength,
quality, characteristics—which provides information on the likely population
response. This is obtained by exposing trained individuals to samples of the
odorous air either in the laboratory or in the field.
5. CHEMICAL ANALYSIS
5.1. Substance-Specific Wet Chemistry Techniques
Substance-specific wet chemistry techniques are usually undertaken at the source

and are considered to be recognized methodologies for assessing compliance
with emission limits. These include:
Odor Measurement 25
Determination of ammonia in air samples by drawing sample gas through

a sulfuric acid solution. The ammonia concentration is determined by
titration after Kjeldahl distillation.
Determination of hydrogen sulfide in air samples by drawing sample gas
through an ammoniacal cadmium chloride solution. The hydrogen sul-
fide concentration is determined by an iodiometric titration of cadmium
sulfide.
The detection limits achieved by these techniques are determined by the vol-
ume of gas sampled. Typically limits of detection vary between 0.1 and 0.5 mg/
m3. In order to achieve low detection limits, large volumes of gas need to be sam-
pled; but the sampling flow rate through typical sampling equipment is limited
to about 2 L/min (IPPC H4-Part 2, 2002). Consequently, long sampling times are
required and peaks in concentration will be missed. In addition, the equipment
is fragile and set up can be time consuming.
5.2. Gas Chromatography
Gas chromatography (GC) is a widely used analytical technique for the sepa-
ration, identification, and quantification of the components of an odorous air
sample. The use of GC allows chromatographic separation of gaseous and liquid
mixtures into individual components. In general, when a sample mixture is intro-
duced into the heated injection port of a GC, it vaporizes. The gaseous sample
mixes with an inert gas (such as helium) also referred to as carrier gas and passes
through a glass or metal tube (column) that contains an absorbent. Because the
various components of the sample interact with the absorbent (stationary phase)
of column to different degrees, compounds will be released from the tube at dif-
ferent specific times. There is a wide selection of column types from a range of
different manufacturers. Nonetheless, in the case of odorous samples, the choice
of a column can be difficult as there are several options or combinations neces-
sary to obtain the optimum information about an odor’s components. Eventu-
ally the components of the injected sample are separated and exit the column at
different times (called “retention times”). These “elution” times are compared to
those of known compounds (analytical standards), thus allowing to some degree
identification.
Once the compounds are separated, they elute from the column and then
enter a detector. The detector is capable of creating an electronic signal when-
ever the presence of a compound is detected, and in most cases this signal is
linearly related to the concentration of the target analyte in the sample. A variety
of detectors are currently available, including mass spectrometer, flame ioniza-
tion, and photoionization detectors (Hobbs, 2001).
Despite the fact that gas chromatographic analyses result in accurate and
reproducible measurements, however, there are some major limitations. First,
the chemical concentrations corresponding to the odor detection thresholds
26 E. Psillakis
cannot be determined due to the synergistic olfactory effects of stimuli compris-

ing complex mixtures of gases (Gostelow and Parsons, 2000). In addition, no
indication is obtained as to the relevance of individual compounds to the odor of
the sample as a whole (Hobbs et al., 1995). Even if individual chemical concentra-
tions and their odor threshold values are known, it is not possible to deduce the
overall sample odor threshold or the odor character of the mixture of odorants.
In addition, direct calibration for analyzing odors can be challenging because
the composition mixture often will be unknown (IPPC H4-Part 2, 2002). Fur-
thermore, in many cases identification remains ambiguous or questionable as a
result of the presence of unknown components at very low concentration level
(Bockreis and Jager, 1999). This is because analysis of odorous samples by direct
injection is in most cases impossible due to the low concentrations of malodor-
ous compounds in these samples (Gostelow et al., 2001; Pillonel et al., 2002).
This problem can be overcome if a preconcentration step prior to analysis is
introduced and will be presented in more details in Chapter 3.
For all these reasons, the characteristics of complex odors cannot be derived
reliably from the individual chemical characteristics and chemical concentrations
of the odorous compounds present in a gas mixture. Nonetheless, gas chroma-
tography in general may help a process design engineer to select equipment if the
type of odor is unknown and may help researchers understand the mechanisms
of odor removal. Longer-term samples will average out any peaks, although this
may be of secondary importance in source/compound identification.
5.2.1. Mass Spectrometry
Mass spectrometers consist of an ion source, a mass-selective analyzer, and an

ion detector and use the difference in mass-to-charge ratio of ionized atoms or
molecules to separate them from each other. Given that molecules have distinc-
tive fragmentation patterns, mass spectrometry is capable of providing essential
structural information to unknown components. Consequently, the power of
this technique lies in the production of mass spectra from each of the analytes
detected instead of merely an electronic signal that varies with the amount of
analyte. It therefore is a powerful tool for the quantitation of analytes and more
importantly for determining chemical and structural information about mole-
cules. Nonetheless, the high cost of instrumentation has limited the wide use of
this detector as compared to the less expensive GC detectors (Hobbs, 2001).
5.2.2. Gas Chromatography Coupled to Olfactometry
An interesting approach that has been used mainly for the determination of
odor-active compounds in food is gas chromatography-olfactometry (GC-O).
First introduced by Fuller et al. (van Ruth, 2001), GC-O uses the human nose
as a detector and has proven to be a valuable method for monitoring the pres-
ence of an odorant in the effluent of a gas chromatograph. Initially, the GC
Odor Measurement 27
effluent was sniffed and when an odor was perceived a description was given for
each retention time, thus associating odor activity with eluting compounds. The
advantages of coupling GC with olfactometry instead of conventional physi-
cal detectors is that although the latter provide relevant information on volatile
composition, many of them are not as sensitive for odor-active compounds as
the human nose (Ferreira et al., 2003).
Furthermore, GC-O allows differentiation between those volatile compo-
nents with a scent from those that do not have one. Early GC-O devices had seri-
ous problems of reproducibility caused by the discomfort from sniffing hot, dry
effluent gases. In an attempt to minimize the discomfort of sniffing, Dravnieks
and O’Donnell (1971) published a GC-O design in 1971, according to which the
hot GC effluent was combined with humidified air to reduce nasal dehydration.
In general, it is very difficult to judge the sensory relevance of volatiles from
a single GC-O run and the method is limited to the screening of odor-active
volatile compounds, unless the chemical stimuli and the assessors’ responses are
quantified (van Ruth and O’Connor, 2001). It should be kept in mind that in
GC-O, single compounds are assessed, and consequently this method does not
provide information on their behavior in a mixture, although it may indicate
their relevance to odor.
5.3. Direct Reading Instrumental Analysis: Colorimetric Detector Tubes
A colorimetric tube is also known as a “length of stain” tube, due to the fact
that the concentration of the chemical being tested produces a color stain in
the analytical material, proportional to the concentration in the air. These tubes
also are called Dräger tubes®, named after the company that introduced them in
1937 (Drägerwerk AG). Colorimetric detector tubes are flame-sealed glass tubes
containing a chemical reagent that reacts with a specific compound or group of
compounds, causing color change. A sample is collected by attaching the detec-
tor tube to a special bellows-type pump that draws a known volume of air with
each stroke. If the target chemical is present, the reagent in the tube changes
color and the length of the color change typically indicates the measured con-
centration. Sampling times are generally short and equipment is portable and
relatively inexpensive, hence they can be used as quick indicators of poor con-
trol to assess a short-term event or for scoping studies. Because they are simple
to use, detector tubes have been used in a broad range of applications. Direct-
reading colorimetric detector tubes also were designed for use in testing workplace
air and for determining compliance with occupational exposure limits, but they
also are used widely to demonstrate compliance with emission limits for spe-
cific substances. Today a considerable number of tubes are available from several
manufacturers covering a wide range of substances, although it is not possible to
differentiate between different chemical species within a generic group. Despite
the simplicity of the method, correct use is essential or results can be highly
misleading. Further, more interference with other chemical species is possible,
28 E. Psillakis
leading to false results. It should be mentioned here that colorimetric tubes are
not suitable for identifying unknown substances, and the correct tube needs to
be selected for a particular situation. Although the limits of detection for this
technique are typically between 0.2 to 1 ppm, detection limit can vary greatly,
depending on the tube used (IPPC H4-Part 2, 2002). According to the manufac-
turers’ literature, the relative standard deviation of detector tubes varies between
5 and 20%. Unused tubes must be stored correctly (some require refrigeration)
and have a finite shelf life, which must be observed.
5.4. Direct Reading Instrumental Analysis: Portable Analyzers
A range of different types of portable instruments suitable for odor measure-

ment are currently available.
5.4.1. Portable Gold Leaf Analyzers
Portable gold leaf analyzers are frequently used to monitor hydrogen sulfide in
the gas phase. This type of detector utilizes the change in resistance of a gold
film sensor caused by adsorption of H2S molecules, within an output propor-
tional to the H2S concentration. Eventually the gold leaf becomes saturated and
has to be regenerated. A common gold film monitor is the Jerome 631-X H2S
analyzer (Able Instruments and Controls Ltd.), which has a reported detection
limit of the order of 0.003 ppm and can measure up to 0.005 ppm H2S. Sampling
and measurement time depend on the level of sulfides present, but typically it is
less than 1 min.
5.4.2. Paper Tape Monitors
Paper tape monitors contain a chemically impregnated tape, which when exposed
to a gas sample changes color in direct proportion to the amount of gas present.
A tape is selected that will react with the gas of interest (IPPC H4-Part 2, 2002).
Today a wide range of compounds can be quantified depending on the instru-
ment selected. A particular advantage is that the monitor can be set to sample
at regular intervals, exposing an unreacted section of the tape each time, and
so leaves a permanent record of the concentration of each sample. However,
given that the sampling times can be of the order of minutes rather than sec-
onds, this type of monitor is not recommended when a large number of samples
are required in a short period of time. Furthermore, the equipment is generally
expensive, and although relatively simple to operate, sufficient information relat-
ing to process parameters/activities at the time of measurement must be col-
lected, thus allowing correct interpretation of the concentration data obtained.
Some instruments are affected by the presence of moisture, which limits their use
for stack monitoring.
Odor Measurement 29
5.4.3. Portable Gas Chromatophy
Portable GCs also exist and can be used for “fingerprinting,” i.e., to analyze
air samples at the complainant’s location in order to ascertain the identity and
concentration of the main odorous components (Santos and Galceran, 2003).
However, the cost of the instrument and the expertise required for analysis
and subsequent evaluation also limit its use as a “quick check” method for
everyday use.
Portable GCs can be coupled to a variety of detectors (including MS, FID,
PID, etc.) (Santos and Galceran, 2002). It should be mentioned here that these
detectors may be used independently as portable instruments for detecting dif-
ferent groups of compounds in air samples.
5.4.4. Flame Ionization Detector
Flame ionization detector (FID) is perhaps the most commonly used detector
for odorants, yielding analytical methodologies with a large dynamic range and
limits of detection in the low nanogram range (Hobbs, 2001). The FID detector
is suitable for odorants mostly composed of hydrogen and carbon, and is not
recommended for sulfides, where sensitivity was found to be less satisfactory.
Inside the FID, fuel (H2) and oxidant (O2 in air) are mixed to create and main-
tain a flame. The gaseous samples entering are burned and charged particles are
formed in that combustion process. This creates a current between the detector’s
electrodes, which is measured by the FID.
5.4.5. Photoionization Detector
This device uses ultraviolet light as a means of ionizing an analyte. The ions
produced by this process are collected by electrodes and the current generated
therefore is a measure of the analyte concentration. PID detectors are used to
analyze volatile organic compounds as well as ammonia and hydrogen sulfide.
Given that the PID responds to compounds with a photoionization potential
equal to or less than that of the energy of the source, appropriate setting of the
lamp voltage will exclude direct detection of the major compounds in the air
including water vapor, methane, and carbon dioxide (Hobbs et al., 1995). PID
detectors have a dynamic detecting range of between a few ppb and 10,000 ppm
and offer instant indicative measurements from typically rugged, lightweight,
handheld instrument.
Given that only a small fraction of the analyte molecules are actually ion-
ized in the PID chamber, it is considered to be a nondestructive GC detector.
Therefore, when the inlet of the PID detector is connected to a portable GC,
its exhaust port can be connected to another detector in series such as an FID
or an electron capture detector (ECD). A schematic diagram of a portable GC
equipped with in-series detectors [PID, FID, and dry electrolytic conductivity
30 E. Psillakis
Figure 3. Schematic of PID, FID, and dry electrolytic conductivity detector (DELCD) in series with
a SRI-8610C portable gas chromatograph. (With permission from Koziel et al., 1999).
(DELCD)] is given in Figure 3 (Koziel et al., 1999). In this way data from
different detectors sensitive to different group of analytes can be taken simul-
taneously providing almost real-time analysis and speciation of a wide range
of compounds. It should be mentioned here that the major challenge in this
instrumental setup is to make the design of the ionization chamber and the
downstream connections to the second detector as low volume as possible so
that peaks that have been separated by the GC column do not broaden out
before detection.
5.4.6. Fourier Transform Infrared Spectrometry
Fourier transform infrared spectrometry (FTIR) is suited to both quantitative

and qualitative measurement. In general, infrared (IR) absorption spectros-
copy is the measurement of the wavelength and intensity of the absorption of
mid-infrared light by a sample. Mid-infrared light (4000–200 cm−1) is energetic
enough to excite molecular vibrations to higher energy levels. The wavelength of
many IR absorption bands is characteristic of specific types of chemical bonds.
Technically, FTIR is well suited to serve as an “optic” nose (van Kempen et al.,
2002).
FTIR instruments are portable and can be set up to monitor either stack
emissions or boundary emissions. In addition, these instruments are adequately
sensitive to pick up odorants of interest (Ni and Heber, 2001). Part of the devel-
opment of these technologies for food industry odor purposes is the identification
of which wavelengths (compounds) meaningfully contribute to the instrument’s
response.
Odor Measurement 31
6. SENSORY METHODS
Sensory measurements employ the human nose as the odor detector. In this sense,
measurements are directly related to the properties of odors as experienced by
humans and on the relationship between psychological and physical attributes
of odor. The problems of complex mixtures, interactions between components,
and detectability below the threshold of smell become irrelevant as the “total
effect” of the overall odor is measured.
There are many factors other than the properties of the odor sample itself
that may influence the perception of an odor. Principal among these is the vari-
ability in the sense of smell between different observers. This may be overcome
to a certain extent by using a panel of several observers, with the result being
expressed as a measure of the central tendency of the individual results. For
repeatable results, great care must be taken in the presentation of samples to
observers. Factors such as the order in which samples are presented, the envi-
ronment in which the testing takes place, and the flow rate of the carrying gas
stream are all important.
Sensory evaluation techniques can be divided into two categories (Gostelow
et al., 2001):
Subjective measurements in which the nose is used without any other

equipment.
Objective measurements that incorporate the nose in conjunction with
some form of dilution apparatus (olfactometry).
6.1. Subjective Measurements
Subjective evaluation measurements have the advantage of being quick and rela-
tively low cost, given that no special equipment is required. Interpretation of
results is difficult and subjective measurements should be handled with caution
due to the inherent variation in odor perception even for well-trained personnel.
Parameters that may be measured subjectively include odor character, hedonic
tone, and intensity. Indeed, for character and hedonic tone, there are no objective
techniques available with the possible exception of the electronic nose (Laing
et al., 1992; Gostelow et al., 2001).
An example of subjective evaluation measurements is the direct scal-
ing technique, which is actually the oldest technique for measuring sensory
stimuli. According to this method, the odor panelist assigns a number to
the odor-containing sample, relative to the referenced standard. For exam-
ple, an odor sample might be compared to a sample of butanol, a typical
reference standard. This provides a good comparison of odor intensity, but
there is no measure of irritation or how objectionable the odor is (Wise
et al., 2000).
32 E. Psillakis
6.2. Objective Measurements
Objective measurements incorporate the nose in conjunction with an instrument

that dilutes the odor sample with odor-free air, usually termed an olfactometer.
In general, dilution may be static or dynamic. Static dilution involves the mix-
ing of fixed volumes of odorous and odor-free air, whereas dynamic dilution
involves the mixing of known flows.
An example of static dilution is the syringe dilution method (ASTM
Method D1391-78), which is also the oldest method of vapor dilution. Accord-
ing to this method, a measured volume of odorous air is transferred from a
graduated hypodermic syringe to a dilution syringe. Odor-free air is then drawn
into the dilution syringe to make the required dilution. The diluted mixture is
then expelled into the observer’s nose. Dilutions to threshold are determined by
varying the dilution ratios. This method was withdrawn as an official ASTM
method in 1986.
Dynamic dilution is superior to static dilution as the effects of sample adsorp-
tion to the internal surfaces of the instrument are minimized (Dravnieks and Jarke,
1980). An additional advantage of dynamic dilution/olfactometers is that the sam-
ple can be delivered to the sniffing port at a constant flow, a factor that has been
shown to improve repeatability of results (Schulz and van Harreveld, 1996).
6.3. Dynamic Olfactometry
There are two categories of dilution-related measurement techniques. The most

common is threshold olfactometry, where the sample is successively diluted until
it can just be detected (i.e., the threshold concentration). The concentration is
then expressed as the number of dilutions required to achieve the threshold con-
centration. Another form of dilution-related measurement is when the sample
odor is compared to a reference odor and the result is expressed as an equivalent
concentration of the reference gas (also termed suprathreshold olfactometry).
The sample or reference odor is diluted until the perceived intensity of each
stream is the same. In both cases, the use of an olfactometer removes (or at least,
reduces) any subjectivity from the measurement.
In general, there are two modes in common use for observers to indicate
whether an odor can be detected at a particular dilution and these are the yes/no
and the forced choice methods (Sneath, 2001). In the yes/no mode, either neu-
tral gas or diluted odor passes from the single port and the panelist is asked to
respond yes or no as to whether they can detect an odor. The panelists are aware
that in some cases blanks will be presented and the samples may be presented
either randomly or in order of increasing concentration. The other method of
dynamic olfactometry is the forced-choice method. In this system, there are two
or more sniffing ports, one of which delivers the diluted odorous sample and the
other(s) odor-free air. The port carrying the odorous sample is chosen randomly
during a test. The panelist is “forced” to choose which port contains the odor
Odor Measurement 33
(hence, the name forced-choice method), using a keyboard. They also indicate
whether their choice was a guess, whether they had an inkling, or whether they
were certain they chose the correct port. Only when the correct port is chosen
and the panelist is certain, the choice is considered to be True.
6.4. Factors Affecting Olfactometry Panel Performance
There are a number of variables involved in olfactometry that will affect measure-
ments. Fatigue from continued exposure to an odor may affect a panelist’s sense of
smell. This phenomenon is called adaptation and may reduce both perceived odor
intensity and perceived odor quality (Dravnieks and Jarke, 1980; Gostelow et al.,
2001). The degree of adaptation will depend on the odor concentration experi-
enced. In general, the weaker the odor concentration of an air sample, the more
does adaptation affect perceived strength. On the other hand, exposure of panel-
ists to strong odors may result in adaptation and affect detection of subsequent
weak odors. Consequently, the presentation schedule of odorous samples may
influence the results. It should be mentioned that in cases where dilutions occur
in a strict order, panelists will begin to expect subsequent samples to be weaker or
stronger and may adjust their responses accordingly. Descending order of presen-
tation may result in olfactory fatigue/adaptation and may obstruct detection of a
weak odor after exposure to a strong odor. In general, presentation of one dilution
series is given either in an ascending or random order of stimuli, thus restraining
adaptation (Dravnieks and Jarke, 1980; Gostelow et al., 2001).
Another beneficial effect of using an ascending schedule of sample presen-
tation is that it also may minimize the effects of adsorption/desorption of mate-
rials commonly leading to sample contamination or alteration. In general, the
use of nonreactive, odor-free materials, minimization of internal surface areas,
and the provision for flushing or easy replacement of flow lines between samples
further reduces problems of adsorption/desorption.
It should be mentioned that the supply of odor-free air is important as any
residual odor in the dilution air would bias the result. The general requirements
for the environment for observations by assessors can be found elsewhere (EN
13725: 2003).
When dynamic olfactometry is used, the flow rates at the sniffing ports can
have a major influence on the reported odor concentration. It should be pointed
out that the airflow is one of the most controversial issues in the effort to reach
international consensus on the standardization of odor measurement.
Other important elements of the test procedure that may influence results
(EN 13725:2003) are the shape/dimensions of the of smelling chamber (olfac-
tometer–nose interface), position of valves of gauge readings on the olfactom-
eter, differences in appearance of sniffing ports where more than one is used,
or even the responses of the operator. It therefore is beneficial if panelists are
isolated from each other, from the olfactometer controls, and from the operator
as shown in Figure 4.
34 E. Psillakis
Figure 4. Representation of a dynamic dilution olfactometer using the ODILE™ of the ODOTEC
Inc. (Montreal, Canada).
Sensitivity to odors is variable between different individuals (Bliss et al.,

1996). Olfactory responses of individuals vary with age. Increasing age is cor-
related with decreasing acuity in odor perception. It is generally accepted that
only persons between 16 and 60 years of age with a normal sense of smell should
be included on an odor panel. Female panelists normally have a greater sensi-
tivity than male panelists from the same age group. It also has been reported
that smokers have less sensitivity than non-smokers. Factors such as health (e.g.,
cold, nasal allergy), personality, education background, and training also may
contribute in some degree to the ability to assess an odor.
It is expected that different observers will report different odor concentra-
tions for the same sample. This effect is minimized by the use of a panel of several
observers and recording the average response. However, a large panel increases
the overall cost of measurement and the time taken to determine sample concen-
trations. Typically, between four (minimum according to EN 13725:2003) and
ten observers (Brennan, 1993) are used with eight being common (Bliss et al.,
1996).
6.5. Portable Sensory Devices
An example of a portable device for field sensory measurements that uses the
dilution to threshold approach is the scentometer, originally manufactured
by Barnebey-Cheney Company; it consists of a simple, handheld odor dilu-
tion device used to measure odor concentration in the field. According to this
technique, the person taking measurements breathes through the scentom-
eter. Gases can pass either directly to the nose or pass through an activated
carbon filter. The analyst chooses dilution factor by selecting the size of the
Odor Measurement 35
hole passing unfiltered air. The advantages of scentometer include portability,

simplicity of use, and ability to give immediate values for odor concentration
and intensity. However, it is difficult to use without experiencing odor fatigue
with increased exposure to samples. The scentometer method has been super-
seded largely by dynamic dilution/olfactometry, although it sometimes may be
used in field studies as it is specifically intended for this purpose (Ritter, 1989;
Sweeten and Miner, 1993).
A more recent device, based on the design of a scentometer, is the nasal
ranger (St. Croix Sensory, Inc. of Lake Elmo, Minnesota), combining the porta-
bility and low cost of a scentometer with the sampling control of more expensive
laboratory olfactometers. This device relies on a pressure transducer to ensure
that assessors maintain the required inhalation rate for the unit (Newby and
McGinley, 2003; Sheffield et al., 2004).
Field measurements have some drawbacks, the most important of which
is that results may be biased due to the fact that the assessor may anticipate
the smell or become desensitized to the odorous air before taking the mea-
surement. These methods also are difficult to verify when it involves only one
person. Some researchers have improved the accuracy of scentometry by using
respirators to avoid odor desensitization as well as by increasing the number
of assessors, thus averaging several measurements for a single observation
(Sweeten and Miner, 1993).
6.6. Standardization of Olfactometry
In the absence of standardized procedures, reported odor concentration levels

simply reflect the experience of the operator, the design of the olfactometer, its
operational mode (manual or automatic), its mixing method, the flow rate pre-
sented to panelists, and the number of panelists employed. Recognition of these
problems has led to the development of standards for olfactometry.
Although olfactometry has been used or researched since the 1890s, it
was not until 1970 that olfactometry was used for the detection of odorants
in environmental samples. The direct consequence was its use in regulatory
issues. The first standards providing a general definition of the method and
process parameters were the VDI 3881 (odor threshold determination, Germany,
1980) and the ASTM E679–91 (determination of odor and taste threshold by
a forced-choice ascending concentration series method of limits, USA, 1991).
The main problem with these standards is that they failed to define a required
level of accuracy of the results. As a result the differences between labora-
tories remained considerable and olfactometry was treated as somewhat of
a black art, escaping the rigor and discipline of statistics and experimental
design (Table 3).
The leaders in the rigorous standardization of olfactometry were
the Dutch in the 1980s, who introduced the Dutch olfactometry standard
36 E. Psillakis
Table 3. Comparison of the Basic Process Parameters of Different Standard Methods
ASTM A&WMA NVN 2820 VDI 3881 EN13725

E679-91 EE-6
Dynamic dilution
Triangular forced-choice method
Binary forced-choice method
Yes/no method
Ascending concentration series
Integral odor free dilution air
Integral dry dilution air

Presentation flow rate (lpm) 5–10 20 20 20
Presentation face velocity (m/s) 0.02–0.05 0.09–0.26 0.2–0.5
Presentation mask diameter (cm) 6–10 4–7 3–5
NVN2820 (Provisional standard: Air quality. Sensory odor measurement

using an olfactometer, Netherlands, 1996). Driven by the need to introduce
a strongly quantitative odor regulatory they were the first to work toward the
standardization of olfactometry which would improve the consistency within a
laboratory (repeatability) as well as between laboratories (reproducibility) and
the closeness of agreement of results to a known reference (accuracy) (Schulz
and van Harreveld, 1996). This standard was subsequently used as the basis
for a European standard for olfactometry EN13725 (air quality-determination
of odor concentration by dynamic olfactometry, Europe, 2003, draft on 1999).
The European standard has become the official olfactometry odor analysis
approach in many countries and dramatically reduced the differences within
and between laboratories that previously existed.
The Air and Waste Management Association’s EE-6 Odor Committee (Sub-
committee on the Standardization of Odor Measurement prepared a document
titled “Guidelines for Odor Sampling and Measurement by Dynamic Dilution
Olfactometry,” USA, 2002) has forwarded its guidelines to the American Society
of Testing Materials (ASTM) as a suggested replacement for ASTM Method
E679–91. These guidelines are similar to the European Standard but they do
allow quite a bit of flexibility in what olfactometer flow rates can be used. This
could potentially be a problem when attempting to compare data and results
from different olfactometry laboratories.
As can be seen, there clearly is a strong need for compatibility between the
odor measurement methods. It is clear that performance-based standardiza-
tion is far easier to implement and practice than equipment-based approaches.
The widespread acceptance of forced-choice dynamic olfactometry as a pre-
ferred measurement method is encouraging, although a practical internation-
ally accepted standard for olfactometry will not occur until testing results are
anchored to a standard reference material.
Odor Measurement 37
A Area
D50 50% of a population that can detect as a sensory stimulus
D/T Dilution to threshold ratio
DELCD Dry electrolytic conductivity
EROM European reference odor mass
ECD EROM European reference odor mass
FID Flame ionization detector
FTIR Fourier transform infrared spectrometry
GC Gas chromatography
GC-O Gas chromatography-olfactometry
LD50 Lethal dose for 50% of a well-defined test population
MS Mass spectometer
OC Odor concentration
OER Odor emission rate
ouE European odor unit
PID Photoionization detector
ppb Parts per billion
ppm Parts per million
ppt Parts per trillion
Q Gas flow or ventilation rate
SOER Specific odor emission rate
TON Threshold odor number
VOC Volatile organic compound
8. REFERENCES
Abbott, J. (1993) “Enclosed wastewater treatment plants-health and safety considerations.” Founda-
tion for Water Research Report FR/W0001.
ASTM Method D1391-78 “Standard method for measurement of odor in atmospheres (dilution
method).” American Society for Testing and Materials (ASTM International) Standards, with-
drawn 1986.
Bockreis, A., Jager, J. (1999) “Odor monitoring by the combination of sensors and neural networks.”
Environmental Modeling and Software, 14, 421–426.
Bonnin, C., Laborie, A., Paillard, H. (1990) “Odor nuisances created by sludge treatment: problems
and solutions.” Water Science and Technology, 22, 65–74.
Bliss, P. J., Schulz, T. J., Senger, T., Kaye, R. B. (1996) “Odor measurement—factors affecting olfac-
tometry panel performance.” Water Science and Technology, 34, 549–556.
Bliss, P.J., Jiang, J.K., Schultz, T.J. (1995) “The development of a sampling technique for determina-
tion of odor emission rate from areal surfaces: II. Mathematical Model.” Journal of Air and
Waste Management Association. 45, 989–994.
Brennan, B. (1993) “Odor nuisance.” Water and Waste Treatment, 36, 30–33.
Dravnieks, A., Jarke, F. (1980) “Odor threshold measurement by dynamic olfactometry: significant
operational variables.” Journal of Air Pollution Control Association, 30, 1284–1289.
38 E. Psillakis
Dravnieks, A., O’Donnell, A. (1971) “Principles and some techniques of high-resolution headspace
analysis.” Journal of Agricultural and Food Chemistry, 19, 1049–1056.
EN 13725(2003). Air Quality-Determination of Odor Concentration by Dynamic Olfactometry. CEN,
Brussels.
Ferreira, V., Pet’ka, J., Aznar, M., Cacho, J. (2003) “Quantitative gas chromatography–olfactometry.
Analytical characteristics of a panel of judges using a simple quantitative scale as gas chroma-
tography detector.” Journal of Chromatography A, 1002, 169–178.
Gholson, A.R., Albritton, J.R., Jayanty, R.K.M., Knoll, J.E., Midgett, M.R. (1991) “Evaluation of
the flux chamber method for measuring volatile organic emissions from quiescent liquid sur-
faces.” Environmental Science and Technology, 21, 519–524.
Gostelow, P., Parsons, S.A. (2000) “Sewage treatment works odor measurements.” Water Science and
Technology, 41(6), 33–40.
Gostelow, P., Parsons, S.A., Stuetz, R.M. (2001) “Odor measurements for sewage treatment works.”
Water Research, 35, 579–597.
Hobbs, P.J., Misselbrook, T.H., Pain, B.F. (1995) “Assessment of odors from livestock wastes by a
photoionisation detector, an electronic nose, olfactometry and gas-chromatography-mass spec-
trometry.” Journal of Agricultural Environmental Research, 60, 137–144.
Hobbs, P. (2001) “Odor analysis by gas chromatography.” In. Stuetz, R.M., Frechen, F.-B. (eds), Odours
in Wastewater Treatment: Measurement, Modeling and Control. London: IWA publishing.
Hwang, S. T. (1985) “Model prediction of volatile emissions: A comparison of several models for pre-
dicting emissions for hazardous waste treatment facilities.” Environmental Progress, 4, 141–144.
Jiang, J.K., Bliss, P.J., Schultz, T.J. (1995) “The development of a sampling technique for determina-
tion of odor emission rate from areal surfaces: I. Aerodynamic performance.” Journal of Air
and Waste Management Association, 45, 917–922.
Jiang, K., Kaye, R. (1996) “Comparison study on portable wind tunnel system and isolation chamber
for determination of VOCs from areal sources.” Water Science and Technology, 34, 583–589.
Jiang, J., Kaye, R. (2001) “Sampling techniques for odor measurement.” In Stuetz, R.M., Frechen,
F.-B. (eds), Odours in Wastewater Treatment: Measurement, Modeling and Control. London:
IWA publishing.
Klenbusch, M. R. (1986) “Measurement of gaseous emission rates from land surfaces using an
emission isolation flux chamber, user’s guide.” EPA/600/8-86/008; US Environmental Protection
Agency.
Koziel, J., Jia, M., Khaled, A., Noah, J., Pawliszyn, J. (1999) “Field air analysis with SPME device.”
Analytica Chimica Acta, 400, 153–162.
IPPC H4-Part2 (2002) Integrated Pollution Prevention and Control (IPPC) Draft: “Horizontal guid-
ance for odor: Part 2—Assessment and control,” Environment Agency. http://www.sepa.org.uk/
pdf/ppc/uktech/ippc_h4_pt2.pdf [2005 Jan 5]
Laing, D.G. Doty, R.L., Breipohl, W. (1992) The Human Sense of Smell. Springer-Verlag Berlin
Heidelberg.
Lindvall. T. (1970) “On sensory evaluation of odorous air pollutant intensities.” Nordisk Hygiejnisk
Tidsskrift, suppl. 2, 1–181.
Lockyer, D.R (1984) “A system for the measurement in the field of losses of ammonia through vola-
tilization.” Journal of the Science of Food and Agriculture, 35, 837–848.
Newby, B.D., McGinley, M.A. (2004) “Ambient odor testing of concentrated animal feeding opera-
tions using field and laboratory olfactometers.” Water Science and Technology, 50(4), 109–114.
Ni, J. Q., Heber, A. J. (2001) “Sampling and measurement of ammonia concentration at animal
facilities— A review.” Proceedings of the American Society of Agricultural Engineers Annual
International Meeting, Sacramento, California (Paper Number: 01–4090).
Pillonel, L., Bosset, J.O., Tabacchi, R. (2002) “Rapid preconcentration and enrichment techniques
for the analysis of food volatile. A review.” Lebensmittel-Wissenschaft und-Technologie, 35(1),
1–14.
Ritter, W.F. (1989) “Odor control of livestock wastes: State-of-the-art in North America.” Journal of
Agricultural Engineering Research, 42, 51–62.
Odor Measurement 39
Santos, F. J., Galceran, M. T. (2002) “The application of gas chromatography to environmental

analysis.” Trends in Analytical Chemistry, 21, 672–685.
Santos, F. J., Galceran, M. T. (2003) “Modern developments in gas chromatography–mass spectrom-
etry-based environmental analysis.” Journal of Chromatography A, 1000, 125–151.
Sarkar, U., Hobbs, S.E. (2002) “Odor from municipal solid waste (MSW) landfills: A study on the
analysis of perception.” Environment International, 27, 655–662.
Schulz, T. J., van Harreveld, A. P. (1996) “International moves towards standardization of odor
measurement using olfactometry.” Water Science and Technology, 34, 541–547.
Sheffield, R., Thompson, M., Dye, B., Parker D. (2004) “Evaluation of field-based odor assessment
methods.” Water Environment Federation/A&WMA Odors and Air Emissions Conference.
http://www.nasalranger.com/Media.cfm [2005 Jan 5]
Sneath, R.W. (2001) “Olfactometry and the CEN standard.” In. Stuetz, R.M., Frechen, F.-B. (eds),
Odours in wastewater Treatment: Measurement, Modeling and Control. London: IWA Publishing.
Sweeten, J.M., Miner, J.R., (1993) “Odor intensities at cattle feedlots in nuisance litigation.” Biore-
source Technology, 45, 177–188.
van Kempen, T.A.T.G., Powers, W.J., Sutton, A.L. (2002) “Fourier transform infrared (FTIR)
spectroscopy as an optical nose for predicting odor sensation.” Journal of Animal Science, 80,
1524–1527.
van Ruth, S.M., O’Connor, C.H. (2001) “Evaluation of three gas chromatography-olfactometry
methods: comparison of odor intensity-concentration relationships of eight volatile compounds
with sensory headspace data.” Food Chemistry, 74, 341–347.
van Ruth, S.M. (2001) “Methods for gas chromatography-olfactometry: a review.” Biomolecular
Engineering, 17, 121–128.
Wise, P.M, Olsson, M.J., Cain, W.S. (2000) “Quantification of odor quality.” Chemical Senses, 25,
429–443.
US EPA Method 0030 (1986) “Volatile organic sampling train.” Test Methods for Evaluating Solid
Waste, Physical/Chemical Methods (SW-846)
Young, P. J. (1984) “Odours from effluent and waste treatment.” Effluent Water Treatment Journal,
24, 189–195.
3
Preconcentration Prior to Gas
Chromatography
Elefteria Psillakis
1. PRECONCENTRATION TECHNIQUES
There are several obstacles impeding complete characterization of odorous

samples when using gas chromatography (GC). In many cases identification
remains ambiguous or questionable as a result of the presence of unknown
components at very low concentration levels. This is due to the fact that direct
injection of odorous samples into a GC is in most cases impossible due to the
low concentrations of malodorous compounds in these samples. This problem
can be overcome if a preconcentration step prior to analysis is introduced. In
general there are three possibilities for enriching components in an air sample:
absorbing the compounds in a suitable liquid, condensing them at low tempera-
tures (cryotrapping), and adsorbing them on a porous solid material.
In the case of liquid absorption, a liquid solution is used to absorb or react
with the target volatile compounds. An aliquot of the solvent solution containing
the entrapped odorous compounds is then injected into the analytical device for
further analyses. There are many drawbacks inherent to liquid absorption. A sol-
vent evaporation step is usually required, leading to large losses of the volatiles
(Pillonel et al., 2002). Furthermore, sample contamination due to contact with
glassware is possible; therefore, a thorough cleaning of all laboratory equipment
coming into direct contact with the sample is required. In cases where impingers
are used, the liquid containing the entrapped analytes may be subject to spillage,
and in these cases, a liquid trap should be used in order to prevent the solution
from getting into the pump. Finally, during analysis, interferences due to solvent
impurities also may obstruct quantification. It also should be mentioned here
that it is generally accepted that it is easier to precondition a solid phase than to
purify a liquid phase.
ELEFTERIA PSILLAKIS Department of Environmental Engineering, Technical University

of Crete, Polytechnioupolis, GR-73100 Chania, Greece e-mail: epsilaki@mred.tuc.gr
41
42 E. Psillakis
In cryogenic trapping, volatile compounds are trapped on an inert surface,

such as glass-fiber wool, glass beads, Tenax®, Porapak Q®, or even activated
carbon. According to this method, during sampling the trap is most frequently
immersed in liquid nitrogen (−196°C) or in liquid argon (−186°C) (Wardencki,
1998). The main advantages of cryogenic trapping versus other trapping tech-
niques are that there are (1) no artifact from thermal desorption, (2) no carryover
between runs, and (3) no breakthrough problem (Pillonel et al., 2002). However,
the additional equipment needed to handle the cryogen is quite expensive and
very sensitive to water. Furthermore, there are another two concerns when using
this preconcentration method. First, if the cooling temperature is too low, liq-
uid oxygen also may be trapped, which can readily oxidize organic compounds,
thus altering the composition of the sample. To overcome this problem, Peltier-
cooling devices can be used providing temperature control within the required
range for primary and secondary refocusing or concentrating a sample (Hobbs,
2001). Second, when samples are too large, then moisture can condense, forming
ice and blocking/restricting the flow of the compounds trapped in the sample.
To prevent clogging of the trap, water has to be removed efficiently from the
charged carrier gas before entering the trap (Pillonel et al., 2002).
More commonly preconcentration is achieved by sampling odorous com-
pounds on a porous material packed in a cartridge or in short columns. In gen-
eral, adsorbents have proved a successful and relatively inexpensive means of
trapping volatile analytes and the associated sorbent tubes are easy to condition
and small in size, facilitating collection, transport, and storage.
The most commonly used sorbents were recently reviewed by Harper (2000)
and the general types are inorganic adsorbents and porous materials based on
carbon and organic polymers. Various adsorbents may be used individually or
in combination and their selection depends on the compounds to be sampled
(e.g., concentration, species and mixtrure) as well as the boiling point involved.
The surface area of an adsorbent also has an impact on the amount of a given
substance that can be withheld by the medium as well as the surface polarity.
Carbon-based adsorbents with a large area are useful to trap very low-boiling
compounds, whereas it gets difficult to desorb substances with higher boiling
points. Porous polymers with a comparatively small surface area allow adsorp-
tion and desorption of lesser volatile components from gaseous samples. The
different characteristics of adsorbents show the need to carefully choose the
right adsorption material for given mixtures (Peng and Batterman, 2000). For
trapping volatile compounds with very different properties, multibed adsorbents
can be helpful. Typical combinations include Tenax TA® or graphitized carbon
and carbon molecular sieve (Harper, 2000). The weaker sorbent (Tenax TA®) is
placed first to trap the heavier molecules and the lighter compounds are retained
on the stronger sorbent located in second position. Desorption always takes
place in the reverse direction to the adsorption step (Pillonel et al., 2002).
Additional sorbent beds may be placed behind the primary bed to guard
against breakthrough. Breakthrough is the appearance of sampled molecules in
Preconcentration Prior to Gas Chromatography 43
the outlet stream, either because of saturation within the bed or displacement by
another chemical. Sampling is no longer efficient when this occurs, and as break-
through progresses the sample will be less and less representative of the external
environment (Harper, 2000). Another concern when using sorbent tubes is the
presence of moisture which inhibits and prejudices the profile of adsorbed
species. This can be overcome by using hydrophobic adsorbents, dry purges, and
other techniques (Peng and Batterman, 2000).
Once target compounds are collected on adsorbent, samples can be trans-
ferred to GC for further analysis. This is accomplished by either thermal or liquid
desorption. There are several reports where the ability of absorbents followed
by solvent or thermal desorption and chromatographic analysis to concentrate
volatile organic compounds (VOCs) from air samples has been demonstrated
(Hobbs et al., 1995; Rabaud et al., 2003; Peng and Batterman, 2000). Thermal
desorption is more popular than solvent extraction due to its speed, ease of
desorption, and minimization of artifact formation. It also offers higher sensi-
tivity given that the sample is not diluted as well as higher recoveries for polar
and reactive compounds which can pose problems for whole air samples. None-
theless, sample degradation due to contact with heated transfer line from the
desorber to the GC inlet is possible and carryover from previous samples can be
minimized or even avoided using short-path desorber systems.
Overall, the performance of a sorbent sampling/thermal desorption method
depends on many factors, including the target compounds (e.g., concentration,
species, and mixture), the method (e.g., sorbent selection, procedures for con-
ditioning, desorption, separation, and analysis of VOCs), and environmental
conditions present during sampling (e.g., temperature and humidity) (Peng and
Batterman, 2000).
A more recent trend in the preconcentration of odorous compounds involves
the use of the solid-phase microextraction (SPME) technique. SPME, developed
by Arthur and Pawliszyn in 1990, addressed the need for simple and rapid sample
preparation. The commercially available SPME unit consists of a short length
of narrow diameter fused-silica optical fiber externally coated with a thin film
of a polymeric stationary phase located inside a syringelike SPME holder. The
coated fiber is exposed to the sample where analytes preferentially partition by
adsorption or absorption (depending on the fiber type) from the sample to the
stationary phase and are concentrated. There are two main types of SPME sam-
pling: immersion sampling where the fiber is immersed into the aqueous solution
and headspace sampling where the fiber is exposed to the headspace above the
liquid (or solid) sample (Figure 1). Immersion sampling is widespread in the
SPME approach, but for volatile compounds the headspace mode is preferred as
it results in faster equilibration times and higher selectivity. After sampling for a
well-defined period of time, the fiber is withdrawn and transferred to the heated
injection port of a gas chromatograph (GC) or to a modified high-performance
liquid chromatography (HPLC) rheodyne valve depending on the target analytes
(Eisert and Pawliszyn, 1997).
44 E. Psillakis
(a) (b)
SPME
holder
SPME
fibre
Sample
Figure 1. Schematic representation of the two SPME sampling modes: (a) headspace and (b) immersion
in the case of liquid samples.
SPME rapidly gained wide acceptance incorporating sampling, extraction,

concentration, and sample introduction into a single solventless step, improv-
ing the method’s detection limits and saving preparation time. Given that the
amounts of analyte partitioned to the coating are proportional to their initial
concentrations in the matrix, SPME is particularly suited for air sampling, where
the effects of partitioning on initial analyte air concentration are negligible. Up
to now, SPME has been successfully applied to a wide variety of compounds
in the gas, liquid, and solid phase for the extraction of volatile and semivolatile
organic compounds from environmental, food, and biological samples (Kataoka
et al., 2000; Kataoka, 2002; Psillakis et al., 2003). Recently, the possibility of
using SPME as a viable tool to monitor trace quantities of odorous compounds
(such as volatile organic compounds and formaldehyde) has been demonstrated
(Béné et al., 2001; Davoli et al., 2003; Koziel et al. 1999).
It is generally accepted that despite the great technological advances there
are some major limitations inherent to gas chromatographic analyses of mal-
odorous air samples. The development of new analytical methodologies presents
a major challenge for many research laboratories around the world in that by
providing new, simple, and economic means for preconcentating odorous com-
pounds will outpace chromatographic analysis of odorous compounds.
GC Gas chromatography
HPLC High-performance liquid chromatography
SPME Solid-phase microextraction
VOCs Volatile organic compounds
Preconcentration Prior to Gas Chromatography 45
3. REFERENCES
Arthur, C.L., Pawliszyn, J. (1990) “Solid-phase microextraction with thermal desorption using silica
optical fibers.” Analytical Chemistry, 62(19), 2145–2148.
Béné, A., Fornage, A., Luisier, J.-L., Pichler, P., Villettaz, J.-C. (2001) “A new method for the rapid
determination of volatile substances: the SPME-direct method: Part I: Apparatus and working
conditions.” Sensors and Actuators B: Chemical, 72(2), 184–187.
Davoli, E., Gangai, M.L., Morselli, L., Tonelli D. (2003) “Characterisation of odorants emissions
from landfills by SPME and GC/MS.” Chemosphere 51(5), 357–368.
Eisert, R., Pawliszyn, J. (1997) “New trends in solid-phase microextraction.” Critical Reviews in
Analytical Chemistry 27(2), 103–135.
Harper, M. (2000) “Sorbent trapping of volatile organic compounds from air.” Journal of Chroma-
tography A, 885(1–2), 129–151.
Hobbs, P., Misselbrook, T. H., Pain, B. F. (1995) “Assessment of odours from livestock wastes by a
photoionization detector, an electronic nose, olfactometry and gas chromatography-mass spec-
trometry.” Journal of Agricultural Engineering Research, 60(2), 137–144.
Hobbs, P. (2001) “Odour analysis by gas chromatography.” In: Stuetz, R.M., Frechen F.-B. (eds),
Odours in Wastewater Treatment. London: IWA Publishing.
Kataoka, H., Lord, H.L., Pawliszyn, J. (2000). “Applications of solid-phase microextraction in food
analysis.” Journal of Chromatography A, 880(1–2), 35–62.
Kataoka, H. (2002) “Automated sample preparation using in-tube solid-phase microextraction and
its application – a review.” Analytical and Bioanalytical Chemistry 373(1–2), 31–45.
Koziel, J., Jia, M. Khaled, A., Noah, J., Pawliszyn, J. (1999) “Field air analysis with SPME device.”
Analytica Chimica Acta, 400(1–3), 153–162.
Peng, C-Y., Batterman, S. (2000) “Performance evaluation of a sorbent tube sampling method using
short path thermal desorption of volatile organic compounds.” Journal of Environmental Moni-
toring, 2(4), 313–324.
Pillonel, L., Bosset, J.O., Tabacchi R. (2002) “Rapid preconcentration and enrichment techniques for
the analysis of food volatile. A review.” Lebensmittel-Wissenschaft und-Technologie, 35(1), 1–14.
Psillakis, E., Mantzavinos, D., Kalogerakis, N. (2003) “Monitoring the sonochemical degradation
ofphthalate esters in water using solid-phase microextraction.” Chemosphere, 54(7), 849–857.
Rabaud, N.E., Ebeler, S.E., Ashbaugh, L.L., Flocchini, R.G. (2003) “Characterization and quanti-
fication of odorous and non-odorous volatile organic compounds near a commercial dairy in
California.” Atmospheric Environment, 37(7), 933–940.
Wardencki, W. (1998) “Problems with the determination of environmental sulphur compounds by
gas chromatography.” Journal of Chromatography A, 793(1), 1–19.
4
The Application of Intelligent
Sensor Array for Air Pollution
Control in the Food Industry
Saverio Mannino, Simona Benedetti, Susanna Buratti,
and Maria Stella Cosio
1. INTRODUCTION
One of the most significant improvements in the food industry during the next
few years is likely to be the development of intelligent sensor array systems that
can give useful information not only about the food quality characteristics but
also on the environment of the food production. For instance, the sensory array
system, often called “electronic nose,” would allow the monitoring of off-odors
and taints, often present at very low levels, or the odor quality of a food from the
raw material to the final product.
2. HISTORY OF THE INTELLIGENT SENSOR ARRAY
The earliest work on the development of an instrument specially to detect odors

probably dates back to Moncrieff in 1961. This was really a mechanical nose,
but the first electronic nose, based on redox reactions of odorants at an elec-
trode, was reported by Wilkens and Hatman in 1964. In 1965, Buck et al. and
Dravieks and Trotter worked on the modulation of conductivity by odorants
and modulation of contact potential by odorants, respectively. However, argu-
ably, the concept of an electronic nose as an intelligent chemical array sensor
system did not emerge until nearly 20 years later from publications by Persaud
and Dodd in 1982 at Warwick University in the United Kingdom. Their purpose
SAVERIO MANNINO, SIMONA BENEDETTI, SUSANNA BURATTI, AND MARIA STELLA

COSIO Department of Food Science, Technology and Microbiology, Univerisity of Milan, Via
Celoria 2, 20133 Milan, Italy; e-mail: saverio.mannino@unimi.it; simona.benedetti@unimi.it; susanna.
buratti@unimi.it; stella.cosio@unimi.it
47
48 S. Mannino et al.
was to model the current conception of the mammalian olfactory system by

demonstrating that a few sensors could discriminate among a larger number of
odorants. They constructed an array of three metal oxide gas sensors, which they
used to discriminate among 20 odorant substances, including essential oils and
pure volatile compounds.
In 1980, in North America, a group led by Stetter et al. started to build an
analytical portable instrument based on a chemical gas sensor array for the US
Coast Guard. This organization needed a portable instrument that would rap-
idly identify and measure volatile chemical vapors in emergency situations.
In Asia, scientists also were beginning to foresee the potential of a sensor
array. In 1984, Iwanaga et al. proposed an instrument employing an array of
metal oxide semiconductor sensors but suggested using simultaneous equations
to compute the relative concentrations of gases in a sample.
Then in 1989, a session at a NATO Advanced Workshop on Chemosensory
Information Processing was dedicated to the topic of artificial olfaction (Schild,
1990) and the design of an artificial olfactory system was further established
(Gardner et al., 1990). Finally, the first conference dedicated to the topic of an
electronic nose was held in 1992 (Gardner and Bartlett, 1992).
At the beginning of the 1990s, the term “artificial” or “electronic nose”
appeared. More extended research began, and applications, especially in food
industry, could be tested. Gardner and Bartlett (1993) defined the electronic nose
as “an instrument, which comprises an array of electronic chemical sensors with
partial specificity and appropriate pattern-recognition system, capable of recog-
nizing simple or complex odors” (p. 212).
3. SENSOR ARRAY COMPARED TO HUMAN NOSE
The sensor array is very far from the human nose, and according to Mielle et al.
(1995) this analytical system is “obviously electronic but not nose.” In fact
the only aspect in common with our odor-sensing organ is its function. There
are striking analogies between the human nose and the electronic nose. Compar-
ing the two is instructive. The human nose uses the lungs to bring the odor to
the epithelium layer; the electronic nose (E-nose) has a pump. The human nose
has mucus, hairs, and membranes to act as filters and concentrators, while the
E-nose has an inlet sampling system that provides sample filtration and condi-
tioning to protect the sensors and enhance selectivity. The human epithelium
contains the olfactory epithelium, which has millions of sensing cells, selected
from 100–200 different genotypes that interact with the odorous molecules in
unique ways. The E-nose has a variety of sensors that interact differently with
the sample. The human receptors convert the chemical responses to electronic
nerve impulses. The unique pattern of nerve impulses is propagated by neurons
through a complex network before reaching the higher brain for interpretation.
Similarly, the chemical sensors in the E-nose react with the sample and produce
The Application of Intelligent Sensor Array 49
electrical signals. A computer reads the unique pattern of signals and interprets
them with some form of “intelligent” pattern classification algorithm.
4. SENSOR ARRAY TECHNOLOGIES
Several commercial intelligent gas sensor array instruments are now available on
the market that cover a variety of chemical sensor principles, system design, and
data analysis techniques. Operationally, an E-nose is a “sensing system” com-
posed of three parts (Fig. 1): a sampling system (a); an array of chemical gas
sensors (b) producing an array of signals when confronted with a gas, vapor, or
odor; and an appropriate pattern-classification system (c).
The ideal sensors to be integrated in an electronic nose should fulfill the
following criteria (Bartlett et al., 1993): high sensitivity toward chemical com-
pounds, that is, similar to that of the human nose (down to 10−12 g/ml), low sen-
sitivity toward humidity and temperature; medium selectivity, they must respond
to different compounds present in the headspace of the sample; high stability;
high reproducibility and reliability; short reaction and recovery time; robust and
durable; easy calibration; easily processable data output; and small dimensions.
By chemical interaction between odor compounds and the gas sensors the
state of the sensors is altered giving rise to electrical signals that are registered
by the instrument. In this way the signals from the individual sensors repre-
sent a pattern that is unique for the gas mixture measured and is interpreted
by multivariate pattern recognition techniques like, for example, the artificial
neural network. Samples with similar odors generally give rise to similar sensor
response patterns and samples with different odors show differences in their
patterns. The sensors of an electronic nose can respond to both odorous and
odorless volatile compounds.
SENSOR RESPONSES
SINGLE SENSOR CHANGE IN
RESISTANCE
Figure 1. Basic elements of an electronic nose.

Various kinds of gas sensors are available, but only four technologies are
currently used in commercialized electronic noses: metal oxide semiconductors
(MOS); metal oxide semiconductor field effect transistors (MOSFET); con-
ducting organic polymers (CP); and piezoelectric crystals [bulk acoustic wave
(BAW)]. Others, such as fiberoptic (Dickinson et al., 1996), electrochemical
(Baltruschat et al., 1997), and bimetal sensors, are still in developmental stage
and may be integrated in the next generation of the electronic noses. In all cases
the goal is to create an array of differentially sensitive sensing elements.
4.1. Metal Oxide Semiconductors
Metal oxide semiconductor sensors (MOS) were first used commercially in the
1960s as household gas alarms in Japan under the names of Taguchi (the inven-
tor) or Figaro (the company’s name). These sensors rely on changes of conductiv-
ity induced by the adsorption of gases and subsequent surface reactions (Kohl,
1992). They consist of a ceramic substrate (round or flat) heated by wire and
coated by a metal oxide semiconducting film. The metal oxide coating may be
either of the n-type [mainly tin dioxide, zinc oxide, titanium dioxide, or iron (III)
oxide], which responds to oxidizing compounds, or of the p-type (mainly cobalt
oxide or nickel oxide), which responds to reducing compounds (Mielle, 1996).
The film deposition technique divides each sensor type into thin (6–1000 nm) or
thick (10–300 μm) film MOS sensors. The first one offers a faster response and
significantly higher sensitivity but is much more difficult to manufacture in term
of reproducibility. Therefore, commercially available MOS sensors often are
based on thick film technologies. Due to the high operating temperature (200–
650 °C), the organic volatiles transferred to the surface of the sensors are totally
combusted to carbon dioxide and water, leading to the change in the resistance.
MOS sensors are extremely sensitive to ethanol which blinds them to any other
volatile compound of interest.
4.2. Metal Oxide Semiconductor Field Effect Transistors
Metal oxide semiconductor field effect transistor (MOSFET) sensors rely on a

change of electrostatic potential. A MOSFET sensor comprises three layers: a
silicon semiconductor, a silicon oxide insulator, and a catalytic metal (usually
palladium, platinum, iridium, or rhodium), also called the gate. When polar
compounds interact with this metal gate, the electric field, and thus the current
flowing through the sensor, are modified. The recorded response corresponds
to the change of voltage necessary to keep a constant preset drain current
(Lundstrom et al., 1990). The selectivity and sensitivity of MOSFET sensors
may be influenced by the operating temperature (50–200°C), the composition
of the metal gate, and the microstructure of the catalytic metal. MOSFET
sensors have a relatively low sensitivity to moisture and are thought to be very
robust.
4.3. Conducting Organic Polymers
Conducting organic polymers (CP) sensors like MOS sensors rely on changes of
resistance by adsorption of gas. These sensors comprise a substrate (such as a fiber-
glass or silicon), a pair of gold-plated electrodes, and a conducting organic polymer
such as polypyrrol, polyaniline, or polythiophene as a sensing element. The polymer
film is deposed by electrochemical deposition between both electrodes previously
fixed to a substrate. When a voltage is passed across the electrodes, a current passes
through the conducting polymer (Amrani et al., 1995). The addition of volatile
compounds to the surface of the sensor alters the electron flow in the system and
therefore the resistance of the sensor. In general, CP sensors show good sensitivities
especially for polar compounds. However, their low operating temperature (<50 °C)
makes them extremely sensitive to moisture. Although such sensors are resistant to
poisoning, they have a lifetime of only about 9–18 months.
4.4. Piezoelectric Crystal Sensors
Piezoelectric crystal sensors are based on the change of the mass, which may be
measured as a change in resonance frequency. These sensors are made of tiny
disks, usually quartz, lithium niobate (LiNbO3), or lithium tantalite (LiTaO3),
coated with materials such as chromatographic stationary phase, lipids, or any
nonvolatile compounds that are chemically and thermally stable (Guilbault and
Jordan, 1988). When an alternating electrical potential is applied at room tem-
perature, the crystal vibrates at a very stable frequency, defined by its mechanical
properties. Upon exposure to a vapor, the coating adsorbs certain molecules,
which increase the mass of the sensing layer, and hence decrease the resonance
frequency of the crystal. This change may be monitored and related to the vola-
tile present. The crystal may be made to vibrate in a bulk acoustic wave (BAW)
or in a surface acoustic wave (SAW) mode by selecting the appropriate combina-
tion of crystal cut and type of electrode configuration. BAW and SAW sensors
differ in their structure: BAW are three-dimentional waves traveling through the
crystal, while SAW are two-dimentional waves that propagate along the surface
of the crystal at a depth of approximately one wavelength. These devices are
also called “quartz crystal microbalance” (QCM or QMB) because, similar to a
balance, their responses change in proportion to the amount of mass adsorbed.
Since piezoelectric sensors may be coated with an unlimited number of materi-
als, they present the best selectivity. However, the coating technology is not yet
well controlled, which induces poor batch-to-batch reproducibility.
5. DATA PROCESSING
The data processing of the multivariate output data generated by the gas sensor
array signals represents another essential part of the electronic nose concept.
The statistical techniques used are based on commercial or specially designed
software using pattern recognition routines like principal component analysis

(PCA), partial least squares (PLS), linear discriminant analysis (LDA), cluster
analysis (CA), fuzzy logic, or artificial neural network (ANN).
Principal components analysis (PCA) is a procedure that allows the extrac-
tion of useful information from the data and exploration of the data structure, the
relationship between objects, the relationship between objects and variables, and
the global correlation of the variables. It was used for explorative data analysis as
it identifies orthogonal directions of maximum variance in the original data, in
decreasing order, and projects the data into a lower-dimensional space formed of
a subset of the highest-variance components. The orthogonal directions are linear
combinations (principal components PCs) of the original variables and each com-
ponent explains in turn a part of the total variance of the data; in particular, the
first significant component explains the largest percentage of the total variance,
the second one, the second largest percentage, and so forth (Beebe et al., 1998).
Partial least square (PLS) regression models describe the dependence between
two variable blocks, e.g., sensor responses and time variables. Let the X matrix
represent the sensor responses and the Y matrix represent time and the X and Y
matrices could be approximated to a few orthogonal score vectors, respectively.
These components are then rotated in order to get as good a prediction of y vari-
ables as possible (Joreskogand and Wold, 1982).
Linear discriminant analysis (LDA) is among the most-used classification
techniques. The method maximizes the variance between categories and mini-
mizes the variance within categories. This method renders a number of orthogo-
nal linear discriminant functions equal to the number of categories minus one
(Meloun et al., 1992).
Cluster analysis (CA) performs agglomerative hierarchical clustering of
objects based on distance measures of dissimilarity or similarity. The hierarchy
of clusters can be represented by a binary tree, called a dendrogram. A final parti-
tion, i.e., the cluster assignment of each object, is obtained by cutting the tree at
a specified level (Gardner and Bartlett, 1992).
The artificial neural networks (ANN) are very sophisticated modeling tech-
niques capable of modeling extremely complex functions (Benedetti et al., 2004).
The basic unit of an ANN is the neuron. A neural network consists of a set of
interconnected network of neurons (Fig. 2). The input layer has one neuron for each
of the sensor signals, while the output layer has one neuron for each of different
sample properties that should be predicted. Usually one hidden layer with a vari-
able number of neurons is placed between the input and the output layer. During
the ANN training phase, the weights and transfer function parameters are adjusted
such that the calculated output value for a set of input value are as close as possible
to the unknown true value of the sample properties. The model estimation is more
complex than for a linear regression model due to the nonlinearity of the model.
Neural networks learn from examples through iteration, without requir-
ing a priori knowledge of the relationship among variables under investigation.
The neural network model mostly used was a multilayer perceptron (MLP) that
Hidden
layer Output
layer w1 Summation
input output
e.g. sensor e.g.conc w2 å
responses
class
w3 Tranfer-function
Weight multiplication
Neuron
Figure 2. Architecture of an artificial neural network (ANN).
learns using an algorithm called backpropagation. Since a neural network can

arrive at different solutions for the same data, if different values of the initial net-
work weights are provided, the network was trained several times. The goal was
to try and find a neural network model for which multiple training approaches
the same final mean squared error (MSE).
6. APPLICATIONS OF INTELLIGENT SENSOR ARRAY
There is a growing emphasis and consensus that an intelligent sensor array or

E-nose is most effective in the quality control of raw and manufactured prod-
ucts, for example, the determination of food freshness and maturity monitoring,
shelf-life investigations, authenticity assessments of products, and even micro-
bial pathogen detection and environmental control. Industries including those
of cosmetics, cars, packaging, tabacco, environment, and food monitoring all
have seen quite heavy investment in recent years.
The E-nose has the interesting ability to address analytical problems that have
been refractory to traditional analytical approaches: for example, the problem of
the volatile compounds from air pollution. The volatile compounds interact in
such a way that no single compound or group of compounds is associated with the
subjective assessment of odor. The odor of “bad” air cannot be traced to a specific
molecule or to a simple list of chemicals by traditional analytical techniques like
gas chromatography or mass spectroscopy. In such cases, the determination of
odor is performed by a panel of skilled and trained human noses that could find
differences in the odor of the samples. The E-nose can learn the fingerprints of air
and easily recognize the differences. The molecular basis for this difference is not
known exactly, but it is possible to affirm that the pattern responses are different
and related to the quality of “good” or “bad” air.
The E-noses have been used to assess livestock wastes (Hobbs et al.,
1995), the application of cattle slurry to grassland (Misselbrook et al., 1997),
airborne pollen levels (Kalman et al., 1997), and jet fuels in soil and water
samples (McCarrick et al., 1996). E-noses also are being applied to bioprocess
monitoring. Microrganisms have been employed increasingly in recent years to

biodegrade biological waste gases, and because of tightening restrictions on
odor emissions the need for reliable gas-monitoring methods has become criti-
cal. Researchers in Germany have demonstrated the use of E-nose techniques
for monitoring such waste-gas-treatment processes (Homan and Fodisch, in
press). Sensor arrays also have been employed in online bioreactor monitoring
by observing changes in the volatiles emitted from cultures during the biodeg-
radation process (Bachinger et al., in press).
Waste from many sources, such as animal factories and fertilizers, many
times is responsible for polluting rivers, streams, and underground water aqui-
fers. The E-nose provides a method for onsite characterization and measuring
the fingerprint of polluting samples (Edward, 1999).
7. THE FUTURE
There is a plethora of potential application for the electronic nose today. Odor
control is of increasing importance in our lives, for examples, in cars, trains,
aircraft, inside, and outside of buildings and factories. This environmental appli-
cation area is particularly important because the E-nose can be trained to rec-
ognize hazardous chemicals as well as odors. Furthermore, with respect to the
human nose, the E-nose does not fatigue as easily, can be placed in hazardous
atmospheres, is less costly, and can travel easily into outer space. It also holds the
promise of being much cheaper, smaller, and easier to use and maintain than a
mass spectrometer.
In conclusion, we hope that what has been written could be a contribution
to those in and allied to this field of work.
MOS Metal oxide semiconductors

MOSFET Metal oxide semiconductor field effect transistors
CP Conducting polymers
BAW Bulk acoustic wave
SAW Surface acoustic wave
QCM Quartz crystal microbalance
QMB Quartz microbalance
PCA Principal component analysis
PLS Partial least square
LDA Linear discriminant analysis
CA Cluster analysis
ANN Artificial neural network
PCs Principal components
MLP Multilayer perceptron

MSE Mean squared error
9. REFERENCES
Amrani, M.E.H., Persaud, K.C., Payne, P.A. (1995) “High-frequency measurements of conducting
polymers : development of a new technique for sensing volatile chemicals.” Meas. Sci. Technol.
6, 1500–1507.
Bachinger T., Martensson P., Mandenius C-F. (in press) Seminars in Food Analysis (vol.2), Chapman &
Hall.
Baltruschat, H., Kamphauser, I., Oelgeklaus, R., Rose, J., Wahlkamp, M. (1997) “Detection of volatile
solvents using potentiodymanic gas sensors,” Analytical Chemistry, 69, 743–748.
Bartlett, P.N., Blair, N. and Gardner, J.W. (1993) Electronic nose. Principles, applications and outlook.
ASIC 15e Colloque, Montpellier, 1993, 478–486.
Beebe K.R., Pell R.J., Seasholtz M.B. (1998) Chemometrics, a practical guide, John Wiley and
Sons, New York.
Benedetti, S., Mannino, S., Sabatini, A.G., Marcazzan, G.L. (2004) “Electronic nose and neural
network use for the classification of honey,” Apidologie 35, 397–402.
Buck, T.M., Allen, F.G., Dalton, M. (1965) “Detection of chemical species by surface effects on metals
and semiconductors,” in T. Bregman and A. Dravnieks (eds.), Surface Effects in Detection, Spartan
Books Inc.
Dickinson, T.A., White, J. Kauer, J.S. and Walt, D.R. (1996) “A chemical detecting system based on
a cross-reactive optical sensor array,” Nature, 382, 697–700.
Dravnieks, A. and Trotter, P.J. (1965) “Polar vapour detection based on thermal modulation of
contact potentials”, J. Sci. Instrum., 42, 624.
Edward, J. S. (1999) EST Public Report. Web site: www.estcal.com
Gardner, J.W. and Bartlett, P.N. (1993) “A brief history of electronic noses”, Sensor and Actuator
B, 18, 211–220.
Gardner, J.W. and Bartlett, P.N. (1992) Pattern recognition in odour sensing. In Sensors and Sensory
systems for an Electronic Nose. Dordrecht: Kluwer Academic Publishers, 212, 161–179.
Gardner, J.W., Bartlett, P.N., Dodd, G.H., Shurmer, H.V. (1990) “The design of an artificial olfactory
system,” in Schild D. (ed.) Chemosensory Information Processing, NATO ASI Series H: Cell
Biology, Vol. 39, Springer, Berlin,1990, 131–173.
Guilbault, G., Jordan, J.M. (1988) “Analytical uses of piezoelectric crystal: a review,” Crit. Rev.
Anal. Chem. 19, 1–28.
Hobbs, P.J., Misselbrook, T.H. and Pain, B.F. (1995) “Assessment of odours from Livestock wastes
by a photoionization detector, an electronic nose, olfactometry and gas chromatography-mass
spectrometry”, J. Agric. Eng. Res. 60, 137–144.
Homan, W.J., Fodisch, H. (in press) Seminars in Food Analysis (vol.2), Chapman & Hall.
Iwanaga S., Sato N., Isegami A., Noro T., Arima H., US pat. 4457161, 1984.
Joreskogand K.G. and Wold H. (eds.) (1982) Systems under indirect observation, Parts I and II,
North Holland, Amsterdam.
Kalman, E-L., Winquist, F. and Lundstrom, I. (1997) “A new pollen detection method based on an
electronic nose“, Atmos. Environ. 31, 1715–1719.
Kohl, D. (1992) “oxidic semiconductor gas sensors”, in Sberveglieri G. (ed.) Gas sensors. Dordrecht:
Kluwer Academic Publishers, 43–88.
Lundstrom, I., Spetz, A., Winquist, F., Ackelid, U. and Sundgren, H. (1990) “Catalytic metals and
field-effect devices useful combination”, Sens. Actuator B, 1, 15–20.
McCarrick, C.W., Ohmer, D.T., Gilliland, L.A., Edwards, P.A. and Mayfield, H.Y. (1996) “Fuel
Identification by Neural Network Analysis of the Response of Vapor-Sensitive Sensor arrays”,
Anal. Chem. 68, 4264–4269.
Meloun M., Militky J., Forina M. (1992) Chemometrics for analytical chemistry. Ellis Horwood,
New York.
Mielle, P. (1996) “Electronic noses: Towards the objective instrumental characterization of food
aroma,” Trends in Food Sci. Technol., 7, 432–438.
Mielle, P., Hivert, B. and Mauvais, G. (1995) “Are gas sensors suitable for on-line monitoring and quan-
tification of volatile compounds ?” In Etievant P. and Schreier P. (eds.) Proceedings of Bioflavour
95 Dijon: INRA editions, 81–84.
Misselbrook, T.H., Hobbs, P.J. and Persaud, K.C. (1997) “Use of an electronic nose to measure odour
concentration following application of cattle slurry to grassland”, J. Agric. Eng. Res. 66, 213–220.
Moncrieff, R.W. (1961) “An instrument for measuring and classifying odours”, J. Appl. Physiol.,16, 742.
NATO Advanced Research Workshop, Reykjavik, Iceland, August 1991, Gardner J.W. and Bartlett
P.N. (eds.), Sensors and Sensory Systems for an Electronic Nose, NATO ASI Series E: Applied
Science, Vol. 212, Kluwer, Dordrecht, 1992.
Persaud, K. and Dodd, G.H. (1982) “Analysis of discrimination mechanisms of the mammalian
olfactory system using a model nose” Nature, 299, 352–355.
Schild, D. (ed.). Chemosensory Information Processing, NATO ASI Series H: Cell Biology, Vol. 39,
Springer, Berlin, 1990.
Stetter J.R., Zaromb S., Findlay M.W. 1984. Sens. Actuator B 6, 269.
Wilkens, W.F. and Hatman, A.D. (1964) “An electronic analog for the olfactory processes”, Ann.
NY Acad. Sci., 116, 608.
5
Electronic-Nose Technology:
Application for Quality
Evaluation in the Fish Industry
Guðrún Ólafsdóttir and Kristberg Kristbergsson
1. INTRODUCTION
The odor of fresh fish is one of the most important quality parameters used to
determine whether fish is acceptable for consumption. The composition of volatile
compounds in fish contributing to the characteristic odors can be determined and
related to quality. Currently there is a need for rapid, automated, and objective
tools for process monitoring and quality assurance of perishable food products.
The possibility of using electronic nose for rapid quality control of fish therefore
is of interest. Knowledge about the composition of the headspace of different fish
products during storage is useful to guide the development of electronic noses for
quality monitoring of fish products. Complicated postmortem processes in the fish
are responsible for the loss of freshness and the onset of spoilage. The spoilage
processes are a combination of physical, chemical, biochemical, and microbio-
logical interactions that are species dependent and additionally various extrinsic
factors such as handling and different storage conditions will further influence the
spoilage pattern and the development of spoilage odors.
2. VOLATILE COMPOUNDS AS INDICATORS OF FISH QUALITY
The odor of fish can be classified as species-related fresh fish odor, microbial
spoilage odor, oxidized odor, processing odor, and environmentally derived
odor, based on the origin of the volatile compounds (Table 1).
GUðRÚN ÓLAFSDÓTTIR AND KRISTBERG KRISTBERGSSON Icelandic Fisheries

Laboratories, Skúlagötu 4, 101 Reykjavík, Iceland and Department of Food Science, University of
Iceland, Hjarðarhaga 2-6, Reykjavík Iceland
57
58 G. Ólafsdóttir and K. Kristbergsson
Because of the nonspecific selectivity and low sensitivity of gas sensors in

electronic noses it is necessary to correlate their responses to the main classes of
quality indicating volatile compounds of the products. Some compounds pres-
ent in concentrations below the detection limit of the gas sensors in electronic
noses may have an impact on the odor and can function as key odorants due to
their low flavor threshold.
Compounds contributing to characteristic fresh fish odors like alcohols,
carbonyls, and dienals causing oxidation odors are present in very low concen-
trations (ppb) but have an impact on the overall odor because of their low odor
threshold (see Table 1). Microbially formed degradation compounds like alco-
hols are present in much higher concentrations (ppm) in the headspace of
fish but their odor threshold is much higher so their impact on the odor is less
(Olafsdottir and Fleurence, 1998). Therefore, it cannot be assumed that the gas
sensors are measuring the overall odor, but rather detecting the components that
are in the highest concentration in the headspace, which may be of both odorous
and nonodorous nature. In some cases the compounds may have little or no effect
on the human olfactory system, but they may be detected by the gas sensors.
Characteristic odor and the progression of odor in fish during storage have
been associated with varying levels of different volatile compounds present in
the headspace of fish (Lindsay et al., 1986; Josephson et al., 1986).
SHORT CHAIN ALCOHOLS CONCENTRATION (PPM)

1000
SULFUR COMPOUNDS
600 30
FRESH FISH ALCOHOLS
CONCENTRATION (PPB)
500
400 20
SWEET ESTERS
AROMATICS
300
200 10
SHORT CHAIN
ALCOHOLS
100 5
FRESH FISH CARBONYLS
DIENALS
1 4 8 12 16 20 24 28 32
PUTRID
SWEET
TIME (DAYS)
FRESH
STALE
FLAT
Figure 1. Generalized changes in concentration of groups of influential aroma compounds and

sensory quality stages of refrigerated whitefish (from Josephson et al., 1986)
Electronic-Nose Technology 59
Table 1. Classes of Odors in Fish and Examples of Compounds Contributing to the

Odors (from Olafsdottir and Fleurence, 1998)
Class of
chemical Examples of Aroma Odor threshold
Fish odor species compounds description in water
Species- C6-C9 alcohols Hexanal / t-2-hexenal, Green, aldehyde-like 4,5ppb / 17ppb a

related and carbonyls 1-octen-3-ol, /1- Mushroom 10ppb/ 0,009ppba
fresh octen-3-one Heavy earthy,
fish odor 1,5-octadiene-3-ol mushrooms 10ppba
1,5-octadiene-3-one Geranium 0,001ppba
2,6-nonadienal Cucumber 0,001ppba
3,6-nonadienol Cucumber, melon-like 10ppba
Bromophenols 2,6-dibromophenol Iodine- and shrimp-like 0,0005μg/kgb
2,4,6-tribromophenol Saltwater fish, brine-like. 0,6μg/kgb
2-bromophenol Sea, marine-like flavor
N-cyclic Pyrrolidine piperidine Earthy
compounds
Short chain ethanol, propanol, Solvent like 1–100 ppmc
alcohols butanol, 3-methyl-1-
butanol
Microbial Short chain acetone, butanone Solvent like
spoilage carbonyls ethanal, propanal
odor 3-methylbutanal Malty 0,06ppm d
2-methylbutanal Malty 0,04ppm d
Amines ammonia, Ammoniacal 110 ppmc
TMA fishy, ammoniacal 30 ppmc
DMA 0,6 ppmc
histamine,
putrecine, Putrid, rotten
cadeverine
Sulfur hydrogen sulfide Sulfury, boiled eggs 5–40 ppb e
compounds methyl mercaptan Rotten, cabbage 0,05 ppb e
methyl sulfide Cabbage-like 0,9μg/kgf
dimethyl disulfide Putrid, onion-like 12 ppb g
dimethyl trisulfide Putrid, cabbage and 0.01ppbg
onion-Garlic like
bis-methylthio 0,3 μg/kgf
methane thioesters
Aromatics phenethyl alcohol Old roses
phenol, Phenolic, 2 ppm
p-cresol Pigpen-odors ,horse 300 μg/kgf
manure
N-cyclic indole Moth ball or fecal like
compounds skatole
acetic acid, 34,2ppmc
Acids butyric acid Sour, rotten, old socks 32,8ppmc
isobutyric acid
Oxidized Unsaturated hexanal green, planty 4,5ppbf
odor aldehydes c4-heptenal cardboard-like, potato-like 0,04ppbh
2,4-heptadienal, fishy oxidised flavor
2,4,7-decatrienal, burnt, fishy, cod-liver
oil-like
(Continued )
Table 1. Classes of Odors in Fish and Examples of Compounds Contributing to the

Odors (from Olafsdottir and Fleurence, 1998)—Cont’d
Processing 2,4-heptadienal and ripened anchovies

odors 3,5-octadien-2-one
methional boiled potato - like odor
2-methyl-3-furanthiol meaty odor in
canned tuna
Environ- methyl sulfide geosmin petroleum odors
mental 2-methyl-iso-borneol earthy, muddy odors
odors
a
Josephson (1991); b Whitfield et al. (1988); c Kawai T. (1996); d Sheldon et al. (1971); e Fazzalari (1978); f Whitfield
and Tindale (1984); g Buttery et al. (1976); h McGill et al. (1974).
Species-related fresh fish odors have been attributed to long-chain alcohols

and carbonyl compounds like 1,5-octadien-3-ol and 2,6-nonadienal, respectively,
which are oxidatively derived from polyunsaturated fatty acids such as eicosapen-
taenoic acid 20:5ω3 (Josephson et al., 1984). Spoilage odors develop as a result of
microbial activity and oxidative degradation of the fish components. Compounds
such as trimethylamine (Oehlenschläger, 1992), short-chain alcohols (Kelleher and
Zall, 1983; Ahmed and Matches, 1983), carbonyls, esters, and sulfur compounds
like hydrogen sulfide, methylmercaptan, dimethyl disulfide, and dimethyl trisulfide
are produced by microbial degradation of fish constituents (Herbert et al.,1975;
Kamiya and Ose,1984). Oxidation of fatty acids contributes to the rancid odors
of fish with the formation of aldehydes like hexanal, 2,7-heptadienal, and 2,4,7-
decadienal (McGill et al., 1974). All these compounds are to some degree volatile
and may be used to monitor freshness and spoilage of fish.
Both single compounds or a combination of compounds representing the
different changes occurring during spoilage have been suggested as indicators for
freshness and spoilage of fish (Lindsay et al., 1986). Ethanol, 3-methyl-1-buta-
nol, 2-methyl-1-propanol, 3-hydroxy-2-butanone, ethyl acetate, and butanoic
acid ethyl ester were the most abundant volatiles in the headspace of haddock
stored in ice (Olafsdóttir, 2003). Similar volatile compounds were found in cold-
smoked salmon during refrigerated storage (Joffraud et al., 2001; Jörgensen et al.,
2001). This is expected since similar profiles of microflora emerge in different
food products when subjected to the same conditions despite being heteroge-
neous initially (Gram et al., 2002). The volatile compounds detected in spoiled
cold-smoked salmon were mainly alcohols produced by microbial activity. Some
of the volatile compounds produced during spoilage of cold-smoked salmon
contributed to the spoilage off-flavor of cold-smoked salmon as confirmed
by gas chromatography-olfactometry. These were trimethylamine, 3-methyl
butanal, 2-methyl-1-butanol, 3-methyl-1-butanol, 1-penten-3-ol, and 1-propanol
(Jörgensen et al., 2001).
It is likely that the same set of sensors can be used for monitoring spoilage
changes in different fish products because similar volatile compounds emerge
in the products. However, because of the complexity of the spoilage processes

caused by the diversity of the microflora and the influence of temperature on
their spoilage potential, it is likely that models to predict the quality or shelf life
of the products based on electronic nose responses will have to be developed for
each product and the respective storage conditions.
3. INDUSTRIAL NEEDS FOR RAPID QUALITY EVALUATION

OF FISH
Although the importance of cultured fish is increasing the fish trade is largely
based on wild fish that has been harvested by various catching methods. Conse-
quently, the quality of fish as a raw material for food production is very variable.
Therefore, the need for rapid evaluation of fish quality is even greater than for
other food commodities where control of the total food chain is better.
Chemical measurements of total volatile bases and total microbial counts
are used in the industry to verify the freshness of fish according to regulations.
Alternative techniques have been developed to monitor changes occurring
postmortem in fish. Detection of ATP metabolites and physical measurements
evaluating changes in texture, microstructure, electrical properties, and color are
well established and newer technologies based on image analysis and spectro-
scopic methods have shown promising results (Olafsdottir et al., 1997c, 2004;
Di Natale, 2003). However, none of these methods have been widely imple-
mented in the industry. This still leaves sensory assessment as the leading method
for freshness evaluation in the fish industry to give the best description of the
actual consumer preferences. Sensory attributes related to appearance, texture,
smell, color, defects, and handling were considered very important in quality con-
trol according to a survey on the importance of various quality attributes of fish
and detection methods used in the European fish sector (Jørgensen et al., 2003).
The results showed general agreement regarding the need for rapid instrumental
methods to measure the overall concepts freshness and quality of fish. The need to
measure objectively the freshness or quality of perishable food products like fish
is of importance both in process management and in the trade of fish. Remote
buying via the internet has become more common and an objective method to
evaluate the quality would facilitate e-commerce. Additionally, the evaluation of
the freshness of various processed fish products is also of interest for the market-
ability of these products.
4. APPLICATION OF ELECTRONIC NOSES FOR FISH
Electronic noses have been suggested for various applications related to quality
evaluation of different food such as monitoring freshness and the onset of spoil-
age or bioprocesses of food. Many of these applications are based on detecting
volatile compounds produced from the growth of fungi, molds, or microbes or

changes occurring in food because of oxidation (Schnürer et al., 1999; Olsson
et al., 2002; Keshri et al., 2002; Pradyumna et al., 1998; McEntegart et al., 2000;
Boothe and Arnold, 2002).
In recent years attempts to use electronic nose technology to track the spoil-
age processes occurring in fish have been reported in numerous papers. Most
of these are feasibility studies, showing the capacity of the electronic nose to
discriminate between different spoilage levels or storage time of samples. Elec-
tronic nose instruments based on various sensor technologies have been used for
fish applications and compared with different reference methods like sensory
analysis, chemical measurements of volatile compounds like TVN or TMA, and
microbial analysis (Table 2).
Different catching methods and handling and storage conditions, in particu-
lar the temperature, will influence the spoilage pattern of fish, and consequently
information about storage time may give ambiguous information about the fresh-
ness stage of the fish (Di Natale et al., 2001). Therefore, a more useful approach is
to compare the electronic nose responses with a reference measurement that gives
information about the microbial and oxidative processes influencing the freshness
quality.
Gas sensors that currently are used in electronic noses are nonselective toward
individual compounds, but some of them are selective toward certain classes of
compounds. Research efforts have focused on developing membranes for selective
detection of important quality-indicating compounds like sensors aimed specifi-
cally at the detection of compounds contributing to freshness odors of fish (Deng
et al., 1996). Many researchers have developed sensors for the selective detection
of microbial metabolites like TMA (Storey et al., 1984; Egashira et al., 1990; Saja
et al., 1999; Zhao et al., 2002).
When monitoring the onset of spoilage of capelin during storage with an
array of electrochemical sensors, the NH3 sensor gave the best results to predict
TVN value of capelin (Olafsdottir et al., 2000). Ohashi et al (1991) evaluated
the performance of a semiconductive trimethylamine gas sensor (In2O3 -MgO)
to detect the freshness of cod, sardine, and yellow tail stored at 10 and 27 °C.
The same group reported further that even a single gas sensor was quite useful
to monitor fish freshness by the selective detection of TMA, DMA, and ammo-
nia (Egashira et al. 1994). Volatiles have been associated with the safety of fish
products like histamine which causes scombroid poisoning, a leading cause of
finfish-borne illness. Histamine traditionally has been used as an indicator for
safety. Selective detection of bioamines has been suggested using an amperomet-
ric detection of histamine with a methylamine dehydrogenase polypyrrole-based
sensor (Zeng et al., 2000).
The development of “smart packaging” has received attention recently.
The potential use of sensors based on pH-sensitive films for monitoring spoil-
age volatiles released into packaged fish headspace were described by Byrne et
al. (2003). The sensors were based on cresol red entrapped within a plasticized
Table 2. Examples of Application of Electronic Noses Based on Various Sensor Technologies to Detect Changes during Storage and
Processing of Different Seafood (adapted from Olafsdottir, 2003)
Type of fish/ application Sensors Instrument Reference method References
Haddock /freshness, storage time Metaloxide sensors FreshSense Sensory analysis Ólafsson et al. (1992)
Cod /freshness, storage time QMB LibraNose TVB-N Di Natale et al. (2001)
Electrochemical sensors FreshSense Sensory analysis Olafsdottir et al. (2003a)
Cod-fillets/ freshness, storage time QMB LibraNose Di Natale et al. (1996)
Electronic-Nose Technology
Sardine/ freshness, storage time Metalloporphyrins and polymers coated EnQbe Sensory (QIM), color, texture Macagnano et al. (2005)
thickness shear mode resonators:
Capelin / freshness, storage time Electrochemical sensors FreshSense TVB-N / sensory Olafsdottir et al.
(1997a,b, 2000)
Herring / freshness, storage time Electrochemical sensors FreshSense TVB-N / sensory Olafsdottir et al. (1998)
Cod roe / ripening stage Sensory
Redfish / storage time TMA / microbial count / sensory Olafsdottir et al. (2002)
Cod fillets / freshness, storage time TVB-N / microbial count / sensory Olafsdottir et al. (2002)
Herring fillets MOSFET and MOS (Taguchi) Lipid oxidation products Haugen and Undeland
Antioxidants/ sensory analyses (2003)
Tuna Conducting polymers AromaScan Microbial count / sensory Du et al. (2001)
Tuna Conducting polymers e-Nose 4000 Microbial count / sensory Newman et al. (1999)
Trout /freshness, storage time Amperometric sensors/ heating Prototype No reference Schweizer-Berberich
filaments et al. (1994)
Salmon fillets Conducting polymers e-Nose 4000 Sensory Luzuriaga and Balaban.
(1999b)
Salmon fillets / freshness, Conducting polymers AromaScan Microbial count / sensory Du et al. (2002)
storage time
Shrimp / freshness storage time Conducting polymers e-Nose 4000 NH3 - electrode / sensory Luzuriaga and Balaban
(1999a)
Smoked salmon / freshness Metal oxide sensors FishNose Sensory/microbial counts Olafsdottir et al. (2005)
storage time AlphaMOS
Cod, sardine, yellow tail/ TMA/ Semiconductor gas sensor Prototype K value, TVB-N Egashira et al. (1994)
DMA/NH3 detection (TiO2 – In2O3-MgO Ohashi et al. (1991)
63
cellulose acetate matrix and a simple illumination source The TVB-N levels
released from orange roughy and black scabbard, deepwater fish, were moni-
tored by color changes in the sensors.
5. CASE STUDY I: FRESHSENSE — AN ELECTRONIC NOSE

PROTOTYPE INSTRUMENT DEVELOPED TO BE SELEC-
TIVE FOR QUALITY INDICATOR COMPOUNDS IN FISH
An electronic nose called FreshSense based on electrochemical gas sensors has

been developed in our laboratory in cooperation with an Icelandic company
(Bodvaki-Maritech). Knowing that odor of fish is one of the most important
quality attributes to determine the freshness quality, the aim was to apply the
instrument for quality monitoring in the fish industry by rapidly measuring
volatile compounds contributing to the spoilage odor of fish. The main classes
of compounds that are found in increasing concentrations in the headspace of
fish with storage time are short-chain alcohols, aldehydes, ketones, esters, sulfur
compounds, and amines (Lindsay et al., 1986) (Figure 2).
As discussed before, the information on the identity and quantity of vola-
tile compounds present in the headspace during storage of fish is essential when
selecting sensors in an array for quality monitoring of fish. Moreover, it is impor-
tant to know the sensitivity of the sensors toward key compounds in the sample
indicating the quality. Selected standard compounds that are representative of
Ammonia odor, dried fish, old fish odor

Amines: ammonia (NH3),
trimethylamine (TMA), dimethylamine (DMA)
Biogenic amines: histamine, Putrid odor, onion like,
cadeverine, putrecine, tyramine old cabbage, rotten egg
Sulfur compounds:
hydrogen sulfide (H2S),
methylmercaptan (CH3SH),
dimethyl disulfide (DMDS),
dimethyl trisulfide(DMTS)
Acids:
Sweet-sour, malty odor acetic acid
Alcohols, aldehydes, ketones, esters:
ethanol, propanol, butanol,
2-me-1-propanol
3-me-1-butanol, 3-methylbutanal
butanone, acetoin
ethyl acetate
Figure 2. Quality indicators of fish. Main classes of compounds contributing to microbial spoilage
odor of fish.
the main classes of compounds causing spoilage odor have been analyzed by the
electronic nose FreshSense (Olafsdottir et al., 1998, 2002). The results showed
that the electrochemical gas sensors (Dräger, Germany: CO, SO2; City Tech-
nology, Britain: NH3) in the FreshSense had different selectivity and sensitivity
toward the compounds selected from these classes. The CO sensor was sensitive
toward alcohols, aldehydes, and esters; the NH3 sensor detected amines like TMA
and ammonia; and the SO2 sensor detected volatile sulfides. Therefore, the main
classes of spoilage indicator compounds present in the headspace can be estimated
based on the individual sensor responses.
A storage study of haddock fillets kept in cold storage (0–2°C) for 3, 7, 10, and
14 days was performed to obtain more detailed information about the identity and
the level of the most abundant volatile compounds present in the headspace dur-
ing storage of fish (Olafsdottir et al., 2003b). Analyses of the headspace volatiles
were conducted using air pump headspace sampling and collection of volatiles
by a TENAX preconcentration technique and analysis by gas chromatography
(GC-MS). The results were compared to the individual responses of the FreshSense
electronic nose sensors and sensory odor evaluation of the fillets. The responses of
the electronic nose sensors were comparable to the results of the analysis of vola-
tile compounds obtained by the GC-MS detection (Figure 3). The sum of the GC
responses of individual compounds belonging to each class of compounds like
the alcohols, aldehydes, and esters corresponded to the CO sensor responses of
1200
CO
Alcohols, aldehydes
and esters SO2
Response sensors (nA)
NH3
800
arbitrary units
Amines (TMA)
400
Sulfur
compounds
0
0 5 10 15
0 5 10 15
(a) Days (b) Days
Figure 3. Sum of the peak areas of compounds representing the three different classes of compounds
detected by GC in the headspace of haddock fillets during storage in ice (a) and responses of the CO,
SO2, and NH3 sensors toward haddock fillets during storage in ice (b) (from: Olafsdottir, 2003)
the electronic nose. The sum of the amines and the sulfur compounds detected by
GC-MS corresponded to the responses of the NH3 and SO2 sensors, respectively.
The response of the CO sensor was the highest and increased early in the
spoilage process while the responses of the NH3 sensor and the SO2 sensors
increased later in the spoilage process. Because of breakthrough of small polar
molecules on the TENAX traps the technique is not suitable for the quantifi-
cation of compounds like ammonia, hydrogen sulfide, methyl mercaptan, and
ethanol that have been known to be present in abundance in the headspace of
spoiled fish. However, the electrochemical sensors can detect these compounds
and therefore the slopes and shapes of the curves are slightly different.
Increasing concentrations of spoilage indicator compounds with storage time
resulted in the development of characteristic odors and simultaneously increased
responses of the electronic nose sensors were observed. On day 3 the odor of the
fresh fillet was very little or neutral and low responses of the sensors were observed.
The first spoilage odors of the fillets were sweetlike odors that were contributed
by the alcohols which give sweet, solventlike odors in combination with aldehydes
giving sweet, rancidlike odors (day 7). The amines contributed to salted fish or
stock fish odor and in combination with the sulfur compounds, cheesy and foul
odors developed and the fillets became stale on day 10. Finally, esters were ana-
lyzed in high levels on day 14 giving characteristic sweet, fruity odors. When these
sweet odors were mixed with the foul smell of sulfur compounds and ammonia-
like stockfish character of the amines the odor of the fillet becomes TMA/ammo-
nialike and sour/putridlike, signaling the overt spoilage.
It is of interest to compare the composition of the headspace of different
fish products during storage to guide the development of an electronic nose for
quality monitoring of fish products. A similar set of sensors can be used for a
variety of fish species that are stored and processed by different methods. The
FreshSense instrument has been used for freshness monitoring of various spe-
cies of fish like haddock, capelin, redfish, and cod that were handled and stored
under different conditions. A similar trend in the responses of the electronic nose
and the development of volatile compounds was observed for the different fish
species (Olafsdottir et al., 1997a, b, 1998, 2000, 2002; DiNatale et al., 2001).
6. CASE STUDY II: ELECTRONIC NOSE MEASUREMENTS

TO EVALUATE QUALITY OF HADDOCK FILLETS
Raw material is often labeled with days from catch; however, because of the
effect of various extrinsic and intrinsic factors, the information about days
from catch is not always reliable for the determination of quality or freshness
of the raw material. Sensory evaluation of fish fillets is more difficult than sensory
evaluation of the whole fish. Therefore, instrumental techniques to evaluate
the freshness quality of the fillets are of special interest. The most obvious
spoilage signs of the fillets are development of spoilage odor, discoloration,
and decreased firmness of the flesh. Various instrumental techniques to detect

these changes have been developed (Olafsdottir et al., 1997c, 2004) but their
implementation in the fish industry has been slow.
Several storage studies of haddock stored both as whole fish on ice and
as fillets have been performed in our laboratory (Tryggvadottir and Olafsdottir,
2000; Olafsdottir et al., 2003b). The main objective of the studies was to evaluate
the possibility of using an electronic nose to monitor spoilage changes in had-
dock fillets and to study the spoilage pattern of fish caught at different seasons
(spring and autumn) using different fishing gear (long line and Danish seine) and
storage conditions.
The fish was stored whole in ice during the first two experiments (May 1999
and Sep 1999), but in the third one (May 2000) the fish was stored as fillets in
Styrofoam boxes in a cold room (0–2°C). Changes of various properties related
to freshness quality of haddock stored in ice were monitored by the electronic
nose FreshSense and compared to traditional chemical [total volatile bases
(TVB-N)], sensory (Torry scheme) and microbial methods [total aerobic viable
count (TVC)] and texture measurements (TPA). Measurements were performed
on fillets on days 1, 4, 6, 8, 11, 13, and 15 counted from the day of catch.
Sensory analysis of the fillets using evaluation of cooked fillets according
to the Torry scheme showed that the spoilage pattern was different in the three
studies (Figure 4). The limit of shelf life was defined as Torry score = 5.5. Fillets
spoil faster than whole fish as expected and seasonal variation influenced the
spoilage rate. Recently spawned fish caught by Danish seine and stored as whole
fish in ice from the spring (May 1999) had shorter shelf life in ice (10 days) than
whole fish caught by long line from the autumn season (Sep 1999) (14–15 days).
10
9
8
Torry sensory score
7
6
5
4 sensory rejection
Torry = 5.5
3
2
MAY 00 MAY 99 SEP 99
1
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Days
Figure 4. Sensory analysis of haddock fillets from storage studies of whole fish [(May 99 (--); Sept
99 (--)] and fillets [May 00 (--)].
Fish stored as fillets at 0–2°C (May 2000) spoiled faster than whole fish and had
a shelf life of only 8–9 days.
The results of the electronic nose measurements and the microbial and
chemical analysis were mostly in agreement with the sensory analysis (Figures 5
and 6). The traditional microbial analysis (TVC) and electronic nose measure-
ments (CO sensor) showed a similar overall trend in the storage studies of whole
fish (May 1999 and Sept 1999) and fillets (May 2000) (Figure 5) and likewise
the results of TVN chemical measurements and the NH3 electronic nose sensor
showed similar overall trends (Figure 6).
The electronic nose measurements showed that the responses of the sen-
sors were highest for the (May 2000) samples indicating the most rapid spoilage
for fish stored as fillets. The results from the sensory analysis, the TVN (total
volatile bases) measurements, and the NH3 sensor showed that recently spawned
fish from the spring season (May 1999), stored as whole fish, had higher spoilage
rate than fish from the autumn season (Sep 1999). Contradictory to these results
the TVC and CO measurements indicated that the spoilage rate appeared to
be higher in the fall (Sep 1999) than in the spring (May 1999). However, based
on the results of sensory analysis it appears that the spoilage potential of the
microflora might have been greater in the spring. This suggests that the role of
specific spoilage organisms producing offensive odors should be emphasized and
the validity of TVC measurement may be questionable as has been suggested by
other researchers (Gram et al., 2002).
When the end of the shelf life was reached, as indicated by sensory analysis,
the responses of the NH3 and SO2 sensors detecting microbially produced amines
and sulfur compounds start to increase rapidly in all experiments. The conclusions
drawn from these results indicate that spoilage patterns are different depending on
season and fillets spoil faster than whole fish, as was to be expected.
9 900
8 MAY 00 800
Response CO sensor (nA)
MAY 00
7 700
SEP 99
TVC log cfu/g
6 600
5 500 SEP 99
4 MAY 99 400
3 300
MAY 99
2 200
1 100
0 0
0 5 10 15 20 0 5 10 15 20
Days Days
Figure 5. Microbial analysis (TVC) and electronic nose measurements (CO sensor) of haddock fil-
lets from storage studies at 0–2 °C for up to 15 days of whole fish [May 99 (--); Sept 99 (--)] and
fillets [May 00 (--)].
70 100
MAY 00
90
Response NH3 sensor (nA)

60 MAY 00
80
TVB-N mg N/100g
50 70
MAY 99
MAY 99 60
40
50
30 40
30
20 SEP 99 20
SEP 99
10 10
0
0 0 5 10 15 20
0 5 10 15 20 Days
Days
Figure 6. Chemical measurements of total volatile bases (TVB-N) and electronic nose measure-
ments (NH3 sensor) of haddock fillets stored at 0–2 °C for up to 15 days, from storage studies of
whole fish [May 99 (--); Sept 99 (--)] and fillets [May 00 (--)].
Principal components analysis is a good way to illustrate the overall results

(Figure 7). The first PC explained 60% of the variation in the data set and the
spoilage level of the samples increased from left to right along PC1. The second
PC explained 16% of the variation in the data and appeared to be discriminating
1.0 PC2 FIRM Bi-plot

1/MAY99
TPA
1/MAY 00
H2S
15/MAY00
0.5 Whole fish
NH3 Fillets
RT-Meter 6/MAY 00 SO3 15/MAY99
1/SEPT99 MAY 1999 MAY 2000
Torry TVB13/MAY99
4/MAY99 TMA
4/MAY 00
0 13/MAY00
8/MAY00
8/MAY99 11/MAY99 11/MAY00
4/SEP99 log TVC
6/MAY99 logH2S
CO
6/SEP99
−0.5 8/SEP99
11/SEP99
15/SEP99
18/SEP99 Whole fish
13/SEP99
SEP 1999
−1.0 PC1
−0.5 0 0.5 1.0
all data, X-expl: 60%, 18%
Figure 7. PCA showing a biplot of sample scores and variable loadings of all data from three stor-
age studies of haddock as whole fish (May 1999 and Sept 1999) and fillets (May 2000) stored at 0–2
°C for up to 15 days. The samples are labeled with storage days and time of the experiment and the
variable loadings are labeled as follows: E-nose: CO, NH3, SO2, and H2S; chemical: TVB/TMA;
microbial: log H2S, log TVC; texture: Firm /TPA.
samples based on the season. The variable loadings show that the CO sensor and
the microbial counts (log H2S and log TVC) are highly correlated and the posi-
tioning of the H2S, SO2, and NH3 sensors on the upper half of the plot show the
contribution of these sensors to discriminate samples from the spring season in
combination with the texture measurement (Firm). Fish spoils faster in the spring
because the fish is recently spawned and the texture of the muscle is softer and
more gaping than in the fall.
Different spoilage patterns are evident depending on the season indicating
that days of storage is not a good estimate of freshness quality. The E-nose sen-
sor responses appear to give information on the quality of fish related to sensory
scores, microbial growth (TVC), and spoilage indicators like TVB-N.
7. CONCLUSIONS
Rapid measurements of the headspace of fish with an electronic nose has the
potential to detect the freshness or quality stage of the fish, if the sensor array
can detect the respective indicator compounds that truly represent the condition
of the fish. The high sensitivity of the electronic nose sensors toward the very
volatile degradation compounds, like alcohols, sulfur compounds, and amines
signaling overt spoilage suggests that the E-nose may be promising for moni-
toring off odors caused by fish and animal waste and off odors caused by the
processing of fish meal or dried products when the raw material is not of optimal
freshness.
8. ACKNOWLEDGMENTS
The case studies were part of projects financed by The Icelandic Centre for
Research: Accurate predictive models—The effect of temperature fluctuations
on microbial growth and metabolites in fish and The European Commission:
MUSTEC project FAIR98 4076.
9. RECOMMENDED READING
The following papers give a good overview of the application of electronic noses
for monitoring food (Garcia-González and Aparicio, 2002; Harper, 2001; Haugen,
2001; Bartlett et al., 1997; Gardner and Bartlett, 1999; Schaller et al., 1998) and Jurs
et al. (2000) give an excellent overview of data processing techniques for electronic
noses.
Information on the website of The NOSE II 2nd Network on Artificial Olfac-
tory Sensing has the aim to stimulate information exchange between scientists,
manufacturers, and end users in Europe in order to develop synergy in the sensor
community, improve the efficiency of R&D, encourage interdisciplinary research,

and promote application of new ideas (http://www.nose-network.org/).
10. REFERENCES
Ahmed, A., Matches, J. R. (1983) “Alcohol production by fish spoilage bacteria.” Journal of Food
Protein 46, 1055–1059.
Bartlett, H.N., Elliott, J.M., Gardner, J.W. (1997). Electronic noses and their application in the food
industry, Food Technology 51(12): 44–48.
Boothe, A.A.J., Arnold, J.W. (2002). ‘Electronic nose analysis of volatile compounds from poultry
meat samples, fresh and after refrigerated storage’, J. Sci. Food Agric.82, 3, 315–322.
Buttery, R. G., Guadagni, D.G., Ling, L.C., Seifert, R.M., Lipton, W. (1976). Additional volatile
components of cabbage, broccoli and cauliflower. J. Agric. Food Chem., 24 (4), 829–832.
Byrne, L.; Lau, K.T.; Diamond, D. (2003). Development of pH sensitive films for monitoring spoil-
age volatiles released into packaged fish headspace. Irish Journal of Agricultural and Food
Research, 42(1) 119–129.
Deng, Z., Stone, D.C., Thompson, M. (1996). Selective detection of aroma components by acoustic
wave sensors coated with conducting polymer films. Analyst 121,671–679.
Di Natale, C., Brunink, J.A.J., Bungaro, F., Davide, F., d’Amico, A., Paolesse, R., Boschi, T., Faccio,
M., Ferri, G. (1996), ‘Recognition of fish storage time by a metalloporphyrins-coated QMB sen-
sor array’, Meas. Sci. Technol. 7: 1103–1114.
Di Natale, C., Olafsdottir, G., Einarsson, S., Mantini, A., Martinelli, E., Paolesse, R., Falconi, C.,
D’Amico, A., (2001). Comparison and integration of different electronic noses for the evalua-
tion of freshness of cod fish fillets. Sensors and Actuators B77,572–578.
Di Natale, C. (2003), Data Fusion in MUSTEC: Towards the Definition of an Artificial Quality Index. In:
Luten J B, Oehlenschlager J and Olafsdottir G (Eds), Quality of Fish from Catch to Consumer: Label-
ing, Monitoring and Traceability, Wageningen Academic Publishers, The Netherlands, 273–280.
Du, W.X., Kim, J., Huang, T.S., Marshall, M.R., Wei, C.I. (2001), ‘Microbiological, sensory and
electronic nose evaluations of yellowfin tuna under various storage conditions’, J Food Prot. 64
(12): 2027–2036.
Du, W.X., Lin, C.M., Huang, T.S., Kim, J., Marshall, M.R., Wei, C.L. (2002), ‘Potential application
of the electronic nose for quality assessment of salmon fillets under various storage conditions’,
J Food Sci. 67 (1): 307–313.
Egashira, M., Shimizu, Y., Takao, Y. (1994), Fish Freshness detection by semiconductor gas sensors,
Olfaction and Taste XI. Proc. Int. Sym. 11, 715–719.
Egashira, M., Shimizu, Y., Takao, Y. (1990) Trimethylamine sensor based on semiconductive metal-
oxides for detection of fish freshness. Sensors and Actuators B-Chem 1 (1–6): 108–112.
Fazzalari, F.A. (1978). Compilation of Odor and Taste Threshold Value Data. American Society for
Testing and Materials, Philadelphia.
Gacrcía-González, D. L. and Aparicio, R. (2002). Sensors: From biosensors to the electronic nose.
Grasas y Aceites, 53(1) 96–114.
Gardner, J.W., Bartlett, P.N. (1999), Electronic noses. Principles and applications, Oxford University
Press. Oxford, England.
Gram, L., Ravn, L., Rasch, M., Bruhn, J.B., Christensen, A.B., Givskov, M. (2002) Food spoilage
- interactions between food spoilage bacteria. Int. J. Food Microbiology 78 (1–2): 79–97.
Harper, W.J. (2001), Strengths and weaknesses of the electronic nose. In: Rouseff and Cadwallader
(Eds) Headspace Analysis of Food and Flavors: Theory and Practice. Kluwer Academic / Plenium
Publishers, New York pp 59–71.
Haugen, J.E. (2001). Electronic noses in food analysis. In: Rouseff and Cadwallader (Eds). Headspace
Analysis of Food and Flavors: Theory and Practice. Kluwer Academic / Plenium Publishers,
New York pp 43–57.
Haugen, J.E., Undeland, I. (2003), Lipid oxidation in herring fillets (Clupea harengus) during ice
storage measured by a commercial hybrid gas-sensor array system, J Agric. Food Chem. 51, 3,
752–759.
Herbert, R.A., Ellis, J.R., Shewan, J.M. (1975), Isolation and identification of the volatile sulfides pro-
duced during chill-storage of North Sea Cod (Gadus morhua). J.Sci. Food Agric., 26, 1195–1202.
Joffraud, J.J., Leroi, F., Roy, C., Berdague, J.L. (2001), Characterisation of volatile compounds pro-
duced by bacteria isolated from the spoilage flora of cold-smoked salmon, Int. J. Food Micro-
biol. 66, 3, 175–184.
Jørgensen, B.M., Oehlenschläger, J., Olafsdottir, G., Tryggvadóttir, S.V., Careche, M., Heia, K., Nes-
vadba, P., Nunes, M.L., Poli, B.M., Di Natale, C., Pérez-Villarreal, B., Ballo, H., Luten, J.B., Smelt,
A., Denton, W., Bossier, P., Hattula, T., Åkesson, G. (2003), A study of the attitudes of the
European fish sector towards quality monitoring and labeling. In: Luten, J.B., Oehlenschläger,
J. and Olafsdottir, G. (Eds) Quality of Fish from Catch to Consumer: Labeling, Monitoring and
Traceability, Wageningen Academic Publishers, The Netherlands, 57–71.
Jørgensen, L.V., Huss, H.H., Dalgaard, P. (2001). Significance of volatile compounds produced by
spoilage bacteria in vacuum-packed cold-smoked salmon (Salmo salar) analyzed by GC-MS
and multivariate regression. J. Agric. Food Chem. 49, 2376–2381.
Josephson, D.B. (1991). Seafood. In Volatile compounds in foods and beverages, H. Maarse (Ed)
Marcel Dekker, New York, 179–202.
Josephson, D.B., Lindsay, R.C., Stuiber, D.C. (1984), Variations in the occurrences of enzymically derived
volatile aroma compounds in salt- and freshwater fish. J. Agric. Food Chem., 32, 1344–1347.
Josephson, D.B., Lindsay, R., Olafsdottir, G. (1986). Measurement of Volatile Aroma Constituents
as a Means for Following Sensory Deterioration of Fresh Fish and Fishery Products. In Pro-
ceedings of an International Symposium, University of Alaska Sea Grant Program, Anchorage,
Alaska, U.S.A., Kramer DE and Liston J (Ed.), Elsevier Science Publishers B.V., Amsterdam
pp. 27–47.
Jurs, P.C., Bakken, G.A., McClelland, H.E. (2000) Computational Methods for the Analysis of
Chemical Sensor Array Data from Volatile Analytes. Chem. Rev. 100 pp. 2649–2678.
Kamiya, A., Ose, Y. (1984). Study of the odorous compounds produced by putrefaction of foods.
V. Fatty acids, sulfur compounds and amines. Chromat., 292, 383–391.
Kawai, T. (1996). Fish flavor. Critical Reviews in Food Science and Nutrition, 36, 3, 257–298.
Kelleher, S.D., Zall, R.R. (1983). Ethanol accumulation in muscle tissue as a chemical indicator of
fish spoilage. J. Food Biochem., 7, 87–91.
Keshri, G., Voysey, P., Magan, N. (2002). Early detection of spoilage moulds in Keshri bread using
volatile production patterns and quantitative enzyme assays. J. Appl. Microbiol. 92, 165–172.
Lindsay, R.C., Josephson, D.B. and Olafsdottir, G. (1986). Chemical and biochemical indices for
assessing the quality of fish packaged in controlled atmospheres. In Proceedings of an Inter-
national Symposium, University of Alaska Sea Grant Program, Anchorage, Alaska, U.S.A.,
Kramer, D.E. and Liston, J. (Ed.), Elsevier Science Publishers B.V., Amsterdam pp 221–234.
Luzuriaga, D.A., Balaban, M.O. (1999a). Evaluation of odor decomposition in raw and cooked
shrimp: Correlation of electronic nose readings, odor sensory evaluation and ammonia levels.
In: Hurst, W.J. (Ed), Electronic nose and sensor array based system, Design and Applications.
Proceedings of the 5th International Symposium on Olfaction and the Electronic Nose. Tech-
nomic Publication Company, Inc, USA, pp.177–184.
Luzuriaga, D.A., Balaban, M.O. (1999b). Electronic nose odor evaluation of salmon fillets stored
at different temperature. In: Hurst, W.J. (Ed) Electronic nose and sensor array based system,
Design and Applications. Proceedings of the 5th International Symposium on Olfaction and the
Electronic Nose. Technomic Publication Company, Inc, USA, pp.162–169.
Macagnano, A., Careche M., Herrero, A., Paolesse, R., Martinelli, E., Pennazza G., Carmonae, P.,
D’Amico A., Di Natale, C. (2005), A model to predict fish quality from instrumental features.
Sensors and Actuators B 111–112 293–298.
McEntegart, C.M., Penrose, W.R., Strathmann, S., Stetter, J.R. (2000). Detection and discrimination
of coliform bacteria with gas sensor arrays. Sensors and Actuators B 70, 170–176.
McGill, A.S., Hardy, T., Burt, R.J. and Gunstone, F. D. (1974). Hept-cis-4-enal and its contribution
to the off-flavor of cold-stored cod. J. Sci. Food Agric., 25, 1477–1481.
Newman, D.J., Luzuriaga, D.A., Balaban, M.O. (1999). Odor and Microbiological Evaluation of
Raw Tuna: Correlation of Sensory and Electronic Nose Data. In: Hurst, W.J. (Ed), Electronic
nose and sensor array based system, Design and Applications. Proceedings of the 5th Interna-
tional Symposium on Olfaction and the Electronic Nose. Technomic Publication Company, Inc,
USA, pp.170–176.
Oehlenschläger, J. (1992). Evaluation of some well established and some underrated indices for the
determination of freshness and/or spoilage of ice stored wet fish’. In: Huss, H.H., Jakobsen, M.,
Liston, J. (Eds), Quality Assurance in the Fish Industry, , Elsevier Science Publisher, Amsterdam,
339–350.
Ohashi, E., Takao, Y., Fujita, T., Shimizu, Y., Egashira, M. (1991). Semiconductive trimethylamine
gas sensor for detecting fish freshness, J. Food Sci. 56(5)1275–1278.
Olafsdottir, G., Martinsdóttir, E., Jónsson, E.H. (1997a). Rapid gas sensor measurements to predict
the freshness of capelin (Mallotus villosus). J. Agric. Food Chem. 45(7) 2654–2659.
Olafsdottir, G., Martinsdóttir, E., Jónsson, E.H. (1997b). Gas sensor and GC measurements of
volatile compounds in capelin (Mallotus villosus). In Seafood from Producer to Consumer,
Integrated Approach to Quality. Eds. Luten, J.B., Börresen, T., Oehlenschläger, J. Amsterdam,
Elsevier, pp 507–520.
Olafsdottir, G., Martinsdóttir, E., Oehlenschläger, J., Dalgaard, P., Jensen, B., Undeland, I., Mackie,
I.M., Henehan, G., Nielsen, J., Nilsen, H. (1997c), Methods to evaluate fish freshness in research
and industry. Trends Food Sci. Technol, 8. 258–265.
Olafsdottir, G., Högnadóttir, Á., Martinsdóttir, E. (1998), Application of gas sensors to evaluate
freshness and spoilage of various sea foods. In: Methods to Determine the Freshness of Fish
in Research and Industry, Proceedings of the Final meeting of the Concerted Action “Evalua-
tion of Fish Freshness” AIR3 CT94 2283. Nantes Nov 12–14, 1997. International Institute of
Refrigeration, pp100–109.
Olafsdottir, G. and Fleurence, J. (1998). Evaluation of fish freshness using volatile compounds -
Classification of volatile compounds in fish. In Methods to Determine the Freshness of Fish
in Research and Industry, Proceedings of the Final meeting of the Concerted Action “Evalua-
tion of Fish Freshness” AIR3 CT94 2283. Nantes Nov 12–14, 1997. International Institute of
Refrigeration, 55–69.
Olafsdottir, G., Högnadóttir, Á., Martinsdóttir, E. and Jónsdóttir, H. (2000). Application of an
Electronic Nose to Predict Total Volatile Bases in Capelin (Mallotus villosus) for Fishmeal
Production, J. Agric. Food Chem. 48 ,6, 2353–2359.
Olafsdottir, G., Li, X., Lauzon, H. L., Jonsdottir, R. (2002). Precision and application of electronic
nose measurements for freshness monitoring of redfish (Sebastes marinus) stored in ice and
modified atmosphere bulk storage. JAFP 11, 3/4, 229–249.
Olafsdottir, G., Di Natale, C., Macagnano, A. (2003a). Measurements of quality of fish by electronic
noses. In: Luten, J.B., Oehlenschläger, J. and Olafsdottir, G. (Eds), Quality of Fish from Catch
to Consumer: Labelling, Monitoring and Traceability, Wageningen Academic Publishers, The
Netherlands, 225–234.
Olafsdottir, G., Jonsdottir, R., Tryggvadottir, S.V., Einarsson, S., Lauzon, H.L., (2003b). Prediction
of freshness quality of haddock fillets using electronic nose, texture and TVN measurements
First Joint Trans-Atlantic Fisheries Technology Conference (TAFT) 33rd WEFTA Meeting
and 48th Atlantic Fisheries Technology Conference, 11–14 June 2003, Reykjavik - Iceland 2003
pp. 148–150. (http://vefur.rf.is/TAFT2003/).
Olafsdottir, G. (2003). Developing rapid olfaction arrays for determining fish quality. In Tothill, I.E.
(Ed) Rapid and on-line instrumentation for food quality assurance. Woodhead Publishing Ltd,
Cambridge, England, pp. 339–360.
Olafsdottir, G., Nesvadba, P., Di Natale, C., Careche, M, Oehlenschläger, J., Tryggvadóttir, S.V.,
Schubring, R., Kroeger, M., Heia, K., Esaiassen, M., Macagnano, A., Jørgensen, B.M. (2004).
Multisensor for fish quality determination. Trends in Food Science and Technology 15, 86–93.
Olafsdottir, G., Chanie, E., Westad, F., Jonsdottir, R., Thalman, C., Bazzo, S., Labreche, S., Marcq,
P., Lundby, F., Haugen, J.E. (2005). Prediction of Microbial and Sensory Quality of Cold
Smoked Atlantic Salmon (Salmo salar) by Electronic Nose. J Food Sci 70 (5) 563–574.
Olafsdottir, G. (2005). Volatile compounds as quality indicators of fish during chilled storage: Evalu-
ation of microbial metbolites by an electronic nose. PhD thesis Faculty of Science, University
of Iceland, 219p.
Olafsdottir, G., Lauzon, H., Martinsdottir, E., Oehlenschläger, J., Kristbérgsson, K. (2006). Eval-
uation of shelf-life of superchilled cod (Gadus morhua) fillets and influence of temperature
fluctuations on microbial and chemical quality indicators. J. Food Sci 71 (2) 97–109.
Ólafsson, R., Martinsdóttir, E., Olafsdottir, G., Sigfússon, Þ. I., Gardner, J. W. (1992), Monitoring
of fish freshness using tin oxide sensors. In: Gardner, J.W. and Bartlett, P. N. (Eds), Sensors
and Sensory Systems for an Electronic Nose, Kluwer Academic Publ. 257–272.
Olsson, J., Börjesson, T., Lundstedt, T., Schnürer, J. (2002), Detection and quantification of ochra-
toxin A and deoxyinvalenol in barley grains by GC-MS and electronic nose. Int. J Food Microb.
72, 203–214.
Pradyumna, K., Namadev, Yair Alroy and Vijay Singh (1998) Sniffing out trouble: Use of an electronic
nose in bioprocesses, Biotechnol. Prog, 14, 75–78.
Saja, R.de, Souto, J., Rodriguez-Mendez, M.L., de Saja, J.A. (1999) Array of lutetium bisphthalocya-
nine sensors for the detection of trimethylamine. Materials Science & Engineering C-Biomimetic
and Supramolecular Systemes 8–9: 565–568.
Schaller, E., Bosse, J.O., Escher, F. (1998) ‘Electronic noses’ and their application to food’, Lebensmittel-
Wissenschaft-und-Technologie, 31 (4) 305–316.
Schnürer, J., Olsson, J. and Börjesson, T. (1999) Fungal volatiles as indicators of food and feeds
spoilage. Fungal Genetics and Biology 27, 209–217.
Schweizer-Berberich, M., Vahinger, S., Göpel, W. (1994) Characterization of food freshness with
sensor array, Sensors and Actuators B, 18–19, 282–290.
Sheldon, R.M., Lindsay, R.C., Libbey, L.M., Morgan, M.E. (1971) Chemical nature of malty flavor
and aroma produced by streptococcus lactis. Applied Microbiol. 22, 263–266.
Storey, r.M., Davis, K., Owen, D., Moore, L. (1984). Rapid approximate estimation of volatiel
amines in fish, J. Food Technol. 19,1–10.
Tryggvadottir, S.V., Olafsdottir, G., 2000. Multisensor for fish: Questionnaire on quality attributes
and control methods -Texture and electronic nose to evaluate fish freshness. Project report for
European Commission (MUSTEC CT-98–4076). RF report 04–00, 36p.
Whitfield, F.B., Shaw, K.J., Tindale, C.R. (1988). 2,6-dibromophenol: The cause of an iodoform-like
off-flavour in some Australian crustacea. J. Sci. Food Agric., 46, 29–42.
Whitfield, F.B., Tindale, C.R. (1984). The role of microbial metabolites in food off-flavours. Food
Technol. Australia, 36, 204–213.
Zeng, K., Tachikawa, H., Zhu, Z.Y., Davidson, V.L. (2000), Amperometric detection of histamine with
a methylamine dehydrogenase polypyrrole-based sensor, Analytical Chemistry (10): 2211–2215.
Zhao, C.Z., Pan, Y.Z., Ma, L.Z., Tang, Z.N., Zhao, G.L., Wang, L.D. (2002), “Assay of fish fresh-
ness using trimethylamine vapor probe based on a sensitive membrane on piezoelectric quartz
crystal,” Sensors and Actuators B-CHEM 81 (2–3): 218–222.
6
Odors Prevention in the Food
Industry
Regina Nabais
1. INTRODUCTION
Taking into account the environmental impact, at the local and regional level, the
control and elimination of the emission of noxious odors in the food-processing
industry is one of that industry’s most difficult problems. The problem normally is
exacerbated not only because of the multiple ways in which it can be approached,
but also because, in spite of many efforts developed, there still is no universally
accepted way to address its solution.
The complexity of the issue of off-odors and the necessary multidisciplinary
expertise entailing a rigorous approach is the reason this chapter is neither exten-
sive nor intended to represent, even at an introductory level, an odor manage-
ment guide for the food industry. The scope is to strengthen the necessity to
reconcile aims, methods, collaborative attitudes, and efforts; to thoroughly out-
line strategies and tactics of scientific, technological, political, and legal inter-
ventions; and to explore the continued improvement and consistency of effective
global and integrated actions that will lessen the general environmental impact,
including odor emissions, of the food-processing industry. Whenever possible,
the main concepts, principles, and fundamentals of a few mentioned suggestions
to odor abatement or control will be mentioned and a list of bibliographic refer-
ences will be provided.
2. CONCEPTIONS AND MISCONCEPTIONS ABOUT ODORS

EMISSION BY THE FOOD INDUSTRY
In Europe, the Charter of Fundamental Rights, as proclaimed at the Nice Euro-

pean Council in December 2000 specifically states: “A high level of environmental
REGINA NABAIS C.E.R.N.A.S. Centro de Recursos Naturais, Ambiente e Sociedade, Escola

Superior Agrária de Coimbra, Instituto Politécnico de Coimbra. Bencanta, 3040–316 Coimbra, Portugal.
75
76 R. Nabais
protection and the improvement of the quality of the environment must be inte-
grated into the policies of the Union and ensured in accordance with the principle
of sustainable development.” This statement reflects the numerous changes in
European politics owing to environmental problems (Mike Holland et al., 2004;
SEPA and NIEHS, 2002 a, b).
Emissions of foul odors, regardless of their origin, are problems of
extreme complexity, assuming an ever-larger role among environmental issues.
This is especially the case when they have a direct, negative impact upon the
populations near an offending industry, whether they are a simple inconve-
nience or a harmful treat to neighboring communities. An odor is recognized
through sensory perception and subjectively characterized either as agreeable
or disagreeable, or even ambivalently perceived, depending upon the specific
circumstances of how an individual’s olfactory sense is stimulated. The olfac-
tory sense probably has not received the attention it deserves, considering that
it influences the senses of touch and taste, with deep repercussions on psycho-
logical behavior, stemming from its possible memorization and affective and
cultural interpretations.
When the sources of pollution emissions are specific industries, such as
food processing, pharmaceutical, or cosmetics, the supposedly marginal ques-
tions of production processes such as esthetics and odor become important
issues in the production process, because they may be reflected in the price of
the products in the marketplace.
Concerning odors, even those generally considered agreeable, under certain
conditions can become unbearable, depending upon the variability of emission
and reception conditions. This particularity is known by the acronym “FIDOL,”
which stands for the following: (1) frequency: how often an individual is exposed
to odour; (2) intensity: the strength of the odor; (3) duration: the length of a
particular odor event; (4) offensiveness/character: the character relates to the
“hedonic tone” of the odor (meaning “smells like”), which may be pleasant, neutral,
or unpleasant; and (5) location: the type of land, use, and nature of human activities
in the neighborhood of an odor source.
3. THE COSTS OF NOT TREATING ODORS
Even if the majority of malodors produced by the food processing industry are
inoffensive, due to low concentrations or reduced toxicity, the industries that
emit them must be prepared to encounter different levels of protest against their
activities:
1. People believe that the odor has a negative effect on their well-being and
health conditions.
2. Environmental and activist political groups often protest against any
activity that may lead to the emission of bad odors.
Odors Prevention in the Food Industry 77
3. The plaintiffs and their lawyers or legal representatives will usually drag
the offending industries through interminable legal battles, extremely
costly in terms of time, money, and resources.
4. The normative legislation at the local, regional, national, and suprana-
tional levels are progressively demanding, regulatory, and restrictive.
5. Other businesses and entities, which complain of the negative or poten-
tially negative impacts on their costumers and their future investments
in the area.
6. Some clients may boycott food products produced by factories that emit
bad odors.
7. The installation of some potentially economically viable activities (e.g.,
commercial ventures, tourism, sporting venues, etc.) would be compro-
mised by locations near odor-emitting industries.
Some activities of the food-processing industry are potential emissaries of “bad”

odors e.g., fish processing, roasting coffee, rendering of fats, extraction of veg-
etable oils, wineries and breweries, chocolate production, preprocessed foods,
even if located some kilometers from inhabited areas, can add significantly to
the energy bills of their neighborhoods, due to the fact that they must maintain
environmental control systems operating, without the need of their utilization
for temperature control.
Table 1 synthesizes what is the meaning and consequences for externalities
(social costs) that are costs not directly considered or paid by the owners of an
economic activity.
The creation of good relations and trust with those who live near offend-
ing emission producers, combined with an ongoing tightening of regulations,
can be decisive factors in the resolution of odor problems. Pushing away the
neighbors from the search for solutions environmental problems caused by an
industry promotes strong feelings of suspicion and helplessness and gives rise to
unbearable resentments, never overcome by developing expensive information
campaigns. In fact, effective public relations and communications tactics help
producers become aware and strengthen and foster favorable actions by poli-
cymakers; otherwise, people tend to exaggerate risks and may be suspicious of
arguments limited to technical information.
Table 1. Relative Cost for Controlling Industrial Environmental Impacts
Enterprise Costs
Quality of the Depreciation of the
prevention or Maintenance and enterprise image and Social Costsa
control systems Initial investment operational costs final property value (externalities)
Improper Inexistent or low High High High

Proper Medium to high Low Null Low
a
Concept explained in (Holland et al. (2004)).
78 R. Nabais
One method to enlist the confidence and trust of the local neighborhood
residents is to offer them a forum in which to explore their apprehensions and
doubts, to offer guided tours of the facility and its production system, and to
include them in both exploring the environmental problems and the quest for
possible solutions. “Science is losing credibility. Conflicts of interest, biased stud-
ies, and secrecy are undermining science’s reputation and its truth-seeking objective
affecting public trust” (Collins, 2000).
Hence, trust and confidence are the main keys to addressing public perception
and understanding. This is not in itself an expensive methodology in monetary
terms, but it demands accomplishing the most difficult tasks human beings can
achieve: cultural changes. Trust enforces the exchange of information and ideas
and increases the likelihood of the community feeling more in control of the solu-
tions for their environmental concerns, instead of promoting overreactions as a
result of exaggerating their fears by useless secrecy or hiding the problems.
4. MAJOR INDUSTRIAL FOOD PROCESSES RESPONSIBLE

FOR ODOROUS SUBSTANCES EMISSIONS
Apart from considering process isolation, capable confinement, and good con-
veyor systems, immediately followed by efficient onsite treatments (and despite
the existence of several exceptions) direct interventions in food process and/or
food process control have poor effect on annoying properties of odors. Food
process residues, their manipulation, transfer operations, treatment and final dis-
posal systems or facilities are also sources of malodors emissions.
A food process residue is usually understood to be an incidental organic
material generated by processing agricultural commodities for human or ani-
mal consumption. The term includes food residuals, food coproducts, food-
processing wastes, food-processing sludge, or any other incidental material
whose characteristics are derived from processing agricultural and fishery prod-
ucts. Examples include process wastewater from cleaning slaughter areas, rinsing
carcasses, or conveying food materials; process wastewater treatment sludge; bone;
fruit and vegetable peels; seeds; shells; pits; cheese whey; off-specification food
products; hide; and hair and feathers.
As any other air pollutant, odor sources may be classified as point source or
disperse source; unfortunately, the food industry may present both odor origins.
A rough organization of improving food industry odors sources may be seen in
Table 2.
Thus, odor can be considered as a “primary air pollutant” emitted by some
food industrial processes: raw materials handling, transport, storage, cooking,
concentration, drying, grilling, fermenting, and roasting processes are examples
of sources of odors emissions. If these emissions are potentially annoying to
their neighbors, their sources must be rethought, controlled, and treated sepa-
rately from the fugitive odors of raw materials.
Table 2. Examples of Food Processes Susceptible of Odors Emissionsa,b
Operations and
processess Utilities operation and maintenance Waste handling and disposal
Cooking; Change, antiquated, inefficient, or improper Remediation and recovery of

drying; odor control systems contaminated land
evaporation; Plan and preview emergency releases Avoid clogged drains, by proper
extraction; procedures
fermentation; Design of heating, combustion, Evaluate emissions from open
frying; or steam generators tanks, lagoons, effluent treatment
roasting plants
Prevention of leaking spills and splashes Convenient planning for solid or
from vessels, pipes, and connections liquid sludge, slurries and wastes
handling, transfer, storage and
final disposal sites
Design appropriate local exhaust ventilation
discharge points
Pumps and flanges sealings
Efficient raw materials logistic organization,
as transports, storage, and handling
Good features for start-up and
shut-down processes
Design of storage tank “breathing” vents
Appropriate installations for end products,
by-products, and waste residues
Efficient transfer processes, including
unloading of vehicles, warehousing
locations, and procedures
Confinement of skips or bins
Adequate design for buildings and
storage areas ventilation
Appropriate washing, disinfection,
cleaning, and maintenance operations
a
From IPPC (2003).
On the other side, in general, liquid wastes produced by food industry have
high contents of organic pollutants including biochemical oxygen demand, ammo-
nia, suspended solids, oils and greases and are commonly acid, biodegradable, and
malodorous (Environmental Protection Agency, 2000). When an unstable organic
waste in liquid, slurry, or sludge starts decomposing, gases are liberated and accu-
mulate in fluid solution phase, until the saturation of the liquid is reached; beyond
this point, smelly vapors are released.
Most of food process treatment systems (ponds, leaching beds, septic tanks,
and trickling filters) are poorly built-up, inefficient, and improperly operated.
Fortunately, there are several highly efficient technologies available for indus-
trial food manufacturing liquid effluents, but the costs are generally unbearable,
considering the narrow marginal profits of these activities. The problem for
80 R. Nabais
small and medium size businesses is to find a technology that is comparatively

affordable, with a short return on the investment relative to its life cycle, capable
of treating high industrial-strength wastewater, with nonodorous treatment,
demanding comparatively little land and flexible enough to be installed as needs
arise (Robèrt et al., 2002; Starkey, 1998).
On other hand, taking into consideration the temperature dependence
of degradation rates, reducing load rates applied to waste treatment features
during cold weather periods is a good practice to decrease off-odors emis-
sion. Another advised procedure for those kinds of biological waste sys-
tems is their continuous loading, at least fed-batch rate additions adapted to
temperature patterns, to avoid the biological agents facing variable nutrient
concentrations or sudden media composition alterations, which can induce
biological nutrient shocks and shifts in metabolic mechanisms to malodorous
processes, as those exemplified in Table 3.
Whatever may be the origin of malodors emissions, it is common to find in the
food industry liquid or gaseous streams containing isolated substances, or in com-
plex mixtures, substances mentioned in the Table 2, Chapter 2, paragraph 3.1. For
instance, at rendering plants, among hundreds of detected odorous compounds
there may be found acetaldehyde, ethyl mercaptan, ammonia, hydrogen sulfide,
butyl amine, indole, butyric acid, methylamine, dimethylamine, methyl mercaptan,
dimethyl sulfide, skatole, dimethyl, qinoline and triethylamine. Quinoline is
classified as a hazardous air pollutant (UKRA, 2004; US Environmental Protec-
tion Agency, 1995).
5. ODOROUS COMPOUNDS AND MIXTURES EMISSIONS IN

FOOD INDUSTRY AND THEIR GENERAL AND SPECIFIC
PHYSICAL AND CHEMICAL CHARACTERISTICS
Odorous substances exhibit common physical and chemical general properties,

all contributing to odor perception; these generic properties are their molecular
weight (less than 300), high volatility, high vapor pressure, and high solubility
(in water needed at the membrane level of the nose and in the lipid components
of nervous cells).
Table 3. Three Dominant Paths of Anaerobic Biological Substrate Biodegradation
Carbohydrates Alcohols, aldehydes, ketones, organic acids Methane, carbon dioxide,

hydrogen sulfide,
ammonia, water and
biomass
Lipids Long and short chain fatty acids, alcohols, acetate
Proteins Peptones, peptides, amino acids, volatile fatty acids,
sulfides, mercaptans, thiols, phenols, indoles
Odorous compounds from most of the food-processing plants are emitted

generally in low concentrations and are simultaneously hydrosoluble and biode-
gradable. Hence, for such pollutants, simultaneously highly soluble and degrad-
able, such as methanol, butanol, tetrahydrofurane, aldehydes, formaldehyde,
carbonic acids, and butyric acid, biological treatment technology appears very
attractive, as its investment and running costs are often lower by a factor of 2 to
10 than the physical–chemical technologies.
6. INDOOR AND OUTDOOR FOOD PROCESSING

FACILITIES: ENVIRONMENTAL QUALITY AND SAFETY,
CONCERNING ODOURS EMISSIONS
Odor emissions in the indoor environment of food industries, affects the indoor
air quality (IAQ). When emissions have negative impacts it is common to observe
a decrease in worker productivity and increases in absenteeism; in extreme con-
ditions, malodorous substances may induce serious illness. Beside these already
mentioned, serious effects, another caution is crucial in IAQ of a food industry
and that is “the taint.” The term taint relates to any foreign smell or taste in
goods, e.g., musky, yeasty, catty, furry, disinfectants, cork, cardboard, hearthy
(geosmin), musty (methylisoborneol), as a result of adsorption or absorption of
volatile compounds, by the commodities materials, throughout the entire pro-
duction line (from raw materials to the final packed products). The causes for
taint occurrences in food products are very diverse, but one of the sources may
be poor indoor air quality.
Anyhow, occurrences like tainting can result in considerable inconveniences
to manufacturers, like lost of production, sales, consumer confidence, damaged
brand names, loss of commercial relationships between suppliers, manufactur-
ers, and retailers, insurance disputes, and in extremely expensive and long legal
battles. The prevention of taint development is vital to any food process, not only
to dignify and preserve product image and stay in business, but also because in
some countries and circumstances (e.g., in United States), if the taint is a result of
intentional or careless manufacturing and cumulatively implies external commerce
depreciation, those legally responsible may be subject to heavy fines and may be
sentenced to up to 3 years in prison (US Food and Drug Administration, 2004).
The three strategies, in order of effectiveness, for reducing pollutants in the
indoor air environment are source control, ventilation, and air cleaning. (1) Source
control eliminates individual sources of pollutants or reduces their emissions, and
is generally the most effective strategy. (2) Ventilation brings outside air indoors.
It can be achieved by opening windows and doors, by turning on local exhaust
fans, or, in some situations, by the use of mechanical ventilation systems, with or
without heat recovery ventilators (air-to-air heat exchangers). However, there are
practical limits to the extent ventilation can be used to reduce airborne pollut-
ants. Costs for heating or cooling incoming air can be significant, and outdoor air
82 R. Nabais
itself may contain undesirable levels of contaminants (Society of Heating, Refrig-

erating, and Air Conditioning Engineers, 1997). (3) Air cleaning may serve as an
adjunct to source control and ventilation. However, the use of air cleaning devices
alone cannot assure adequate air quality, particularly where significant sources are
present and ventilation is insufficient.
The balance of the amount and quality of outdoor air getting in and indoor
sources of emissions is of major importance of indoor air quality; this balance
is identified by indoor/outdoor ratios, usually in the range of 0.5 to 1.7. Unfor-
tunately, indoor odor consequences are not yet considered when ventilation sys-
tems are projected; in fact, they are addressed to provide makeup air, controling
ambient temperature, and or relative humidity control. If odor emission is a
problem, designing for this objective gives the worker a comfortable environment
to work in, minimizes concerns related to worker safety, and improves worker
attitudes, thus, increasing productivity.
The ventilation system layout for nonworker cover or enclosure areas gener-
ally includes an exhaust point(s). Instead, a makeup air access is used and posi-
tioned strategically away from the exhaust to achieve the desired sweep direction
of airflow motion. To avoid odors releasing through the makeup air inlet, this
access must automatically be shut, in the event that positive pressure builds up
in the enclosed headspace.
In some instances, the makeup air source is not necessary, as other sources of
air can replace it. The air contained in the headspace of the drains and sewer com-
monly provides makeup air; in this case one must be attentive to taint prevention.
For worker space areas, the ventilation system layout is site specific, but a few
design criteria always need to be considered, including the following aspects:
Position ventilation system exhaust points at or near the source of the

odors; avoid fouling worker space areas by alternate exhaust locations.
Hydrogen sulfide is heavier than air; hence, its exhaust points must be
located near the floor. This gas is commonly a prevalent odor in waste-
water treatment systems. As a result, the makeup air supply should be
located high, near the ceiling of the enclosure, while the odorous exhaust
is low, at the floor level where the source of the odors is generally found.
Avoid shortcuts between the makeup air entrances and exhausting system;
this is an important consideration to provide indoor air quality to workers.
The required internal summer ventilation rate should be at least 10 times
greater than the minimum continuous rate. This is the prescriptive approach
of the American Society of Heating, Refrigerating and Air-Conditioning
Engineers (ASHRAE); current ventilation standards are America National
Standards Institute (ANSI) and are ANSI/ASHRAE Standard 62–2001
(ASHRAE, 1999) plus ASHRAE BOD approved for publication Addenda
and ASHRAE Standard 62.2-2003 (ASHRAE, 2003). In fact, these docu-
ments provide techniques to establish minimum ventilation requirements
(liter/sec) per person or per surface unit (liter/sec per m2), but the minimum
values must be specifically estimated according to specific conditions, such

as internal and external temperature, relative humidity, and recommended
CO2 (or a specific targeted pollutant) concentration. Anyhow, just to give
an order of magnitude, the continuous rate may be 10 liter/sec per person,
at 22°C and 40% relative humidity. ASRHRAE standards may be a useful
start-up approach for each food industry internal environment, if the gas-
problem concentrations within the building must be kept at a known level
during the summer in a properly designed facility. It worth reminding that
the creation of ASHRAE standards and guidelines are recommendations,
determined solely by their usefulness, hence, the conformance to them so
far is completely voluntary.
The use of forced air supply and exhaust for worker space areas (forced
air supply together with exhaustion system) provide the best ventilation
coverage of enclosures. Equipping the makeup air supply and exhaust
registers with volume dampers to control the rate of air. Volume dampers
are useful with field flow balancing the ventilation system. Makeup air
flux supply should be somewhat less than exhaust to create the negative
air pressure within the enclosure.
Air changes are defined two different ways: First, how many times the air
in the space circulates through the mechanical equipment. Second, how
many times the air in the space is replaced with outside air.
Finally, another perspective, ACE (that is, the ratio of the effective ven-
tilation rate at the breathing zone divided by the effective ventilation rate
which would occur with perfect mixing while at the same ventilation rate)
shall be at least 1.0 to 1.2.
In summary, ventilation rate procedure prescribes:
the outdoor air quality for acceptable ventilation;

outdoor air treatment when necessary, adapted to ventilation rates for
residential, commercial, institutional, vehicular, and industrial spaces;
criteria for reduction of outdoor air quantities, when recirculated air is
treated by contaminant-removal equipment;
criteria for variable ventilation when the air volume in the space can be
used as a reservoir to dilute contaminants.
7. MAIN PREVENTIVE TECHNOLOGIES TO CONTROL

ODOR EMISSIONS
Malodorous emissions are never isolated problems of the food industry envi-
ronment. Indeed, odorant emissions only reflect environmental mismanagement
faults and procedures.
84 R. Nabais
According to the United Nations Environment Program (UNEP) (Penn-

sylvania State University, 2002; UNEP Working Group, 2004a, b) in what con-
cerns general environmental industrial management and cleaner technologies,
there must be a consideration of several issues (multicriteria system), which
despite the difficulties in being correctly conceptualized, may result in very
positive impacts on odor emissions abatement. Among possible ways are auto-
matic control systems, best practices implementation, energy control, financial
planning, features “housekeeping,” improved chain management, inventory
management, key performance indicators, preventive maintenance, materials
accounting, modify inputs, onsite treatment, primary screening/segregation,
process control/improvement, recycling, redesign reuse, wastes management,
and storm water Management. The great difference between pollution control
and cleaner production is that pollution control is an after-the-event, “reactive
actions” approach, whereas cleaner production reflects a proactive, “anticipate
and prevent” philosophy.
Most importantly, performing good practices in those aspects afterwards
can be proudly announced as promotional advertisements for manufactured
products or services. Companies therefore are adopting a more systematic
approach to environmental management, sometimes through the environmental
management system (EMS).
EMS is a decision-making structured tool to project an action program,
bringing cleaner production into the company’s strategy of development, man-
agement, and day-to-day operations; in fact, it provides a continuous globally
integrated approach to progressively improve the industrial activities by EMS,
depicted in Figure 1.
An example of an element of this integrated management model, but applied
within a scale of a regional development strategic vision, may be achieved by
strategically installing a biorefinary, projected to become a friendly technology,
materially and energetically sustained, which enables exploitation of biological
raw and natural materials in the form of green or processed, wet and waste wet
biomass from a targeted sustainable the regional land utilization, representing
the interesting concept of industrial symbiosis (Lifset, 1999; Realff and Abbas,
2003).
7.1. Production Shifting, Reformulation, and or Retrofitting
Production shifting, reformulation, and retrofitting, said “frontend” control mea-

sures, are best received by industries, because they should be implemented only
after favorable or unfavorable economic return is correctly estimated. If careful
studies are done, the final results are usually very positive for both perspectives:
profits for the company’s stockholders and also important environment gains
(Pinchin, 2004). Just as an example, in the case of fisheries there are several tips
to consider when processing poor quality fish, such as (1) the temperature in
6. Continuous
improvement
5. Managing
review
1. Environmental
principles
2. Planning
3. Implementation
and operation
4. Checking and
corrective actions
Figure 1. General sketch of an environmental management system.
indirect steam dryers must be controlled to avoid scorching of the fish (overdry-
ing, losing weight) and amplifying odor emission problems; (2) increasing the
pressure of the presses will improve the recovery of malodorous constitution
solutions; and (3) the size of the holes in the pre-strainer must be increased,
because if the fishery processes poor quality raw materials, these tend to produce
a slurry that clogs up the outlets of the presses, diminishing its efficiency.
7.2. Storages and Warehousing
When handling material that is date sensitive or perishable, which is the case in
the food industry, several requirements must be foreseen:
The storage spaces must be clean and carefully organized. All points of
the storage must be easily accessible and offer safe conditions for fre-
quent inspections and surveillance procedures; refrigeration units should
be cleaned and maintained on a preset periodic basis.
For refrigerated and frozen products, temperature and moisture is crit-
ical. Given the effects of temperature on the safety conditions of bio-
logical goods and raw materials, alarm devices are necessary to monitor
temperatures.
86 R. Nabais
If refrigeration systems are used, precautions must be taken to avoid drips

over the products or raw materials.
First-in-first-out rotation (FIFO) principle is the removal of inventory
from storage in a systematic way where earlier stock items are used first.
This can be accomplished by date coding the inventory according to the
date of receipt. Warehousing, mainly, involves three activities (receiving,
storage and shipping) that are included in a quality control program. The
receiving operation is the foundation for processing finished food prod-
ucts of a designated quality (GDV, 2003 and Tybor et al., 1995). The
main recommendations, to improve organization of stocks: (1) minimize
queues’ waiting periods and dock time; (2) move refrigerated or frozen
items directly into storage; (3) compare the incoming shipments with the
order specifications; and (4) date code all incoming shipments directly on
the container or pallet load for stock rotation.
Certain supplies or ingredients may require segregation (e.g., acetic acid,

sodium hydroxide), even for safety recommendations; as a rule one should always
consult material safety data sheets (MSDS). The MSDSs were conceived, devel-
oped, and determined by the Occupational Safety and Health Administration
(OSHA) in their Hazard Communications Standard 29CFR 1910.1200. This
standard is commonly referred to as the “right-to-know law”, requiring compa-
nies to provide specific information about ingredients, hazards, and precautions
on all hazardous material labels. The MSDS must be developed by the product
manufacturer and is mandatorially sent, whenever requested, if not promptly
sent with the product.
Shipping is the final step in which a food business can have direct control
on product quality. Ship items on a first-in–first-out basis and use the same
guidelines in shipping that you followed in receiving. If not properly managed,
storage items may become ruined and smelly. Figure 2 displays an improper
condition of biological residues storage because the bags are not closed. For
example, onions have an unpleasant, pungent odor. An increase in odor levels
indicates improper storage that promotes self-heating, which is associated with
the loss of essential oils, sugar, and vitamins. Damage to onions, caused by
setting them down roughly, for example, results in greater respiration intensity.
The cleaning and sanitizing operations should address three basic areas: inter-
nal facility including floors, walls, ceilings, and ventilation system; exterior
facility; and grounds.
7.3. Process Modification
A good example of a process modification can be evaluated by the following

description. There are two basic types of yeast dough mixing processes used in
bakeries: sponge dough and straight dough. For the purpose of estimating emis-
sions, the length of the fermentation time is the critical difference between these
Figure 2. Residues improperly maintained in open bags, waiting for reutilization or destruction.
Permission by courtesy of Portary Gestão de Resíduos Lda (2004).
two processes. It is during the fermentation process that the Volatile organic
compounds (VOC) are produced. The sponge dough process, which is most
commonly used by commercial bakeries, produces the largest amount of VOC
emissions because the required fermentation time can be 5 hours or more. By
cutting spongeous fermentation period, from 4 hours to just 1, through use of
a preferment concentrate, the bakeries can lower ethanol emissions by half and
more.
In a British bakery, two depanners use a vacuum system to remove bread
from the baking trays and tins. The vacuum is generated in vacuum chambers
by fans driven by standard 15 kW AC motors. With these improved conditions,
having a payback period of 1.3 years, technical and economical advantages were
registered: (1) energy savings: about 135,000 kWh/year; (2) reduction in CO2
emissions: 67,500 kg/year; (3) damage of bread rolls were minimized; and (4)
lower overall maintenance cost was obtained.
Also, to control VOC emissions from commercial and industrial boilers, no
auxiliary equipment is needed; properly maintaining the burner/boiler package
will keep VOC emissions at a minimum. Maintenance services must include keep-
ing the air/fuel ratio at the manufacturer’s specified setting, having the proper air
and fuel pressures at the burner, and maintaining the atomizing air pressure on
oil burners at the correct levels. An improperly maintained boiler/burner pack-
age can result in VOC levels over 100 times the normal levels of specific pro-
cesses, and the energetic efficient decreases.
Another good example in the food industry is deep frying processes. It is
important to choose oil with a high smoke point (the temperature at which it
starts to smoke). When oil smokes, it begins to decompose and, some experts say
many of its unsaturated fatty acid molecules become saturated. Frying oils are
exposed to atmospheric oxygen and moisture at elevated temperatures during
88 R. Nabais
frying. The oil gets deteriorated, as the triglycerides in the oil undergo oxidative,
hydrolytic, and thermal alterations producing more than 400 various chemical
compounds.
Breakdown of the glycerol molecules in the fat creates acrolein, an obnox-
ious-smelling compound that easily inflames. This means that perfectly safe fat,
with a flash point of 260 °C, might burst into flames at a dangerously low tem-
perature after having been used, just once in high temperature.
Avoid adding salt to food before frying, because salt draws moisture to the
food’s surface, which will splatter when the food is added to the hot oil. Salt also
lowers the smoke point and breaks down the oil.
7.4. Capturing, Conveying, Confining and Removing Odors
The most common odor management control for wastewater systems is a float-
ing cover made from polyethylene (PE) membrane suspended on cables and/or
supported by floatation devices, with perforated gas collection pipes, weighted
sunken troughs to collect water, sump pumps, and so on. The main function of
such cover is to prevent vector migration of vapors or gases into the atmosphere
and to collect methane gas to be disposed by flaring or other means to reduce
odor pollution. No matter how elaborate installation procedures and the meth-
odologies, efficiency, the puncture of PE membrane, wind, hail, snow, ice vulner-
ability, explosive hazards due to static electricity and accumulation of volatile
gases, vandalism, dependency on electric power supply and reliance on profes-
sional maintenance of electric motors and pumps contribute to a substantial
initial investment and a high operating cost.
However, the best technology in the world or the most costly odor control
system can only be effective in reducing odorous emissions if the odors reach the
control devices. If the containment and ventilation systems are not able to keep
the nuisance smells from escaping, then the money spent on the odor control has
been wasted.
7.5. In Situ Compounds Recovering
Solving malodorous emissions constantly implies improvements in industrial

performance, greater efficiency in resources consumption, and reduction in
the levels of waste. At the microlevel, the industrial ecology concept demands
close attention to unit processes, facility operations, and industrial behavior
and organization.
The food industry emits gaseous streams, in many of their processes, such
as storage vents, distillation vents, reactor vents, mixers, pan coolers, spray dry-
ers, and granulators.
Gaseous compounds may be separated and recovered from the whole gas
stream by absorption and adsorption, distillation, extraction and refrigeration
or condensation based systems. Of the techniques used in the final separation of

gaseous compounds, Table 4 shows comparison of three different vapor recovery
systems.
Adsorption plays an important role in industry, both in the removal of con-
taminants from a product and as a direct means of recovery of substances. Its most
important application is in the removal of organic contaminants from polluted
sources.
Another example is the case of citrus oil, produced by pressings citrus
peels (to decrease liquid content before the peels can be reused, i.e., pellets pro-
duction) which is usually in a form of emulsion of water oil; the oil may be
recovered by vacuum distillation process or fractional distillation if we desire
different grades of oil. With this procedure, the liquid effluents will have a lower
biological oxygen demand (BOD) concentration, acquiring better adaptation
to be biologically treated, reducing malodors emissions, both from liquid and
gaseous sources.
Table 4. Comparing Performances of Refrigeration, Absorption, and Adsorption for

in situ Compounds Recoverya
Refrigeration Absorption Adsorption
Application Higher concentration Low concentration—0.5% Low concentration—0.5%

process Condensation Separation and Separation and
condensation condensation
Operational temp. Very low temp. Room and above Room and above
Pressure drop Very low Low High
Vacuum Not required Required Required
Operations Fully automatic/ Full automatic/ Automatic batch process
continuous continuous
Maintenance Low maintenance Low maintenance High maintenance
Power cost Power used when Continuous Continuous
required
Installation cost Very low Low Large
Concentration Not required Required for outlet gas Required for outlet gas
measurement
Space required Smallest Small Large
Steam Note required Required Required
Inert gas Not required Special circumstances Required for safe
operations
Carbon Not required Not required Required periodic
replacement replacement
Solvent Not required Required Not required
replacement
Safety Intrinsically safe Safe Precautions required
a
From Mohunta (2004).
90 R. Nabais
7.6. Managing Onsite Good Technological Practices in All Industrial Process:

Preventing and Suppressing Odorous Emissions
Whatever may be the industrial manufacturer, good practices always require

eliminating unintentional holes in buildings and keeping doors and windows
closed, avoiding the storage of odorous materials outside the building and the
transport of materials between buildings in open containers, and so on. Con-
tainment of highly odorous process gases, i.e., keeping it separate from less
odorous streams, may reduce the capital and operating costs of the required
abatement system(s). As a general rule all odor control systems should exert
a negative pressure locally or within process buildings to prevent odorous air
leaking out.
The most cost-effective approach to odor control may be to examine the
operation and maintenance practices at the processing facility and to introduce
prevention pollution techniques such as recovery, product design changes, tech-
nology or process modifications, input or material changes, in-process recycling
and reuse, and good operating practices, e.g., waste segregation, preventive main-
tenance, training and awareness programs, and production scheduling (Holland
et al., 2004).
Automation of transfer processes and computerized raw material handling

and pumping system, totally enclosed, help maintain a cleaner environment
(Tybor et al., 1995).
Installation of removable insulation on–off valves, pipes, and fittings
to reduce losses in the process heat distribution system, with potential
energy savings of 2–5% and prevent scale accumulation by ensuring water
treatment systems are operating effectively.
Scale buildup in boiler feed water tubes inhibits both throughput and
heat transfer, with potential gains in boiler efficiency of 10–12%. When-
ever heat production increases its efficiency, decreases the combustion
operations, and hence potential gaseous emissions also diminish, and in
consequence all gaseous pollutants decrease as well.
A comprehensive option for reducing odor was proposed by one rendering

industry, involving a process change to replace the existing dryer with a steam-
heated indirect dryer, improvements to ventilation of the rendering spot, and
replacing the scrubber with a biological filter.
Whatever may be the alternatives to reduce or solve odor problems, there are
not many options outside of the schemes suggested by Figure 3 and Figure 4, as
follows:
Analysis of the Figure 3 shows that we only have two options to deal with
industrial smelly problems, but many times they are complementary to each
other: emission and reception. Malodor control may be addressed through the
following alternative ways:
Odor formation process
Transfer to air
Release to atmosphere
Dispersion
Frequency, Duration, Exposure

Intensity of Exposure
Detection and perception

Time of the day, activity
context, relation to source
and association with
odor Appraisal by the receptor
Other ambient stressors
Individual perception
(noise, dust crowding)
Annoyance/Nuisance
Access to complaint channels,

Legal instruments and
Expected results
COMPLAIN ACTION
Figure 3. Origins of odors complaints. From Van Harreveld (2001).
Reduction of odor production, extraction of ventilation air with treat-

ment to reduce odor concentration. Using miscellaneous additives that
can react or mask odors.
Reduction of transfer rate (i.e., for instance pH, temperature emissions
control, covering the surface).
Reduction of exposed area of slurry, including storage, soiled surfaces,
grids, and so on.
Variable dilution in the atmosphere through turbulent dispersion (turbu-
lence or stability of the boundary layer, wind direction, wind speed, etc.).
Adapt the emissions to the neighbors’ characteristics.
Indeed, it is clear that high waste production level is an indication of process

inefficiency and losts of profits, which precisely shows the expenditure of valu-
able resources to produce a product that cannot be sold: waste.
Figure 5 displays the relative distribution of preferences calculated, over
8 fundamental principles, considering 123 different specific clean technolo-
gies options, preferred by 48 food industries, located in 21 different countries,
resulting from an inventory realized by the International Cleaner Production
92 R. Nabais
Raw materials
Extraction and
beneficiation Manufacturing Consumers
process
Residues/Emissions
Recycling Retrofitting Reusing
Environment impact
Environment (Externalities-
SocialCosts)
Figure 4. Cleaner production principles.
Information Clearinghouse (ICPIC), which underlines the high importance

of process modification and housekeeping improvements, in what concerns
cleaner technologies. Housekeeping involves special training, improving
management to avoid leakages, periodic maintenance of equipment, strict
control of washing water to reduce wastewater, reuse of cooling water where
possible, obtaining guarantees of raw material quality, and improving raw
material storage.
7.6.1. Process Modification, Recovery, Reuse, and Recycle
Some food processing wastes such as tomato and apple pressings, corn cobs
and husks, sweet corn cannery waste, stale bakery products, and spent brewery
mash may be used as feed complements to cattle, helping farmer to reduce
Reuse
13%
Housekeeping
24%
Recycle
13%
Material
Substitution
2%
Recovery New technology

13% 9%
Product
Modification Process
2% Modification
24%
Figure 5. Relative frequency of cleaner technologies principles used for pollution abatement in food
industry. From International Cleaner Production Information Clearinghouse (2004).
overall feeding costs. Also, feeding these recyclable wastes to animals reduces
the volume of solid waste that food processors must dispose of at an approved
sanitary landfill or apply as a soil amendment.
In conclusion, any material that does not become a part of the final products
must be promptly recycled into or off the process, reuse to some other by-product,
or ultimately adequately disposed. The costs of odors treatment can be minimized
by focusing on the previous options and cleaner production.
7.6.1a. Reduce. The next step is to minimize the use of all materials in the
process. This can include reducing errors in batch preparation, optimizing clean-
ing operations to reduce the volume of water used, and turning off equipment
that is not in use.
7.6.1b. Reuse. There are many opportunities to reuse waste products in the
food manufacturing industry. This will reduce the demand for raw materials and
the cost of treatment and disposal. It may be possible to reuse some clean pro-
cess waters in the boiler or install drip trays to catch product that eventually
may return to the process in next batches. To enable biobased materials produc-
tion in the food industry, it is only necessary to obey preestablished rules of
good waste management. An interesting example in this domain is Cargill Dow’s
polylactide (PLA), which is a versatile biopolymer, made from 100% renewable
94 R. Nabais
resources like crop residue (stems, straw, husks, and leaves) from corn or other
crops (Vinka et al., 2003).
7.6.1c. Recycle. Are the wastes identified by your assessment really wastes?
Can some of these be reclaimed through a simple treatment process, which
enables them to be reused on-site. Other by-products that cannot be used on-site
may be recycled off-site. In these cases there may be the potential to sell recy-
clable items and also save by the avoiding disposal costs.
7.6.1d. Treat and Dispose. This option should only be considered after the
other options have been exhausted. Generally these options are typically a cost
to industry. However, it may be essential to consider this as a part of your overall
cleaner production strategy. The costs of treatment can be minimized by focus-
ing on described options, which may be specifically applied to different food
industries and services (UNEP, 2004a,b).
According to the Ontario Ministry of Agriculture (1995), many of those food
wastes are wet and/or decompose readily. Leachates and polluted or contaminated
runoffs may easily develop problems with odors. To avoid such problems, farmers
should tailor handling and storage arrangements to their situation and the waste
or by-product they use. Some general guidelines to minimize odors and water
quality problems include:
Provide adequate storage facilities (preferably covered) appropriate for

the material. Locate storage facilities away from and where possible
downwind of neighbors’ residences and buildings.
Wherever possible, keep the volume of stored waste to a minimum and feed
the material to animals as quickly as possible; this is especially important
for wastes that decompose readily or that produce particularly strong odors
(e.g., vegetables and fruits in general, such as onions and cabbages).
Provide the feed in a trough or receptacle that will allow all the feed to be
eaten rather than spread into the ground.
Avoid water quality problems through collections; store and properly dis-
pose of leachates or liquor from the stored material; they can be heavily
concentrated pollutants, and give rise to smelly emissions.
7.6.2. Enabling Technology
Summarizing what has been said, several alternatives should be considered

when choosing technological options to environmental integrated manage-
ment system (United Nations Environment Programme, 2004): (1) Good oper-
ating practices (e.g., waste segregation); (2) improved housekeeping (e.g., keep
containers covered); (3) material substitution (e.g., water-soluble or no-clean
fluxes); (4) product redesign (e.g., design for the environment); (5) product
reformulation; (6) in-process recycling; (7) process or equipment modifications;

(8) recycling and recovering wastes (e.g., biodiesel production from wastes).
An example of a recycling option, for potential smelly residues of edible
oils and fats, is to collect and reprocess them to produce biofuel. Previously, it is
necessary to remove any water and/or water vapor bubbles present in the waste
oil by heating the waste oil at 105 °C. From all the industrial processes two main
products result: methyl esters, the chemical name for biodiesel, smelling like
French fries, and glycerine (a valuable by-product that can be used in soaps and
other products), with the additional advantage of the possibility of easy biodeg-
radation. Smelly waste refuse of vegetable oils, used for frying, is an attractive
source of biodiesel, but also is difficult to convert because it contains 2–10% free
fatty acids (which are, precisely, the cause of the rancid odors and taste).
7.6.3. Preventive Managing Procedures
Cleaner production aims to prevent pollution, reduce the use of energy, water,
and material resources, and minimize waste, profitably and without reducing
production capacity (Environmental Protection Agency, 2000; ICPIC, 2004;
Pennsylvania State University, 2002; Robert et al., 2002). Strict inventory con-
trol, e.g., minimizing expired shelf-life materials in storage dependences and
avoiding long periods of permanence of the products in the expedition spaces,
is good practice.
Pollution prevention processes are one of the major procedures of cleaner
technologies. Cleaner technology main tasks are:
1. Energy conservation.
2. Water conservation.
3. Maintenance of the equipment in good repair.
4. The cleaning and sanitizing procedure prevents the build up of dirt
and debris. Cleaning and sanitizing must address three basic set points:
exterior facility access, pavements, grounds, and connections to sewers;
internal facility and utilities, including floors, walls, ceilings, and ventila-
tion system; and drains and equipment.
As illustration, one of the important issues for many food industries is oil and
grease residue handling, their removal or treatment systems, and final discharge
or disposal devices; in fact, grease tends to solidify downstream, causing frequent
blockages and/or backups. They also may cling to wastewater ducts and reduce
their flow capacity in the long term. All degreaser equipment only separates oil
and grease onto the surface of the bulk effluent and at best relocates it.
By accidental overflow onto pavements, in many cases it remains as an oily
film that poses a risk for slips and falls. Then, there is the problem with cleaning
and disposal. On the other hand, oily waste cannot be dumped down the drain,
because haul-off can be a very difficult and expensive consequence.
96 R. Nabais
Interior grease traps should be evaluated for cleaning needs on a weekly

basis. Frequency of and proper maintenance is the only sure way to know if the
traps are operating properly. Using hot water (120°C or higher) to wash or rinse,
the grease stays emulsified, not allowing its separation. Many food treatment
facilities require cooling jets that are activated to cool the water before entering
the interceptors. Even if the grease amount is not much, as it cools it add and
accumulates on the interceptor pipe walls and causes a plug.
To really remove oil and grease, gravity separation may be used, provided
the oil particles are large enough to float toward the surface and are not emul-
sified. If this is the case, the emulsion must be first broken by simultaneously
decreasing of temperature and pH.
Another special recommendation for food industries is to assure proper
drains, and if odors develop, to reduce the amount of standing water of the pave-
ments and floors and by changing water-based cleaning process to dry systems.
In general, reducing biological solids content (soluble or insoluble) of waste
waters produced by any food industry is always a good principle, to achieve
potential pollution abatement and the unpleasant odors emissions resulting
from their uncontrolled biological degradation.
7.6.4. Planning Rigorous Maintenance Proceedings
Inspection and maintenance programs embody the idea that some organizing
procedure mechanism must be put into place to ensure that all installations and
equipment must be kept in nonpolluting condition. Measuring resource effi-
ciency is the key to making incremental improvements in performance. Examples
of environmental performance indicators, for companies, are the amount and
the type of energy used, carbon dioxide emissions, water consumption, effluent
discharges, waste management, and environmental incidents registers.
Avoid leaks from gaskets and seals. Spill, splash discharges, and leak preven-
tion is one good tool to prevent odors emissions, while usually also meaning eco-
nomical gains (e.g., these interventions must take part in good practice checklist
procedures of any preventative detailed maintenance plan).
7.7. Special Measures : Developing Correct Urban Development Plans

and Adequate Layout Design of Buffer Areas
Proper land use planning avoids allowing incompatible land uses in close prox-
imity and is one of the most important tools in odor management. Many odor
problems can be avoided by appropriate placement of new facilities.
The lack of consistent methods to deal with odors prevents authorities from
establishing policies and standards, which would eliminate, or at least minimize,
community odor nuisances [Danish Environmental Protection Agency (DEPA),
2002; New South Wales Department of Environment an Conservation (NSW),
2004; UNEP Working Group for Cleaner Production in the Food Industry
(UNEPTIE), 2004a,b; Robert et al., 2002].
According to ZERI (Zero Emissions Research Initiative, of United Nations
University, 1994), in the future industries will be organized into clusters in a way
that waste from one industry becomes input for other industries, forming an inte-
grated system. This simple, genial concept imitates natural ecosystems, where no
real wastes exist. Industrial ecology shares with ZERI by anticipating the environ-
mental consequences of production, consumption, and waste management, inte-
grating overall balanced activities for ecoindustrial parks and product life-cycles,
“taking a systems perspective on the environmental consequences of production
and consumption” (Lifset, 1999; Robèrt et al., 2002).
It also is urgent to ensure compatibility of industrial activities with adjacent
land uses, adequate separation distances to allow for process upsets, and adopt
the principles of waste minimization, cleaner production, and best practice con-
trol technology. These objectives-oriented interventions will minimize or elimi-
nate the release of odors from the site, strictly oriented by the specificity of the
industrial activity (UNEPTIE, 2004a, b).
Avoiding unregulated development zoning serves principally to protect
property owners from the negative externalities (social costs) of new develop-
ments. The zoning plans are based on the following considerations:
Designated land use of the site and the surrounding land under the dis-
trict plan.
Confine location of activities within the site and their orientation in rela-
tion to prevailing winds.
Consider sensitivity uses of the downwind receptors.
Respect buffer distances from the site boundaries to sensitive land uses.
Use screening, such as by earth bunds, shelter belts, or natural topogra-
phy, that functions as obstacles for odors transference.
Odor is transported primarily by dispersion and diffusion. Dispersion is the

gradual dilution of the odor by mixing it with ambient air, transporting it with
the bulk flows assisted by the wind. Once odors are released, they are transferred
into the surrounding airspace, where air currents and turbulence disperse and
dilute them.
Besides its concentration, the noxious effect of odorous gases in the air
depends if they are in mixtures, its composition, moisture content, temperature,
pH, oxygen concentrations, and weather and environmental conditions (season,
wind patterns) wind direction, wind speed, weight of gas and adsorption to dust
particles.
Ammonia (NH3) is lighter than air; results from anaerobic or aerobic activ-
ity; and is soluble in water. Hydrogen sulfide (H2S) is heavier than air; it presents
a low odor threshold and a high solubility in water; one serious problem of this
98 R. Nabais
gas is its toxicity increases with higher concentrations, while simultaneously it

loses its odor character.
Weather patterns, humidity, and temperature largely determine odor trans-
port and detection, all of which can change with the season, the day, or even
the time of day. Warm temperatures and high humidity increase the potential
for odor nuisances, while cold, dry conditions reduce the potential for nuisance
complaints. Odors disperse better under warm, windy conditions compared with
still conditions. Odor generally does not disperse much early in the morning
or early in the evening, particularly in cool weather. Odorous flow also tends
to drain down the valley (katabatic drift) from areas of higher altitude and/or
higher pressures to areas of lower altitude and/or lower pressures (New South
Wales Department of Environment and Conservation, 2004).
Considering layout implementation, effective odor control depends on four
factors:
1. good planning and good industrial management;

2. good relations and communication with neighbors;
3. appropriate control and treatment technology selection; and
4. sensitivity of the region.
New facilities should be carefully sited so that prevailing winds do not carry
odors to sensitive neighboring sites. However, no setback distance will work for
every site. Setback requirements are site-specific and should be calculated using
complete information about local conditions. For example, setback calculations
should include the size and type of operation, the facilities covered by the setback
(such as lagoons, fields, and houses), as well as the proposed sites and methods
for applying wastes to land. Local vegetation, prevailing winds, weather pat-
terns, and neighboring land uses also should be considered in the calculations.
An efficient planning option is confining odoros activities to certain districts
(by effective zoning regulations). Examples of sensitivity activities are residen-
tial dwellings, visitor accommodations (hotels, bed and breakfast, guesthouses),
hospitals or nursing homes, schools, churches, holiday and weekend dwellings,
shelters, campsites, caravan parks, and sports facilities.
Figure 6 gives an idea of how an emission source can have a favorable lay-
out previously roughly settled, by considering its location related to the relative
frequency of dominant wind-flow directions.
The setback distance, or buffer zones, of implantation of the facilities must
considerer not only the frequency and intensity and dominant wind, but also the
exposure angle of the facilities (Heber, 1997; Schauberger and Piringer, 1997;
Pennsylvania State University, 2002). Several European countries have established
such setback models, such as the known Austrian guideline that is based on
quantitative odor evaluations and neighbor surveys and is a tool to assess air-
borne emissions resulting from livestock husbandry and the pollution caused by
these emissions.
N
100
80
NW NE
60
40
20
W 0 E
SW SE
Figure 6. A hypothetical way to draw a wind pattern relative to frequency and regional vicinities
borders (….), of a well-located emission site related to wind-orientation dominance, spotted in the
center of the winds-rose.
Percentile lines are commonly used to describe the odor concentrations on a

map of the surroundings of the plant. Percentile lines (or isoconcentration lines) con-
nect places in a contour plot at which a certain threshold concentration is exceeded,
during a certain percentage of the year; this information may be used to indirectly
estimate odor concentrations, which will be mentioned in the next chapter.
Properly planted and developed, dense vegetable curtains can filter dust
particles and reduce or modify wind direction. Fast-growing evergreen trees
have good windbreaker properties in buffer zones, around an odorant unit, and
help disperse odors. The planting of trees around facilities that emit odorous or
VOC compounds is currently being used for odor abatement. Trees will help to
a small extent to capture the odorous compounds and particularly the dust. On
the other hand, vegetation curtains provide a visual barrier that reduces odor
perception for residents and also, under adequate cold weather, may serve as a
condenser element for organic vapors.
Technically, one must definitely realize that tree barriers are not the solu-
tion, but up to a point they contribute to soften the opinions about food industry
activities as being dirty, smelly, noxious and undesirable. Besides if they are cho-
sen among pleasant odorant species, their odor may hide or overcome unpleas-
ant smelly emissions.
Organic matter decomposes anaerobically in standing water, releasing odors
when disturbed or tracked onto public roads. The landscape should be properly
graded and maintained so that water does not stand in access roads and around
production facilities.
An interdisciplinary team should establish these criteria in cooperation with
industry, local communities, and the appropriate environment policy authorities.
100 R. Nabais
Using geographic and meteorological data, maps can be drawn up to identify lev-
els of exposure and predict odor annoyance impacts on the residents near those
areas.
For each of the currently available management options described in this
chapter, performance criteria should be developed and data collected so that
management improvements can be recommended.
In summary, odor management strategies are shifting their approaches,
from old systems—complaint driven—to the new systems based on anticipation
measures and precautionary inspection systems.
7.8. Using Conventional Plume Dispersion and/or Virtual Chimneys
Odors are often low-density gases. Once released into the environment they are
transported by wind and diluted and dispersed by atmospheric turbulence.
According to theoretic common concepts, odorous gases require special consid-
eration and treatment. Warren Spring Laboratory in the United Kingdom (Williams
and Thomson, 1985) suggests the following technique to give rough estimations of
the uncorrected chimney height hu (meters), by one of two alternatives:
hu = (0.1DQ)0.5
where
D is the number of dilutions or odor units required to disperse the emis-

sion source to the TOC50%, and
Q is the volumetric flow rate in m3/sec at 0°C and 760 mm Hg (1 atmo-
sphere standard, or 1.013 bar)
Another equivalent expression is:

hu = (0.1Mo/TOC50%)0.5
where
Mo is the mass emission rate of the odorous gas in g/sec.

TOC50%is the result of an odor panel test giving the number of dilutions
(odor units) required to disperse the odorous source to levels below the
detection threshold odor concentration, for 50% of the odor panel.
It is worth a reminder that these two equations were empirically developed;

therefore, they must be applied only respecting the units and the dimensions of
all variables.
The detection threshold levels normally considered are 5, 10, 20, and 50
OU. The level 5 OU is considered to be the minimum for detection in the open
environment, where inhabitants are subject to a wide range of natural odors and
consequently these levels correspond to worst-case situations. The 10 and 20 OU
are levels that, if occurring frequently enough, may be sufficiently detectable to
constitute a nuisance. Finally, investigation of a 50 OU detection threshold gives

information about the extent of high levels of impact.
It is generally not practical to disperse more than 200 OU/m3 per sec of
exhaust flow without correctly designed chimneys. Anyhow, dispersion of odor-
ous emissions via a chimney is not recommended, unless a well-trained odor
panel previously evaluates the emission source.
Wind is responsible for the rapid horizontal transport of humidity, warm
air, pollutants, and odors, while turbulence is responsible for vertical transport.
Wind turbulence can be visualized as eddies of different sizes that cause fluctua-
tions in concentration over short time intervals.
Internal concentration variations within a plume are created by puffs of
air. The turbulence caused by these eddies diffuse the plume. Large eddies cause
meandering of the plume near the source, normally in a lateral direction. As
the plume is torn by the introduction of pure air puffs, a concentration profile
develops within the plume.
Rough terrain, valleys, and other topographical features can increase the
complexity of airflow patterns. For the usually very low concentration of odor-
ants pollutants in the food processes emissions, and because if odor criterion
needed for the analysis is of the order of seconds, then the time-averaged formu-
las could underestimate a shorter-term peak odor effect. Most odor sampling,
however, is also time-averaged to collect sufficient sample for analysis. This
makes validation of proposed averaging-time adjustments difficult. Straight
standard Gaussian plume model does not accurately simulate atmospheric pol-
lutant transport over rough terrain profile, irregular slopes, and surface singu-
larities, as vegetation or protective fences effects.
If advisable, but never if potential health nuisance consequences are pre-
viewed, enhanced dispersion is a cost-effective option, which may be achieved by
increased exhausting stacks height, increased exhausting gas velocity, or moving
the source away from critical reception spots.
A promising proprietary technology/device that deserves close observation,
both in terms of theoretic principles and technologic instruments, is “extended-
stacks” or virtual chimney: a wind-based odor diversion system, simulating the
natural transport phenomenon caused by the wind. The odorous gaseous mass is
blown vertically through the atmospheric layers while diluting the odors (Finn et
al., 2004). This curious proprietary patented system theoretically seems to work
beautifully, by the simplicity of the basic theory: when one increases longitudi-
nal transfer velocity in a “flowing wind-polluted tube,” the pressure inside the
axe of the “emission wind tube” will decrease, according with Bernoulli equa-
tion. The horizontally pointed outer ring of fans blows inward at a high veloc-
ity, creating an internal low-pressure zone. The atmospheric air, from the outside
environment of “the windy polluted tube,” attempts to equalize this internal
underpressure of the windy tube, so the fresh air continuously moves toward
the center low-pressure zone. The fresh air mixes with the odorous molecules,
diluting them. The diluted air stream is ejected vertically upward, by one vertically
pointed fan in the center. This continuous process not only (potentially) causes
102 R. Nabais
massive dilution, and simultaneously “virtually increases a fluid stack height” (to
discharge odorous gas stream as far as possible, from the ground level).
Enhanced spatial dilution of odorants, as the technique just described, must
not be considered to be a “miraculous” solution to any air pollution problem,
especially if it is considered to be a nuisance. If pollutant concentrations meet
all emissions regulations and the legal compliance requirements are met, but
the emissions remain a malodorous nuisance to the surrounding neighborhood,
further dilution of the pollutants should be sought.
8. CONCLUSIONS
By accomplishing environmental management system integrative principles or

at least all achievable pollution prevention procedures, industries and commu-
nities may prevent expressions of pollution, including odor disturbances, thus
minimizing the need and the costs for control, treatment, and disposal. Indeed,
these are still the best options to rethink traditional industrial processes; hence,
we must support environmental policies that require creative ecoefficient strat-
egies to address the continuous improvement, sustained development, neces-
sity of organization’s public commitment to environmental compliance and
protection, and ultimately, the assurance of our own quality of life.
9. AKNOWLEDGMENTS
The author deeply expresses gratitude to: Doctor Xavier Nicolay, of Institut
Meurice–UBT, for his strong partnership conscience, team responsibility and
consistent coordination, in putting together this “many-hands–multicultures
knitted book”; but, specially, for his kind understanding, restless support, car-
ing spirit, endless patience, attentive readings, helpful and valuable advices about
this chapter content. However, copyrights remain the author’s responsibility for
all the still remaining imprecision or errors.
10. REFERENCES
Anonymous (2002) Cleaner production tools: LCA and beyond. Editorial. Journal of Cleaner
Production 10 403–406.
ASHRAE (2003) HVAC Applications Handbook. (IP). Item 210–8103-03. Publisher: ASHRAE.
ISBN: 1–931862–22–2.
ASHRAE (1999) Ventilation for Acceptable Indoor Air Quality, Standard 62–1999, Atlanta, GA,
American Society of Heating, Refrigeration, and Air Conditioning Engineers.
Collins, R. (2000) CSPI PROJECT “Science Conflicted: Restoring Trust in the National Academy of Sci-
ences”. Science Conflicted: Restoring Trust in The National Academy of Sciences”, Integrity in Science
Project, Center for Science in the Public Interest. Baltimore sun Journal. Editorial of August, 2000.
Danish Environment Protection Agency (DEPA) (2002) “Industrial odour control.” Environmental
Guidelines No. 9 2002. Vejledning fra Miljøstyrelsen Environmental Guidelines no. 9, 2002.
http://www.mst.dk/udgiv/publications/2002/87-7972-297-0/html/helepubl_eng.htm
Environment Agency for England and Wales in collaboration with the Scottish Environment Protec-
tion Agency (SEPA) and the Northern Ireland Environment and Heritage Service (NIEHS) (2002a).
Technical Guidance. Note IPPC H4. “Integrated Pollution Prevention and Control – IPPC DRAFT.
Horizontal Guidance for Odour.” Part 1 – Assessment and Control. www.environment-agency.gov.
uk/commondata/105385/ h4_odour_part1.pdf.
Environment Agency for England and Wales in collaboration with the Scottish Environment Protec-
tion Agency (SEPA) and the Northern Ireland Environment and Heritage Service (NIEHS) (2002b)
“Technical Guidance. Note IPPC H4. Integrated Pollution Prevention and Control (IPPC) DRAFT.
Horizontal Guidance for Odour Part 2 – Assessment and Control.” www.environment-agency.gov.
uk/ commondata/ 105385/ h4_odour_part2.pdf.
Environment Protection Authority, Victoria, Australia (2004) “Strategies For Promoting Cleaner
Production In The Food Industry.” Rowan Mackenzie and Scott Hamilton. http://www.uneptie.
org/pc/cp/library/training/howtoCPC/manual_cdrom/CPlinks/pdfs/Hamilton.pdf.
Environmental Protection Agency (2000) “Biosolids and Residuals, Management Fact Sheet, Odor
Control in Biosolids Management,” EPA 832-F-00–067, United States, 2000. http://www.epa.
gov/owmitnet/mtb/odor_control-biosolids.pdf.
Ericsons Naturnomics Company (2004) “Odor Answers. Stack Extenders.” http://www.eric-sons.
com/ne4/odoranswers/stackextenders/index.htm.
Erwin, T.H., Vinka, Rabago, K.R., Glassner, D.A., Gruber, P.A. (2003) “Applications of life cycle
assessment to NatureWorks polylactide (PLA) production. Polymer Degradation and Stability.”
80: 403–419. www.cargilldow.com/corporate/life_cycle/docs/PLA_Article.pdf. 2003.
European Commission Directorate-General Jrc. Joint Research Centre. Institute for Prospec-
tive Technological Studies Technologies for Sustainable Development. European IPPC
Bureau. (2003) “Integrated Pollution Prevention and Control, BREF (Best available tech-
niques REFerence document) on the Slaughterhouses and Animal by-products according to
Article 16(2) of Council Directive 96/61/ EC. 2003.” http://www.umweltdaten.de/nfp-bat-e/
slaughterhousesbref-e.pdf.
Gesamtverband der Deutschen Versicherungswirtschaft e.V. (GDV, German Insurance Association)
(2003) Wild, Y. “Refrigerated containers,” in Chapter 7. Container Handbook. 2003. http://www.
containerhandbuch.de/
Heber, A.J. (1997) “Setbacks for sufficient swine odor dispersion and dilution.” Paper presented at
the Livestock and Environment Symposium, University of Nebraska Cooperative Extension
Service, New World Inn, Columbus, Nebraska, December 10–11, 1997. Modified Dec. 16, 1997.
http://pasture.ecn.purdue.edu/~heber/setba.htm.
Holland, M., Hunt, A., Hurley, F., Watkiss, P. (2004) Draft final methodology paper for “Service
contract for carrying out cost-benefit analysis of air quality related issues, in particular in the
clean air for Europe (CAFE) programme.” 5/ 07/ 04 AEAT/ ED51014/ Methodology Paper
Issue 3. http://europa.eu.int/comm/environment/air/pdf/cba_methodology_issue3.pdf.
Institute for Prospective Technological Studies Technologies for Sustainable Development.
European IPPC Bureau. European Commission Directorate-General Jrc. Joint Research
Centre. 2003. “Integrated Pollution Prevention and Control, BREF (Best available techniques
REFerence document) on the Slaughterhouses and Animal by-products according to Artical
16(2) of Council Directive 96/61/EC.”
International Cleaner Production Information Clearinghouse (ICPIC) (2004) Environmental Man-
agement Center. Technical Case Studies. Manufacturing. “Manufacture of Food Products and
Beverages.” http://www.emcentre.com/unepweb/tec_case/ .
Lifset, R. (1999) United Nations University. Zero Emissions Forum. The Industrial Ecology of
Renewable Resources. ZEF-EN-1999–2. http://www.unu.edu/zef/publications-e/ZEF-EN-1999-
02-D.pdf
Mohunta, D. M. (2004) “Recovery of volatile organic compounds from gaseous streams.” http://
www.ccdcindia.com/index2.php?act=voc&subact=vocpaper. Commercial, Chemical and Devel-
opment Company.
New South Wales Department of Environment and Conservation (NSW) (2004) “Odour control.”
http://www.epa.nsw.gov.au/mao/odourcontrol.htm
104 R. Nabais
Ontario Ministry of Agriculture, Food and Rural Affairs and Ontario Ministry of Environment and
Energy (1995) “The Guide to Agriculture Land Use and Minimum Distance Separation.” http://
www.gov.on.ca/OMAFRA/english/landuse/facts/guidealu.pdf.
Pennsylvania State University in cooperation with the Pennsylvania Department of Agriculture
(2002) Odor Management in Agriculture and Food Processing. Odor Management in Pennsyl-
vania. A Reference Manual. Revised printing June, 2004. http://server.age.psu.edu/fac/Elliott/
EntireOdorManual.pdf.
Pinchin Environmental Newsletter (2004) Ventilation & Emission Control Group. http://www.
pinchin.com/newsletters/air/odourcomplaints.htm.
Portary (2004) Recolha e Transporte de Resíduos Industriais Banais Perigosos. http://www.portaryresiduos.
pa-net.pt/html/gest.htm.
Realff, M.J., Abbas, C. (2003) “Industrial symbiosis. refining the biorefinery.” Journal of Industrial
Ecology. Vol. 7, Issues 3–4.
Robèrt, K.H. (2002) Schmidt-Bleek, B., Aloisi de Larderel, J., Basile, G., Jansen JL, Kuehr R, Price
Thomas P, Suzuki M, Hawken P, Wackernagel M. (2002) “Strategic sustainable development —
selection, design and synergies of applied tools.” J Cleaner Prod.; 10(3): 197–214.
Schauberger, G., Piringer, M. (1997) Guideline to assess the protection distance to avoid annoyance by
odour sensation caused by livestock husbandry. Proceedings of the Fifth International Livestock
Environment Symposium, May 29–31. 1997.
Society of Heating, Refrigerating and Air-Conditioning Engineers (1997) “Handbook of Funda-
mentals.” http://www.containerhandbuch.de/chb_e/ wild/index.html.
Starkey, R. (1998) Centre for Corporate European Environment Agency. (CCEM). Internet Gateway
for Cleaner Production, Pollution Prevention and Sustainable Business. “Environmental man-
agement, Tools for SMEs: A handbook.” Environmental Issues Series. Edited for the EEA.1998.
http://reports.eea.eu.int/GH-14-98-065-EN-C/en/enviissu10.pdf.
Sustainable Agri-Food Production and Consumption Forum (2004) UNEP Publications. United
Nations Environment Programme Division of Technology, Industry and Economics. “Industry
Sector Guide Lines.” Ref. UNEPTIEb, 2004.
Tybor, P.T., Hurst, W.C., Reyknolds, A.E., Schuyler, G.A. (1995) Quality Control. A Model Program
for the Food Industry. University of Georgia College of Agricultural & Environmental Sciences.
Cooperative Extension Service, Extension Food Science. http://pubs.caes.uga.edu/caespubs/
pubcd/b997-w.html.
US Food and Drug Administration (2004) Federal Anti-Tampering Act U.S.C. Title 18 – “Crimes
And Criminal Procedure. Part I – Crimes Chapter 65 - Malicious Mischief. § 1365. Tampering
with consumer products.” http://www.fda.gov/opacom/laws/fedatact.htm.
US Environmental Protection Agency (1995) Technology Transfer Network. Clearinghouse for
Inventories & Emission Factors Compilation of Air Pollutant Emission Factors, “AP-42, Fifth
Edition, Volume I: Stationary Point and Area Sources.” Research Triangle Park, NC. 1995.
UNEP Working Group for Cleaner Production in the Food Industry (2004a) “A Guide to Cleaner
Production in the Food Industry.” http://www.geosp.uq.edu.au/emc/cp/Res/guides/GUIDE1.
HTM to http://www.geosp.uq.edu.au/emc/cp/Res/guides/GUIDE7.HTM.
UNEP Working Group for Cleaner Production in the Food Industry (2004b) “Fact sheets.” http://
www.geosp.uq.edu.au/emc/CP/Res/xfactsheets.htm. The University of Queensland and The
United Nations Environment Programme.
United Kingdom Renderers Association (UKRA) (2004) http://www.ukra.co.uk/index.htm.
United Nations Environment Programme — The UNEP Working Group for Cleaner Production in
the Food Industry (2004) “A Guide to Cleaner Production in the Food Industry.” http://www.
geosp.uq.edu.au/emc/cp/res/yfood_manual.htm. (access on Wednesday, August 25, 2004).
Van Harreveld, A.P. (2001) “From odorant formation to odour nuisance: new definitions for discuss-
ing a complex process.” Water Science & Technology. Vol. 44. No 9. pp 9–15 © IWA.
Williams, M.L., Thompson, N. (1985) “The effects of weather on odour dispersion from livestock
buildings and from fields.” In: Odor Prevention and Control or Organic Sludge and Livestock
Farming. V.C. Nielsen, J.H. Voorburg, P. L’Hermite. Elsevier Applied Science Publishers, New York.
pp. 227–233.
7
Odors Treatment:
Physicochemical Technologies
Regina Nabais
7.1. INTRODUCTION
Taking into consideration simple economic factors and/or social responsi-

bility, nowadays, every industry tends to analyze its productive organization
through total quality management, which includes environmental and safety
issues such as avoiding annoyances or nuisances by odor emissions; in this,
the food industry is no exception. However, even though many efforts have
been developed to prevent and control odorous emissions at the production
unit level, these measures usually are not enough to assure good quality solu-
tions for all kinds of odor problems. When this happens, effective treatment or
abatement measures must be considered. This chapter’s scope is to describe the
main technologies available to reduce and control volatile organic compounds
(VOC) and malodorous emissions from industrial processes, especially those of
interest for the food industry [Environmental Protection Agency (EPA), 1995].
The main concepts, principles, fundamentals, advantages, and disadvantages
of each technology will be generally referred and discussed.Indeed, controlling
odors is an important consideration for protecting the environment and the
community’s good will, but it is worth remembering that operational improve-
ment usually requires strong competencies, capabilities, and experienced tech-
nological assistance.
When deciding technological investments, technology isn’t the solution;
it only enables people to find good answers for solving problems. Among all
technologies, simplicity and adequacy must always be our priority; in accord
with the philosophic and scientific principle of Occam’s razor—“Pluralitas
non est ponenda sine necessitas”—everything should be made as simple as
possible, but not simpler.
REGINA NABAIS CERNAS, Centro de Recursos Naturais, Ambiente e Sociedade, Escola

Superior Agraria de Coimbra, Instituto Politecnico de Coimbra, Bencanta, 3040-316 Coimbra,
Portugal
105
106 R. Nabais
7.2. BASIC PRINCIPLES UNDERLYING ODOROUS EMISSIONS

TREATMENT AND/ OR ABATEMENT
Despite the complexity of the problems of handling off-odors, there are only two
ways of management of their emissions: (1) source control, by selecting appropri-
ate industrial processes able to reduce emissions in the source, and (2) treatment
of the final resultant emissions.
When applied to industrial processing, the goal of cleaner production, is
to avoid pollution generation in the first place,which frequently reduces risks,
cuts costs, and identifies new business opportunities. Cleaner production aims
to reduce waste and inefficiency at the source and also can help to develop the
most efficient ways to operate processes, produce better products, and provide
new services.
Adequate control equipment and effective treatments may be based on cap-
ture recovery (nondestructive technologies) or on destruction. Some treatment
technologies use combinations of both capture recovery and destruction. Crite-
ria for the selection of the best choice for odor control and treatment must be
carefully chosen to assure their effectiveness at sustainable costs.
Even so, whatever may be the selected treatment, it will require high initial
cost to install, efforts to implement, and careful planning to start up and main-
tain, and together with the important capital investments will require careful
criteria that considers heat conservation/recovery. Large airflow rates with low
volatile pollutant concentrations frequently are encountered in food industry
processes, considering VOC and odors emissions.
Regardless of the selected option to deal with off-odors, rigorous, careful,
and strict maintenance programs will need to be designed and implemented,
especially if nonthermal emission control devices are chosen. Usually, all tech-
niques need replacement of costly consumables (i.e., adsorbents, reactants, fuel
etc.) and require energetic provisions to regenerate subsidiary materials (e.g.,
adsorbents) and storage of spare parts.
Consideration also must be given to the legal requirement of using either
the “best available control technology” (the best possible choice of competing
processes for reduction of a specific pollutant given economic, chemical, and
physical constraints), or the “maximum available control technology” (the process
that produces the greatest emissions reduction, regardless of all constraints), or
“reasonable available technology” ( a balanced process that considers equal con-
cerns to abatement technology that is more reasonable, economical, and proven
in practice), or even the “best available technology not entailing excessive costs”
(best at preventing pollution and available in the sense that it is procurable by
the operator of the activity concerned, including a balance between environ-
mental benefit and financial costs) [Integrated Pollution Prevention and Control
(IPPC), 2003].
However, before one starts to identify and choose any suitable technology for
air pollution control, including odor abatement, there are variables and procedures
Odors Treatment: Physicochemical Technologies 107
to consider: nature, concentration, physical, and chemical properties of the pol-

lutants to remove, considering corrosiveness, abrasiveness, toxicity and flamma-
bility, temperature, relative humidity, pressure, flow rate, and particles content of
the emissions; overall appraisal of the process efficiency, regarding compliance
with regulations, legislation, recommendations, and guidelines; feasibility and
economic interest of recycling or recovery of the pollutants compared with final
disposal costs; power and subsidiary requirements; foreseen initial and operational
costs; and finally, planning convenient training programs for the personnel who
will operate the system.
This chapter presents the most important VOC abatement technologies
adaptable to odors treatment, especially in the food industry. All available treat-
ment technologies are effective to removing odorous compounds when cor-
rectly applied, but each has specific performances and constraints inherent to
the process. In many situations, employing different technologies in a sequential
configuration provides synergetic effects, while enhancing the best of each tech-
nology. Operational plants confirm that for waste airstreams of food industrial
process affected by unpleasant odorant pollutants, such as amines, mercaptans,
aldehydes, etc., to be treated at an acceptable cost, enough to obtain sufficiently
low pollutant levels of outlet gas, require sequences and selected combinations
of more than one of the aforementioned technologies. Examples of common
technologies used for industrial proposes including food processes are listed in
Table 7.1 [New South Wales Department of Environment and Conservation
(NSW), 2004].Because of their low operating costs, if properly operated and
maintained biological technologies are usually popular among food industries.
These technologies are comprehensively described and discussed in Chapter 8,
this volume.
7.3. MAIN TECHNOLOGIC TREATMENT SYSTEMS FOR

GASEOUS POLLUTANTS
Whenever odor emissions cannot be prevented, containment and often some

form of treatment are followed by release to the atmosphere. Reliance is placed
on sufficient dispersion taking place before sensitive receptors are reached.
Odor control methods may be generally classified as: nondestructive or
destructive.
Nondestructive methods can be very effective in reducing gaseous emissions
from process stream vents, but they do not really destroy the VOC or odorant
pollutants. Non-destructive methodologies are advantageous if the recovery
substances are valuable to the industry itself and the recovery process is located
near the points of generation, e.g., recovery of ethanol in food-grade production
factories, at the headspace of fermentation closed devices (fermenters). How-
ever, if not carefully conceived, in most cases efforts and resources will be wasted
as odor-causing pollutants only trade places, from one medium to another.
108 R. Nabais
Table 1. Examples of Odor Control Technological Options for Food Process Industries
Rendering industry
Grease trap waste

Chicken feathers
Slaughterhouses
Fermentation
incineration
Fish meal
Animal
Coffee
Boiler
Wet scrubbing. Absorption: moderate to high costs,
three stages often required. Needs careful selection of
Ammonia
Amines,
aerosols
scrubber liquor and usually trials. Not always success-
Amines
Amines
ful, requires regular maintenance and daily tests of
active agent and pH control in some cases.
Thermal oxidizer (direct). Temperatures between
600°C and 1000°C with residence time of 0.3 to 1
second. Doubling of residence time may enable after-
burner temperature to be reduced 20°C to 100°C.
Requires careful design to reduce air volume to a
minimum. Capital cost high and cost of running is
high for large air volumes. Needs further control if
Aldehydes, amines
Aldehydes, amines
sulfur or chlorine present in exhaust gases.
Amines, aerosols
Thermal Oxidizer (catalytic). Temperature 400°C.
Lower temperature operation than direct method,
Amines
but the catalyst may be destroyed if not operated and

maintained correctly. This is frequently a problem.
Aldehydes,
Carbon adsorption. Needs regenerating at regular

Amines,
aerosols
Alcohol
intervals. Can be effective but expensive for large air
Amines
Amines
Amines
amines
volumes. Small operations are reasonably inexpensive

to deal with.
Aldehydes,
Best management practices. Cleanliness and avoiding

spills requires human effort but is relatively inexpen-
amines
sive.
Condensers. Condensable liquids removed from the

Aldehydes,
exhaust gas stream, reduces the amount of odorous

Amines
amines
gas to be treated.
Destructive methods are those, when properly operated, that are not producers
of secondary organic waste—they destroy the pollutants or contaminants by
oxidation of pollutants or malodorous compounds.
Table 7.2 shows the technologies commonly used to control VOC emissions,
which may be easily adapted to food industry odor emissions reduction and control.
Table 2. Odor and VOC Control Technologies
Destructive technologies Nondestructive technologies
Biofiltration; Absorption; adsorption;

thermal oxidation (catalytic, regenerative, recuperative) condensation
7.3.1. Absorption
In general, absorption is the transference of one substance into or through

another of a different state. When applied to air pollution this technology means
that gaseous components are transferred from a gas stream to a liquid phase.
The liquid phase may be water or aqueous solutions of chemicals that react with
the absorbed components. According to the adequacy of their application and
maintenance, the efficiency of these systems may be in a range of 70 to over 99%
[United States Corps of Engineers (USACE), 2004].
Absorption commonly refers to the intimate mixing of a gaseous exhaust
stream with a liquid. The liquid must dissolve or absorb specific constituents
of the stream. Wet scrubbers are examples of absorbers, whereby the exhaust
stream is routed through a vessel and forced into contact with a circulating liquid
solution.
Low cost and space requirements make this technology attractive for odor
control under certain circumstances; secondary treatment or disposal of the
effluent, depending on the gas stream constituents, corrosion, and fouling, may
impact maintenance costs.
This technology is particularly efficient for the treatment of water-soluble
odorous compounds like ammonia, amines, ketones, acetates, etc.
The liquid phase can be either an aqueous solution (acidic or alkaline) or a
solvent solution, depending on the solubility of the compounds. Chemical oxi-
dation is carried out either in a liquid phase or in a gas phase. The most com-
monly used oxidizing agents are sodium hypochlorite (NaClO) and hydrogen
peroxide (H2O2) although other oxidants can be used (EPA, 1995). Both have
advantages and disadvantages: sodium hypochlorite is cheaper but the scrubber
effluent contains chlorine and chloride. Hydrogen peroxide has no spent product
but is more expensive.
There are several types of equipment designs for contact phases. Scrub-
bers include packed beds, spray towers, tray towers, absorption towers (see Fig-
ure 7.1), Venturi scrubber, and static mixers. Most of these configurations also
require the installatation of mist eliminators before the treated air is released
into the atmosphere.
Tables 7.3 and 7.4 list general aspects of air pollutant treatment system,
based on absorption technology and the major advantages and disadvantages of
this technology, respectively.
Contaminants transferred to the liquid phase are removed from the scrub-
ber by continuous or periodic overflow of blowdown.
All chemical scrubbers are designed for a large contact area between the
gas and liquid phases. Odor control, however, requires some dedicated design
features. In particular, the air residence time in the scrubber must be sufficient,
typically around 3 seconds.
The scrubber and related ductwork must be cleaned to remove gelatinous
substances and milk stone deposits, especially if this option is considered by
dairy industries.
110 R. Nabais
Table 3. General Characterization of Absorption
Controls of the systems and

Performance factors Malfunction conditions usual checklists
Temperature Inadequate liquid flow Temperature gradient between

pH Inadequate pH control inlet and outlet of the gaseous
Gas-liquid contact Inadequate chemical selection or stream
Adequate liquid flow rate addition Liquid flow rates and pressures
Adequate chemical Excessive temperature of Gaseous effluent opacity
solutes additions the liquid
Plugged beds or mist eliminators
Corrosion
From Crowder (2004).
Table 4. Major Advantages and Disadvantages, of the Absorption Technology
Process Advantages Disadvantages

Asorption Highly efficient removal of Requires the use of large quantities of chemicals;
scrubbers odorous compounds Chemicals most commonly used (sodium hydroxide
(up to 99.0% depending and sodium hypochlorite) are expensive and
upon system design and hazardous to handle;
chemicals employed.); Treatment of multiple compounds generally
Small footprint compared requires multiple stages and different chemicals;
to most equivalent odor For H2S removal, improper handling of the
control technologies (face recycled liquid stream can result in liberation of H2S;
velocities can approach High pH operation causes scaling (fouling formation/
120 meters per minute); precipitation of solids) of the packing in packed
Ability to handle very high tower systems;
H2S loadings (upwards Scrubbers are complex systems requiring frequent
to 200 ppm); maintenance. Many sensors and moving parts that
It even treats fumes on line; need replacement or calibration;
Increased relative velocity High pressure drops required;
between scrubbing the fluid Internal plugging, corrosion, erosion;
and gas stream, increases Increased need for internal inspection;
efficiency for solids. Gas and liquid chemistry control important;
Particulate matter is not recycled;
High corrosion potential;
Liquid waste stream;
High energy consumption.
Wet scrubbing using packed towers has proved to be the most reliable means
of removing odorous pollutants and odor reduction, handling, for example, a
wide variety of sulfides in inlet concentration. Alkaline (NaOH) scrubbing can
achieve highly efficient removal of hydrogen sulfide (H2S) but the solution pH
must be maintained at a high level to prevent release of the H2S. In most cases
an oxidant such as NaOCl is added to prevent this from occurring. Highly vol-
atile sulfides such as dimethyl sulfide (DMS) and dimethyl disulfide (DMDS)
Mist eliminator
Outlet gas
Scrubbing
liquid
distribution
Packing material
Make up
water or
solution
Inlet gas
Figure 1. Sketch of an absorption tower; wet scrubber system.
typically require two stages of scrubbing to emsure removal below detectable

odor thresholds.
Design and operating parameters include scrubber geometric shape, liquid
spray or injection locations, gas residence time, gas velocities, gas and liquid tem-
peratures, gas and liquid pressure drop, and liquid–gas flow rate ratio.
Air flow rate and or residence time are important parameters to get an initial
idea of the required (1) footprint of the equipment, (2) admissible inlet pollut-
ant concentrations, and (3) overall removal efficiency. The order of magnitudes
that we may expect are: for air velocity through a chemical wet scrubber washer,
2–3 m/s; Venturi scrubbers operate at 45 to 120 m/s. By comparison, traditional
activated carbon systems operate between 0.25 a 0.5 m/s, and a typical biofilter
operates much closer to 0.025 m/s.
For rendering industries, multistage wet scrubbers using various scrubbing
agents are the primary alternative to incinerators for high-intensity odor control.
Wet scrubbers also generally are used to control whey dryer emissions in cheese
manufacturing.
7.3.2. Adsorption
Adsorption is a surface phenomenon where volatile compounds are selectively

adsorbed on the surface of solids with large surface-to-volume ratios. There are
112 R. Nabais
two major classes of adsorption processes: chemical adsorption (chemisorption)

and physical adsorption (physisorption).
Physisorption is an adsorption in which the forces involved are intermo-
lecular forces (weak forces); i.e., no significant change in the electronic orbital
patterns of the species involved. In the physisorption, the adsorbed species is
chemically identical with those in the fluid phase, so that the chemical nature
of the compound is not altered by adsorption and subsequent desorption.
Adsorption enthalpy in the physisorption is related to factors such as molecu-
lar mass and polarity, typically 5–40 kJ mol-1, while in the chemisorption it
is related to the chemical bond strength, and typically 40–800 kJ mol-1. Usu-
ally, the kinetics of physisorption is also faster; however, there is no definitive
distinction between the two systems, because intermediate cases are common.
Because of adsorbent regeneration constraints, chemical adsorption is not
very common in industrial air pollution control systems. Furthermore, when
desorption is considered, after adsorption processes with higher activation
energy, it requires significant activation energy to remove the chemisorbed spe-
cies from the adsorbent surface. This is a costly operation because it must be
performed under very high temperature or high vacuum conditions or by some
suitable chemical treatment of the surface; therefore, comparatively chemi-
sorption is a more expensive operation.
One may choose between many adsorbent substances, such as activated
alumina, activated carbon (the most common), synthetic resins, molecular
sieves, peat, polymers, silica gel, zeolites, or simply soil. Larger molecules
adsorb better than smaller molecules. Nonpolar molecules adsorb better than
polar ones. Nonsoluble or slightly soluble molecules adsorb better than highly
soluble molecules.
Based on the polarity or solubility of the molecule being adsorbed,
pH may have an influence on the extent of adsorption. Alcohols are poorly
adsorbed, because they are very soluble and highly polar. Aldehydes are
highly polar molecules, but as their molecular weight increases, the polar-
ity decreases, and adsorbability related to alcohols is higher. Adsorption of
amines is limited by polarity and solubility. Higher molecular weight organic
compounds will generally be more adsorbable because adsorptive attraction
increases with the size.
Temperature increases the rate of diffusion through the stream phase to the
adsorption sites, but since the adsorption process is exothermic, increases in tem-
perature may reduce the degree of adsorption. This temperature effect is negli-
gible in water treatment applications and ambient vapor phase applications.
Additionally, if odor is the only environmental problem of the airstream,
the malodorous constituents represent only a fraction of the total organic com-
pounds that would be adsorbed by the activated media. Because of the large
loading of other organic substances, media such as activated carbon would have
to be regenerated in situ or frequently replaced to be regenerated elsewhere.
Either way, sorbent regeneration is still a difficult and dirty task.
In regenerative systems, the VOC is removed from the air or process

stream and then is recovered by directly heating the adsorbent and using
this energy to overcome the intermolecular attractive forces. The resulting
vapor stream is condensed, allowing for recovery and reuse of the solvent
(Calgon Carbon Corporation, 2002). Continuous and automatic processing
of solvent-laden air is achieved using a minimum of two adsorber vessels,
one of which is steam stripped while the other is adsorbing solvent onto the
carbon bed.
Low-pressure steam may be used to strip (remove) adsorbed components
from the carbon.
Tables 7.5 and 7.6 list general aspects of air pollutant treatment system,
based on adsorption technology and the major advantages and disadvantages
of this technology:
Other considerations in carbon adsorption applicability include moisture
content and temperature of the emission stream. As water vapor competes
with pollutants for adsorption sites on the carbon bed, relative humidity levels
must be kept under 50% for usual applications. Adsorption is a highly efficient
technology to remove odorous compounds (up to 99.0% depending on sys-
tem design and chemicals employed) and has the ability to handle very high
H2S loadings (upward to 200 ppm). On the other hand, it sustains face veloci-
ties approaching 2m/s and has a very small footprint compared to the closest
equivalent odor control technologies. Adsorption on activated carbon is useful
for recovery of VOC with intermediate molecular weights. Smaller compounds
do not adsorb well and larger compounds cannot be removed during regen-
eration, which typically is by steam stripping. The pollutants will remain in
the bed after each adsorption cycle, ultimately requiring a replacement of the
carbon bed.
Activated carbon systems have interesting applications in removing VOC
from relatively dry airstreams (Figure 7.2).
Table 5. General Characterization of Adsorption Technology

Temperature Infrequent adsorption; Inlet gas temperature and

Pressure Physical deterioration of flow rates;
Organic concentration adsorbent; Pressure drop across the
Contaminant molecular weight Plugging of the adsorbent; adsorbent;
Moisture High temperatures; Regeneration frequency
Presence of particulate matter High organic vapor concen- and cycle period
in the system tration
From Air Pollution Training Institute (2004).

114 R. Nabais
Table 6. Major Advantages and Disadvantages of the Adsorption Technology
Adsorption Activated carbon is applied to off-

gasses containing low VOC
concentrations (less than 50 ppm).
Is an interesting solution for small gas
fluxes.
Adsorption1 Simple, passive system design, Limited capacity for H2S that translates
Unimpreg- with few moving parts and few into frequent change-outs and poor
nated acti- maintenance concerns. economics;
vated Effective removal capacity of Media change-out is time consuming,
systems organics and low levels of H2S. labor-intensive, and dirty.
(carbons, Safe operation with no hazardous Spent media must be landfilled or sent
activated materials. off-site for thermal reactivation.
alumina, Provides an effective second stage Careless system operation can produce a
silica gel) treatment to other technologies. fire hazard.
Adsorption 2 Highly efficient removal of Reduced capacity for organics removal
Caustically odorous compounds (up to 99.0% due to presence of impregnated substance;
impregnated depending on system design and Caustic regeneration is costly and
activated chemicals employed); hazardous, and creates spent caustic
carbons Small footprint compared to disposal problems;
most equivalent odor control Spent carbon cannot be thermally
technologies (face velocities can reactivated and must be landfilled;
approach 2m/s); Media change-out is time consuming,
Ability to handle very high H2S labor-intensive, and a very dirty
loadings (upward to 200 ppm); operation;
High removal capacity for H2S; Careless system operation can produce a
Simple maintenance procedures. fire hazard.
From Van Deuren et al. (2002); US Army Corps of Engineers (2004).
Clean
gas
Polluted
outlet
gas
Blower
Adsorbent
material
Figure 2. Sketch of an activated carbon adsorption system. Adapted from Croll-Reynolds Clean Air
Technologies (1996).
7.3.3. Condensation
Condensation is a separation process in which gaseous components of a gaseous

stream are removed by cooling down the gas stream, causing a changing state
phase, converting vapor components to liquid, and then removing the liquid
phase. Air-to-air or air-to-liquid heat exchangers are commonly used.
The most common equipments are shell and tube heat exchangers (see
Figure 7.3). The contaminant removal efficiency of the systems is about 90%,
and is higher for higher vapor pressure of the contaminants.
There are three types of condensers:
1. Conventional (–18 ºC to 4º C),using brine coolants. Conventional con-

densers are of two kinds: direct (mixing the condensed gaseous compo-
nents with the coolant, increasing wastewater treatments or components
recovery procedures) or surface condensers.
2. Refrigeration (–100 ºC) ,using common compressed industrial coolants.
3. Cryogenic (–195.5 ºC), using liquefied gases such as nitrogen and carbon
dioxide.
Condensation is typically combined with particulate removal equipment (i.e.,

mist eliminators, electrostatic precipitators, etc.) to separate the condensed vapor
from the gas stream after which the effluent is reprocessed, reused, or disposed.
Odor removal efficiency of condensation systems should be cautiously con-
sidered since odors can be associated with gaseous phase constituents that are
not readily condensed.
Depending upon the nature of the exhaust gas stream, fouling of the heat
exchangers and maintenance may become problematic. Finally, since condensate
is created or is a valuable product, its disposal or recycling costs must be con-
sidered. Tables 7.7 and 7.8 list general aspects of air pollutant treatment system,
based on condensation technology and the major advantages and disadvantages
of this technology.
Cooling Inlet air

water
Hot water
outlet
Condensate
Figure 3. Shell and tube condenser.

116 R. Nabais
Table 7. General Characterization of Condensation Technology

Gaseous compo- High outlet temperatures Outlet gas stream temperature

nents concentration Low coolant flow rates Gaseous stream and coolant
Coolant inlet and Inadequate heat transfer conditions flow rates
outlet temperatures Buildup of ice-frozen organic compounds Coolant flow rate
From Air Pollution Control Technology, (2004).
Experiments are described in which the carbon content of vapors given

off during wort boiling at the Bavarian State Brewery, Weihenstephan, are con-
tinuously collected and condensed by means of a plate heat exchanger and the
remainder is piped into the furnace of the brewery boiler (Muller, 1990; Muller
and Meyer-Pittroff, 1990).
Condensation of components of a gaseous stream may be achieved by using
indirect cooling of the gas to condense water vapor onto the collecting surfaces.
Because condensed water is the medium that carries away the entrained
particles there are available in the market entirely self-cleaning devices with low
pressure drop.
At least, with this technology, if condensed pollutants are removed from the
exhaust gas stream, the amount of odorous gas to be treated downstream will
be reduced. In biological industries (for example, rendering works) it is usual to
place the biofilter after the condenser (NSW, 2004).
Condensers are convenient equipment for simultaneously removing con-
taminants from the gas phase and quenching steam and often are the first stage
in an air-cleaning system.
Table 8. Major Advantages and Disadvantages of the Condensation Technology
Condensation Ability to cost-effectively The coolant fluid depends on the vapor pressures
recover and reuse con- of the organic contaminants to be isolated;
densed components from Thus, the identity of the VOC pollutants dictates
this process increases the the operating conditions for the control process;
technique’s value; this may be difficult to achieve in gaseous streams
Condensation method composed of complex odors mixtures;
will work with any The relatively low concentration of solvents in air
volatile compound, in requires very low temperatures resulting in high
any concentration, from energy costs (electrical power), and hence, poor
streams with different economics;
flow rates. Recovery by cryogenic condensation will condense
and freeze water vapor from the airstream requir-
ing complex exchanger thawing techniques.
From Van Deuren et al. (2002).

7.3.4. Oxidation
There are a set of technologies based in physicochemical oxidation processes, and

despite intrinsic differences of operational systems, they are generally referred to
as “incineration.”
7.3.4.1. Incineration Generalities
In vapor incineration processes, organic components of gaseous stream are oxi-

dized and form carbon dioxide and water vapor. If organic components of gas
stream contain chlorine, fluorine, or sulfur, then hydrochloric acid, hydrofluoric
acid, sulfur dioxide, and other compounds may be formed. Like other combus-
tion control devices, thermal incinerators operate on the principle that any gas-
eous organic compound heated up to a high enough temperature in the presence
of sufficient oxygen will be oxidized to carbon dioxide and water. The theoretical
combustion temperature for thermal oxidation depends on the properties of the
VOC to be combusted. General aspects of air pollutant treatment systems based
on incineration (thermal oxidation) are listed in Table 7.9.
Odor abatement by incineration in combustion plants is well suited in plants
with large consumption of energy (fish meal, meat meal, bone meal factories, etc.);
in this case, heating the boilers or cooking devices can enhance the exhaust gas heat,
whereas for plants with low energy consumption the operation costs of incineration
are often unbearable. In this case the best and cheapest solution is often a combina-
tion of ad/absorption.
Table 9. General Characterization of Incineration (Thermal Oxidation Technology)
Controls of the
systems and
Variations in chamber Low combustion temperature; Outlet gas

temperature, residence time, Inadequate residence time; stream VOC
inlet VOC concentration; Inadequate mixing of the gases; concentration;
Vent streams with a low heat Generation of additional pollutants in the Inlet and outlet
content need extra fuel for incinerator; gas stream
flame stability and their Loss of catalyst activity for catalytic processes; temperature
combustion; Operating temperatures must be higher than (thermal
Dimension parameters include 800 ºC, to avoid carbon monoxide formation; oxidizers);
variable waste flow rate, Thermal incinerators can achieve more than Stack effluent
calorific value, temperature 98% destruction efficiency at combustion opacity;
and oxygen content, and chamber temperatures ranging from 700 to Variations in
residence time; 1,300º C and residence times of 0.5 to 1.5 the gas flow
Appropriate turbulence design. seconds. rate.
From Van Deuren et al.( 2002).

118 R. Nabais
7.3.4.2. Thermal Oxidation
Thermal oxidation, by definition, is an exothermal reaction, by which a hydrocarbon is

converted in carbon dioxide and water vapor, in the presence of oxygen, as follows:
Cn H2m + (n + m/2) O2 => n CO2 + m H2O + Heat
The number of oxygen atoms converted into n molecules of carbon dioxide and
m molecules of water vapor and heat, which is given off in the exothermic reactor.
In general, a thermal oxidizer is a specially engineered furnace or chamber
where materials are burned, combusted, and/or oxidized. The furnace or cham-
ber operates at relatively high temperatures with a sufficient retention (residence)
time and optimum turbulence to destroy combustible material.
With thermal incineration odorous chemicals are oxidized into less odor-
ous or non-odorous substances, for instance, carbon dioxide, water and sulfur
dioxide. The incineration temperature normally ranges between 500 and 1,200
ºC. Thermal oxidizers also have fast startups; therefore, they are good choices
for batch working or shift processes. The air required for the boiler plant can be
taken from the exhaust system. This method allows the factory to clean large
volumes of air at low costs.
Considering combustion heat recovery possibilities, thermal oxidation sys-
tems may operate by three modes: direct flame, recuperative, and regenerative.
Direct flame systems or flares operate by contact of the waste stream with a
flame to achieve oxidation of the VOCs. These systems are the simplest thermal
oxidizers and the least expensive to install, but require the greatest amount of
auxiliary fuel to maintain the oxidation temperature, thus entailing the highest
operating cost. Flares are very useful for destruction of intermittent streams;
this is a very common situation in the food industry. Direct fire is essentially
a combustion chamber with a burner and the appropriate control system. The
exhaust from a direct-fired unit is typically at the combustion temperature with
no primary or secondary heat recovery. This is used where heat recovery is not
required (e.g., when fuel for the burner is free or very cheap). Flares are used
for gas streams with organic vapor contents greater than two- or threefold their
lower explosive limits; if the gas stream does not have enough heat content, addi-
tional fuel is added to the gaseous stream.
Recuperative thermal oxidation systems (see Figure 7.4). In this configura-
tion, an oxidation chamber is coupled to a heat exchanger where combustion
heat of exhaust gas is transferred to inlet air. Thermal recovery efficiencies are
limited to 40–70% of lower explosive limit to prevent autoignition in the heat
exchanger. Supplemental fuel therefore is usually required to maintain a high
enough temperature for the desired destruction efficiency. Recuperative systems
are more expensive to install than flares, but have lower operating costs.
Regenerative thermal oxidizer (RTO) is a developing technique based on
the thermal oxidation, by operating with great fuel efficiency. An RTO consists
of two or more heat exchangers connected by a common combustion chamber
or zone (see Figure 7.5). The heat exchangers consist of beds filled with media,
Gas
Heat
outlet
exchanger
Gas
inlet
Burner
Oxidation Chamber
Figure 4. Recuperative thermal oxidizer. Adapted from Virtuous Cycle (2004).
which will allow air to pass while serving as a mass to store heat. The media
material selection, size, and shape can vary greatly and substantially impact the
design and utility efficiency of the RTO.
In comparison to both thermal and catalytic oxidation, RTOs have the advan-
tage of the VOC application flexibility and destruction of a thermal oxidizer with
better fuel efficiency than a catalytic oxidizer without the risk of poisoning or fouling
expensive catalyst.Regenerative thermal oxidation systems are the most expensive
thermal oxidizers to build, but savings in auxiliary fuel pays the added initial costs.
It is important to reduce the moisture content of any gas stream requir-
ing incineration (above 400ºC) in order to reduce fuel consumption, because
Supplement fuel
flame
Combustion Chamber
Stack
Bed packing -
ceramic heat
accumulator
media
Blower to
return
exhausted air
Blower of from the stack
Waste Air
Stream Treated Air to the stack
Figure 5. Sketch of a regenerative thermal oxidizer.

120 R. Nabais
Table 10. Relative Costs of Incinerating

Odorous Vapors Saturated with Water Vapor at
Various Temperatures
Saturation temperature Cost ratio
400C 1.00
500C 1.05
600C 1.15
700C 1.36
800C 1.81
850C 2.22
From NSW (2004).
incineration costs increase with the moisture content of the airstream to be

incinerated, as shown in Table 7.10.
Food and animal products industries are a major area where thermal oxidizers
are used to eliminate odors.
7.3.4.3. Catalytic Oxidation
“Catalysis” was a term used by Berzelius in 1835 to describe the ability of a chemi-
cal element to increase the rate of a chemical reaction between other compounds
without it being appreciably consumed by the reaction. Catalysts do this by lower-
ing the amount of chemical energy required for a given reaction to take place.
Catalytic oxidation converts VOC into carbon dioxide and water, as do
other oxidation processes, but it can be planned to avoid by-products requiring
treatments or final disposal. The advantage of this technology, in comparison
with noncatalytic oxidation, is a substantial reduction of process temperature
(315–455°C); hence catalytic systems represent important energy savings.
Typical catalytic oxidizer components include the catalyst housing, blower,
burner, heat exchanger, controls, and stack (see Figure 7.6). Catalytic oxidation
Treated air
Catalyst
Fuel
Air stream Pre heater

to treat
Blower
Figure 6. Catalytic thermal oxidation. Adapted from (2004).

is well suited to applications with VOC concentrations ranging up to 25% of the

lower explosion limit. With proper selection of catalyst, operating conditions,
and equipment design, catalytic oxidation can attain VOC conversions of up
to 99%. Advantages of this technology are low fuel usage, particularly with the
proper choice of heat exchanger, given low operating temperatures, and little for-
mation of partial oxidation products, such as carbon monoxide and aldehydes.
Disadvantages include susceptibility to catalyst poisons and the sensitivity of
the catalysts to high temperatures. Tables 7.11 and 7.12 list general aspects of
air pollutant treatment system, based on catalytic oxidation technology and the
major advantages and disadvantages of this technology.
Table 11. General Characterization of Catalytic Oxidation Technology

Catalytic oxidizers are used Low combustion temperature; Inlet and outlet gas tempera-
to treat gas streams that Inadequate residence time; tures;
have organic vapor concen- Poor mixture of gaseous Accumulation of particles,
trations under the lower compounds; followed by emissions opacity
explosive limit Fouling or plugging;
Loss of catalyst activity
Table 12. Major Advantages and Disadvantages of the Catalytic Oxidation Technology
Advantages Disadvantages
Lower fuel costs than thermal Higher capital costs than thermal oxidation;
oxidation; Operates at lower Less fuel is required compared with other thermal systems
temperatures than thermal because they operate at lower temperatures: from 315 ºC to 455ºC;
oxidation Catalytic oxidizers are not effective with high concentrations of
solids or liquid particles;
Frequent cleaning is desired, so the restoration of catalytic
activity can rise to up 90%;
Catalyst must be optimized for the contaminant present in the
off-gas stream (catalytic conversion is species-dependent);
Catalyst must be cleaned regularly (12- to 24-month intervals);
Several compounds influence catalysts performances (e.g.,
fluorine and sulfur);
Require safety precautions;
Catalysts operate typically for less than 3,000 ppm pollutants
concentrations;
Catalysts must be replaced periodically; Catalyst material may
be hazardous and may require disposal as contaminated waste;
Phosphorous, heavy metals (zinc, lead), sulfur compounds,
chlorine, bromine, iodine, fluorine, and any particulate can
result in shortening the life of the catalyst.
Adapted from Van Deuren et al.(2002).

122 R. Nabais
7.4. REFERENCES
Air Pollution Control Technology Verification Center (1999) APC. VOHAP. “Control technologies.”
January 28, 1999.
Air Pollution Training Institute (APTI) (2004) Air Pollution Control Technology Series Training
Tool. Control Technology Series. http://www.epa.gov/air/oaqps/eog/utrain.html
Calgon Carbon Corporation (2002)_ ES-OC-TP-005. “Options in odor control. New alternatives
and traditional technologies”, June 2002.
Croll-Reynolds Clean Air Technologies (1996) “Odor control in edible oil processing. Reprinting
of Pollution Engineering”, CRS-11-96.
Crowder, J. (2004) “Carbon adsorption control devices”, in: Air Control Techniques, P C. Air Pollu-
tion Control Technology APTI, 2004. Control Technology Series.
Davis, R. J., Zeiss, R.F. (1998) “Cryogenic condensation: A cost-effective technology for control-
ling VOC emissions”, Proceedings of the 1998 Annual Meeting of the American Filtration and
Separations Society, May 1998, St. Louis, Missouri.
Environmental Protection Agency (1979) Snyder, W.H. Technical Report. “Guideline for fluid mod-
elling of atmospheric diffusion”, EPA-450/4-79-016 (Draft) 1979.
Environmental Protection Agency. (1995) “Compilation of air pollutant emission factors”, AP-42,
Chapter 9: Food and Agricultural Industries. Fifth Edition, Volume I. United States. http://
www.epa.gov/ttn/chief/ap42/
Environmental Protection Agency (2000) Biosolids and residuals, Management fact sheet. “Odor con-
trol in biosolids management.” http://www.epa.gov/owmitnet/mtb/odor_control biosolids.pdf.
Environmental Protection Agency (1993)“Guideline on air quality models” (revised): Appendix
W of 40 CFR Part 51 (supplement b). (AH-FRL-7478-3). RIN 2060-AF01). Revision to the
Guideline on Air Quality Models. http://www.weblakes.com/lakeepa6.html#GAQM.
Environmental Protection Agency (1992)Technical Support Div. Cox, W.M. Technical Report. “Pro-
tocol for determining the best performing model”, PB-93-226082/XAB; EPA-454/R-92/025.
Integrated Pollution Prevention and Control (IPPC) (2003) European Commission Directorate-Gen-
eral Jrc. Joint Research Centre. Institute for Prospective Technological Studies Technologies for
Sustainable Development. European IPPC Bureau. “BREF (Best available techniques. REFerence
document) on the Slaughterhouses and Animal by-products according to Article 16(2) of Council
Directive 96/61/EC.” http://europa.eu.int/comm/environment/ippc/brefs/sa_bref_1103.pdf.
Julie Van Deuren, Teressa Lloyd, Shobha Chhetry, Raycharn Liou, James Peck. Federal Remedia-
tion Technologies Round Table. “Remediation Technologies Screening Matrix and Reference
Guide, Version 4.” http://www.frtr.gov/matrix2/section3/3_14.html.
Muller, K., Meyer-Pittroff, R. (1990) Master Breweries Association of the Americas. “Condensation
and thermal treatment of brew house vapors”, Technical Quarterly. 27(1), 1-4.
Muller, K. (1990) Master Breweries Association of the Americas. “Combustion of obnoxious sub-
stances contained in brewhouse vapors through the boiler firing system.” Technical Quarterly
27(2), 52-55.
New South Wales Department of Environment and Conservation (2004) “Odour control.” http://
www.epa.nsw.gov.au/mao/odourcontrol.htm.
New South Wales Environment Protection Authority (2001) “Approved Methods and Guidance, For
the Modelling and Assessment of Air Pollutants in New South Wales.” http://www.environment.
nsw.gov.au/resources/amgmaap.pdf.
Rafson, H.J. (ed) (1998) Odor and VOC Control Handbook. McGraw-Hill, New York.
US Department of Energy, Office of Environmental Management and Office of Science and Tech-
nology (2001)“Membrane system for the recovery of volatile organic compounds from remedia-
tion off-gases. Summary report.” Industry Programs and TRU and Mixed Waste Focus Area.
http://apps.em.doe.gov/ost/pubs/itsrs/itsr266.pdf.
US Environmental Protection Agency Research Triangle Park. Control Technology Center (1995) Infor-
mation Transfer and Program Integration Division. Office of Air Quality Planning and Standards.
R. A. Zerbonia. J. J. Spivey. S. K. Agarwal, A. S. Damle. C. W. Sanford., “Vapor gas streams, control
technology center, Survey of control technologies for low concentration organic vapor gas streams”,
EPA-456/R-95-003.1995. http://home.triad.rr.com/sanfordconsult/low_vo.pdf.
US Environmental Protection Agency (2002)Air Pollution Control Cost Manual, Sixth edition,
EPA/452/B-02-001. 2002. http://www.epa.gov/ttn/catc/dirl/c_allchs.pdf.
US Environmental Protection Agency (1994) Volatile Organic Compound Removal from Air
Streams by Membranes Separation. Membrane Technology and Research, Inc. United States
Environmental Protection Agency. EPA/540/F-94/503.
US EPA Model (2004) Regulatory Models - ISCST3 | ISCLT3 | ISC-PRIME | AERMOD | CALINE3
| CALINE4 | CAL3QHC | FDM | UAM - downloads. Environmental Protection Agency,“40 CFR
Part 51(AH-FRL-7478-3)”. RIN 2060-AF01. 04/17/03.http://www.weblakes.com/lakeepa3.html.
United States Environmental Protection Agency (1986) Guideline on Air Quality Models (Revised),
EPA-450/2-78-027R, July 1986.
US Army Center for Health Promotion and Preventive Medicine (2005) Introduction to Air Pollution
Control. http://chppm-www.apgea.army.mil/desp/pages/samp_doc/aircontrol_intro/22.2005).
US Army Corps of Engineers. Engineering and Design (2004)Adsorption Design Guide. DG 1110-
1-2. Chapter 3.http://www.usace.army.mil/usace-docs/design-guides/all.htm.
US Army Corps Of Engineers (2001) Engineering and Design–Air Stripping. CECW-E 1110-1-3. 31.
Oct 2001. http://www.usace.army.mil/usace-docs/design-guides/all.htm.
Van Deuren, J., Lloyd, T., Chhetry, S., Liou, R., Peck, J. (2002) Federal Remediation Technologies
Roundtable, “Remediation technologies screening matrix and reference guide, version 4”, http://
www.frtr.gov/matrix2/section3/3-14.html.
Virtuous Cycle (formerly BRM Business) (2004)Major Components of Recuperative Thermal
Oxidizer. http://www.virtuouscycle.co.uk.
8
Odors Treatment: Biological
Technologies
Bram Sercu, João Peixoto, Kristof Demeestere,
Toon van Elst, and Herman Van Langenhove
1. INTRODUCTION
Physical–chemical waste gas cleaning techniques have proven their efficiency

and reliability and will continue to occupy their niche, but several disadvantages
remain. Among them are high investment and operation costs and the possible
generation of secondary waste streams. With biological waste treatment tech-
niques, reactor engineering is often less complicated and consequently costs are
less. In addition, usually no secondary wastes are produced. Biological meth-
ods are nonhazardous and benign for the environment. Possible drawbacks are
restricted knowledge about the biodegradation processes, limited process con-
trol, and relatively slow reaction kinetics. Anyway, the biological methods for
the removal of odors and volatile organic compounds (VOCs) from waste gases
are cost-effective technologies, when low concentrations (below 1–10 g/m−3) are
to be dealt with (Kosteltz et al., 1996). Therefore, decision making can be based
merely on economical analysis. Like the treatment of liquid effluents, gaseous
streams will be more often considered for biological treatment. For organic com-
pounds, the biological reaction can be described as:
CHO + O2 + nutrients C5 H7 O2 N (cell dry weight) + CO2 + H2O + heat
When heteroatoms are present (e.g., chlorine, sulfur), end-products like HCl
or H2SO4 can be formed. For efficient pollutant removal, target pollutants have
to be sufficiently biodegradable and bioavailable. A major advantage in the case
of odor treatment is that biocatalysts have high affinity for the substrates, which
BRAM SERCU, JOÃO PEIXOTO, KRISTOF DEMEESTERE, TOON VAN ELST, AND
HERMAN VAN LANGENHOVE EnVOC Research Group, Department of Organic Chem-
istry, Faculty of Bioscience Engineering, Ghent University, 9000 Gent, Belgium. Department of
Biological Engineering, CEB, University of Minho, 4710–057 Braga, Portugal. Project Research
Gent nv, 9030 Gent, Belgium.
125
126 B. Sercu et al.
allows efficient treatment of low influent concentrations. Biocatalysts also oper-

ate at room temperature and they have innocuous final products (e.g., carbon
dioxide and water). Provided that you have the right inocula, microorganisms
can metabolize almost every compound there is. In general, odors consist of a
very complex mixture of volatile organic as well as inorganic compounds. The
most relevant compounds regarding odors in the food industry are nitrogen-
containing compounds (ammonia, amines, amides, and more complex molecules
like indole and scatole), reduced sulfur compounds (hydrogen sulfide and vola-
tile organic sulfur compounds) and VOCs like alcohols, aldehydes, volatile fatty
acids, and phenols. Most are easily biodegradable.
2. GENERAL OVERVIEW OF TECHNIQUES
Several biological waste gas treatment reactor concepts exist. They can be distin-
guished according to their filter material (organic or inorganic) and the type of
liquid phase (noncontinuous or continuous). Both characteristics influence mass
transfer and the presence or type of biofilm. In most cases, pollutants are first
transferred from the gas phase to the liquid phase and subsequently to the biofilm
(Figure 1). It has been argued, however, that pollutants can directly be transferred
from the gas phase to the biofilm when no water film is present, or that fungal
mycelia, protruding in the gas phase, can directly take up substrates without a dis-
solution step (Engesser and Plaggemeier, 2000). Optimal conditions for the organ-
isms (micro)environment should be provided. Temperature, pH, water activity,
nutrient availability, oxygen concentration, and osmotic potential are important
parameters. In practice, most of these parameters can be controlled in an accept-
able range by reactor choice, operation, and control. However, sometimes waste
gas characteristics like high or fluctuating temperatures and low oxygen concentra-
tions are more difficult to adjust, leading to limitations in the application area of
biological waste gas treatment techniques.
Laboratory studies provide much information about the removal of single
compounds or simple mixtures, but especially for VOCs, often relatively high
Gas phase Liquid layer Biofilm Support
Cg C1
Cthr
Figure 1. Pollutant transfer during biological waste gas cleaning (Waweru et al., 2000).
Odors Treatment: Biological Technologies 127
influent concentrations are investigated, conditions not prevailing in odorous

waste gases. Although the experimental conditions and the influent gas can be
controlled carefully in lab studies, it is sometimes difficult to extrapolate results
to real situations, especially when the composition of the waste gas is very com-
plex or unknown.
Other studies investigate the removal of odors in full-scale installa-
tions or use pilot-scale waste gas treatment units receiving waste air from
industrial plants. In this case, data with real complex odorous mixtures are
obtained. However, only in a few cases the waste gas characteristics in two
different plants are much alike. Therefore, before installing a biological waste
gas treatment reactor, often a series of pilot-scale experiments is conducted
on-site to investigate performance and design criteria. Devinny et al. (1999)
recommended the following steps in a protocol for biofilter design and imple-
mentation: (1) preliminary investigations of the waste stream (e.g., pollutant
concentrations, process flow rate, relative humidity, temperature), (2) litera-
ture research and modeling to determine if biofiltration is an appropriate
technique, and (3) possibly further bench- and pilot-scale experiments to
obtain a final reactor design.
Many classifications and denominations of bioreactors have been appear-
ing in the literature. Here we shall classify them in five main groups: biofilters,
biotrickling filters, biowashers, and two more recent techniques: the biological
plate tower and the membrane bioreactor. Many associations of different reac-
tors are also possible. Stability, adaptability, and low equipment and operational
costs are the requests they all must meet.
2.1. Biofilters
Biofiltration is often used with a broader meaning, referring to all the waste air
biological technologies. Strictly it means the most used technology about this
matter: the biofilters. They made their appearance in the 1950s in the deodor-
ization of air from wastewater treatment or composting plants. Nowadays they
are aimed not only at air polluted with organic gases, such as VOCs and many
other hydrocarbons, but also ammonia or H2S. They are more efficient with
low molecular weight gases, with high solubility in water and simpler molecular
bonds. Biofiltration has been extensively used, because waste gases in this case
generally contain low concentrations of well biodegradable organic and inor-
ganic compounds.
Basically, a biofilter is a layer of biologically active media (an organic filter
matrix), usually of natural origin. The filter particles are typically soil, compost,
peat, wood chips, tree bark, and heather. Granular activated carbon and plas-
tic material are also used. One kind or several combinations of particles have
been used. The media must provide a large surface area, nutrients and moisture
(around 50% of the media) for the microbial activity, and adsorption/absorption
of the odorous molecules. The microflora for the degradation of odors—mainly
128 B. Sercu et al.
bacteria and fungi—is part of the package. There is no continuous water phase.
For better results, the addition of nutrients containing nitrogen and phosphorus
must be considered, although this will add some cost to the process. The pres-
ence of bulking inerts usually calls for the addition of nutrients, mainly with
high load regimes (Devinny et al., 1999). Adequate porosity (around 0.50) is
essential for low pressure drop (power requirements).
To build a conventional open-bed filter (Figure 2), in the early ages of the
technique, a hole was excavated in the ground (around 1.0 m deep) and filled up
with a bed of the selected media. Nowadays, synthetic material or concrete is
used. Perforated piping or other systems are used for gas distribution under the
bed. The waste air flow, combined with the void fraction, causes the residence
time to be normally between 15 and 60 sec, the time it takes for the odors to be
absorbed and metabolized through the filter. Surface loading rates are about
1.2 m3 m−2 min−1 (Devinny et al., 1999). Impermeableness is desirable to avoid
liquid leaching.
For optimal long-term operation of biofilters, next to controlling the biofilter
moisture content, precautions should be made to prevent acidification if sulfur
or nitrogen-containing compounds are present. This can be accomplished by
buffering, e.g., by adding CaCO3 (Rafson, 1998), or regular replacement of the
filter material (every 1 to 5 years, depending on the loading rate). The latter treat-
ment is also needed to remove other accumulated intermediates or end-products,
to prevent high pressure drops, and to prevent nutrient limitation if nutrients
are not provided during biofilter operation. Indeed, removal efficiencies go from
60% to 100%, depending on the media and the pollutants contaminating the
air. Initial performance is very good but as time goes on problems may occur,
leading to severe degradation of efficiency. Clogging and channeling are likely
to appear.
Modular closed systems are commercially available. They minimize the
surface for installation because they are stacks of trays that can be set in
series or parallel arrangements, or combinations of both. The usual time of
operation, with good removal characteristics, for conventional systems (2 to
Clean air
Filter media
Polluted air Excess water
Figure 2. Schematic of a conventional open-bed biofilter.

4 years, according to Devinny et al., 1999), is extended due to selected media,

uniform distribution, and the inclusion of controls for temperature (usually
around 37 °C), pH, moisture, and airstream relative humidity (it must be
near saturation).
2.2. Biotrickling Filters
Biotrickling filters are single unit operation reactors (for both capture and
destruction), like biofilters. A packed column is inoculated with microorgan-
isms that attach to the particles. Biofilms grow using nutrients supplied by
the contaminated airstream and by a liquid flow that trickles down the pack-
ing, continuously or periodically. The liquid moving phase and the inorganic
nature of the media particles are the most important differences between
biofilters and biotrickling filters (Figure 3). Unlike biofilters that use natural
materials, most particles for biotrickling filters are built with plastic, steel,
or ceramic material. The simplest of all is the Raschig ring. Many particle
designs have been used. The odor is first transferred from the air to the circu-
lating water. Next it must diffuse to the biofilm. Finally the microorganisms
oxidize the compounds.
In previous studies using a biotrickling filter for VOC removal with Pseu-
domonas putida as the biodegrading bacteria, several packing materials were
checked to try to avoid channeling and clogging to block the process (Peixoto
and Mota, 1998). It was proved that it is very difficult to overcome the men-
tioned problems, even when using 20-mm Raschig rings. The presence of the
four phases (gas–vapor, liquid, biological, and solid) involved in the process
Clean air
Packed
column
Fresh
water
Polluted Purge
Nutrients
air
Figure 3. Schematic of a biotrickling filter.

130 B. Sercu et al.
makes things very hard to deal with. Even with a high surface area and poros-
ity, the bacterial growth reduced significantly those parameters in a short time.
Ultimately the flow used only one last channel (Figure 4), with evident poor effi-
ciency and higher pressure drop, and the process had to stop. Clogging is a direct
result of the bacterial (fungal) growth. Growth means that the microorganisms
are metabolizing the pollutants, as they are meant to. Therefore, it does not seem to
make sense to try to limit the growth to avoid clogging and channeling. A way
to remove the exceeding biomass seems to be the natural answer but very tough to
find and a settling tank would be needed after the reactor. Backwashing and high
shear stress do not seem to solve the problem.
2.3. Biowashers
In biowashers the biomass is suspended in the liquid phase. The waste content
from the air is first washed and next oxidized by the suspended microorganisms.
There may be one or two separate units for absorption and metabolization.
Classically, the airstream is washed in a spray chamber (scrubber; packed
bed scrubber) and then the liquid phase is sent to an activated sludge tank, where
the pollutants are oxidized (Figure 5).
Together, these two units form a bioreactor which is commonly named bio-
scrubber. Instead of the activated sludge tank, the contaminated liquid may be
sent to an airlift reactor (Ritchie and Hill, 1995; Rittmann et al., 2000). This
kind of reactor, also named circulating-bed biofilm reactor, is known for its good
Liquid In Liquid In
Air/Vapor Out Air/Vapor Out
Air/Vapor In Air/Vapor In
Liquid Out Liquid Out
Figure 4. Evolution of a biotrickling filter packed with Raschig rings, after colonization by bacteria,
due to clogging and chaneling.
Clean air Clean air
Airlift reactor
Washing
chamber
(scrubber)
Activated sludge
tank Polluted air,
Polluted fresh water
air Air (O2) and nutrients
Figure 5. Schematic of biowashers: to the left, a two unit reactor with scrubber and activated sludge;
to the right, a single unit airlift reactor.
mass transfer properties. Although the biomass in an airlift reactor is attached to

particles, these are still suspended in the circulating liquid phase.
In a simpler way, absorption and oxidation may happen in a single unit
operation. The airstream is directly bubbled to the liquid phase of the activated
sludge tank or the airlift reactor (Figure 5), where regeneration takes place. In
both cases, the air is washed and at the same time it supplies oxygen for the
aerobic oxidation and mechanical power for the suspension and agitation of the
particles/biomass in the bulk water. For this form of operation, the liquid phase
never leaves the reactor, only being added to replace losses, due to evaporation
and sludge purge. The air (or pure oxygen) for the oxygenation of the tank water
is replaced by the polluted airstream. This simple solution, if fit, may be eco-
nomically interesting.
Bioscrubbers are only sporadically used, mainly for removing high concen-
trations of highly water-soluble compounds. They have been used in the treat-
ment of waste gases from incinerators and foundry industry (amines, phenol,
formaldehyde, ammonia).
2.4. Comparison of Technologies
Table 1 summarizes the properties of the bioreactors described above and

includes the biological plate tower (BPT). When the environmental conditions
become toxic or aggressive, the biofilms play an important role in protecting the
microorganisms in them. Therefore, attached biomass is less prone to inhibition
and destruction. Besides, the growth rate may exceed the dilution rate without
washing out the biomass. Biofilters, biotrickling filters, and airlift reactors, all
grow attached biofilms.
132 B. Sercu et al.
Table 1. Four entry characterization of biological systems for air pollution control
LIQUID PHASE
Circulating Stationary
Biowasher Spray/
Dispersed
MICROBIAL shower
COLUMN TYPE
FLORA Biotrickling Biofilter Packing;
Attached
filter; BPT plates
Supplied with Part of the
water packing media
MINERAL NUTRIENTS
Compact biofilters reduced significantly the footprint of conventional ones

and have shortened the disadvantage related to the need of a huge application
area. The absence of a moving liquid phase is its big disadvantage. For the control
of the temperature, pH, and dissolved oxygen, and the supply of nutrients, the
presence of liquid water is a major advantage. One risk for open-bed biofilters is to
get flooded by rain or dried out by the sun. Anyway, when the application area is
not a problem, biofilters are still the most-used technology, due to the accumulated
knowledge, their low cost, and simplicity to operate. One concern related to open-
bed biofilters is the release of microorganisms to the surrounding air. Food and
fermentation industries need an environment with a controlled amount of micro-
organisms because of the nature of their processes. Van Groenestijn and Hesselink
(1993) refer to emissions up to 104 colony-forming units per cubic meter of treated
gas. To minimize the risk to the process, an enclosed biofilter with an induced
draft system (vacuum), as described in Devinny et al. (1999), should be preferred.
This way only clean air will be drawn into the system if leaks are present. The final
ventilation ducting must be positioned accordingly.
Biotrickling filters are used because of their ability for process control and
are recommended when high concentrations of (acidifying) compounds have
to be treated, or when only limited area space is available on site. Biowashers
are favored when the pollutants have high solubility in water. Otherwise they
are water-consuming and less attractive. If an existing activated sludge tank can
receive the waste air from another process, that fact may be very profitable. It is
necessary to guarantee that the new pollutants will not interfere with the acti-
vated sludge performance. Stripping of dissolved odors, flora inhibition, and
filamentous bulking are some of the problems that may appear or get out of
control.
Contacting times, application area, toxicity of pollutants, acid production
during degradation, clogging, chaneling, solubility in water, adhesion to pack-
ings, costs of technology, one or several pollutants, selected species or natural
consortia, pressure drop, selection of support media, open or closed-beds, and
so on, are variables to be considered. With so many variables, there is still a long
way to go before one can be sure of what solution is the best. For the same rea-
son, modeling is also very difficult. For further information about reactor design
and modeling, books of Devinny et al. (1999) and Kennes and Veiga (2001) are
recommended.
2.5. New Technologies
2.5.1. Biological Plate Tower (BPT)
The big mistake in the transposition from physical–chemical to biological reac-

tors is to forget that the presence of biofilms completely changes the behavior and
performance of the reactor. Biofilm growth is chaotic and never tridimensionally
homogeneous on a random packing. Oriented packing has better results but not
yet good enough. A good physical–chemical reactor does not have to be a good
biological one and it is indeed a poor option in many situations. Four-phase reac-
tors always bring about hydrodynamic problems. Sloughing, channeling, and clog-
ging always occur. A good efficiency of removal makes it happen faster. To solve
these problems and make the process easy to operate steadily for a long time, a new
concept of reactor was designed and tested with air polluted with VOCs. Pseu-
domonas putida was the selected inoculum. The effluent simulation was achieved
with the mixing chamber described in Peixoto and Mota (1997).
The observation of the growth on the plane surfaces (top liquid distribu-
tor, base plate) of a biotrickling filter (Figure 6) suggested a design based on
Figure 6. Liquid distributor showing the growth of biomass on the lower surface. On the top side,
that occurrence was even stronger.
134 B. Sercu et al.
horizontal surfaces. Basically, the BPT is a pile of parallel circular plates with a
single hole on the border. The plates are placed in such a way that the holes will
alternate (180°) from one to the next plate. In this way, a cascade of liquid will go
downward, changing direction from plate to plate. The gaseous stream follows
the opposite direction, upward. The bacteria attach to their top surface. Figure 7
shows the schematic of the flows and biofilm growth on the plates.
The reactor is a four-module (about 28.8 dm3 each) BPT with 20 plates in
each module. An individual plate surface area (top face) is about 40 195 mm2.
The scratched surfaces of the plates were intended to make the bacterial adhe-
sion easier. Only two or three of the four modules are operated continuously.
The other(s) is kept free and ready to replace any one that reaches saturation
with biomass. In this way the operation can be kept going virtually forever.
The performance is quite stable (the biofilm activity, surface-dependent, is
kept approximately constant) and the constant surface contact area makes it
easy to model and scale-up the process. The total surface area and the space
between plates can be designed for the desired operating time. In theory, the
available surface in a BPT is a tenth of the surface in a biotrickling filter, consid-
ering the same total volume.
The new design proved to ensure a stable operation for longer periods, as well
as high VOC removal (92 % removal for inlet toluene concentration of 10 g m−3
and empty bed residence time, EBRT = 108 sec). It has very good hydrodynamic
performance and operates continuously without problem. In the long term, the
short area is compensated by the steady operation.
The disposal of the newly formed biomass is also much easier than in the
biotrickling filter. Unlike biofilters whose packing has to be rejected after a certain
Air/Vapor Liquid Liquid
Air/Vapor Air/Vapor
Liquid Liquid
Figure 7. Simplified schematics of the BPT, with only five plates, to better visualize the directions of
both flows and the attached biofilm on the upper surface of the plates.
time of operation, BPT biomass is withdrawn as a water-rich solid phase—the

biofilm attached to the plates—and quite easy to handle. When the thickness
of the biofilm reaches the maximum value allowable, the set of plates is simply
replaced by a clean one and the biomass is dealt with outside the reactor. Sam-
pling the biofilm for analysis is very easy. It does not oblige the operation to be
stopped, or severely shaken as it happens with biotrickling filters, and any plate
can be sampled. Even operation demands a constant surface of biofilm. Oxygen
uptake rate measurements were made to find out if there were great activity dif-
ferences between different plates and between the surface and inside the biofim.
The respiratory activity was similar (about 0.11 mg g−1 s−1, mass of oxygen per
mass of volatile solids per time) for the superficial samples of all plates, show-
ing some difference (up to about 20 %) for the lower ones where it was higher.
The middle samples had almost zero activity (0.01 mg g−1 s−1 or less) and none
of the base samples showed any activity. For the respirometry tests, the carbon
source was phenol. The plates at the bottom of each module had thicker biofilms
than the upper ones, due to the higher concentration of the carbon source
and oxygen in the entrance. The first module, which receives the higher dose, is
the one that shows the thickest films, reaching over 15 mm until needing to be
replaced (Figure 8).
The research on the BPT is still ongoing. Assays to quantify VOC and odor
removals are now being planned. In the future, different bacteria, plate shapes, and
distances between plates will be tested. The bacterial growth hanging from holes in
the plates (sieve trays, similar to Figure 6) also will be investigated. The possibility
of using different bacteria in different modules also will be considered.
Figure 8. Photograph of the bottom plates of the first module showing the biofim growth on the
BPT plates. The huge biofilm does not endanger the permeability of the system.
136 B. Sercu et al.
2.5.2. Membrane Bioreactor
In the membrane bioreactor concept, one side of the membranes is dry and acts
as a surface for uptake of pollutants from the air flowing along the membranes,
while the other side is kept wet and covered by a biofilm. In Figure 9, a flat mem-
brane bioreactor with a composite membrane is shown, but also other configu-
rations like hollow fiber membranes modules can be applied. Pollutants diffuse
through the membrane and are subsequently degraded by the microorganisms in
the biofilm or in the recirculating aqueous phase. By continuous recirculation of
the aqueous phase, the microbial degradation process can be easily controlled. The
main advantages of membrane bioreactors for waste gas treatment include the
high specific surface area, the ability to prevent clogging, the good reactor control,
the physical separation of gas and biofilm, the low pressure drop, the absence of
channeling, and the independent control of gas and liquid phase (De Bo, 2002).
Potential disadvantages are the high investment costs, the additional mass transfer
resistance caused by the membrane, a decreased biofilm activity as the biofilm
ages, and clumping of hollow fiber membranes at high biofilm growth. The reac-
tor concept, although not implemented in practice yet, has potential to eliminate
VOCs characterized by poor water solubility, by lack of biodegradability, and by
toxicity (Reij et al., 1998). Recently a flat membrane reactor was developed and
applied for the degradation of DMS and toluene as single compounds (De Bo
et al., 2002, 2003). In this case a composite membrane was used, guaranteeing a
Biofilm
Membrane
Substrate
Nutrients
Oxygen
pH-buffers
Mineral
Waste
medium
air
Dense Porous
Figure 9. Scheme of a flat membrane bioreactor for waste gas treatment (De Bo, 2002).
stable long-term reactor performance because clogging of the porous membrane

was prevented. For DMS, an ECmax of 4.8 kg m−3 d−1 was obtained, which was
higher than any reported figure for biofilters or biotrickling filters.
3. REMOVAL OF COMMON ODORS
Most of the work has been done about removal of ammonia, reduced sulfur
compounds (either single or as a mixture), and odorous VOCs, especially with
biofiltration. About compounds like amides, indole, scatole, and pyridine no
information was found. However, it can be assumed that when bioreactors can
operate with high odor reduction efficiency, these compounds are also suffi-
ciently degraded.
3.1. Removal of Ammonia
Within the group of odorous nitrogen-containing compounds, ammonia is by

far the most investigated and documented compound with respect to its removal
from waste gases by biological treatment technologies. Although ammonia has
a rather high odor threshold value (3.9 ppmv to 5.8 ppmv) compared to many
other odorous (nitrogen) compounds (Weckhuysen et al., 1994; Devos et al.,
1990), and consequently dilutes rapidly to below detection downwind from the
emission source, its very sharp and unpleasant odor can cause a severe odor
nuisance nearby its emission source. Among biotechnological waste gas treat-
ment systems, mainly biofilters have been used to control emissions containing
ammonia. The main processes taking place during biofiltration of ammonia are
presented in Figure 10.
GAS PHASE BIOFILM FILTER MATERIAL
I II
NH3 NH3 NH4+ NH4+
Nitrosomonas
III
NO2 −
Nitrobacter
NO3−
Figure 10. Main processes taking place during ammonia biofiltration (I: absorption; II: adsorption;
III: nitrification).
138 B. Sercu et al.
Since ammonia has a low Henry’s law constant (H20 °C = 5.6 × 10−4) (Perry
and Green, 1984) and a protonation constant pKa, 20 °C of 9.23 (Weast et al.,
1984), in biofilters it is partly retained by adsorption onto the carrier material
and by absorption into the water fraction of the carrier material. In this context,
Shoda (1991) reports a maximum volumetric NH3 elimination capacity of 15 g
m−3 d−1 in a peat biofilter due to these physical–chemical transfer processes. In
a compost biofilter, Smet et al. (2000) obtained an NH3 adsorption and absorp-
tion capacity, per volume of compost, of 490 g m−3 and 47 g m−3, respectively,
at an NH3 inlet concentration of 159 ppmv and a compost moisture content
of 40%. Next to these physical–chemical processes, nitrification by the autotro-
phic bacteria Nitrosomonas and Nitrobacter is generally considered as the main
microbiological process for the degradation of NH3 (Terasawa et al., 1986; Van
Langenhove et al., 1988; Williams, 1995). More recently, the autotrophic genera
Nitrosospira and Nitrospira also are reported to be responsible for nitrification
(Schramm et al., 2000; Regan et al., 2002).
As a result of these phase transfer and (micro)biological processes, NH3 has
been removed efficiently at concentrations up to 50 ppmv. Using a wood bark
biofilter, Van Langenhove et al. (1988) obtained removal efficiencies of at least
90% at concentrations between 6 ppmv and 17 ppmv and at NH3 mass loading
rates (Bv) up to 58 g m−3 d−1. At similar concentrations (4 ppmv to 16 ppmv),
Weckhuysen et al. (1994) observed elimination efficiencies of 83% or higher at
NH3 mass loading rates between 6.8 g m−3 d−1 and 27.2 g m−3 d−1. In an inoculated
peat biofilter, NH3 elimination capacities (EC) up to 41 g m−3 d−1 are reported
at an inlet concentration of 20 ppmv (Hartikainen et al., 1996). More recently,
removal efficiencies as high as 99.5% were obtained for 100 days in an inoculated
perlite biofilter, at concentrations of 50 ppmv and NH3 loading rates between
8.6 g m−3 d−1 and 21.5 g m−3 d−1 (Joshi et al., 2000).
Due to the sensitivity of nitrifying microorganisms, however, biofiltration of
waste gases containing high ammonia concentrations (above 50 ppmv) has been
reported to be questionable. Don (1985) and Hartikainen et al. (1996) reported the
biofilter removal efficiency for NH3 dropped drastically at waste gas concentrations
exceeding 35 ppmv to 60 ppmv. However, more recently, Liang et al. (2000) could
obtain NH3 removal efficiencies of at least 95% at inlet concentrations between
20 ppmv and 500 ppmv in compost biofilter in which active carbon was added to
reduce compaction and channeling, as well as to increase the reactive surface and
durability of the biofilter. Similarly, removal efficiencies over 90% were achieved
by Kim et al. (2002) at inlet concentrations up to 150 ppmv in a biofilter system
packed with small cubes of polyurethane sponge coated with a powder mixture
of activated carbon and natural zeolite. Kalingan et al. (2004) obtained complete
NH3 removal in a peat biofilter containing inorganic supporting material at a NH3
concentration of 200 ppmv, while Smet et al. (2000) observed no NH3-toxicity in a
compost biofilter at concentrations up to 775 ppmv.
Besides the ammonia input concentration, the mass loading rate of a bio-
filter seems to be critical for an efficient performance. In an inoculated activated
carbon biofilter, Yani et al. (1998) found a complete NH3 removal up to Bv = 95 g

m−3 d−1, whereas the elimination efficiency decreased at higher loading rates. The
highest EC observed by these authors was 220 g m−3 d−1 at Bv = 250 g m−3 d−1
and at an EBRT of 52 s. Smet et al. (2000) and Demeestere et al. (2002) obtained
elimination capacities up to 350 g m−3 d−1 in a compost biofilter at EBRT = 131 s
and 21 s, respectively. EC peak values of 530 g m−3 d−1 and 1285 g m−3 d−1 were
reported at inlet concentrations of 250 ppmv (Bv = 600 g m−3 d−1) and 450 ppmv
(Bv = 1329 g m−3 d−1), respectively (Demeestere et al., 2002). According to these
authors, the cumulative loading (mass of NH3 per filter material volume, g m−3)
is the limiting factor for a NH3 degrading biofilter. Smet et al. (2000) observed
a sharp reduction in elimination after a cumulative NH3 removal of 6000 g m−3.
Osmotic effects, due to the accumulation of NH4NOx at concentrations (mass
of NH4NOx-N per mass of compost) higher than 4 g kg−1, were found to be the
reason for the inhibition of NH3 removal (Smet et al., 2000; Demeestere et al.,
2002). However, a subsequent loading of the biofilter with a carbon source like
methanol could regenerate the biofilter material, due to methylotrophic conver-
sion of NH4+ and NOx− into biomass (Demeestere et al., 2002).
In order to achieve optimum NH3 removal in biofilters, the moisture content
of the filter material should be between 40% and 60%, the temperature between
30°C and 35°C and the pH between 7 and 8 (Van Lith et al., 1997; Warren et al.,
1997). With respect to the latter parameter, acidification of the filter material due
to the accumulation of nitrite and/or nitrate can inhibit the long-term stability
of a NH3 degrading biofilter, as reported by Heller and Schwager (1996). On the
other hand, no acidification was observed by Don (1985) and Smet et al. (2000),
who attributed that effect to the establishment of an equilibrium between NH3
absorption increasing the pH and nitrification decreasing the pH.
Next to the removal of NH3 in waste gases by biofiltration, bioscrubbers and
biotrickling filters also have been used for NH3 degradation. Due to the presence
of a recirculating water phase, both these techniques allow to drain off accumu-
lating toxic compounds, to control the pH, and to add nutrients. As an example,
Smits et al. (1995) obtained a biological elimination capacity of 96 g m−3 d−1 in
a pilot-scale biotrickling filter at superficial gas and liquid velocities of 1300 m
h−1 and 2.5 m h−1, respectively. No gas-to-liquid mass transfer limitation was
observed under these conditions. Due to its low Henry’s law coefficient, efficient
NH3 scrubbing from the gas phase can be obtained. However, it was observed
by the authors that up to 70% of the ammonia removed from the waste gas was
not nitrified but removed with the drain water. Consequently, although the use
of both biotrickling filters and bioscrubbers can be very efficient to remove NH3
from the waste gas, it implicates the subsequent treatment of the NH4+-loaded
drain water in a wastewater treatment plant. Another drawback of these tech-
nologies is the relative low removal of less water-soluble odorous compounds in
bad-smelling waste gases.
Although it is shown by some authors (Tang et al., 1996; Chou and Shiu,
1997; Busca and Pistarino, 2003; Chang et al., 2004) that also other odorous
140 B. Sercu et al.
nitrogen-containing compounds than ammonia can be efficiently removed by

biological waste gas treatment technologies, fewer experimental data are pub-
lished so far in that field. For example, as far as we know, there is no information
available about the biotechnological removal of nitrogen compounds like amides,
indole, scatole, and pyridine. Nevertheless, some reports deal with the removal of
gaseous amines by biofiltration. Amines are bad-smelling compounds that often
are present in waste gases arising from fish markets, meat treatment industries,
and other food industries (Busca and Pistarino, 2003). According to Chou and
Shiu (1997), methylamine (MA) can be successfully removed, i.e., hydrolyzed to
ammonia and nitrified to nitrate and/or incorporated into microbial biomass,
in peat biofilters at mass loading rates up to 160 g m−3 d−1, at a pH between 7.5
and 8.5, and at a moisture content between 55% and 60%. Tang et al. (1996)
investigated the removal of triethylamine (TEA, 78 ppmv to 841 ppmv) in a bio-
filter consisting of a mixture of compost and chaff particles and obtained the
highest TEA elimination capacity of 3360 g m−3 d−1 at an inlet concentration of
550 ppmv, above which substrate inhibition occurred. However, a comprehensive
picture cannot be drawn for these compounds due to scarceness of the literature
available (Busca and Pistarino, 2003).
3.2. Removal of Hydrogen Sulfide
Next to ammonia, H2S biofiltration has been studied extensively, because it is

one of the most frequently produced odorous compounds in industrial processes
like petroleum refining, rendering, wastewater treatment, food processing, and
paper and pulp manufacturing (Yang and Allen, 1994a). The bacteria respon-
sible for H2S degradation in biofilters mostly belong to the genera Thiobacillus
(e.g., T. thioparus) and Acidithiobacillus (e.g., A. thiooxidans) and can be either
neutrophilic or acidophilic. Under optimal conditions, H2S is oxidized to sulfu-
ric acid, but during stress conditions (high loads, oxygen limitation) accumula-
tion of elemental sulfur has been observed.
Because H2S is very biodegradable, most investigations report very effi-
cient H2S removal in a wide concentration range. Yang and Allen (1994a), for
instance, observed higher than 99.9% removal efficiencies for H2S inlet concen-
trations ranging from 5 ppmv to 2650 ppmv. However, because sulfuric acid is
produced, acidification of the filter material will inevitably occur during the bio-
filtration process, its rate depending on the buffer capacity of the filter bed and
the amount of H2S removed. Degorce-Dumas et al. (1997) found that buffering
the packing to a near neutral pH doubled the length of the period during which
> 95% H2S removal efficiency was obtained. When the pH decreased below 6.6,
the H2S removal efficiency started to decrease, together with the number of non-
acidifying thiobacilli. Instead, acidifying thiobacilli became dominant. There-
fore, a correlation between the number of nonacidifying thiobacilli and the H2S
removal efficiency was suggested. Other authors, however, observed a smaller
effect of acidic pH values on the H2S removal efficiency. Yang and Allen (1994a),
for instance, found almost equal H2S removal efficiencies at pH values between
3.2 and 8.8. Only at pH = 1.6, the removal efficiency decreased to 15%. The high
H2S removal efficiency at pH = 3.2 was attributed to the abundance of acido-
philic sulfur oxidizing bacteria. Also other studies did not report decreased H2S
removal efficiencies at pH values as low as 3 (Wada et al., 1986; Cook et al., 1999)
or even 1.2 (Yang et al., 1994). During biofiltration, the pH will first decrease at
the inlet side of the biofilter, where most of the H2S is oxidized and the low pH
front will consequently move to the deeper parts of the biofilter (Cook et al.,
1999). In general, it should be sufficient to maintain a pH value higher than 3 for
efficient H2S removal. However, it could be useful to maintain neutral pH values
to prevent inhibition of the removal of other compounds present in the waste
gas, corrosion, and increased filter medium degradation. To increase the pH of
the biofilter material, washing can be applied (Yang and Allen, 1994b), although
only small pH increases are usually obtained. Smet et al. (1996b) observed that
regeneration of an acidified biofilter (pH = 4.7) was not possible by trickling
tap water or buffer solution over the bioreactor, because most of the sulfate was
leached as the corresponding sulfate salts and not as sulfuric acid. In addition,
leaching caused washout of essential microbial elements. Alternatively, the use
of more concentrated buffer solutions in combination with a complete mineral
medium or mixing with limestone powder was recommended.
Next to acidification, the accumulation of elemental sulfur and sulfate in the
filter material can potentially inhibit microbial activity. Yang and Allen (1994b)
found the highest concentrations of both compounds at the inlet side of the
biofilter. Elemental sulfur was present because it was formed as an intermedi-
ate during incomplete H2S oxidation after exposure to high H2S concentrations.
By adding increasing amounts of sulfate to different biofilters, Yang and Allen
(1994a) observed that concentrations (mass of S per mass of compost) exceed-
ing 25 mg g−1 were inhibitory for H2S removal, probably due to toxic effects. This
inhibition effect, however, was not confirmed by Jones et al. (2003), for sulfate
concentrations up to 100 mg g−1. In general, it is recommended to evaluate the
expected H2S loading rate before designing a biofilter. If it is assumed that all
sulfur entering a biofilter will ultimately accumulate as sulfate, its cumulative
concentration can be calculated to assess the long-term deactivation of a biofil-
ter, e.g., with a threshold of 25 mg g−1.
Because H2S is very biodegradable, EBRTs can be rather low, e.g., 15 s
(Yang and Allen, 1994a) without affecting the H2S removal efficiency. Possibly
other, less biodegradable or water-soluble compounds present in the waste gas
will determine the lower limit of the EBRT. Next to organic materials like com-
post, peat, or wood bark, different alternative carrier materials were described
for H2S biofiltration, being rockwool, fuyolite, and ceramics (Kim et al., 1998),
a pelletized mixture of pig manure and sawdust (Elías et al., 2000), pellets of
agricultural residues (Elías et al., 2000), porous lava inoculated with Thiobacillus
thiooxidans (Cho et al., 2000), and microorganisms immobilized in Ca-alginate
(Chung et al., 1996a,b, 1997, 1998; Huang et al., 1996; Park et al., 2002).
142 B. Sercu et al.
Next to biofilter applications, more recent articles describe H2S removal

with biotrickling filters. Their main advantage is optimal control of pH, nutri-
ents, and accumulation products, although of course treatment costs are higher.
At an EBRT between 30 s and 120 s, high H2S removal efficiencies (> 95%) eas-
ily can be obtained for H2S concentrations between 200 ppmv and 2000 ppmv
(Ruokojärvi et al., 2001; Sercu et al., 2005b). At lower influent concentrations,
lower EBRTs can be used at high removal efficiencies. Gabriel and Deshusses
(2003) described the retrofitting of existing chemical scrubbers for H2S removal
to biotrickling filters, maintaining an EBRT between 1.6 s and 2.2 s. Removal
efficiencies > 98% were commonly reached for 30 ppmv inlet concentrations,
with decreases to 90% at 60 ppmv peak concentrations. The removal of volatile
organic sulfur compounds in the same reactor was lower, however, e.g., 35%
± 5% for carbon disulfide. The authors attributed the residual odor after the
biotrickling filter mainly to the persistence of these compounds. Also Wu et al.
(2001) obtained > 95% H2S removal efficiency at EBRT = 5 s, at < 6 ppmv influ-
ent concentrations in a pilot-scale biotrickling filter. At 20 ppmv influent concen-
tration the removal efficiency decreased to about 89%.
3.3. Removal of Ammonia and Hydrogen Sulfide
A number of studies have been performed regarding the simultaneous removal

of H2S and NH3, because both can constitute an important part of odorous gas
mixtures. Similarly as with the removal of the separate compounds, high removal
efficiencies can be obtained during simultaneous dosing of both compounds, at
concentration levels usually occurring in odorous mixtures (< 50 ppmv). At higher
H2S and NH3 concentrations, inhibition of the NH3 removal can occur. Kim et
al. (2002), for instance, obtained higher than 99% and 92% removal efficiencies
for H2S and NH3, respectively, in a wood chips biofilter, at influent concentra-
tions of about 50 ppmv (EBRT = 1 min). At concentrations exceeding 200 ppmv,
however, H2S inhibited the NH3 removal, which decreased to 30%, but this effect
was reversible when the H2S concentration decreased again. By using a granulated
activated carbon biofilter, the inhibition during H2S peak loadings decreased due
to buffering effects. An important aspect of simultaneous H2S and NH3 biofil-
tration is that the extent of the pH decrease, caused by production of sulfuric
and nitric acids, can decrease, because the accumulation of acidic products can
be small due to (NH4)2SO4 formation. At NH3 concentrations equal or higher
than H2S (on volumetric basis), Chung et al. (2000), for instance, observed no
acidification. Recently, Chung et al. (2004) showed that acidification during H2S
and NH3 removal could further be decreased considerably by selection of hetero-
trophic bacteria (Pseudomonas putida CH11 for H2S and Arthrobacter oxydans
CH8 for NH3). Heterotrophs oxidize H2S and NH3 mainly to elemental sulfur
and organic nitrogen, causing only very small production of acidic end-products.
In the activated carbon biofilter, H2S and NH3 concentrations between 20 ppmv
and 120 ppmv could be very efficiently removed during 180 d. A carbon source
had to be supplied every two weeks to support growth of the heterotrophic organ-
isms. A number of researchers used microorganisms immobilized in Ca-alginate
to remove mixtures of H2S and NH3, although the performance of these reactors
was somewhat lower than with the more traditional biofilters (Chung et al., 2000,
2001a,b). Possible advantages, however, are increased possibilities for pH control
and removal of metabolic products (elemental sulfur and (NH4)2SO4), as is also
the case with biotrickling filters.
3.4. Removal of Volatile Organic Sulfur Compounds
Volatile organic sulfur compounds (VOSCs) include compounds like dimethyl

sulfide (DMS), dimethyl disulfide (DMDS), mercaptans, and carbon disulfide.
These compounds have been related to odor complaints in some studies. A direct
correlation could even be established between the total odor concentration and
the concentration of VOSCs in waste gases of rendering plants (Defoer et al.,
2002). Van Langenhove et al. (1992) compared a full-scale biotrickling filter and a
biofilter for treating rendering emissions. Both techniques removed alkanals very
efficiently, but organic sulfur compounds were much less efficiently removed.
This was attributed to an insufficient development of microorganisms capable
of degrading these compounds. Goodwin et al. (2000) also observed problems
removing reduced sulfur compounds with a biofilter at a biosolids composting
facility. Increasing the EBRT from 20 s to 32 s improved the removal efficiency
somewhat. In contrast, VOCs like methane, formaldehyde, isopentanal, N,N-
dimethyl methenamine, and dimethylamine were removed for more than 95% in
all cases at average inlet concentrations of 15 ppmv.
Different reasons can explain the relation of VOSCs and odor nuisance.
First of all, VOSCs combine a very bad smell with very low odor threshold val-
ues, for instance 0.1 ppbv to 3.6 ppbv for DMDS and 0.9 ppbv to 8.5 ppbv for
methyl mercaptan (Smet et al., 1998). This means that to prevent odor nuisance
only very low concentrations can persist in the treated gas stream. Second, com-
pared with H2S, VOSCs are less biodegradable. Degradation rates decrease in the
order H2S > MM > DMDS > DMS (Cho et al., 1991; Smet et al., 1998). There-
fore, it is recommended to inoculate biofilters to shorten the start-up period
and to remove high concentrations of these compounds. Smet et al. (1996a), for
instance, increased the maximal DMS elimination capacity from 10 g m−3 d−1
to 680 g m−3 d−1 after inoculation of a compost biofilter with Hyphomicrobium
MS3. Also other authors used inocula (e.g., Hyphomicrobium spp., Thiobacillus
spp.) to remove VOSCs in biofilters (Cho et al., 1991, 1992; Zhang et al., 1991;
Park et al., 1993). However, in full-scale applications, the use of inoculation is
not well documented. Smet (1995) reported successful removal of organic sulfur
compounds in a full-scale biofilter treating emissions from mushroom compost-
ing, after inoculation with a specialized strain. Fifty days after inoculation, the
total sulfur removal efficiency (excluding H2S-S) in the inoculated biofilter sec-
tion had increased to 99% compared with 68% in the noninoculated section. But
144 B. Sercu et al.
even when inoculation is used, in a mixture of reduced sulfur compounds, H2S is

preferentially degraded over dimethyl sulfide or other organic sulfur compounds
(Cho et al., 1992; Wani et al., 1999; Zhang et al., 1991). This occurs because H2S
oxidation yields most energy for the microorganisms (Smet et al., 1998). There-
fore, the bioreactor has to be designed large enough to allow H2S degradation at
the inlet side of the biofilter and degradation of the remaining VOSCs deeper in
the biofilter bed. Finally, when a biofilter is designed properly to remove VOSCs,
there is still a change of long-term decrease in removal efficiency because of
acidification. Similarly as for H2S, sulfuric acid is formed after complete oxida-
tion of VOSCs. Microorganisms degrading the VOSCs, however, are much more
sensitive to low pH values than H2S oxidizing bacteria. Smet et al. (1996b), for
instance, observed a decreased DMS elimination capacity when the compost pH
decreased below 5. To prevent problems due to acidification, the bioreactor has
to be designed large enough, and for high influents loadings pH control should
be included. Alternatively, two-stage systems have been proposed, first remov-
ing H2S and subsequently VOSCs (Kasakura and Tatsukawa, 1995; Park et al.,
1993; Ruokojarvi et al., 2001; Sercu et al., 2005b). Ruokojarvi et al. (2001), for
instance, developed a two-stage biotrickling filter for sequential removal of H2S,
methyl mercaptan (MM) and DMS. Two bioreactors connected in series were
inoculated with enriched activated sludge, the first operating at low pH for H2S
removal and the second at neutral pH for DMS removal. MM was removed in
both reactors. H2S, DMS and MM elimination capacities (as S) as high as 47.9 g
m−3 h−1, 36.6 g m−3 h−1 and 2.8 g m−3 h−1, respectively, were obtained for the
entire two-stage biotrickling filter at > 99% removal efficiencies and the reactor
showed a good long-term stability.
3.5. Removal of Odorous VOCs
Generally, odorous VOCs are biodegradable in biofilters (Van Langenhove et

al., 1989b, 1992; Goodwin et al., 2000). In most cases these compounds are not
the cause of odor problems when biofilter malfunctioning occurs, and therefore
literature data about the removal of low concentrations of these compounds are
less available than, for example, ammonia and hydrogen sulfide.
The removal of aldehydes, alcohols, and fatty acids is generally very good
in biofilters (Kiared et al., 1997; Mohseni and Allen, 2000; Otten et al., 2004;
Sheridan et al., 2003; Weckhuysen et al., 1993). For methanol, for instance,
it was, found that concentration step changes and periods without methanol
loading did not affect its removal efficiency in biofilters (Mohseni and Allen,
1999), which was attributed to the good biodegradability and high water solu-
bility of methanol. In some studies it was shown that nutrient addition could
enhance VOC elimination capacities during longer periods, e.g., in the case of
butanal (Weckhuysen et al., 1993) or butyric acid (Sheridan et al., 2003). It has
been observed that in the case of aldehydes, the corresponding organic acids
can accumulate during biofiltration, especially at higher influent concentrations
(Weckhuysen et al., 1993; Sercu et al., 2005a). This can lead to a pH decrease,
potentially limiting the removal efficiencies of other compounds in the waste
gas. Next to biofilters, biotrickling filters have been used to remove odorous
VOCs. Again, high removal efficiencies have been obtained for aldehydes, alco-
hols, and volatile fatty acids (Chang and Lu, 2003; Chua et al., 2000; Ibrahim
et al., 2001; Kirchner et al., 1991), even at low EBRT. Kirchner et al (1987), for
instance, showed > 90% removal efficiencies for compounds like aldehydes and
alcohols at 5 ppmv to 40 ppmv influent concentrations and 2.4 s EBRT. Ibrahim
et al. (2001) found 92% and 95% removal efficiencies for 10 ppmv acetaldehyde
and propionaldehyde inlet concentrations, respectively, in a column packed with
immobilized activated sludge beads at EBRT = 12.4 s. At higher influent con-
centrations, the removal of both compounds decreased, however, due to inhibi-
tory effects. For higher influent concentrations, higher EBRT values are needed,
as shown by Chang and Lu (2003). They found nearly complete isopropanol
removal efficiencies in a biotrickling filter, operated between 20 s and 90 s EBRT
time at influent concentrations between 100 ppmv and 500 ppmv. When too high
influent loadings are applied, accumulation of compounds or intermediates can
occur. Chua et al. (2000) found > 99% removal efficiencies for butyric and valeric
acid in a biotrickling filter, at mass loadings between 4.8 g m−3 h−1 and 37.8 g m−3
h−1 (0.05 g m−3 to 0.86 g m−3). However, at loading rates exceeding 32 g m−3 h−1,
the maximal biodegradation capacity was reached and accumulation of volatile
fatty acids in the liquid phase was observed.
3.6. Removal of Odor Mixtures
From the previous sections, it is clear that most of the components present in
odorous mixtures can be removed efficiently with biological waste gas cleaning
techniques, when properly operated, even at relatively high influent concentra-
tions. Also, in industrial applications treating mixtures of compounds, often high
(odor) removal efficiencies can be obtained. Park et al. (2001), for instance, used
a biotrickling filter packed with ceramics and inoculated with activated sludge to
remove odors at a composting facility. After a 30 d acclimation period, > 95%
removal efficiencies were obtained for NH3 and H2S during about 60 d of opera-
tion. Also, at a biosolids composting facility, Goodwin et al. (2000) found effi-
cient odor removal with a biofilter (> 95%) after about 3 months of operation at
EBRT = 20 s, as determined with olfactometric analyses. Luo (2001) observed >
98% odor reduction with wood bark biofilters treating rendering emissions during
a period of 3 years, at EBRT = 6.8 min. Reducing the EBRT to 1.7 min did not
affect the odor removal efficiencies during the first 3 months of operation. After
22 months, however, the odor removal efficiency was 99.1% at EBRT = 6.8 min
and only 29.7% at EBRT = 1.7 min. This clearly shows that regular filter medium
replacement is necessary, especially when lower EBRT values are used.
When a complex mixture of odorous compounds has to be treated, removal
efficiencies of the single compounds can be smaller than expected. This can be
146 B. Sercu et al.
caused by, e.g., toxic effects of substrates or metabolites. Van Langenhove et al.
(1989a) compared the applicability of a tree bark biofilter for removing odors
from a vegetable processing industry, mainly emitted during the blanching pro-
cess. The main odorous compounds identified were sulfides, isothiocyanates,
nitriles, and aldehydes. In pilot-scale experiments all compounds were removed
with > 95% removal efficiencies at a volumetric loading rate of 200 m3 m−2 h−1.
However, a full-scale biofilter, designed according to the results obtained from
the pilot-scale studies, had lower removal efficiencies after 6 months of opera-
tion (45% to 65% for sulfides). This was found to be caused by the accumulation
of isothiocyanates, which was not observed during the short-term pilot-scale
experiments. For hexanal, Van Langenhove et al. (1989b) observed 85% removal
efficiency in a wood bark biofilter, at 10 ppmv inlet concentration and EBRT =
0.33 min. To simulate emissions from a food processing plant, 40 ppmv SO2 was
added to the waste stream, leading to a drastic decrease of the hexanal removal
efficiency to 40%. Next to toxic effects, preferential degradation of easily biode-
gradable compounds can inhibit the removal of other compounds. Smet et al.
(1997), for instance, found that isobutanal was preferentially degraded before
DMS, in a biofilter inoculated with Hyphomicrobium MS3, when both com-
pounds were simultaneously dosed. This could cause low removal of DMS when
a biofilter is designed too small.
4. CASE STUDIES
4.1. Methodology
In all case studies mentioned in this paragraph, samples have been taken of the
untreated and the treated airflows, in order to determine important parameters.
First, the chemical composition of the airflow was revealed using GC-MS analy-
sis. These data are very useful for the determination of the total chemical load
going to the bioreactor, as well as for improving the working efficiency of it,
being able to indicate the compounds or groups of compounds that are degraded
insufficiently. The second type of analyses used is the determination of the total
odor concentration, using dynamic olfactometry. These data are used to deter-
mine the total odor removal efficiency of the bioreactor, which is the final wanted
effect of the use of a bioreactor in case of odor problems.
4.1.1 GC-MS Analysis
4.1.1a. Sampling procedure. The gases were sampled using a method that
involved preconcentration on an adsorbent. This preconcentration step was
carried out at the sampling location. Tenax TA was used as adsorbent. Tenax
TA is a porous polymer based on 2,6-diphenylene oxide. It has been specifically
designed for the trapping of volatiles and semivolatiles (SIS, 2000). The collected
waste gases were cooled at about 4 °C before adsorption. This cooling stage
was used in order to increase the breakthrough volume and in order to separate
excess water vapor. The sampled adsorption tubes were filled with approximately
750 mg Tenax. Sampling rate was 200 ml min−1 and sampled volume varied
between 50 ml and 10 L.
4.1.1b. Analysis. The analysis of the VOCs present in the sample was done
in different steps, including desorption from the adsorbent, separation by gas
chromatographic techniques, quantification by flame ionization detection, and
subsequent identification through mass spectrometry. The desorption step con-
sisted of a thermal desorption. A second preconcentration (cryogenic trapping
of the VOCs) was necessary in order to achieve good chromatographic separa-
tion. The cryogenically concentrated samples were introduced immediately into
the GC by rapid heating of the trap. The different compounds were separated
in a gas chromatograph (Varian 2700) with a 100% polydimethylsiloxane apolar
column (type DB-1, 30 m × 0.53 mm, film thickness 5 μm, J&W Scientific). The
mass spectrometer used was a Finnigan MAT 112 S with an electron impact ion
source and a magnetic sector analyser.
4.1.2. Olfactometry
4.1.2a. Sampling procedure. The gases were sampled using the static
sampling method. In this method, a sample is collected and transferred into
a sampling container (bag). Collecting the sample was done with the “lung
principle,” where the sample bag is placed in a rigid container and the air is
removed from the container using a vacuum pump. The partial vacuum cre-
ated in the container causes the bag to fill with a volume of sample equal to
the volume that was removed from the space around the bag in the rigid con-
tainer. In some sampling points, where a risk of condensation in the sampling
bag existed due to high humidity and high temperatures, a predilution was
applied using dry odor-free nitrogen. Sampling materials used were Teflon™
for tubing and disposable sampling bags made of Nalophan™ film. All samples
were analyzed by the accredited odor laboratory of PRA OdourNet bv (The
Netherlands) within 30 hours after sampling. During transportation, samples
were not exposed to direct sunlight. All measurements of the odor concen-
trations were executed by dynamic olfactometry according to the EN13725
(CEN, 2003).
4.1.2b. Principle of dynamic olfactometry. The odor concentration of a

gaseous sample of odorants is determined by presenting a panel of selected and
screened human subjects with that sample, varying the concentration by dilut-
ing with neutral gas, in order to determine the dilution factor at the 50% detec-
tion threshold. At that dilution factor, the odor concentration is 1 ouE m−3 by
definition. The odor concentration of the examined sample is then expressed
148 B. Sercu et al.
as a multiple of one European odor unit per cubic meter (ouE m−3) at standard
conditions for olfactometry.
4.2. Odor Removal at a Vegetable Oil Extraction Plant
4.2.1. Background (Van Elst and Van Langenhove, 2001)
Crushing and extraction plants for vegetable oils often cause considerable emis-
sions of odor, which may cause offense in nearby residential areas. The type of
oil seed processed partially determines the amount and type of odor released.
Only limited information is found in literature on the composition of these
waste gases, on source strength classification, and on possible treatment meth-
ods. Lacoste et al. (1996) performed a quantitative study of odorous compounds
in gas effluents from three rapeseed crushing plants. Olfactometry was used to
determine odor concentrations of gaseous effluents. Chemical analyses revealed
the presence of nitriles, aldehydes, and sulfur compounds like mercaptans in
conditioning and pressing emissions, while hydrogen sulfide and acetaldehyde
were the major odorant compounds in the absorption unit effluents.
An oil crushing and extraction plant, situated in an industrial area in the
northern part of France, mainly processes soybean, sunflower, and rapeseed to
produce vegetable oils. Significantly higher odor emissions occurred when pro-
cessing rapeseed (colza) compared to those associated with other types of oil
seed. Because of the growing number of complaints arising from the surrounding
residential area at distances of more than one kilometre, the plant management
decided to tackle the odor problem. The process of improving the odor situation
in the vicinity of this plant was a process that took several years and included
various types of measurements and interim evaluations. The measurements were
a combination of chemical analyses, olfactometry, and field panels. Chemical
measurements (gas chromatography, combined with mass spectrometry) were
mainly used to get a better understanding of the composition of the different
waste gas streams on the plant. Different gas streams have been sampled to iden-
tify and quantify the VOCs present. In the interpretation, special emphasis was
put on compounds with a low odor threshold. Olfactometric measurements were
used to determine the total amount of odour present in the waste gas stream.
These data were very useful to make a classification of the different sources in
order to set priorities for abatement but also to calculate the total odor abate-
ment efficiency of treatment systems. Field panel measurements determined the
impact of the total odor emission on the vicinity of the plant.
4.2.2. Identification of the Main Sources
The first step in the processing of the seeds consists of a number of physical
treatments, like cleaning, crushing, heating up to 60°C, pressing, and cooling.
The main odor sources in this treatment are the hot and humid vapors that
arise at the heating stage and during pressing. Also the cooling of the material
before entering the extraction unit, which takes place on the open conveyor belt
between the crushing and the extraction unit, can be considered as an impor-
tant source. The residual oil is then extracted from the flakes with hexane in an
extraction unit. Hexane and oil are separated in a distillation unit. The residual
fraction of the seed is treated in a desolventizer to remove hexane. Before being
vented in the atmosphere, the vapors of the extraction process pass through an
absorption system with mineral oil to recapture hexane. The emissions of the
absorption contain high concentrations of hydrogen sulfide, and thus represent
an important odor source. The extracted residue of the seed is dried and cooled
and sold as livestock feed. Large amounts of fresh air are used in this process
and are emitted to the atmosphere, loaded with odorous components. Olfacto-
metric emission measurements of a selected number of odor sources resulted in
the following emissions (see Table 2).
The vapors of the conditioners were already incinerated in both steam boil-
ers with an odor removal efficiency higher than 90%. The resulting calculated
emissions in European odor units per hour demonstrate the importance of the
emission of the absorption unit on one hand (72%) and of the drying-cooling
unit on the other hand (18%).
Chemical measurements were carried out on the six sampled emission points.
Table 3 gives an overview of the compounds per chemical group. The data in Table 3
are the emitted mass flows per hour for five sampling points (conditioners not
included). The results of the chemical measurements show an important contribu-
tion of mainly organic sulfur compounds and hydrogen sulfide to the total odor
concentration, considering their low odor threshold. Though no data were found
on the odor threshold of the specific nitriles and 4-isothiocyanato-1-butene,
there might be an important influence of these compounds to the global odor
concentration. The high hydrocarbon content in absorption, conveyor belt, and
drying-cooling are mainly caused by the presence of hexane as extraction solvent
(hexane, 2-methylpentane, 3-methylpentane, cyclohexane, methylcyclopentane).
Table 2. Odor Concentrations Measured at Six Points in the Process
Odor Total flow Total odor Percentage of

concentration rate per emission measured
Emission point (ouE m−3) source (m3 h−1) (ouEh−1) emission (%)
Vapors of conditioners 46 × 103 10300 - -

Exhaust steam boiler 0.9 × 103 32200 29 × 106 0.9
Presses 425 × 103 195 83 × 106 2.7
Absorption unit 5564 × 103 400 2225 × 106 72.1
Conveyor belt from
extraction to drying unit 515 × 103 350 180.106 5.8
Drying/cooling unit 8.8 × 103 64700 570 × 106 18.5
150 B. Sercu et al.
Table 3. Mass Flows in g h−1 for the Different Emission Points
Boiler Drying-
exhaust Presses Absorption Conveyor belt cooling
Hydrocarbons 0.27 3.1 470 395 3940

Aldehydes 1.2 12 25 7.1 60
Ketones 0.35 1.5 1.2 1 16
Alcohols - 7.7 - 0.05 3
Nitriles 0.35 17 - 151 3600
Organic sulfur compounds - 6.5 35 0.4 15
4-isothio-cyanato-1-butene - 0.7 - 0.2 7
Hydrogen sulfide - 0.89 1822 - -
4.2.3. Abatement Techniques
As a first abatement step, some high concentrated streams limited in volumet-

ric flow were chosen to be incinerated in the existing steam boilers (i.e., waste
gases coming from the absorption unit and presses). As a second step, a suitable
technique was chosen for the flow coming from the drying-cooling unit. This
waste gas stream, relatively low in concentration but high in volumetric flow, was
decided to be treated in a biofilter system after doing some pilot tests. A biofilter
combined with a scrubber, designed for a flow of 100.000 m3 h−1, was constructed
by Monsanto EnviroChem systems and operation started in April 1998.
A collecting chamber was installed to receive all waste gas flows. The pur-
pose of this chamber was to create a velocity drop and separate residual dust.
The scrubber had three main purposes: capture of small dust particles, humidi-
fication of the airstream up to 100%, and cooling up to 37 °C. The biofilter itself
is a closed, top-down model. The biofilter material consists of small polystyrene
balls surrounded with compost. Inside of the filter are two stages, each subdi-
vided in different compartments, with load measuring cells. These cells measure
the weight of the compartment, and depending on the weight, additional water
can be sprinkled if dehydration is stated. Figure 11 shows the scrubber, the front
side of the biofilter, and the extraction fan and the silencer, both positioned
behind the biofilter.
Triplicate olfactometric control measurements (June and July 1998) of the
ingoing and outgoing odor concentrations, as well as the chemical composition
of both flows, confirmed the good odor removal efficiency of the complete sys-
tem, with a low residual odor concentration (see Table 4). The chemical compo-
sition of the waste gas at the outlet of the biofilter only revealed hydrocarbons
(hexane, etc.) above the detection limit. Afterward (August 1998), some smaller
but concentrated waste gas streams were added to the collecting chamber, which
resulted in a complete solution of the odor problem after three years of analyses
and investments. Table 4 shows that though the inlet concentration increased over
the different measurements, the outlet concentrations were relatively constant.
Figure 11. Scrubber and biofilter (left); fan and silencer (right).
Table 4. Average Odor Concentrations Measured at the Biofilter
Concentration Concentration at Removal

Date at inlet scrubber (ouE m-3) outlet biofilter (ouE m-3) efficiency (%)
June 1998 3 935 622 84.2

July 1998 11 787 1 670 85.8
July 1998 33 914 1 964 94.2
Sept 1998 165 537 1 277 99.2
A value between 1000 and 2000 (ouE m-3) can be considered as a normal back-
ground value for the typical “own smell” of a good working biofilter. In some
cases, still lower values are possible (up to 500 ouE m-3).
4.3. Odor Removal at an Animal Rendering Plant
Rendering is the transformation of animal by-products into stable products

mainly by evaporation of the water and separation of the fat. Fresh animal
by-products start to decompose as soon as the animal has been slaughtered
into mainly volatile substances through anaerobic processes often initiated by
the bacteria of the stomach and intestinal contents. The volatile substances
are set free when the raw material is heated and dried (water evaporation).
They are found in the water vapor and part of them are condensed with the
water to be treated in wastewater treatment plants, whereas others remain in
the gaseous phase (noncondensables) depending on the vapor pressure under
the given condensation conditions (Oberthür and Vossen, 2001). The odorous
substances from rendering originate mainly from the proteins in the animal by-
products through anaerobic decomposition. The main constituents of render-
ing odors are hydrogen sulfide, ammonia, organic sulfides, aldehydes, organic
acids, and other minor compounds, which due to their low odor threshold,
152 B. Sercu et al.
however, might contribute in a characteristic way to the rendering odor (VDI,

1996). Due to the nature of the material processed, animal rendering activities
thus result in the emission of volatiles and disgusting odors, causing nuisance
in the factory’s neighborhood (De Roo and Van Langenhove, 2000). Several
technologies such as thermal or catalytic combustion, stage scrubbers, and bio-
filters may be used for the elimination of volatiles from waste gases.
In the investigated rendering plant, the noncondensable gases are inciner-
ated in the steam boilers, resulting in a highly efficient odor removal. The other
odorous waste gases are treated in different biofilters, each preceeded with a
scrubber using normal water as scrubbing liquid.
The so-called category one material (cadavers, destruction blood, slaughter-
house by-products) is processed in one production line (ca. 300.000 ton year–1).
On this line, two large, conventional biofilter units are in use. Biofilter 1 treats the
waste air coming from the “clean zone.” General building extraction is used to
avoid diffusive emissions, as well as point suction on all process units. Biofilter 1 is
divided in two parts (1A and 1B); both airflows could be monitored separately.
Biofilter 2 treats the waste air coming from the “unclean zone” (building
extraction and point suction on breakers, pasteurization tanks, buffer tanks,
etc.). In a separate production line, animal by-products of poultry are processed
for use in petfood (ca. 100.000 tonnes year–1). Biofilter 3 treats the air coming
from this separate poultry line.
In the period between 1998 and 2002 the in- and effluent gas flows have been
monitored by olfactometry and GC-MS analyses (unpublished reports Project
Research Gent). Table 5 gives an overview of the different chemical compounds
found in the waste gas streams.
Defoer et al. (2002) showed that the presence and concentration of the
organic sulfur compounds is determining for the total odor concentration of
the flow. A direct correlation could be established. For this reason, only the total
Table 5. Overview of Different Chemical Compounds in Rendering Air
Compound class Chemicals identified
Hydrocarbons Pentane, hexane, heptane, octane, nonane, decane, undecane, dodecane,

tridecane, 2-methylpentane, 3-methylpentane, benzene, toluene, ethylben-
zene, o, m, p-xylene, methylethylcyclohexane, methylcyclopentane
Alcohols Ethanol, 3-methylbutanol
S-Compounds Dimethyl sulfide, dimethyl disulfide, dimethyl trisulfide, carbon disulfide
Halogenated VOCs Trifluoromethylbenzene, 1-chlorobutane, tetrachloroethylene,
dichloromethane
Ethers 2-Methyl-1,3-dioxolane
Furanes Furane, 2-methylfurane
Ketones Acetone, 2-butanone
Aldehydes 3-Methylbutanal, 2-methylbutanal, n-hexanal, isobutyraldehyde,
benzaldehyde
Table 6. Overview of the Results
Odour in Odour out VOSCs in VOSCs out

(103 ouE m−3) (μg m−3)
BF 1A 1998 638 14 1815 200

2000 118 1.7 322 12
BF1B 1998 267 5 130 200
2000 48 3.8 100 10
BF2 2000 172 48 569 438
2002 163 4 400 7
BF3 9/1999 1339 297 4599 3277
12/1999 128 3.5 193 15
2000 240 84 931 531
2002 19 6 Not measured
odor concentration (expressed in ouE m-3) and the concentration of volatile

organic sulfur compounds (expressed in ouE m-3) are shown in Table 6.
This set of data shows that even with high influent odor concentrations, low
outlet concentrations can be reached. However, it seems to be difficult to reach a
“normal” background value situating between 1000 and 2000 ouE m-3. This can
be caused by two factors:
the influent concentrations are quite variable due to different processes;

peak loads can be negative for the efficiency of the biofilter; and
the presence of the volatile organic sulfur compounds cause the typical
smell of the waste air; as mentioned above VOSCs are less biodegradable,
unless the biofilter is inoculated with sulfur-degrading microorganisms.
Bv Mass loading rate

BPT Biological plate tower
DMS Dimethyl sulphide
DMDS Dimethyl disulphide
EBRT Empty bed residence time
EC Elimination capacity
MM Methyl mercaptan
ppbv Parts per billion volume
ppmv Parts per million volume
TEA Triethyl amine
VOC(s) Volatile organic compound(s)
VOSC(s) Volatile organic sulphur compounds(s)
154 B. Sercu et al.
6. REFERENCES
Busca, G., Pistarino, C. (2003) “Abatement of ammonia and amines from waste gases: a summary.”
16:157.
CEN (2003) EN 13725, Air quality — determination of odour concentration by dynamic olfactom-
etry, European Standard, Brussels, Belgium.
Chang, C.-T., Chen, B.-Y., Shiu, I.-S., Jeng, F.-T. (2004) Biofiltration of trimethylamine-containing
waste gas by entrapped mixed microbial cells, Chemosphere 55:751.
Chang, K., Lu, C. (2003) Biofiltration of isopropyl alcohol in a trickle-bed air biofilter, Biodegrada-
tion 14:9.
Cho, K.-S., Hirai, M., Shoda, M. (1991) Degradation characteristics of hydrogen sulfide, methane-
thiol, dimethyl sulfide and dimethyl disulfide by Thiobacillus thioparus DW44 isolated from peat
biofilter, J. Ferment. Bioeng. 71:384.
Cho, K.-S., Hirai, M., Shoda, M. (1992) Enhanced removability of odorous sulphur-containing
gases by mixed cultures of purified bacteria from peat biofilters, J. Ferment. Bioeng. 73:219.
Cho, K.-S., Ryu, H. W., Lee, N. Y. (2000) Biological deodorization of hydrogen sulfide using porous
lava as a carrier of Thiobacillus thiooxidans, J. Biosci. Bioeng. 90:25.
Chou, M.-S., Shiu, W.-Z. (1997) Bioconversion of methylamine in biofilters, J. Air Waste Manage.
Assoc. 47: 58.
Chua, H., Li, X. Z., Yu, P. F. H., Tam, C. Y., Huang, Y. L., Yang, S. T. (2000) Design and perfor-
mance of a fibrous bed bioreactor for odor treatment, Appl. Biochem. Biotechnol. 84–86:469.
Chung, Y.-C., Huang, C., Liu, C. H., Bai, H. (2001a), Biotreatment of hydrogen sulfide- and ammonia-
containing waste gases by fluidized bed bioreactor, J. Air Waste Manage. Assoc. 51:163.
Chung, Y.-C., Huang, C., Pan, J. R., Tseng, C.-P. (1998) Comparison or autotrophic and mixotrophic
biofilters for H2S removal, J. Environ. Eng. April 1998:362.
Chung, Y.-C., Huang, C. P., Tseng, C.-P. (1996a), Biodegradation of hydrogen sulfide by a laboratory-
scale immobilized Pseudomonas putida CH11 biofilter, Biotechnol. Prog. 12:773.
Chung, Y.-C., Huang, C. P., Tseng, C.-P. (1996b), Operation optimization of Thiobacillus thioparus
CH11 biofilter for hydrogen sulfide removal, J. Biotechnol. 52:31.
Chung, Y.-C., Huang, C. P., Tseng, C.-P. (1997) Removal of hydrogen sulphide by immobilized Thio-
bacillus sp. strain CH11 in a biofilter, J. Chem. Technol. Biotechnol. 69:58.
Chung, Y.-C., Huang, C., Tseng, C.-P. (2001b), Biological elimination of H2S and NH3 from waste
gases by biofilter packed with immobilized heterotrophic bacteria, Chemosphere 43:1043.
Chung, Y.-C., Huang, C. P., Tseng, C.-P., Pan, J. R. (2000) Biotreatment of H2S- and NH3-contain-
ing waste gases by co-immobilized cells biofilter, Chemosphere 41:329.
Chung, Y.-C., Lin, Y.-Y., Tseng, C.-P. (2004) Operational characteristics of effective removal of H2S
and NH3 waste gases by activated carbon biofilter, J. Air Waste Manage. Assoc. 54:450.
Cook, L. L., Gostomski, P. A., Apel, W. A. (1999) Biofiltration of asphalt emissions: full-scale opera-
tion treating off-gases from polymer-modified asphalt production, Environ. Prog. 18 :178.
De Bo, I. (2002) Membrane biofiltration of single-compound waste gas streams, PhD thesis, Ghent
University, Belgium.
De Bo, I., Heyman, J., Vincke, J., Verstraete, W., Van Langenhove, H. (2003) Dimethyl sulfide removal
from synthetic waste gas using a flat poly(dimethylsiloxane)-coated composite membrane biore-
actor, Environ. Sci. Technol. 37:4228.
De Bo, I., Jacobs, P., Van Langenhove, H. (2002) Removal of toluene and trichloroethylene from
waste air in a membrane bioreactor, Environ. Sci. Pollution Res. Special Issue 3:28.
Defoer, N., De Bo, I., Van Langenhove, H., Dewulf, J., Van Elst, T. (2002) Gas chromatography-
mass spectrometry as a tool for estimating odour concentrations of biofilter effluents at aerobic
composting and rendering plants, J. Chromatogr. A 970:259.
Degorce-Dumas, J. R., Kowal, S., Le Cloirec, P. (1997) Microbiological oxidation of hydrogen
sulphide in a biofilter, Can. J. Microbiol. 43:264.
Demeestere, K., Van Langenhove, H., Smet, E. (2002) Regeneration of a compost biofilter degrading
high loads of ammonia by addition of gaseous methanol, J. Air Waste Manage. Assoc. 52:796.
De Roo, K., Van Langenhove, H. (2000) Emissions from animal rendering, in: Proceedings of the 7th
symposium on analytical and environmental problems, Z. Galbacs, ed., Szeged, Hungary, October
2, 2000, p. 38.
Devinny, J. S., Deshusses, M. A., Webster, T. S. (1999) Biofiltration for Air Pollution Control. CRC
Press LLC, Boca Raton.
Devos, M., Patte, F., Rouault, J., Laffort, P., Van Gembert, L. J. (1990) Standardized Human Olfac-
tory Thresholds, Oxford University Press, Oxford, 165 p.
Don, J. A. (1985) The rapid development of biofiltration for the purification of diversified waste gas
streams, VDI Berichte 561:63.
Elías, A., Barona, A., Ríos, F. J., Arreguy, A., Munguira, M., Peñas, J., Sanz, J. L. (2000) Application of
biofiltration to the degradation of hydrogen sulfide in gas effluents, Biodegradation 11:423.
Engesser, K.-H., Plaggemeier, T. (2000) Microbiological aspects of biological waste gas purification,
in: Biotechnology, 2nd edition, Volume 11c. Environmental Processes III, Solid waste and waste gas
treatment, preparation of drinking water, J. Klein, J., and J. Winter, eds., Wiley-VCH, Weinheim,
pp. 275–302.
Gabriel, D., Deshusses, M. A. (2003) Retrofitting existing chemical scrubbers to biotrickling filters
for H2S emission control, Proc. Natl. Acad. Sci. U. S. A. 100:6308.
Goodwin, J. P., Amenta, S. A., Delo, R. C., Del Vecchio, M., Pinnette, J. R., Pytlar, T. S. (2000) Odor
control advances at cocomposting facility, Biocycle January 2000:68.
Hartikainen, T., Ruuskanen, J., Vanhatalo, M., Martikainen, P. J. (1996) Removal of ammonia from
air by a peat biofilter, Environ. Technol. 17:45.
Heller, W., Schwager, H. (1996) Luftreinigung durch Kompost-Biofiltration, Agrarforschung 3:520.
Huang, C. P., Chung, Y. C., Hsu, B. M. (1996) Hydrogen sulfide removal by immobilized autotrophic
and heterotrophic bacteria in the bioreactors, Biotechnology Techniques 10:595.
Ibrahim, M. A., Mizuno, H., Yasuda, Y., Fukunaga, K., Nakao, K. (2001) Removal of mixtures of
acetaldehyde and propionaldehyde from waste gas in packed column with immobilized activated
sludge gel beads, Biochem. Eng. J. 8:9.
Jones, K., Martinez, A., Rizwan, M., Boswell, J. (2003) Evaluation of sulfur toxicity and media capac-
ity for H2S removal in biofilters packed with both natural and commercial media, in: Proceedings
of the 96th annual Air & Waste Management Association Conference, June 22–26, San Diego, CA.
Joshi, J. A., Hogan, J. A., Cowan, R. M., Strom, P. F., Finstein, M. S. (2000) Biological removal of
gaseous ammonia in biofilters: space travel and earth-based applications, J. Air Waste Manage.
Assoc. 50:1647.
Kalingan, A. E., Liao, C. M., Chen, J. W., Chen, S. C. (2004) Microbial degradation of livestock-
generated ammonia using biofilters at typical ambient temperatures, J. Environ. Sci. Health Part
B-Pestic. Contam. Agric. Wastes 39:85.
Kasakura, T., Tatsukawa, K. (1995) On the scent of a good idea for odour removal, Water Quality
International 2:24.
Kennes, C., Veiga, M.C. (2001) Bioreactors for Waste Gas Treatment, Kluwer Academic Publishers,
Dordrecht.
Kiared, K., Wu, G., Beerli, M., Rothenbühler, M., Heitz, M. (1997) Application of biofiltration to
the control of VOC emissions, Environ. Technol. 18:55.
Kim, N. J., Hirai, M., Shoda, M. (1998) Comparison of organic and inorganic carriers in removal of
hydrogen sulfide in biofilters, Environ. Technol. 19:1233.
Kim, H., Xie, Q., Kim, Y. J., Chung, J. S. (2002) Biofiltration of ammonia gas with sponge cubes
coated with mixtures of activated carbon and zeolite, Environ. Technol. 23:839.
Kirchner, K., Hauk, G., Rehm, H. J. (1987) Exhaust gas purification using immobilized monocul-
tures (biocatalysts), Appl. Microbiol. Biotechnol. 26:579.
Kirchner, K., Gossen, C. A., Rehm, H. J. (1991) Purification of exhaust air containing organic pol-
lutants in a trickle-bed bioreactor, Appl. Microbiol. Biotechnol. 35:396.
156 B. Sercu et al.
Kosteltz, A. M., Finkelstein, A., Sears, G. (1996) What are the ‘real opportunities’ in biological gas
cleaning for North America?, in: Proceedings of the 89th Annual Meeting & Exhibition of Air &
Waste Management Association, A&WMA, Pittsburgh, PA; 96-RA87B.02.
Lacoste, F., Bosque, F., Brenne, E. (1996) Monitoring atmospheric emissions from industrial colza
trituration plants, Oleagineux Corps Gras Lipides 3:369.
Liang, Y. K., Quan, X., Chen, J. W., Chung, J. S., Sung, J. Y., Chen, S., Xue, D. M., Zhao, Y. Z.
(2000) Long-term results of ammonia removal and transformation by biofiltration, J. Hazard.
Mater. 80:259.
Luo, J. (2001) A pilot-scale study on biofilters for controlling animal rendering process odours, Wat.
Sci. Technol. 44:277.
Mohseni, M., Allen, D. G. (1999) Transient performance of biofilters treating mixtures of hydro-
philic and hydrophobic volatile organic compounds, J. Air Waste Manage. Assoc. 49:1434.
Mohseni, M., Allen, D. G. (2000) Biofiltration of mixtures of hydrophilic and hydrophobic volatile
organic compounds”, Chem. Eng. Sci. 55:1545.
Oberthür, R., Vossen, F. (2001) Intrinsic odour emission from rendering and the odour history of a ren-
dering plant, in: Proceedings of the 1st IWA International Conference on Odour and VOCs: Mea-
surement, Regulation and Control Techniques, The University of NSW, Sydney, Australia, 25 – 29th
March, pp. 131–138.
Otten, L., Afzal, M. T., Mainville, D. M. (2004) Biofiltration of odours: laboratory studies using
butyric acid, Adv. Environ. Res. 8:397.
Park, S.-J., Hirai, M., Shoda, M. (1993) Treatment of exhaust gases from a night soil treatment plant
by a combined deodorization system of activated carbon fabric reactor and peat biofilter inocu-
lated with Thiobacillus thioparus DW44, J. Ferment. Bioeng. 76:423.
Park, S.-J., Nam, S.-I., Choi, E.-S. (2001) Removal of odor emitted from composting facilities using
a porous ceramic biofilter, Wat. Sci. Technol. 44:301.
Park, D. H., Cha, J. M., Ryu, H. W., Lee, G. W., Yu, E. Y., Rhee, J. I., Park, J. J., Kim, S. W., Lee, I.
W., Joe, Y. I., Ryu, Y. W., Hur, B. K., Park, J. K., Park, K. (2002) Hydrogen sulfide removal utiliz-
ing immobilized Thiobacillus sp IW with Ca-alginate bead, Biochem. Eng. J. 11:167.
Peixoto, J., Mota, M. (1997) Volatile organic compounds evaporation chamber for the simulation of
gas effluents in laboratory research, Biotechnology Techniques 11:1.
Peixoto, J., Mota, M. (1998) Biodegradation of toluene in a trickling filter, Bioproc. Eng. 19:393.
Perry, R. H., Green, D. (1984) Perry’s Chemical Engineers’ Handbook, 6th ed., McGraw-Hill Interna-
tional Editions. Chemical Engineering Series, New York, 2234 p.
Rafson, H. J. (1998) Odor and VOC Control Handbook . McGraw-Hill, New York.
Regan, J. M., Harrington, G. W., Noguera, D. R. (2002) Ammonia- and nitrite-oxidizing bacterial
communities in a pilot-scale chloraminated drinking water distribution system, Appl. Environ.
Microbiol. 68:73.
Reij, M. W., Keurentjes, J. T. F., Hartmans, S. (1998) Membrane bioreactors for waste gas treatment,
J. Biotechnol. 59:155.
Ritchie, B. J., Hill, G. A. (1995) Biodegradation of phenol polluted air using an external loop airlift
bioreactor, J. Chem. Technol. Biotechnol. 62:339.
Rittmann, B. E., Schwarz, A. O., Sáez P. B. (2000) Biofilms applied to hazardous waste treatment, in:
Biofilms II, J. Bryers, ed., John Wiley & Sons, Inc., New York, pp. 207–234.
Ruokojärvi, A., Ruuskanen, J., Martikainen, P. J., Olkkonen, M. (2001) Oxidation of gas mixtures
containing dimethyl sulfide, hydrogen sulfide, and methanethiol using a two-stage biotrickling
filter, J. Air Waste Manage. Assoc. 51:11.
Schramm, A., De Beer, D., Gieseke, A., Amann, R. (2000) Microenvironments and distribution of
nitrifying bacteria in a membrane-bound biofilm, Environ. Microbiol. 2:680.
Sercu, B., Demeestere, K., Baillieul, H., Verstraete, W., Van Langenhove, H. (2005a), Degradation of
isobutanal at high loading rates in a compost biofilter, J. Air Waste Manage. Assoc. 55:1217.
Sercu, B., Núñez, D., Van Langenhove, H., Aroca, G., Verstraete, W. (2005b), Operational and
microbiological aspects of a two-stage biotrickling filter removing hydrogen sulfide and dimethyl
sulfide, Biotechnol. Bioeng. 90:259.
Sheridan, B. A., Curran, T. P., Dodd, V. A. (2003) Biofiltration of n-butyric acid for the control of
odour, Bioresource Technol. 89:199.
Shoda, M. (1991) Methods for the biological treatment of exhaust gases, in: Biological Degradation
of Wastes, A. M. Martin, ed., Elsevier Science Publishers, New York, pp. 31–46.
SIS, (2000) Scientific Instruments Services: adsorbent resins, general listing and description of seven
resins; http://www.sisweb.com/index/referenc/resin2.htm
Smet, E. (1995) Evaluatie werking biofilter – De Kleijn (Maaseik), Unpublished Report, Ghent University.
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 of methyl sulphides, Appl. Microbiol.
Biotechnol. 45:293.
Smet, E., Lens, P., and Van Langenhove, H. (1998) Treatment of waste gases contaminated with
odorous sulphur compounds, Crit. Rev. Environ. Sci. Technol. 28:89.
Smet, E., Van Langenhove, H., and Verstraete, W. (1996b) Long-term stability of a biofilter treating
dimethyl sulphide, Appl. Microbiol. Biotechnol. 46:191.
Smet, E., Van Langenhove, H., Maes, K. (2000) Abatement of high concentrated ammonia loaded
waste gases in compost biofilters, Water Air Soil Poll. 119:177.
Smet, E., Van Langenhove, H., and Verstraete, W. (1997) Isobutyraldehyde as a competitor of the
dimethyl sulfide degrading activity in biofilters, Biodegradation 8:53.
Smits, M. C. J., Hoek, A. P., Osinga, B., Ottengraf, S. P. P., and Wijngaard, M. H. (1995) Removal of
ammonia with odour from composting facility waste gas with a bio-trickling filter, Meded. Fac.
Landbouww. Gent 60:2313.
Tang, H.-M., Hwang, S.-J., and Hwang, S.-C. (1996) Waste gas treatment in biofilters, J. Air Waste
Manage. Assoc. 46:349.
Terasawa, M., Hirai, M., and Kubota, H. (1986) Soil deodorization systems, Biocycle 27:28.
Van Elst, T., and Van Langenhove, H. (2001) Case study: odour emission at a vegetable oil plant in
France, in: Proceedings of the 1st IWA International Conference on Odour and VOCs: Measure-
ment, Regulation and Control Techniques, The University of NSW, Sydney, Australia, 25–29th
March, pp. 91–398.
Van Groenestijn, J. W., and Hesselink, P. G. M. (1993) Biotechniques for air pollution control,
Biodegradation 4:283.
Van Langenhove, H., Bendinger, B., Oberthür, R., and Schamp, W. (1992) Organic sulfur com-
pounds: persistant odourants in the biological treatment of complex waste gases, in: Biotech-
niques for Air Pollution Abatement and Odour Control Policies, A. J. Dragt, and J. van Dam, Eds.,
Elsevier, Amsterdam, pp. 309–313.
Van Langenhove, H., Cornelis, C., Schamp, N., and Wuyts, E. (1989a) Biofiltration in a vegetables
processing industry, Meded. Fac. Landbouww. Gent 54:1471.
Van Langenhove, H., Lootens, A., and Schamp, N. (1988) Elimination of ammonia from pigsty
ventilation air by wood bark biofiltration, Meded. Fac. Landbouww. Gent 53:1963.
Van Langenhove, H., Lootens, A., and Schamp, N. (1989b) Inhibitory effects of SO2 on biofiltration
of aldehydes, Water Air Soil Poll. 47:81.
Van Lith, C., Leson, G., and Michelsen, R. (1997) Evaluating design options for biofilters, J. Air
Waste Manage. Assoc. 47:37.
VDI (1996) Emissionsminderung, Anlagen zur Verwertung und Beseitigung von Tierkörperteilen und
tierischen Erzeugnissen, Düsseldorf, Verein Deutscher Ingenieure
Wada, A., Shoda, M., Kubota, H., Kobayashi, T., Katayama-Fujimura, J., and Kuraishi, H. (1986)
Characteristics of H2S oxidizing bacteria inhabiting a peat biofilter, J. Ferment. Technol. 64: 161.
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.
Warren, J., Loerh, C., and Fellow A. S. C. E. (1997) Biofiltration: fundamentals, design and opera-
tion principles and applications, J. Environ. Eng. 123:538.
Waweru, M., Herrygers, V., Van Langenhove, W., and Verstraete, W. (2000) Process engineering of
biological waste gas purification, in: Biotechnology, 2nd edition, Volume 11c. Environmental Pro-
158 B. Sercu et al.
cesses III, Solid Waste and Waste Gas Treatment, Preparation of Drinking Water, J. Klein, and
J. Winter, eds., Wiley-VCH, Weinheim, pp. 260–273.
Weast, R. C., Astle, M. J., and Beyer, W. H. (1984) CRC. Handbook of Chemistry and Physics, 64th
ed., CRC Press, Inc., Boca Raton, Florida, 2303 p.
Weckhuysen, B., Vriens, L., and Verachtert, H. (1993) The effect of nutrient supplementation on the
biofiltration removal of butanal in contaminated air, Appl. Microbiol. Biotechnol. 39:395.
Weckhuysen, B., Vriens, L., and Verachtert, H. (1994) Biotreatment of ammonia-containing and
butanal-containing waste gases, Appl. Microbiol. Biotechnol. 42:147.
Williams, T. O., 1995, Odors and VOC emissions control methods, Biocycle 36:49.
Wu, L., Loo, Y.-Y., and Koe, L. C. C. (2001) A pilot study of a biotricking filter for the treatment of
odorous sewage air, Wat. Sci. Technol. 44:295.
Yang, Y., and Allen, E. R. (1994a) Biofiltration control of hydrogen sulfide 1. Design and opera-
tional parameters, J. Air Waste Manage. Assoc. 44:863.
Yang, Y., and Allen, E. R. (1994b) Biofiltration control of hydrogen sulfide 2. Kinetics, biofilter
performance, and maintenance, J. Air Waste Manage. Assoc. 44:1315.
Yang, Y., Togna, A. P., and Skladany, G. (1994) Treatment of CS2 and H2S vapors by biofiltration, in:
Proceedings of the 49th Purdue Industrial Waste Conference, Purdue University, pp. 449–456.
Yani, M., Hiral, M., and Shoda, M. (1998) Ammonia gas removal characteristics using biofilter with
activated carbon fiber as a carrier, Environ. Technol. 19:709.
Zhang, L., Hirai, M., and Shoda, M. (1991) Removal characteristics of dimethyl sulfide, methane-
thiol and hydrogen sulfide by Hyphomicrobium sp. I55 isolated from peat biofilter, J. Ferment.
Bioeng. 72:392.
Index
Page numbers followed by f and t indicate figures and tables, respectively.
Absorption, 30, 41, 43, 81, 88, 89t, 108t, Biofiltration, 108t, 127, 137–142
109, 110t, 111f, 117, 127, 130–131, Biomass, 80t, 84, 130–131, 133f, 134–135,
138–139, 148–150 139–140
Activated carbon, 34, 42, 111–113, 114t, Bioscrubber, 130–131, 139
127, 138, 142 Biotrickling filter, 129–130
Adaptation, 2, 33 Biowasher, 127, 130–132
Adsorbent, 42–43, 112–113, 114f, Biological plate tower (BPT), 127, 131,
146–147 133–135
Adsorbing, 41, 113 Brewery, 92, 116
Adsorption, 28, 32–33, 42–43, 81, 89, 108, Buffer areas, 96–100
111–113, 114f, 146–147 Building Sources, 20–21
Advertisements, 84 Bulk acoustic wave, 50–51
Air pollution, 47–54, 102, 106, 109, 112, 132t
Alcohols, 11, 58, 59t, 60, 64–66, 70, 80, Cane sugar processing, 9t, 11
112, 126, 144–145, 150t, 152t Capelin, 62, 63t, 66
Aldehydes, 11, 23t, 59t, 60, 64–66, 80t, Capture-recovery, 106
81, 107, 108t, 112, 121, 126, 144–146, Case study, 64–70
148, 150t, 151, 152t Catalytic metal, 50
Amines, 6, 10–11, 23t, 59t, 62, 64–66, 68, Catalytic oxidation, 119–121
70, 107, 108t, 109, 112, 126, 140 Catalytic oxidizer, 119–121
Ammonia, 6–8, 10, 18, 23t, 25, 29, 62, Catalytic combustion, 152
65–66, 79–80, 97, 109, 127, 137–140, Chemical analysis, 24, 68
142–143 Chemical compounds, 49, 88, 152
Ammonia-like stockfish odor, 64f Chemical gas sensors, 49, 64–65
Animal feeding, 7, 9–10 Cleaner production, 84, 92f, 93–95, 97, 106
Animal rendering, 10, 151–153 Clogging, 42, 128–130, 132–133, 136–137
Area sources, 18, 20 Cluster analysis, 51–52
Artificial neural network, 51–54 Cod, 59t, 62, 63t, 66
Cod roe, 63t
Biodegradation, 54, 80t, 95, 125, 145 Colorimetric Detector Tubes, 27–28
Biofilm, 126f, 136f–137f Condensation, 18, 88–89, 108t, 115–116,
Biofilter, 128f, 129–130, 137f 147, 151
159
160 Index
Conducting polymer sensors (CP), 51 Herring, 63t

Conducting polymers, 51 High-performance liquid chromatography
Cryotrapping, 41 (HPLC), 43
Human health, 3
Deodorization, 127 Human olfactory system, 2, 21, 58
Destructive methods, 107–108 Human response, 1–3, 24
Development plans, 96–100 Hydrogen sulfide, 64f
Dilution to threshold (D/T) ratios, 22, 34
Direct scaling technique, 31 Incineration, 108t, 117–120
Dispersion, 91, 97, 101–102, 107 Indoor air quality, 81–82
Dynamic olfactometry, 32–33, 36, Indoors, 3, 81
146–147 Industry activities, 8, 9t, 99
Isolation chamber, 18–19
Electrochemical gas sensors, 64–65
Electronic nose, 49f, 50–51, 57–70 Ketones, 11, 23t, 64, 80t, 109, 150t, 152t
Emission rates, 18–21
Environmental control, 53, 77 Linear discriminant analysis, 51–52
Environmental management systems
(EMS), 84 Manure processing, 9–10
Esters, 58f, 60, 64–66, 95 Mass spectrometry, 26, 147–148
European reference odor mass (EROM), Meat rendering plants, 9t, 10–11
22–23 Membrane bioreactor, 127, 136–137
Externalities, 77, 92f, 97 Metal oxide semiconductor field effect
transistors, 50
F.I.D.O.L., 76 Metal oxide semiconductors, 50
Fatigue, 2, 33, 35, 54 Metaloxide sensors (MOS), 50–51, 63t
FIFO, 86 Microbial counts, 61, 63t, 70
Fingerprints, 53 Micro-organisms, 126, 129–132, 136, 138,
Fish, 9t, 57–70 141, 143–144, 153
Fish Canning, 9, 11 Multivariate analysis, 49, 51
Fish fillets, 66–70
Flame ionization detector (FID), Nitrification, 135f, 138, 139–140
29, 30f Non destructive methods, 107
Food industry, 1–12, 47–54, 75–102 Nuisance, 1, 3–6, 10–12
Fourier transform infrared spectrometry
(FTIR), 30 Objective measurements, 31–32
Fresh fish odor, 59t Odor character, 22, 26, 31, 98
Freshness, 53, 57, 60–62, 63t, 64, Odor concentration, 2, 16, 20–24, 33–35, 91,
66–67, 70 99, 100, 143, 146–150, 151t, 152–153
Front-end control, 84 Odor control, 12, 15, 54, 79t, 88, 90, 98,
Frying, 79t, 87–88, 95 106–107, 109, 111, 113
Odor guidelines, 3–6
Gas chromatography (GC), 25–27, 41–44, Odor intensity, 2, 12, 21–22, 31, 33
60, 65, 148 Odor mixtures, 145–146
Gas chromatography-mass spectrometry Odor removal, 26, 115, 135, 145–146, 148–153
(GC-MS), 65–66, 146–147 Odor sampling, 16, 17f, 36, 101
Gas chromatography- olfactometry (GC- Odor sources, 16, 78, 148–149
O), 26–27 Odor unit, 4, 22–24, 100, 148–149
Index 161
Off-odor, 47, 75, 80, 106 Retrofitting, 92f

Olfactometer, 20–21, 32–33, 34f, 35–36 Reuse (of solvent), 113
Olfactometry, 20, 22, 26–27, 32–33, 35–36, RTO, 118–120
60, 146–148, 152
Olfactometry Panel, 33–34 Salmon, 60, 63t
Organic sulfur compounds, 7, 126, Scentometer, 34–35
142–144, 149, 150t, 152–153 Scrubber, 90, 108t, 109, 110t, 111,
Organoleptic, 1, 21 130–131, 139, 142, 150, 151f, 152
Outdoors, 81–83 Seasonal variation, 67
Oxidation, 58, 60, 62, 63t, 108–109, Sensing system, 49
117–121, 131, 141, 144 Sensor array, 47–54
Sensory analysis, 62, 63t, 67–68
Paper tape monitors, 28 Sensory methods, 24, 31–37
Partial least square, 51–52 Shifting, 84–85
Pattern recognition, 48–49, 52 Shrimp, 63t
pH sensitive films, 62–64 Solid-phase microextraction (SPME),
Phenolic compounds, 8 43–44
Photoionization detector (PID), 25, 29–30 Sorbent bed, 42
Point sources, 17–18, 20 Sorbent sampling, 43
Portable analyzers, 24, 28–30 Sour and putrid odors, 23t, 58f, 59t,
Portable gold leaf analyzers, 28 64f, 66
Preconcentration, 41–44, 65, 146–147 Specific odor emission rate, 20–21
Pre-concentration Spoilage, 57, 59t, 60–62, 64–70
Prevention, 75–102 Storage studies, 67–68, 69f
Preventive odor emission control tech- Subjective measurements, 31
nologies, 83–102 Sulfur compounds, 7, 58f, 64f–65f, 66, 68,
Preventive managing procedures, 95–96 70, 126, 137, 142–144, 149
Primary pollutants, 10, 78 Sulphur oxidation, 141, 144
Principal component analysis (PCA), Surface acoustic wave, 51
51–52, 69f Sweet- and fruity odors, 66
Properties of odors, 31, 78
Tainting, 81
Quality, 57–70, 81–83 Texture measurements, 67
Quality indicators, 64f Threshold odor numbers (Ton), 22
Quartz microbalance, 51, 63t TMA (trimethylamine), 11, 23t, 59t, 60,
Quartz microbalance sensors (QMB), 51 62, 63t, 64f, 65–66, 69f
Total volatile basic nitogen (TVB-N), 63t,
Rapid evaluation of fish quality, 61 64, 67, 69f, 70
Recuperative thermal oxidation systems, Trapping techniques, 42
118, 119f Trout, 63t
Recycling, 84, 90, 92f, 95, 107, 115 Tuna, 60t, 63t
Reformulation, 84–85, 94–95
Regeneration, 112–113, 114t, 131, 141 Vegetable oil extraction, 148–151
Regenerative thermal oxidizer (RTO), Ventilation, 20–21, 81–83, 86, 88, 90–91,
118–120 95, 132
Regulations, 3–6, 61, 77, 102, 107 Volatile compounds, 27, 41–43, 48–49, 51,
Rendering, 10–11, 80, 90, 111, 140, 143, 53, 57–62, 64–66, 81, 111
151–153 Volatile fatty acids (VFA), 7–8, 126, 145
162 Index
Volatile organic compound (VOC), Volatiles, 27, 50, 54, 60, 62, 65, 146, 152
7–8, 10–12, 29, 43–44, 87, 107, 108t,
116–117, 119–121, 125, 135 Waste gas, 125–127, 131, 136–141, 143,
Volatile organic sulphur compound 145, 147–148, 150, 152
(VOSC), 143–144, 153 Wind tunnel system, 18–20

Elefteria Psillakis, Vassilis Gekas (Auth.), Xavier Nicolay (Eds.) - Odors in The Food Industry-Springer US (2006)

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Elefteria Psillakis, Vassilis Gekas (Auth.), Xavier Nicolay (Eds.) - Odors in The Food Industry-Springer US (2006)

Загружено:

Авторское право:

Доступные форматы

Integrating Safety and Environment Knowledge

Into Food Studies

Volume 1 FOOD SAFETY: A Practical and Case Study Approach

Volume 2 ODORS IN THE FOOD INDUSTRY

Volume 3 UTILIZATION OF BY-PRODUCTS AND TREATMENT OF

Volume 4 PREDICTIVE MODELING AND RISK ASSESSMENT

Volume 5 EXPERIMENTS IN UNIT OPERATIONS AND PROCESSING

Volume 6 CASE STUDIES IN FOOD SAFETY AND ENVIRONMENTAL

Library of Congress Control Number: 2006926455

©2006 Springer Science+Business Media, LLC

March 2006 Kristberg Kristbergsson

Olfactory nuisances, to the same extent as noise pollution, appear to be a modern

January 2006 Xavier Nicolay

1. ODOR PROBLEMS IN THE FOOD INDUSTRY . . . . . . . . . . . . . . . . . . . 1

3. PRECONCENTRATION PRIOR TO GAS CHROMATOGRAPHY . . . . 41

4. THE APPLICATION OF INTELLIGENT SENSOR ARRAY FOR AIR

5. ELECTRONIC-NOSE TECHNOLOGY: APPLICATION FOR

6. ODORS PREVENTION IN THE FOOD INDUSTRY . . . . . . . . . . . . . . . . 75

7. ODORS TREATMENT: PHYSICOCHEMICAL TECHNOLOGIES. . . . 105

8. ODORS TREATMENT: BIOLOGICAL TECHNOLOGIES . . . . . . . . . . . 125

Maria Stella Cosio

Toon van Elst

Herman Van Langenhove

With growing population, industrialization, and urbanization, the odor problem

2. HUMAN RESPONSE TO ODORS

ELEFTERIA PSILLAKIS AND VASSILIS GEKAS Department of Environmental Engi-

model (Frechen, 1994) describing human odor perception it is the physiological

3. ODORS AND HUMAN HEALTH

An important area of research is whether odors are simply a nuisance or a legiti-

4. ODOR GUIDELINES AND REGULATIONS

4.2. The Netherlands

4.3. United Kingdom

4.4. United States of America

Table 1. Examples of Ambient Standards for Odor-Causing Compounds [all agencies

Location Compound Ambient Odor Standard

California Hydrogen sulﬁde 30 ppbva (1-hr average)

4.5. World Health Organization Guideline Values

For substances with malodorous properties at concentrations below those where

The microbial decomposition of organic nitrogen compounds under both aerobic

5.2. Sulfur Compounds

5.3. Volatile Organic Compounds

During the degradation of organic matter, volatile organic compounds (VOC)

5.4. Phenolic Compounds

6. FOOD INDUSTRY COMPONENTS

Grower/ Processing Wholesale Service/ Consumption

Table 2. Selected Food Industry Activities

Industry Activities Description

Livestock & poultry feed operations • Cattle feedlots

7. SELECTED CASE STUDIES OF ODOR PROBLEMS IN

7.1. Livestock Enterprises (including animal feeding operations and manure

Odors and gases emitted are a by-product of the microbial decomposition

7.2. Meat Processing

7.3. Meat-Rendering Plants

Meat-rendering plants process animal by-product materials for the production

7.4. Fish Canning

7.5. Cane Sugar Processing

One of the main environmental problems associated with sugar processing is

9. ACRONYMS AND ABBREVIATIONS

mve Mestvarkeneenheden or fattener units

It generally is recognized that for effective odor control measures to be implemented

ELEFTERIA PSILLAKIS Department of Environmental Engineering, Technical University

The sampling of an environmental parameter, such as odor, is an extremely com-

2.1. General Considerations for Sampling Odorous Compound