Emerging Contaminants from Industrial and Municipal Waste: Removal technologies (The Handbook of Environmental Chemistry)

The Handbook of Environmental Chemistry Editors-in-Chief: O. Hutzinger · D. Barceló · A. Kostianoy Volume 5 Water Pollu...

Author: Damia Barceló | Mira Petrovic

73 downloads 1166 Views 4MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

The Handbook of Environmental Chemistry Editors-in-Chief: O. Hutzinger · D. Barceló · A. Kostianoy Volume 5 Water Pollution Part S/2

Advisory Board: D. Barceló · P. Fabian · H. Fiedler · H. Frank · J. P. Giesy · R. A. Hites M. A. K. Khalil · D. Mackay · A. H. Neilson · J. Paasivirta · H. Parlar S. H. Safe · P. J. Wangersky

The Handbook of Environmental Chemistry Recently Published and Forthcoming Volumes

Environmental Specimen Banking Volume Editors: S. A. Wise and P. P. R. Becker Vol. 3/S, 2009

Estuaries Volume Editor: P. J. Wangersky Vol. 5/H, 2006

Polymers: Chances and Risks Volume Editors: P. Eyerer, M. Weller and C. Hübner Vol. 3/V, 2009

The Caspian Sea Environment Volume Editors: A. Kostianoy and A. Kosarev Vol. 5/P, 2005

The Black Sea Environment Volume Editors: A. Kostianoy and A. Kosarev Vol. 5/Q, 2008 Emerging Contaminants from Industrial and Municipal Waste Removal Technologies Volume Editors: D. Barceló and M. Petrovic Vol. 5/S/2, 2008 Emerging Contaminants from Industrial and Municipal Waste Occurrence, Analysis and Effects Volume Editors: D. Barceló and M. Petrovic Vol. 5/S/1, 2008 Fuel Oxygenates Volume Editor: D. Barceló Vol. 5/R, 2007 The Rhine Volume Editor: T. P. Knepper Vol. 5/L, 2006

Marine Organic Matter: Biomarkers, Isotopes and DNA Volume Editor: J. K. Volkman Vol. 2/N, 2005 Environmental Photochemistry Part II Volume Editors: P. Boule, D. Bahnemann and P. Robertson Vol. 2/M, 2005 Air Quality in Airplane Cabins and Similar Enclosed Spaces Volume Editor: M. B. Hocking Vol. 4/H, 2005 Environmental Effects of Marine Finﬁsh Aquaculture Volume Editor: B. T. Hargrave Vol. 5/M, 2005 The Mediterranean Sea Volume Editor: A. Saliot Vol. 5/K, 2005

Persistent Organic Pollutants in the Great Lakes Volume Editor: R. A. Hites Vol. 5/N, 2006

Environmental Impact Assessment of Recycled Wastes on Surface and Ground Waters Engineering Modeling and Sustainability Volume Editor: T. A. Kassim Vol. 5/F (3 Vols.), 2005

Antifouling Paint Biocides Volume Editor: I. Konstantinou Vol. 5/O, 2006

Oxidants and Antioxidant Defense Systems Volume Editor: T. Grune Vol. 2/O, 2005

Emerging Contaminants from Industrial and Municipal Waste Removal Technologies Volume Editors: Damià Barceló · Mira Petrovic

With contributions by A. Agüera · D. Barceló · J. M. Bayona · G. Buttiglieri · M. S. Díaz-Cruz A. R. Fernández-Alba · E. García-Calvo · M. D. Hernando P. Jovanˇci´c · T. P. Knepper · K. Koˇsuti´c · B. Kunst P. Letón · S. Malato · V. Matamoros · M. Matoˇsi´c · M. Mezcua I. Mijatovi´c · J. A. Perdigón-Melón · M. Petrovi´c · M. Radeti´c J. Radjenovi´c · A. Rodríguez · R. Rosal

123

Environmental chemistry is a rather young and interdisciplinary ﬁeld of science. Its aim is a complete description of the environment and of transformations occurring on a local or global scale. Environmental chemistry also gives an account of the impact of man’s activities on the natural environment by describing observed changes. The Handbook of Environmental Chemistry provides the compilation of today’s knowledge. Contributions are written by leading experts with practical experience in their ﬁelds. The Handbook will grow with the increase in our scientiﬁc understanding and should provide a valuable source not only for scientists, but also for environmental managers and decision-makers. The Handbook of Environmental Chemistry is published in a series of ﬁve volumes: Volume 1: The Natural Environment and the Biogeochemical Cycles Volume 2: Reactions and Processes Volume 3: Anthropogenic Compounds Volume 4: Air Pollution Volume 5: Water Pollution The series Volume 1 The Natural Environment and the Biogeochemical Cycles describes the natural environment and gives an account of the global cycles for elements and classes of natural compounds. The series Volume 2 Reactions and Processes is an account of physical transport, and chemical and biological transformations of chemicals in the environment. The series Volume 3 Anthropogenic Compounds describes synthetic compounds, and compound classes as well as elements and naturally occurring chemical entities which are mobilized by man’s activities. The series Volume 4 Air Pollution and Volume 5 Water Pollution deal with the description of civilization’s effects on the atmosphere and hydrosphere. Within the individual series articles do not appear in a predetermined sequence. Instead, we invite contributors as our knowledge matures enough to warrant a handbook article. Suggestions for new topics from the scientiﬁc community to members of the Advisory Board or to the Publisher are very welcome.

ISBN 978-3-540-79209-3 DOI 10.1007/978-3-540-79210-9

e-ISBN 978-3-540-79210-9

The Handbook of Environmental Chemistry ISSN 1433-6863 Library of Congress Control Number: 2008937808 c 2008 Springer-Verlag Berlin Heidelberg This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, speciﬁcally the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microﬁlm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable to prosecution under the German Copyright Law. The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a speciﬁc statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Cover design: WMXDesign GmbH, Heidelberg Typesetting and Production: le-tex publishing services oHG, Leipzig Printed on acid-free paper 9876543210 springer.com

Editors-in-Chief Prof. em. Dr. Otto Hutzinger

Prof. Andrey Kostianoy

Universität Bayreuth c/o Bad Ischl Ofﬁce Grenzweg 22 5351 Aigen-Vogelhub, Austria [email protected]

P.P. Shirshov Institute of Oceanology Russian Academy of Sciences 36, Nakhimovsky Pr. 117997 Moscow, Russia [email protected]

Prof. Dr. Damià Barceló Dept. of Environmental Chemistry IIQAB – CSIC Jordi Girona, 18–26 08034 Barcelona, Spain [email protected]

Volume Editors Prof. Dr. Damià Barceló

Mira Petrovic

Dept. of Environmental Chemistry IIQAB – CSIC Jordi Girona, 18–26 08034 Barcelona, Spain [email protected]

Dept. of Environmental Chemistry IIQAB – CSIC Jordi Girona, 18–26 08034 Barcelona, Spain [email protected]

Advisory Board Prof. Dr. D. Barceló

Prof. Dr. H. Frank

Dept. of Environmental Chemistry IIQAB – CSIC Jordi Girona, 18–26 08034 Barcelona, Spain [email protected]

Lehrstuhl für Umwelttechnik und Ökotoxikologie Universität Bayreuth Postfach 10 12 51 95440 Bayreuth, Germany

Prof. Dr. P. Fabian

Prof. Dr. J. P. Giesy

Lehrstuhl für Bioklimatologie und Immissionsforschung der Universität München Hohenbachernstraße 22 85354 Freising-Weihenstephan, Germany

Department of Zoology Michigan State University East Lansing, MI 48824-1115, USA [email protected]

Dr. H. Fiedler Scientiﬁc Affairs Ofﬁce UNEP Chemicals 11–13, chemin des Anémones 1219 Châteleine (GE), Switzerland hﬁ[email protected]

Prof. Dr. R. A. Hites Indiana University School of Public and Environmental Affairs Bloomington, IN 47405, USA [email protected]

VI

Prof. Dr. M. A. K. Khalil

Prof. Dr. Dr. H. Parlar

Department of Physics Portland State University Science Building II, Room 410 P.O. Box 751 Portland, OR 97207-0751, USA [email protected]

Institut für Lebensmitteltechnologie und Analytische Chemie Technische Universität München 85350 Freising-Weihenstephan, Germany

Prof. Dr. D. Mackay Department of Chemical Engineering and Applied Chemistry University of Toronto Toronto, ON, M5S 1A4, Canada

Department of Veterinary Physiology and Pharmacology College of Veterinary Medicine Texas A & M University College Station, TX 77843-4466, USA [email protected]

Prof. Dr. A. H. Neilson

Prof. P. J. Wangersky

Swedish Environmental Research Institute P.O. Box 21060 10031 Stockholm, Sweden [email protected]

University of Victoria Centre for Earth and Ocean Research P.O. Box 1700 Victoria, BC, V8W 3P6, Canada wangers@telus. net

Prof. Dr. J. Paasivirta Department of Chemistry University of Jyväskylä Survontie 9 P.O. Box 35 40351 Jyväskylä, Finland

Prof. Dr. S. H. Safe

The Handbook of Environmental Chemistry Also Available Electronically

For all customers who have a standing order to The Handbook of Environmental Chemistry, we offer the electronic version via SpringerLink free of charge. Please contact your librarian who can receive a password or free access to the full articles by registering at: springerlink.com If you do not have a subscription, you can still view the tables of contents of the volumes and the abstract of each article by going to the SpringerLink Homepage, clicking on “Browse by Online Libraries”, then “Chemical Sciences”, and ﬁnally choose The Handbook of Environmental Chemistry. You will ﬁnd information about the – – – –

Editorial Board Aims and Scope Instructions for Authors Sample Contribution

at springer.com using the search function. Color ﬁgures are published in full color within the electronic version on SpringerLink.

Preface

This book on “Emerging Contaminants from Industrial and Municipal Waste” is based on the scientiﬁc developments and results achieved within the European Union (EU)-funded project EMCO (reduction of environmental risks posed by emerging contaminants, through advanced treatment of municipal and industrial wastes). One of the key elements of the EMCO project was to provide support to the various Western Balkans countries involved in the project as regards the implementation of the Water Framework Directive (WFD) (2000/60/EC). A regional network, as proposed by the EMCO project, aiming to ensure the comparability (and reliability) of measurement data obtained by screening methodologies for water quality management, would support the EU Water Initiative, which aims to promote co-operation between countries in order to better manage their water resources. The EMCO project addressed basically two directives: Directive 91/271/EEC to reduce the pollution in Community surface waters caused by municipal waste and the IPPC Directive (Directive 96/61/EC). This Directive expands the range of pollutants that should be monitored in industrial efﬂuent discharges like those from the paper and pulp industry, reﬁneries, textiles and many other sectors. The EMCO project has devoted its attention to the wastewater treatment technologies, especially in the Western Balkan countries. It is obvious that building up and improving wastewater treatment plant performance in the public and private sectors will avoid direct pollution of receiving waters by urban and industrial activities. The book is divided into two volumes: Vol. I—Occurrence, Analysis and Effects, and Vol. II—Removal Technologies. Volume I is structured in several chapters covering advanced chemical analytical methods, the occurrence of emerging contaminants in wastewaters, environmental toxicology and environmental risk assessment. Advanced monitoring analytical methods for emerging contaminants cover the use of liquid chromatography combined with tandem mass spectrometric detection or hybrid mass spectrometric techniques. It is certainly known that without these advanced mass spectrometric tools it would not be possible to investigate the fate and behaviour of emerging pollutants at the wastewater treatment plants and receiving waters at the nanogram per litre level. Ecotoxicology is also a very relevant aspect that should be taken into consideration for emerging

X

Preface

contaminants, and it is also covered in this book. Risk assessment methodologies will allow us to critically establish the good performance of an appropriate wastewater treatment technology for the removal of urban, agricultural and industrial wastewaters. Volume II covers different treatment options for the removal of emerging contaminants and includes membrane bioreactors (MBR), ozonization and photocatalysis, and advanced sorbent materials together with more conventional natural systems, such as artiﬁcial recharge and constructed wetlands. The MBR is an emerging technology based on the use of membranes in combination with traditional biological treatment. It is considered as a promising technology able to achieve more efﬁcient removal of micro-pollutants in comparison to conventional wastewater treatment plants. Other examples reported in the book are advances in nanomaterials, also an emerging ﬁeld in wastewater treatment, which are providing great opportunities in the development of more effective wastewater treatment technologies. Overall, this book is certainly timely since the interest in emerging contaminants and wastewater treatment has been growing considerably during the last few years, related to the availability of novel treatment options together with the advanced and highly sensitive analytical techniques. This book can also be considered, in a way, the follow-up of two previous books in this series entitled Emerging Organic Pollutants in Waste Waters and Sludge, Vols. 1 and 2 (5 1 and 5 0), published in 2004 and 2005. The present book is complementary to these volumes since here much more attention has been devoted to wastewater treatment systems, which are a key part of this book. The book will be of interest to a broad audience of analytical chemists, environmental chemists, water management operators and technologists working in the ﬁeld of wastewater treatment, or newcomers who want to learn more about the topic. Finally, we would like to thank all the contributing authors of this book for their time and effort in preparing this comprehensive compilation of research papers. Barcelona, September 2008

D. Barceló M. Petrovic

Contents

Removal of Emerging Contaminants in Wastewater Treatment: Conventional Activated Sludge Treatment G. Buttiglieri · T. P. Knepper . . . . . . . . . . . . . . . . . . . . . . . .

1

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology J. Radjenovi´c · M. Matoˇsi´c · I. Mijatovi´c · M. Petrovi´c · D. Barceló . . . .

37

Removal of Emerging Contaminants in Water Treatment by Nanoﬁltration and Reverse Osmosis B. Kunst · K. Koˇsuti´c . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 Ozone-Based Technologies in Water and Wastewater Treatment A. Rodríguez · R. Rosal · J. A. Perdigón-Melón · M. Mezcua A. Agüera · M. D. Hernando · P. Letón · A. R. Fernández-Alba E. García-Calvo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 Removal of Emerging Contaminants in Waste-water Treatment: Removal by Photo-catalytic Processes S. Malato . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 Behavior of Emerging Pollutants in Constructed Wetlands V. Matamoros · J. M. Bayona . . . . . . . . . . . . . . . . . . . . . . . . 199 Input of Pharmaceuticals, Pesticides and Industrial Chemicals as a Consequence of Using Conventional and Non-conventional Sources of Water for Artiﬁcial Groundwater Recharge M. S. Díaz-Cruz · D. Barceló . . . . . . . . . . . . . . . . . . . . . . . . 219 Advanced Sorbent Materials for Treatment of Wastewaters P. Jovanˇci´c · M. Radeti´c . . . . . . . . . . . . . . . . . . . . . . . . . . . 239 Conclusions and Future Research Needs D. Barceló · M. Petrovic . . . . . . . . . . . . . . . . . . . . . . . . . . . 265

XII

Contents

Erratum to Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology J. Radjenovi´c · M. Matoˇsi´c · I. Mijatovi´c · M. Petrovi´c · D. Barceló . . . . 275 Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281

Contents of Volume 5, Part S/1 Emerging Contaminants from Industrial and Municipal Waste Occurrence, Analysis and Effects Volume Editors: Barceló, D., Petrovic, M. ISBN: 978-3-540-74793-2

Emerging Contaminants in Waste Waters: Sources and Occurrence M. Petrovic · J. Radjenovic · C. Postigo · M. Kuster · M. Farre M. L. de Alda · D. Barceló Analysis of Emerging Contaminants of Municipal and Industrial Origin M. Gros · M. Petrovic · D. Barceló Acute and Chronic Effects of Emerging Contaminants T. Smital Traceability of Emerging Contaminants from Wastewater to Drinking Water M. Huerta-Fontela · F. Ventura Impact of Emergent Contaminants in the Environment: Environmental Risk Assessment J. Blasco · A. DelValls

Hdb Env Chem Vol. 5, Part S/2 (2008): 1–35 DOI 10.1007/698_5_098 © Springer-Verlag Berlin Heidelberg Published online: 29 November 2007

Removal of Emerging Contaminants in Wastewater Treatment: Conventional Activated Sludge Treatment G. Buttiglieri1 · T. P. Knepper2 (u) 1 DIIAR-Environmental

Section, Politecnico di Milano, P.za Leonardo da Vinci, 32, 20133 Milan, Italy 2 Europa University of Applied Science Fresenius, Limburger Strasse 2, 65510 Idstein, Germany [email protected] 1

Deﬁnition of Conventional Activated Sludge Treatment . . . . . . . . . . .

3

2

Legislation in Europe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6

3 3.1 3.2 3.3 3.4

Micropollutants Removal in CAS . . . . . . Pesticides . . . . . . . . . . . . . . . . . . . . Pharmaceuticals and Personal Care Products Surfactants and Their Metabolites . . . . . . Other Contaminants . . . . . . . . . . . . . .

. . . . .

11 11 16 25 29

4

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

32

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

33

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

Abstract Persistent organic pollutants released from wastewater discharges into the environment can persist in the environment, bioaccumulate through the food web, enter drinking water production and pose a risk to human health and the environment. Conventional activated sludge (CAS) treatments are usually designed to remove or to decrease the concentrations of pathogens and the loads of the bulk organic but generally they are not designed to remove residues of trace organics. In the present work the presence of pesticides, pharmaceuticals and personal care products, surfactants and other contaminants and their removal in CAS systems are presented and discussed. The capacity to eliminate micropollutants in CAS depends on various factors, including physico-chemical properties, biological persistence of the individual compound and the technology and process conditions (e.g. temperature and seasonal variability, hydraulic and sludge retention time applied). Sludge retention time, though not exclusively, has been revealed as one of the most important process parameters. The relative importance of sorption, more relevant for lipophilic compounds and some hydrophilic compounds (e.g. surfactants), as compared to biodegradation can increase when the residence time in CAS is too short for implementing an efﬁcient degradation. For high polar substances (e.g. most pharmaceuticals) the most important removal process is biological transformation or mineralization by microorganisms. Finally, present levels of knowledge about the degradation pathway in CAS is often not complete and formation of human and natural metabolites should be continuously and carefully monitored as they can be, occasionally, more toxic than the parental compounds.

2 Keywords CAS · Pesticides · Pharmaceuticals · Surfactants Abbreviations ABS alkylbenzene sulfonate ADBI celestolide AEO alcohol ethoxylates AES alkyl ether sulfate AHDI phantolide AHTN tonalide APEO alkylphenol ethoxylates AS alkyl sulfate ATII traseolide BE benzoylecgonine BPA bisphenol A BOD biochemical oxygen demand BTri benzotriazole CAS conventional activated sludge COD chemical oxygen demand DEET N,N-diethyl-m-toluamide DCPEG dicarboxylated metabolite of MCEP DPMI cashmeran DTPA diethylenetriamino pentaacetate DWD drinking water directive EDPP primary metabolite of methadone EDTA ethylenediamino tetraacetate EQS environmental quality standards HHCB galaxolide HRT hydraulic retention time LAS alkylbenzene sulfonate MCPEG monocarboxylated polyethylene glycols MCPP mecocrop NP nonylphenol NPEC nonylphenoxy carboxylates NPEO nonylphenol ethoxylates OC octocrylene OMC octyl-methoxycinnamate OPEO octylphenol ethoxylates OT octyl-triazone PAA peroxyacetic acid PCM polycyclic musk PEG polyethylene glycol PPCP pharmaceuticals and personal care products SRT sludge retention time TCEP tris-2-chloroethyl phosphate TCPP tris-2-chloropropyl phosphate TSS total suspended solids TTri 4-/5-tolyltriazole VSS volatile suspended solids WFD water framework directive

G. Buttiglieri · T.P. Knepper

Removal of Emerging Contaminants in Wastewater Treatment: CAS Treatment

3

WWTP wastewater treatment plant 4-MBC 3-(4-methylbenzylidene) camphor

1 Definition of Conventional Activated Sludge Treatment Activated sludge is the biomass produced in wastewater by the growth of organisms in aeration tanks in the presence of dissolved oxygen. The term “activated” derives from the high presence of bacteria and other micro-organisms in such biomass living on the incoming wastewater. Sewage and activated sludge are mixed and aerated as oxygen is required by biomass to live, grow, and multiply in order to reduce the dissolved organic content in the wastewater. A Conventional Activated Sludge (CAS) treatment is usually designed to remove or to decrease the concentrations of pathogens and the loads of the bulk organic and inorganic constituent that may otherwise pollute the receiving waters and lead to eutrophication. It is, generally, made up of a water treatment line (for removal of pollutants from the water phase with production of sediments with high water content) and a sludge line (for treatment of separated activated sludge, produced in the water line, in order to make them compatible with the ﬁnal disposal). A water treatment line in CAS, Fig. 1, usually includes the following phases: • Mechanical preliminary treatments (or pre-treatments): griding, removal of sand and oils, equalization tanks; • Mechanical treatments (or primary): primary settling, eventually after chemical-physical treatment of precipitation and ﬂocculation; • Biological treatment (or secondary): activated sludge (or attached biomass) necessarily followed by secondary settling. Optional phases, Fig. 1, may be included: • Tertiary treatment: nitriﬁcation-denitriﬁcation, phosphate removal. Nitrate removal is biological while phosphorus is usually removed by chemical precipitation; • Quaternary treatments: ﬁltration, chemical oxidation (ozonization and similar), adsorption on activated carbon, wetlands, etc.; • Final disinfection (among quaternary treatments): ﬁnal step in order to improve microbiological quality of efﬂuent water (chlorination, UV, or peroxyacetic acid PAA) in terms of bacteria, helminths eggs, virus, etc. [1]. Out of the water treatment line sludges are liquids with a high content of suspended solids (around 10–40 g L–1 ), usually characterized by high putrescibility and so needing further treatments.

4

G. Buttiglieri · T.P. Knepper

Fig. 1 Scheme of an activated sludge treatment plant. Quaternary treatments and phases indicated in the bottom are optional. In the upper side sludge, and other materials, productions are indicated in grey boxes. Adapted from [1]

The sludge line is, commonly, made up of: • Concentration: additional sedimentation in order to increase the solid percentage; • Aerobic stabilization–anaerobic digestion: to reduce putrescibility usually with biological processes; • Hygienization: not frequent step to eliminate pathogens; • Sludge conditioning and disidratation: conditioning to increase disidratation characteristics of the sludge to get easier ﬁltration and centrifugation, disidratation till a high dry content; • Final disposal: usually ﬁnal destinations include agricultural reuse, incinerator, and landﬁll. Concerning the use of sludge in agriculture seven EU Member States (Belgium-Wallonia, Denmark, Spain, France, Ireland, the UK, and Hungary) reported that they apply 50% or more of the sludge they generate on land (while the others report minor or no sludge spread on agricultural land, 86/278/EEC and Report [2]). The standard parameters that are commonly checked in a CAS treatment are listed below. • Physical parameters: pH, temperature, and dissolved oxygen (DO). These parameters inﬂuence the biological activities of the biomass (and the rate of pollutant removal kinetics) as well as the chemical removal in primary and tertiary treatments. As a general rule these parameters need to be optimized depending on the speciﬁc inﬂuent characteristics and removal necessities. Extreme conditions of pH can inhibit or be harmful to bacteria. Lower temperature (e.g. during winter or in European northern

Removal of Emerging Contaminants in Wastewater Treatment: CAS Treatment

•

•

•

•

•

5

countries compared to southern ones) decreases the removal kinetics, in particular for aerobic biomass (e.g. nitriﬁcant bacteria) [3–6]. As a consequence, with lower temperature, it is necessary to design bigger tanks for activated sludge in order to achieve high removal of organic compounds or for nitrogen removal (in particular for ammonia removal in tertiary treatment). Oxygen must be checked in aerated tanks (as a common rule should be higher than 1.5–2 mg L–1 ) as well as its absence in the anoxic tank for tertiary treatment of nitrogen (where present). Hydraulic Retention Time (HRT) and Sludge Retention Time (SRT). HRT is a measure of the average length of time that a soluble compound remains in a built reactor. It is usually between a few hours and very few days depending on the load inﬂuent and the design of the plant [3]. It is an important control parameter in many wastewater treatment processes and has a profound inﬂuence on the contact time of the compounds within the reactor as well as on the design of the dimensions of the tanks. As a consequence to optimize process performance, to have a sufﬁcient contact time within the reactors and not too big volumes, HRT should be carefully selected and maintained. SRT is a measure of the time that sludge remains in the water treatment line. In a CAS it affects the total suspended solid concentration (and the sludge to be withdrawn out of the systems) as well as the percentage of compounds adsorbed to sludge itself. Lower SRT leads to higher production of sludge to be treated and may lead to incomplete removal of organics and nutrients (as discussed in the following). Total Suspended Solids (TSS) and Volatile Suspended Solids (VSS). TSS indicates the amount of solids suspended in the water; VSS represents the content of volatile compounds and is usually related to biomass content. VSS is a important indicator measured to follow biomass growth in CAS wastewater treatment systems. BOD (Biochemical Oxygen Demand) and COD (Chemical Oxygen Demand). These parameters indicate the content of organic matter and the biodegradability of the inﬂuent wastewaters. Higher BOD/COD ratio indicate more easily biodegradable sewage and is common for domestic sewage; lower ratios show higher presence of recalcitrant compounds and possibly the sewage origin is industrial or mixed. Nutrient content: Nitrogen and Phosphorus. Total Nitrogen content as well as Nitrate and Ammonia content are to be taken into account to design correctly the aerated and anoxic phase, and their sequence, in the biological section (particularly if the discharge of treated water is inside a sensitive area). Phosphorus is, generally, of major concern especially for sensitive areas and has to be controlled for reduction of eutrophication into receiving waters. Pathogens and other microbiological parameters. For discharging water, content of E. coli, Helminths eggs and other pathogens need to be checked particularly for water reuse.

6

G. Buttiglieri · T.P. Knepper

2 Legislation in Europe Since 2000 the Water Framework Directive (WFD) has been in place as the main European legislation to protect our water resources and the water environment of Europe. It requires managing the river basin so that the quality and quantity of water does not affect the ecological services of any speciﬁc water body. Viewing rivers as integrated systems this EU framework aims to achieve good chemical and biological water status and to restore every river, lake, groundwater, wetland and other water body across the community to a “good status” by 2015. The only exception should apply to water bodies designated by their government as “heavily modiﬁed” and where over-riding socio-economic reasons prevent the necessary improvements. Ecological status is based on biological elements and supported by chemical and physicochemical elements, as well as hydro-morphological elements. Chemical status refers to speciﬁc pollutants (e.g. priority substances) for which environmental quality standards are proposed. The European Commission adopted a proposal for a new Directive to protect surface water from pollution on 17th July 2006 [7], which will set limits on concentrations in surface waters of 41 dangerous chemical substances (including 33 priority substances and eight other pollutants) that pose a particular risk to animal and plant life in the aquatic environment and to human health (Tables 1 and 2). The priority hazardous substances are identiﬁed, taking into account the precautionary principle, relying in particular on the determination of any potentially adverse effects of the product and on a scientiﬁc risk assessment [8]. With regard to pollution prevention and control, Community water policy is based on a combined approach. According to article 10 (WFD 2000/60/EC) for point and diffuse sources Member States shall ensure the establishment and/or implementation of the emission controls based on best available techniques (BAT) or the relevant emission limit values or in the case of diffuse impacts the controls including best environmental practices (BEP) and of environmental quality standards (EQS) for water, sediment, and biota. The environmental quality standards should be complied with by 2015 at the latest and direct discharges of pollutants into surface water should cease by 2025. Beside the WFD the urban wastewater treatment directive (UWWT), the bathing water directive, the nitrates directive and the drinking water directive will stay in place. The objective of the Drinking Water Directive (98/83/EC) is to protect the health of the consumers in the European Union. In the DWD a total of 48 microbial and chemical parameters must be monitored and tested regularly. The parametric values of the DWD are based on the scientiﬁc knowledge available and the precautionary principle (e.g. pesticides) has also been taken into account. There is also concern over endocrine-disrupting chemicals but at

Removal of Emerging Contaminants in Wastewater Treatment: CAS Treatment

7

Table 1 Environmental Quality Standards (EQS) for priority substances in surface water [7] (Unit: [µg/l]) No. Name of substance

1 2 3 4 5 6

7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

Alachlor Anthracene Atrazine Benzene Pentabromodiphenylether Cadmium and its compounds (depending on water hardness) C10–13 Chloroalkanes Chlorfenvinphos Chlorpyrifos 1,2-Dichloroethane Dichloromethane Di(2-ethylhexyl) phthalate(DEHP) Diuron Endosulfan Fluoranthene Hexachlorobenzene Hexachlorobutadiene Hexachlorocyclohexane Isoproturon Lead and its compounds Mercury and its compounds Naphthalene Nickel and its compounds Nonylphenol Octylphenol Pentachlorobenzene Pentachlorophenol Polyaromatic hydrocarbons (PAH)

CAS number

AA-EQS Inland surface waters1

AA-EQS Other surface waters1

15972-60-8 120-12-7 1912-24-9 71-43-2 32534-81-9

0.3 0.1 0.6 10 0.0005

7440-43-9

< 0.08–0.25

MAC-EQS Inland surface waters2

MAC-EQS Other surface waters2

0.3 0.1 0.6 8 0.0002

0.7 0.4 2.0 50 n.a.

0.7 0.4 2.0 50 n.a.

0.2

< 0.45–1.5

85535-84-8

0.4

0.4

470-90-6 2921-88-2 107-06-2 75-09-2 117-81-7

0.1 0.03 10 20 1.3

0.1 0.03 20 20 1.3

1.4

1.4

0.3 0.1 n.a. n.a. n.a.

0.3 0.1 n.a. n.a. n.a.

330-54-1 115-29-7 206-44-0 118-74-1 87-68-3

0.2 0.005 0.1 0.01 0.1

0.2 0.0005 0.1 0.01 0.1

1.8 0.01 1 0.05 0.6

1.8 0.004 1 0.05 0.6

608-73-1

0.02

0.002

0.04

0.02

34123-59-6 7439-92-1

0.3 7.2

0.3 7.2

7439-97-6

0.05

0.05

91-20-3 7440-02-0 25154-52-3 1806-26-4 87-86-5 87-86-5

2.4 20 0.3 0.1 0.007 0.4

1.2 20 0.3 0.01 0.0007 0.4

1.0 n.a. 0.07

1.0 n.a. 0.07

n.a. n.a.

n.a. n.a.

2.0 n.a. n.a. 1

2.0 n.a. n.a. 1

8

G. Buttiglieri · T.P. Knepper

Table 1 (continued) No.

29 30 31 32 33 1 2

Name of substance CAS number

AA-EQS Inland surface waters1

AA-EQS Other surface waters1

Benzo(a)pyrene Benzo(b)ﬂuoranthene Benzo(k)ﬂuoranthene Benzo(g,h,i)perylene Indeno(1,2,3-cd)pyrene Simazine Tributyltin compounds Trichlorobenzenes (all isomers) Trichloromethane Triﬂuralin

0.05 = 0.03

0.05 0.1 = 0.03 n.a.

50-32-8 205-99-2 207-08-9 191-24-2 193-39-5 122-34-9 688-73-3

= 0.03 = 0.002 = 0.002

MAC-EQS Inland surface waters2

MAC-EQS Other surface waters2 0.1 n.a.

= 0.03 n.a.

n.a.

= 0.002 n.a.

n.a.

= 0.002 n.a.

n.a.

1 0.0002

1 0.0002

4 0.0015

4 0.0015

12002-48-1

0.4

0.4

n.a.

n.a.

67-66-3 1582-09-8

2.5 0.03

2.5 0.03

n.a. n.a.

n.a. n.a.

This parameter is the Environmental Quality Standard expressed as an annual average value (EQS-AA). This parameter is the Environmental Quality Standard expressed as a maximum allowable concentration (EQS-MAC). Where the MAC-EQS are marked as “n.a.” = “not applicable”, the Annual Average-EQS values are also protective against short-term pollution peaks since they are signiﬁcantly lower than the values derived on the basis of acute toxicity.

present no parametric values are established. The revised Bathing Water Directive (Directive 2006/7/EC) is based on microbiological indicators of faecal contamination and it reduces the list of parameters to just two Escherichia coli and Intestinal Enterococci. The Groundwater Directive (2006/118/EC) sets underground water quality standards and introduces measures to prevent or limit inputs of pollutants into groundwater only for chemical parameters in response to the requirements of the Water Framework Directive. For emission limitation the Urban Wastewater Treatment Directive (UWWT 91/271/EEC) aims to protect the environment from the adverse effects of urban, and from certain industrial sectors, wastewater discharges and concerns the collection, treatment, and discharge of domestic wastewater, mixture of wastewater and wastewater from certain industrial sectors. The main focus of the Directive is the prevention of eutrophication or oxygen depletion and the monitoring of the performance of treatment plants and re-

Removal of Emerging Contaminants in Wastewater Treatment: CAS Treatment

9

Table 2 Environmental Quality Standards (EQS) for other Pollutants [7] (Unit: [µg/l]) No. Name of substance

1 2 3 4 5 6 7 8 1 2

DDT total para-para-DDT Aldrin Dieldrin Endrin Isodrin Carbontetrachloride Tetrachloroethylene Trichloroethylene

CAS number

n.a. 50-29-3 309-00-2 60-57-1 72-20-8 465-73-6 56-23-5 127-18-4 79-01-6

AA-EQS Inland surface waters1

AA-EQS Other surface waters1

MAC-EQS Inland surface waters2

0.025 0.01 = 0.010

0.025 n.a. 0.01 n.a. = 0.005 n.a.

n.a. n.a. n.a.

12 10 10

12 10 10

n.a. n.a. n.a.

n.a. n.a. n.a.

MAC-EQS Other surface waters2

This parameter is the Environmental Quality Standard expressed as an annual average value (EQS-AA). This parameter is the Environmental Quality Standard expressed as a maximum allowable concentration (EQS-MAC). Where the MAC-EQS are marked as “n.a.” = “not applicable”, the Annual Average-EQS values are also protective against short-term pollution peaks since they are signiﬁcantly lower than the values derived on the basis of acute toxicity.

ceiving waters is consistent with that goal. The IPPC Directive (96/61/EC) is about minimizing pollution from various industrial sources [9]. Among the others speciﬁc regulation (2004/850) on persistent organic pollutants, limitation in the production and use, is also to be considered: “The Community is seriously concerned by the continuous release of persistent organic pollutants into the environment. These chemical substances are transported across international boundaries far from their sources and they persist in the environment, bioaccumulate through the food web, and pose a risk to human health and the environment. Further measures need therefore to be taken in order to protect human health and the environment against these pollutants”. The increasing extent and level of municipal wastewater treatment in Europe in the past decades has signiﬁcantly improved the quality of surface waters, even though obligations set for the European Union are not equally fulﬁlled by all its members (Table 3). Oxygen depletion is now largely under control in many places. Phosphate concentrations in European rivers have been regulated and eutrophication of lakes and coastal waters has been reduced as a result even if problems remain. According to EEA (2003) nitrogen pollution, particularly from agriculture, has remained constant and there is no evidence of changes in nitrate levels in groundwater [10]. Organic matter is removed effectively in wastewater treatment plants (WWTP) as well as many organic trace pollutants. Pollution of rivers by heavy metals and some

10

G. Buttiglieri · T.P. Knepper

Table 3 Waste water treatment in agglomerations affected by sensitive areas and organic loads – Situation at 1. January 2002 [11] Member state

Belgium Denmark Germany 3 Greece Spain France Ireland Italy Luxembourg Netherlands

Articles applied1

5(8) 5(4)

5(8), 5(4)4 5(8), 5(4)

Austria 5(8)5 Portugal Finland 5(8) Sweden United Kingdom Total MS not applying Article 5 1

2 3

4 5

Agglomeration Complying concerned treatment level Load [p.e.] %2 8 952 516 6 698 384 124 876 488 609 400 5 740 260 16 728 379 3 362 856 3 024 094 804 500 15 906 991 1 851 885 1 372 700 6 377 300 7 672 670 6 221 177 210 199 600 69 416 121

29 96 P-Reduction 90% N-Reduction 74% 40 25 36 8 72 14 P-Reduction 79% N-Reduction 66% 100 11 7 73 29 – 42

Non complying treatment level % 71 4 – 60 75 64 92 22 86 – 0 90 93 27 71 – 58

According to Article 5(8), a Member State does not have to identify sensitive areas for the purpose of the Directive if it implements the treatment established under paragraphs 2, 3 and 4 of the Directive over all its territory. The option of Article 5(4) of the Directive exempts a Member State from the provisions for individual treatment plants with more than 10 000 p.e. according to Article 5(2) and 5(3), but it has to show that a minimum percentage of reduction in the overall load entering a treatment plant in that area is at least 75% for total phosphorus and 75% for total nitrogen. Percentage in relation to the total organic load affected in the Member State. Germany did not include the waste water load of their entire territory, but only the load of agglomerations above 2000 p.e. In Germany the load of agglomerations below 2000 p.e. represents about 2% of the entire waste water load. Luxembourg applies Article 5(4), but wishes to be evaluated according to Article 5(2) and 5(3) until it achieves full compliance with Article 5(4). As Austria applies Article 5(8) from the end of 2002 onwards. The current evaluation includes only agglomerations discharging into the catchment areas of sensitive areas identiﬁed by other Member States.

other heavily regulated chemicals is generally decreasing but data availability for many other pollutants is too weak to make assessments and there is a lack of comparable data on the European scale.

Removal of Emerging Contaminants in Wastewater Treatment: CAS Treatment

11

3 Micropollutants Removal in CAS Regarding poorly degradable compounds several of them may be discharged with WWTP efﬂuents into receiving waters. In many cases the receiving waters are used for drinking water production, e.g. via bank ﬁltration. In this way a partly closed water cycle is obtained, in which treated municipal wastewater is used indirectly for drinking water production. In partly closed water cycles, however, poorly biodegradable polar compounds can be quite problematic as they may travel along the water path from wastewater to the raw water used for drinking water production. If polar compounds are not biodegradable they are not fully retained by any of the natural or technical barriers foreseen in such a cycle, and dilution may be the only mechanism by which their concentration decreases [12]. Polar compounds may occur in WWTP efﬂuents because they are truly persistent under the conditions of an activated sludge process or because their microbial degradation was not fast enough to be completed within a low hydraulic retention time. An incomplete degradation implies, however, that this compound could be further degraded after its discharge in the receiving water body. The concentration of a metabolite may increase in activated sludge treatment even though it is degradable, provided that its formation proceeds faster than its further transformation. The occurrence and removal of many polar high production volume chemicals has been studied far less thoroughly. It is often not clear whether a certain class of compounds is widespread in municipal wastewaters, to which extent they are removed in WWTP and whether yet unknown polar metabolites are being formed [13]. The emerging contaminants to be considered have different physicochemical properties (from strongly acidic to neutral and weakly basic) and may originate from different wastewater sources such as household, industry, and surface runoff. In the present work the removal in a CAS treatment plant of the following classes of compounds will be taken into account: Pesticides, Pharmaceuticals and Personal Care Products, Surfactants and their metabolites and other contaminants. 3.1 Pesticides Pesticides are used in order to protect plants or plant products against harmful organisms or prevent the action of such organisms. Many of them are not fatal for the target organisms but may also harm organisms in the environment and in this way affect the natural balance. A wide variety of types of plant protection have been developed to protect plants. Among these, herbicides are quantitatively the most important class (more than 50% of all

12

G. Buttiglieri · T.P. Knepper

pesticides used). The other main classes of pesticides are the insecticides and fungicides, followed by the acaricides, nematicides, molluscicides, and rodenticides. Herbicides can be classiﬁed under different aspects such as selectivity, chemical structure, the mode and mechanism of the biological action, and the types of plants to be controlled. They are used wherever the growth of undesired weeds has to be inhibited and are used extensively in North America, Western Europe, Japan, and Australia. Consumption per km2 can vary widely [14] and the concentrations of the active ingredients, even of the same herbicide, vary strongly and depend mostly on the effectiveness of the agent, the time of application, the ﬁeld conditions, and the technical way of spraying. For example, some amounts are presented in the literature for diuron: in the UK 6.7 kg ha–1 , in France 1.8 kg ha–1 , and in Germany 0.6–4.8 kg ha–1 are applied in pre-emergence mode and 10–30 kg–1 as a total herbicide [15, 16]. After contamination of the aquatic environment was observed maximum application rates were limited to 1 kg ha–1 and the use of atrazine was even banned in many European countries. To reduce their environmental impact, modern herbicides are expected to decompose within a relatively short time after application without forming persistent metabolites or degradation products [17]. Removal in CAS By cleaning pesticide spraying equipment on the farmyard instead of washing it at the ﬁeld edge, highly concentrated solutions of pesticides may enter the sewerage system if the farmyard is connected to it [18]. The same will happen when herbicides are applied at sealed surfaces in residential areas. In this way, municipal wastewater may be signiﬁcantly loaded with herbicides, which rain causes to be washed into the sewage drains, during the period of herbicide application [19]. In a municipal CAS treatment plant, the polar herbicides are not always adequately removed. Therefore, in CAS efﬂuents, high concentrations of different herbicides can be observed, which contribute to the surface water contamination. The relative importance of sorption as compared to biodegradation can increase when the residence time in CAS is too short for implementing an efﬁcient degradation process or there is a high nutrient level for microorganisms and even the ones specialized in the biodegradation of herbicides ﬁnd an alternative more easily accessible. In this case a concentration decrease in CAS is often due to adsorption onto suspended solids rather than biodegradation [20]. Nevertheless, for the less hydrophobic compounds removal by sorption might be negligible [21]. Studies on the behavior of pesticides in water treatment processes refer mostly to laboratory tests or investigations of semi-technical plants. In add-

Removal of Emerging Contaminants in Wastewater Treatment: CAS Treatment

13

ition, water supply companies have collected a lot of data on the behavior of pesticides during the drinking water puriﬁcation process, but have seldom published their results. However, information on the behavior of polar herbicides is needed, because the limited number of reports available that are based on realistic operating conditions or which reproduce practical conditions are several years old or performed at concentration levels that do not correspond to practical conditions [22]. As to the triazine herbicide atrazine and simazine, and their degradation products desethylatrazine and desisopropylatrazine, primary wastewater treatment processes (evaluated by Meakins et al. during bench-scale CAS) appear to be ineffective in signiﬁcantly reducing the concentrations. The contribution of biological degradation to the removal of triazine is also negligible and the removal from wastewater is facilitated by the wastage of excess sludge; the maximum adsorption of any of the herbicides studied was < 40% [20]. In untreated municipal wastewater samples, in the western Balkan region, atrazine and simazine were present usually in rather low concentrations (< 250 ng L–1 ). A high concentration of atrazine (28 µg L–1 , Table 4) was observed only in the wastewater of the city of Sisak (Croatia), probably due to the contribution of industrial waste from the manufacture of herbicides at that location. Terbutylazin was detected up to 0.10 µg L–1 and metamitron, together with other pesticides, was not detected at all [23]. The occurrence and removal of 19 organochlorine pesticides during the CAS treatment process was investigated in the wastewater treatment plant, receiving mainly domestic wastewater and urban runoff, of the city of Thessaloniki, northern Greece [21]. The most frequent and abundant pesticide in untreated, primarily treated, and secondarily treated wastewater was Heptachlor-exo-epoxide (Table 4). Quintozene was also highly frequent, as expected since it is the only pesticide, among those studied, which is still registered for use in Greece. DDT was detected in about 30% of untreated wastewater samples; however, it was not found in the other stages of the treatment suggesting transformation to DDE and DDD (DDT metabolites). Other chlorinated pesticides, which despite their ban were still detected, are Dieldrin, Aldrin, HCHs, and Heptachlor. As expected, Dieldrin was always in higher concentrations than Aldrin, as a result of the potential transformation of the latter to the former. On the contrary, Hexachlorobutadine, Dichlobenil, Isodrin, and Heptachlor-endo-epoxide were never detected. The primary removal of organochlorine pesticides exhibited lower correlation with the log Kow showing that it cannot be attributed only to sorption, but to other mechanisms as well [21]. Degradation of acidic pesticides under laboratory conditions is well studied, but there are few publications dealing with their behavior in real WWTPs [19, 24, 25]. Generally, CAS was found to be ineffective in removing chlorinated phenoxyacid herbicides from settled sewage. However, under laboratory conditions mecoprop (MCPP) proved to be biodegradable (nearly

14

G. Buttiglieri · T.P. Knepper

Table 4 Mean removal of pesticides by CAS processes [13, 21, 23–25, 28] Compound

Inﬂuent concentration [µg L–1 ]

Efﬂuent concentration [µg L–1 ]

Elimination in CAS %

Triazine-derivatives and metabolites Atrazine Not detected 1 < LOD-28 5 Desethylatrazine Not detected 1 Simazine Not detected 1 < LOD-0.50 5 Terbutylazine Not detected 1 < LOD-0.10 5 Phenylurea-derivatives Isoproturon

Not detected 1

Phenoxyalkane carboxylic acids 2,4-D Not detected 1 2,4,5-T MCPA Not detected 1 MCPP Organochlorine compounds Hexachlorobutadine Dichlobenil Hexachlorobenzene Quintozene Isobenzan a-Endosulfan a-HCH b-HCH Γ -HCH Aldrin Isodrin Dieldrin Endrin Heptachlor Heptachlor-exo-epoxide Heptachlor-endo-epoxide p-p -DDE p-p -DDD p-p -DDT

18.2 7

< LOD-0.23 1 0.05–0.1 6

< 20 2 < 20 2 20–400 2,3 10.6 7
Not detected 4 Not detected 4 0.02 ± 0.016 4 0.06 ± 0.04 4 Not detected 4 0.051 ± 0.095 4 0.039 ± 0.033 4 0.026 ± 0.071 4 0.001 ± 0.006 4 0.010 ± 0.025 4 Not detected 4 0.027 ± 0.020 4 0.002 ± 0.007 4 0.046 ± 0.062 4 0.330 ± 0.290 4 Not detected 4 0.012 ± 0.018 4 0.022 ± 0.034 4 0.007 ± 0.017 4

Not detected 4 Not detected 4 0.0017 ± 0.0039 4 0.014 ± 0.014 4 Not detected 4 0.003 ± 0.005 4 0.006 ± 0.008 4 0.006 ± 0.013 4 0.001 ± 0.001 4 Not detected 4 Not detected 4 0.009 ± 0.010 4 0.003 ± 0.006 4 0.006 ± 0.010 4 0.025 ± 0.030 4 Not detected 4 0 ± 0.001 4 0.006 ± 0.012 4 Not detected 4

0–38 106

1

90 ± 15 4 75 ± 23 4 84 ± 17 4 79 ± 19 4 84 ± 9 4 82 ± 11 4 86 ± 9 4 77 ± 18 4 81 ± 5 4 91 ± 14 4 87 ± 16 4 88 ± 9 4 78 ± 18 4

Removal of Emerging Contaminants in Wastewater Treatment: CAS Treatment

15

Table 4 (continued) Compound

Other pesticides Metamitron Bentazone 1 2 3 4 5 6 7

Inﬂuent concentration [µg L–1 ]

Efﬂuent concentration [µg L–1 ]

Not detected 1,5 Not detected 1

7.96 7 9.88 7

Elimination in CAS %

[24]; [28]; [28]: sum of MCPA and MCPP; [21]; [23]; [13]; [25] (maximum detected concentration).

100%) even if this requires a long adaptation time (lag-phase) of activated sludge [26]. In real WWTPs, this presents a major difﬁculty since, like the majority of herbicides, MCPP is applied only during a short growth period of plants, which means that WWTPs that contain a non-adapted activated sludge, receive shock-loads of herbicides, which will not be eliminated (Table 4). A long acclimatization period (about 4 months) was also observed in a bench-scale study using sequencing batch reactors before 2,4-D—another acidic pesticide—biodegradation was established [27]. Subsequently, at steady-state operation, all reactors achieved practically complete removal (> 99%) of 2,4-D [28]. Bernhard et al., studying the presence and removal of different pesticides, found a lower concentration of MCPP (in the inﬂuents at concentrations from below LOD to 0.23 µg L–1 , Table 4) but conﬁrmed a low removal (in corresponding CAS-efﬂuents the concentrations were in a range from < LOD to 0.18 µg L–1 with a removal between 0% and 38%) [24]. Other pesticides—as there is not much agriculture around Wiesbaden—could not be detected while MCPP comes from leaching of roof isolation [29]. Another pesticide, isoproturon, was spiked in a membrane bioreactor (not CAS treatment with much longer time availability for adaptation); its removal started after 15 days of spiking and 50% of degradation was achieved after 19 days. This conﬁrms the need for adaptation of the microorganisms to particularly persistent compounds, which rarely occurs in the wastewater of the CAS. Even if raw water used for drinking water production is contaminated by polar herbicides, the drinking water itself is, usually, not affected. Technical ﬁltration and oxidation steps or combined puriﬁcation techniques are able to solve the problems. Therefore, in Europe only minor cases of exceeding the drinking water standard of 0.1 µg L–1 are reported [30, 31]. Despite the

16

G. Buttiglieri · T.P. Knepper

removal efﬁciencies observed, WWTPs remain potential sources of toxic contaminants to the aquatic environment, therefore the need of controlling their efﬂuents is essential. 3.2 Pharmaceuticals and Personal Care Products Pharmaceuticals and Personal Care Products (PPCPs) are a class of new, socalled emerging contaminants that have raised great concern in recent years. They deserve attention because of their continuous introduction into the environment via efﬂuents from sewage treatment and the concern about their effect on ecosystem and human health directly via drinking water and/or indirectly via the food chain. Residues of pharmaceuticals administered in large quantities to humans or animals have been detected in all compartments of the aquatic environment. The main sources of pharmaceutical residues in the environment are human medicine, the pharmaceuticals industry, and veterinary medicine. Discharge of therapeutic agents in efﬂuents from production facilities, hospitals, private households, improper disposal of unused drugs, and direct discharge of veterinary medicines lead to contamination of surface water, ground water, and eventually drinking water [32, 33]. Moreover, in the case of pharmaceuticals, they are developed with the intention of performing a biological effect. Often they are persistent in order to avoid the substance from becoming inactive before having a curing effect and have lipophilicity properties in order to be able to pass through membranes and to be taken up easily by the human or animal body. Finally, they are used by man in rather large quantities. As a consequence they can be referred to as “pseudo-persistent” contaminants as their transformation and/or removal rates are compensated by their continuous introduction into the environment [34]. Depending on the pharmacology, a medical substance will be excreted as a mixture of metabolites, as unchanged substance, or conjugated with an inactivating compound attached to the molecule. When they enter a wastewater treatment plant, xenobiotics are not usually completely mineralized. They are either partially retained in the sludge, or metabolized to a more hydrophilic but still persistent form and, therefore, pass through the wastewater-treatment plant and end up in the receiving waters. Their removal in WWTPs is variable and depends on the properties of the substance and process conditions (e.g. sludge retention time, hydraulic retention time, temperature) [35]. Levels of many pharmaceutically active compounds are barely reduced and they are, therefore, detected in WWTP efﬂuents. Although present at low concentrations in the environment, drugs can have adverse effects on organic organisms. These effects are chronic rather than acutely toxic, and depend on exposure (bioavailability), susceptibility to the compound in question, and the degradability of the substances [36].

Removal of Emerging Contaminants in Wastewater Treatment: CAS Treatment

17

Removal in CAS Residues of pharmaceuticals from human health applications are frequently detected in efﬂuent and efﬂuents from municipal CAS treatment plants. The amounts of pharmaceuticals administered to the patients and the degree of elimination from the human body are decisive for the probability of detection and the concentration levels of such residues in raw sewage. In general, municipal WWTP have been not designed to remove residues of trace organics and the ability of drug residues to pass through CAS depends on various factors, including the chemical and biological persistence of the individual compound, its sorption behavior, its evaporation via water vapor and the technology used for sewage puriﬁcation. For most pharmaceuticals the extent of stripping by aeration is negligible because of their low volatility. Sorption to suspended solids in the wastewater and subsequent removal by sedimentation as primary and secondary sludge comprise another removal process which is relevant for lipophilic compounds (log Pow > 3 at ambient pH). On the basis of the high polarity of most pharmaceuticals and the corresponding metabolites, signiﬁcant sorption by non-speciﬁc interactions can mainly be ruled out. Hence, the transformation or even mineralization of pharmaceuticals by microorganisms in CAS treatment remains the decisive removal process. Several compounds, such as for example paracetamol, are readily biodegradable during CAS while others pass through without reduction of their concentrations as shown in the following [17]. Aerobic and anaerobic biodegradation have been reported to be the most important processes for removal of pharmaceutical products from the dissolved phase, the amount removed increasing with hydraulic retention time and with the age of the sludge in the activated sludge treatment. For example, diclofenac has been shown to be signiﬁcantly biodegraded only when the sludge retention time was at least 8 days [37]. Abiotic transformation of pharmaceutical products in surface water or wastewater may occur by hydrolysis and photolysis. Because pharmaceutical compounds are usually resistant to hydrolysis, the extent of this reaction can be regarded as negligible for most environmentally relevant human drugs whereas direct and indirect photolysis is a primary pathway for their abiotic transformation in surface waters [37]. Seasonal variability and temperature, in pharmaceutical concentration and removal, is among the parameters to be taken into account. The total load and removal of 26 pharmaceutical substances, e.g., belonging to 11 therapeutic classes, in six CAS Italian treatment plants are considered [38]. The measured load values, comparable with other data [39] are in the range of 1.2–4.6 g d–1 1000 inhabitants–1 (Fig. 2). A comparison between winter and summer loads could be done for the Varese Olona plant (the only one sampled in both the seasons). In this plant the total inﬂuent load in summer was about half that in winter (1.5 vs. 3 g d–1 1000 inhabitants–1 ) probably because of higher removal rates and less

18

G. Buttiglieri · T.P. Knepper

Fig. 2 Removal rates in CASs. Total loads of the pharmaceuticals were normalized for population equivalents [38]

use of pharmaceuticals, particularly antibiotics, in summer than winter. For several pharmaceuticals the extent of contamination can be anticipated continuously during the year, for others seasonal variability can be expected. Considering the total loads, removal rates were generally less than 40% with the exception of the plant in Varese Lago (64%). When a comparison was possible (CAS in Varese Olona), removal rates were higher in summer than winter (31% vs. 0%), in line with a temperature-dependent increase of microbial activity (average temperature 9.7 ◦ C in winter and 18.6 ◦ C in summer) [38]. Effect of SRT: a study, conducted by Radjenovic et al. is presented as an example of CAS working at low SRT [36]. The behavior of several pharmaceutical products in different therapeutic categories (analgesics and antiinﬂammatory drugs, lipid regulators, antibiotics, etc.) was monitored during treatment in a conventional CAS process, Rubì WWTP, designed for 125 550 inhabitant equivalents. During the sampling program the WWTP was operating with an average daily ﬂow of 22 000 m3 day–1 . A mixture of municipal, hospital, and industrial wastewater is treated. The hydraulic retention time of CAS treatment was approximately 12 h and SRT approximately 3 days. Sampling points were primary sedimentation tank efﬂuent (inﬂow to the CAS) and CAS efﬂuent. The highest inﬂuent concentrations (µg L–1 ) were measured for the analgesic and anti-inﬂammatory drugs naproxen, ibupro-

Removal of Emerging Contaminants in Wastewater Treatment: CAS Treatment

19

fen, ketoprofen, diclofenac, acetaminophen, the antihyperlipoproteinaemic drugs gemﬁbrozil and bezaﬁbrate, the β-blocker atenolol and the diuretic hydrochlorothiazide. For other compounds input concentrations measured were sometimes close to the limits of quantiﬁcation [36]. Mean removal, calculated for each of the pharmaceutical compounds, is presented in Table 5. The removal (term used here for conversion of a micropollutant to compounds other than the parent compound) was greater than 75% for naproxen, ibuprofen, acetaminophen, hydrochlorothiazide, and paroxetine. Most other compounds were eliminated with a wide range of efﬁciencies. Carbamazepine, atenolol, and metoprolol were not removed at all. Rates of removal of the antibiotic sulfamethoxazole were very variable in both treatments investigated. According to Drillia et al. its microbial degradation will depend on the presence of readily biodegradable organic matter in wastewater [41]. Also, the possible conversion from human metabolite back to the original compound during treatment should be taken into account [42]. Efﬁciency of removal of other compounds (e.g. pravastatin, erythromycin, and indomethacin) varied in the CAS treatment but no explanation was suggested. In general great ﬂuctuations were observed and removal efﬁciency was found to be much more sensitive to changes in operating conditions (temperature, ﬂow rate, etc.) [36] Similar removal, Table 5, for pharmaceuticals compounds (ibuprofen, diclofenac, cloﬁbric acid, and carbamazepine) was performed in Wiesbaden WWTP (ﬂow of about 60 000 m3 d–1 , 90% municipal and 10% industrial) with a 22 HRT and a low SRT (12–14 d). 2,4-Dichlorobenzoic acid was measured too and detected in inﬂuent concentration varying from 0.1 to 3.5 µg L–1 . The removal of 2,4-Dichlorobenzoic acid ranged from 36% to 96% and it was found to be dependent on the concentrations of the compounds in the inﬂuents. At concentrations higher than 0.4 µg L–1 the removal reached levels of about 80% whereas at lower concentrations removal less than 80% was obtained [24]. A study conducted by Kosjek et al. focused on removal, with high SRT, of commonly used non-steroidal anti-inﬂammatory drugs (ibuprofen, naproxen, ketoprofen, diclofenac) and cloﬁbric acid in two specially designed small-scale pilot wastewater treatment plants (one at high concentration and the other at low concentration of micropollutants). The ﬁrst daily feed rate was 2.0 L per day, HRT of 48 h, and estimated effective SRT for most of the experimentation of about 100 days. Except for diclofenac, steady-state removal over a two-year monitoring period has been achieved. Elimination of the compounds was > 87% for ibuprofen, naproxen, and ketoprofen but only 49–59% for diclofenac and 30% of cloﬁbric acid removal with no sign of adaptation by the biomass. According to the literature [24, 43–45] a biomass adaptation time of several weeks to months is needed to achieve stable removal. Higher removal efﬁciency for the reactor fed with higher concentration of pharmaceuticals was achieved

20

G. Buttiglieri · T.P. Knepper

Table 5 Mean removal of selected pharmaceuticals by CAS processes [13, 23, 24, 32, 35, 36, 40] Compound

Inﬂuent concentration [µg L–1 ]

Elimination (%) in CAS

Analgesic and anti-inﬂammatory drugs 8–16 ÷ < LOD-1.55 5 1.2–2.7 ÷ < LOD-1.52 5 11–19 ÷4.9–12.3 2 1.2–3.7 3 ÷ 0.15–0 – 9 4 ÷ < LOD-11.9 5 3.4 ± 1.7 6 Diclofenac 1.1–3-5 ÷ 1.6–4.4 2 0.9–4.1 3 ÷ 1–2 4 ÷ 0.05–4.205 2.0 ± 1.5 6 Indomethacine 0.08–0.16÷ < LOD-0.24 5 Acetaminophen 7–26 Mefenamic acid 0.015–0.035 Propyphenazone 0.03–0.08 ÷ < LOD-0.46 5 Naproxen Ketoprofen Ibuprofen

85.1 ± 11.4 ÷ 90.1 ± 9.5 1 ÷ 45–50 5 51.5 ± 22.9 ÷ 90.4 ± 8.5 1 ÷ 55–99 5 82.5 ± 15.8 ÷ 91.2 ± 9.5 1 ÷ 97 2 ÷ 97.6 ± 3.3 3 ÷ > 90 4 ÷ > 99 5 96 6 50.1 ± 20.1 ÷ 54.2 ± 28.5 1 ÷ 24 2 ÷ 30.3 ± 30.6 3 ÷ 3 4 ÷ 0–30 5 33 6 23.4 ± 22.3 98.4 ± 1.72 29.4 ± 32.3 42.7 ± 19.0

Anti-ulcer agents Ranitidine

0.1–0.8 ÷ < LOD-0.758

5

Psychiatric drugs Paroxetine 0.03–0.07

42.2 ± 47.0 ÷ 0–45

5

90.6 ± 4.74

Antiepilectic drugs Carbamazepine Antibiotics Oﬂoxacin Sulfamethoxazole Erythromycin

0.18–0.3 ÷ 0.81–1.46 2 ÷ 0.3–1.93 No elimination ÷7 2 ÷ 5 4 ÷ 10–30 1–1.54 ÷ 0.12–1.55 5 ÷ 2.0 ± 1.3 6 0.2–0.5 0.2–1.3 ÷ 0.019–11.60 5 ÷ 0.8 ± 0.23 6 0.08–0.25 ÷ 0.024–0.420 5 ÷ 0.83 ± 0.27 6

23.8 ± 23.5 55.6 ± 35.4 ÷ 5–90 5 ÷ 24 6 23.8 ± 29.2 ÷ 0–5 5 ÷ 25 6

B-blockers Atenolol Metoprolol

0.2–3 ÷
Diuretics Hydrochlorothiazide

5–10

Hypoglycaemic agents Glibenclamide 0.050–0.075

76.3 ± 6.85

44.5 ± 19.1

5

Removal of Emerging Contaminants in Wastewater Treatment: CAS Treatment

21

Table 5 (continued) Compound

Inﬂuent concentration [µg L–1 ]

Lipid regulator and cholesterol lowering statin drugs Gemﬁbrozil 3–5.5 ÷ < LOD-1.70 5 Bezaﬁbrate 1–3 ÷ 1.5–7.6 3 ÷ < LOD-0.26 5 4.9 ± 4.1 6 Cloﬁbric acid 0.05–0.27 ÷ 0.06–0.15 2 ÷ 0.03–0.06 4 0.25 ± 0.09 6 Pravastatin 0.1–0.4 ÷ < LOD-1.17 5 1 2 3 4 5 6

Elimination (%) in CAS

38.8 ± 16.9 ÷ 60–80 5 48.4 ± 33.8 ÷ 75 ± 30.3 3 97 6 27.7 ± 46.9 ÷ 26 2 ÷ 55 4 ÷ 0–40

5

52 6 61.8 ± 23.6

[32], synthetic feed (5–50 µg L–1 ); [24]; [35]; [13]; [23]; [40].

but stopped at similar concentrations (around 0.1–0.4 µg L–1 ) in both reactors [32]. Such high SRT are, however, difﬁcult to be maintained in CAS and can be more suitably applied to other wastewater technologies (e.g. membrane bio-reactors). Nevertheless, it is important to remember that not only SRT might be considered for the removal of pharmaceuticals, and in general of emerging contaminants. Clara et al. investigated several compounds in selected full-scale treatment plants of different size and with different treatment efﬁciencies, lab and pilot scales [35]. The investigated substances include, among others, the pharmaceuticals ibuprofen, bezaﬁbrate, diclofenac, and carbamazepine (Table 5). The investigated micropollutants showed different behaviors during the wastewater treatment process. Some of the compounds were eliminated in dependency on the SRT, whereas the antiepileptic drug carbamazepine was not affected during the treatment. For diclofenac contradictory results were obtained and beside the SRT other inﬂuences seem to be of importance. With regard to ibuprofen and bezaﬁbrate a strong correlation between achievable efﬂuent concentrations and the SRT10 ◦ C was observed. For these substances a critical value was identiﬁable for the SRT10 ◦ C amounting to approx. 10 days [35]. The presented observations lead to the conclusion that for CASs operated at SRTs10 ◦ C higher than 10 days low efﬂuent concentrations can be achieved, for some compounds, leading to a signiﬁcant reduction of emissions. Compared to design guidelines, the recommended minimum sludge age corres-

22

G. Buttiglieri · T.P. Knepper

ponds to the requirements for nitrogen removal (nitriﬁcation, denitriﬁcation). According to the urban wastewater directive of the European Community (91/271/EEC) this recommended treatment efﬁciency corresponds to the standards for CASs situated in sensitive areas. It can be concluded that CASs with nitrogen removal may also efﬁciently remove degradable micropollutants. CASs designed for BOD removal only show a high optimization potential in regard to the removal of conventional wastewater parameters (e.g. nitrogen) as well as of micropollutants. Sensitive area requirements according to 91/271/EEC not only strongly reduce oxygen consumption for C and N oxidation in receiving waters and contribute to eutrophication abatement by N and P elimination, but may also achieve efﬁcient removal of micropollutants [35]. Personal Care Products (PCPs) Activities such as cleaning, washing of textiles, personal care or material protection may generate the release of chemicals incorporated in products into wastewater. Polycyclic musk compounds are used frequently as fragrances in washing powders, shampoos, and other consumer products, which are supposed to smell nicely. UV ﬁlters and biocides in plastics, coatings, or textiles may leach or volatilize from the matrix and end up in the sewer system with surface runoff. The consumption of organic UV ﬁlters, used in sunscreen products and considered as endocrine active chemicals, is increasing due to the growing awareness of hazards posed by UV radiation and recommendations for prevention of skin cancer. To improve product stability, UV ﬁlters are used as additives for cosmetics, plastics, clothing, or varnish. UV ﬁlters have been detected in ﬁsh, lake water, and wastewater [46, 47]. For PCP in general the transport through urban drainage systems and wastewater treatment plants (WWTPs) to the receiving waters is an important pathway for contamination of the aquatic environment. Because of high consumption volumes and low degradability, they have been detected in treated wastewater, in surface waters, in ﬁsh, and in sediments [48, 49]. Samples from a CAS (23 000 inhabitants served and SRT of 16 days) were taken to examine the fate and removal of several polycyclic musks (PCMs) (Galaxolide, Tonalide, Celestolide, Phantolide, Traseolide, Cashmeran) and UV ﬁlters [(3-4-methylbenzylidene) camphor (4-MBC), Octylmethoxycinnamate (OMC), Octocrylene (OC), Octyl-triazone (OT)] for different treatment steps [50]. The concentration of inﬂuents and efﬂuents are presented in Table 6. The concentration in the water line of Galaxolide and Tonalide decreased by 81% and 77%, respectively, whilst the reduction observed for the other PCMs ranged between 72 and 86%. UV ﬁlters were reduced between 92 and 99%. The mass balance calculated for the entire water line showed that more than 50% of PCMs, OC, and OT were sorbed onto the sludge, in contrast

Removal of Emerging Contaminants in Wastewater Treatment: CAS Treatment

23

to other micropollutants as shown in the previous and sorption onto solids was the dominating factor driving elimination for PCMs, OC, and OT (supported by measured Kd, sorption coefﬁcient, higher than 5000 L kgSS–1 ). The portion in the efﬂuent was < 25% for all compounds. Signiﬁcant degradation occurred for 4-MBC, OMC, and OC (45–90%) (Fig. 3). The sorption process is conﬁrmed to be the dominant process in other studies (among the others Zeng et al. 2007; Bester 2004, Table 6). Signiﬁcant removal, for galaxTable 6 Inﬂuent concentration and removal of polycyclic musks, UV ﬁlters, and insect repellents [13, 23, 40, 48, 50–54] Compound

Polycyclic musks (PCMs) Galaxolide [HHCB]

Tonalide [AHTN]

Celestolide [ADBI] Phantolide [AHDI] Traseolide [ATII]

Cashmeran [DPMI] UV ﬁlters 3-(4-methylbenzylidene) camphor [4-MBC] Octyl-methoxycinnamate [OMC] Octocrylene [OC] Octyl-triazone [OT]

Inﬂuent concentration [µg L–1 ]

Efﬂuent concentration [µg L–1 ]

4.42 ± 1.38 11.5–147 1 1.94 ± 0.35 2 0.43–7 3 0.83–4.44 5 0.030–2.67 7 1.3 ± 0.7 8 1.430 ± 0.284 0.890–3.470 1 0.583 ± 0.103 2 0.110–5.400 3 0.210–1.106 5 0.052–0.860 7 0.33 ± 0.20 8 0.20 ± 0.10 not detected 1 0.09 ± 0.01 not detected 1 0.090 ± 0.005 not detected 1

0.770 ± 0.177 1.680 ± 0.480 1 0.695 ± 0.058 2 0.010–0.098 3 0.451–0.870 5

< LOD-0.340 Not detected 0.21–0.69 1

0.73 ± 0.24 8 0.320 ± 0.112 0.110 ± 0.020 1 0.212 ± 0.017 2 0.013–0.230 3 0.144–0.170 5 0.10 ± 0.05 8 < LOQ not detected 1 < LOQ not detected 1 < LOQ not detected 1

7

Not detected 0.08 ± 0.01 1

0.960 ± 0.245

0.070 ± 0.019

20.070 ± 13.728

0.030 ± 0.008

1.680 ± 0.818 0.720 ± 0.185

< LOQ < LOQ

24

G. Buttiglieri · T.P. Knepper

Table 6 (continued) Compound

Insect repellents Bayrepel DEET 1 2 3 4 5 6 7 8 9

Inﬂuent concentration [µg L–1 ]

Efﬂuent concentration [µg L–1 ]

0.6–1.4 4 ÷

4,6

[51]; [48]. Most of the AHTN (∼ 80%) is transferred to the sludge and about 50% for HHCB. Biodegradation of HHCB may account to about 7% of the total. [52]. Log Kow = 5.9 for HHCB and 5.7 for AHTN. [53]; [54]; [13]; [23]; [24]. Bayrepel-acid, Bayrepel metabolite, with efﬂuent concentrations between 0.20 and 0.55 µg L–1 . [40].

olide and tonalide, up to 70% in the temperature range 12–25 ◦ C, was found during coagulation-ﬂocculation [55]. Nevertheless mass increment, in a study by Horii et al., of HHCB-lactone suggested formation of a lactone through the oxidation of HHCB [52]. Changes in the market may inﬂuence the presence and removal of emerging contaminants making harder the identiﬁcation of new compounds, their removal and metabolite formation. With any new xenobiotic introduced into the environment, in order to truly determine the toxicological and environmental risk, the persistence and fate of it, and any resulting metabolites, is also of extreme importance. An example is the insect repellent N,N-diethylm-toluamide (DEET). DEET has been detected at low µg L–1 levels in many different water bodies everywhere in the world, with peak concentrations of DEET in the summer and autumn months [53]. It could be shown, that the entry of DEET into the aquatic environment was mainly via wastewater, and that the degradation only occurred after an ambient adaptation time and at concentration levels exceeding a threshold value of approximately 1 and up to 2.5 µg L–1 [24, 53]. Since 1998, at least in the products produced by Bayer, the new active substance Bayrepel (1-piperidinecarboxylic acid, 2-(2-hydroxyethyl), 1-methylpropyl ester; KBR 3023) has gradually been introduced to all markets world-wide. Bayrepel was present in all analyzed inﬂuent samples taken with values in the range 0.6 and 1.4 µg L–1 (Table 6). In all corresponding WWTP efﬂuents Bayrepel was not present at all (< LOD).

Removal of Emerging Contaminants in Wastewater Treatment: CAS Treatment

25

Fig. 3 Removal of PMCs and UV ﬁlters in the entire water line [50]

It could be shown in laboratory microbial degradation experiments that the primary aerobic biodegradation of Bayrepel is very rapid. This was assumed to be also the case of the WWTP presented here, where the processes used include primary settling, activated sludge, and nitriﬁcation steps. Because of this rapid transformation, potential metabolites of Bayrepel need to be investigated and, in order to assess the fate of a product in the environment, it is by no means sufﬁcient to only monitor for the parent compound, but it is also essential to analyze for stable metabolites being formed [53]. 3.3 Surfactants and Their Metabolites Surfactants are a broad group of chemicals that play important and vital roles in a great variety of ﬁelds. Applications as diverse as cleaning, food, metallurgy, pharmacy, medicine, paints and varnishes, mining and many others utilize the characteristic properties provided by the surfactants. Following their use in aqueous systems as surface-active compounds, surfactants reach the CAS treatment plant but often they cannot be completely mineralized because of short retention times in the CAS. Therefore, together with their biochemical degradation products (metabolites), they arrive in surface waters. In addition, relevant compounds adsorb to the biological sewage sludge in the CAS and then, in original or partly degraded form, are disposed of along with excess sludge when this is removed.

26

G. Buttiglieri · T.P. Knepper

Today, these effects are not observed as often as in the past, as the non-biodegradable compounds of the branched alkylbenzene sulfonate type (ABS), formerly widely applied, have been replaced by compounds that are more easily biodegradable. Nevertheless, some surfactants can still be found in relatively high concentrations in efﬂuents of CAS and in relatively remote, estuarine sediments across Europe. Therefore, it is important that the fate of the corresponding metabolites should also be studied. The most widely used ionic and non-ionic surfactants should have the highest priority [56]. Under optimized conditions, more than 90–95% of surfactants can be eliminated by CAS treatment. Even if such high elimination rates are achieved, the principal problem is the formation of recalcitrant metabolites out of the parent surfactants. Some compounds, e.g. alkylphenol ethoxylates (APEOs) are eliminated incompletely, forming metabolites that are more resistant to further degradation and more toxic than the parent compounds. There is no complete knowledge about the degradation pathway of the main surfactants, especially for less used or new surfactants. As a consequence, very little data about their occurrence in wastewaters and elimination in CAS are available. On the other hand, some surfactants, such as APEOs and linear alkylbenzene sulfonates (LASs), have been frequently analyzed during the past two decades [56]. Ionic Surfactants Concentrations of linear alkylbenzene sulfates (LAS), among the anionic surfactants, are reported in raw inﬂuent in the range from 0.5 to more than 15 mg L–1 . The removal efﬁciency was generally higher than 95% (average removal was 98.5–99.9%) and the corresponding efﬂuent values were from < 5 to 1500 µg L–1 (Table 7). The results of ﬁve national pilot studies on LAS [57] showed that the differences observed in the main operating characteristics of the CAS (i.e. treatment type, plant size, SRT, HRT, and temperature) were not found to greatly inﬂuence the removal of LAS. In CAS a substantial percentage (15–35%) of LAS is physically removed by precipitation during the primary settling step. Alkyl ether sulfates (AES) and alkyl sulfates (AS), another important group of anionic surfactants, are used in laundry and cleaning detergents as well as in cosmetic products. The maximal reported concentration in the raw municipal wastewaters was 6 mg L–1 , while in treated water it ranged from 3 to 15 µg L–1 . Eichhorn et al. studied the change in AES concentration in a WWTP that discontinuously received heavily contaminated wastewaters from a company formulating personal care products, such as shower gels, shampoos, and foam baths [58]. High removal rates were found in seven CAS

Removal of Emerging Contaminants in Wastewater Treatment: CAS Treatment

27

Table 7 Inﬂuent, efﬂuent concentration and removal of surfactants in CASs [56, 58– 60, 64–70] Compound

Inﬂuent [µg L–1 ]

Efﬂuent [µg L–1 ]

IONIC SURFACTANTS Anionic surfactants LAS 500–15 000 AES Up to 6000 1 100–1300 2 400–4500 3 AS < 20–620 3 Cationic surfactants 2.4–325 NON-IONIC SURFACTANTS NPEO < 30–2120 NPEC < 0.2–219 NP < 0.5–251 Alcohol ethoxylates 125–3600

4

5 5

Removal [%]

< 5–1500 3–15 1 1.2–12.1 < 13 < 13 < 0.1–19

> 95 (average 98.5–99.9) > 99 99.2 2 99.9 3

2

99.9 3 95–99

4

< LOD-49 0.6–113 < LOD-225 < 0.1–509

4

81–99.5 5

97.0 ± 4.4 6 98–99.9 7

PEG and their derivatives PEG MCPEG DCPEG Non-ionic surfactants

85–3720 8 22–85 9 0.5–7.7 9 10–100 9 < 0.2–5.8 9 containing an amide group

CDEAs (coconut diethanolamides) Alkyl glucamides (AGs) Alkyl polyglucosides (APG)

111–124

10

14 10

26–45

1

< LOD-0.2 1

7–13

1

Not detected 1

81–98

8

∼ 90 10

∼ 100 1

1

[58]; [59]; 3 [60]; 4 [69]; 5 Higher value in WWTP receiving > 60% of industrial wastewater (mainly textile). 6 [64]; 7 [58, 65, 66]; 8 [58, 66–68]; 9 [68]; 10 [70]. 2

in The Netherlands with an average value of 99.2% (Table 7) [59] as well as in Germany, where elimination rates were higher than 99.9% [60].

28

G. Buttiglieri · T.P. Knepper

Cationic surfactants are used widely as fabric softeners, anti-static agents, antiseptic components, and as the main ingredients in rinse conditioner products. However, there is little data on their occurrence in wastewater and removal during wastewater treatment. Nevertheless, high removal, often higher than 99%, is reported (Table 7) [56]. Non-Ionic Surfactants Alkylphenol ethoxylates (APEO) are among the most widely used surfactants, although their environmental acceptability is strongly disputed. The widespread occurrence of APEO-derived metabolites in treated wastewaters raised concerns about their impact on the environment following disposal of efﬂuents into aquatic systems. Studies have found that their neutral and acidic metabolites are more toxic than the parent compounds. These ﬁndings have raised public concerns and a voluntary ban on APEO use in household cleaning products began in 1995 and restrictions on industrial cleaning applications in 2000. In western Europe and the USA APEOs have been completely replaced by alcohol ethoxylates (AEO). As a result, the general trend of declining APEO concentrations was observed. However, mainly because of its lower price, APEOs are still being used in substantial amounts in institutional and industrial applications [56]. The concentrations of nonylphenol ethoxylates (NPEOs), as the strongly prevalent sub-group of APEOs, determined in the inﬂuents of CAS varied widely among various plants from < 30 to 1035 µg L–1 (Table 7). However, values can go up to 22.5 mg L–1 in industrial wastewaters [56]. Levels of octylphenol ethoxylates (OPEOs) are signiﬁcantly lower, comprising approximately 5–15% of total APEOs in CAS inﬂuents. NPEO metabolites, nonylphenol (NP) and nonylphenoxy carboxylates (NPECs) could be detected in CAS inﬂuents at levels up to 40 µg L–1 (high values, in Table 7, are detected in industrial wastewaters). As for the elimination efﬁciency of NPEO during wastewater treatment, values from 81 to 99.5% were calculated. However, the removal of parent NPEOs led to the formation of transformation products that are much more resistant to microbial degradation and overall elimination of nonylphenolic compounds during sewage treatment is very low. Ahel et al., estimated that approximately 60–65% of all nonylphenolic compounds introduced to CAS are discharged into the environment: 19% in the form of carboxylated derivatives, 11% as lipophilic NP1 EO and NP2 EO, 25% as NP and 8% as untransformed NPEO. Furthermore, they also estimated that 92–96% of NP is discharged from the CAS via digested sludge and only 4–8% via secondary efﬂuents [61]. The oligomeric distribution of NPEOs in primary efﬂuent is characterized by a typical mixture that was slightly altered by biological degradation. Besides a maximum at NPEOs with 8–10 EO units (typical commercial mixture), another maximum appears at lower oligomers (nEO = 1–3), which are considered to be stable metabolic products. Ahel et al.

Removal of Emerging Contaminants in Wastewater Treatment: CAS Treatment

29

found that the elimination of higher oligomers was almost complete in an activated sludge treatment and higher oligomers (nEO > 8) were not present in the secondary efﬂuents [62]. Biologically treated wastewaters contained only traces of oligomers with nEO = 3–8. A signiﬁcant portion of NP remains in the solid phase and the contribution, in a study of De Voogt et al., of NP in the water fraction of inﬂuent, efﬂuent, and primary sludge was 52, 86, and 9% and of NPEOs 82, 86, and 11%, respectively [63]. Alcohol ethoxylates are widely accepted as environmentally safe surfactants and today they are often used as an alternative to APEOs. AEOs were detected in inﬂuent waters at concentrations of 125–3600 µg L–1 whereas concentrations up to 509 µg L–1 were found for efﬂuent samples [56]. Mc Avoy et al., found that the overall average removal of AEOs by trickling ﬁlter treatment was 89.9 ± 6.4% and in CAS was 97.0 ± 4.4% [64]. No apparent change in alkyl-chain length distribution was observed during primary and secondary treatment. Several studies also reported elimination efﬁciency ranging from 98 to 99.9%, irrespective of the type of sewage treatment, operating conditions, or inﬂuent concentration [58, 65, 66]. AEOs, especially those with a long alkyl chain, showed a great tendency to adsorb onto particulate matter. Polyethylene glycols (PEG) are widely used in different household products and several studies report concentrations of PEGs in WWTP inﬂuent, ranging from 85 to 3720 µg L–1 , [58, 67, 68] and the elimination efﬁciency ranged from 81 to 98%. Differences in the removal may partly be a result of differences in operational conditions during treatment processes and also the formation of PEGs as primary degradation products from APEOs. Analyses of inﬂuent and efﬂuent at CAS showed that even high molecular mass PEGs were very efﬁciently removed from wastewater. However, as a consequence of biodegradation processes occurring in the sewer and CASs, relatively high concentrations of acidic metabolites were found in raw and treated wastewaters. In CAS inﬂuent monocarboxylated polyethylene glycols (MCPEGs) were determined at levels of 22–85 µg L–1 and dicarboxylated metabolites (DCPEGs) at levels of 10–100 µg L–1 . In efﬂuents MCPEGs and DCPEGs were detected at concentrations of 0.5–7.7 and < 0.2–5.8 µg L–1 respectively [6]. Other surfactants (non-ionic surfactants containing an amide group and alkyl polyglucosides) inﬂuent, efﬂuent concentrations and removals are presented in Table 7. 3.4 Other Contaminants Synthetic Chelating Agents Ethylenediamino tetraacetate (EDTA) and diethylenetriamino pentaacetate (DTPA) are utilized in many industrial applications, e.g. in the textile, photo, and pulp and paper industries as well as in galvanic enterprizes and house-

30

G. Buttiglieri · T.P. Knepper

holds and have been proven to be widely distributed in aquatic systems. Concentration data regarding DTPA are quite scarce [71, 72]. Bernhard et al. noticed no removal for EDTA. CAS inﬂuent and efﬂuent was at an almost steady concentration level ranging from 107 to 134 µg L–1 (Table 8) [24]. Low removal is conﬁrmed by a study by Knepper et al. in which inﬂuent and efﬂuent samples taken from different CAS treatment plants in Spain, Germany, Austria, The Netherlands, and Belgium were analyzed for EDTA and DTPA. EDTA could be quantiﬁed in all investigated wastewater samples in concentrations between 4 and 970 µg L–1 whereas DTPA was not present in all the samples. The analyzed concentrations of DTPA were also signiﬁcantly lower with values between 1 and 155 µg L–1 [71]. Flame Retardants The ﬂame retardants TCEP and TCPP showed a slight reduction in CAS. TCEP was detected at concentrations from 0.20 to 0.60 µg L–1 in inﬂuent samples and between 0.14 and 0.36 µg L–1 in the efﬂuents. TCPP, which has replaced TCEP, could be detected in the inﬂuents at concentrations from 0.46 to 1.20 µg L–1 and between 0.50 and 0.99 µg L–1 in CAS efﬂuents (Table 8). Removals of TCEP by AST varied between 14% and 42%. The removals of TCPP of the WWTP ranged between 0% and 35%. An improved removal of both analytes by increasing SRT could not be observed [24]. Bisphenol A Bisphenol A is used by the manufacturers as an intermediate in the production of polycarbonate and epoxy resins, ﬂame retardants, and other specialty products. Final products include adhesives, protective coatings, powder paints, automotive lenses, protective window glazing, building materials, compact disks, optical lenses, etc [73]. High removal (> 90%) is usually detected (Table 8) [54, 74, 75]. Considering the low efﬂuent concentrations, biodegradation/biotransformation processes are the main removal pathway for BPA. In one of the CAS analyzed low removal was observed (inﬂuent 1.710 ng L–1 , efﬂuent 1.530 ng L–1 ), indicating a dependency of degradation on the SRT as this plant was operating at SRTs between 1 and 2 days [13, 54]. Drugs of Abuse Some drugs of abuse are present and measurable in surface waters of populated areas. Cocaine and its metabolite, for example, were detected in the largest Italian river, the Po, with a ﬁve million people catchment basin (steadily carried the equivalent of about 4 kg cocaine per day) [76]. Cocaine and its metabolite benzoylecgonine (BE) were present, and measurable, in surface

Removal of Emerging Contaminants in Wastewater Treatment: CAS Treatment

31

Table 8 Inﬂuent, efﬂuent concentration and removal of other contaminants in CASs [13, 23, 24, 54, 71, 74, 76, 77] Compound

Chelating agents EDTA

DTPA Flame retardants TCEP

TCPP

Bisphenol A

Inﬂuent [µg L–1 ]

107–134 4–970 100 5 1–155

Efﬂuent [µg L–1 ]

1

Benzoylecgonine Morphine Temazepan EDDP Benzotriazoles BTri TTri

0.20–0-60 1 0.1–0.25 5 < LOD-0.5 8 0.46–1.20 1 0.5–0.7 5 < LOD-2.5 8 0.2–0.5 3 0.72–2376 4

1 2 3 4 5 6 7 8

11 05

0.14–0.36

1

14–42 20 5

1

0.50–0.99

1

0–35

1

92–94 > 95 4

3

0.01–0.15 3 0.026–0.125

4

8

0.042–0.120 6 0.489 ± 0.117 7 0.390–0.750 6 0.290 ± 0.011 7 Not detected 7 0.320 ± 0.056 7 Not detected 7 7–10 5 2–3 5

1

2

< LOD-2.06 Drugs of abuse Cocaine

107–134

2

Removal [%]

0.138 ± 0.020 7 0.022 ± 0.004 7 < LOQ 7 0.126 ± 0.014 7 0.048 ± 0.001 7 35 5 10 5

[24]; [71]; [74]; [54]; [13]; [76]; [77]; [23].

waters of populated areas and in several Italian CASs wastewater samples tested (Table 8). As expected, the parent drug levels were much lower than the metabolite, their ratio in wastewater samples being in accordance with the known metabolic fate of cocaine in humans.

32

G. Buttiglieri · T.P. Knepper

Bones et al. analyzed a wastewater treatment plant in Dublin, Ireland looking for cocaine, BE, EDDP (primary metabolite of methadone), morphine, and temazepan. EDDP was detected in a number of efﬂuent samples in low ng L–1 concentrations, although methadone itself was not detected in any of the collected samples. Morphine was not detected in the inﬂuent of the CAS but was present at a detectable level below the LOQ in the corresponding efﬂuent sample. Such an observation may occur due to cleavage of glucuronide metabolites of morphine during the treatment process [77]. Benzotriazoles Benzotriazoles are a class of corrosion inhibitors in de-icing ﬂuids and dishwashing agents in industrial and household applications. This class consists of benzotriazole (BTri) and tolyltriazole isomers (TTri). Benzotriazoles have only recently been detected as regular contaminants of municipal wastewater. Reemtsma et al. evaluated its presence and removal in two monitoring campaigns from eight WWTP located in four countries (Austria, Belgium, Germany, Spain). Concentrations and removal are shown in Table 8 [13].

4 Summary The capacity to eliminate micropollutants depends on various factors, including the chemical and biological persistence of the individual compound, its sorption behavior, its volatility, and the technology and operation conditions (e.g. temperature, hydraulic retention time, and sludge retention time applied). Sorption to suspended solids in the wastewater and to sludge particles and subsequent removal by sedimentation is relevant for lipophilic compounds (log Pow > 3 at ambient pH) and some hydrophilic compounds that can interact in a speciﬁc way (e.g. surfactants). For high polar substances (such as most pharmaceuticals) and the corresponding metabolites the most important removal process is the biological transformation or mineralization by microorganisms, which strongly depends upon the treatment technology and operation conditions. Removal rates showed a great variability according to sewage treatment plants and types of treatments. The mere occurrence of a contaminant in CAS efﬂuents, anyway, does not imply that it would spread in the aquatic environment. Only when this polar pollutant was stable would this risk have to be considered. Therefore, a moderate efﬂuent concentration of a poorly degradable compound that did not experience any concentration decrease in CAS is to be considered more problematic than the same efﬂuent concentration of a compound that was degraded extensively from a much higher inﬂuent concentration [13].

Removal of Emerging Contaminants in Wastewater Treatment: CAS Treatment

33

Formation of human and natural metabolites, ﬁnally, should be continuously and carefully monitored as they can result, on some occasions, in metabolites that are more toxic than the parental compounds. Quaternary treatments (e.g. adsorption on activated carbon and disinfection) in addition to CAS, or different wastewater treatments, might be helpful to remove or further degrade these compounds, nevertheless taking into account the risks for possible toxic by-products (e.g. chlorine by-products in disinfection).

References 1. Bonomo L (2005) Wastewater treatment plant (Impianti di trattamento delle acque di riﬁuto). Course issue. Politecnico di Milano, Italy 2. Report from the Commission to the Council and the European Parliament on implementation of the community waste legislation for the period 2001–2003 [SEC(2006)972] 3. Metcalf E (1991) Wastewater Engineering. Mc Graw-Hill, New York 4. Auriol M, Meknassi YF, Adams CD, Tyagi RD (2006) Water Res 40:2847 5. Salvetti R, Azzellino A, Canziani R, Bonomo L (2006) Water Res 40:2981 6. Lishman L, Smyth SA, Saraﬁn K, Kleywegt S, Toito J, Peart T, Lee B, Servos M, Beland M, Seto P (2006) Sci Total Environ 367:544 7. Proposal for a Directive of the European parliament and of the council on environmental quality standards in the ﬁeld of water policy and amending Directive 2000/60/EC [COM(2006)398 ﬁnal] 8. TGD (2003) Technical guidance document on risk assessment in support of Commission Directive 93/67/EEC on risk assessment for new notiﬁed substances, Commission regulation (EC) No 1488/94 on risk assessment for existing substances, Directive 98/8/EC of the European Parliament and of the Council concerning the placing of biocidal products on the market. http://ecb.jrc.it/cgi-bin/reframer.pl?A=ECB&B=/tgdoc 9. Feuerhacker M (2007) Proceedings from Micropol & Ecohazard 2007. 5th IWA specialised conference on assessment and control of micropollutants/hazardous substances in water. 17–20 June 2007, Frankfurt, Germany 10. EEA (2003) EEA Brieﬁng status of Europe’s water No 1/2003 11. Report from the commission to the council, the European Parliament, the European economic and social committee and the committee of the regions. Implementation of Council Directive 91/271/EEC of 21 May 1991 concerning urban waste water treatment, as amended by Commission Directive 98/15/EC of 27 February 1998. [COM(2004) 248 ﬁnal] 12. Knepper TP, Barcelò D, Lindner K, Seel P, Reemtsma T, Ventura F, De Wever H, van der Voet E, Gehringer P, Schönerklee M (2004) Water Sci Technol 50:195 13. Reemtsma T, Weiss S, Müller J, Petrovic M, Gonzalez S, Barcelò D, Ventura F, Knepper TP (2006) Environ Sci Technol 40:5451 14. EDS (2004) Applied active agents of herbicides. EDS European Data Service, Federal Statistical Ofﬁce, Germany (available at http://www.welt-in-zahlen.de) 15. Goody DC, Chilton PJ, Harrison I (2002) Sci Total Environ 297:67 16. Tixier C, Bogaerts P, Sancelme M, Bonnemoy F, Twagilimana L, Cuer A, Bohatier J, Veschambre H (2000) Pest Manag Sci 56:455 17. Reemtsma T, Jekel M (eds) Organic pollutants in the water cycle. Properties, Occurrence and Environmental Relevance of Polar Compounds. Wiley-VCH, Weinheim, Germany

34

G. Buttiglieri · T.P. Knepper

18. 19. 20. 21. 22.

Carter A (2000) Pestic Outlook 11:149 Seel P, Knepper TP, Gabriel S, Weber A, Haberer K (1996) Vom Wasser 86:247 Meakins NC, Bubb JM, Lester JN (1994) Chemosphere 28:1611 Katsoyiannis A, Samara C (2004) Water Res 38:2685 Heinz I, Flessau A, Zullei-Seibert N, Kuhlmann B, Schulte-Ebbert U, Michels M, Simbrey J, Fleischer F (1995) Directive 80/778/EEC; part III: the parameter for pesticides and related products. Final report for the European Commission – DG XI, Brüssel Terzic S, Senta I, Ahel M, Gros M, Petrovic M, Barcelò D, Knepper T, Mitjans D, Ventura F, Jovancic P, Jabucar D (2007) Sci Total Environ (Submitted) Bernhard M, Müller J, Knepper T (2006) Water Res 40:3419 Peschka M, Müller J, Knepper TP, Seel P (2006) Trends in pesticide transport into the river Rhine. In: Hdb Env Chem Vol. 5, Part L: pp 155–175. Springer, Berlin Heidelberg New York Nitschke L, Wilk A, Schüssler W, Metzner G, Lind G (1999) Chemosphere 39:2313 Mangat SS, Elefsiniotis P (1999) Water Res 33:861 Petrovic M, Gonzalez S, Barcelò D (2003) Trends Anal Chem 22:685 Bucheli TD, Müller SR, Voegelin A, Schwarzenbach RP (1998) Environ Sci Technol 32:3465 Stackelberg PE, Gibs J, Furlong ET, Meyer MT, Zaugg SD, Lippincott RL (2007) Sci Total Environ 377:255 Rodriguez-Mozaz S, Lopez de Alda MJ, Barcelò D (2004) J Chromatogr 1045:85 Kosjek T, Heath E, Kompare B (2007) Anal Bioanal Chem 387:1379 Joss A, Zabczynski S, Anke Göbel A, Hoffmann B, Löfﬂer D, McArdell CS, Ternes TA, Thomsen A, Siegrist H (2006) Water Res 40:1686 Barcelò D, Petrovic M (2007) Anal Bioanal Chem 387:1141–1142 Clara M, Kreuzinger N, Strenn B, Gans O, Kroiss H (2005) Water Res 39:97 Radjenovic J, Petrovic M, Barcelò D (2007) Anal Bioanal Chem 387:1365 Nikolau A, Meric S, Fatta D (2007) Anal Bioanal Chem 387:1225 Castiglioni S, Bagnati R, Fanelli R (2006) Environ Sci Technol 40:357 Ternes TA (1998) Water Res 32:3245 Ternes TA, Bonerz M, Herrmann N, Teiser B, Andersen HR (2007) Chemosphere 66:894 Drillia P, Dokianakis SN, Fountoulakis MS, Kornaros M, Stamatelatou K, Lyberatos G (2005) J Hazard Mater 122:259 Göbel A, McArdell CS, Suter MJ, Giger W (2004) Anal Chem 76:4756 Tauxe-Wuersch A, De Alencastro LF, Grandjean D, Tarradellas J (2005) Water Res 39:1761–1772 Zwiener C, Frimmel FH (2003) Sci Total Environ 309:201 Heberer T (2002) J Hydrol 266:175 Poiger T, Buser HR, Balmer ME, Bergqvist PA, Müller MD (2004) Chemosphere 55:951 Balmer ME, Buser HR, Müller MD, Poiger T (2005) Environ Sci Technol 39:953 Bester K (2004) Chemosphere 57:863 Heim S, Schwarzbauer J, Kronimus A, Littke R, Woda C, Mangini A (2004) Org Geochem 35:1409 Kupper T, Plagellat C, Brändli RC, De Alencastro LF, Grandjean D, Tarradellas J (2006) Water Res 40:2603 Zeng X, Sheng G, Gui H, Chen D, Shao W, Fu J (2007) Chemosphere 69:1305 Horii Y, Reiner JL, Loganathan BG, Kumar KS, Sajwan K, Kannan K (2007) Chemosphere 68:2011 Knepper TP (2004) J Chromatogr 1046:159

23. 24. 25.

26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53.

Removal of Emerging Contaminants in Wastewater Treatment: CAS Treatment

35

54. Clara M, Strenn B, Gans O, Martinez E, Kreuzinger N, Kroiss H (2005) Water Res 39:4797 55. Carballa M, Omil F, Lema JM (2005) Water Res 39:4790 56. Knepper TP, Barcelò D, De Voogt P (2003) Analysis and fate of surfactants in the aquatic environment. Wilson & Wilson’s comprehensive analytical chemistry, vol 40. Elsevier, Amsterdam 57. Waters J, Feijtel TC (1995) Chemosphere 30:1939 58. Eichhorn P, Petrovic M, Barcelò D, Knepper TP (2000) Vom Wasser 95:245 59. Feijtel TC, Struijs J, Matthijs E (1999) Environ Toxicol Chem 18:2645 60. Schröder FR, Schmitt M, Reichensperger U (1999) Waste Manage 19:125 61. Ahel M, Giger W, Koch M (1994) Water Res 28:1131 62. Ahel M, Giger W, Molnar E, Ibric S (2000) Croat Chem Acta 73:209 63. De Voogt P, Kwast O, Hendriks R, Jonkers N (2000) Analysis 28:776 64. McAvoy DC, Dyer SD, Fendiger NJ, Eckhoff WS, Lawrence DL, Begley WM (1998) Environ Toxicol Chem 17:1705 65. Kiewiet AT, van der Steen JMD, Parsons JR (1995) Anal Chem 67:4409 66. Crescenzi C, Di Corcia A, Samperi R (1995) Anal Chem 67:1797 67. Barcelò D, Dachs J, Alcock S (eds) (2000) BIOSET: Biosensors for evaluation of the performance of wastewater treatment works. Final report. CSIC, Barcelona, Spain 68. Crescenzi C, Di Corcia A, Marcomini A, Samperi R (1997) Environ Sci Technol 31:2679 69. Waters J, Lee KS, Perchard V, Flanagan M, Clarke P (2000) Tenside Surf Det 71:161 70. Schröder HF (1993) Vom Wasser 81:299 71. Knepper TP, Werner A, Bogenschütz G (2004) J Chromatogr 1085:240 72. Lee HB, Peart TE, Kaiser KLE (1996) J Chromatogr 738:91 73. Staples CA, Dorn PB, Klecka GM, O’Block ST, Harris LR (1998) Chemosphere 36:2149 74. Nakada N, Tanishima T, Shinohara H, Kiri K, Takada H (2006) Water Res 40:3297 75. Ivashechkin P, Corvini P, Fahrbach M, Hollender J, Konietzko M, Meesters R, Schröder HF, Dohmann M (2004) In: Conference Proceedings of the Second IWA Leading-Edge Conference on Water and Wastewater Treatment Technologies, 1–4 June 2004, Prague, Czech Republic 76. Zuccato E, Chiabrando C, Castiglioni S, Calamari D, Bagnati R, Schiarea S, Fanelli R (2005) Environ Health 4:14 77. Bones J, Thomas KV, Paull B (2007) J Environ Monit 9:701

Hdb Env Chem Vol. 5, Part S/2 (2008): 37–101 DOI 10.1007/698_5_093 © Springer-Verlag Berlin Heidelberg Published online: 6 November 2007

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology Jelena Radjenovi´c1 (u) · Marin Matoˇsi´c2 · Ivan Mijatovi´c2 · Mira Petrovi´c1,3 · Damià Barceló1 1 Department

of Environmental Chemistry, IIQAB-CSIC, c/ Jordi Girona 18–26, 08034 Barcelona, Spain [email protected] 2 Faculty of Food Technology and Biotechnology (PBF), Pierottijeva 6, Zagreb, Croatia 3 Institució

Catalana de Reserca i Estudis Avanzats (ICREA), Barcelona, Spain

1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

40

2

Membrane Classiﬁcation . . . . . . . . . . . . . . . . . . . . . . . . . . . .

40

3

Types of Membrane Bioreactor Conﬁgurations . . . . . . . . . . . . . . . .

41

4 4.1 4.2 4.3

Hydraulics of Membrane Bioreactor General . . . . . . . . . . . . . . . . Membrane Fouling . . . . . . . . . Methods to Control Fouling . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

45 45 47 52

5 5.1 5.2 5.3 5.4

Biological Performance of Membrane Bioreactor . . Microbiological Aspects . . . . . . . . . . . . . . . . . Nitriﬁcation/Denitriﬁcation and Phosphorus Removal Removal of Organic Matter and Suspended Solids . . Bacteria and Virus Removal . . . . . . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

54 54 58 62 65

6 6.1 6.2 6.3 6.4

Removal of Trace Organic Compounds by a Membrane Bioreactor Removal of Pharmaceutically Active Compounds . . . . . . . . . . . Removal of Hormones . . . . . . . . . . . . . . . . . . . . . . . . . Removal of Surfactants and Their Degradation Products . . . . . . Removal of Sulfonated Organic Compounds, Pesticides, Musk Fragrances and Other Micropollutants . . . . . . .

. . . .

. . . .

. . . .

. . . .

69 70 77 80

. . . .

82

7

Advantages and Drawbacks of MBR Technology . . . . . . . . . . . . . . .

85

8

Application and Cost Analysis of a Membrane Bioreactor . . . . . . . . .

90

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

94

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

Abstract The development and application of a membrane bioreactor (MBR) for fullscale municipal wastewater treatment is the most important recent technological advance in terms of biological wastewater treatment. The MBR is a suspended growth-activated sludge system that utilizes microporous membranes for solid/liquid separation instead of secondary clariﬁers. It represents a decisive step forward concerning efﬂuent quality by

38

J. Radjenovi´c et al.

delivering a hygienically pure efﬂuent and by exhibiting a very high operational reliability. Advanced MBR wastewater treatment technology is being successfully applied at an ever-increasing number of locations around the world. In this chapter, the authors have covered several aspects of MBR, with an exhaustive overview of its operational and biological performance. Different conﬁgurations and hydraulics of MBR are presented, with attention given to the fouling phenomenon and strategies for reducing it. Also, the high quality of MBR efﬂuent is discussed, whereas in comparison with CAS removals of organic matter, ammonia, phosphorus, solids, bacteria and viruses are signiﬁcantly enhanced. Emphasis has been given to the improved capability of MBR to remove organic contaminants present at trace concentration levels (ng L–1 , µg L–1 and mg L–1 ), which to the authors’ knowledge represents a ﬁrst attempt to summarize the published literature on this subject. Finally, advantages and disadvantages of MBR over CAS are concerned. In conclusion, MBR represents an efﬁcient and costeffective process that copes excellently with the growing needs for transforming wastewater into clean water that can be returned to the hydrological cycle without detrimental effects. Keywords Biological performance · Membrane bioreactor (MBR) · Trace organic pollutants · Virus removal · Wastewater treatment

Abbreviations ABT Aminobenzothiazole AEO Alcohol ethoxylates AHTN 6-acetyl-1,1,2,4,4,7-hexamethyltetraline AOB Ammonium oxidizing bacteria APs Alkylphenols APEOs Alkylphenol ethoxylates BOD Biological oxygen demand BPA Bisphenol-A BT Benzothiazole CAS Conventional activated sludge COD Chemical oxygen demand 2,4-D 2,4-dichlorobenzoic acid DBPs Disinfection by-products DEET Meta-N,N-diethyl toluamide DO Dissolved oxygen 2,4-DP 2-(2,4-dichlorophenoxy) propionic acid E1 Estrone E2 17β-estradiol EBPR Enhanced biological phosphorus removal ED Electrodialysis EDCs Endocrine disrupting compounds EDI Electro deionization EDTA Ethylenediamino tetraacetate EE2 17α-ethinylestradiole EPS Extracellular polymeric substances F/M Food-to-microorganism ratio FS Flat sheet

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology GAOs HF HHCB HRT LAS MBR MBT MCPA MCPP MF MLSS MLVSS MTBT MWCO NF NOB NP NPECx NPEOx NSAx OHBT OP OPEOs P3 PAC PAHs PAN PAOs PE PES PhACs pI PP PVDF RO SMP SPC SRT SS SVI TC TCEP TCPP TKN TN TSS UF VSS WWTP

Glycogen accumulating organisms Hollow ﬁber 1,3,4,6,7,8-hexahydro-4,6,6,7,8,8-hexamethylcyclopenta-gamma-2-benzopyran Hydraulic retention time Linear alkylbenzene sulphonates Membrane bioreactor Mercaptobenzothiazole (4-chloro-2-methylphenoxy)-acetic acid Mecoprop Microﬁltration Mixed-liquor suspended solids Mixed-liquor volatile suspended solids Methylthiobenzothiazole Molecular weight cut-off Nanoﬁltration Nitrate oxidizing bacteria Nonylphenol (x = 1, 2) nonylphenol mono, di-carboxylates (x = 1, 2, 3) nonylphenol mono, di, tri-ethoxylates (x = 1, 2) naphthalene mono, di-sulfonates Hydroxybenzothiazole Octylphenol Octylphenol ethoxylates persistent polar pollutants Powdered activated carbon Polycyclic aromatic hydrocarbons Polyacrylonitrile Phosphate accumulating microorganisms Polyethylene Polyethylsulphone Pharmaceutically active compounds Iso-electric point Polypropylene Polyvinylidene diﬂuoride Reverse osmosis Soluble microbial products Sulfophenyl carboxylates Solids retention time Suspended solids Sludge volume index Total Coliform Tris-(2-chloroethyl) phosphate Tris-(chloropropyl) phosphate Total Kjeldahl nitrogen Total nitrogen Total suspended solids Ultraﬁltration Volatile suspended solids Wastewater treatment plant

39

40

J. Radjenovi´c et al.

1 Introduction Membrane bioreactor (MBR) technology, which combines biological-activated sludge process and membrane ﬁltration has became more popular, abundant, and accepted in recent years for the treatment of many types of wastewaters, whereas the conventional activated sludge (CAS) process cannot cope with either composition of wastewater or ﬂuctuations of wastewater ﬂow rate. MBR technology is also used in cases where demand on the quality of efﬂuent exceeds the capability of CAS. Although MBR capital and operational costs exceed the costs of conventional process, it seems that the upgrade of conventional process occurs even in cases when conventional treatment works well. It can be related with increase of water price and need for water reuse as well as with more stringent regulations on the efﬂuent quality. Along with better understanding of emerging contaminants in wastewater, their biodegradability, and with their inclusion in new regulations, MBR may become a necessary upgrade of existing technology in order to fulﬁll the legal requirements in wastewater treatment plants (WWTPs). The idea for coupling the activated sludge process and membrane separation was ﬁrstly reported by research conducted at Rensselaer Polytechnic Institute, Troy, New York, and Dorr-Oliver, Inc. Milford, Connecticut, US [1, 2]. The ﬁrst MBR installation (Membrane Sewage System-MST) was made by Dorr-Oliver, Inc., with ﬂat sheet ultraﬁltration plate and frame membrane. It did not gain much interest in North America but it had considerable success in Japan in the 1970s and 1980s. Before the 1990s, most of the installed MBRs were used for industrial water treatment. With the development of submerged membranes, ﬁrstly introduced by Yamamoto et al. [3], the number of MBRs treating municipal wastewater increased while the MBR market is currently experiencing accelerated growth. The global MBR market doubled over the last 5-year period and in 2005 it has reached a market value of $217 million in 2005 with a projection for the year 2010 of $360 million [4]. The MBR process can be conﬁgured in many different ways depending on project-speciﬁc nutrient removal objectives. Anoxic zones before or after the aerobic treatment may be used for denitriﬁcation, depending on the efﬂuent nitrate and total nitrogen requirements. Anaerobic zones may be used to achieve enhanced biological phosphorus removal in any of its possible conﬁgurations.

2 Membrane Classification The membrane process is a very important separation process in water and wastewater technology, which becomes increasingly competitive and is supe-

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

41

rior to the traditional water technology with proven performance and process economics. The most widely applied membrane separation processes are microﬁltration (MF), ultraﬁltration (UF), nanoﬁltration (NF), reverse osmosis (RO), electrodialysis (ED) and electro deionization (EDI), whereas the ﬁrst four processes produce permeate and concentrate. The separation ranges are as follows: 100 to 1000 nm for MF, 5 to 100 nm for UF, 1 to 5 nm for NF, and 0.1 to 1 nm for RO. Firstly, application of membrane-based technologies in wastewater treatment was focused on tertiary treatment of secondary efﬂuent, so as to obtain a high-quality ﬁnal efﬂuent that can be reused for different purposes. However, over the past 10 years, MBRs have emerged as an effective secondary treatment technology, whereas membranes applied are usually in the range of those of MF and UF. Membranes are usually made from different plastic and ceramic materials, but metallic membranes also exist. The most widely used materials are celluloses, polyamides, polysulphone, charged polysulphone and other polymeric materials such as polyacrylonitrile (PAN), polyvinylidene diﬂuoride (PVDF), polyethylsulphone (PES), polyethylene (PE), and polypropylene (PP). All of these polymeric materials have a desirable chemical and physical resistance. They are also hydrophobic, and it is known that hydrophobic membranes are more prone to fouling than hydrophilic ones due to the fact that most interactions between the membrane and the foulants are of hydrophobic nature [5, 6]. All commercially available membranes are therefore modiﬁed by chemical oxidation, organic chemical reaction, plasma treatment, or by grafting to achieve more hydrophilic surface. This modiﬁcation process usually differs one membrane from another together with the method of fabrication of the membrane module.

3 Types of Membrane Bioreactor Configurations Membrane separation is carried out either by pressure-driven ﬁltration in side-stream MBRs (Fig. 1) or with vacuum-driven membranes immersed directly into the bioreactor, which operates in dead-end mode (Fig. 2) in submerged MBRs. The more common MBR conﬁguration for wastewater treatment is the latter one, with immersed membranes, although a side-stream conﬁguration is also possible, with wastewater pumped through the membrane module and then returned to the bioreactor. The energy consumption required for ﬁltration in submerged MBR is signiﬁcantly lower (Table 1). Both conﬁgurations need a shear over the membrane surface to prevent membrane fouling with the constituents of mixed liquor. Side-stream MBRs provide this shear through pumping, as with most other membrane processes, whereas immersed processes employ aeration in the bioreactor to provide it. Shear enhancement is critical in promot-

42

J. Radjenovi´c et al.

Fig. 1 Side-stream MBR with external pressure-driven membrane ﬁltration

Fig. 2 Submerged MBR with internal vacuum-driven membrane ﬁltration Table 1 Comparison of ﬁltration conditions for tubular and immersed MBRs [8]

Manufacturer Model Surface area [m2 ] Permeate ﬂux [L m–2 h–1 ] Pressure [bar] Air ﬂow rate [m3 h–1 ] Energy for ﬁltration [kWh m–3 ]

Side-stream tubular membrane

Submerged membrane

Zenon Permaﬂow Z-8 2 50–100 4 – 4–12

Zenon ZeeWeed ZW-500 46 20–50 0,2–0,5 40 0.3–0.6

ing permeate ﬂux and suppressing membrane fouling, but generating shear also demands energy, which is probably the reason for submerged conﬁguration predominance. Also, in side-stream MBR module fouling is more pronounced, due to its higher permeate ﬂux. Pumping of activated sludge induces shear stress to microbial ﬂocs, causing them to break-up, which leads to a decrease in particle size and releasing of foulant material from the ﬂocs [7].

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

43

This may signiﬁcantly promote the membrane fouling rate. In Table 1 ﬁltration conditions are presented for tubular and immersed MBRs [8]. The conﬁgurations of MBR are based on either planar or cylindrical geometry. There are ﬁve principal membrane conﬁgurations currently employed in practice: 1. 2. 3. 4. 5.

Hollow ﬁber (HF) Spiral-wound Plate-and-frame (i.e., ﬂat sheet (FS)) Pleated ﬁlter cartridge Tubular.

In the HF module, large amounts of HF membranes make a bundle, and the ends of the ﬁbers are sealed in epoxy block connected with the outside of the housing (Fig. 3). The water can ﬂow from the inside to the outside of the membrane, and also from the outside to the inside, which is produced differently by different manufacturers. These membranes can work under pressure and under vacuum (Fig. 4). The spiral-wound conﬁguration is mostly used for the NF and RO process. The membranes are wound around the perforated tube through which permeate goes out (Fig. 5). The spiral-wound modules are manufactured in standard dimensions by all major manufacturers, which makes their installation easier and membrane production less costly. Many membrane modules can be installed together in series or parallel in plants with high capacity (Fig. 6). Plate-and-frame membrane modules comprise of FS membranes with separators and/or support membranes. The pieces of these sheets are clamped onto a plate. The water ﬂows across the membrane and permeate is being collected through pipes emerging from the interior of the membrane module in a process that operates under vacuum (Fig. 7). There are also membrane conﬁgurations such as plated ﬁlter cartridge and tubular module, but they are not so widely used as the other three mod-

Fig. 3 Hollow-ﬁber (HF) membrane module

44

J. Radjenovi´c et al.

Fig. 4 Hollow-ﬁber (HF) immersed membrane module (Zenon, Canada) ﬁltrating activated sludge under vacuum

Fig. 5 Spiral-wound membrane part

ules. Typically, tubular membranes are encased in pressure vessels and mixed liquor is pumped to them and they are predominantly used for side-stream conﬁgurations. HF and FS modules are mostly immersed directly in mixed liquor with permeate drawn through the membranes using vacuum pumps. In the case of HF membranes, use of 0.8 mm to 1.5 mm ﬁne screen upstream of membranes is recommended to protect the membranes from hair

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

45

Fig. 6 Connection of the spiral-wound elements

Fig. 7 Plate-and-frame immersed membrane module (Kubota, Japan)

and other stringy materials that can result in excessive cleaning frequencies. A ﬁne screen of 2–3 mm is usually employed for FS membrane systems.

4 Hydraulics of Membrane Bioreactor 4.1 General During MBR wastewater treatment, solid–liquid separation is achieved by MF or UF membranes. The basic principle is that the feed water passes over the membrane surface and the product is called permeate, whereas the re-

46

J. Radjenovi´c et al.

Fig. 8 Basic principle of membrane ﬁltration

jected constituents form concentrate or retentate (Fig. 8). A membrane is simply a two-dimensional material used to separate components of ﬂuids usually on the basis of their relative size or electrical charge. The capability of a membrane to allow transport of only speciﬁc compounds is called semipermeability. This is a physical process, where separated components remain chemically unchanged. Components that pass through membrane pores are called permeate, while rejected ones form concentrate or retentate. Mass balance of the solute in the process can be presented by the equation: Qf cf = Qp cp + Qc cc ,

(1)

where Qf – feed ﬂow rate; cf – solute concentration in feed ﬂow; Qp – permeate ﬂow rate; cp – solute concentration in permeate; Qc – solute concentration in concentrate; cc – solute concentration in concentrate. Membrane rejection of solutes can be calculated according to the following equation: cf – cp R= , (2) cf where Cf represents concentration of solute in feed ﬂow and Cp represents its concentration in permeate. The fraction of feed ﬂow converted to permeate is called yield, recovery or water recovery (S). Water recovery of the membrane process is given with the equation: Y=

Qp , Qf

where Qp is the permeate ﬂow and Qf is the feed ﬂow.

(3)

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

47

Recovery is normally close to 100% for dead-end ﬁltration, while it varies signiﬁcantly for cross-ﬂow ﬁltration depending on the nature and design of membrane process. Permeate ﬂux (usually denoted as J) is the volume of water passed through a unit area of membrane per unit of time and it is often normalized to a standard temperature. The common unit for J is usually L m–2 day–1 , and most of the available data for MBR is given in that manner rather than in SI units. MBR membranes generally operate at ﬂuxes between 10 and 100 L m–2 h–1 . The ﬂux is related to its driving force which is transmembrane pressure (TMP or ∆P) while the membrane performance can be estimated from the membrane permeability (K), which is calculated as permeate ﬂux per unit of TMP and is usually given as L m–2 h–1 bar–1 . 4.2 Membrane Fouling A decrease in the permeate ﬂux or increase in TMP during a membrane process is generally understood by the term “fouling”. Fouling occurs as a consequence of interactions between the membrane and the mixed liquor, and is one of the principal limitations of the MBR process. There has been done a lot of research on this subject [9], so it is of interest to describe the fouling in more details. Fouling of membranes in MBRs is a very complex phenomenon with diverse relationships among its causes, and it is very difﬁcult to localize and deﬁne membrane fouling clearly. The main causes of membrane fouling are: 1. 2. 3. 4.

Adsorption of macromolecular and colloidal matter Growth of bioﬁlms on the membrane surface Precipitation of inorganic matter Aging of the membrane

As a measure of fouling, resistance (R), which is inversely related to K, is often used. R is given by: R=

∆P , ηJ

(4)

where η stands for permeate viscosity in (kg m–1 s–2 ). This total ﬁltration resistance consists of a number of components, which can be divided as: membrane resistance, resistance of the fouling layer on the membrane surface, and resistance offered by the membrane-solution interfacial region. Membrane resistance is a function of characteristics of membrane material such as thickness and pore size, and it determines the ﬂux through the membrane for ﬁltration of one-component liquid, i.e., clean water. For MBRs, membrane resistance is often given as its inverse value called “clean water permeability”, which is normally within the range of a few hundred to a few thousand L m–2 h–1 bar–1 . The resistance of the fouling layer is associated with the ﬁl-

48

J. Radjenovi´c et al.

tration mechanism, which is dependent on the membrane characteristics and characteristics of ﬁltered solids, while membrane-solution interfacial region resistance is associated with concentration polarization. Concentration polarization is a phenomenon of solute tendency to accumulate within the boundary liquid layer of near-stagnant ﬂuid adjacent to membrane surface. Since liquid velocity within this layer is close to zero, the only mode of mass transport is diffusion, which is signiﬁcantly slower then convective transport in the bulk solution. As a consequence, resistance to ﬁltration occurs. The thickness of the boundary layer is dependent on system’s hydrodynamics and can be decreased by promoting the turbulence of liquid ﬂow.

Fig. 9 Fouling mechanisms

According to recognized mechanisms (Fig. 9), the fouling on the membrane occurs as: 1. Complete blocking caused by occlusion of pores by the particles with no particle superimposition 2. Intermediate blocking caused by occlusion of pores by particles with particle superimposition 3. Standard blocking where particles smaller than the membrane pore size deposit onto the pore walls thus reducing the pore size 4. Cake ﬁltration where particles larger than the membrane pore size deposit onto the membrane surface Many authors tried to divide total ﬁltration resistance (Rt ) on three components [5, 10–13] (Eq. 5): Rt = Rm + Rc + Rf ,

(5)

where Rm is the membrane resistance, Rc is the cake resistance and Rf is the fouling resistance. It is assumed that fouling consists of two separate processes, one being cake fouling caused by suspended particles that form a cake layer on the membrane surface, and the other type associated with adsorption of smaller colloid and soluble matter on the membrane surface and in the membrane pores. To therefore determine the divided resistances, the measurement of Rm with clean water and Rt with activated sludge is done. Then another subsequent clean water ﬁltration with fouled membrane is performed, which gives Rm + Rf value and the Rc can then be calculated by

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

49

subtraction. This model neglects the coupling of synergistic effects on fouling, which may occur among different activated sludge components. Similar to the above approach, in an attempt to make a distinction among different constituents of activated sludge, relative contribution of biomass fractions on fouling has been intensively investigated by fractionation of activated sludge into suspended solids, colloids, and solutes. This fractionation is usually performed by centrifugation of biomass and subsequent ﬁltration of the supernatant through 0.45 µm ﬁlter for separation of colloid from soluble matter. Again, this approach does not take into account all the interactions among fractions, but it represents an interesting approach to clariﬁcation of the fouling phenomenon. The results of these studies, recently reviewed by Judd [4], vary signiﬁcantly in assessment of the relative contributions of components to fouling. Despite variations in the published results, it is generally accepted that fouling caused by suspended solids is less than that of supernatant. Moreover, Chang et al. [13] reported a faster fouling rate with the efﬂuent of biological process than with the activated sludge and concluded that suspended solids can act as a dynamic layer over the membrane surface, which can slow the penetration of soluble fouling species that cause fouling. Ng et al. [14] also reported greater ﬁltration resistance from mixed-liquor supernatant than from the biomass. With regards to the fouling mechanism, it is generally assumed that soluble and colloid materials are responsible for pore blocking, while suspended solids contribute mainly to cake layer resistance. Again, one has to be aware of the fact that biomass itself is responsible for a composition of soluble and colloid material in the liquid phase, and that composition of supernatant is a function of the physiological state of the suspended phase, i.e., biomass. Among the constituents of the supernatant, extracellular polymeric substances (EPS) have gained a lot of attention as possible important foulants in MBR [5, 15, 16]. EPS is a general term for various types of macromolecules found outside the cell surface, most commonly carbohydrates and proteins. They are normally produced by microorganisms as a construction material necessary for the development of microbial aggregates, such as bioﬁlms or ﬂocs, or used as a protective barrier around the bacteria. Apart from the EPS that are bound in microbial ﬂocs, EPS can be found in the water phase as free EPS. Substances in this category originate from the break up of ﬂocs and cell lyses, or they can be introduced by the inﬂuent. Another group of substances overlapping the EPS is called “soluble microbial products” (SMP). This group contains a wider range of substances that can also be deﬁned as soluble EPS. The presence of EPS is desirable in CAS because it helps formation of microbial ﬂocs and makes them easier to settle, but due to its heterogeneous nature EPS can form hydrated gel which can act as a barrier to permeate ﬂow in MBR. Little is known about the circumstances that inﬂuence EPS production and their possible release to the water phase. Many operating parameters including substrate composition and organic loading

50

J. Radjenovi´c et al.

rate appear to affect EPS, with solid retention time (SRT) probably being the most signiﬁcant factor [17]. Generally taken, the EPS level may be lower when fewer disturbances, such as starvation, oxygen or essential nutrient depletion, are introduced to the biomass. Rosenberger et al. [18] found a correlation between polysaccharide concentration in the supernatant of mixed liquor and high fouling rates of submerged membrane. The concentration of polysaccharides of microbiological origin was inﬂuenced by temperature and stress situations to microorganisms. Fouling can be divided from the practical point of view on: 1. Reversible fouling that can be removed from the membrane by physical cleaning 2. Irreversible fouling removed by chemical cleaning 3. Irrecoverable fouling that cannot be removed by any cleaning Physical cleaning in MBRs is normally achieved either by back-ﬂushing or by relaxation (stopping the permeate ﬂow and continuing to scour the membrane with air bubbles). Physical cleaning is a simple and short method (usually lasting less then 2 min) of fouling suppression which demands no chemicals and generally it is less likely that it will affect the membrane material. The latest published data [4] on the cleaning regime of full-scale plants suggests that presently most of MBR facilities use relaxation rather than backﬂushing. However, by means of physical cleaning it is not possible to remove all the material deposited on the membrane. Chemical cleaning is a more effective method, which is able to remove more strongly the adsorbed deposits. Chemical cleaning is carried out mostly with sodium hypochlorite and sodium hydroxide for organic deposits removal, or with acidic solutions for removal of lime or other inorganic deposits. Cleaning is performed by soaking the membrane in the cleaning solution or by adding the cleaning agent into the back ﬂush water. Most MBRs employ chemical maintenance cleaning on a weekly basis, which lasts 30–60 min, and recovery cleaning when ﬁltration is no longer sustainable, which occurs once or twice a year. Deposits that cannot be removed by available methods of cleaning is called “irrecoverable fouling”. This fouling builds up over the years of operation and eventually determines the membrane life-time. Development of the fouling given as pressure transient for these three types of fouling for the continuous operation are presented on Fig. 10. MBRs are normally operated under a constant ﬂux. Since the fouling rate increases roughly and exponentially with the ﬂux, MBR plants operate at modest ﬂuxes and preferably below the so-called critical ﬂux. The critical ﬂux concept, ﬁrstly introduced by Field et al. [19], assumes that in MF/UF processes exists a ﬂux below which a decline of permeability with time does not occur, and above which fouling occurs. In MBR operations, critical ﬂux is normally deﬁned as the highest ﬂux under which a prolonged ﬁltration with constant permeability is possible. Critical ﬂux is often determined by the ﬂux-

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

51

Fig. 10 Fouling rates for different types of fouling (tphys – duration of physical cleaning cycle, tchem – duration of chemical cleaning cycle)

step method, in which the ﬂux is incrementally increased in number of steps with ﬁxed duration, and the increase in TMP is recorded. It is then possible to observe the apparent ﬂux where fouling occurs, observed as a signiﬁcant TMP increase or deviation in linearity of K. The critical ﬂux deﬁnition in its strong form demands that K obtained during ﬁltration in sub-critical conditions equals K obtained during clean water ﬁltration. It is possible to achieve such results when ﬁltration media has deﬁned characteristics regarding the size of the particles. In MBR operation, however, due to the complexity of the mixed liquor, some irreversible fouling constantly occurs, which makes it impossible to achieve the sub-critical conditions as for the strong form of the critical ﬂux. Nevertheless, this concept is widely accepted in MBR operations as a tool that can provide useful guide value for the appropriate operating ﬂux. Pollice et al. [20] reviewed the sub-critical fouling phenomenon in MBR. From the reviewed data it is evident that even sub-critical operation inevitably leads to fouling. This fouling is often reported to follow a two-stage fouling pattern [21, 22], which includes slow TMP increase over a long period of time, followed by a rapid increase after some critical time period. In the work of Zhang et al. [23], this pattern is extended with an initial period of conditioning fouling. In cited work, fouling in MBR under sub-critical conditions three stages are introduced, which include: 1. Initial conditioning fouling 2. Slow fouling 3. Sudden TMP jump

52

J. Radjenovi´c et al.

During the initial conditioning fouling reported also by Ognier et al. [24] and Jiang et al. [25] interactions take place between the membrane surface and soluble components of the mixed liquor. This fouling is usually rapid (measured in hours), irreversible by nature, and occurs even for zero ﬂux operation [23]. In the second stage, slow fouling, the membrane surface is gradually covered by biopolymers such as EPS, which changes the properties of the membrane surface and makes attachment of the microbial ﬂocs to the membrane surface easier. Thus, bioﬁlm growth on the membrane surface may be promoted. Over time, complete or partial pore blocking takes place. This blocking is expected to be inhomogeneous since the air and the liquid ﬂow are distributed unevenly in MBR. With regions of membrane more fouled than others, ﬂux locally varies, thus exceeding the critical ﬂux in some areas of the membrane surface, which then leads to a sudden TMP jump characteristic for operation above the critical ﬂux. The other explanation for the sudden TMP jump may be the change of properties of the fouling cake on the membrane surface due to its compression. 4.3 Methods to Control Fouling To control the fouling that inevitably occurs in MBR operation, several key parameters can be modiﬁed. The most important strategies are concentration polarization suppression, optimization of physical and chemical cleaning protocols, pre-treatment of feed wastewater, and mixed-liquor modiﬁcation. Fouling related to concentration polarization can be reduced either by promoting turbulence or by reducing ﬂux. As mentioned above, both MBR conﬁgurations need shear over the membrane surface to prevent this type of fouling. As with most other membrane processes, side-stream MBRs provide shear through pumping, which increase cross-ﬂow velocity, whereas immersed processes employ aeration around the membrane to provide shear stresses. Aeration intensity over the submerged membrane surface is recognized as the key operational parameter in preventing cake formation on the membrane surface in the submerged conﬁguration [26–28]. Membrane permeability and critical ﬂux have been shown to increase roughly linearly with aeration rate, up to a level above which no further increase is observed. Liu et al. [29] investigated critical aeration intensities for different sludge concentration and ﬂuxes, drawing a quantitative correlation between those parameters. However, increasing membrane aeration is normally prohibitively expensive. Since membrane aeration contributes signiﬁcantly to the energy demand, much development has been focused on reducing aeration whilst maintaining membrane permeability. A development has been achieved in aeration efﬁciency with new jet aeration and cyclic aeration systems. It is often in practice to use different aeration systems for biological system and for

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

53

membrane fouling control, in order to insure most efﬁcient energy use for both systems demands. Reducing the permeate ﬂux always reduces fouling, but at the same time this strategy demands a more installed membranes, which then contributes to the capital cost of MBR installation. In other words, ﬂux can be maintained below the critical value to ensure stable operation with little or negligible increase in TMP, thus decreasing cleaning frequency and consumption of chemicals, or total installed membrane area can be reduced on the behalf of frequent cleaning. The latter strategy is called intermittent operation. In practice, most submerged MBRs treating municipal wastewater operate at net ﬂuxes of 20–30 L m–2 h–1 with the relaxation period every 10 min and periodical maintenance chemical cleaning every few months. As mentioned above, pre-treatment of feed wastewater through screening is necessary for both HF and FS membrane modules. HF membranes have a tendency for aggregates of hair and other debris to collect at the top of the membrane elements. These aggregates usually cannot be signiﬁcantly removed by back-ﬂushing. FS modules are somewhat less prone to such clogging, but they too need a feed-water pre-treatment, though with coarser screens. The properties of the mixed liquor that affect the fouling propensity can also be altered to minimize fouling. The production and release of EPS could be inﬂuenced by changing the biological state of biomass, usually by SRT modiﬁcation as mentioned above, but it is rarely done in practice. To decrease the ESP or SMP concentration, mixed liquor is mostly modiﬁed through addition of chemicals. The use of ﬂocculants and coagulants such as aluminum or ferric chloride has been investigated in an attempt to minimize fouling [10, 30]. Also, the addition of adsorbent reagents such as powdered activated carbon (PAC) has been found to improve the membrane performance by decreasing the level of organic compounds with potential for membrane fouling [14, 31]. Recently, a commercial product, a cationic polymer-based compound called MPE50 manufactured by Nalco Company has been developed and tested at full-scale to enhance membrane performance. The cleaning protocol is mainly dictated by designed operational net ﬂux as explained above. Usually the protocol suggested by the manufacturer is followed as a guideline, and the majority of the installed plants work in the sub-critical regime. However, cleaning protocol has been studied intensively by many researchers where the key parameters of interest are duration and frequency of the cleaning and the back-ﬂush ﬂux. Less frequent, longer back-ﬂushing (600 s ﬁltration/45 s back-ﬂushing) has been found to be more efﬁcient then more frequent but shorter back-ﬂushing (200 s ﬁltration/15 s back-ﬂush) [25]. To optimize the back-ﬂush duration, Smith et al. [32] developed a generic control system based on TMP monitoring. Membrane relaxation, which is the most common practice for fouling control, encourages the diffusive back-transport of foulants away from the membrane surface, which is enhanced by air scouring. Relaxation allows longer ﬁltration periods

54

J. Radjenovi´c et al.

between chemical cleanings, and despite some reports that it may not be a feasible practice for large-scale MBRs [33], it is widely used in practice. Although intensive research has been done on this subject, membrane fouling in MBRs needs further attention in order to understand the complex interactions among biologically active and constantly changing ﬁltration media, hydrodynamic conditions of the ﬁltration process, and the membrane itself.

5 Biological Performance of Membrane Bioreactor 5.1 Microbiological Aspects In the biochemical stage of wastewater treatment, organic carbon and nutrients are removed from wastewater by microbes. These microbes live and grow enmeshed in EPS that bind them into discrete micro-colonies forming three-dimensional aggregated microbial structures called ﬂocs. The ability of microorganisms to form ﬂocs is vital for the activated sludge treatment of wastewater. The ﬂoc structure enables not only the adsorption of soluble substrates but also the adsorption of the colloidal matter and macromolecules additionally found in wastewaters [34, 35]. The diversity of microbial community in activated sludge is very large, containing prokaryotes (bacteria), eukaryotes (protozoa, nematodes, rotifers), and viruses. In this complex microsystem, bacteria dominate the microbial population and play a key role in the degradation process [34]. MBR technology with biochemical and sludge-separation stage integrated into one step implies a continuous generation of new sludge with the consumption of feed organic materials, while some sludge mass is decayed by endogenous respiration. Endogenous respiration involves consumption of cell-internal substrate, which leads to a loss of activity and slightly reduced biomass. Endogenous respiration implies all forms of biomass loss and energy requirements not associated with growth by considering related respiration under aerobic conditions: decay, maintenance, endogenous respiration, lyses, predation, and death. It can be both aerobic and anoxic, though under anoxic conditions it is a lot slower and especially protozoa are considerably less active under denitrifying conditions (slower predation) [36]. Endogenous respiration of a microbial community in MBR can be encouraged by very high sludge age, i.e., high sludge concentration. The energy available to microorganisms is determined by the supply of substrate. By increasing the SRT, which increases biomass concentration, it would be theoretically possible to reach a situation where the amount of energy provided equals the maintenance demand. This concept was ﬁrst introduced

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

55

by Pirt [37], where maintenance energy is deﬁned as the amount of biochemical energy strictly necessary for sludge endogenous respiration. Microorganisms satisfy their maintenance energy requirements in preference to producing additional biomass. Therefore, under the conditions of decreased nutrient supply, external substrate is used only for the upkeep of bacterial vial functions, while the amount of bacteria is not changed. Moreover, the higher the biomass concentration, the lower the sludge loading, i.e., foodto-microorganism (F/M) ratio (g COD gTSS–1 day–1 ) becomes [38]. When the sludge loading becomes low enough, little or no excess sludge is produced [3, 39–42]. Low et al. [41] reported a sludge production reduced by 44% when increasing a biomass concentration in an MBR from 1.7 to 10.3 g L–1 . In various studies on applications of MBR in wastewater treatment, zero sludge production was established at different F/M ratios, obviously depending on feed compositions which determine the growth of microbial populations [43–45]. However, there is an optimal biomass concentration (i.e., SRT) for a successful operation of MBR. Biomass retention results in a slow-growing population with high sludge ages, where cell dormancy and death reduce the viability of population [46, 47]. An example of changes in sludge yield and biomass concentration with sludge age are presented on a Fig. 11 (HRT— hydraulic retention time, k-rate constant for endogenous metabolism, kd -rate constant for biomass decay) [48]. Several explanations are suggested for this phenomenon. Since MBR sludge acts as a non-Newtonian ﬂuid by increasing the mixed-liquor suspended solids (MLSS) concentration, the viscosity of sludge increases exponentially. This results in mass transfer limitations for both the oxygen and substrate, which increases aeration costs as well as causing extensive membrane fouling [49]. On the other side, at lower MLSS concentrations, more speciﬁc surface area is available for the uptake of a substrate and enzyme production, and the enzymatic activity is higher. Thus, when operating at

Fig. 11 Net observed yield · · · and biomass concentration as a function of sludge age in a MBR, HRT=2.7 h, Y=0.4, k = 0.07 d–1 , kd = 0.06 d–1 [48]

56

J. Radjenovi´c et al.

low SRT response of the system in degradation of xenobiotic waste should be faster. Moreover, the chances of genetic mutation and adaptation of microorganisms to different organic loadings should be greater [46]. Horan et al. [50] also noted that at high sludge ages the solubility of substrate becomes ratelimiting. Ng et al. [51] studied the performance of MBR at low SRT (0.25–5 days). They indicated that modiﬁcation of sludge morphology, i.e., proliferation of non-ﬂocculating microorganisms, could have a positive impact on removal performance. In addition, recent works of Wilen et al. [52] showed that the surface properties and the structure of biological ﬂocs in activated sludge are correlated to the chemical constituents of EPS, and can be signiﬁcantly inﬂuenced by the operating condition. However, some investigations have given completely opposite results. Massé et al. [53] observed a decrease in ﬂoc size at higher SRTs. This could be due to lower production of EPS, which is responsible for the formation of ﬂocs or other cell aggregates. Moreover, growth of non-ﬂocculating bacteria is enhanced because they are more exposed to the present substrate than when they are arranged into macro-ﬂocs. Some authors believe that there should be a minimal rate of sludge wasting in order to keep an optimal range of sludge concentration in MBR [46, 54– 56]. When no sludge is withdrawn from the reactor, accumulation of inorganic compounds can be expected [16, 56–58]. Retention and accumulation of non-biodegradable compounds in the bioreactor could lead to microbial inhibition or toxicity, which limits the alternatives available for excess sludge disposal. Several works have described a possible negative long-term effect of accumulation of recalcitrant compounds on process stability [40, 59–62]. Non-biodegradable solids (solids that are not metabolized under present conditions) are either present in the inﬂuent or they are produced in the microbial process. Their forming can also be a result of protozoan activity, which may not degrade the bacterial cell walls fully, leaving behind the inert material. However, inerts are not ultimately inert: it is possible that degradation of inert material occurs by slow-growing bacteria, which will depend on the SRT [63]. Many studies have reported a stable performance of MBR during long operating periods, with a dynamic balance of active biomass and inorganic fraction during long-term operating periods [40, 58, 64]. In these studies, the mixed-liquor volatile suspended solids (MLVSS) to MLSS (MLVSS/MLSS) ratio was used as an indicator of the amount of viable sludge, and it was found to be relatively constant. Pollice et al. [65] explained this phenomenon by a possible hydrolysis and/or enzymatic solubilization of inert matter. Taking into account all of the above-mentioned aspects of SRT control in MBR, SRT should be chosen in such a way to avoid both the adverse effects of accumulated non-biodegradable substances resulting from low sludge discharge and also an excessive production of sludge at low sludge ages. High sludge ages are one of the main advantages of MBR, consider-

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

57

ing that in conventional treatment processes long SRTs are impossible because of bad settling ability of sludge at high concentration and withdrawal of suspended solids with the efﬂuent. Typical values for MLSS concentration in MBR vary from 10 to 25 g MLSS L–1 , while in CAS they are around 1.5–5 mg MLSS L–1 [44, 66]. Besides the “prolonged SRT” strategy, sludge decay rate in MBR could be boosted by disintegration of some part of sludge. The most common way for achieving this is sludge lyses. Lyses imply death and the breaking apart of cells, and therefore loss of bacteria. The autochthonous substrate formed contributes to the organic loading and is reused in microbial metabolism. Since the biomass growth on this substrate cannot be distinguished from the growth on the original organic substrate, it is called cryptic growth, and it was ﬁrst introduced by Ryan et al. [67]. Limiting step for cell lyses is the degradation of the cell wall, and in order to accelerate it, physical or chemical treatment can be used [68]. Canales et al. [69] managed to improve the endogenous metabolism in an MBR by inducing cell death and lyses with a thermal treatment. Biomass was extracted and treated at three different temperatures (50, 70, and 90 ◦ C) (see Fig. 12), while the hydrolysates were recycled to the bioreactor. Thus, improvement of endogenous metabolism was obtained by cryptic growth with both HRT and SRT were very low (2 and 10 h, respectively). Other techniques for establishing cryptic growth in an MBR are ultrasound disintegration [70], ozone-induced biodegradation [71–73], and alkaline treatment [74]. Experimental results showed that by ozonization, excess sludge production could be reduced signiﬁcantly, and biological performance of mineralization and nitriﬁcation would not be inhibited [71]. The ozone

Fig. 12 Death kinetics of P. Fluorescens at different temperatures [69]

58

J. Radjenovi´c et al.

dosing rate is directly proportional to the amount of sludge to be treated. For example, in a study of Sakai et al. [72], it was found to be 0.034 kg O3 per kg of suspended solids (SS). On the other side, growth of controllable predators has been successfully tested for reducing excess sludge production in bench and pilot-scale reactors [75]. Grazing fauna mainly consists of protozoa and metazoa. These higher organisms consume bacteria, while decomposition of substrate remains unaffected. During energy transfer from low to high trophic levels, energy is lost due to inefﬁcient biomass conversion. Under optimal conditions, the total loss of energy will be maximal, and the total biomass production thus will be minimal [49]. Environmental factors that inﬂuence and limit microbial growth are temperature and pH value, i.e., the acidity or alkalinity of the aqueous environment. Temperature has a profound effect not only in governing the rate of the treatment but it also affects bacterial composition. Chiemchaisri et al. [76] investigated performances of MBRs at various temperatures and noted a reduction in the number of strict aerobic bacteria when temperature was lowered, suggesting a limited oxygen transfer, partly due to reduced viscosity of mixed liquor at lower temperatures. The temperature range for optimal performance of MBR was found to be from 15 to 25 ◦ C, while the treatment efﬁciency deteriorated as the temperature decreased to 10 ◦ C. As far as pH is concerned, autotrophic metabolism is considered impaired outside the optimal pH range (7.2–8.5) [77]. How to operate MBR systems efﬁciently remains a topic of argument because there is a lack of information on the development of microbial community structure in the reactor [57]. The characteristics of sludge morphology (dispersed bacteria, lower amount of large ﬁlamentous bacteria, ﬂoc densiﬁcation) certainly play an important role in the removal efﬁciencies, but they also affect sludge ﬁlterability and fouling mechanisms. Under the high organic loading conditions (i.e., low SRT), foaming and sludge bulking may rise. In particular, the modiﬁcation of sludge structure induced by membrane separation compared to a settling separation is still unclear. Because MF and UF membrane retain dispersed bacteria as well as colloidal and supracolloidal material, the biological medium in MBR can be signiﬁcantly different from those produced in an activated sludge [78]. It can be assumed that if operated at high sludge ages, bacteria in MBR face conditions of extreme competition for the inﬂowing substrates. However, microbiology and physiology of MBR are far from being understood. 5.2 Nitrification/Denitrification and Phosphorus Removal An irrational use of agricultural fertilizers and pesticides and the discharge of incompletely treated industrial and municipal wastewater results in high ni-

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

59

trogen and phosphate concentration in surface water and groundwater. This enrichment usually leads to an excessive eutrophication of lakes and other water bodies, promoting an excessive growth of certain weedy species. Since both nitrates and phosphates are rate-limiting for the process of eutrophication (extraordinary growth of algae), their removal is of crucial importance for successful wastewater treatment. Nitrate and nitrite-contaminated water supplies are also related to several diseases such as methemoglobinemia occurring in infants, also called “blue baby disease” [79]. Moreover, these two compounds can induce mutations of DNA, causing gastric cancer [80]. Biological nitriﬁcation is an oxic process of conversion of ammonia to nitrite (NO2 ) and then to nitrate (NO3 ). Following nitriﬁcation, nitrogen can be removed from wastewater by reducing nitrate to nitrogen gas (N2 ) in the process of anoxic denitriﬁcation. Because of the low growth rate and poor cell yield of nitrifying bacteria, nitriﬁcation is generally a rate-limiting step in biological nitrogen removal performance. The key requirement for nitriﬁcation to occur is that the net rate of accumulation of biomass (and hence the net rate of withdrawal of biomass from the system) is less than the growth rate of nitrifying bacteria [81]. Long SRTs applied in MBR prevent nitrifying bacteria from being washed out from the bioreactor, improving the nitriﬁcation capability of the activated sludge. Moreover, nitriﬁers are less endangered by faster-growing heterotrophic bacteria, which are better competitors for the ammonia nitrogen (NH3 -N). Many studies have proven that MBR can operate as a high-rate nitrifying technology that can be applied in the nitriﬁcation of wastewater containing a high concentration of ammonia nitrogen [82]. Chiemchaisri [82] and Muller [40, 83, 84] found that more than 80% of the inﬂuent total Kjeldahl nitrogen (TKN) could be nitriﬁed to NO3 in an MBR. On the other hand, the denitriﬁcation process requires anoxic conditions in order to occur. To enhance denitriﬁcation, usually an anoxic tank is added upstream from the aerated tank. Anoxic conditions can also be introduced by operating MBR in an intermittent aeration mode, even when regarding submerged MBR, which needs permanent bubbling. In the intermittently aerated MBR, ammonium is nitriﬁed mostly to nitrate and most phosphates are removed during the aerobic period (aeration), where the accumulated nitrate is completely denitriﬁed during the anoxic period (non-aeration), and phosphorus (P) is taken up. The net P removal is achieved by wasting sludge after the aerobic period when the biomass contains a high level of polyphosphates (polyP) [85]. P is found in wastewater as phosphates (orthophosphates, condensed phosphates, organic phosphate fractions), and it can be eliminated either by precipitation and/or adsorption, or by luxury uptake. Only a small amount of phosphorus is used for cell metabolism and growth (1–2% of the total suspended solids (TSS) mass in the mixed liquor) [86]. Precipitation and adsorption processes require an appropriate pH, the presence of iron or calcium ions, etc. In WWTPs, luxury uptake of P is accomplished by the introduc-

60

J. Radjenovi´c et al.

tion of an anaerobic phase in the wastewater treatment line ahead of the aerobic phase and recycling of sludge through the anaerobic and aerobic phase [87]. Exposing mixed liquor to an anaerobic/aerobic sequence selects phosphate accumulating microorganisms (PAOs) due to a competition between PAOs and other aerobic organisms. This competition mechanism is based on a complete anaerobic uptake of the lower fatty acids by the polyP organisms (i.e., PAOs), which assures that in the aerobic phase, no fatty acids are left. The polyP organisms use the stored internal substrate during aerobic conditions while other aerobic organisms are lacking substrate. This process is usually referred to as the enhanced biological phosphorus removal (EBPR) process. EBPR process can be established in MBR treatment unit by operating it in intermittent aeration mode. Moreover, phosphorus removal will be signiﬁcantly improved in an MBR by a physical retention of PAOs, whose size is typically larger than 0.5 µm. Since an MF membrane (0.2 µm) will act as a physical barrier to retain the PAOs in the reactor, sufﬁcient biomass is provided for the EBPR mechanism to take place [88, 89]. Intermittently aerated MBR can achieve nitrogen and phosphorus removal by a simultaneous nitriﬁcation and denitriﬁcation, P-uptake and P-release in the same reactor in accordance with time cycle of aeration and non-aeration. However, even though intermittent aeration was successful in removing nitrogen, P removal was difﬁcult to achieve at a higher level [83, 90]. This is probably due to the inhibition by nitrate. In the anaerobic stage, nitrate reduces phosphate release, and in the aerobic stage it diminishes its uptake. Denitriﬁcation has more capability than phosphorus release with respect to the competition of substrate [91]. This is because nitrate will be utilized as a ﬁnal electron acceptor in the growth of non-polyP heterotrophs. Thereby, the amount of substrate available for polyP organisms is reduced and hence the removal of phosphorus is lowered. There are some studies that conﬁrm the ability of polyP organisms for denitriﬁcation, however, not all PAOs can use nitrate as an electron acceptor [92]. In addition, intermittently aerated MBR showed an unstable nitrogen removal in its application to treat domestic sewage of rural settlements because of incomplete denitriﬁcation [93]. Stable nitriﬁcation can be maintained as long as the oxygen concentration is high enough [94]. Chiemchaisri et al. [83] found that by lowering the dissolved oxygen (DO) concentration, nitriﬁcation was signiﬁcantly inhibited, although it recovered completely after raising the DO concentration to 1 mg/L. Nitrosomonas and Nitrospira-like bacteria were identiﬁed as the predominant ammonium (AOB) and nitrate-oxidizing bacteria (NOB) in an MBR, respectively [95]. Both of these are obligate aerobic and under anoxic conditions are unable to store or utilize their substrate. The absence of oxygen may provoke stress and damage of their metabolism, NOB being the more sensitive [96]. There is much research on the effect of SRT on MBR performance as far as nitriﬁcation/denitriﬁcation and phosphorus removals are concerned. Ac-

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

61

cording to Cicek et al. [46], there is a decrease in nitriﬁcation rate at very low SRT (2 days), supposedly due to a partial loss of nitrifying microorganisms. On the other side, Li et al. [57] observed a decreasing trend of nitriﬁers when increasing the sludge concentration, i.e., solids retention time. Another study conﬁrmed a negative inﬂuence of long SRT on nitriﬁcation performance [56], which was explained by impeded oxygen and substrate transfer owing to an increase in MLSS concentration. Similar observations on the effect of SRT on phosphorus removal were taken by Yoon et al. [97], who recorded a decreased P removal at long SRT. In this case, it is probably due to a fact that PAOs undergo competitive conditions with glycogen-accumulating organisms (GAOs) at SRTs longer than 20 days [98]. These ﬁndings indicate that a compromise should be found between a sufﬁciently long SRT necessary to prevent the washout of nitriﬁers, and a negative inﬂuence of too long SRT (decreased mass transfer due to poor aeration, competition with GAOs, etc.). However, over 90% removal of NH3 -N is usually achieved in MBR systems, almost independent on the SRT [56, 65]. Pollice et al. [65, 99] investigated a performance of an MBR system which start-up was done without any sludge inoculum. The system was fed on municipal sewage in order to favor biomass selection based on the imposed operating conditions. Biodegradation of the inﬂuent chemical oxygen demand (COD) and complete nitriﬁcation were consistently obtained already in the ﬁrst days of operation. The ammonium oxidation performance over the whole experiment showed a typical nitriﬁcation start-up curve with initial N-NO2 production followed by complete nitriﬁcation that occurred only 10 days after the start-up of the plant. As far as HRT is concerned, several studies noted a complete nitriﬁcation in an MBR operating with a HRT as low as 2 h [39, 100]. Other important factors that are to be considered for nitrogen removal are alkalinity, temperature, and organic and nitrogen loads (C/N ratio, i.e., biological oxygen demand (BOD) to total nitrogen (TN) ratio) [101, 102]. The BOD/TN ratio must be high enough to denitrify the nitrogen to be nitrated. Though this ratio is an important factor to be considered for successful nitrogen removal, it depends upon the components of organic matter that were readily degradable, such as volatile fatty acids [101]. As far as temperature is concerned, it is considered that it has to be maintained below 40 ◦ C to ensure sufﬁcient nitriﬁcation. If the temperature is controlled, a nitriﬁcation rate usually over 99% can be gained in spite of variations of inﬂow TN concentration [101]. Aerated MBR offers two major advantages in the elimination of phosphorus: complete removal of all particles (containing usually up to 0.1 mg of P per mg of total solids (TS)), and aeration, which prevents the phosphate release that occurs under anoxic conditions. Furthermore, there is an increasing interest for the application of MBR as a technology for phosphorus recycling, since the P-content of sludge is expected to increase when prolonging SRT [86].

62

J. Radjenovi´c et al.

Much research has conﬁrmed that MBR is a highly viable wastewater treatment technology regarding nitriﬁcation-denitriﬁcation and phosphorus removal. With optimized design and operating parameters it warrants high efﬂuent quality in terms of ammonia, nitrates, and phosphates present in wastewater. Current European regulation describes guidelines for total phosphorus and nitrogen in treated efﬂuent to 1–2 mg L–1 and 10–15 mg L–1 , respectively [86]. More stringent regulations are expected to come into force soon in some countries, which fulﬁllment will require improvements in the existing treatments and the implementation of additional ones. 5.3 Removal of Organic Matter and Suspended Solids Knowledge about COD removal mechanisms that occur when mixing an activated sludge with real wastewater is still scarce. The microbial response to dynamic conditions in a real wastewater treatment unit can be different from a simple increase in cell number (i.e., growth of microbial population), and include other substrate-removal mechanisms like sorption, accumulation, and storage [103]. There have been several investigations on treatment efﬁciencies of MBR and CAS processes operating under comparable conditions that have shown signiﬁcantly improved performance of an MBR in terms of COD, NH3 -N and SS removals [3, 30, 40, 51, 83, 104–111]. There are several factors that may contribute to the lower organic carbon content of MBR efﬂuents as compared to CAS processes, like longer retention times, smaller ﬂoc sizes, etc. Côté et al. [100, 112] attributed the improved COD removal to the avoidance of biomass washout problems commonly encountered in activated sludge process, as well as to complete particulate retention by the membrane. Membrane rejection of a signiﬁcant amount of soluble organic molecules and colloids makes their removal more effective due to a higher lyses activity in the reactor induced by elevated concentrations of these compounds. Higher sludge ages that are achieved by long SRTs allow more complete mineralization of biodegradable raw water organics, but also an adaptation of microorganisms to less biodegradable compounds. Therefore, biomass can acclimatize to wastewater without being restricted to fast-growing and ﬂoc-forming microorganisms. In a study of Al-Malack et al. [95], COD removal efﬁciency in immersed MBR was found to increase signiﬁcantly with increase in MLSS concentration, however, the effect of SRT on permeate COD became insignificant for MLSS concentrations above of 3 g L–1 , which probably means that the organic loading rate was not high enough to show a signiﬁcant difference at higher biomass concentrations. Since typical sludge concentrations for immersed MBRs are between 15 and 25 g L–1 , elimination of organic matter and turbidity is almost independent on SRT, and average removals normally achieved for COD and SS are over 90 and nearly 100%, respectively [112].

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

63

Better performance of MBR operated at long SRTs can also be explained by the presence of dispersed bacteria that are advantageous in the overall population competition when substrate concentration becomes very low, i.e., at low F/M ratio and high sludge age. Flocs in a bioreactor were found to be smaller (Fig. 13) [56], which can explain enhanced mass transfer for both oxygen and carbon, thus enabling a higher removal rate and more adaptability to changes in the inﬂuent quality and quantity [53, 113]. In another study it was demonstrated that the ﬂocs were more active and displayed greater species diversity [104]. The overall capacity of biomass to degrade different carbon substrates does not change signiﬁcantly at different SRTs, which conﬁrms that MBR is capable of degrading a wide variety of carbon substrates in a similar fashion. This robustness of MBR treatment regarding turbidity and organic matter removals was conﬁrmed in several studies [114]. Xing et al. [58] recorded high treatment efﬁciency regardless of the absolute level of sludge concentration in the MBR, and unaffected by variations in SS and volatile suspended solids (VSS) inﬂuent concentrations. In another study, in spite of large ﬂuctuations in the inﬂuent, COD efﬂuent COD was always low and extremely stable, because upon the addition of organic substrates, biomass responded immediately with increased respiration activity [114]. It is assumed that there is an upper limit for organic loading rate in an MBR under which degradation performance is independent of biomass concentration and organic loading rate. Rosenberger et al. [44] found that for organic loading rates lower than 7 kg COD m–3 day–1 , COD removal in MBR was high and stable regardless of MLSS concentration and composition of microbial culture. Moreover, another study reported that the mineralization process was not impaired nor with the shifts in the morphological composition of microbial population, or even with the occurrence of high numbers of ﬁlamentous bacteria [44]. Pollice et al. [65] tested a performance of MBR when its start-up was done without any sludge inoculum. Biodegradation of the inﬂuent COD as well as complete

Fig. 13 Sludge particle size distributions at different SRTs [56]

64

J. Radjenovi´c et al.

nitriﬁcation were consistently obtained already in the ﬁrst days of operation, which demonstrated the high responsiveness of MBR. In aerated MBR, COD loss is also a result of the production of volatile compounds that are released from the system under aerated conditions [92]. Concerning turbidity removal, due to a complete retention of particulate matter by the membrane, there are no suspended solids found in the MBR efﬂuent, unlike the efﬂuent of a conventional process. The UF/MF membrane can capture all SS in the reactor because of its ﬁne pore size [115]. Therefore, non-biodegradable organic compounds are removed through ﬁltration of particulates and discharged with the sludge. Gander et al. [116] reported that membrane contribution to the removal of organic matter was approximately 30%, this roughly equating to the insoluble fraction that was removed via active biomass. In another study with an external membrane module, total COD removal was 97% on average, where 85% was removed by the bioreactor and only 12% resulted from membrane separation [58]. As far as HRT is concerned, results of Sun et al. [113] indicated a clear inﬂuence of the operation time on biomass concentration (Fig. 14). Short HRT brings up a higher concentration of biomass because the volumetric organic loading is bigger (more food is supplied to microorganisms), although the oxidation of organic matter might not be complete. On the other hand, Chaize et al. [39] recorded a complete nitriﬁcation and organics removal at HRT of only 2 h.

Fig. 14 Growth of MLSS concentrations in submerged MBR at different HRTs [113]

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

65

Aeration ﬂow is also one of the main factors that affect the biochemical process of BOD and COD removals. The right amount of oxygen needs to be provided to the microorganisms in response to their three speciﬁc demands: 1. carbonaceous BOD (conversion of the carbonaceous organic matter in wastewater to cell tissue and various end products), 2. nitrogenous BOD (in the process of nitriﬁcation ammoniacal nitrogen is oxidized to the intermediate product nitrite, which is then converted to nitrate), 3. inorganic COD (oxidation of reduced inorganic compounds within the wastewater) [117]. Biomass characteristics such as SMP and EPS strongly inﬂuence the oxygen transfer, so therefore they will determine the rate of organics removal. These compounds are also widely recognized as the main membrane foulants [118]. SMP is soluble and thus is in the liquid phase and EPS is bounded to cells and makes a part of the solid phase. In order to reach the active sites on the bacterial cell membrane, the oxygen needs to penetrate the liquid ﬁlm surrounding the ﬂocs and then diffuse through the ﬂoc matrix (EPS) [117]. EPS amounts differ with changes in microbial state and operating conditions of the bioreactor. In the intermittently aerated MBR they increase in proportion to the non-aeration time [119]. Nevertheless, Ujang et al. [89] reported no signiﬁcant difference in COD removal efﬁciency when varying aeration and non-aeration time, indicating that in intermittently aerated MBR organic matter can be degraded both under aerobic and anaerobic conditions. Also, over-aeration can bring about poor sludge characteristics such as bad ﬂoc structure and rather low sludge volume index (SVI), which can then be related to fouling [101]. In conclusion, immersed MBR is strongly capable of resisting shockloadings, and variations in the inﬂow turbidity and organic matter content have no effect on their removal efﬁciencies. The removal of organic pollutants in terms of COD and SS has been proven to be very high and a good-quality efﬂuent can be achieved during long-term operation. However, how to operate MBR systems efﬁciently remains a topic of argument because there is a lack of information on the development of microbial community structure in MBRs during nitriﬁcation [57]. It is important to distinguish among these different contributions to the overall COD removed, in order to better understand the dynamics of the process and to build-up a useful basis for process designing and modelling. 5.4 Bacteria and Virus Removal Microbiological water quality is usually measured by monitoring the organisms that might indicate that the water is contaminated with fecal material

66

J. Radjenovi´c et al.

or that disinfection is inadequate. These organisms are referred to as “indicator organisms”, and they are not harmful to health but they coexist in high quantities where pathogens are present. The most common indicator organisms used are total coliform (TC) in drinking water quality control, while for wastewater evaluation fecal coliform and bacteria Escherichia coli are used. On the other side, viruses are expected to be more suitable indicator organisms than bacteria since they are much smaller and harder to straining than bacteria, and also considered to be more resistant to common disinfectants [120]. The removal of enteric viruses requires speciﬁc attention, given their low infective dose, long-term survival in the environment, and low removal efﬁciency in the conventional wastewater treatment. Due to the difﬁculty in assaying animal viruses, bacteriophages have been suggested as viral indicators because they closely resemble enteric viruses in terms of structure, morphology, size, and behavior [121]. A variety of bacteriophages have shown potential as model organisms for monitoring virus-removal in drinking water treatment plants [122], such as F-speciﬁc RNA coliphages (viruses infecting E. coli). MS-2 is the most-studied F-RNA coliphage, and it is often used as an indicator because of its being the smallest (0.02–0.025 µm) among viruses and relative hydrophobicity, which makes it a good worst-case strain and therefore representative to address the ability of pathogen removal. If these pathogens (i.e., microorganisms capable of causing diseases) are not removed by water treatment or disinfection and stay present in water, consumers may suffer infectious diseases like cholera, polio, typhoid, hepatitis, and a number of other bacterial, viral, and parasitic diseases. Sewage treatment may reduce the numbers by ten to ten thousand-fold, depending on the nature and degree of treatment. However, even tertiary treatment of sewage will not eliminate all viruses. In well-functioning biological plants, as many as 106 CFU L–1 resistant coliform bacteria were found in the efﬂuent [123, 124], while this number is much larger when counting with the existence of smaller viruses. In the production of potable water, a limited number of bacteria are acceptable, which also depends on the type of bacteria. The average content of viruses in drinking water should be only around 10–8 viruses L–1 . Hence, given their low infective dose, long-term survival in the environment and low removal efﬁciency in the conventional wastewater treatment, enteric virus removal requires speciﬁc attention. For their elimination, the most important step in potable and wastewater treatment is disinfection, during which the number of pathogenic organisms in water is lowered to an acceptable value. Primary methods of disinfection are chlorination, chloramines, ozone, and ultraviolet light. Other disinfection methods include chlorine dioxide and treatment with potassium permanganate. These processes are often accompanied by mutagenic/carcinogenic and toxic disinfection by-products (DBPs), which are potentially harmful to humans and aquatic organisms. Another disadvantage of chemical sterilization methods is that they kill the present microorganisms without removing the dead ones,

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

67

which are a source of pyrogens (compounds that can cause a rise of a body temperature). One of the most advanced options for disinfection is MBR treatment, where removal of microorganisms is achieved by ﬁltration. Membrane treatments have been proved to be very efﬁcient in reducing vegetative bacteria [125–127]. Ottoson et al. [126] noted a 5 log removal of E. coli and 4.5 log removals of enterococci. These reductions are in line with the MBR removal rates of Faecal coliforms and Faecal streptococci (up to 7 log) reported by Ueda et al. [128]. Phages and spores are not as efﬁciently removed as bacteria, though their elimination can be successfully increased by submerging a membrane module in the reactor for a few weeks, which allows a membraneattached bioﬁlm to develop [126, 129, 130]. Some investigations have shown that membranes were capable of removing viruses completely (UF) or signiﬁcantly (MF) under appropriate conditions [116]. If the pore sizes of the membrane are smaller, then viruses can be removed by size exclusion. Cooper and Straube [131] found that RO can effectively remove viruses from wastewater without any additional treatment. It was found in these studies that the main role in the removal of bacteria as well as viruses plays a bioﬁlm formed on the membrane surface. In the absence of bioﬁlm, virtually no phage removal was observed, while in its presence better phage removal was observed at higher sludge concentrations. It has been suggested that mechanisms for this removal comprise a physical component due to pore size reduction, a chemical component due to viral adsorption on the bioﬁlm, and a biological component resulting from the predation of phage by other microorganisms. Given enough time for a bioﬁlm to develop, the removal improves signiﬁcantly because the membrane surface gets fully covered with gel layers, while internal blockage and partial coverage (presumably by EPS) are observed in the membrane pores [95, 117]. In the study performed by Shang et al. [117] the membrane alone showed poor virus removal. This was to be expected since the average pore size of the membrane ﬁbers (0.4 µm) was much larger than the size of the bacteriophage MS-2 (0.02 µm). The overall removal increased substantially with the presence of biomass and bioﬁlm. Similar results were found by Lv et al. [132] for the elimination of the phage T4 in two membrane modules with pore sizes of 0.22 and 0.1 µm. In the second one, a membrane alone could block most phages from leaking by direct membrane interception, while cake and gel layer played a signiﬁcant role on the phage removal for the 0.22 µm membrane module. Phages are also expected to associate with biomass ﬂocs and then get removed by ﬂocculation or cell adsorption in the aeration tank [117]. There is evidence that pore size does not necessarily describe the ability of a ﬁlter to remove particles from solutions [127]. Besides gel layer at membrane surface, important factors for adsorption of viruses on membranes are chemical composition of membrane, ratio of membrane pore diameter to virus diameter and hydrophobic and electrostatic interactions. The charge of

68

J. Radjenovi´c et al.

most viruses will be negative under the conditions present in most wastewater efﬂuents (i.e., pH 6–7). Neutral net charge at the iso-electric point (pI) of a virus leads to maximum virus–virus coagulation. Aggregation may therefore further promote virus retention by membranes. It has also been noted that the presence of particular ions promotes virus aggregation compared to buffers at low pH alone. For electrostatic interaction of viruses, the thickness of the double-layer as described by Gerba [129, 133] plays the most important role, which is governed by the pH and the presence of salts in the bulk solution (Fig. 15). Increasing the salt concentration (for example its thickness) is reduced (Gouy layer around the virus is decreased) and thus virus adsorption to membrane surfaces is facilitated. Gerba showed that certain salts have a positive effect on both electrostatic and hydrophobic interactions. pI of a virus is a parameter relevant for its electrostatic interactions and relative hydrophobicity. By knowing a virus’s pI it is possible to predict the likelihood of its adsorption to a charged surface as long as the suspending conditions are known, at the ﬁrst place pH of a bulk solution. Van Voorthuizen et al. [129] observed that the retention of bacteriophage MS2 at its pI and in the presence of salts was signiﬁcantly higher when using a hydrophobic membrane compared to the hydrophilic one. If a solution’s pH is greater than the pI of a virus, hydrophobic interaction could play the major role in maintaining virus-ﬁlter adsorption due to the increased electrostatic repulsion at higher pH levels [133]. On the other hand, the pI of a virus can differ within the same strain due to the fact that different viruses have different protein coatings that surround the virus. In addition, metals and other substances present in water could form complexes with these protein coatings, which will have an impact on adsorption characteristics and the measured pI [129].

Fig. 15 Schematic illustration of virus structure with electrokinetic double layers [129]

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

69

The presence of human enteric viruses is a major risk associated with wastewater reuse. As sewage mixes with the receiving water, viruses are carried downstream and the length of time they remain detectable depends on temperature, their degree of absorption into sediments, penetration of sunlight into the water, pH, and other factors. Consequently, enteric viruses can be found in sewagepolluted water at the intakes to water treatment plants. In recent years, the reuse of wastewater for non-potable reuse has gained much attention. MBR efﬂuents were found to be compliant with the EU Bathing Water Directive (EC/160/75) including parameters such as total coliforms, Faecal coliforms, Streptococcus faecalis as Salmonella spp. and Coliphages [134]. However, how to cope with wastewater when different types of viruses coexist and how to dispose the virus-bearing excess sludge must be studied further.

6 Removal of Trace Organic Compounds by a Membrane Bioreactor WWTPs treating wastewater from municipalities and industries have been shown as major sources of many environmental pollutants. These pollutants usually originate from synthetic chemicals that have been used widely for industrial, agricultural, and household purposes. Compounds like pharmaceutically active compounds (PhACs), industrial chemicals, and pesticides are produced worldwide on a 100 000 t scale. After their usage for the intended purpose, a large fraction of these substances will be discharged into the wastewater unchanged or in the form of degradation products that are often hardly eliminable in conventional WWTPs. Depending on the efﬁciency of the treatment and chemical nature of a compound, they reach WWTP efﬂuents and surface waters in certain concentration. In the worst case, they are present in drinking water, in spite of expensive treatment steps. Although the exact effect of consistent exposure to trace organic contaminants is still unclear, there is no more doubt that it has signiﬁcant adverse consequences for public health. For example, antibiotics and their metabolites can signiﬁcantly increase antibiotic resistance in the population. Synthetic hormones can act as endocrine disruptors by mimicking or blocking hormones and disrupting the body’s normal functions. Due to their polarity, they can be eliminated during wastewater treatment only incompletely. Polar poorly degradable compounds were detected in high and comparable concentrations in the efﬂuents of numerous WWTPs all over Europe. A proper wastewater treatment as mandatory in the European Union due to the Urban Wastewater Treatment Directive (91/271/EEC) will not eliminate polar pollutants completely [135]. Therefore, to avoid such contaminants, emissions with WWTP efﬂuents would have to be reduced by their advanced treatment or by avoidance and replacement measures for the respective pollutant. One of the

70

J. Radjenovi´c et al.

most promising technologies is MBR technology. The potential of MBR to efﬁciently remove hazardous substances from wastewater is often highlighted. Besides the fact that there is a physical retention of all the molecules larger than the molecular weight cut-off (MWCO) of the membrane, hydrophobic substances also tend to accumulate onto the sludge and therefore they are removed from the efﬂuent. Furthermore, as all the bacteria are held back, there are better adapted to mineralizing of micropollutants present in the reactor. 6.1 Removal of Pharmaceutically Active Compounds Most pharmaceutical substances are by nature biologically active and hydrophilic so that the human body can absorb them easily, and persistent in order to avoid the degradation before having a curing effect. Depending on the pharmacology of a medical substance, it will be excreted as a mixture of metabolites, as unchanged substance or conjugated with an inactivating compound attached to the molecule [136]. Once they enter a WWTP, pharmaceutical residues are usually not completely degraded or retained by adsorption to sludge. Hence, they pass through wastewater treatment and end up in the receiving waters in certain percentage. Their removal in WWTPs is variable and depending on the properties of the substance and process parameters (i.e., SRT, HRT, and temperature) [137, 138]. A large number of PhACs are hardly eliminated and therefore detected in WWTP efﬂuents. The presence of PhACs in surface, drinking, and wastewaters is well documented in literature [136, 139–147]. Although present in low environmental concentrations, drugs can have adverse effects on aquatic organisms. These effects are rather chronicle than acute toxic effects, depending on the exposure factor (bioavailability), degradability, and susceptibility of the compound in question [148]. The results reinforce concerns about excreted pharmaceutical compounds from wastewater systems that may end up in the water supply, potentially resulting in adverse effects for humans and the environment. The most important removal pathways of organic compounds during wastewater treatment are biotransformation/biodegradation, adsorption to the sludge (excess sludge removal), and stripping by aeration (volatilization). Considering low values of Henry coefﬁcients (KH ) of most of the PhACs detected in wastewater streams [149], stripped fraction removed by volatilization can be neglected [150]. In most of the studies, two processes of abiotic (adsorption) and biotic degradation (transformation by microorganisms) could not be distinguished, and the term “removal” usually refers to a conversion of a certain micropollutant to compounds other than the parent compound. The membrane-activated sludge process is expected to enhance traceorganics removal to a greater extent than the conventional treatment (Fig. 16). There are many reasons for this assumption: higher sludge age, higher

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

71

Fig. 16 Elimination rates of PhACs in MBR and CAS treatment [156]

biomass concentration, complete retention of solids and microorganisms, etc. Several studies have been conducted that conﬁrmed an advantage of MBR over CAS when reduction of pharmaceuticals is concerned [151–155]. Radjenovi´c et al. [155, 156] found signiﬁcantly improved removal of lipid regulators and cholesterol lowering statin drugs (gemﬁbrozil, bezaﬁbrate, cloﬁbric acid and pravastatin), β-blockers (atenolol and metoprolol), antibiotics (oﬂoxacin and erythromycin), anti-ulcer agent (ranitidine) and some analgesics and anti-inﬂammatory drugs as well (propyphenazone, mefenamic acid, and diclofenac). Still, some authors report comparable elimination rates for these two processes [137, 146, 157]. The current understanding of biotransformation of PhACs in WWTPs and MBRs, and of their biodegradation pathways and mechanisms is still incomplete. Although the biodegradation of some pharmaceuticals in batch reactors has been described, it is still unclear how this information relates to biotransformation processes under real conditions of WWTPs. Qunitana et al. [154] investigated microbial degradation of several pharmaceuticals and performance of MBR in their elimination from wastewater. They found a formation of potentially stable metabolites during ketoprofen and bezaﬁbrate transformation, which may deserve further attention when analyzing removal of PhACs in wastewater treatment technologies. In laboratory degradation experiments, ketoprofen yielded two metabo-

72

J. Radjenovi´c et al.

lites formed along the biphenyls, biphenyl ethers and related compounds pathway [154, 158]. Bezaﬁbrate was hydrolytically cleaved along the amide bond yielding one well-degradable metabolite (4-chlorobenzoic acid) and another metabolite that was not mineralized [154]. Ibuprofen, bezaﬁbrate and naproxen were degraded only with the addition of external carbon source (co-metabolism), whereas diclofenac was not transformed [154]. In the same study, the only two metabolites found in wastewater were hydroxyl-ibuprofen and 4-chlorobenzoic acid detected in the MBR inﬂuent, while in the efﬂuent they were not present. Considering the fact that it is still uncertain what kind of adverse effect on humans and the environment these compounds can have, monitoring of these pharmaceutical by-products should be established in WWTPs. Due to the lack of knowledge about the metabolites of PhACs, the non-existence of adequate analytical methods, and a possible sampling inaccuracy, usually no ﬁrm conclusion about their biotransformation can be made. Joss et al. [150] performed batch biodegradation experiments with CAS and MBR sludge. Based on the average results for kinetic biodegradation constants (Kbiol , kgSS –1 d–1 ), they established three different classes of compounds according to their susceptibility to biological degradation: 1. compounds with Kbiol < 0.1 kgSS –1 d–1 , that have no removal (e.g., carbamazepine, diclofenac, diazepam), 2. partially removed compounds, 0.1 < kbiol < 10 kgSS –1 d–1 , (e.g. roxythromycin, fenoprofen, acetylsalicilic acid, naproxen, bezaﬁbrate, cloﬁbric acid, fenoﬁbric acid, gemﬁbrozil, piracetam, and some iodinated contrast agents), 3. compounds removed with more than 90% efﬁciency, kbiol > 10, (e.g., ibuprofen and acetaminophen). However, there are many literature data contradictory to these results. Good elimination in CAS has been reported for both indomethacine and diclofenac [143]. Removals of bezaﬁbrate were over 80% in some investigations [137, 143]. The discrepancy in results of existing publications on this matter could be due to different pharmaceuticals concentrations and/or sludge origin (sludge age, wastewater composition, etc.). When performing batch experiments, their outcome will also depend on the way of handling sludge prior to the experiment. Many studies have conﬁrmed a complete biodegradation of a non-steroidal anti-inﬂammatory drug ibuprofen to hydroxyl-ibuprofen and carboxyibuprofen in WWTP and MBR, whereas removals higher than 95% have been reached [106, 137, 141, 146, 154, 159]. Although a very high elimination of this drug during wastewater treatment has been reported by several authors, it is frequently found in surface and ground waters [160–162]. This is no surprise considering its wide usage and high therapeutic doses prescribed for the treatment of pain, dysmenorhea, inﬂammation, and fever, which makes

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

73

it one of the most important pharmaceutical contaminants found in WWTPs. Also, Stumpf et al. [163] found that hydroxyl-ibuprofen was quite stable during conventional treatment in WWTP, while carboxy-ibuprofen, the main metabolite in humans, disappeared. Moreover, pharmacologically active isomer of ibuprofen is the most prominent form detected in environmental samples, which could be explained by the fact that microorganisms mainly utilize its inactive isomer [141]. This change in enantiomeric composition points to biological dissipation of ibuprofen rather than other processes such as sorption and uptake by the sludge. In short-term biodegradation tests with pilot sewage plant and bioﬁlm reactor, cloﬁbric acid and diclofenac turned out to be very persistent [153, 164, 165]. However, such pilot plants may not be adequate simulations of the actual processes occurring during wastewater treatment. In some WWTPs, attenuation of 50–70% of diclofenac has been reported [106, 137, 143, 145, 166]. Ternes et al. [143] recorded 51% elimination of cloﬁbric acid during conventional wastewater treatment. However, literature data on this subject is still very contradictory. Many studies have reported extremely low efﬁciency of conventional treatment in the removal of diclofenac (only 10–30%) [137, 142, 146, 167]. Cloﬁbric acid has also been found to be a refractory contaminant during wastewater treatment [145]. Kimura et al. [125] related the persistence of diclofenac and cloﬁbric acid in both MBR and CAS processes to the presence of chlorine in their structures, which makes them hardly degradable. Moreover, in batch experiments with MBR and CAS sludge Joss et al. [150] showed no difference between their biodegradation constants (Kbiol ) of diclofenac, which were low in both cases (≤ 0.1 L gSS –1 d–1 ), whereas for cloﬁbric acid Kbiol determined with CAS sludge was greater than with MBR sludge (0.3–0.8 and 0.1–0.23 L gSS –1 d–1 , respectively). On the other hand, Radjenovi´c et al. [155] noted a signiﬁcant improvement in removal of these compounds when using MBR unit. Elimination in MBR of diclofenac and cloﬁbric acid were 87% and 72%, compared to 50% and 28% found in CAS, respectively. Clara et al. [106] noted an improvement in elimination of diclofenac during MBR treatment with prolonging the SRT. González et al. [168] suggested that faster diminution of diclofenac was because of better acclimation of microorganisms to the MBR inﬂuent water. Besides possible changes in microbial consortia during its adaptation to wastewater contaminants, another explanation for better performance of MBR could be higher sorption potential of sludge, as the organic matter content is higher with respect to the CAS sludge. Moreover, according to the results of the EU project POSEIDON [169], sorption processes are relevant for the elimination of diclofenac. Cloﬁbric acid can perhaps be sorbed to sewage sludge particles in acidic conditions, since considering its pKa value (4.91) it will exist in non-ionic form, which makes it more hydrophobic [170]. Carbamazepine is an established drug used for the control of psychomotor epilepsy. It is also used in the treatment of trigeminal neuralgia and

74

J. Radjenovi´c et al.

bipolar depression. Concentrations to several hundred nanograms per liter of this pharmaceutical have been detected in different surface waters [142]. Poor elimination of this neutral drug has been reported previously by many authors [144, 146, 157, 171]. Carbamazepine does not adsorb onto the sludge [169]. Also, the pore size of MF membranes usually applied in wastewater treatment processes does not allow any further retention of the molecule. Therefore, carbamazepine passes both WWTP and MBR without any reduction and efﬂuent concentration in the range of inﬂuent ones were measured in many studies [137, 145, 146, 157, 172, 173]. As is often the case, efﬂuent concentrations of carbamazepine are detected to be greater than the inﬂuent ones. This could be explained by the presence of its input conjugate compounds that are being retransformed during treatment into the original compounds [143]. Unexplained variations of concentration over time can also be observed for sulphonamide antibiotics, probably also because of unknown conjugation and deconjugation processes that may occur during contact with activated sludge. For example, a signiﬁcant amount of sulfametoxazole enters WWTP in metabolized form as N4 -acetyl-sulfamethoxazole that can be converted back to the original compound [174]. Variable elimination percentages of sulfamethoxazole can also be associated with the dependence of its biotic degradation on the presence of easily biodegradable organic matter in wastewater, which is submitted to changes in both MBR and CAS systems [175]. Theoretically, trace organic removal should be better in MBR than in CAS because of high SRT and sludge retention on membranes. This enables biological adaptation and changes in microbial consortia, whereas synthesis of specialized enzymes for biodegradation of micropollutants is induced. Data from literature show that higher sludge age often reached in MBRs may signiﬁcantly improve the removal of speciﬁc compounds [146, 152]. For compounds like trimethoprim and macrolide antibiotics azithromycin, erithromycin, and clarithromycin, a clear increase in transformation was found at sludge ages of 60–80 days [152]. In the same study, higher reduction of roxythromycin was observed already at 33 days SRT. Clara et al. [106] reported higher removal of diclofenac when increasing SRT in an MBR. However, Joss et al. [146] found no improvement in degradation of micropollutants with increased SRT (i.e., carbamazepine, naproxen, diclofenac, ibuprofen, roxythromycin, and bezaﬁbrate). In general, the more hydrophobic the chemical is, the amount adsorbed will be greater. However, there are many factors that may contribute to the ultimate concentration of organic pollutants in sewage [176]. Urase et al. [170, 177] reported higher elimination of diclofenac, indomethacin, and some other acidic PhACs (cloﬁbric acid, ibuprofen, ketoprofen, fenoprofen, gemﬁbrozil, naproxen) during CAS and MBR treatment in acidic operational conditions. It was postulated that this was due to their increased hydrophobicity since these compounds did

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

75

not exist in ionic form in the acidic pH, which resulted in adsorption onto sludge particles. For example, the elimination rate of diclofenac was more than 90% when pH was in the range of 4.3–5.0, compared to around 10–15% removal at neutral pH. On the other hand, propyphenazone and carbamazepine do not have functional groups to form ions, and their removal was not affected by the pH. Jones et al. [178] detected no adsorption of ibuprofen, salbutamol, acetaminophen, and propranolol hydrochloride to sewage sludge. In the same study, some removal to solids of mefenamic acid was indicated. However, this compound has frequently been found in WWTP efﬂuents. There are several mechanisms responsible for the sorption of a certain organic compound onto the activated sludge: adsorption onto bacterial lipid structure, sorption onto polysaccharide structures outside bacterial cells, and chemical binding to bacterial proteins and nucleic acids [179]. Kumagai et al. [180] reported greater biosorption of pharmaceuticals having a more signiﬁcant protein binding. However, in real sewage, pharmaceuticals are likely to be out-competed for sorption sites by other hydrophobic contaminants, which means that a greater proportion of them will remain in the aqueous phase than the expected one. Besides hydrophobic processes that are taking place, a number of other reactions like complex formations with metal ions, ion exchange, and hydrogen bindings decide about the partition of the organic compound between the solid and the liquid phase [181]. Since for adsorption processes the organic fraction, i.e., VSS of the sludge is relevant [182], MBR sludge is expected to have higher sorption potential, as the organic matter content is higher in respect to the sludge of the CAS. Literature data indicates irrelevant sorption coefﬁcients (Kd ) values for most of the pharmaceuticals investigated by now, with the exception of macrolide antibiotics azithromycin and clarithromycin [183, 184]. For other PhACs like acetaminophen, naproxen, indomethacin, ibuprofen, fenoprofen, diclofenac, roxithromycin, bezaﬁbrate, cloﬁbric acid, fenoﬁbric acid, gemﬁbrozil and N4 acetyl-sulfamethoxazole, sorption effects can be neglected [150]. In this case, biological transformation can be estimated by direct comparison of the soluble concentration in the inﬂuent and efﬂuent. Joss et al. [146] used an approximation to assess if the sorbed amount is signiﬁcant or not (Eq. 6): Lsorbed ≤ 0.1Lsol, out ,

(6) (g m–3 WW, treated ),

where Lsorbed is the load of the compound sorbed onto sludge and Lsol, out is the soluble load in the efﬂuent (g m–3 WW, treated ). In the same study, biodegradability of pharmaceuticals was estimated in MBR and CAS sludge. Transformation rate constants were similar in two types of sludge for most of the compounds. However, in some cases they differed quite a lot, like for gemﬁbrozil and fenoﬁbric acid, where kbiol in MBR was ten times smaller than for CAS sludge, and also MBR degradation of piracetam and bezaﬁbrate was conducted at a signiﬁcantly higher rate (Fig. 17).

76

J. Radjenovi´c et al.

Fig. 17 Kinetic degradation constants of pharmaceuticals observed in sludge from nutrient-removing municipal WWTPs: the average of CAS and MBR batch experiments is indicated. The error bars indicate the 95% conﬁdence interval. The lines at Kbiol 0.1 and 10 L gSS –1 d–1 indicate limits for less than 20% and more than 90% removal expected from nutrient-removing municipal wastewater treatment. The faded columns indicate values from which experimental resolution allows only identifying an upper limit for Kbiol (upper error bar) (adopted from Joss et al. [150])

Knowledge about the removal of trace organic compounds by MBR is very limited. There have been several investigations conducted on the efﬁciency of WWTPs in removing PhACs from wastewater and its comparison with advanced MBR treatment [106, 137, 146, 150, 151, 153–155, 177]. For most of the investigated PhACs, MBR efﬂuent concentrations were usually signiﬁcantly lower than in the efﬂuent of a conventional treatment. They are removed from wastewater during membrane treatment by sorption, degradation, or a combination of both. Better removal of readily biodegradable micropollutants in the MBR could be due to smaller ﬂock size of sludge, which enhances mass transfer by diffusion and therefore increases the elimination. Considering the

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

77

composition of sludge originating from a membrane bioreactor (specialized microorganisms, large portion of active biomass in suspended solids) improved removal is to be expected. In general, no relationship has been found so far between the structures of micropollutants and their removal during wastewater treatments. Urbanization and constant population growth is likely to keep increasing the quantity of wastewater discharged to WWTPs. Also, considering fast development of pharmaceutical industry and general aging of population, it can be assumed that PhACs will be more consumed and with a more diverse array, with development of new compounds that have unknown fate and effects on the environment. At the same time, the demand for clean water increases as well. Therefore, new technologies for wastewater treatment like MBR will have increasing interest. Although the efﬁciency of MBR as a barrier for micropollutants such as PhACs is still not clear, it seems to be a promising mean for their removal. 6.2 Removal of Hormones Estrogenic substances have been identiﬁed and quantiﬁed in a wide variety of environments associated with industrial and municipal efﬂuents, as well as urban and agricultural runoffs [185–191]. Negative adverse health effects on aquatic organisms which could be attributed to endocrine disrupting compounds (EDCs) are reported by several authors [192, 193]. EDCs are substances that interfere with the hormone system of animals and human beings. When absorbed into the body, they either mimic or block hormones and disrupt the body’s normal functions. This disruption can happen through altering normal hormone levels, halting or stimulating their production, or changing the way they travel through the body, thus affecting the functions that these hormones control. According to a description by the European Commission (CEC 1999), an endocrine disruptor is “an exogenous substance or mixture that alters function(s) of endocrine system and consequently causes adverse health effects in an intact organism, or its progeny, or (sub) populations”. Two different classes of substances causing endocrine disruption can be identiﬁed: natural substances, including natural sexual hormones (estrogens, progesterone and testosterone) and phytoestrogens (chemicals produced by plants that act like estrogens in animal cells and bodies), xenobiotic substances, including synthetic hormones as the contraceptive 17αethinylestradiole (EE2) as well as man-made chemicals and their by-products (e.g., pesticides, cleaning agents, ﬂame retardants, etc.). Sumpter et al. [194] documented feminization of male and sexually immature ﬁsh caused by stimulation of vitellogenin synthesis in male ﬁsh by EDCs. Guillette et al. [195] noted changes in gonads of alligators. Several studies conﬁrm that human estrogens, mainly 17β-estradiol (E2), estrone (E1) and

78

J. Radjenovi´c et al.

EE2 are responsible for a signiﬁcant part of the endocrine-disrupting effects seen in the aquatic environment [185, 196]. They are known to have a strong biological impact already at very low concentrations of 0.1–0.5 ng L–1 [197]. Xenoestrogens like nonylphenol (NP) and bisphenol-A (BPA) are not as estrogenically active as for example E2, but due to their widespread production, they can be found in the environment at much higher concentrations [185]. Human excretion is thought to be the principal source of estrogens and progesterone. Although they undergo various transformations in liver of humans and mammalians (oxidation, deoxydation, hydroxylation and methylation), estrogens are principally excreted as inactive polar conjugates of glucuronide acid and sulphate [198]. However, these conjugates can be cleaved in WWTP due to the presence of β-glucoronidase and arylsulfatase enzymes of a bacterial sludge, and they reform original compounds [186, 199, 200]. Therefore the estrogenicity of the efﬂuent is greatly increased, since the estrogenic potentials of conjugated forms of estrogens are clearly much lower [201]. Since EDCs are suspected to enter rivers, streams, and surface water through the efﬂuents of WWTPs, the elimination of these substances in these treatment plants is of elementary interest. Different processes of varying efﬁciency are applied. Regarding conventional wastewater treatment, some authors consider that in current European activated sludge treatment plants with a HRT no greater than 14 h, the elimination of estrogens and progesterons present in the inﬂuent is not complete [186, 199, 202–204]. However, results obtained by now are very diverse, and efﬁciencies over 90% have been reported for E1, E2, and EE2 in various municipal WWTPs [151, 184, 185, 205–208]. Slightly lower removal of E1 is frequently detected in WWTPs, which is probably due to the conversion of E2 to E1 during the treatment and the cleavage of glucuronides [186, 209]. Advanced water-puriﬁcation techniques like UV radiation, ozonization, or activated charcoal may signiﬁcantly improve the removal of endocrine disrupters, but these techniques are not broadly applied because of their high cost. Membrane bioreactor technology is a possibility to enhance the removal of EDCs. Factors like high sludge ages and low organic loads could not yet be correlated to improved degradation capacity, although biomass compositions might inﬂuence hormone removal in MBRs [210]. Servos et al. [201] found that EE2 was not degraded under non-nitrifying conditions, while nitrifying sludge could oxidize it to more hydrophobic compound. Vader et al. [201] also showed that the EE2 degradation capability of sludge was associated to the nitrifying activity. This phenomenon is probably a reﬂection of improved biological diversity and growth conditions in these systems resulting in increased biological transformations. Synthetic estrogens in general exhibit greater resistance to the activated sludge process [211]. Besides aerobic biodegradation, sorption is also hypothesized as one of the main mechanisms for the elimination of hormones in the aquatic en-

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

79

vironment. However, there is no clear agreement about its relative importance [211]. Although E1 and E2 are relatively water soluble, a signiﬁcant fraction can be associated with organic particles or colloids in the treatment systems, potentially inﬂuencing their degradation and ultimate fate [56]. Clara et al. [212] investigated sorption behavior of BPA, E2, and EE2. High adsorption potential to sewage sludge could be observed for these substances, with no saturation levels detected. However, at high pH values typical for WWTPs (due to the application limestone or milk of lime for sludge conditioning) it was observed a release of the adsorbed fractions of all three investigated substances. Several researchers have shown that higher pH values (pH > 11) led to an almost complete desorption of EDCs (ex. BPA), which generally exhibits phenolic character [212, 213]. Also, in the experiments of Reddy et al. [200] acidic pH showed to be the best method in preventing dissociation of the steroid conjugates to free steroids. In a study by Joss et al. [205], the MBR elimination of natural estrogens E1 and E2 was seen to be higher than in the CAS sludge by the factor of 2– 3. Higher removal rates in MBR compared to CAS could be explained by the smaller ﬂock size of MBR sludge: it was measured to be 10–100 µm for MBR ﬂocks and 100–500 µm for CAS [56]. The thickness of the boundary layer is estimated to be 10–20 µm for MBR and 20–30 µm for CAS ﬂocks [205]. Although there is no information available as to where various estrogen degradations take place throughout the ﬂock, on the ﬂock surface (e.g., on the outer bioﬁlm layer) or in the bulk medium (e.g., catalyzed by extracellular enzymes), it is considered that the size of ﬂocks may also contribute to degradation activity of the sludge. Among natural hormones, E2 is considered as highly biodegradable, while EE2 is slowly biodegradable [186, 214]. The removal of E2 was always recorded to be high regardless of pH [177]. In another study, the same authors found pH conditions irrelevant for the sorption of E1, E2, EE2, and BPA [170]. These estrogens do not have hydrophilic functional groups and therefore the decrease in pH has no effect on their water–sludge partition coefﬁcient (Kp ) [170]. A linear relationship was found between their log Kp and logarithm of Kow (octanol-water partition coefﬁcient, log Kow ) values, which means that adsorption of compounds increases linearly with an increase in its hydrophobicity [170]. Generally, the micro and ultraﬁltration membranes do not display a barrier effect to hormones, but compared to conventional secondary and tertiary systems, a high removal can be expected due to full particle retention promoting the adsorption onto the sludge ﬂocks. Schäfer et al. [213] showed that particles of activated sludge could not only act like adsorbents of hormones but also like a dynamic membrane or rejecting layer for micropollutants. Although a certain estrogenic activity could be expected in MBR efﬂuent (due to the cleavage of estrogen conjugates, or dissolution of particles due to digestion process that may release estrogens by desorption), membrane tech-

80

J. Radjenovi´c et al.

nology represents a good combination of different mechanisms for removal of endocrines. On one side, smaller ﬂoc sizes and higher sludge activity enhance biodegradation, and on the other particle size exclusion enables retention of the adsorbed compounds. 6.3 Removal of Surfactants and Their Degradation Products Surface active substances are an extensively used group of chemicals, e.g., domestic detergents, pesticide formulations, industrial products, etc. Several main classes of surfactants (e.g., linear alkylbenzene sulphonates (LAS), alkylphenol ethoxylates (APEOs) and alcohol ethoxylates (AEO)) have shown very high ubiquity in the environment, thus presenting a serious environmental problem. Among the APEOs, octylphenol ethoxylates (OPEOs) and nonylphenol ethoxylates (NPEOs) are the two most common surfactants on the market. NPEOs account for about 80% of the total APEOs consumption: they are widely used in industry, agriculture, and households as detergents, emulsiﬁers, wetting agents, spermicides, and pesticides, etc [215]. Approximately, 500 000 tons are produced annually worldwide, 60% of which ends up in the aquatic environment [216, 217]. Anionic surfactants LAS are mainly used in laundry detergents and cleaning agents. Numerous data on primary and ultimate biodegradation of LAS have been reported. It is expected for these compounds to undergo a primary degradation of up to 93–97%. Very high levels of biodegradation (97–99%) have been found in some WWTPs using aerobic processes [105, 218–220]. Therefore, LAS are generally regarded as biodegradable surfactants, with a breakdown mechanism that involves degradation of the straight alkyl chain, the sulphonates group and ﬁnally the benzene ring [221, 222]. Yet, biological degradation of LAS in WWTP is not complete since aerobic breakdown intermediates sulfophenyl carboxylates (SPC) are regularly found in WWTP efﬂuents. In a study that covered eight municipal WWTPs in Western Europe, SPCs were detected in median efﬂuent concentrations of 57 µg/L [223]. Besides their biological removal, the process of adsorption to sludge particles also occurs in WWTP. Berna et al. [224] reported that a signiﬁcant proportion of LAS in raw sewage (10–35%) adsorbs to particulate matter. Moreover, longer alkyl chains confer greater hydrophobicity, thus increasing their adsorptive tendency [225]. LAS elimination in MBR unit has been reported to be very similar to a conventional treatment by several authors [105, 226, 227]. Both LAS and SPCs were reported to be removed to a high extent in these two treatment processes (96–98%) [228]. However, Bernhard et al. [228] studied elimination of persistent polar pollutants (P3 ) pollutants in MBR and CAS, whereas MBR showed a signiﬁcant improvement when removal of LAS is considered, and a slightly better performance regarding the attenuation of the concentration of SPC.

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

81

WWTP can eliminate parent compounds of NPEOs rather efﬁciently [227, 229–231]. However, in WWTPs APEOs degrade into more toxic shorter-chain APEOs and alkylphenols (APs) such as NP, octylphenol (OP), NP mono-, di-, and triethoxylates (NPEO1, NPEO2, and NPEO3) and NP carboxylates (NPEC1 and NPEC2). Many studies have reported on their wide occurrence in the environment [232–235]. It was observed a change in distribution of NP, NPEOs, and NPECs between WWTP inﬂuent and efﬂuent [236, 237]. Ahel at al [237] noted that in the primary efﬂuent NPEOs were the most abundant class (82.4%), while in the secondary efﬂuent over 70% were metabolic products, the most abundant being NP1EC and NP2EC (46.1%) (Fig. 18). Therefore, whereas in the inﬂuent NP and NP1EO are the main fractions, in the efﬂuent NP1EC and NP2EC are the predominant ones [137, 237]. There is a concern that the concentration levels of these metabolites present in the environment may be sufﬁcient to have endocrine disrupting effect on wildlife and humans [187, 238]. Moreover, during chlorination of wastewater, the residues of alkylphenolic compounds can be transformed into even more persistent halogenated derivatives that can reach drinking water systems [139]. Compared to conventional treatment, MBR technology has the advantage of giving an efﬂuent with lower concentration of lipophilic metabolites. This is probably due to a better adaptation of microbial consortia which become more capable of degrading persistent NPEO oligomers (NPEO1 and NPEO2) [226]. The elimination efﬁciencies of parent nonylphenolic compounds in MBR unit and CAS treatment have been reported to be similar (> 90%) [226, 239]. In a study by González et al. [227], CAS treatment was

Fig. 18 Relative abundance of NPEOs and their metabolites in primary and secondary efﬂuents (weight-based average value of 11 WWTPs in the Glatt Valley, Switzerland) [237]

82

J. Radjenovi´c et al.

Fig. 19 Boxplot (calculated on a molar basis) and average composition of NP compounds in inﬂuent, CAS efﬂuent and MBR efﬂuent [227]

found to be generally inefﬁcient in removing nonylphenolic compounds with overall elimination around 54%, while the ultimate elimination efﬁciency in the MBR reached 94%. The distribution of the nonylphenolic compounds in the MBR efﬂuent showed rather proportional percentages of all species (23% of the parent compounds, 35% NPECs, 34% short ethoxy chain NPEOs and 8% NP) with overall much lower concentration of potentially estrogenic metabolites as compared to the CAS efﬂuent (Fig. 19). Considering that the primary biodegradation of APEOs results in the formation of various persistent metabolites that are usually poorly removed even in most efﬁcient WWTPs and have a signiﬁcantly enhanced removal in the MBR, membrane technology gives room for expectation that with this alternative wastewater treatment the ecological risk associated with alkylphenolic compounds as well as other ionic and non-ionic surfactants can be drastically reduced. The utilization of MBRs in municipal WWTPs may ensure efﬁcient elimination and biodegradation of APEO-derived EDCs, thus reducing the ecotoxicity of efﬂuents posed by those compounds. 6.4 Removal of Sulfonated Organic Compounds, Pesticides, Musk Fragrances and Other Micropollutants Sulfonated organic compounds are frequently used in industrial processes [240–243] and often found in municipal efﬂuents [240, 244]. Reemtsma

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

83

et al. [245] studied the performance of MBR system in eliminating naphthalene sulfonates and benzothiazoles. Technical mixtures of naphthalene monosulfonates (NSAs) and disulfonates (NDSAs) are widely used as dispersants in industrial processes. Benzothiazoles are a class of industrial chemicals that are used as fungicides, corrosion inhibitors and vulcanization accelerators. They reported very similar efﬁciencies of MBR and CAS treatment in eliminating naphthalene sulfonates for most of isomers, except for 1,7- and 2,7-NDSA who were subjected to more complete degradation in MBR system, probably due to low substrate supply and a high sludge age [245]. In both treatments disulfonates were poorly removed, while the removal of NSAs was a lot better (1-NSA and 2-NSA were had removal rates greater than 99%). This was probably due to better capability of MBR sludge to biodegrade these micropollutants, since due to their great polarity and low molecular weight removals by sorption and membrane rejection can be excluded [245]. In the same study, benzothiazoles were removed by membrane treatment with an average of 87%. However, strong differences were reported for various benzothiazoles: while the concentration of benzothiazole (BT) was lowered for only 37%, up to 99% of mercaptobenzothiazole (MBT) was degraded. Practically no removal was noted for aminobenzothiazole (ABT), hydroxybenzothiazole (OHBT), and methylthiobenzothiazole (MTBT). In another study, two MBRs operating in parallel with CAS treatment were monitored in Berlin, Germany [246]. Performance of two membrane units regarding the removal of benzothiazoles was signiﬁcantly better than in the treatment (43% and 10%, respectively). The greatest difference in efﬁciency between MBR and CAS was detected for BT, which was removed signiﬁcantly better in the MBR (70%) [246]. The process of the aerobic degradation benzothiazoles is largely unknown: although some benzothiazoles are degraded in activated sludge system, most of laboratory data on the pathways of their biodegradation are inconclusive. Pesticides continue to be the focus of many environmental studies and contamination of water resources by pesticide residues is one of the major challenges for the preservation and sustainability of the environment. Inappropriate use of pesticides can give rise to severe and long-lasting ecological damage through pesticide-containing wastewaters that enter the environment. Acidic herbicides such as phenoxy acids (MCPP, MCPA, 2,4-D, 2,4-DP) and bentazone represent an important class of pesticides used not only in the control of weeds in crops but also as algicides in paints and coatings and roof-protection agents. Bernhard et al. [228] compared removals of several persistent polar pollutants in MBR and CAS. In the case of pesticides, some of them were found to be poorly degradable in both WWTP and CAS (e.g., atrazine, bentazone, isoproturon), while 2,4-dichlorobenzoic acid was rapidly eliminated in both treatments. González et al. [168] also noted a good MBR removal of 2,4-D and persistence of bentazone on the other side. For compounds like MCPP and MCPA, the importance of acclimation period was emphasized, con-

84

J. Radjenovi´c et al.

sidering that their degradation was signiﬁcantly higher after a lag period of microorganisms [168]. However, the advantage of MBR in comparison with other investigated processes is high unit-volume removal rate. For example, in other investigated treatments of 2,4-D polluted wastewaters such as conventional activated sludge processes [247, 248], sequential batch reactors [249], and the anaerobic ﬂuidized bed reactor [250], it was found to be very low, even at highest elimination efﬁciencies (0.02 to 0.3 kg 2,4-D m–3 day–1 ). Signiﬁcant improvement in biodegradation when using MBR instead of CAS treatment was reported for insect repellents and metabolites Bayrepel, Bayrepel acid and DEET [228]. These compounds were very persistent during conventional treatment, but achieved high removals in the membrane unit. Polar compounds can spread along a partially closed water cycle after discharge with municipal wastewater and occur in raw waters used for drinking water production. For compounds like ethylenediamino tetraacetate (EDTA), TCPP, and TCEP, their potential to spread with the water along its ﬂow path and penetrate into groundwater from inﬁltrated surface waters has been shown in several studies [251–253]. EDTA is utilized in many industrial applications and in households and has been proven to be widely distributed in aquatic systems [254]. EDTA was detected in very high medium concentration in European municipal WWTP efﬂuents and surface waters, 60 µg L–1 and 3.7 µg L–1 , respectively [223]. This compound showed an extremely recalcitrant behavior during conventional and membrane treatment as well [228]. Another poorly degradable compound in both CAS and MBR is ﬂame retardants (TCEP, TCPP) [228]. Moreover, no enhancement in their removal could be observed when increasing SRT [228]. For the compounds that cannot be removed effectively, new emission limits should be set and also certain strategies for their avoidance, in order to prevent their entrance into closed water cycles. Benzotriazoles are a class of high-production-volume chemicals that are used as corrosion inhibitors in various industrial processes and in households. They were detected in WWTP efﬂuents and surface waters [140, 255]. Treatment of municipal wastewater by MBR instead of CAS also improves the removal of benzotriazoles (61 and 37% on average, respectively) [256]. Also, MBR was able to cope with elevated inﬂuent concentrations without responding with elevated efﬂuent ones, and stability of performance with changes in temperature between summer and winter was found to be greater in the membrane unit [256]. Submerged MBR also showed to be very efﬁcient when treating polymeric industrial wastewater, like high-strength acrylonitrile-butadiene-styrene wastewater [257]. Moreover, very high removal of metals and polycyclic aromatic hydrocarbons (PAHs) was noted in submerged ultraﬁltration MBR operating in alternate aerobic/anoxic cycles mode [258]. Musk fragrances tonalide and galaxolide are generally removed to the same extent (85%) in MBR and CAS, with slightly lower efﬂuent concentrations in the MBR unit [106, 137]. As both substances are very hydrophobic, for

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

85

an estimation of the mass ﬂux, speciﬁc Kd s have to be included. Sorption coefﬁcients of these compounds were determined to be 5200 and 7900 L kg–1 for galaxolide in CAS and MBR sludge, respectively, and 10 800 and 16 000 L kg–1 for tonalide, respectively [259]. Diurnal variation pattern of tonalide (AHTN) and galaxolide (HHCB) were paralleled by the nitrogen load, suggesting that human excretion was a major source of these micropollutants, as well as it was the case with PhACs [146]. Their elimination was around 50% during biological wastewater treatment, which was estimated to be mainly due to their sorption onto sludge particles [146]. Polar compounds may occur in WWTP efﬂuents because of their persistence during the activated sludge treatment or because of their incomplete microbial degradation. Sorption of these polar pollutants to wastewater solids can be neglected. However, there can be present signiﬁcant ionic interactions, especially for organic cations. Moreover, concentrations of metabolic products may increase after wastewater treatment, even though the compound is degradable, provided that their formation proceeds faster than its further transformation. General concern about the presence and behavior of organic micropollutants and restrictive legislation on their management and ﬁnal destination invite to enhance the actual treatment processes and to ﬁnd a reliable alternative.

7 Advantages and Drawbacks of MBR Technology A need for the development of MBR technology arose mainly from the limitations of the CAS process. It is of interest here to describe CAS in more detail, since it has been used successfully for almost a century in wastewater treatment. CAS in its most simple manner consists of a primary physical treatment that includes screening of gross solids and sedimentation of settleable solids followed by biological treatment with activated sludge, and subsequent secondary sedimentation where activated sludge in the form of ﬂocs is separated from treated water by gravitational force. The biological step is carried out in an aerated bioreactor in the presence of mixed microbial culture, where pollutants from water are degraded by microorganisms and turned into microbial biomass and gases such as carbon dioxide, water, and inorganic nitrogen products. This stage may include an anoxic zone preceding the aerobic zone within the single reactor or a separate post-denitriﬁcation reactor to achieve a complete nitrogen removal. Also, preliminary anaerobic zone for biological removal of phosphorus is available. The settled sludge is returned to the bioreactor while excess sludge additionally grown during the process is being constantly removed. Essentially, the process uses aerobic microbial biodegradation of organic matrix presented in wastewater in the same manner as the natural microbial community in the water bodies would, if given enough time

86

J. Radjenovi´c et al.

and oxygen. The ﬁnal products of the process are treated water and excess sludge. Treated water is usually discharged into water bodies such as lakes and rivers, while excess sludge ends up mostly as a fertilizer in agriculture or it is disposed of on land. Some countries like Germany and Switzerland forbade the use of secondary sludge in agriculture and excess sludge is incinerated together with hazardous wastes. In any case, the processing of this sludge, which includes operations like thickening, anaerobic stabilization, chemical conditioning, dewatering and thermal reduction [260], represents a cost and a problem that has to be dealt with. Land application of sewage sludge in agriculture is very restricted owing to the presence of potentially toxic substances, i.e., heavy metals, pathogens, persistent organic pollutants, etc. Critical shortage of available land coupled with new, more-stringent regulations for design and operation of landﬁlls have caused prices of their sighting, building, and operating to rise sharply. Incineration is usually the ﬁnal option for sewage sludge treatment due to an abundant ash generation, which has a high content of heavy metals and is generally toxic. Therefore, high sludge production is one of the main drawbacks of CAS. Currently, reduction of sludge wasting is a major challenge of biological wastewater treatment. Excess sludge processing and disposal could account for about 50–60% of the total cost of wastewater treatment [261, 262]. The ideal way to solve the problem of sludge post-treatment and disposal is to reduce its production. To reduce the production of biomass, the wastewater process must be engineered in such a way that substrate utilization is diverted from assimilation of carbon for biosynthesis to non-growth activities of a microbial community. In activated sludge plants, the sludge-yield coefﬁcient (Y) is typically 0.5. [263]. According to Urbain et al. [264], the yield coefﬁcient for an aerobic membrane separation process treating municipal wastewater (488 ± 143 mg COD/L) was 0.23 kgSS kgCODremoved–1 . Pollice et al. [99] reported a production of sludge in an MBR of 0.12 gVSS gCODremoved–1 , which was in accordance with previously reported yields for MBRs [34, 42, 65, 263]. This advantage of MBR, together with the abandonment of energydemanding sludge recirculation loop in CAS, contribute to better competitiveness of membrane technology compared to the conventional one. The limiting step in the conventional treatment is the separation of sludge from the treated water. Without a good sedimentation in secondary settler, parts of the sludge end up in treated water, which leads to poor removal efﬁciency. It should be noted that during the normal operation of CAS per each kg of BOD5 removed up to 0.6 kg of microbial biomass is formed [265], so if the separation of activated sludge is not properly carried out, the whole treatment process loses its purpose. Sedimentation of sludge is inﬂuenced by the characteristics of microbial ﬂocs as a function of their physiological state. In other words, the biological process must be operated in such a way to allow the formation of easily settling microbial ﬂocs. Sludge with poor settling characteristics is often called “bulking sludge”, and in most cases this prob-

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

87

lem occurs due to the growth of ﬁlamentous bacteria. In ﬁlamentous growth, bacteria form ﬁlaments of single-cell organisms that attach end-to-end and normally protrude out of the sludge ﬂoc. Common ﬁlamentous organisms are Sphaerotilus natans, Microthrix parvicela and Thiothrix spp. The main reasons for bulking are low DO concentration, low F/M ratio, and nutrient deﬁciency. At such conditions of low substrate concentration, ﬁlamentous organisms, due to their increased surface-to-mass ratio can compete better for substrate and overgrow the ﬂoc-forming organisms. While DO concentration can be provided by a proper aeration system, problems with variations in wastewater ﬂow rate and composition can seriously affect the CAS process. The usual measures for bulking control are the addition of ﬂocculants like ferric chloride and aluminum sulphate to the settler, or chlorination of return sludge (0.002 to 0.008 kg of chlorine per kg of MLSS per day) since ﬁlamentous bacteria is more sensitive to oxidative agents. Also, if the design of the plant allows, bulking can be mitigated by setting the operational parameters (such as the F/M ratio) high enough to enhance the growth of ﬂoc-forming organisms. By doing so, microorganisms of the activated sludge are kept in the exponential growth phase in which they produce a large amount of excess biomass. To achieve high F/M, the MLSS in the aeration basin has to be kept low (around 3–5 g L–1 dry mass weight) while the concentration of the organic matter in the feedwater needs to be high. These conditions are usually easy to achieve with municipal wastewater with a small amount of industrial wastewater and drainage water. In the cases where drainage water dilutes the wastewater signiﬁcantly, or industrial wastewater adds its components to the inﬂuent, the efﬁciency of CAS can be seriously lowered due to a poor sedimentation of microbial ﬂocs. The quality of CAS efﬂuent is another important issue. Firstly, microbiological contamination of the efﬂuent may be signiﬁcant since there is no physical barrier between activated sludge and treated water. A correlation has been reported between the occurrence of eye and ear infections in humans and their contact with water where recreational use occurs (e.g., rivers or lakes) that has been receiving CAS efﬂuents [266]. This problem is even more pronounced if hospitals discharge their wastewater into sewage without treatment, because of the increased number of pathogens that may be found in raw sewage and in the efﬂuent. Also, there is a problem with speciﬁc compounds whose biodegradation depends on specialized microbial species. If such species have a slow growth rate they will be washed out with the excess sludge during the constant and fast sludge disposal rate (i.e., short SRT) of CAS treatment. As a consequence, specialized slow-growers may not develop in sufﬁcient number to degrade efﬁciently some trace pollutants. Emerging contaminants in municipal wastewater and their fate in the environment have become an issue of importance for the legislators and decision-makers. Since the design of most municipal WWTPs does not allow operation at longer SRTs, they may not be

88

J. Radjenovi´c et al.

suitable for degradation of some organic micropollutants. To overcome the limitations of conventional treatment with activated sludge, MBR technology can be successfully employed. While bacteria in activated sludge decompose and degrade organic matter from the wastewater, membrane separates them from the treated water, thus replacing the secondary settler used in CAS. The simple change from one physical separation technique to another leads to quite complex changes in the process characteristics. It affects the sludge characteristics in several ways. During CAS wastewater treatment, the bacterium present can survive in the bioreactor only in the form of ﬂocs because the ones that do not settle are discharged with the treated water. Also, due to its short SRT, it is necessary for all microorganisms to grow fast or otherwise they will be washed out from the bioreactor. In other words, microbial population in CAS is selected among fast-growing and ﬂoc-forming species. On the contrary, in the case of MBR, the bacterial ability to settle and to grow fast is of negligible importance. MBR works at much longer SRTs, which can be measured in months rather than days. As an illustration, the SRT for a full-scale MBR for municipal water treatment (Porlock, UK, 1900 m3 /d) was reported to be 30–60 days [4] in comparison with the typical SRT of the conventional process with activated sludge, which ranges from 3 to 15 days [260]. In such conditions, slowergrowing species with the ability to decompose less-biodegradable compounds have the opportunity to proliferate. In other words, in MBR selection among microorganisms is primarily made by their capability to degrade the substrate, which is also the primary purpose of the treatment process. Without demand for settling of the sludge, the F/M ratio can be set much lower, thus allowing operation at much higher MLSS concentrations, which consequently leads to higher volumetric efﬁciency of the process. Given the reduction in bioreactor volume, the elimination of secondary clariﬁers and the elimination of granular media ﬁlters, MBR typically has a much smaller footprint relative to CAS, when achieving the same discharge limits. Due to this footprint reduction, other concerns such as esthetics and odors can be more easily addressed. A low F/M ratio means that less substrate is available per unit of biomass. According to the maintenance concept introduced by Pirt [37], part of the energy contained in the supplied substrate is used for maintenance functions that are independent of growth rate. When the energy supplied to the bioreactor is lowered, the biomass ceases to grow and to utilize the substrate for maintenance. In this manner, the sludge production in the process is much lower, or does not occur at all. Very low sludge production in pilot MBR operations are reported [16], but it is often impractical for full-scale operations to keep F/M too low. The design of such plants would include very high MLSS concentrations that can promote membrane clogging, or large bioreactors, which contributes to the initial capital cost. Moreover, high MLSS concentration reduces aeration efﬁciency, which is possibly the most signiﬁcant problem with maintenance of

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

89

high MLSS concentration. Nevertheless, due to the low F/M ratio, there is a signiﬁcant decrease of sludge production in MBR in comparison to CAS, which then decreases the cost of excess sludge handling. As water reuse and reclamation increases, MBR technology can make reclaimed water more accessible by achieving the reclaimed water treatment standards in nearly a single step, thus reducing the complexity of these systems. Further, the use of reclaimed water reduces the stresses on other water bodies by reducing the need for water withdrawals and by reducing pollutant loading. In the United States, reclaimed water is being used to augment drinking water supplies. Required treatment involves multiple steps, typically culminating in RO. In wastewater treatment, RO typically is preceded by MF or UF to reduce RO membrane fouling. Therefore, implementation of an MBR process provides the ﬂexibility to install RO without the expense of a pre-treatment process. With the use of UF membranes (effective pore size of 0.04 µm) instead of CAS, most of the pathogens of concern in wastewater can be signiﬁcantly removed from the efﬂuent. The membranes provide an additional barrier to Faecal coliform, Cryptosporidium, and Giardia even in processes that use MF membranes due to a dynamic ﬁlm layer over the membrane that reduces the effective ﬁltration pore size [267]. In addition, the clarity of the efﬂuent produced by the MBR process is consistently below 0.1 nephelometric turbidity units (NTU), which is comparable to drinking water standards. This low turbidity can result in an efﬂuent highly amenable to ﬁnal disinfection using ultraviolet light. Membrane ﬁltration followed by ultra-violet (UV) treatment results in a highly disinfected efﬂuent. MBR systems do not require any more signiﬁcant operational attention, in each case much less than CAS process. A process control of an MBR system is reduced to monitoring the MLSS concentration, occasional adjustments of the chemical feed rates, and scheduling membrane recovery cleaning. Therefore, MBR is a much better solution for the small plants where CAS is non-feasible due to its requirement for constant attention and monitoring. On the other side, the cost of oxygen demand is superior in MBR. Energy consumption of MBR comes from power requirements for pumping feed water, recycling retentate, permeate suction (occasionally) and aeration [268]. The two MBR conﬁgurations have substantial differences in terms of aeration. In the side-stream conﬁguration, aeration is supplied by ﬁne bubble aerators that are highly efﬁcient for supplying oxygen to the biomass. In submerged MBRs, the aeration mode is turbulent and cross-ﬂow is generated, which scours the membrane surface and provides oxygen to the biomass. Aeration cost in the latter-mentioned conﬁguration represents around 90% of the total costs, whereas in side-stream MBR, only ∼20% derives from it [269]. However, energy consumption of the side-stream system is usually two orders of magnitude higher than that of submerged systems. These low costs of submerged MBRs are associated with low ﬂuxes, which in turn increase capital costs and

90

J. Radjenovi´c et al.

footprints. Also, packing density inﬂuences the ﬁnal cost of MBR: low packing densities of membrane modules mean that higher speciﬁc area of membrane is required to produce the same ﬂux, which increases the energy requirements. There are certain drawbacks for wider implementation of MBR technology. MBR is widely viewed as being a state-of-the-art technology but is also sometimes seen as high-risk and prohibitively costly compared to CAS and other more established technologies. MBRs were historically perceived as suitable only for small-scale plants with high operator skill requirements, and the key operating expenditure parameters such as membrane life unknown [4]. Many of these drawbacks are no longer true. Perhaps the biggest challenge to companies active in the market is to persuade decision-makers of the capability of MBRs and what beneﬁts they will undoubtedly bring to the customer. In the past, there were an insufﬁcient number of full-scale MBR treatment plants to convince decision-makers of the reliability of this advanced treatment. Presently, there are a number of examples of successful implementation of MBRs across the range of applications, and there is certainly less reason to be suspicious of this technology.

8 Application and Cost Analysis of a Membrane Bioreactor MBRs became commercially available more than 10 years ago, and their market has continued to grow. In the beginning of their application, the customers were put-off by the high-costs, appropriate operating skills, and high-level maintenance labor. However, the technology was improved, and now there are many manufacturers of MBRs, including Zenon Environmental, Kubota, Ionics/Mitsubishi rayon, USFilter, Aqua-Aerobics7Pall and Norit X-ﬂow. There are more than 2200 MBR installations in operation or under construction worldwide [270]. The main world MBR providers and characteristics of their systems are presented in Table 2. Kubota (Japan) has installed most of the world’s MBRs while Zenon (Canada) dominates in regard to installed capacity having almost four times more water treated through their membranes than Kubota [4]. Zenon has installed about 85% of North American installations, which comprise about 11% of the world’s MBR market. Asian markets (mostly in Japan and South Korea) have employed MBR technology mostly for small-scale domestic applications. In general, most of the MBRs in operation are medium or smallscale plants. More than 85% of Kubota’s MBRs have ﬂows less than 200 m3 d–1 while out of 219 MBR plants that treat municipal wastewater in North America, only 17 exceed 10 000 m3 d–1 . The largest capacity plant in operation is in North America (Traverse City, MI), which operates at 26 900 m3 d–1 , while the largest MBR worldwide currently is in Kaarst, Germany (48 000 m3 d–1 ) with total membrane area of 84 480 m2 [4]. Both of these plants operate with Zenon

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

91

Table 2 Comparison of the main MBR systems (adapted from Yang et al. [270])

Number of installationsa Membrane Conﬁguration Pore size (µm) Material Module size (m2 ) Cleaning method Cleaning frequency (min/min) Recovery method a

Kubota (Japan)

Mitsubishi-Rayon (Japan)

Zenon (Canada)

1538 (1138 + 400)

374 (170 + 204)

331 (204 + 127)

FS Vertical immersion 0.4 Chlorinated PE 0.8 Relax 1/60

HF Horizontal immersion 0.1/0.4 PE 105 Relax 2/12

HF Vertical immersion 0.04 PVDF 31.6 Backpulse and relax 0.5/15

Chlorine backwash

Chlorine backwash

Chemical soak

Municipal WWTPs + Industrial WWTPs

membranes. Leading manufacturers have exponential growth in the number of installed MBRs and their cumulative capacity in the last decade. Although the market is still dominated by Zenon and Kubota, there is a wide range of MBR systems available, however most are still at the development stage. The photographs of typical Zenon and Kubota membranes are presented in Figs. 20 and 21, respectively.

Fig. 20 Kubota ﬂat sheet MBR

92

J. Radjenovi´c et al.

Fig. 21 Zenon hollow-ﬁber MBR, ZeeWeed 500a membranes

The side-stream MBRs that were predominant before the 1990s are still present on the market but they hold a smaller share. The main manufacturers of side-stream MBR systems are Norit X-Flow, Millenniumpore and NovasepOrelis. Most of the side-stream MBRs today treat industrial wastewaters or landﬁll leachates. Municipal wastewater treatment is both the earliest and largest application of MBR, and it is predicted that this will continue to be its primary use. Due to its small footprint and potential for reuse of high-quality efﬂuent, MBR is capable of coping with population growth and limited space. For industrial applications where more stringent regulations are imposed, it provides an efﬂuent that can be safely discharged into the environment. The main applications of membrane technology reported in industry are for treatments of heavily loaded wastewaters such are oily wastewaters [62], or discharges from tanneries [245] and textile industries [271]. Promising applications also exist in treating landﬁll leachate, chlorinated solvents in manufacturing wastewater, and for groundwater remediation. Energy usage for membrane aeration is a signiﬁcant operating cost for any membrane bioreactor facility. Yoon et al. [272] calculated the total variable operational cost of MBR by summing the decreasing sludge-treatment cost and increasing aeration cost (see Fig. 22). Since minimized sludge production implies maximized aeration cost, and vice versa, they considered the existence of an optimum point between these two extreme cases, where

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

93

Fig. 22 Aeration demand for biodegradation of organic matters as a function of target MLSS and HRT. Flow rate and COD of inﬂuent were 1000 m3 day–1 and 400 mg L–1 , respectively [272]

the total operational cost is minimized. They concluded that for reasonable ranges of HRT and MLSS sludge treatment cost overwhelms aeration cost, so the most adequate strategy for MBR cost reduction would be maintenance of low sludge production conditions. High SRT in an MBR means a high MLSS concentration and low F/M ratio, which enables application of short HRT. However, sludge production is obviously inversely proportional to HRT when MLSS is mixed. The shortest HRT and the minimum sludge production cannot be achieved simultaneously. Furthermore, overall membrane cost has decreased exponentially over time for all main manufactures. Relative decrease for Kubota MBR systems and cost of Kubota membrane over the years is given in Fig. 23 [273]. As can be seen, the membrane whole-life costs decreased more than eight times in last 15 years, which has considerably closed the gap in prices between ASP and MBR technologies. Since the expected membrane lifetime has increased and enough full-scale plants have been successfully operated and proven to be reliable, the MBR technology is becoming increasingly competitive, and its future market position should be guaranteed. Despite its relative youth, MBR technology has developed over a decade to a mature product available for all sizes of application, in domestic, municipal, or industrial sector. Further improvement of the process will increase

94

J. Radjenovi´c et al.

Fig. 23 Relative cost decrease of Kubota membranes and MBR systems (adopted from Kennedy et al. [273])

its cost-effectiveness and MBR technology is expected to play a key role for wastewater treatment in the next years, in Europe as well as worldwide. To date, European countries with the highest number of full-scale MBR plants are England, France, Germany, Belgium, and the Netherlands. MBR markets are expected to open in other countries as well: in dry southern states like Spain, Greece, and Italy, due to their water shortages, and in Central and Eastern European countries (such as Hungary, Poland, Bulgaria, etc.) that will be obligated to develop their wastewater treatment technologies and adapt them to the standards and environmental legislation of the European Union. Acknowledgements This work was ﬁnancially supported by the European Union EMCO project [INCO-CT-2004-509188] and by the Spanish Ministry of Education and Science, projects EVITA (CTM2005-24254-E) and CEMAGUA (CGL2007-64551/HID). J.R. gratefully acknowledges the I3P Program (Itinerario integrado de inserción profesional), co-ﬁnanced by CSIC (Consejo Superior de Investigaciones Cientíﬁcas) and European Social Funds, for a pre-doctoral grant.

References 1. Hardt FW, Clesceri LS, Nemerow NL, Washington DR (1970) J Water Pollut Control Fed 42:2135 2. Smith CV, Gregorio D, Talcott RM (1969) 24th Annual Purdue Industrial Waste Conference. Purdue University, Lafayette, Indiana, USA 3. Yamamoto K, Hiasa M, Mahmood T, Matsuo T (1989) Water Sci Technol 21:43

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44.

95

Judd S (2006) The MBR Book. Elsevier, Oxford, UK Chang IS, Lee CH (1998) Desalin 120:221 Choi JG, Bae TH, Kim JH, Tak TM, Randall AA (2002) J Membr Sci 203:103 Wisniewski C, Grasmick A (1998) Coll Surf A: Physicochem Eng Aspects 138:403 Côté P, Thompson D (2000) Water Sci Technol 41:209 Chua HC, Arno TC, Howell JA (2002) Desalin 149:225 Lee J (2001) Water Res 35:2435 Chang IS, Lee CH, Ahn KH (1999) Separ Sci Technol 34:1743 Jiang T, Kennedy MD, van der Meer WGJ, Vanrolleghem PA, Schippers JC (2003) Desalin 157:335 Chang IS, Kim SN (2005) Proc Biochem 40:1307 Ng HY, Hermanowicz SW (2005) Water Environ Res 77:187 Cho BD, Fane AG (2002) J Membr Sci 209:391 Rosenberger S (2002) Water Res 36:413 Hernandez Rojas ME, Van Kaam R, Schetrite S, Albasi C (2005) Desalin 179:95 Rosenberger S, Laabs C, Lesjean B, Gnirss R, Amy G, Jekel M, Schrotter JC (2006) Water Res 40:710 Field RW, Wu D, Howell JA, Gupta BB (1995) J Membr Sci 100:259 Pollice A, Brookes A, Jefferson B, Judd S (2005) Desalin 174:221 Brookes A, Jefferson B, Judd S (2004) IWA 4th World Water Congress, Marrakech, Marruecos Wen X, Bu Q, Huang X (2004) Water Environment-Membrane Technology Conference, Seoul, Korea Zhang J, Chua HC, Zhou J, Fane AG (2006) J Membr Sci 284:54 Ognier S, Wisniewski C, Grasmick A (2002) Membr Technol 147:6 Jiang T, Kennedy MD, Guinzbourg BF, Vanrolleghem PA, Schippers JC (2005) Water Sci Technol 51:19 Ueda T, Hata K, Kikuoka Y, Seino O (1997) Water Res 31:489 Le-Clech P (2003) J Membr Sci 218:117 Howell JA, Chua HC, Arnot TC (2004) J Membr Sci 242:13 Liu R, Huang X, Sun YF, Qian Y (2003) Proc Biochem 39:157 Holbrook RD, Massie KA, Novak JT (2005) Water Environ Res 77:323 Kim JS, Lee CH (2003) Water Environ Res 75:300 Smith PJ, Vigneswaran S, Ngo HH, Ben-Aim R, Nguyen H (2005) J Membr Sci 255:99 Hong SP, Bae TH, Tak TM, Hong S, Randall A (2002) Desalin 143:219 Michael LS, Fikret K (2002) Bioprocess Engineering: Basic Concepts, 2nd edn. Prentice-Hall International, Upper Saddle River, NJ, USA Liwarska-Bizukojc E, Bizukojc M (2005) Proc Biochem 40:2067 Gujer W, Henze M, Mino T, Van Loosdrecht M (1999) Water Sci Technol 39:183 Pirt SJ (1965) The maintenance energy of bacteria in growing cultures. Proc R Soc Lond B Biol Sci 163:224 Shahalam ABM, Al-Smadi BM (1993) J Environ Sci Health A Environ Sci Eng 28:1751 Chaize S, Huyard A (1991) Water Sci Technol 23:1591 Muller EB (1995) Water Res 29:1179 Low EW, Chase HA (1999) Water Res 33:847 Stephenson T, Judd S, Jefferson B, Brindle K (2000) Membrane bioreactors for wastewater treatment. IWA Publishing, London Liu R, Huang X, Xi J, Qian Y (2005) Proc Biochem 40:3165 Rosenberger S, Witzig R, Manz W, Szewzyk U, Kraume M (2000) Water Sci Technol 41:269

96

J. Radjenovi´c et al.

45. Yoon SH, Kim HS, Chung YC (2002) Water Res 36:2695 46. Cicek N, Macomber J, Davel J, Suidan MT, Audic J, Genestet P (2001) Water Sci Technol 43:43 47. Macomber J, Cicek N, Suidan MT, Davel J, Ginestet P, Audic JM (2005) J Environ Eng 131:579 48. Howell JA, Arnot TC, Liu W (2003) Ann NY Ac Sci 984:411 49. Wei Y (2003) Water Res 37:4453 50. Horan N (1990) Biological Wastewater Treatment Systems: Theory and Operation. Wiley, Chichester 51. Ng HY, Hermanowicz SW (2005) Water Res 39:981 52. Wilen BM, Balmer P (1999) Doktorsavhandlingar vid Chalmers Tekniska Hogskola 391 53. Massé A, Spérandio M, Cabassud C (2006) Water Res 40:2405 54. Hasar H, Kinaci C, Ünlü A, Toðrul H, Ipek U (2004) Biochem Eng J 20:1 55. Lubbecke S, Vogelpohl A, Dewjanin W (1995) Water Res 29:793 56. Huang X, Gui P, Qian Y (2001) Proc Biochem 36:1001 57. Li H, Yang M, Zhang Y, Yu T, Kamagata Y (2006) J Biotechnol 123:60 58. Xing CH (2000) J Membr Sci 177:73 59. Knoblock MD, Sutton PM, Mishra PN, Gupta K, Janson A (1994) Water Environ Res 66:133 60. Minami K (1994) Desalin 98:273 61. Ross WR, Barnard JP, Strohwald NKH, Grobler CJ, Sanetra J (1992) Water Sci Technol 25:27 62. Zaloum R, Lessard S, Mourato D, Carriere J (1994) Water Sci Technol 30:21 63. Van Loosdrecht MCM, Henze M (1999) Water Sci Technol 39:107 64. Rosenberger S, Kraume M (2002) Desalin 146:373 65. Pollice A, Laera G (2005) Water Sci Technol 52:369 66. Metcalf and Eddy Inc (1991) Wastewater Engineering; Treatment, Disposal, and Reuse, 3rd edn. McGraw-Hill Inc, New York 67. Ryan FJ (1959) J Gen Microbiol 21:530 68. Mason CA, Hamer G (1987) Appl Microbiol Biotechnol 25:577 69. Canales A, Pareilleux A, Rols JL, Goma G, Huyard A (1994) Water Sci Technol 30:97 70. Yoon SH, Kim HS, Lee S (2004) Proc Biochem 39:1923 71. He SB, Xue G, Wang BZ (2005) J Dong Hua Uni (Eng Ed) 22:100 72. Sakai Y, Fukase T, Yasui H, Shibata M (1997) Water Sci Technol 36:163 73. Song KG, Choung YK, Ahn KH, Cho J, Yun H (2003) Desalin 157:353 74. Yeom IT, Lee KR, Choi YG, Kim HS, Kwon JH, Lee UJ, Lee YH (2005) Water Sci Technol 52:201 75. Zhang B, Yamamoto K (1996) Water Sci Technol 34:295 76. Chiemchaisri C, Yamamoto K (1994) J Membr Sci 87:119 77. Marsili-Libelli S, Tabani F (2002) Water Res 36:1181 78. Gao M, Yang M, Li H, Yang Q, Zhang Y (2004) J Biotechnol 108:265 79. Doyle MP, Herman JG, Dykstra RL (1985) J Free Radic Biol Med 1:145 80. Hansson LE, Nyren O, Bergstrom R, Wolk A, Lindgren A, Baron J, Adami HO (1994) Int J Cancer 57:638 81. Barnes D, Bliss PJ (1983) In: Spon E, Spon FN (eds) Biological control of nitrogen in wastewater treatment. Spon Pr, New York 82. Gao M, Yang M, Li H, Wang Y, Pan F (2004) Desalin 170:177 83. Chiemchaisri C, Wong YK, Urase T, Yamamoto K (1993) Filtr Sep 30:247 84. Muller EB (1994) PhD Thesis, University of California, Santa Barbara, CA

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

97

85. Chung J, Kim Y, Lee DJ, Shim H, Kim JO (2006) J Environ Eng Sci 23:981 86. Lesjean B, Gnirss R, Adam C, Kraume M, Luck F (2003) Water Sci Technol 48:87 87. Smolders GJF, Van der Meij J, Van Loosdrecht MCM, Heijnen JJ (1995) Biotechnol Bioeng 47:277 88. Beun JJ, Heijnen JJ, Van Loosdrecht MCM (2001) Biotechnol Bioeng 75:82 89. Ujang Z, Salim MR, Khor SL (2002) Water Sci Technol 46:193 90. Seo GT, Lee TS, Moon BH, Lim JH, Lee KS (2000) Water Sci Technol 41:217 91. Chuang SH, Ouyang CF, Wang YB (1996) Water Res 30:2961 92. Barker PS, Dold PL (1996) Water Sci Technol 34:43 93. Ueda T, Hata K, Kikuoka Y (1996) Water Sci Technol 34:189 94. Rosenberger S, Krüger U, Witzig R, Manz W, Szewzyk U, Kraume M (2002) Water Res 36:413 95. Al-Malack MH (2006) J Membr Sci 271:47 96. Chen GH, Yip WK, Mo HK, Liu Y (2001) Water Res 35:1029 97. Yoon TI, Lee HS, Kim CG (2004) J Membr Sci 242:5 98. Wang JC, Park JK, Whang LM (2001) Water Environ Res 73:704 99. Pollice A, Laera G, Blonda M (2004) Water Res 38:1799 100. Côté P, Buisson H, Pound C, Arakaki G (1997) Desalin 113:189 101. Kim H, Kim HS, Yeom IT, Chae YB (2005) Water Sci Technol 51:465 102. Choi E, Eum Y, Gil KI, Oa SW (2004) Water Sci Technol 49:97 103. Carucci A, Dionisi D, Majone M, Rolle E, Smurra P (2001) Water Res 35:3833 104. Cicek N, Franco JP, Suidan MT, Urbain V, Manem J (1999) Water Environ Res 71:64 105. De Wever H, Van Roy S, Dotremont C, Müller J, Knepper T (2004) Water Sci Technol 50:219 106. Clara M, Strenn B, Kreuzinger N, Ausserleitner M (2004) Water Sci Technol 50:29 107. Rui L, Xia H, Ruopeng L, Yi Q (2001) Huanjing Kexue (China Environ Sci) 22:20 108. Wagner J, Rosenwinkel KH (2000) Water Sci Technol 41:251 109. Dufresne R, Lavallee HC, Lebrun RE, Lo SN (1996) International Environmental Conference, TAPPI Press, Norcross, GA, USA 110. Suwa Y, Suzuki T, Toyohara H, Yamagishi T, Urushigawa Y (1992) Water Res 26:1149 111. Bailey AD, Hansford GS, Dold PL (1994) Water Res 28:297 112. Côté P, Buisson H, Praderie M (1998) Water Sci Technol 38:437 113. Sun DD, Hay CT, Khor SL (2006) Desalin 195:209 114. Fan XJ (1996) Water Sci Technol 34:129 115. Chu L, Li S (2006) Separ Puriﬁc Technol 51:173 116. Gander MA (2000) Water Sci Technol 41:205 117. Germain EE (2005) Biotechnol Bioeng 90:316 118. Nagaoka H, Ueda S, Miya A (1996) Water Sci Technol 34:165 119. Lim BS, Choi BC, Yu SW, Lee CG (2007) Desalin 202:77 120. Leong LYC (1983) Water Sci Technol 15:91 121. Maier RMP, Jan L, Gerba CP (2000) Environmental Microbiology. Academic Press, San Diego 122. Jofre J, Olle E, Ribas F, Vidal A, Lucena F (1995) Appl Environ Microbiol 61:3227 123. Reinthaler FF, Posch J, Feierl G, Wust G, Haas D, Ruckenbauer G, Mascher F, Marth E (2003) Water Res 37:1685 124. Koivunen J, Siitonen A, Heinonen-Tanski H (2003) Water Res 37:690 125. Madaeni SS (1995) J Membr Sci 102:65 126. Ottoson J, Hansen A, Bjorlenius B, Norder H, Stenstrom TA (2006) Water Res 40:1449 127. McGahey C, Olivieri VP (1993) Water Sci Technol 27:307 128. Ueda T, Horan NJ (2000) Water Res 34:2151

98 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161. 162. 163. 164. 165.

J. Radjenovi´c et al. van Voorthuizen EM, Ashbolt NJ, Schafer AI (2001) J Membr Sci 194:69 Farahbakhsh K, Smith DW (2004) Water Res 38:585 Cooper RC, Straube D (1979) Water Sew Works 166 Lv W, Zheng X, Yang M, Zhang Y, Liu Y, Liu J (2006) Proc Biochem 41:299 Gerba CP (1984) Adv Appl Microbiol 30:133 Günder B (1999) PhD Thesis. Das membranbelebungsverfahren in der kommunalen Abwasserreinigung, University of Stuttgart, Stuttgart European Council (1991a) Directive concerning urban wastewater treatment. Off J Eur Comm L135, (91/271/EEC) Halling-Sørensen B, Nors Nielsen S, Lanzky PF, Ingerslev F, Holten Lützhøft HC, Jørgensen SE (1998) Chemosphere 36:357 Clara M, Strenn B, Kreuzinger N, Kroiss H, Gans O, Martinez E (2005) Water Res 39:4797 Vieno NM (2005) Environ Sci Technol 39:8220 Petrovi´c M, González S, Barceló D (2003) Trends Anal Chem 22:685 Kolpin DW, Furlong ET, Meyer MT, Thurman EM, Zaugg SD, Barber RB, Buxton HT (2002) Environ Sci Technol 36:1202 Buser HR (1999) Environ Sci Technol 33:2529 Heberer T (2002) Toxicol Lett 131:5 Ternes TA (1998) Water Res 32:3245 Metcalfe CD, Koenig BG, Bennie DT, Servos M, Ternes TA, Hirsch R (2003) Environ Toxicol Chem 22:2872 Castiglioni S, Bagnati R, Fanelli R, Pomati F, Calamari D, Zuccato E (2006) Environ Sci Technol 40:357 Joss A, Keller E, Alder AC, Göbel A, McArdell CS, Ternes T, Siegrist H (2005) Water Res 39:3139 Giger W, Alder AC, Golet EM, Kohler HPE, McArdell CS, Molnar E, Siegrist H, Suter MJF (2003) Chimia 57:485 Jjemba PK (2006) Ecotoxicol Environ Saf 63:113 http://www.syrres.com/; last visited: February 2007 Joss A, Zabczynski S, Göbel A, Hoffmann B, Löfﬂer D, McArdell CS, Ternes TA, Thomsen A, Siegrist H (2006) Water Res 40:1686 Lesjean B, Gnirrs R, Buisson H, Keller S, Tazi-Pain A, Luck F (2005) Water Sci Technol 52:453 Göbel A, McArdell CS, Joss A, Siegrist H, Giger W (2007) Sci Total Environ 372:361 Kimura K, Hara H, Watanabe Y (2005) Desalin 178:135 Quintana JB (2005) Water Res 39:2654 Radjenovi´c J, Petrovi´c M, Barceló D (2007) Anal Bioanal Chem 387:1365 Radjenovi´c J, Petrovi´c M, Barceló D (2007) Trends Anal Chem, in press [available online 22 October 2007] Clara M (2004) Water Res 38:947 Higson FK (1992) Adv Appl Microbiol 37:135 Winkler M, Lawrence JR, Neu TR (2001) Water Res 35:3197 Bound JP, Voulvoulis N (2006) Water Res 40:2885 Gros M, Petrovi´c M, Barceló D (2006) Talanta 70:678 Schwab BW, Hayes EP, Fiori JM, Mastrocco FJ, Roden NM, Cragin D, Meyerhoff RD, D’Aco VJ, Anderson PD (2005) Regul Toxicol Pharmacol 42:296 Stumpf M, Ternes TA, Haberer K, Baumann W (1998) Vom Wasser 91:291 Zwiener C, Glauner T, Frimmel FH (2000) J Sep Sci 23:474 Zwiener C, Frimmel FH (2003) Sci Total Environ 309:201

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology

99

166. Thomas PM, Foster GD (2005) Environ Toxicol Chem 24:25 167. Heberer T (2002) J Hydrol 266:175 168. González S, Müller J, Petrovi´c M, Barceló D, Knepper TP (2006) Environ Poll 144: 926 169. http://poseidon.bafg.de/; last visited: February 2007 170. Urase T, Kikuta T (2005) Water Res 39:1289 171. Strenn B, Clara M, Kreuzinger N, Gans O (2004) Water Sci Technol 50:269 172. Nakada N, Tanishima T, Shinohara H, Kiri K, Takada H (2006) Water Res 40:3297 173. Carballa M, Omil F, Lema JM (2005) Water Res 39:4790 174. Göbel AMCS, Suter MJF, Giger W (2004) Anal Chem 76:4756 175. Drillia P (2005) J Hazard Mat 122:259 176. Khan SJ, Ongerth JE (2002) Water Sci Technol 46:105 177. Urase T, Kagawa C, Kikuta T (2005) Desalin 178:107 178. Jones OAH, Voulvoulis N, Lester JN (2006) Arch Environ Contam Toxicol 50:297 179. Lester JN, MacLeod C, Birkett JW (1999) Microbiology and Chemistry for Environmental Scientists and Engineers. E and FN Spon, London 180. Kumagai T, Inoue T, Mihara Y, Ebina K, Yokota K (2006) Biol Pharm Bull 29:183 181. Boxall ABA, Blackwell P, Cavallo R, Kay P, Tolls J (2002) Toxicol Lett 131:19 182. Schwarzenbach RP, Gschwend PM, Imboden DM (1993) Environmental Organic Chemistry. Wiley, New York 183. Göbel A, Thomsen A, McArdell CS, Joss A, Giger W (2005) Environ Sci Technol 39:3981 184. Ternes TA (2004) Water Res 38:4075 185. Körner W, Bolz U, Süßmuth W, Hiller G, Schuller W, Hanf V, Hagenmaier H (2000) Chemosphere 40:1131 186. Ternes TA, Kreckel P, Mueller J (1999) Sci Total Environ 225:91 187. Jobling S, Nolan M, Tyler CR, Brighty G, Sumpter JP (1998) Environ Sci Technol 32:2498 188. Loos R, Hanke G, Umlauf G, Eisenreich SJ (2007) Chemosphere 66:690 189. Larsson DGJ, Adolfsson-Erici M, Parkkonen J, Pettersson M, Berg AH, Olsson PE, Forlin L (1999) Aq Toxicol 45:91 190. Routledge EJ, Sheahan D, Desbrow C, Brighty GC, Waldock M, Sumpter JP (1998) Environ Sci Technol 32:1559 191. Ahel M, Molnar E, Ibric S, Giger W (2000) Water Sci Technol 42:15 192. Sumpter JP (1998) Toxicol Lett 102–103:337 193. Sumpter JP (1998) Arch Toxicol Suppl 20:143 194. Sumpter JP (1995) Toxicol Lett 82–83:737 195. Guillette LJ Jr, Gross TS, Masson GR, Matter JM, Percival HF, Woodward AR (1994) Env Health Persp 102:680 196. Snyder SA, Kelly KL, Grange AH, Sovocool GW, Snyder EM, Giesy JP (2001) ACS Symp Ser 791:116 197. Länge R, Hutchinson TH, Croudace CP, Siegmund F, Schweinfurth H, Hampe P, Panter GH, Sumpter JP (2001) Environ Toxicol Chem 20:1216 198. Lauritzen C (1988) Natural and synthetic sexual hormones-biological basis and medical treatment principles. Grundlagen und Klinik der menschlichen Fortpﬂanzung, Berlin 199. D’Ascenzo G, Di Corcia A, Gentili A, Mancini R, Mastropasqua R, Nazzari M, Samperi R (2003) Sci Total Environ 302:199 200. Reddy S, Iden CR, Brownawell BJ (2005) Anal Chem 77:7032

100

J. Radjenovi´c et al.

201. Servos MR, Bennie DT, Burnison BK, Jurkovic A, McInnis R, Neheli T, Schnell A, Seto P, Smyth SA, Ternes TA (2005) Sci Total Environ 336:155 202. Tilton F, Benson WH, Schlenk D (2002) Aq Toxicol 61:211 203. Matsui S, Takigami H, Matsuda T, Taniguchi N, Adachi J, Kawami H, Shimizu Y (2000) Water Sci Technol 42:173 204. Witters HE, Vangenechten C, Berckmans P (2001) Water Sci Technol 43:117 205. Joss A, Andersen H, Ternes T, Richle PR, Siegrist H (2004) Environ Sci Technol 38:3047 206. Murk AJ, Legler J, Van Lipzig MMH, Meerman JHN, Belfroid AC, Spenkelink A, Van der Burg B, Rijs GBJ, Vethaak D (2002) Environ Toxicol Chem 21:16 207. Andersen H, Siegrist H, Halling-Sørensen B, Ternes TA (2003) Environ Sci Technol 37:4021 208. Kirk LA, Tyler CR, Lye CM, Sumpter JP (2002) Environ Toxicol Chem 21:972 209. Carballa M, Omil F, Lema JM, Llompart M, García-Jares C, Rodríguez I, Gómez M, Ternes T (2004) Water Res 38:2918 210. Wintgens T, Gallenkemper M, Melin T (2004) Water Sci Technol 50:1 211. Kuster M, López De Alda MJ, Barceló D (2004) Trends Anal Chem 23:790 212. Clara M, Strenn B, Saracevic E, Kreuzinger N (2004) Chemosphere 56:843 213. Schäfer AI, Mastrup M, Jensen RL (2002) Desalin 147:243 214. Baronti C, Curini R, D’Ascenzo G, Di Corcia A, Gentili A, Samperi R (2000) Environ Sci Technol 34:5059 215. Ying GG, Williams B, Kookana R (2002) Environ Internat 28:215 216. Renner R (1997) Environ Sci Technol 31:316A 217. Solé M, López De Alda MJ, Castillo M, Porte C, Ladegaard-Pedersen K, Barceló D (2000) Environ Sci Technol 34:5076 218. Brunner PH, Capri S, Marcomini A, Giger W (1988) Water Res 22:1465 219. Ruiz F, Prats D, Rico C (1989) Organic Contaminants in Wastewater, Sludge and Sediment, p 124. In: Proc COST Workshop, Brussels, 1988 220. De Henau H, Matthijs E, Namkung E (1989) Organic contaminants in waste water, sludge and sediment, p 5. In: Proc COST Workshop, Brussels, 1988 221. Hashim MA, Kulandai J, Hassan RS (1992) J Chem Technol Biotechnol 54:207 222. Perales JA, Manzano MA, Sales D, Quiroga JM (1999) Bull Environ Contam Toxicol 63:94 223. Reemtsma T, Weiss S, Mueller J, Petrovi´c M, González S, Barceló D, Ventura F, Knepper TP (2006) Environ Sci Technol 40:5451 224. Berna JL, Moreno A, Ferrer J (1991) J Chem Technol Biotechnol 50:387 225. Prats D, Ruiz F, Vazquez B, Zarzo D, Berna JL, Moreno A (1993) Environ Toxicol Chem 12:1599 226. Terzi´c S, Matosi´c M, Ahel M, Mijatovi´c I (2005) Water Sci Technol 51:447 227. González S, Petrovi´c M, Barceló D (2007) Chemosphere 67:335 228. Bernhard M, Müller J, Knepper TP (2006) Water Res 40:3419 229. Li HQ (2000) J Chromatogr 889:155 230. González S, Petrovi´c M, Barceló D (2004) J Chromatogr A 1052:111 231. Komori K, Okayasu Y, Yasojima M, Suzuki Y, Tanaka H (2006) Water Sci Technol 53:27 232. Ahel M, Giger W (1985) Anal Chem 57:1577 233. Bennie DT, Sullivan CA, Lee H-B, Peart TE, Maguire RJ (1997) Sci Total Environ 193:263 234. Blackburn MA, Kirby SJ, Waldock MJ (1999) Marine Pollut Bull 38:109 235. Ferguson PL, Iden CR, Brownawell BJ (2001) Environ Sci Technol 35:2428

Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology 236. 237. 238. 239. 240. 241. 242. 243. 244. 245. 246. 247. 248. 249. 250.

251. 252. 253. 254. 255. 256. 257. 258. 259. 260. 261. 262. 263. 264.

265. 266. 267. 268. 269. 270. 271. 272. 273.

101

Ying GG, Williams B, Kookana R (2002) Environ Internat 28:215 Ahel M, Giger W, Koch M (1994) Water Res 28:1131 Soto AM, Justicia H, Wray JW, Sonnenschein C (1991) Env Health Persp 92:167 Schröder HF (2002) Water Sci Technol 46:57 Altenbach B, Giger W (1995) Anal Chem 67:2325 Storm T, Reemtsma T, Jekel M (1999) J Chromatogr A 854:175 Jungclaus GA, Games LM, Hites RA (1976) Anal Chem 48:1894 Reemtsma T, Fiehn O, Kalnowski G, Jekel M (1995) Environ Sci Technol 29:478 Reemtsma T (2000) Rapid Comm Mass Spectrom 14:1612 Reemtsma T, Zywicki B, Stueber M, Kloepfer A, Jekel M (2002) Environ Sci Technol 36:1102 Kloepfer A, Gnirss R, Jekel M, Reemtsma T (2004) Water Sci Technol 50:203 Hill NP, McIntyre AE, Perry R, Lester JN (1986) Water Res 20:45 Saleh FY, Lee GF, Wolf HW (1980) J Water Pollut Control Fed 52:19 Mangat SS, Elefsiniotis P (1999) Water Res 33:861 Wilson GJ, Suidan, MT, Maloney SW, Brenner RC (1997) Proc of WEFTEC ’97 70th Annual Conference and Exposition of the Water Environment Federation, Chicago, IL Knepper TP, Sacher F, Lange FT, Brauch HJ, Karrenbrock F, Roerden O, Lindner K (1999) Waste Manage 19:77 Fries E, Püttmann W (2001) J Environ Monit 3:621 Lange FT, Wenz M, Brauch HJ (1995) J High Resol Chromatogr 18:243 Knepper TP (2003) Trends Anal Chem 22:708 Glassmeyer ST, Furlong ET, Kolpin DW, Cahill JD, Zaugg SD, Werner SL, Meyer MT, Kryak DD (2005) Environ Sci Technol 39:5157 Weiss S, Jakobs J, Reemtsma T (2006) Environ Sci Technol 40:7193 Chang JS, Chang CY, Chen AC, Erdei L, Vigneswaran S (2006) Desalin 191:45 Fatone F, Bolzonella D, Battistoni P, Cecchi F (2005) Desalin 183:395 McArdell CS, Joss A, Alder AC, Felis E, Giger W, Göbel A, Lofﬂer D, Siegrist H, Zabczynski S, Ternes TA (2005) SETAC Europe Meeting, Lille, France Tchobanoglous G, Burton FL, Stensel HD (2003) Wastewater Engineering: Treatment and Reuse. Metcalf & Eddy Inc, 4th edn. McGraw Hill, Toronto Davis RD, Hall JE (1997) Eur Water Pollut Control 7:9 Spellman FR (1997) Wastewater biosolids to compost. Technomic Publishing Co, Lancaster, PA Metcalf and Eddy Inc (1991) Wastewater Engineering: Treatment, Disposal and Reuse, 3rd edn. McGraw-Hill, New York Urbain V, Trouve E, Manem J (1994) Membrane bioreactors for municipal wastewater treatment and recycling, p 317. In: Proc 67th Annual Conference and Exposition of the Water Environmental Federation, Vol 1 Guender B (2001) The Membrane Coupled Activated Sludge Process in Municipal Wastewater Treatment. Technomic Publishing Co, Lancaster, PA Prüss A (1998) Int J Epidemiol 27:1 Ueda T (2001) Bull Nat Inst Agric Sci (Japan) 40:1 Owen G, Bandi M, Howell JA, Churchouse SJ (1995) J Membr Sci 102:77 Gander M, Jefferson B, Judd S (2000) Separ Puriﬁc Technol 18:119 Yang W, Cicek N, Ilg J (2006) J Membr Sci 270:201 Rozzi A, Malpei F, Bianchi R, Mattioli D (2000) Water Sci Technol 41:189 Yoon SH, Kim HS, Yeom IT (2004) Water Res 38:37 Kennedy S, Churchouse SJ (2005) Wastewater Europe Conference. Milan, Italy

Hdb Env Chem Vol. 5, Part S/2 (2008): 103–125 DOI 10.1007/698_5_100 © Springer-Verlag Berlin Heidelberg Published online: 8 January 2008

Removal of Emerging Contaminants in Water Treatment by Nanofiltration and Reverse Osmosis Branko Kunst (u) · Kreˇsimir Koˇsuti´c Faculty of Chemical Engineering and Technology, University of Zagreb, PO Box 177, 10000 Zagreb, Croatia [email protected] 1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

103

2

Reverse Osmosis and Nanoﬁltration . . . . . . . . . . . . . . . . . . . . . .

105

3

Predictive Modelling of RO and NF . . . . . . . . . . . . . . . . . . . . . .

106

4 4.1 4.2 4.3

Solute Rejection Mechanisms by the NF/RO Membrane Size Exclusion . . . . . . . . . . . . . . . . . . . . . . . Charge Exclusion . . . . . . . . . . . . . . . . . . . . . Physicochemical Interactions . . . . . . . . . . . . . . .

. . . .

. . . .

108 108 111 112

5 5.1

. . . . . . . . . . .

116

5.2 5.3

Recent Findings . . . . . . . . . . . . . . . . . . . . . . Removal of Emerging Contaminants at Submicrogram per Liter Concentrations . . . . . . . Removal of Antibiotics . . . . . . . . . . . . . . . . . . The Latest Results on the Organics Removal by NF/RO

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

116 118 119

6

Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

123

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

123

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

Abstract The general rules established in abundant studies on removal of conventional pollutants from waters by reverse osmosis and nanoﬁltration were reconsidered in this contribution in order to determine their applicability for the removal of emerging contaminants. The new results on removal of organic pollutants at trace concentrations from waters and wastewaters are presented, together with the assessment of the extent to which the general ﬁndings and rules of NF/RO are valid and conﬁrmed in the recent studies. Some speciﬁc ﬁndings and conclusions from studies on the removal of emerging micropollutants are also highlighted. Keywords Removal of emerging contaminants · Water treatment · Nanoﬁltration · Reverse osmosis · Rejection mechanisms

1 Introduction Emerging contaminants are organic compounds found in the commercial products used in large quantities in everyday life, such as pharmaceuti-

104

B. Kunst · K. Koˇsuti´c

cally active compounds (PhACs), endocrine disrupting compounds (EDC), pesticides, human and veterinary drugs, surfactants, textile dyes, etc. Their widespread presence in low concentrations in efﬂuents and generally in the aquatic environment has recently become an environmental problem, and the public has expressed human-health concerns regarding the occurrence of these contaminants in water reuse projects. Some of them are highly persistent. For others, the high degradation rate has been counterbalanced by their continuous introduction into the aquatic environment. Although the health effects of the consumption of these micropollutants at low concentration levels as of yet have not been fully elucidated, it is agreed that drinking water should be relatively free of such compounds. It is therefore essential to prevent the emerging contaminants from entering the environment, or to decrease signiﬁcantly their concentrations by adequate treatment methods. As conventional water treatment processes generally seem to be ineffective in the removal of emerging contaminants, advanced methods such as membrane processes, carbon adsorption, and oxidative treatments should be used for this purpose. The ability of membrane separation processes, especially reverse osmosis (RO) and nanoﬁltration (NF), to remove efﬁciently and economically various substances makes them a natural choice for removal of the emerging contaminants from water. Membrane separation processes have been used in drinking water treatment for almost 30 years, and in the last decade they have been increasingly applied in process water and wastewater treatments, particularly where high quality standards of a product are required [1–6]. The effective removal of organic compounds from various waters by NF/RO has been a major challenge for membrane scientists, as conﬁrmed by the numerous studies that have been recently reviewed [7, 8]. These studies investigated the removal of different solutes and parameters that inﬂuence the removal, and contributed to better understanding of membrane retention mechanisms. The majority of the studies on the removal of conventional pollutants from various waters by NF/RO were limited to some targeted compounds and they usually covered binary aqueous solutions and synthetic model waters. Wastewaters as a rule contain a great variety of organic micropollutants in very low, submicrogram per liter concentrations. Therefore, in studying the rejection efﬁciency of NF/RO membranes in removing organic micropollutants, experiments should be conducted at environmentally realistic concentrations [9], and using the recently developed analytical methods capable of determining many trace compounds present in wastewaters. With such a distinction in mind, it seems natural to reconsider the ﬁndings and general rules that were obtained and established in the abundant studies on removal of conventional pollutants from waters by reverse osmosis and nanoﬁltration. After that, new results on removal of micropollutants (EMCOs) from wastewaters are presented, together with the assessment of the extent

Removal of Emerging Contaminants by NF/RO

105

to which the general ﬁndings and rules of NF/RO are valid and conﬁrmed in the latest studies. Some speciﬁc ﬁndings and conclusions from studies on the removal of emerging micropollutants are also highlighted.

2 Reverse Osmosis and Nanofiltration In pressure-driven membrane processes, feed water is forced through a membrane by high pressure exerted on the feed water membrane side, and separated into two streams, permeate and retentate (concentrate). The two resulting streams have widely different solute concentrations, the permeate being almost pure water and the retentate containing all the suspended particles and most of the dissolved components of the feed water. A good separation membrane, as was ﬁrstly engineered by Loeb and Sourirajan [10], has an asymmetric (anisotropic in vertical cross section) structure with a thin dense selective layer (skin) on the membrane surface, supported by the much thicker porous layer made from the same or different material. The ﬁrst successful membranes were made from cellulose acetate, and later on various polymers and their combinations known as composite membranes, built from the other cellulosics, aromatic polyamides, polysulfone, polyethersulfone, polyacrylonitrile, polyvinylideneﬂuoride, and inorganic materials such as ceramics, were utilized. According to the properties (porosity, electrical charge) of the used membrane, there are four pressure-driven membrane processes: microﬁltration, ultraﬁltration, nanoﬁltration, and reverse osmosis. In microﬁltration (MF), larger suspended particles are retained on the membrane, in ultraﬁltration (UF) smaller suspended particles and colloids are separated out, and in nanoﬁltration (NF) and reverse osmosis (RO) larger or smaller dissolved particles, ions and molecules, are removed from water. The use of membranes of various porosities calls for adequate operating pressures, so the applied transmembrane pressure in a membrane operation also typiﬁes the kind of membrane process. The economical arguments in the process and membrane selection are very important. It is sometimes better, for example, to use a reliable nanoﬁltration membrane at slightly less than maximal possible solute rejection and at appreciably lower operating pressures to save energy than it is to use a more tight reverse osmosis membrane. As the pore size range of the nanoﬁltration membranes’ skin (approximately 0.5–2 nm) does not signiﬁcantly differ from the pore size range of the reverse osmosis membranes’ skin (approximately 0.2–1.0 nm), the basic difference between an NF and a RO membrane is in their rejection of ions. It is typical for NF membranes to reject almost completely di- and multivalent ions, with the retention of univalent ions being less than 70%. The rejection of different ions by a RO membrane varies less, and is mostly governed by the

106

B. Kunst · K. Koˇsuti´c

ions’ hydrated size. In the case of organic molecules, there is no clear boundary between the rejections of nanoﬁltration and reverse osmosis membranes, so in this contribution they will be considered together as NF/RO membranes. During more than 30 years of reverse osmosis and nanoﬁltration usage for the water treatment purposes, much work has been done on the preparation of the highly selective and productive membranes, on solving the solute permeation mechanism through a membrane, and on selected applications of the membrane processes mostly in the ﬁeld of treatments for drinking water. Still, there are unsolved problems and ambiguities concerning mostly the retention mechanisms of different solutes and their modelling in case of NF/RO membranes. Early expectations and efforts of some researchers to ﬁnd a membrane that would completely reject solute molecules and ions were illusory. There is not a process of membrane separation that can produce pure water. The membrane is a physical barrier, which allows permeation, under a driving force, of more or less of each of the feed mixture components, with solutes permeating the membrane much slower than water. This means that permeate will always contain traces of the solutes from the feed, i.e., that a truly semipermeable membrane does not exist. S. Sourirajan, one of the inventors [2] of the ﬁrst asymmetric membrane, even stated that “the word semipermeable contributes nothing to the scientiﬁc understanding of reverse osmosis”.

3 Predictive Modelling of RO and NF In research on the membrane separation process performance, scientists tried very early to ﬁnd theoretical expressions and equations that would describe and predict solute rejection by the NF/RO membranes, as well as parameters affecting the separation process. Such helpful theoretical expressions, called models, would be based on available characteristics of a membrane and solutes to be rejected. A successful model would give a satisfactory agreement of the data calculated by theoretical equations with those obtained experimentally. A basic problem in modelling the NF/RO separation process is a description of the NF/RO membrane itself: is it a porous, or is it a homogeneous phase. Mulder [3] describes the NF/RO membrane as being intermediate between porous and nonporous type barrier. If so, how can hydrodynamic and charge interactions inside very narrow pores (voids) be described? A consequence of this uncertainty is various models proposed for the solute separation by NF/RO membranes. Early attempts to describe mass transport through a membrane were done by means of irreversible thermodynamics. Kedem and Katchalsky [11] and Spiegler and Kedem [12] developed the equations on material trans-

Removal of Emerging Contaminants by NF/RO

107

port through a membrane considering membrane system as a whole (without speciﬁed parameters), behaving as a black box. This means that basic physical properties like membrane pore structure, charge density, and a mechanism of solute transport were not deﬁned in this approach. The surface force–pore ﬂow model [2, 13] quantitatively represents the notion that the membrane separation mechanism is determined by both the surface effects and the ﬂuid transport through the pores. The preferentially sorbed water on the membrane surface is forced through the membrane’s capillary pores by the applied pressure, and solute molecules are transported through the pores by diffusion under the chemical potential gradient. Some early models proposed for RO, called solution-diffusion models, were based on the idea of dissolution of the solute molecules in a barrier membrane, followed by their diffusion through it. These models were recently reviewed and modiﬁed [14] including intermediate steady-state or unsteady-state organic solute adsorption. The modiﬁed models allow for the understanding of the mechanisms of organics–membrane interactions. The models using a hydrodynamic approach assume a geometrical model of a membrane and solutes (size, shape and charge of the membrane pores and of the solutes molecules), and describe the transport equations based on the model assumptions. One such model [15] considers a porous membrane as a bundle of narrow cylindrical pores through which the transport of rigid solute molecules occurs hindered by steric interactions due to molecular friction in the constrained space of a pore. In case of the charged NF membranes and transport of the charged particles assuming to be the point charges, the transport along the pores is described by the extended Nernst–Planck equation. Such space-charge models [16] and their modiﬁcations when surface charge density is small and pores are sufﬁciently narrow, were found convenient for the narrow pore NF membranes [17–19]. Van der Bruggen and Vandecasteele [20] recently proposed a porous membrane model based on Spiegler–Kedem transport equations describing the ﬂux of uncharged organics through NF membranes. Retention curves were determined from the membrane pore size distribution, the experimental membrane water ﬂux, and the experimentally found diffusion parameter of a membrane. As an appropriate size parameter was needed, the molecular weight values related to the effective solute diameters were used in the calculations. Bowen and Wellfoot [21] critically assessed previous theoretical descriptions of membrane nanoﬁltration. A new two-parameter model (pore radius and membrane charge) for electrolyte rejection has been developed, and it includes dielectric exclusion in the form of an energy barrier to ion partitioning into the pores. The predicted values of effective membrane charge density were signiﬁcantly reduced in magnitude, and their variation with concentration for divalent salts was in better agreement with physical models of ion adsorption.

108

B. Kunst · K. Koˇsuti´c

Each one of the models can be applied more or less successfully to predict mass transfer of some group of solutes through certain membranes, but selecting the appropriate modelling approach for the NF/RO separation process is not an easy task. In spite of the abundant studies of the organic solutes, rejection by the NF/RO membranes and the encouraging results by the proposed modelling efforts, the knowledge necessary for the truly successful prediction model is still limited. Therefore, as Bellona et al. stated in their review [7] “a comprehensive understanding of the factors affecting the mass transfer of solutes through high-pressure membranes is invaluable to the development of a predictive model”. And in this respect the authors expressed a belief that “an assessment of the knowledge base regarding the factors affecting the transfer of solutes through NF/RO membranes was helpful”. A main question that should be answered is one of the complex rejection mechanism of a solute by a NF/RO membrane, which would take into account the speciﬁc physical properties of both, the solute and the membrane.

4 Solute Rejection Mechanisms by the NF/RO Membrane This part of the paper deals with basic mechanisms controlling the rejection of various solutes by the NF/RO membrane. It is well established that they can be described by one of the following three principles: • size exclusion (sieving, steric effects), • charge (electrical, Donnan) exclusion, and • rejection governed by physicochemical interactions between solute, solvent and a membrane. In reality, there is always an interaction of these effects, although in many cases one of them prevails. Some rejection mechanisms are quite well explained; for example, the major rejection mechanism of solutes larger than the membrane pore size is physical sieving. Other mechanisms of rejection such as charge exclusion are known to dominate in case of inorganic ions as solutes. The rejection of ionizing organics by the charged NF membranes is not quite clearly understood yet, as well as the very important adsorptive and hydrophobic interactions between membrane and solute, operative in uncharged organics rejection. 4.1 Size Exclusion Size exclusion (sieving, steric effects) is the basic mechanism controlling the rejection of organic solutes by a RO/NF membrane. It is based on the simple notion that solutes larger than the pore size of the membrane cannot perme-

Removal of Emerging Contaminants by NF/RO

109

ate through it. The assumptions for this mechanism would be the knowledge of data on spherical size of the solute molecules and data on the diameter of cylindrical pores of various membranes. As such simpliﬁed assumptions in reality do not hold, both actual, complex quantities should be more clearly described. There are several parameters representing each of these quantities and before their use, their signiﬁcance together with the methods of their determination should be explained. The size of a solute molecule can be expressed by its molecular mass (weight), by its diameter, by the effective molecular size (which takes into account the shape of the molecule), by the Stokes (hydrodynamic or solvated) diameter, or it might be given by some other size and morphology parameters like molecular width or molecular length. Molecular weight is available for all the solutes, and it was shown [22] that the rejection of a number of undissociated organic compounds by the ultra-low pressure RO membranes increases linearly with the molecular weight. Van der Bruggen et al. [23] asserted that molecular weight is not a suitable size parameter for nanoﬁltration, because it does not represent the geometry of a molecule. However, it might be used if a correlation between molecular weight and effective diameter is determined [19]. For inorganic ions, the data on crystal ionic radia as the size parameter are available [24]. The crystal ionic radia can be used only approximately for the comparison’s sake, because ions in aqueous solutions are more or less hydrated, so the size of hydrated ions, listed in the same article, should be considered in case of ions’ transport through a membrane. A quantity deﬁning the size of small organic molecules more accurately should take into account the molecular shape and its structure. This can be done by calculations based on a simulated molecular structure. Van Bruggen et al. [23] calculated the effective molecular diameter from molecular structure and shape using an energetic optimization procedure. Later on, they [25] noticed that a calculated hydrated diameter might be a better parameter to describe the retention behavior of an organic compound. The morphology of the solute molecule is described by the molecular length and molecular width, as deﬁned by Kiso et al. [26] and used by Agenson et al. [27]. The molecular length was determined as the distance between the two most distant atoms in the molecule, and the molecular width was calculated as half of the square root of the surface area of the rectangle enclosing the projection of the molecule on a plane perpendicular to the axis of the molecule, i.e., the straight line connecting the two most distant atoms. The often used size parameter is the Stokes law diameter (the apparent rigid sphere diameter) calculated from the solute’s diffusivity [28–31]. It is sometimes assumed that the Stokes law diameter in addition to molecular shape factor might include also a hydration layer of the molecule (ion). There are additional factors that can inﬂuence the size of a molecule such as interaction with other molecules (of natural organic matter, for example), and changes in the solution chemistry affecting the geometry of a solute molecule.

110

B. Kunst · K. Koˇsuti´c

A membrane pore size is the second quantity necessary for the meaningful analysis of the solutes rejection by a membrane according to the size exclusion mechanism. Having in mind the asymmetric membrane structure, the pore size of the membrane’s skin (thin retention determining layer) is the needed quantity, and its determination is not simple. The membrane’s skin structure, morphology and porosity have been usually described by the models, either as bundles of straight, narrow, cylindrical pores (the Poiseuille equation) or the voids between the close-packed spheres of equal diameter (the Kozeny-Carman equation), deﬁning the water ﬂux, Jw , through a membrane that is proportional to the applied pressure. Such hydrodynamic models with the idealized pores could not fully explain experimental data on size exclusion, because skin morphology and structure in reality deviate from their simpliﬁed descriptions. The pores in the membrane’s skin should be better envisaged [32] as material-free void space in a dense polymer membrane layer through which ﬂuid transport takes place under a driving force. Narrow “pores” (voids) or density defects form naturally in solid polymers by the irregular packing of almost randomly kinked stiff-chain polymer molecules. The “active” pores (voids) consequently represent continuous tortuous paths for the solute and solvent molecules through the membrane’s skin. The real size of an open, active pore is mostly given by the pore shape and geometry along the tortuous path network. Hence there is always a distribution of pore (void) sizes in the membrane’s skin. The porosity of such speciﬁc structure of the membrane’s skin could not be measured so far in a direct experiment. The often used membrane characterization method in recent works, atomic force microscopy (AFM), gives primarily information on surface morphology (roughness), i.e., tells about size of pores on the membrane surface [8, 33, 34], and not about the size of pores in the skin. Therefore the porosity of the membrane’s skin has to be determined indirectly, by the solute transport methods [28, 29], having in mind that these methods allow ﬁnding the minimal size of the pore (void) constriction present along the pores open to ﬂux. The pore (void) dimensions of the membranes’ skin obtained by this method were always smaller than the surface pore sizes. In this way, one of the skin pore size parameters such as: the effective membrane pore radius, rp , the membrane molecular weight cutoff (MWCO), and the pore size distribution (PSD) in the skin layer can be described. Matsuura et al. [28, 29] characterized the membranes’ thin layer porosity by determining their pore size and PSD using the transport data of polyethylene glycols and polyethylene oxides of various molecular masses as the standard compounds used for MWCO determination. Koˇsuti´c and Kunst [35–37] determined the membranes’ PSD based on retentions of several disk-like markers (mostly cyclic ethers) by the RO and NF membranes. Yoon [31] estimated the effective pore radia for the membranes using a model of steric

Removal of Emerging Contaminants by NF/RO

111

interaction of hard spheres in cylindrical pores [30, 38]. Nghiem et al. [39] determined the average pore size from the retention data of inert organic markers of various molecular masses and a pore transport model that incorporates steric exclusion and hindered convection and diffusion. Boussou et al. [40] characterized the examined membrane porosities by the MWCO obtained from ﬁltration experiments with six organic compounds, and ﬁtting the retention curve with the log-normal model. Kimura et al. [41] emphasized, however, that MWCO cannot be used for precise prediction of rejection by RO membranes since properties of standard compounds used for MWCO determination (a certain Mw range of polyethylene glycols) and those of target compounds might be remarkably different. It also does not provide information on the retention for the molecules having a molecular weight below the MWCO. 4.2 Charge Exclusion Charge exclusion is the next important rejection mechanism of NF membranes. It was mainly developed and understood in the case of inorganic ions as solutes. The charge exclusion of ionic solutes is always present in nanoﬁltration, due to electrical charges carried by a membrane. They have been deliberately built in the skin layer of a nanoﬁltration membrane in form of sulfonic, carboxylic or ammonium groups covalently bonded to a membrane matrix. In aqueous solutions, these functional groups dissociate (depending on pH), leaving a membrane negatively or positively charged and repelling ions from the membrane. A sign of the membrane surface charge depends on the pH of the feed water, and for the most thin layer composite NF membranes the charge in neutral aqueous solutions is negative due to the dissociation of the sulfonic and carboxylic groups of the membrane’s skin. The surface charge of an NF membrane is often quantiﬁed by zeta potential measurements, and these measurements are usually done at pH 6–8. The zeta potential values change with pH, and at isoelectric point, which for the most NF membranes is at pH 3–4, the zeta potential value is zero. The membrane surface charge determines the rejection of dissolved ions, with anions repelled and cations attracted by a negatively charged membrane. This means that the charge exclusion mechanism is controlled by electrochemical laws, such as Donnan effect, the Nernst-Planck equation, and the general principle of conserving electroneutrality across a membrane. As a result, a coupled transport of ions takes place [42], i.e., individual ions get partitioned into the solvent-permeated membrane and migrate to the permeate side in a manner that the negative charge of an ion is balanced by the equivalent positive charge of an accompanying cation. The interdiffusion coefﬁcient of the electrolyte governs the permeation rate according to the coupled-transport mechanism.

112

B. Kunst · K. Koˇsuti´c

The charge exclusion mechanism seems simple to apply, and the experiments with the small inorganic ions conﬁrm its basic postulates. However, discrepancies appear in the results of the electrostatic rejections of multivalent ions, as well as in case of electrostatic interactions between charged organic compounds and the membrane charged surface. A part of the problem regarding the charge exclusion mechanism is in the porosity of the tight nanoﬁltration membranes, which is close to that of the reverse osmosis membranes, so the charge exclusion mechanism is always combined with the size exclusion. Two mechanisms cannot be completely separated, which is often reﬂected in discussion of the experimental results. The problem has also been encountered in modelling efforts of nanoﬁltration, where the predictive models were developed including both basic membrane rejection mechanisms. Such models were already mentioned and will not be repeated here. 4.3 Physicochemical Interactions In addition to size and charge exclusion mechanisms, the rejection of uncharged organic solutes by RO/NF membranes is frequently controlled by physicochemical properties of the system. These include general concepts on physicochemical interactions among membrane material, solute and water molecules, and more speciﬁcally quoted effects such as solute adsorption on and in the membrane, which is often connected to hydrophobic nonspeciﬁc interactions or to hydrogen bonding between membrane and solutes. The solute molecular polarity expressed by its dipole moment may also inﬂuence the solute rejection by a membrane, and the equilibrium between ionized and nonionized species in the aqueous solution, given by the solute dissociation constant value (pKa ), is also often described as the physicochemical factor affecting the organics rejection by the NF/RO membranes. Sourirajan and Matsuura [2] in their fundamental work postulated that the membrane separation effect is governed by the physicochemical phenomena at the membrane–solution interface. A preferential sorption of one of the constituents (usually water) of the solution at the interface is the overall result of mutual interactions arising from short-range forces depending on the individual contributions of the polar-, steric-, and nonpolar-character of the nonionized solute, solvent and the membrane material involved. Water is a polar solvent, a membrane surface made from aromatic polyamide chains has both polar groups and less polar backbone, and the nonionized organic molecule may be polar or nonpolar, or both. If the system consists of the same solvent, the same membrane material, and different pollutant molecules, the polarities of the latter will indicate the relative values of mutual interactions. Means to express quantitatively the polarity of a nonionized solute molecule are:

Removal of Emerging Contaminants by NF/RO

113

• the solute molecule acidity or basicity values; and • the Taft or the Hammet numbers for the solute molecule. The solute molecule acidity or basicity represents a hydrogen bonding ability of a solute organic molecule; a lower basicity value of the organic molecule, ∆ν (bas), expresses its lower proton accepting power, and a lower acidity value points to a solutes higher proton accepting ability. In presence of a molecule characterized by the lower basicity value, water will be less preferentially sorbed at the membrane–solution interface, and the solute rejection factor R will be lower. Similarly, Taft parameters, σ ∗ , represent the polar character of solute molecules. It is obtained by adding up the Taft number values of the substituent groups of a molecule. The Taft number concept has a ﬁrm thermodynamic base, it is generally applicable, and the extensive numerical data on ∗ in the literature. A lower Taft σ , measured and/or calculated are available number value for the organic molecule, σ ∗ expresses its lower proton donating power. In the presence of such a molecule, the preferential sorption of water at the interface will be stronger and the solute separation factor R will be higher. Hydrophobic nonspeciﬁc interactions between solutes and a membrane were frequently found to affect the solutes rejection by the NF/RO membrane. Hydrophobic interactions are closely related to adsorption effects on the membrane surface, and in order to estimate this type of physicochemical interaction, hydrophobicity of both the membrane and the solute should be known. A relative hydrophobicity of the membrane surface is commonly represented by the water contact angle values, as the means of quantifying the hydrophobicity. It is usually measured by the sessile drop method [40, 43, 44], and the low contact angle value represents low membrane surface hydrophobicity, and the higher contact angle values correspond to higher membrane hydrophobicities. A solute’s hydrophobicity, according to Kiso et al. [26, 45, 46] is correlated to its octanol-water partition coefﬁcient, log KO/W . They measured the rejection and adsorption of alkyl phtalates, non-phenylic and aromatic pesticides onto RO and NF membranes and found that their rejection depends on two factors: their adsorption on the membrane and the molecular shape of the solute. The adsorption properties, expressed by the partition coefﬁcient for a solute between the membrane and bulk solution, are affected by the membrane material, and by a solute’s hydrophobicity determined by its octanolwater partition coefﬁcient. The more hydrophobic solute, represented by the higher log KO/W value, the better would be its rejection by a membrane. This means that the adsorption properties of the examined compounds in the membrane were controlled by the hydrophobic interactions. Van Bruggen et al. [47–49] conﬁrmed the validity of the log KO/W value reﬂecting hydrophobic interactions between membrane and solute, and showed

114

B. Kunst · K. Koˇsuti´c

that its value correlates well with adsorption on the membrane for molecules with a molecular weight below the molecular weight cut-off of the membranes. They also found that adsorption of organic molecules from an aqueous solution on the membrane surface or inside the pores may result in supplementary effects such as the membrane ﬂux decline. The more of a given component is adsorbed on the membrane, the larger the membrane’s ﬂux decline. The ﬂux decline actually occurs by a combination of two effects: the molecule should have the appropriate size to ﬁll the membrane pores, and the adsorption process is enhanced by the hydrophobicity of the component. The membrane ﬂux decline actually points out to the phenomenon of membrane fouling which is partly caused by the adsorption of the organic compounds on the membrane. Braeken et al. [25] recently gave another detailed explanation of the inﬂuence of hydrophobicity on retention of organic compounds with a molecular weight below the MWCO of the used membrane. A good correlation was found between the hydrophobicity and the retention: the hydrophobic molecules generally had a lower retention, while hydrophilic molecules showed a higher retention. This was attributed to an inﬂuence of the solute hydrophobicity on the molecule hydration, and via its hydrated, effective size on the solute rejection. Hydrophobic compounds, namely, have less polar groups and are less solvated. Because of their smaller size, they can more easily enter the membrane pores and permeate the membrane. Hydrophilic molecules generally have more OH or O groups, which can form hydrogen bonds with the water molecules. Due to such higher afﬁnity with the water phase, hydrated hydrophilic molecules are larger permeating in a smaller amount through the membrane structure. Kimura et al. [9, 50] also conﬁrmed the ﬂux decline effect and emphasized its relevance at low feed concentrations, at the submicrogram per liter level. Under such conditions the adsorption of hydrophobic compounds in the membrane pores could lead to an overestimation of the retention. The adsorption of hydrophobic solutes contributes to rejection, but the adsorption capacity of a membrane is ﬁnite. Therefore, an accurate calculation of the rejection of the hydrophobic compounds is possible when the adsorption equilibrium of the membrane with the compounds of interest is achieved. Hydrophobicity of the solute also brings about the evolution of the permeate concentration [25], showing a decrease of the solute retention as a function of time due to the solute adsorption on the membrane surface or inside the membrane pores. The rejection of hydrophobic compounds at higher concentrations decreases until saturation of the membrane was reached. This retention value was named the “steady-state retention”. Another physicochemical property, that according to Van der Bruggen et al. [23, 48], could inﬂuence a rejection by the NF membrane is the solute polarity expressed through its dipole moment. The uncharged molecules

Removal of Emerging Contaminants by NF/RO

115

with the high permanent dipole moment (above 3 Debye) were shown to exhibit lower retentions than the molecules of approximately same size but with lower dipole moments. The effect was explained by the preferable orientation of polar molecules towards the charged membrane surface. The polar molecule is by electrostatic attraction directed towards the charged membrane with the side of the dipole having the opposite charge closer to the membrane pores. Thus the polar molecule enters more easily into the membrane, resulting in more polar molecules permeating the membrane compared to non-polar molecules of the same size. It was however, stated that such a result should be taken with a reserve due to the effect of a solvent on the solute dipole moment. Namely, dipole moments are usually determined in apolar solvents and their values might be different in aqueous solutions. In listing various effects, which have the impact upon the NF/RO membranes rejection, operating conditions of the membrane system should also be mentioned. Chellam and Taylor [51] showed, for example, that the feed water recovery inﬂuenced the rejection of solutes. A decrease of rejection was observed for an increase in recovery. The higher recovery would increase the concentration difference across the membrane and consequently the potential of diffusion. Similarly, Chen et al. [52] conﬁrmed that the rejection of some pesticides was dependent on operational ﬂux and recovery. For a particular pesticide the highest percent rejection occurred at high ﬂux and low recovery, and the lowest percent rejection occurred at low ﬂux and high recovery, which indicated the basic diffusion-control theory. The diffusion control may dominate a solute rejection if the molecule is in a size range where molecular charge, Van der Waal’s forces or other surface interactions can affect the interaction between the membrane surface and the solute molecule during mass transfer. In the preceding part the important effects inﬂuencing organic solutes removal by the NF/RO membranes were registered. A contribution of each of them differs from case to case, and it is hard to predict, without experimental evidences, which one would prevail in any speciﬁc case. Thus, the effort of Bellona et al. [7], who tried to organize the inﬂuences of physicochemical properties of solutes and membranes important for rejection, is worthy of attention. They drew a “rejection diagram” for organic non-charged and negatively charged solutes, with compounds grouped according to distinctive physicochemical characteristics in order to discern the mechanisms responsible for rejection. The underlying concept behind the rejection diagram is that for any given compound, if the physicochemical characteristics of the solute and membrane are known, the driving factors of rejection can be predetermined and the rejection predicted. By applying this diagram, a qualitative estimate on a degree of rejection is obtained, and the experimental results should verify the usefulness of this diagram and quantify rejection of a certain compound by a particular membrane type.

116

B. Kunst · K. Koˇsuti´c

5 Recent Findings As mentioned in the Introduction, the ﬁndings and general rules on removal of different solutes and conventional pollutants from waters by nanoﬁltration and reverse osmosis were obtained in studies limited to some targeted compounds and covered mostly binary aqueous solutions and synthetic model waters. As recent studies investigating the organic micropollutants retention were conducted at an environmentally realistic, sub microgram per liter concentrations, and on wastewaters containing a great variety of compounds, the new ﬁndings and general conclusions might deviate from the previously acquired knowledge. Thus, recent results and newly planned studies should show to what extent the earlier knowledge can be applied, in order to explain the removal of organic trace pollutants from various waters. To illustrate recent ﬁndings on the removal of emerging contaminants by the NF/RO, the results in this part are presented in three groups: • the removal of emerging contaminants from wastewaters at low feed concentrations (submicrogram per liter); • the removal of a speciﬁc group of micropollutants, antibiotics; • the results of the latest studies on the organics removal by NF/RO. 5.1 Removal of Emerging Contaminants at Submicrogram per Liter Concentrations Kimura et al. [50] investigated the rejection of selected trace organics and surrogate compounds by NF/RO membranes as a function of their physicochemical properties at initially very low (100 ng/l to 100 µg/l) micropollutants concentrations in feed water. The experiments showed that charged compounds could be rejected to a great extent (i.e., > 90%) regardless of physicochemical properties of the tested compounds. In contrast, the rejection of non-charged compounds was inﬂuenced mainly by the size of the compounds. It was also found that the concentration range of solutes might affect the rejection efﬁciency of a membrane. The following study [41] examined the ability of RO membranes to retain uncharged trace organics. The examined compounds were chosen to cover a certain range of molecular weights and KO/W . Two membranes made of different materials (polyamide and cellulose acetate) were tested. The comparison of the results showed that the polyamide membrane generally exhibited better rejections. However, the rejection was incomplete, amounting to 57–91%, which could be correlated to solute molecular weight, conﬁrming that the size exclusion dominated the retention by the polyamide membrane. In case of the CA membrane, the compounds polarity was better able to describe the retention trend of the tested compounds. The results imply

Removal of Emerging Contaminants by NF/RO

117

that the dominant rejection mechanism could be different depending on the membrane material and the physicochemical properties of the target organic micropollutants. In the work of Yoon et al. [53] removal of 52 emerging contaminants from one model water and three surface waters by NF and UF was examined. The source waters were spiked with tested compounds at very low concentrations (2–250 ng/l). The results showed that all the tested compounds could be divided into two groups: the more polar, less hydrophobic compounds were less retained than the less polar, and more hydrophobic compounds. Although the detailed explanations of the retention mechanisms for various compounds could not be given, the authors concluded that the retention was governed by hydrophobic adsorption. When the steady-state operation is achieved, size exclusion can be dominant for solute retention. The retentions of the NF membrane were greater than those of the UF membrane, implying that retention was governed by the membranes’ pore size. In addition, the retention of solutes appears to be affected by source water chemistry. In continuation of the previous work, the removal of the more polar, less hydrophobic compounds was further examined [54]. Experiments were performed at environmentally relevant solute concentrations ranging from 2 to < 150 ng/l. The general separation trend due to hydrophobic adsorption, was observed. The results showed again that both size exclusion and hydrophobic adsorption mechanisms were operative at the solutes’ retention by the NF membrane, while the UF membrane retained typically hydrophobic solutes owing mainly to the hydrophobic adsorption. Snyder et al. [55] investigated the efﬁcacy of membranes and activated carbons for the removal of emerging contaminants at pilot- and/or full-scale. Structurally diverse target compounds at trace concentrations were evaluated. Several membrane types and applications were used including: microﬁltration (MF), ultraﬁltration (UF), nanoﬁltration (NF), reverse osmosis (RO), electrodialysis reversal, membrane bioreactors (MBR), and combinations of membranes in series. MF and UF were found to reject very few target compounds, and the MBR system removed some of the micropollutants. It showed better results when followed by the RO. NF and RO rejected almost all target compounds. Also, granular activated carbon was highly effective at removing all target chemicals. However, breakthrough curves demonstrated that compounds with greater hydrophilicity breach activated carbon faster than hydrophobic compounds. The ﬁndings conﬁrm that membrane processes and carbon adsorption greatly reduce the concentrations of emerging contaminants; however, several compounds are still detectable in membrane permeate and carbon efﬂuent. The results were conﬁrmed in the study on the removal of pharmaceuticals and endocrine disruptors from South Korean waters [56]. Rejection of trace organic compounds, disinfection by-products and chlorinated solvents, by thoroughly characterized RO and NF membranes from

118

B. Kunst · K. Koˇsuti´c

aqueous solution at low feed concentrations (100 µg/l) was studied by Kim et al. [57]. The rejections of compounds of intermediate hydrophobicity by the candidate membranes were less than salt rejections for these membranes. Such a result suggests that transport of the examined solutes through these membranes is facilitated by solute–membrane interactions. It was supposed that hydrogen bonding plays a role in solute transport through the membranes. Xu et al. [58] investigated the rejection of emerging contaminants at very low concentrations by a variety of commercial NF/RO membranes, in order to simulate operational conditions for drinking water and wastewater treatment. Also, the effects of various water quality matrices were tested. A range of organic micropollutants of various physicochemical properties from hydrophilic ionic, hydrophilic nonionic to hydrophobic nonionic were examined. It was found that tight NF membranes at lower operating pressure behave likewise to RO membranes, but at higher pressures, the membrane surface charge is more important than the MWCO. The tight NF membranes at lower feed pressure provided a product water quality similar to a RO membrane. The presence of efﬂuent organic matter improved the rejection of negatively charged compounds. Hydrophobic nonionic compounds were only partially removed by RO membrane, and a tight NF membrane can achieve a similar effect. An increase of process recovery did not essentially inﬂuence the permeate concentration, and the presence of efﬂuent organic matter completely neutralized the inﬂuence of hydrodynamic conditions. 5.2 Removal of Antibiotics Several recent papers reported on the removal of antibiotics, a special group of pharmaceutically active compounds, from waters by the NF/RO membranes. Adams et al. [59] evaluated the effectiveness of the conventional drinking water treatment processes under typical plant conditions in the removal of seven common antibiotics: carbadox, sulfadimethoxine, sulfamethazine, sulfamerazine, sulfathiazole, sulfachlorpyridazine and trimethoprim. Experiments were carried out with spiked aqueous solutions prepared from deionized water and water from the Missouri River. The antibiotics were effectively removed by reverse osmosis, sorption on activated carbon, and by oxidation. Conversely, coagulation/ﬂocculation/sedimentation with alum and iron salts, excess lime/soda ash softening, ultraviolet irradiation at disinfection dosages, and ion exchange were all relatively ineffective methods of antibiotic removal. Li et al. [60] reported on membrane treatment of the oxytetracycline waste liquor. Using RO process a volume reduction coefﬁcient was 3.5, oxytetracycline in retentate was concentrated more than three times, and oxytetracycline in permeate was decreased from more than 1000 mg/l to lower than

Removal of Emerging Contaminants by NF/RO

119

80 mg/l. A permeate could be later treated by the biological process, owing to the major reduction in both the organic concentration and antibiotic toxicity. The RO ﬂux capacity did not decrease for repeated operations, and used membranes could be regenerated by chemical cleaning. As oxytetracycline could not be successfully recovered from the RO retentate by simple crystallization, the additional treatment of the RO retentate by ultraﬁltration and subsequent crystallization resulted in a recovery ratio of more than 60% and the product purity higher than 80%. Therefore, the RO-UF membrane process can be an effective way for the antibiotic wastewater treatment as well as the antibiotics recovery from the waste liquor. Koˇsuti´c et al. [61] investigated the removal of antibiotics by NF/RO from two model wastewaters of a manufacturing plant producing pharmaceuticals for veterinary use. The following antibiotics were used: levamisole, sulfaguanidine, sulfamethazine, sulfadiazine, trimethoprim, praziquantel, enroﬂoxacin and oxytetracycline. The rejections by the selected RO and the tight NF membranes were acceptably high, exceeding in most cases 98%. The loose NF membrane retained the smaller molecules less effectively. Relating the solute rejections to the membranes’ porosity has shown that size exclusion was the prevailing rejection mechanism by all membranes. The rejection of the low molecular organics by all the examined membranes was lower, ranging from 0.517 to 0.976 for the RO and tight NF membranes, and from 0.247 to 0.506 for the loose NF membrane. They follow the order of the solutes’ molecular size, though the speciﬁc physicochemical effects might inﬂuence the rejection of some low molecular organics. Some papers in this ﬁeld report on the use of NF/RO membranes in the antibiotics production processes mainly to improve recovery and purity of the product simultaneously with a treatment of the process wastewaters. Such are the papers on separation and puriﬁcation of benzylpenicillin produced by fermentation [62, 63]. 5.3 The Latest Results on the Organics Removal by NF/RO A removal of natural hormones: estradiol, estrone, testosterone, and progesterone by two characterized nanoﬁltration membranes was investigated by Nghiem et.al. [39]. The hormones were undissociated at the pH of the experiment, so steric exclusion and adsorptive effects dominated in the solute rejection. The hormones’ molecular size was larger than the average pore size of both membranes, and their initial rejection was almost complete and dominated by the adsorption of hormones to the membrane matrix. However, the initial high retention decreased with time. This was explained by the reached adsorption equilibrium, after which a diffusion of hormones via hydrogen bonding to membranes functional groups takes place, ultimately resulting in a lower retention.

120

B. Kunst · K. Koˇsuti´c

The contribution of electrostatic interactions to the rejection of ionizable trace pollutants was studied by Bellona and Drewes [64] and Nghiem et al. [65, 66]. Bellona and Drewes [64] examined the rejection of selected organic acids by NF membranes. The rejection of negatively charged acids was signiﬁcantly inﬂuenced by the effective surface charge. The rejection, primarily driven by the membrane surface charge, correlated well with the degree of ionization of the solutes. The rejection of ibuprofen, an organic acid with hydrophobic properties, was also pH dependent, but at pH values below the pKa ibuprofen adsorbed and partitioned through the membrane. At pH above the pKa , adsorption of ibuprofen was minimal and the electrostatic repulsion dominated. The presence of calcium in the feed water lowered the effective membrane surface charge of both tested membranes, however, rejection of negatively charged organic solutes decreased only for membranes with a MWCO larger than the solute size. Nghiem et al. [65, 66] reported on the retention of three pharmaceuticals: ibuprofen sulfamethoxazole, and carbamazepine, representing three different drug categories. These small organic molecules, if electrically charged, differ in their transport through a NF membrane from the behavior of proteins. In the ﬁrst of two papers [65] retentions of the compounds by two well-characterized NF membranes were examined. The retention of all of them was quite high by a tight NF membrane, i.e., it was dominated by size exclusion, whereas both electrostatic repulsion and steric exclusion govern the retention of ionizable solutes by a loose NF membrane. A decrease of solution pH diminished the rejection of ibuprofen and especially sulfamethoxazole, which transformed from the negatively charged to neutral species as the solution pH decreased. The rejection of uncharged pharmaceuticals is inﬂuenced by their intrinsic physicochemical properties such as their dipole moments. This is evident in case of cylindrical sulfometoxazole molecule, which in uncharged state permeates the loose membrane due to its high polarity. The second paper [66] dealt with the role of electrostatic interactions in the retention of pharmaceutically active contaminants by a loose NF membrane. Retention of the nonionizable molecule, carbamazepine, was independent of the solution chemistry. The inﬂuence of the solution pH, ionic strength and the presence of divalent cations, on retention of sulfamethoxazole and ibuprofen was strong. The retention increased dramatically as the compound transforms from a neutral to a negatively charged species when the solution pH increases above its pKa value. In contrast, solution ionic strength reduced the effectiveness of electrostatic interaction as a major retention mechanism by the loose NF membranes, due to the restrained double layer. The inﬂuence of a divalent salt concentration is stronger than that of sodium chloride as the background electrolyte. Ben David, Freger et al. [67, 68] found that thermodynamic (sorption) and size effects are superimposed in RO/NF, which leads to complex behavior that is difﬁcult to interpret in terms of simple concepts, such as the pore size or the

Removal of Emerging Contaminants by NF/RO

121

molecular weight cut-off. In considering sorption in the membrane, quantitative sorption data of various small organic compounds by the polyamide active layer (skin) and not by the supporting membrane structures are required and crucial for understanding the effect. They presented a new method for direct measurements of a true partitioning coefﬁcient of a solute between the aqueous phase and the material of the polyamide active layer that is free of artifacts related to possible sorption on supporting structures. Their results on sorption of alcohols, polyols and some aromatic solutes on the active layers of polyamide RO/NF membranes show that the very use of the effective pore size or MWCO for RO and NF membranes requires much care and should always address the issues of thermodynamic afﬁnity. The fouling tendency in nanoﬁltration of organics is important for the membrane transport of water and solutes. A fouling study of three fully characterized nanoﬁltration membranes [40] showed a remarkable inﬂuence of the membrane properties. The lowest fouling was observed for the most hydrophilic membrane having the largest pore size, the high negative charge in case of ﬁltrating salts and a relatively smooth surface in case of ﬁltrating silica. The effect of the feed properties was shown by using aqueous feed solutions containing uncharged organics, salts or silica colloids. Colloidal fouling in the presence of different NaCl concentrations indicated that fouling increased with increasing ionic strength due to the shielding effect of the membrane and the colloid charge. Another paper [69] described the effects of membrane fouling on rejection of trace organics by NF/RO membranes, using a secondary efﬂuent from a municipal wastewater treatment plant as the feed. Fouling resulted in more negative zeta potential and changes of hydrophobicity for ﬁve examined membranes. The rejection of organic micropollutants by CTA, NF, and ULPRO membranes was especially inﬂuenced, while it was less important in the case of RO membranes. The effect was stronger for ionic organic micropollutants, due to a more negative surface charge. The hydrophobic non-ionic solutes were also better rejected. However, the increasing surface charge resulted in a larger MWCO of a fouled membrane due to membrane swelling. This could lead to lower rejection for hydrophilic non-ionic solutes, especially where NF membranes of a larger MWCO were used. The effect of the feed solution (distilled water, tap water, and river water) on the removal of pesticides by NF was investigated by Zhang et al. [70] Atrazine or simazine was added to different water matrices at a concentration of 5 mg/l and 100 µg/l, and rejection of pesticides and the water ﬂux of NF membranes were measured. The rejection of pesticides was higher in river and tap waters than in distilled water, but the water ﬂux was lower. This was mainly explained by ion adsorption inside the membrane pores, although the presence of natural organic matter (NOM) enhanced the size exclusion due to formation of NOM-triazine complex. No signiﬁcant effect of the pesticides concentration was observed. Considerable rejections of nitrate and

122

B. Kunst · K. Koˇsuti´c

organic matter were found, in particular for the membrane with the smallest pores. This conﬁrmed the assumption that pore narrowing by ion adsorption affected the rejections and water ﬂuxes. Koˇsuti´c et al. [71] proposed an experimental porosity characterization of the NF/RO membranes’ active layer without resorting to theoretical models. From the membrane’s ﬂux dependency on pressure, two characteristic parameters were determined: (1) the susceptibility of the active layer structure under pressure and (2) the membrane’s pure water permeability. The significance of the characteristic parameters was conﬁrmed by relating them to the pore size distributions (PSD) and to the effective number of pores in the active membrane layer. The examined RO/NF membranes were classiﬁed according to the values of the characteristic parameters, and the classiﬁcation was illustrated by the membranes’ rejection for various, mostly organic solutes. Santos et al. [72] proposed the “geometric model” of a solute for a description and better understanding of the rejection behavior of elongated molecules by nanoﬁltration membranes. They represented the solute molecular geometry with a prolate revolution ellipsoid, and the solute characteristics: size, shape, symmetry and location of functional groups determine how the molecule orients during uptake in the membrane and transport through it, and consequently how the membrane would reject it. Braeken et al. [73] investigated the relative contribution of convection, diffusion and charge effects during transport of dissolved organics through NF membranes. The diffusion was analyzed by the Donnan model and the charge effect was investigated by changing the pH of the feed solution, which inﬂuenced both the membrane and the compound charge. Convection was found to be the dominant mechanism for transport of organics, and it depended on the compound properties. The retention of uncharged solutes slightly decreased at high pH, due to the enlargement of the membrane pores caused by an increased repulsion between the membrane charged groups. The retention of charged compounds increased with increasing the pH, owing to an increased repulsion between the charge of compound and membrane. Diffusion through tight membranes became more important as the molecular weight of the compound increased, because steric hindrance in narrow pores mostly affected convective transport. In the paper of Ali et al. [74] an attempt was made to grasp the inﬂuence of NF membrane characteristics on the cost of a membrane separation system. This was achieved through combining a rejection simulation with an economic model for membrane system. Based on the simulation, it was found that the looser the pore structure of a membrane, the greater the membrane charge capacity needed in order to achieve a desired application. Guidelines for selection of the most suitable membranes for a speciﬁc rejection can be developed from the predicted rejection. From the cost point of view, the membrane charge density did not have any effect on the cost. However,

Removal of Emerging Contaminants by NF/RO

123

increasing the membrane pore radius will lead to a decrease in system separation cost. This study also discovered that by using much looser membrane, a signiﬁcant cost can be saved.

6 Concluding Remarks In this contribution, the ability of reverse osmosis/nanoﬁltration process to remove efﬁciently the emerging organic contaminants from water was evaluated. The assessment of the previously gathered knowledge on the basic rules for the rejection of conventional organic pollutants by NF/RO, and on the solute and membrane characteristics have indicated that such knowledge might be generally applied for the removal of emerging contaminants. This has been repeatedly conﬁrmed in the recent studies that deal with the retention of micro-pollutants at very low, environmentally realistic concentrations, as long as the single, targeted organic micropollutant is investigated, and under the assumption that the steady-state adsorption of the compound in a membrane was reached. On the other hand, when numerous constituents were present in wastewaters, the results were limited to the mere qualitative facts of higher or lower removal without detailed explanations on the retention mechanisms and the factors affecting the membrane’s retention. The validity of the reviewed knowledge and the basic rules of the NF/RO separations in such cases of complex mixtures of solutes could not be conﬁrmed yet due to insufﬁcient experimental results and quite recent developments of the adequate analytical techniques that could not have been widely used so far. Acknowledgements This work has been supported by EU project EMCO – (INCO CT 2004-509188) – Reduction of environmental risks, posed by Emerging Contaminants, through advanced treatment of municipal and industrial wastes, and by the Croatian Ministry of Education, Science and Sport (Project No. 0125017).

References 1. Belfort G (1984) Synthetic membrane processes: fundamentals and water applications. Academic Press Inc, Orlando FL 2. Sourirajan S, Matsuura T (1985) Reverse osmosis/ultraﬁltration process principles. National Research Council Canada, Ottawa 3. Mulder M (1996) Basic principles of membrane technology. Kluwer Academic Publishers, Dordrecht 4. Buckley CA, Brouckaert CJ, Kerr CA (1993) Reverse osmosis application in brackish water desalination and in the treatment of industrial efﬂuents. In: Amjad Z (ed) Reverse osmosis, Chap 9. Van Nostrand Reinhold, New York, p 275

124

B. Kunst · K. Koˇsuti´c

5. Bhattacharyya D, Williams ME (1992) Reverse osmosis, theory. In: Ho WSW, Sirkar KK (eds) Membrane handbook, Chaps 21–25. Van Nostrand Reinhold, New York, p 263 6. Schäfer AI, Fane AG, Waite TD (2005) Nanoﬁltration – principles and applications. Elsevier, Oxford 7. Bellona C, Drewes JE, Xu P, Amy G (2004) Water Res 38:2795 8. Hilal N, A1-Zoubi H, Darwish NA, Mohammad AW, Abu Arabi M (2004) Desalination 170:281 9. Kimura K, Amy G, Drewes JE, Watanabe Y (2003) J Membr Sci 221:89 10. Loeb S, Sourirajan S (1962) Adv Chem Ser 38:117 11. Kedem O, Katchalsky A (1963) Trans Faraday Soc 59:163 12. Spiegler KS, Kedem O (1966) Desalination 1:311 13. Matsuura T, Sourirajan S (1981) Ind Eng Chem Proc Des Dev 20:273 14. Williams ME, Hestekin JA, Smothers CN, Bhattacharyya D (1999) Ind Eng Chem Res 38:3683 15. Deen WM (1987) AIChE J 33:1409 16. Wang XL, Tsuru T, Nakao S, Kimura S (1995) J Membr Sci 103:117 17. Wang XL, Tsuru T, Nakao S, Kimura S (1997) J Membr Sci 135:19 18. Combe C, Guizard C, Aimar P, Sanchez V (1997) J Membr Sci 132:109 19. Bowen WR, Mohammad AW, Hilal N (1997) J Membr Sci 126:91 20. Van der Bruggen B, Vandecasteele C (2002) Water Res 36:1360 21. Bowen WR, Wellfoot JS (2002) Chem Eng Sci 57:1121 22. Ozaki H, Li HF (2002) Water Res 36:123 23. Van der Bruggen B, Schaep J, Wilms D, Vandecasteele C (1999) J Membr Sci 156:29 24. Nightingale ER (1959) J Phys Chem 63:1381 25. Braeken L, Ramaekers R, Zhang Y, Maes G, Van der Bruggen B, Vandecasteele C (2005) J Membr Sci 252:195 26. Kiso Y, Kon T, Kitao T, Nishimura K (2001) J Membr Sci 182:205 27. Agenson KO, Oh J, Urase T (2003) J Membr Sci 225:91 28. Singh S, Khulbe KC, Matsuura T, Ramamurthy P (1998) J Membr Sci 142:111 29. Khayet M, Matsuura T (2003) Desalination 158:57 30. Bowen WR, Mohammad AW (1998) Chem Eng Res Des 76:885 31. Yoon Y, Lueptow RM (2005) J Membr Sci 261:76 32. Meares P (1976) The physical chemistry of transport and separation by membranes. In: Meares P (ed) Membrane separation processes, Chap 1. Elsevier Scientiﬁc Publishing Comp., Amsterdam, p 1 33. Bowen WR, Doneva TA (2000) Desalination 129:163 34. Hilal N, Al-Zoubi H, Darwish NA, Mohammad AW (2005) Desalination 177:187 35. Kaˇstelan-Kunst L, Koˇsuti´c K, Danani´c V, Kunst B (1997) Water Res 31:2878 36. Koˇsuti´c K, Kaˇstelan-Kunst L, Kunst B (2000) J Membr Sci 168:101 37. Koˇsuti´c K, Kunst B (2002) Desalination 142:47 38. Lee S, Lueptow RM (2001) Environ Sci Technol 35:3008 39. Nghiem LD, Schäfer AI, Elimelech M (2004) Environ Sci Technol 38:1888 40. Boussu K, Zhang Y, Cocquyt J, Van der Meeren P, Volodin A, Van Haesendonck C, Martens JA, Van der Bruggen B (2006) J Membr Sci 278:418 41. Kimura K, Toshima S, Amy G, Watanabe Y (2004) J Membr Sci 245:71 42. Mukherjee P, SenGupta AK (2006) J Membr Sci 278:301 43. Zhang W, Wahlgren M, Sivik B (1989) Desalination 72:263 44. Wintgens T, Gallenkemper M, Melin T (2002) Desalination 146:387 45. Kiso Y, Nishimura Y, Kitao T, Nishimura K (2000) J Membr Sci 171:229

Removal of Emerging Contaminants by NF/RO 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74.

125

Kiso Y, Sugiura Y, Kitao T, Nishimura K (2001) J Membr Sci 192:1 Van der Bruggen B, Vandecasteele C (2001) Environ Sci Technol 35:3535 Van der Bruggen B, Braeken L, Vandecasteele C (2002) Separ Purif Technol 29:23 Van der Bruggen B, Braeken L, Vandecasteele C (2002) Desalination 147:281 Kimura K, Amy G, Drewes JE, Heberer T, Kim TU, Watanabe Y (2003) J Membr Sci 227:113 Chellam S, Taylor JS (2001) Water Res 35:2460 Chen S, Taylor JS, Mulford LA, Norris CD (2004) Desalination 160:103 Yoon Y, Westerhoff P, Snyder SA, Wert EC (2006) J Membr Sci 270:88 Yoon Y, Westerhoff P, Snyder SA, Wert EC, Yoon J (2007) Desalination 202:16 Snyder SA, Adham S, Redding AM, Cannon FS, DeCarolis J, Oppenheimer J, Wert EC, Yoon Y (2007) Desalination 202:156 Kim SD, Cho J, Kim IS, Vanderford BJ, Snyder SA (2007) Water Res 41:1013 Kim T-U, Amy G, Drewes JE (2005) Water Sci Technol 51:335 Xu P, Drewes JE, Bellona C, Amy G, Kim TU, Adam M, Heberer T (2005) Water Environ Res 77:40 Adams C, Wang Y, Loftin K, Meyer M (2002) J Environ Eng-ASCE 128:253 Li SZ, Li XY, Wang DM (2004) Separ Purif Technol 34:109 Koˇsuti´c K, Dolar D, Aˇsperger D, Kunst B (2007) Separ Purif Technol 53:244 Nabais AMA, Cardoso JP (2000) Bioproc Engin 22:41 Tessier L, Bouchard P, Rahni M (2005) J Biotechn 116:79 Bellona C, Drewes JE (2005) J Membr Sci 249:227 Nghiem LD, Schäfer AI, Elimelech M (2005) Environ Sci Technol 39:7698 Nghiem LD, Schäfer AI, Elimelech M (2006) J Membr Sci 286:52 Ben-David A, Bason S, Jopp J, Oren Y, Freger V (2006) J Membr Sci 281:480 Ben-David A, Oren Y, Freger V (2006) Environ Sci Technol 40:7023 Xu P, Drewes JE, Kim TU, Bellona C, Amy G (2006) J Membr Sci 279:165 Zhang Y, Van der Bruggen B, Chen GX, Braeken L, Vandecasteele C (2004) Separ Purif Technol 38:163 Koˇsuti´c K, Dolar D, Kunst B (2006) J Membr Sci 282:109 Santos JLC, de Beukelaar P, Vankelecom IFJ, Velizarov S, Crespo JG (2006) Separ Purif Technol 50:122 Braeken L, Bettens B, Boussu K, Van der Meeren P, Cocquyt J, Vermant J, Van der Bruggen B (2006) J Membr Sci 279:311 Ali N, Mohammad AW, Ahmad AL (2005) Separ Purif Technol 41:29

Hdb Env Chem Vol. 5, Part S/2 (2008): 127–175 DOI 10.1007/698_5_103 © Springer-Verlag Berlin Heidelberg Published online: 7 February 2008

Ozone-Based Technologies in Water and Wastewater Treatment A. Rodríguez1 · R. Rosal1 · J. A. Perdigón-Melón1 · M. Mezcua2 · A. Agüera2 · M. D. Hernando1 · P. Letón1 · A. R. Fernández-Alba2 (u) · E. García-Calvo1 1 Department

of Chemical Engineering, University of Alcalá, 28871 Alcalá de Henares, Madrid, Spain

2 Department

of Analytical Chemistry, University of Almería, 04120 Almería, Spain [email protected]

1 1.1 1.2 1.3 1.4 1.5

Fundamentals of Ozonation Processes . . The Molecule of Ozone . . . . . . . . . . . Solubility of Ozone in Water . . . . . . . . Ozone Mass Transfer . . . . . . . . . . . . Decomposition of Ozone in Water . . . . . Ozone Reactions with Organic Compounds

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

129 129 130 131 134 138

2 2.1 2.2 2.3 2.4 2.4.1 2.4.2 2.5 2.6

Ozone Uses in Water Treatment . . . . . . . . . . . . . . . . . . . . Precipitation of Oxides . . . . . . . . . . . . . . . . . . . . . . . . . Disinfection of Drinking Water . . . . . . . . . . . . . . . . . . . . . Natural Water and Wastewater Treatment . . . . . . . . . . . . . . . Catalytic Ozonation . . . . . . . . . . . . . . . . . . . . . . . . . . . Homogeneous Catalytic Ozonation . . . . . . . . . . . . . . . . . . Catalysis by Metals and Metal Oxides . . . . . . . . . . . . . . . . . Applications in the Treatment of Industrial Wastewater . . . . . . . Removal Efﬁciency of Pharmaceuticals in Wastewater: A Case Study

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

140 141 143 145 153 155 156 163 166

3

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

171

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

172

Abstract Ozone is a strong oxidant that can be used in the potabilization of surface or ground water as well as in wastewater treatment to remove microorganisms, inorganic ions and organic pollutants. The oldest use of ozone is as a biocide in drinking water potabilization. The integral ozone exposure required for a given degree of disinfection can be calculated from the deactivation kinetic constant of the microorganism. Ozone removes iron, manganese and arsenic from water by oxidation to an insoluble form that is further separated by ﬁltration. Both processes require ozone in molecular form, but the removal of organic pollutants that are refractory to other treatments can be possible only by exploiting the indirect radical reactions that take place during ozonation. Ozone decomposes in water, especially when hydrogen peroxide is present, to yield the hydroxyl radical, the strongest oxidizer available in water treatment. Models for the ozonation process are required to adjust the ozone dosing to the desired degree of removal of a given pollutant or an aggregate measure of pollution. Mineralization, deﬁned as the removal of organic carbon, has been accomplished in wastewaters from urban and domestic treatment plants. The results show that the logarithmic decrease of TOC as a function of the integral ozone exposure usually presents two zones with different kinetic parameters.

128

A. Rodríguez et al.

Among advanced oxidation processes, a promising alternative currently under development is the use of ozone in combination with solid catalysts. The mechanism of catalytic ozonation is not clear, but in the case of metal oxides, the adsorption of ozone or organic compounds on Lewis acid sites is only possible near the point of zero charge of the surface. Activated carbon seems to behave as an initiator of ozone decomposition, a role that may also occur with other types of catalysts. Some results on the mineralization of water with the drugs naproxen (non-steroidal anti-inﬂammatory) and carbamazepine (anticonvulsant) are presented using titanium dioxide as catalyst. Keywords Advanced oxidation processes · Disinfection · Kinetic models · Ozonation · Solid catalysts

Abbreviations a Speciﬁc gas–liquid interfacial area [m–1 ] Alk Alkalinity [mg CaCO3 L–1 ] cA Concentration of a given compound [M] Concentration of dissolved ozone in water [M] CO3 ∗ Equilibrium concentration of dissolved ozone in water [M] CO 3 cs Bulk concentration of catalyst [kg m–3 ] Concentration of surface sites of catalyst [mol kg–1 ] ct ctO3 , ct10 Concentration–time exposure parameter for ozone [M s] Bubble diameter [m] db Diffusivity of oxygen [m2 s–1 ] DO 2 DO 3 Diffusivity of ozone [m2 s–1 ] E Enhancement factor Ha Hatta number Henry’s law constant [atm mole fraction–1 ] He i Ionic strength [M–1 ] Rate constants for the catalytic decomposition of ozone [m3 kg–1 s–1 ] k1 , k2 –1 Kinetic constant of adsorption [L kg–1 ka cat s ] –1 k–a Kinetic constant of desorption [mol kg–1 cat s ] –1 Kinetic constant of catalytic ozonation [L kg–1 kc cat s ] kd Kinetic constant of ozone decomposition [units depending on the order of reaction] Kinetic constants for direct reaction with ozone [L mol–1 s–1 ] kD , kDi kHO· Kinetic constant for reactions with hydroxyl radical [L mol–1 s–1 ] Kinetic constants of the hydroxide initiation of ozone decomposition [M–1 s–1 ] kHO– – kHO2 Kinetic constants of the hydroperoxide initiation of ozone decomposition [M–1 s–1 ] kL Liquid phase individual mass transfer coefﬁcient [m s–1 ] Volumetric mass transfer coefﬁcient [s–1 ] kL a Kinetic constant for microorganism deactivation [M–1 s–1 ] kN –1 ko Kinetic constant of the surface oxidation process [L kg–1 cat s ] –1 –1 Kinetic constant for direct reaction with ozone [L mol s ] kO3 kr Kinetic constant of termination reactions [L mol–1 s–1 ] Adsorption equilibrium constant [L mol–1 ] Ka Equilibrium constant for the surface oxidation process [L mol–1 ] Kox NO3 Absorption rate or ﬂux of ozone [mol m–2 s–1 ] pH of the point of zero charge of a surface pHPZC

Ozone-Based Technologies in Water and Wastewater Treatment PO3 rd R Rct Sc TOC TOCc TOC∗c TOCo TOD ug X z

129

Partial pressure of ozone in gas [Pa] Rate of ozone decomposition [mol m–3 s–1 ] Kinetic constant for TOC removal during ozonation [L mol–1 s–1 ] Hydroxyl ozone ratio deﬁned by Eq. 29 –1 Schmidt number [µL ρ–1 L DO3 ] Total organic carbon [mg L–1 ] Organic carbon refractory to ozonation [mg L–1 ] Organic carbon in oxalate, acetate and formiate [mg L–1 ] Initial total organic carbon [mg L–1 ] Total ozone dose transferred [mol L–1 ] Superﬁcial gas velocity [m s–1 ] Ozone dose transfer at the beginning of the ozonation [mol L–1 ] Stoichiometric coefﬁcient

Greek letters εg Gas holdup Liquid viscosity [kg m–1 s–1 ] µL ρL Liquid density [kg m–3 ] Surface tension [N m–1 ] σL τ Hydraulic retention time [s] θ Unit fraction of catalyst occupied sites

1 Fundamentals of Ozonation Processes 1.1 The Molecule of Ozone Ozone is a bluish coloured gas with a boiling point of 161.3 K (–111.9 ◦ C) and a melting point of 80.7 K (–192.5 ◦ C). Mixtures of ozone and oxygen with more than 20% ozone become explosive. In practice, the risk of explosion does not exist because corona discharge commercial ozone generators produce much lower concentrations. From microwave spectroscopy, it has been shown that the molecule of ozone has an O–O distance of 127.2 pm and an O–O–O angle of 116.78◦ . The structure of the ozone molecule has been represented by resonance theory by two main equal weighting open structures plus a cyclic form revealed by the electron diffraction method (Fig. 1).

Fig. 1 Resonance structures for the molecule of ozone

130

A. Rodríguez et al.

Table 1 Oxidation potential for common oxidants referred to a normal hydrogen electrode Oxidant

Potential E0 (V, 25 ◦ C)

Fluoride Hydroxyl radical Atomic oxygen Ozone Hydrogen peroxide Permanganate Chlorine dioxide Hypochlorous acid Chlorine Oxygen

3.06 2.80 2.42 2.07 1.78 1.68 1.57 1.49 1.36 1.23

The central atom in the open structures forms an sp2 hybridization with one lone pair and positive charge that explains the strong electrophilic behaviour of the molecule. Ozone has a dipole moment of 0.5337 D, a consequence of the electron density of the open structures that strongly inﬂuences the chemistry of ozone. Ozone is a very reactive molecule, with a redox oxidation potential of 2.07 V. In fact it is one of the strongest oxidizers available for water treatment (Table 1). 1.2 Solubility of Ozone in Water The rate and extent of oxidation/mineralization of water pollutants depends on the concentration of dissolved ozone, CO3 . It is, therefore, an essential parameter in the design of water treatment facilities. The ozone mass balance in a volume element of aqueous phase during an isothermal ozonation process controlled by the chemical step is shown in Eq. 1. The value of CO3 ∗ , the volumetric transis determined by the ozone solubility in water, CO 3 fer coefﬁcient, kL a, and the ozone decomposition kinetic constant, kd , due to reactions between ozone and water and the compound dissolved in it: dCO3 ∗ n = kL a CO – C . – k d CO O 3 3 3 dt

(1)

At low pressure, ozone is only slightly soluble in water and if ideal gas behaviour and a negligible ozone transfer resistance in the gas phase are assumed, the relationship between the partial pressure of ozone, PO3 , and its solubility in water can be expressed by Henry’s law: ∗ PO3 = CO H . 3 e

(2)

Ozone-Based Technologies in Water and Wastewater Treatment

131

Due to decomposition of ozone in water, the experimental determination of ∗ by means of Eq. 2 parameters from Eq. 1 is not easy. It is usual to calculate CO 3 taking H e values from published correlations, such as those of Andreozzi et al. and Rischbieter et al. [1, 2]: B log H e = A – , (3) T where A and B are parameters that depend on the ionic strength of the solution; Roth and Sullivan [3], where H e (atm mole fraction–1 ) is expressed as a function of temperature and pH of water: 2428 7 0.035 , (4) H e = 3.84 × 10 COH– exp – T or Sotelo et al. [4], in which H e (kPa mole fraction–1) depends on temperature, pH, ionic strength (i) and type of salt dissolved in water: 2118 9 0.012 exp(0.96 i) . (5) H e = 1.03 × 10 COH– exp – T Equation 5 corresponds to sodium phosphate solutions and 0 ≤ T ≤ 20 ◦ C, ∗ values estimated from these 2 ≤ pH ≤ 8.5, and 10–3 M ≤ i ≤ 10–1 M. The CO 3 or similar equations are close to the ozone solubility values in real wastewater, although in cases where an important deviation between the estimated and real solubility values is expected, the Henry’s law constant must be experimentally measured [5]. 1.3 Ozone Mass Transfer The absorption rate of ozone in water, NO3 , can be expressed as: ∗ NO3 a = kL a CO – C O 3 , 3

(6)

where kL is the ozone mass transfer coefﬁcient and a the speciﬁc gas–liquid interfacial surface inside the ozonation reactor. As indicated in the preceding section, the concentration of dissolved ozone depends also on the rate of ozone decomposition, r d : n . r d = k d CO 3

(7)

Expressions similar to Eq. 7 permit an easy estimation of the ozone consumption rate in complex systems, such as those that occur in the ozonation of wastewater (see Sect. 2.4). The parameter k d in Eq. 7 is not a real kinetic constant because, besides temperature, its value depends on the properties of the water matrix: organic and inorganic matter dissolved, pH, alkalinity and ionic strength.

132

A. Rodríguez et al.

The ozonation kinetics may be controlled either by physical absorption or by chemical reaction. The value of the Hatta number gives a rule to determine the rate-controlling process for a set of given conditions [6]. The Hatta number is calculated with the values of kL and k d and assuming the double ﬁlm model of Lewis–Whitman [7]: 2 n–1 n + 1 DO3 k d CO3 , (8) Ha = kL where DO3 is the ozone diffusivity in water and n the ozone decomposition kinetic order. For Ha < 0.3, the rate of ozone absorption is higher than the ozone decomposition rate and, therefore, chemical kinetics controls the ozonation process (Eq. 1). Operation conditions should guarantee that the process is controlled by the chemical reaction step in order to provide a maximum ﬂow of oxidant. At operational conditions with Ha > 0.3 values, the ozonation decomposition rate is so high that the concentration of ozone in water cannot be measured and the absorption step controls the overall ozonation process. In these cases, the ozone transfer model must take into account the contribution of the chemical reaction to the absorption expressed by the enhancement factor (E), either calculated by the general approach [7, 8] or by means of experiments [9]: ∗ NO3 a = kL aCO E. 3

(9)

The diffusivity of ozone can be calculated by the Wilke–Chang type correlation of Haynuk et al. or by means of ozone-speciﬁc expressions, such as those proposed by Matrozov et al. (A = 4.27 × 10–10 ) or Jonson and Davis (A = 5.9 × 10–10 ) with the following expression [10]: DO3 = A

T , µL

(10)

where DO3 is in m2 s–1 , T in K and µL , the solution viscosity, in poise. The mass transfer coefﬁcient, kL , can be estimated from equations such as that proposed by van Dierendonck for stirred tanks in which µL and ρL , the viscosity and density of the aqueous solution, are expressed in SI units: µL g –0.5 kL = 0.42 3 Sc . (11) ρL In bubble columns and for bubble sizes db < 2 mm, Calderbank proposed the same equation to estimate kL and Eq. 12 for bubbles with db > 2 mm: kL = kL(db =2 mm) 500 db .

(12)

The bubble diameter can be estimated from the operation parameters ug (superﬁcial gas velocity) and εg (gas holdup into the column) and the liquid-

Ozone-Based Technologies in Water and Wastewater Treatment

phase properties ρL (density) and σL (surface tension): 6 1 – εg ug ρL g 0.5 =2 0.25 . db σL σL g

133

(13)

ρL

If all bubbles are spheres with the same size, a can be calculated by: a=

6εg . db

(14)

Although in Eqs. 10–14 the contribution of pollutants present in wastewater has not been taken into account, the values of kL , DO3 and a obtained from them will be used to characterize the transfer phenomenon taking place in ozonation processes. According to Beltrán [11], kL , DO3 and a can be experimentally determined in the wastewater where ozonation processes take place provided the appropriate kinetic regime is chosen. The experiments performed to determine kL a and k d consist in bubbling a continuous gas ﬂow containing ozone through the wastewater in a stirred tank or bubble column where ozonation takes place. Figure 2 shows the concentration of dissolved ozone during the ozonation of a wastewater (Table 5, D070208) from the secondary clariﬁcation of a municipal wastewater treatment facility. The experiment was carried out at 25 ◦ C in a 5-L stirred tank agitated at 1000 rpm with a four-blade turbine. The gas, a mixture of ozone and oxygen with a 45.9 g Nm–3 ozone concentration, was bubbled at a rate of 0.36 Nm3 h–1 . During the experiment the pH was in the range 8.04–8.25. Three different zones can be appreciated in Fig. 2. Zone I is characterized by a strong increase in ozone dissolved concentration and is followed by zone II, where the ozone concentration reaches a stationary value, CO3 s . In zone III the concentration of ozone decays after stopping the gas ﬂow. Assuming that the decomposition of ozone follows ﬁrst-order kinetics, Eq. 1 applied to zones II and III leads to the following expressions: ∗ 0 = kL a CO – CO3 s – k d CO3 s , (15) 3 dCO3 =– k d CO3 . (16) dt The integration of Eq. 16 yielded k d . The solubility of ozone was calculated as indicated in Sect. 1.2 and the value of kL a was obtained from Eq. 15. The experimental value of kL a can be up to two or ﬁve times higher than the corresponding estimation from literature correlations due to the speciﬁc composition of wastewater [5]. For the operational conditions of the experiment that is represented in ∗ , k a and k were 0.247 mM, 0.614 min–1 Fig. 2, the calculated values of CO d 3 L –1 and 0.139 min , respectively. (The equation from Rischbieter et al. (A = 5.12, B = 1230 K–1 ) was used to determine He .)

134

A. Rodríguez et al.

Fig. 2 TOC and CO3 values during the ozonation of D070208 wastewater (Table 5). pH: 8.04–8.25, T: 25 ◦ C, gas ﬂow rate: 0.36 Nm3 h–1 , gas ozone concentration: 45.9 g/Nm3 , kL a = 0.614 min–1

The decomposition of ozone in water makes the experimental determination of kL a complex. To overcome this, and taking the surface renewable theories into account, the ozone mass transfer coefﬁcient can be based on the corresponding value of some less reactive compound such as oxygen at the same pressure and temperature: kL a O = kL a O 3

2

DO2 DO3

0.5 .

(17)

1.4 Decomposition of Ozone in Water Ozone in aqueous solution decomposes through a complex mechanism initiated by reaction with a hydroxide ion and followed by formation of several radical oxidizing species, such as HO, HO2 and HO3 . The structures of ozone and HOx in liquid water remain uncertain. Chalmet and Ruiz-López [12] combined quantum and classical computer simulations and showed that even if ozone undergoes electron polarization, it does not participate in hydrogen bonds with liquid water. In contrast, HOx form strong hydrogen bonds, being better proton donors but weaker proton acceptors than water. Their electronic and geometrical structures are signiﬁcantly modiﬁed by the solvent, suggest-

Ozone-Based Technologies in Water and Wastewater Treatment

135

ing that water plays a crucial role in oxidation mechanisms initiated by ozone in liquid water. Concerning the mechanism and kinetics of ozone decomposition, the reaction follows a chain process extensively studied by Buhler et al. [13], Staehelin et al. [14], Tomiyasu et al. [15] and Hoigné [16]. In the absence of UV radiation or solid catalysts, the initiation takes place through a reaction between ozone and the hydroxide ion to yield a hydroperoxide (HO2 · ) and a superoxide radical ion (O2 · – ): ki,1

O3 + HO– –→ HO2 · + O2 · – kOH– = 70 M–1 s–1 . In the presence of hydrogen peroxide, initiation takes place by reaction of ozone with the hydroperoxide ion, HO2 – , the conjugate base of hydrogen peroxide: ki,2

O3 + HO2 – –→ HO2 · + O3 · – kHO2 – = 2.2 × 106 M–1 s–1 . Propagation involves the formation of ozonide radical ion O3 · – , the radical species HO3 · and HO4 · and several reactions of hydrogen peroxide, an intermediate product of the degradation chain: HO2 · O2 · – + H+ , O3 + O2 · – → O3 · – + O2 , O3 · – + H+ HO3 · , HO3 · → HO· + O2 , O3 + HO· → HO4 · , HO4 · → HO2 · + O2 , HO2 – + H+ H2 O2 , HO· + H2 O2 → HO2 · + H2 O , HO· + HO2 – → HO2 · + HO– . Homogeneous termination takes place by reactions consuming radicals: → O3 + HO– , HO· + O3 HO4 · + HO4 · → H2 O2 · + 2O3 (tentatively proposed) , HO4 · + HO3 · → H2 O2 · + O2 + O3 (tentatively proposed) . There are a wide variety of compounds able to promote or inhibit the chainreaction processes. Promoters of the free-radical reaction are substances capable of regenerating the superoxide anion from the hydroxyl radical. Common organic promoters include formic and glyoxylic acids, primary alcohols and humic acids. The inhibitors of the free-radical reaction are compounds capable of consuming hydroxyl radicals without regenerating the superoxide anion. These include bicarbonate and carbonate ions, tertiary alcohols like tert-butanol and some humic substances [11, 17].

136

A. Rodríguez et al.

The formation of hydroxyl radicals from ozone can be enhanced by the presence of solid catalysts. In the case of metal oxides, heterogeneous ozone decomposition is determined by the presence of surface hydroxyl groups acting as Brönsted acid sites. These sites also determine the charge of the surface as a function of pH, and therefore the ion-exchange behaviour of the catalyst. In addition to this, metal oxides have Lewis acid sites that, in an aqueous solution, allow water molecules to coordinate on their surface [18]. The adsorption of ozone requires the displacement of coordinated water and is strongly dependent on the presence of other bases. In the case that a Lewis site is accessible to ozone, the mechanism for its adsorption/decomposition on a catalytic surface would follow a mechanism similar to that used for explaining gas-phase decomposition [19]: O3 → (O3 )ads (O3 )ads → (O)ads + O2

(i) (ii) .

The interaction of the ozone molecule with an oxidized site may yield adsorbed or non-adsorbed oxygen: O3 + (O)ads → 2O2 , O3 + (O)ads → O2 + (O2 )ads → 2O2

(iii) .

In aqueous solution, the hydroxide ion is expected to act as a strong inhibitor of the adsorption ability of the catalyst by blocking Lewis acid sites. Therefore, the catalytic activity at high pH should proceed by a redox mechanism involving surface hydroxyl groups. Ozone would react with them to yield an ozone anion radical or another active species able to oxidize organic compounds either in solution or on the surface. Activated carbon is particularly efﬁcient as an initiator in the decomposition reaction of ozone in the liquid phase [20]. The capacity of activated carbon to transform ozone into hydroxyl radicals depends on its surface properties. It has been demonstrated that metal centres, electrons from graphenic layers and basic surface groups like chromene and pyrone are active sites for ozone adsorption [21]. These basic Lewis sites are located at π-electron-rich regions and behave as a Lewis base in aqueous solution [22]: πC + 2H2 O → πC–H3 O+ + OH– . The molecule of ozone may attack the basic delocalized π-electron system or lone pairs in pyrrolic groups with the generation of hydroxyl radicals [23]: · · ·NH + O3 → · · ·N+ O– + HO2 · , HO2 · → H+ + O2 · – , O2 · – + O3 → O3 · – + O2 , O2 · – + H+ → HO· + O2 .

Ozone-Based Technologies in Water and Wastewater Treatment

137

The generation of radicals from the interaction between ozone and activated carbon has been studied by the Rct methodology using pCBA as probe compound [24]. Sánchez-Polo et al. [21] showed that the interaction between ozone and groups on the surface of activated carbon leads to an increase of the concentration of superoxide radical ion enhancing ozone transformation into hydroxyl radicals. As the activity of activated carbon decreased with ozone exposure, it has been suggested that activated carbon does not behave as a true catalyst but rather as a conventional initiator or promoter for the ozone transformation into radicals. Figure 3 shows the transient response of dissolved ozone concentration after charging a semicontinuous reactor with a catalyst concentration of 0.5 g/L. The mixture of ozone and oxygen was bubbled into the liquid by means of a porous glass disc with a total gas ﬂow of 240 NL/h. The catalysts used were titanium dioxide Degussa P25 and activated carbon (AC). The concentration of ozone in the liquid was measured using a Rosemount 499A OZ ozone amperometric sensor equipped with Pt 100 RTD temperature compensation and checked against the Indigo Colorimetric Method (SM 4500-O3 B). The signal was continuously monitored by means of a Rosemount 1055 Dual Input Analyzer connected to an Agilent 34970A data acquisition system. The unsteady-state catalytic decomposition of ozone can be modelled assuming that simultaneous non-catalytic reaction follows a ﬁrst-order kinetic expression. Ozone was supposed to adsorb on the surface of titanium dioxide,

Fig. 3 Transient state decomposition of ozone at 25 ◦ C after introducing 0.5 g/L of TiO2 and AC while bubbling ozone (40–42 g/Nm3 at 240 NL/h and pH = 5)

138

A. Rodríguez et al.

so that its decomposition takes place according to the mechanism based on gas-phase reaction and described before. In the absence of data on adsorbed and non-dissociated ozone intermediates, the ﬁrst reaction can be described as: O3 →(O)ads + O2 . A further ozone molecule reacts with the oxidized site to yield non-adsorbed products. The concentration of ozone can be calculated by solving the following system of differential equations: dCO3 ∗ = kL aE CO – C (18) O3 – k1 cs CO3 (1 – θ) – k2 cs CO3 θ – k d CO3 , 3 dt dθ (19) – ct = – k1 CO3 (1 – θ) + k2 CO3 θ , dt where θ is the fraction of catalyst occupied surface sites, k1 and k2 are the rate constants for the catalytic reactions (i + ii) and (iii) and cs is the bulk concentration of solids. Rosal et al. [25] reported the following kinetic constants at pH = 5 and 20 ◦ C for ozone decomposition on titanium dioxide: k1 = 7.21 × 10–3 ± 3.1 × 10–4 m3 kg–1 s–1 , k2 = 2.73 × 10–4 ± 2.5 × 10–5 m3 kg–1 s–1 and k d = 8.74 × 10–3 ± 1.3 × 10–4 s–1 . Lin et al. [19] compared average rates of decomposition of aqueous ozone, showing that oxides with lower ∆hf exhibit higher activities but always lower than those of noble metals and much lower than those of activated carbon. Another factor that has been pointed out is the fact that ﬁne catalyst particles may enhance the absorption of ozone by a “shuttle” mechanism involving the physical adsorption of ozone on the surface of particles [26]. For P25 titanium dioxide the maximum enhancement, denoted by E in Eq. 18, represented three times the mass transfer rate of ozone in a particle-free liquid [25]. 1.5 Ozone Reactions with Organic Compounds Ozonation may take place by the direct reaction of the ozone molecule with the target compound or by means of hydroxyl radicals produced from the decomposition of ozone in aqueous media. It has already been stated that hydroxyl radicals are strong secondary oxidants produced as a consequence of ozone self-decomposition in water. In practice, both direct and indirect reactions take place simultaneously, but when an oxidation process is specifically designed to enhance the concentration of HO· radicals in a solution, one speaks of an advanced oxidation process (AOP). The data of Acero and von Gunten [27] and Bufﬂe et al. [28, 29] allow some insight into the order of magnitude of the concentration of both oxidants in an ozonation process. These researchers found that the ratio of the concentration of hydroxyl radicals to dissolved ozone ranges from 10–6 to 10–8 , the former value being encountered in AOP while the latter is typical from the ozonation of drinking water. The hydroxyl concentration during the early stages of ozone

Ozone-Based Technologies in Water and Wastewater Treatment

139

decomposition in water is greatly enhanced by the presence of amines or phenols through the formation ozonide (O3 · – ) or superoxide (O2 · – ) radical anions [23]. The direct reactions of ozone with organic compounds in aqueous solutions are essentially limited to those taking place with unsaturated and aromatic compounds and are governed by the dipolar structure of the ozone molecule. The 1,3-cycloaddition to unsaturated compounds leads to the formation of a primary ozonide:

In a protonic solution, the primary ozonide decomposes via a zwitterion that yields a hydroperoxide. This three-step process is called the Criegge mechanism.

Aromatic compounds do not undergo cycloaddition. Instead, the ozone molecule attacks electrophilic positions in the aromatic ring. Electrondonating groups like –OH or –NH2 induce a high electronic density in the ortho and para positions and, consequently, in these positions aromatic compounds react actively with ozone. Electron-withdrawing groups such as –COOH deactivate the aromatic ring for the substitution reaction. The reaction is favoured by a resonance of the intermediate. For example, the attack in the ortho position of phenol takes place by the following mechanism [30]:

Hydroxyl radicals initiate oxidative degradation by three reactions: hydrogen abstraction, hydroxyl addition and electron transfer. A saturated organic compound may be attacked by a hydroxyl radical and may undergo hydrogen abstraction, a negligible reaction in compounds with aromatic rings and

140

A. Rodríguez et al.

double bonds [31]. It has been reported that the indirect oxidation of methyltert-butyl ether starts with the abstraction of an α-hydrogen to form an organic radical, which reacts with oxygen to yield a peroxy radical with a large (∼ 109 M–1 s–1 ) second-order rate constant [32]. The peroxy radical can abstract hydrogen to form α-hydroperoxy methyl-tert-butyl ether. In aqueous solution, the reaction continues with the hydrolysis of the oxygen–oxygen bond to produce tert-butyl formate and, subsequently, formic acid and tertbutyl alcohol: O2

(CH3 )3 COCH3 + HO· → (CH3 )3 COC· H2 –→ (CH3 )3 COCH2 OOH , –H2 O

(CH3 )3 COCH2 OOH –→ (CH3 )3 COCHO , +H2 O

(CH3 )3 COCHO –→ (CH3 )3 COH + HCOOH . Unsaturated and aromatic compounds undergo hydroxyl addition, a reaction with a very high rate (109 –1010 M–1 s–1 ) and a product distribution markedly affected by substituents. The hydroxyl radical is a strong electrophile and, in the case of aromatic rings, preferably adds at electron-rich sites [33]. For example, the attack of hydroxyl on aniline leads to ortho- and para-hydroxy compounds [34]. The stabilization of radical intermediates produced during the addition of hydroxyl radicals may take place by hydrogen abstraction or by electron transfer and proton elimination. Further reactions lead to ring opening and the formation of open conjugated structures.

Electron transfer is the other mechanism of hydroxyl oxidation, commonly encountered in oxidation of transition metal ions, which is also described in organic compounds in which large substituents avoid addition reaction [35].

2 Ozone Uses in Water Treatment Ozone is used as the only oxidant or in association with other oxidants or energy (AOPs) in surface water, ground water or wastewater treatments. The ozone-based technologies have the common objective of optimizing the use of ozone to improve the disinfection or the removal of the pollutants present in water. The reason is not only the fact that it is an expensive oxidant, but also that it induces the generation of toxic oxidation intermediates. To reach this

Ozone-Based Technologies in Water and Wastewater Treatment

141

goal, it is necessary to develop models whose level of complexity depends on the knowledge of the processes. 2.1 Precipitation of Oxides Iron and manganese are undesirable in drinking water because of their effect on the appearance and taste of the water, their ability to cause black or reddish staining or the formation of sediments. The rusty or brown stains on fabrics are of concern because they are not removed by usual detergents. Iron oxide deposits on tanks, water heaters and pipelines create problems of water supply related to equipment maintenance. These pollutants are not health threatening so the EPA does not set a mandatory water quality standard. The guideline standards for both these metals have been established in their soluble states, taking into account the growth of iron- and manganese-oxidizing bacteria that strongly affects the overall water quality. In the United States, the National Secondary Drinking Water Regulations include secondary maximum contaminant levels (SMCLs) as a guideline to avoid aesthetic effects related to odour, taste and colour. Current SMCLs are 0.3 mg/L for iron and 0.05 mg/L for manganese. The Council Directive 98/83/EC on the quality of water intended for human consumption includes iron and manganese in Annex I, Part C, with values ﬁxed only for monitoring purposes of 0.2 mg/L for iron and 0.05 mg/L for manganese. The way iron and manganese should be removed depends on their oxidation state and concentration. Both can be present in water in dissolved form with oxidation states that depend on pH (Fe2+ , Mn2+ , Fe3+ , Mn4+ ) or in colloidal particle suspension. Ground waters, being anaerobic, have higher iron and manganese contents than aerated water. In the latter case, the redox potential of the water allows the oxidation of reduced ionic forms into insoluble oxides. As concerns the oxidation mechanism, there is a certain controversy over whether it consists of an oxygen transfer from ozone to the reduced metal or an electron transfer from the reduced metal to ozone [36]. Iron and manganese are usually removed by oxidation of the dissolved forms into an insoluble form by aeration or by chemical oxidization followed by sand ﬁltration. The success of removal by oxidation depends not only on the oxidant used and its concentration, but also on pH and on the presence of natural organic matter. Oxidation takes place at a faster rate at higher pH values and the presence of organic matter makes removal more difﬁcult. Both iron and manganese tend to form bonds with humic acids and other natural organic matter compounds present in water. When air is used as oxidant this causes removal difﬁculties and in this case the oxidation with ozone is recommended. In general, the removal of iron is normally easier than that of manganese, but a high content of iron requires treatment with several tanks in series. It has been stated that the oxidation of iron with ozone is rapid,

142

A. Rodríguez et al.

but tends to form colloidal particles difﬁcult to remove by sand or anthracite ﬁltration [37]. The ozone dose required for oxidation can be estimated stoichiometrically as 0.43 mg/mg iron and 0.88 mg/mg manganese, the latter for 8.0 < pH < 8.5, from the following reactions [38]: 2Fe2+ + O3 + H2 O → 2Fe3+ + 2HO– + O2 , 2FeO + O3 + 3H2 O → 2Fe(OH)3 + O2 , Mn2+ + O3 + H2 O → Mn4+ + 2HO– + O2 , MnO + O3 → MnO2 + O2 . Other oxidants may be used to remove iron and manganese by oxidation, but the dose of oxidant is higher. Table 2 shows the usual values for precipitation of iron and manganese from drinking water as a function of the oxidant [39]. The removal of oxides of iron and manganese may be carried out using different ﬁltration media such as conventional beds of anthracite and sand with chemically bonded manganese oxide. The most suitable, however, is manganese greensand, a granular form of the zeolite mineral glauconite coated with manganese oxide that bonds due to the ion-exchange properties of glauconite [40]. This manganese-modiﬁed ﬁltration medium also exhibits a catalytic effect in the chemical oxidation of iron and manganese removal. If necessary, the coating is regenerated by addition of potassium permanganate to oxidize the MnO to MnO2 . Backwashing of the greensand removes the precipitated oxides from the bed. A preﬁlter to remove most of the precipitated iron prior to the manganese greensand also prolongs the service run and reduces the pressure drop on the bed of greensand. As concerns residual waters, ozone has been proposed to remove arsenic from the wastewater of nonferrous metallurgical industries [41]. Arsenic is a constituent of most sulphide ores and concentrates processed in nonferrous metallurgical industries. Process wastes have to be treated in an environmentally acceptable manner because of the environmental legal regulations. The National Primary Drinking Water Regulations (EPA) limit the levels of arsenic in drinking water to below 0.010 mg/L. Community water systems exceeding 0.005 mg/L (one half of the arsenic MCL) must notify their customers in their

Table 2 Commonly accepted dosing of oxidants required to remove iron and manganese from drinking water Oxidant

Iron (mg/mg Fe)

Manganese (mg/mg Mn)

Chlorine Chlorine dioxide Potassium permanganate Ozone

0.62 1.21 0.94 0.43

1.27 2.45 1.92 0.88

Ozone-Based Technologies in Water and Wastewater Treatment

143

annual reports. The Council Directive 98/83/EC also includes a maximum level of 0.010 mg/L (Annex I, Part B, Chemical Parameters). The ﬁrst step in the removal of arsenic takes place by precipitation by sulphide formation. The solubility of arsenic sulphide is about 30 mg/L, high enough to require a further treatment prior to discard to the environment. Ozone can be used to oxidize As(III) to As(IV), which in the presence of Mn(II) forms a precipitate with a Mn/As mole ratio around unity, believed to be MnAsO4 ·nH2 O. The residual arsenic concentration depends on the initial manganese and iron concentrations and can be brought below the mandatory limit of 0.010 mg/L. The precipitation of arsenic with manganese by ozonation is also effective for removing arsenic in the pH range of 1–2 where ferric arsenate and ferric hydroxide do not precipitate. It has been reported that the conversion of As(III) to As(V) was fast with ozone with simultaneous oxidation of iron and manganese. The sequestering effect of the resultant As(V) played an important role. The sorption of freshly precipitated Fe(OH)3 was also signiﬁcant and estimated to be 15.3 mg As/g Fe(OH)3 [42]. 2.2 Disinfection of Drinking Water Ozone has been used since 1919 in drinking water disinfection. It is a strong biocide which is able to deactivate resistant pathogen microorganisms resistant to chlorine and chlorine dioxide, such as Cryptosporidium parvum oocysts. The ozone-based technologies for drinking water disinfection try to provide operation conditions which do not favour indirect ozone reaction via hydroxyl radicals [43]. Ozone doses should eliminate and/or reduce the concentration of faecal microorganisms (faecal coliforms and Escherichia coli) to values that exclude any risk to human health. In case some pathogen microorganism refractory to ozone treatment exists, it is necessary to speciﬁcally determine the required ozone doses. With this object the integral ct-exposure parameter (ct) is deﬁned by multiplying the disinfectant concentration in water by the time that the microorganism is in contact with it: ctO3 = CO3 dt . (20) According to the microorganism deactivation model of Chick–Watson and assuming ﬁrst-order deactivation kinetics, ct O3 determines the reduction of viable microorganisms from an initial concentration No to a ﬁnal concentration N. For a given ozonation time in batch or plug ﬂow:

N log No

t CO3 dt .

= – kN 0

(21)

144

A. Rodríguez et al.

Equation 21 allows the calculation of the ozone necessary to obtain an efﬁcient disinfection for a given microorganism with deactivation constant kN [44]. The ﬂow model permits the determination of the ozone concentration proﬁle and therefore, the value of ctO3 [45, 46]. When this information is not accessible, the usual solution is to multiply the concentration of ozone at the exit of the reactor by the time that 10% of an inert tracer injected by pulse is inside the reactor, ct10 . The relationship between the concentration of ozone dissolved, CO3 , and the total ozone dose (TOD) transferred can be obtained from Eq. 1, assuming ﬁrst-order kinetics for ozone decomposition [47]: t CO3 = TOD – X – k d

CO3 dt ,

(22)

∗ kL a CO – C O 3 dt , 3

(23)

0

t TOD = 0

where X is the TOD at the beginning of the ozonation, where it is possible that the ozone decomposition is higher than the ozone absorption rate (mass transfer control) and the ozone dissolved is not detected. Solving Eq. 22 the following expression for X is obtained: t X = TODi =

∗ kL aCO . 3

(24)

0

From Eq. 22 and for a process in which ﬂow follows the continuous stirred tank reactor (CSTR) model, the concentration of ozone inside the reactor is given by: CO3 =

TOD – X , 1 + kdτ

(25)

where τ is the hydraulic retention time (HRT). From the Chick–Watson model and Eq. 25, the extent of the ozone disinfection in a CSTR, expressed as the relationship between the actual (N) and initial (No ) concentration of microorganisms, is: N 1 . = No 1 + k TOD–X τ N 1+k τ

(26)

d

This equation connects the operational parameters TOD, X and τ and the kinetic parameters of the process (kN , k d , kL a) with the required extent of disinfection. The deactivation constants for E. coli, Bacillus subtilis spores, Rotavirus, Giardia lambia cysts, Giardia muris cysts and Cryptosporidium

Ozone-Based Technologies in Water and Wastewater Treatment

145

parvum oocysts, which are refractory to ozonation, have been reported [43] but in general, data on microorganism deactivation kinetics are scarce. 2.3 Natural Water and Wastewater Treatment The objectives of ozone-based treatments of surface water, ground water and wastewater are disinfection and the elimination of dissolved organic matter. Water dissolved organic compounds may present a huge variability in proportion and nature of pollutants (persistent organic pollutants (POPs), personal care products (PCPs), endocrine disruptors) depending on their source. The ozone-based technologies for natural water and wastewater treatment provide operation conditions favouring the direct or radical ozone reactions (AOPs). Ozone can be used as the only technology or in combination with other processes with the aim of improving coagulation–ﬂocculation or biodegradability, to remove pollutants in natural water treatments or as a tertiary treatment in association with biological wastewater treatment [48]. Although the number of papers published on the efﬁciency of ozone to eliminate POPs and PCPs is considerable, the use of ozone is less extended in wastewater treatments than in disinfection or in natural water treatments [49–57]. Ozone is an expensive oxidant and the necessary doses in wastewater treatments are higher than in natural water, thus increasing operational costs. However, the ability of ozone to mineralize organic matter, alone or in association with other oxidants such as hydrogen peroxide, makes it especially attractive for new developments, in particular those for which the objective is the reuse of wastewater. The efﬁciency of the use of ozone requires new ozone generators as well as models of the ozonation process to optimize the ozone doses, thus reducing operational costs and avoiding toxic intermediates. Figure 4 shows the evolution of total organic carbon (TOC) during the ozonation (O3 /H2 O2 system) of aqueous solutions of a number of pollutants and their mixtures classiﬁed as PCPs: analgesics (dipyrone, diclophenac, acetyl salicylic acid), anti-inﬂammatories (ibuprofen), antiseptics (triclosan), antibiotics (tetracycline), antineoplastics (cyclophosphamide), anxiolytics (carbamazepine), hormones (oestradiol) and diagnostic compounds (acetamide). The ozonation processes were carried out in a semicontinuous mode in a 20-L bubble column reactor. Gas ﬂowed at a rate of 0.36 Nm3 h–1 (kL a = 5.6 × 10–3 s–1 ) with an ozone concentration of 29.7–40.3 g/Nm3 . The experiments were performed at pH values in the range 7.5–8.5, temperatures of 20–30 ◦ C and a concentration of hydrogen peroxide of 1.0 × 10–4 M. In all cases TOCo was reduced at least 50% during the ﬁrst 30 min of ozonation. Table 3 shows representative results and operational conditions. With the aim of determining the nature of the ﬁnal refractory organic carbon (TOCc ), the contribution of oxalate, acetate and formiate, the common ozonation end

146

A. Rodríguez et al.

Fig. 4 Ozonation of PCPs in a 20-L bubble column reactor (kL a = 5.6 × 10–3 s–1 ). Experimental values of TOC versus time t for several compounds and mixtures. pH = 7.5–8.5, T = 20–30 ◦ C, [H2 O2 ] = 1.0 × 10–4 M; gas ﬂow rate: 0.36 Nm3 h–1 , gas ozone concentration: 29.7–40.3 g/Nm3

Table 3 Experimental data for the ozonation of PCPs and mixtures Compound

TOCo × 104 O3g (M) (mM)

T (◦ C)

pH

(TOCc / (TOCc∗ / TOCo ) TOCo )

k (min–1 )

Acetyl salicylic acid Carbamazepine Diclophenac Dipyrone IBU+DCF 10% pot IBU+DCF 20% pot Ibuprofen Tetracycline Triclosan ASA+DYP+CZP+ IBU+DCF Cyclophosphamide β-Oestradiol ASA+acetamide

9.7 7.8 6.7 9.4 14.5 16.3 7.8 7.7 3.0 32.4

0.84 0.84 0.84 0.84 0.62 0.84 0.84 0.84 0.84 0.84

24 24 24 23 24 22 20 23 30 22

7.8 7.8 7.6 8 7.8 7.8 7.8 7.8 7.7 7.8

0.21 0.42 0.29 0.30 0.35 0.25 0.27 0.18 0.30 0.15

0.107 0.123 0.037 0.049 0.035 0.051 0.062 0.070 0.083 0.023

8.8 2.0 17.3

0.84 0.84 0.84

24 24 24

7.8 7.8 7.8

0.26 0.39 0.53

0.30 0.22 0.27 0.17 0.17 0.17

0.40

0.071 0.026 0.071

Ozone-Based Technologies in Water and Wastewater Treatment

147

products, were measured (TOC∗c in Table 3). The experimental data of TOC were ﬁtted with pseudo ﬁrst-order kinetic expressions like Eq. 25 and kinetic constants are shown in Table 3: TOC = kt . (27) ln TOCo As in disinfection of drinking water, the ozonation models for wastewater must relate process conditions and kinetic parameters with the ozone dose required to remove pollutants. The supply of ozone is determined from a selected parameter whose value must be reduced. Depending on the information available about the wastewater, the object will be the reduction of one of various speciﬁc pollutants or to lower global parameters such as TOC or chemical oxygen demand (COD). The basic knowledge about the ozonation process determines how close the model is to reality. It is possible that in wastewater from a given industrial process the ozonation kinetic constants of the main pollutants can be available. In most cases the reaction paths of compounds present in wastewater matrices and their elimination kinetic constants are not known and therefore a global approach is normally preferred. In the experiment of Fig. 2, where ozonation by a radical pathway is not favoured, an elimination of TOC close to 10% was observed during the ﬁrst 4 min of reaction (zone I). The ozone decomposition kinetic constant was k d = 0.139 min–1 obtained by solving Eq. 28 as in Sect. 1.3: dCO3 = – k d CO3 . (28) dt The TOC removal kinetic constant was k = 0.78 min–1 , obtained by ﬁtting the experimental TOC values of Fig. 2 with a pseudo ﬁrst-order kinetic equation. Although this approach can characterize the kinetics of the process, it does not relate the ozone dissolved concentration with TOC in order to calculate the ozone dosing. Taking into account that the elimination of pollutants in water by ozone is due to direct and indirect (radical) reactions, Elovitz and von Gunten [24] proposed a model for the removal of speciﬁc micropollutants in which the ozonation process is characterized by a parameter Rct deﬁned as the relationship between the integral ct-exposure to ozone and the hydroxyl radical, the two principal oxidants in the system:

CHO· dt Rct =

(29) CO3 dt The Rct parameter characterizes the ozonation process and allows estimation of the concentration of the hydroxyl radical in water from the concentration of dissolved ozone. The balance of a determined pollutant (P) with CPo initial concentration in water in a volume element of the reactor either in batch or plug ﬂow during an ozonation process follows the expression of Eq. 30.

148

A. Rodríguez et al.

The kinetic constants kO3 and kHO· are linked with direct and indirect ozone reactions, respectively. Rct connects the extent of decontamination with the integral ct-exposure of ozone: CP ln = kO3 CPo

t

t CO3 dt + kHO·

0

0

CHO· dt = kO3 + Rct kHO·

t CO3 dt (30) 0

From Eqs. 30 and 22 the ozone requirements (concentration of ozone in water and TOD) can be linked with the elimination level of pollutants. The low concentration values of the hydroxyl radical in water (CHO· ≤ 10–12 M) make its direct measurement practically impossible. However, the integral

CHO· dt may be determined by means of probe compounds [58, 59], such as p-chlorobenzoic acid (pCBA), whose direct and indirect kinetic constants are known (kO3 /pCBA ≈ 0.15 M–1 s–1 , kHO· /pCBA ≈ 5 × 109 M–1 s–1 ). A balance to pCBA leads to the following expression to the integral ct-exposure to HO· :

pCBA

ln pCBAo kHO·

=

CHO· dt .

(31)

Fig. 5 Evolution of TOC (◦) and ozone concentration during treatment of D070208 wastewater (Table 4) with O3 /H2 O2 . pH = 8.04–8.25, T = 25 ◦ C; gas ﬂow rate: 0.36 Nm3 h–1 , gas ozone concentration: 45.9 g/Nm3 ; kL a = 0.614 min–1 and injection of 0.15 mL of H2 O2 (30% w/v) every 5 min

1.010 1.039 1.021 1.019 1.021 1.000 1.035 1.039 1.071 1.034 1.035 1.031

15.51 12.77 12.17 8.45 14.04 11.89 8.98 8.56 11.85 15.54 8.76 8.38

D070206f D070208f D070308 D070417 D070419 U070205f U070208f U070222 U070305 U070308 U070416 U070419

91.3 85.8 62.7 85.4 76.4 90.5 100 92.1 88.4 85.6 94.2 95.0

TOCo TOC CO3 gas (ppm) removed (mM) (%)

Sample

0.233 0.240 0.236 0.235 0.236 0.231 0.239 0.240 0.247 0.239 0.239 0.238

CO3 ∗ (mM)

0.038 0.075 0.120 0.049 0.050 0.027 0.037 0.053 0.054 0.038 0.043 0.051

CO3 I (mM)

3.26 1.36 0.42 1.51 2.25 3.20 3.20 2.20 2.25 2.04 2.77 2.48

kdI (min–1 )

0.823 1.084 0.996 0.337 0.496 1.226 2.130 1.744 0.513 0.220 1.270 0.525

0.050 0.091 0.134 0.070 0.078 0.043 0.056 0.090 0.067 0.051 0.065 0.068

RI CO3 II (mM–1 min–1 ) (mM)

2.32 1.01 0.33 0.94 1.22 1.85 1.91 1.04 1.69 1.42 1.63 1.69

kdII (min–1 )

0.125 0.375 0.084 0.192 0.047 – – 0.261 – 0.165 0.159 –

0.198 0.204 0.190 0.199 0.206 0.202 0.202 0.209 0.210 0.188 0.201 0.205

RII CO3 III (mM–1 min–1 ) (mM)

0.112 0.109 0.104 0.072 0.088 0.061 0.107 0.093 0.111 0.105 0.115 0.109

kdIII (min–1 )

Table 4 Reaction conditions and kinetic parameters of the ozonation (O3 /H2 O2 ) of domestic (D) and urban (U) wastewater at T = 25 ◦ C, pH = 7.6–8.2 and KL a = 0.614 min–1 (f: sample ﬁltered before ozonation)

Ozone-Based Technologies in Water and Wastewater Treatment 149

7.98 8.04 7.76 8.12 8.38 8.01 8.11 7.61 7.66 7.59 7.65 7.94

12.8 12.8 10.1 0.55 7.45 3.6 12.4 – 11.6 0.65 – –

D070206 D070208 D070308 D070417 D070419 U070205 U070208 U070222 U070305 U070308 U070416 U070419

– – 368 631 579 – – 407 492 446 601 548

Suspended Conductivity pH solids (µS) (mg/L)

Sample

44 44 182 81 66 42 35 62 71 54 67 52

COD (mg L–1 )

528.14 495.52 419.37 579.93 521.38 539.23 507.67 454.57 497.44 415.29 475.17 344.64

2.56 2.12 2.12 3.01 5.43 14.69 12.45 11.57 14.46 47.82 95.83 30.52

Alkalinity NO3 – (ppm CaCO3 ) (ppm)

Table 5 Characterization of domestic (D) and urban (U) wastewaters

9.64 7.97 5.16 12.91 11.61 8.18 13.03 7.09 6.39 4.87 2.53 1.34

PO4 3– (ppm)

111.46 112.42 84.93 86.39 80.69 126.5 131.59 102.83 114.26 90.60 116.25 101.46

SO4 2– (ppm)

59.80 62.24 51.84 54.59 54.54 77.49 75.78 67.84 73.64 61.36 78.47 81.35

Cl– (ppm)

41.27 – 36.04 49.76 42.56 – – 28.32 30.25 16.05 1.14 9.39

NH4 + (ppm)

150 A. Rodríguez et al.

Ozone-Based Technologies in Water and Wastewater Treatment

151

Figure 5 shows the evolution of ozone and TOC during the ozonation of the same wastewater of Sect. 2.3 (D070208, Table 5). Reaction conditions were the same as in Sect. 2.3 but equal volumes (0.15 mL) of hydrogen peroxide (30% w/v) were injected every 5 min in order to favour the TOC elimination by a radical pathway. Three zones that correspond with three different ozone decomposition and TOC elimination kinetics can be observed. The TOC that remained in zone III is refractory to ozonation. In zone III injections of hydrogen peroxide were stopped to avoid ozone decomposition by that compound, thereby allowing the concentration of ozone to reach a stationary state (CO3 = 0.204 mM). In the later part of zone III, the concentration of ozone decreased once gas ﬂow was stopped. The conditions in zone III allow the determination of kdIII = 0.109 min–1 and kL a = 0.619 min–1 . In zones I and II the ozone ﬂuctuations are due to the decomposition induced by hydrogen peroxide. In these two zones two minimums CO3 I = 0.075 mM and CO3 II = 0.091 nM may be identiﬁed as indicated in Fig. 5. Assuming that these two values could correspond with two stationary states of the process, the ozone decomposition kinetic constants for each zone can be calculated: kdI = 1.36 min–1 and kdII = 1.01 min–1 . Assuming that, due to reaction conditions, the ozonation proceeds by a radical mechanism, Eq. 30 applied to TOC leads to the following expression: TOCo ln = Rct kHO· TOC

CO3 dt = R

CO3 dt .

(32)

Figure 6 shows the logarithmic plot of TOC removal as a function of the integral ozone exposure following Eq. 32. Two different TOC removal zones, identiﬁed with zones I and II in Fig. 5, can be observed. The corresponding slopes are RI = 1.084 mM–1 s–1 and RII = 0.375 mM–1 s–1 . The parameter R allows a kinetic characterization of the ozonation processes even though global parameters to measure the contamination in water such as TOC or COD are used. As said before, from Eqs. 32 and 22 the ozone requirements for a given degree of TOC removal can be linked. Table 4 shows the TOC removed and the experimental values of R, k d and ozone concentration at the different steps of ozonation of two kinds of wastewaters: domestic (D) and urban (U) from the secondary clariﬁer of two municipal wastewater treatment plants. Wastewaters were treated by ozonation processes with the O3 /H2 O2 system as indicated before. With the aim of reﬂecting the seasonal variability of wastewater, the samples were collected at different dates. The main characteristics of the wastewater are shown in Table 5. Yurteri and Gurol [60] related the ozone decomposition kinetic constant k d with pH, alkalinity (Alk) and TOC. These researchers found a deviation within ±25% between k d estimated by the empirical Eq. 33 and the experimental values determined by Eq. 34 in ozonation processes of surface water

152

A. Rodríguez et al.

Fig. 6 Determination of kinetic parameters R for the ozonation. Treatment of D070208 wastewater (Table 4) with O3 /H2 O2 . pH = 8.04–8.25, T = 25 ◦ C; gas ﬂow rate: 0.36 Nm3 h–1 , gas ozone concentration: 45.9 g/Nm3 ; kL a = 0.619 min–1 and injection of 0.15 mL of H2 O2 (30% w/v) every 5 min

and wastewater (pH = 6.8–9.0, TOC = 0.3–5.3, Alk = 10–500 mg/L CaCO3 ). log k d = – 3.98 + 0.66pH + 0.61 log TOC – 0.42 log(Alk/10) dCO3 = k d CO3 – dt

(33) (34)

The removal of speciﬁc pollutants during an ozonation process can be performed by considering separately the direct reaction with ozone and the radical chain propagation with hydroxyl radical: dCM = – (zkD CO3 + kHO· CHO· )CM . dt

(35)

The corresponding balance to ozone may take into account the initiation and termination reactions as well as reaction with the organic intermediates: dCO3 ∗ =kL a(CO – CO3 ) – kHO– CHO– + kHO2 – CHO2 – CO3 3 dt – (kD CM + kDi CMi )CO3 – kr Cr CO3 .

(36)

Ozone-Based Technologies in Water and Wastewater Treatment

153

The usefulness of these models to determine the requirement of ozone depends on knowledge about (1) the stoichiometry of the direct ozone reaction, (2) the direct kinetic constants of ozone with M and with reaction products Mi, (3) the kinetic constant of hydroxyl reaction with M and (4) the ozone decomposition kinetic constants due to other radical species in water, kr . Glaze and Kang [61, 62] and Beltran et al. [63] solved a similar set of equations to determine the removal of low molecular weight halogenated compounds, polynuclear aromatics and nitroaromatic hydrocarbons. The concentration of radical species in solution was determined in all cases by assuming a stationary state. 2.4 Catalytic Ozonation The homogeneous rate of production of hydroxyl radicals from ozone is strongly dependent on pH, since the active species in the initiation of the ozone decomposition mechanism are HO– and HO2 – , the concentrations of which are directly related to the concentration of hydroxide [15, 16]. However, the ozonation under alkaline conditions presents an important drawback in the case of water with bromine levels higher than 50 µg/L due to the formation of bromate as oxidation by-product [43]. Excessive bromate formation is a major concern due to its potential carcinogenicity, which imposed a limit of 10 µg/L in drinking water standards both in the United States and Europe. Besides limiting ozone exposure, a recommended strategy to reduce bromate formation is to use pH < 7 because bromate formation is also strongly pH dependent [64]. On the other hand, under acidic conditions the formation of hydroxyl radicals and the rate of mineralization are much lower than in conventional ozonation. In this situation, a catalyst may be used to promote ozone decomposition, ozonation reactions or both. Other advanced technologies for water and wastewater treatment currently under development may avoid bromate formation. Sonochemical methods, photolysis or photocatalysis, Fenton processes or certain combinations with conventional technologies like adsorption, wet oxidation, membrane separations or biological treatment might compete with ozonation. To date, however, only ozone-based technologies have been widely used in water treatment plants, which justiﬁes the effort to develop catalytic processes able to operate in acidic media. Another important drawback of conventional homogeneous ozonation that might be overcome by using a catalyst is the inhibition due to the presence of carbonates, bicarbonates and other radical scavengers. The case of carbonates and bicarbonates is especially obvious not only from their presence in natural water and wastewater, but also because they are products of the mineralization reactions. Inhibitors reduce the efﬁciency of an ozonation process and cause a poor mineralization due to the persistency of intermediate oxidation products. Short-chain carboxylic acids represent a class of

154

A. Rodríguez et al.

organic compounds particularly refractory to the oxidation by ozone. Acids such as pyruvic, glyoxalic or oxalic are normally produced during the ozonation of complex organic molecules, their refractory character being responsible for most of the organic content of treated wastewaters. A number of attempts have been made to remove these compounds during ozonation using catalysts in either the homogeneous or heterogeneous phase [65–69]. An important question about the behaviour of carboxylic acids in a heterogeneous catalytic system is the ability of the catalyst to adsorb the organic substrate. This point will be discussed below. It has been suggested that the combination of ozonation and adsorption on activated carbon in a single process is an alternative to the treatment of wastewaters containing organic contaminants [70]. As revealed before, the ozonation efﬁciency for carbon removal is limited due to the formation of refractory short-chain carboxylic acids. On the other hand, activated carbon becomes saturated easily when treating wastewaters with high organic content, requiring frequent regeneration or replacement [71]. The ozonation on activated carbon may allow these limitations to be overcome because of a high adsorption capacity combined with high surface area and catalytic activity due surface metals and other surface chemical properties. The catalytic mechanism of ozonation on activated carbon is still unclear, but most results suggest that the role played by carbon is essentially to promote the decomposition of ozone with a subsequent increase in the production of radicals. The hydroxyl radicals formed would not be bonded to the surface, being free to react in the aqueous phase. Therefore, activated carbon would behave as an initiator of the radical-type chain reaction that transforms ozone into hydroxyl radicals. In what follows, attention will be focused on homogeneous catalytic systems and the ozonation on metals and metal oxides. By far the most commonly tested catalysts for the ozonation of organic compounds are supported and unsupported metals and metal oxides, especially titanium oxide and manganese oxide [66, 67, 72]. There is a certain controversy on the mechanism of ozonation on ionizable surfaces. Some authors assume the formation of surface oxidation sites able to interact with organic compounds [73]. Ma and Graham [74] suggested a mechanism based on the initiation of ozone decomposition by hydroxide ions linked to the negatively charged surface of manganese oxide. The interaction of organic solutes with charged surfaces must be relevant and will be discussed below. Ozonation on supported metals has been less studied. Lin et al. [19, 75] reported a considerable efﬁciency for the removal of formic acid on Pd and Pt on alumina. Noble metals, especially when supported on SiO2 , also showed appreciable activity for the decomposition of ozone in water and are candidates to catalyse an ozonation process. The ozonation of carboxylic and chlorinated carboxylic acids on Ru/CeO2 and Ru/CeO2 –TiO2 have been reported by Karpel et al. [76] and Fu et al. [77]. The results of ozonation on metals and metal oxides have in common a strong de-

Ozone-Based Technologies in Water and Wastewater Treatment

155

pendence of the reaction rate on the mode of preparation of the catalyst and on support pre-treatment. These variables should affect the interaction of the molecule with the surface and its adsorption capability for ozone or organic molecules. 2.4.1 Homogeneous Catalytic Ozonation Earlier works showed that certain metals in solution are able to increase the removal of organics from aqueous solution with respect to non-catalytic ozonation [78]. The catalytic activity of Fe(II), Fe(III), Mn(II), Ni(II), Cr(III), Ag(I), Cu(II), Co(II), Zn(II) and Cd(II) have been reported [79]. It has been proposed that the mechanism of homogeneous ozone–metal systems is based on the generation of hydroxyl radicals through an ozone decomposition reaction [80]: Fe2+ + O3 → FeO2+ + O2 , FeO2+ + H2 O → Fe3+ + HO· + HO– . On the other hand, Novell and Hoigné [81] indicated that the production of hydroxyl radicals cannot be directly related to the interaction of ozone with the transition metal and the latter reaction should be substituted by the following one: FeO2+ + Fe2+ + 2H+ → 2Fe3+ + H2 O . In addition to this, the iron-catalysed ozonation may share reactions with the classical Fenton homogeneous process. The interaction of ozone and water is known to produce hydrogen peroxide, which may produce hydroxyl radicals: Fe2+ + H2 O2 → Fe3+ + HO· + HO– , Fe2+ + HO· → Fe3+ + HO– , RH + HO· → H2 O + R· , R· + Fe3+ → R+ + Fe2+ . Organic radicals should play an important role in the reduction of Fe(III), but the regeneration of the catalyst may take place by other mechanisms with the intervention of hydroperoxy radicals, HO2 · [82]. Oxalic acid tends to form complexes with transition metals such as manganese, iron and cobalt. The formation of these complexes plays an important role in the catalytic mechanism of ozonation [83]. Andreozzi et al. [66] studied the ozonation of glyoxalic acid catalysed by manganese salts and suggested a mechanism of oxidation mediated by Mn(III): Mn(II) + O3 + 2H+ → Mn(IV) + O2 + H2 O , Mn(II) + Mn(IV) → 2Mn(III) .

156

A. Rodríguez et al.

Mn(III) reacts with an acid moiety originating the abstraction of CO2 with the consequent reduction of Mn(III) to Mn(II). Mn(III) + HOOC–R → Mn(III) – OOC–R + H+ Mn(III)–OOC–R → Mn(II) + CO2 + R· The formation of complexes has also been proposed for the ozonation of oxalic acid using Co(II) by Pines and Reckhow [83]. In this case the Co(II) complex is oxidized by ozone to a Co(III) form with subsequent elimination of Co(II), resulting in an oxalate radical anion which further decomposes. 2.4.2 Catalysis by Metals and Metal Oxides The mechanism of catalytic ozonation on oxides and metals may involve the adsorption of ozone, but not necessarily the adsorption of organic pollutants. It has been demonstrated that dissolved ozone adsorbs and decomposes on many solid surfaces other than activated carbon, the resulting radicals being responsible for indirect oxidation reactions [25, 75]. This mechanism relies on the well-known result that gas-phase ozone adsorbs on solid surfaces to yield different molecular or ionic species. Dhandapani and Oyama [84] reported that ozone decomposition on p-type oxides is consistent with the formation of superoxide (O2 – ) or peroxide (O2 2– ) species on the surface. Bullanin et al. [85] suggested that on the stronger Lewis sites ozone dissociates after adsorption to yield a surface oxygen atom. With weaker sites, ozone molecules form a coordination bond via one of the terminal oxygen atoms. Another possibility is the formation of weak hydrogen bonds with surface OH groups: O3 + S O2 + O–S , O3 + S O=O–O–S , O3 + SOH O=O–O· · ·HO–S . In the case of metals and metal oxides, the catalytic reaction may also involve the adsorption of organic molecules or ions on surface sites leading to two additional mechanisms. The ﬁrst possibility is an Eley–Rideal interaction between an adsorbed organic molecule and ozone or radicals from the bulk aqueous phase. On the other hand, an adsorbed organic molecule can react with adsorbed ozone or the products of surface ozone decomposition. Legube and Karpel [36] proposed a redox version of the latter for heterogeneous ozonation on metals, in which the ozone oxidizes surface metal atoms with the generation of hydroxyl radicals. Organic molecules are oxidized by electron transfer from the catalyst yielding back the reduced form of the metal

Ozone-Based Technologies in Water and Wastewater Treatment

157

and organic radical species: Mered n+ + O3 → Meox n+1 + O3 · – , O3 · – + H+ HO3 · , HO3 · → HO· + O2 , Meox n+1 + A → Meox n+1 – A → Mered n + A· . The adsorption of neutral organic compounds on Lewis acid sites is difﬁcult due to the competitive adsorption of water molecules on the surface. Moreover, at basic pH, the hydroxide anion should prevent any adsorption on Lewis sites. Surface interaction is easier for ionizable organic molecules in aqueous solution if the surface is charged and allows ion exchange. The surface of metal oxides exhibits ion-exchange properties and the hydroxyl groups formed behave as Brönsted acid sites and dissociate depending on the pH of the solution. M–OH + H+ M–OH2 + M–OH + HO– M–O– + H2 O If K1 and K2 are the ionization constants for the preceding surface equilibria, the point of zero charge (PZC) represents the pH at which the surface is not charged: pK1 + pK2 . (37) 2 A neutral organic compound may adsorb on metal oxide surfaces provided it is a strong enough Lewis base and the pH of the solution is near pHPZC of the oxide. Otherwise, it is reasonable to assume that only ionizable substances would be capable of interacting with charged surfaces. Carboxylates, for example, adsorb on positively charged surfaces by exchanging the corresponding counteranion [18]. The kinetics of adsorption may play an important role in the ozonation process. Figure 7 shows some results for the adsorption of naproxen (pKa = 4.60) and carbamazepine (pKa = 14.0) on TiO2 Degussa P25 with pHPZC of 6.8 [86] and TiO2 /Al2 O3 with pHPZC of 8.3, prepared by impregnation with titanium isopropoxide following the method of Zhaobin et al. [87]. As expected from the PZC of the surface, adsorption is favoured for naproxen under acidic conditions at which the surface behaves as an anion exchanger [88]. Carbamazepine does not dissociate in acidic solutions and consequently its adsorption pattern is similar for pH values in the range 3–7. Similar results have been published for other acidic solutes [67, 89]. The results indicate that the rate of adsorption may be slow enough to control the overall kinetics. For the drugs mentioned above, the adsorption ranged only 5–15% from their equilibrium value during the ﬁrst hour. Even for naproxen at pH = 3, below its pKa the adsorption was slow and equilibrium required a day or more in most pHPZC =

158

A. Rodríguez et al.

Fig. 7 Dimensionless concentration of naproxen during adsorption on TiO2 /Al2 O3 at pH = 5 (), on TiO2 P25 at pH = 3 (◦), on TiO2 P25 at pH = 7 () and carbamazepine on TiO2 P25 at pH = 5 (•) and pH = 7 (). Temperature 25 ◦ C, catalyst loading 1 g/L, initial concentration 6.0–6.5 × 10–5 M

cases. This agrees with several pieces of data showing that the adsorption of acid pollutants on metal catalysts supported on alumina is slow, taking from hours to days to complete [68, 89]. A kinetic model of ozonation should take into account this possibility. Without explicit consideration of the surface charge, the ozonation mechanisms that involve the adsorption or ion exchange of an organic compound start with the bonding of the adsorbate to a vacant site (S). The adsorbate (A) would displace coordination water and bond to the surface: A + S A–S . The reaction may then take place between adsorbed organic solute and an oxidized site on the catalytic surface following a Langmuir–Hinshelwood mechanism: O–S + A–S → AOx + 2S . Alternatively, dissolved ozone or hydroxyl radicals in solution may react with the adsorbed organic compound through an Eley–Rideal interaction: HO· + A–S → AOx + S . Some organic compounds react with dissolved ozone at a high rate. This is the case for drugs such as naproxen and carbamazepine mentioned above. Anyway, the mineralization rate is slow in non-catalytic systems in acidic

Ozone-Based Technologies in Water and Wastewater Treatment

159

conditions, so that the design of catalytic reactors is focused on refractory compounds under conditions in which direct reactions can be ignored. The rate of reaction of an organic compound combines the homogeneous reaction with hydroxyl radicals and the heterogeneous reaction of hydroxyl (or ozone) with solute with at least one reagent adsorbed. If the reaction takes place between adsorbed species and hydroxyl radicals from the bulk, the Eley–Rideal rate expression would be as follows: dCA (38) = kHO· CHO· CA + kc Cs CHO· θ , dt where θ is the fraction of surface sites occupied by adsorbate and Cs the bulk concentration of solids. Accepting the Rct concept of Elovitz and von Gunten [24], the ratio of hydroxyl radicals and ozone at any time is constant over a wide range of the ozonation process: –

Rct =

CHO· . CO3

(39)

If the catalytic reaction with the adsorbed organic compound is the ratelimiting process, an adsorption equilibrium exists at any time. Using ka and k d to denote the adsorption and desorption kinetic constants, Eq. 38 can be rewritten as follows: ka CA dCA = kHO· Rct CO3 CA + kc Cs Rct CO3 . (40) – dt ka CA + k–a If surface coverage is low, Eq. 40 can be simpliﬁed by assuming that ka CA k–a : dCA – = kHO· Rct + kc Cs Rct Ka CO3 CA . (41) dt The adsorption equilibrium constant, Ka , becomes included in a group of constants with a linear dependence on catalyst load. Equation 41 can be integrated to obtain explicitly the time-integrated concentration of ozone: CA,o ln = kHO· Rct + kc Cs Rct Ka CO3 dt . (42) CA A similar result would be obtained for a reaction between adsorbed organic compounds and oxidized catalyst sites, provided that the elementary surface step is rate controlling and an adsorption equilibrium exists at any time for both reagents: –

kc Cs CO3 CA dCA , = kHO· Rct CO3 CA + dt 1 + Ka Cs 1 + Kox CO3

(43)

where Kox represents the equilibrium constant of the surface oxidation step. If the equilibrium constants are small, the surface step would be ﬁrst order in

160

A. Rodríguez et al.

the oxidant and in the organic compound and the differential and integrated rate equation would be similar to Eqs. 41 and 42, respectively. A surface redox mechanism such as that involved in ozonation has sometimes been described by means of a Mars–van Krevelen rate expression. The rate of catalytic reaction would depend on the rate of the oxidation process, ko , and the rate of the organic compound with the oxidized catalyst, kc . Assuming that the ozone is the oxidant and ignoring the surface stoichiometric coefﬁcient, the rate of organic depletion would be as follows: –

ko kc Cs CO3 CA dCA = kHO· Rct CO3 CA + . dt ko CO3 + kc CA

(44)

If the rate of catalyst oxidation is low, the catalytic reaction would be zero order in the organic compound. Conversely, for a high rate of surface oxidation, the reaction rate would be independent of the concentration of oxidant and ﬁrst order in the oxidized organic compound. The resulting equations can be integrated and yielded equations somewhat different from Eq. 42. Although relatively common in catalysis, the Mars–van Krevelen approach recently received some criticism concerning its fundamental background. Vannice [90] proved that the kinetic expression is incorrect and that the obtained reaction orders can be derived from the Langmuir–Hinshelwood equation under more transparent assumptions. On the other hand, if adsorption of organics is so slow that it controls the overall kinetics, the rate of the catalytic process would be independent of the concentration of ozone or other oxidants: dCA = kHO· Rct CO3 CA + ka Cs CA . – (45) dt The integration of Eq. 45 leads to an expression in which the logarithmic decrease of the organic compound is not linear in the time-integrated concentration of ozone: CA,o · ln = kHO Rct CO3 dt + ka cs t . (46) CA The mechanisms described in Eqs. 42 and 46 can be discriminated by using kinetic data. For example, Eq. 42 predicts that a change in the concentration of ozone should have no effect in the logarithmic decrease of the concentration of a given compound, while Eq. 46 suggests that decreasing the ozone dose would lead to a time-integrated concentration in more time and, therefore, to a greater conversion of the organic compound. If an aggregate such as TOC is used instead of the concentration of a single compound, the kinetic expressions would show the evolution of mineralization with the time-integrated concentration of ozone. Figure 8 shows experimental data corresponding to the ozonation of naproxen (6.5 × 10–5 M in pure water) using titanium dioxide Degussa P25 as catalyst and in noncatalytic runs performed in the same conditions. Ozone was continuously

Ozone-Based Technologies in Water and Wastewater Treatment

161

Fig. 8 Mineralization during the ozonation of naproxen (6.5 × 10–5 M) on TiO2 P25 at pH = 3 () and pH = 5 (•) at 25 ◦ C and catalyst loading of 1 g/L. Empty symbols correspond to non-catalytic runs under the same conditions. The units of the integral ozone exposure are mM s

bubbled from a corona discharge ozone generator and the steady-state concentration of ozone in the liquid was 0.230 mM. Data show the existence of two different mineralization periods. During the ﬁrst period most of the TOC decay takes place, the reaction being considerably accelerated by the presence of catalyst. The second period was essentially independent of pH and reﬂects the slow mineralization of refractory compounds. The similitude between catalytic and non-catalytic plots hides at least one fundamental difference between both processes. Figure 9 shows the concentration of oxalate measured by ion chromatography (DIONEX, DX120 ion chromatograph) and reveals that, even on a neutral surface, oxalate is mineralized in conditions at which the rate of reaction is very slow. In fact, the higher degree of reaction of oxalate takes place with neutral surface charge and neutral pH at which the mineralization is not particularly deep. For the ozonation of naproxen in the runs reported in Fig. 9, the global extent of mineralization was about 50% in non-catalytic runs at pH 5–7 and reached over 60% using catalyst. The ozonation of carbamazepine allowed a deeper mineralization with 73% TOC reduction after 120 min. In fact, most of the TOC decay takes place during the ﬁrst 10–20 min where the removal of the more oxidizable compounds takes place. In non-catalytic ozonation, oxalate accounted for as much as 30% of the organic carbon remaining in the

162

A. Rodríguez et al.

Fig. 9 Concentration of oxalate during the non-catalytic ozonation of naproxen at pH = 5 () and the catalytic ozonation on TiO2 P25 of naproxen at pH = 3 (), carbamazepine at pH = 3 (◦) and carbamazepine at pH = 5 (•). Catalyst loading 1 g/L; initial concentration of naproxen 6.5 × 10–5 M

reaction mixture, in contrast with a maximum of 12% encountered in noncatalytic runs. This pattern reveals that the use of a catalyst favours not only the reactions leading to oxalate but also the mineralization of oxalate itself. In fact, oxalate was not detected in runs performed at pH = 7 with a mineralization degree of about 50%. At pH = 3, at which the surface of P25 is positively charged and may interact with oxalate anion in solution, neither the rate of mineralization nor the removal of oxalate from the solution are particularly high Therefore, the mineralization of oxalate is not favoured by positive surface charge, a result that seems to exclude a mechanism based on the ion exchange of oxalate. The results also proved that the rate of ozone decomposition is inhibited by the catalyst over a neutral surface. At pH = 7, the homogeneous rate constant for the ozone self-decomposition is 8.83 × 10–3 s–1 which lowers to 1.27 × 10–3 s–1 in the presence of 1 g/L of P25 TiO2 . The catalytic mineralization rate is a maximum at pH = 5, and the reaction is also inhibited by higher pH values. Anyway, the inhibition effect of increasing the concentration of hydroxide anion is much greater on the mineralization reaction than on ozone decomposition. The best results for the removal of reaction intermediates were obtained for slightly positive surface charge, suggesting that the adsorption of organics on Lewis sites may be the mechanism of the catalytic ozonation of naproxen and carbamazepine. In catalytic runs, the degree of mineralization was not directly linked to the accumulation of low molecular weight carboxylic acids. In all runs, low levels of acetate, formiate and other

Ozone-Based Technologies in Water and Wastewater Treatment

163

low weight acids were detected in most cases, but without an accumulation pattern linked to the evolution of TOC. These results point towards a stronger surface interaction with the ﬁrst ozonation products than with the more oxidized carboxylic acids. The dependence of the degree of mineralization on the rate of adsorption expressed by Eq. 46 was not tested and could be conﬁrmed by experiments performed in conditions at which the integral ozone exposure is not linear with time. 2.5 Applications in the Treatment of Industrial Wastewater The use of ozone-based technologies for the treatment of pollutants in water has been the focus of attention in the literature during the last few years. Examples of their application as eco-effective alternatives have been presented for several types of contaminants in water, in particular for pharmaceutical residues or PCPs. Data in scientiﬁc publications concerning the use of ozone for industrial wastewater treatments can be found to a lesser extent. As has been mentioned before, ozone is an expensive oxidant and the treatment of industrial or wastewater efﬂuents needs doses higher than those for the treatment of natural waters. In consequence, its use may be limited, but the ability of ozone to mineralize organic matter, alone or in association with other oxidants such as hydrogen peroxide, may be attractive for new developments such as in the reuse of wastewater. Combinations with other oxidation techniques, such as UV irradiation or ultrasonic techniques, can also be of interest due to a higher efﬁciency and lower cost. The effectiveness of ozone-based technologies has been evaluated in scientiﬁc publications for different industrial sectors, such as textiles, petroleum reﬁneries/phenols, pulp and paper, and electroplating wastes. Recent publications compile and review the studies performed for the treatment of recalcitrant pollutants by the use of AOPs, where possible mechanisms responsible for synergistic action are described [91]. The advantages of combined treatments are the reduction in the time of treatment and higher removal efﬁciency. In fact, most common approaches have made use of chemical (O3 or H2 O2 ) or photochemical (UV) based processes where the oxidation power for the degradation of organic pollutants can be signiﬁcantly enhanced. Table 6 shows a summary of the reviewed literature concerning AOPs applied to the treatment of industrial wastewater efﬂuents. The effectiveness of combined photocatalytic and ozonation processes has been probed for textile efﬂuents [92]. The removal efﬁciency of phenols from wastewater using a UV/H2 O2 /O3 process (pH = 7, c(H2 O2 ) = 10 mM) is complete (100%) within a period of 30 min of treatment [93]. The colouring matter is almost completely eliminated, achieving a high reduction of TOC.

164

A. Rodríguez et al.

Table 6 Applications of ozone-based processes for the treatment of industrial wastewater Industrial activities and others

Pollutants/related pollutants

Treatment

Removal efﬁciency (%)

Petroleum reﬁneries Textiles

Phenols

UV/H2 O2 /O3

Colouring matter

TiO2 /UV/O3 -BAC

Colouring matter

O3 and electroﬂocculation TiO2 /UV/O3 -BAC

100% Phenol [91] 58% TOC 90% Dyes [93] 50% TOC > 80% [94]

Paper

Domestic/ other industries Agriculture Pharmacy

PAEs (DMP, DEP, DBP, DEHP)a POPs (HCB, PBB003, PBB10, PBB18, PBB52, PBB103)b HMWc Lignin products (e.g. phenols) Organic load due to starch (e.g. saccharides, carboxylic acids) LAS Pesticides Antibiotics Steroid hormones, beta-blockers, X-ray contrast media

TiO2 /UV/O3 -BAC

Refs.

> 94.9%

[95]

> 89.3%

[95]

O3 /biological treatment 80% O3 /UV Complete

[96] [97]

Partial (no data) O3 O3 O3 /UV O3 /H2 O2 O3

Complete (aqueous solution) Complete (aqueous solution) > 90%

[98] [99] [100]

a

Dimethyl phthalate (DMP), diethyl phthalate (DEP), dibutyl phthalate (DBP) and di(2ethylhexyl) phthalate (DEHP) b Hexachlorobenzene (HCB), 4-bromobiphenyl (PBB003), 2,6-dibromobiphenyl (PBB10), 2,2 ,6-tribromobiphenyl (PBB18), 2,2 ,5,5 -tetrabromobiphenyl (PBB52) and 2,2 ,4,5 ,6pentabromobiphenyl (PBB103) c HMW: organic compounds of high molecular weight

Organic pollutants such as phthalate esters (PAEs) and other POPs in raw water are also efﬁciently eliminated by various processes which combine the use of catalyst, UV radiation and biological activated carbon, TiO2 /UV/O3 BAC [94]. The treatment of wastewater efﬂuents generated in the industry of pulp and paper also needs the use of advanced technologies. Pulp mill wastew-

Ozone-Based Technologies in Water and Wastewater Treatment

165

ater contains a signiﬁcant amount of complex organic compounds of high molecular weight (MW > 1000 Da). It means that the treatment is not feasible by biological methods. The use of ozone-based processes for this type of efﬂuent has demonstrated the capacity of this technology in enhancing the biodegradability, decreasing the toxicity and increasing the removal of organic pollutants from the efﬂuents [95]. The advantage of the combination of a pre-treatment by ozone (dosage of 0.7–0.8 mg O3 /mL wastewater) followed by biological treatment allows the conversion of organic compounds of high molecular weight (HMW) to low molecular weight (LMW), increasing the biodegradability from 5 to 50% [96]. An important factor in this process is the effect of pH. For the treatment of alkaline bleach plant efﬂuent, a superior performance of the ozonation process is under basic pH conditions. It is due to the reaction of organic compounds with molecular ozone and with oxidizing radicals, including the hydroxyl radical, which are effectively formed at high pH. When the paper industry does not employ wood to obtain the pulp and uses recycled paper, the composition of the efﬂuent and pH is not the same, it is less basic. Another example of the application of advanced technologies using ozone is the treatment of efﬂuents which contain organic compounds arising from the degradation of starch (e.g. saccharides, carboxylic acids), phenolic compounds derived from lignin and smaller amounts of other pollutants that can be persistent in the environment and are detected in fresh water (e.g. surfactants). The studies performed with O3 and/or UV have shown the utility of this approach. By this procedure, the complete degradation of lignin products and their diminishment has been demonstrated, but the organic load due to starch is not removed. The results obtained in these studies show that toxic or inhibitory compounds (e.g. phenols) are more easily oxidized than the highly biodegradable ones (e.g. glucose, fatty acids) by ozone-based technologies [97]. Linear alkylbenzene sulphonates (LASs) are anionic surfactants which are discharged into wastewaters through different sources (domestic or industrial), reaching aquatic compartments given their wide use. As a reference, in 1995 the world production of LASs was ca. 2.8 × 106 tonnes but now more than 4 × 106 tonnes are consumed globally every year. Few reports can be found on LAS degradation. AOPs have been considered as strong oxidation procedures for the degradation of such organic contaminants. The use of ozonation has been proved as the most efﬁcient approach for degrading the typical LAS mix present in municipal and industrial wastewaters where the typical pH values are slightly basic [98]. Hazardous organic contaminants such as pesticides when discharged into the environment represent a risk for human health and for the ecosystem due to their toxicity and persistence. Several publications have shown the effectiveness of ozone for removing pesticides in aqueous solution [99]. Ozonation appears to be a more efﬁcient technique which can be easily implemented

166

A. Rodríguez et al.

with UV and/or H2 O2 for treating wastewater with high organic loads. The efﬁciency depends, to a great extent, on the nature of the pollutant, and up to now few experiences have been explored and documented for real conditions [99]. AOPs are technologies whose position in the water treatment processes for industry still need to demonstrate levels of reliability and fullscale implementation. For the assessment of the removal efﬁciency of AOP processes, common procedures are based on the measurement of global parameters (i.e. TOC). The use of techniques such as gas chromatography–mass spectrometry (GC-MS) or liquid chromatography–mass spectrometry (LCMS) provides analytical information appropriate for further efﬁciency evaluation and facilitates identiﬁcation of by-products which can be of environmental concern. The use of toxicity assays in combination with chemical analysis has been considered as a strategic approach for overall assessment [99]. The efﬂuents originating from the pharmaceutical industry can show low biodegradability since they contain active substances. In particular, certain antibiotics, anti-tumour agents and analgesics are neither degradable nor adsorbable on sewage sludge. AOPs applied to remove pharmaceuticals based on the use ozone [100] are able to completely oxidize recalcitrant compounds, rendering them less harmful and forming easily biodegradable components, but also combinations of AOPs have enlarged the possibilities to treat target recalcitrant pollutants [100]. Investigations carried out with antibiotics, steroid hormones, beta-blockers or X-ray contrast media have demonstrated the removal capacity of ozone-based processes, achieving signiﬁcant elimination of those pharmaceutics in efﬂuents (> 90%). 2.6 Removal Efficiency of Pharmaceuticals in Wastewater: A Case Study As has been commented above, it is well established that urban wastewaters, which include domestic and some industrial waters, among others, represent a signiﬁcant source of contamination with a strongly contaminating effect on natural aquatic systems [101–103]. Even when they are submitted to biological treatment, it has been demonstrated by many studies that multiple organic compounds, such as pharmaceuticals, PCPs, hormones and other disrupting compounds, escape conventional wastewater treatments and some of them are becoming ubiquitous in the environment [104]. The presence of these contaminants in treated waters also endangers their reuse in diverse applications, an aspect which is of special interest since the availability of water of good quality is a critical issue and represents an essential component for sustainable socio-economic development. Consequently, the application of more exhaustive wastewater treatment protocols, including the use of new and improved technologies, the application of wider and integrated quality control strategies comprising chemical, microbiological and toxicological analysis, and the study and development of wastewater reuse strategies are

Ozone-Based Technologies in Water and Wastewater Treatment

167

goals which it is necessary undertake [105]. With this aim, an ambitious programme (TRAGUA) ﬁnanced by the Spanish Government has been initiated, which attempts to tackle the different aspects involved in the reuse of wastewater coming from sewage treatment plants (STPs). As a part of this programme, results regarding the efﬁciency of O3 and O3 /H2 O2 treatments in the removal of organic contaminants in a municipal wastewater treatment plant (WWTP) efﬂuent are presented. A suitable analytical methodology was developed in order to obtain an adequate evaluation of the processes. Two LC-MS systems equipped with modern and sensitive mass spectrometers, hybrid triple-quadrupole linear ion trap (QTRAP) and time-of-ﬂight (TOF), were used with this aim. The joint application of both techniques provided very good results in terms of accurate quantiﬁcation and unequivocal conﬁrmation of the organic pollutants present in the samples. Quantiﬁcation was performed by LC-QTRAP-MS operating in the selected reaction monitoring (SRM) mode with both positive and negative electrospray ionization, in order to cover a broad range of analytes. Limits of detection reached by the optimized method were between 0.04 and 50 ng L–1 , thus guaranteeing an exhaustive evaluation of the samples. Unequivocal analyte conﬁrmation was provided by LC-TOF-MS analysis, which allows accurate mass measurements of the identiﬁed compounds to be obtained with errors lower than 2 ppm. Figure 10 shows as an example the identiﬁcation by TOF-MS of codeine and acetaminophen, based on the accurate mass of their molecular ions and of their main fragments. With the application of the developed method, up to 40 compounds were identiﬁed in a wastewater efﬂuent after the application of a conventional biological treatment. They include mainly pharmaceuticals of different therapeutic groups, such as analgesics/anti-inﬂammatories, antibiotics, lipid regulators, beta-blockers, antidepressants, anti-epileptics/psychiatrics, ulcer healing compounds, diuretics and bronchodilators. The occurrence of many of these compounds has already been reported in environmental waters [106, 107]. Also of interest was the presence of some of their metabolites, such as 1,7-dimethylxanthine (paraxanthine) or fenoﬁbric acid, and especially the metabolites of the antipyretic drug dipyrone and its active product 4-methylaminoantipyrine (4-MAA), such as N-acetyl4-aminoantipyrine (4-AAA), N-formyl-4-aminoantipyrine (4-FAA) and antipyrine, which were detected at a high level of concentration. Finally, the disinfectant chlorophene and the pesticide diuron completed the group of detected compounds. All of them are listed in Table 7, where the concentrations found in an efﬂuent sample of a municipal WWTP are also shown. Concentration values ranged from 2 to 6590 ng L–1 . The stimulant caffeine, the diuretic hydrochlorothiazide, the beta-blocker atenolol, the analgesic/anti-inﬂammatory naproxen, the antibiotic ciproﬂoxacin and the metabolites of dipyrone (4AAA and 4-FAA) were the compounds present at the highest concentration

168

A. Rodríguez et al.

Fig. 10 Example of the identiﬁcation by TOF-MS of codeine and acetaminophen in an efﬂuent sample based on accurate mass measurements

Ozone-Based Technologies in Water and Wastewater Treatment

169

Table 7 Compounds and concentrations present in a WWTP efﬂuent and their removal efﬁciency after O3 and O3 /H2 O2 treatment Compound

Concentration in the efﬂuent (ng L–1 )

Removal efﬁciency (%) O3

Removal efﬁciency (%) O3 /H2 O2

Erythromycin Ciproﬂoxacin Sulfamethoxazole Mepivacaine Caffeine Omeprazole Carbamazepine Codeine Ketorolac Paraxanthine Atenolol Naproxen Indomethacin Propanolol 4-MAA Diazepan Metoprolol Ranitidine Fluoxetine Trimethoprim Metronidazole 4-FAA Antipyrine 4-AAA Oﬂoxacin Salbutamol Ketoprofen Mefenamic acid Sotalol Terbutaline Fenoﬁbric acid Furosemide Diclofenac Benzaﬁbrate Gemﬁbrozil Hydrochlorothiazide Chlorophene Diuron Ibuprofen

341 2559 243 2 600 104 140 657 465 132 1443 1990 37 59 18 5 53 336 782 157 185 3191 17 6590 316 10 590 64 25 11 165 531 33 61 143 1310 88 9 52

98 95 97 90 93 98 98 100 98 87 99 98 100 96 100 100 100 99 91 98 100 100 100 98 97 100 98 96 100 100 100 98 100 90 88 96 86 89 84

100 99 100 100 98 100 100 100 100 98 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 99 100 100 100 100 99 100 96 100 100 87 95 96

170

A. Rodríguez et al.

Fig. 11 Total charge of contaminants present in a WWTP efﬂuent before and after the application of O3 and O3 /H2 O2 treatments

levels (> 1 µg L–1 ). This group of compounds represents about 75% of the total charge of these contaminants in the efﬂuent, as can be observed in Fig. 11. The characterized efﬂuent from the secondary clariﬁcation of a municipal wastewater treatment facility (Table 7) was submitted to treatment with O3 and O3 /H2 O2 . The O3 treatments were carried out at 25 ◦ C in a 5-L stirred tank agitated at 1000 rpm with a four-blade turbine. The gas, a mixture of ozone and oxygen with a 45.9 g Nm–3 ozone concentration, was bubbled at a rate of 0.36 Nm3 h–1 . During the experiment the pH was in the range 8.04–8.25. The same experimental conditions were maintained in O3 /H2 O2 , but now equal volumes (0.15 mL) of hydrogen peroxide (30% w/v) were injected every 5 min in order to favour pollutant elimination by the radical pathway. The results obtained demonstrated that ozonation of wastewaters degrades pharmaceuticals with a high efﬁciency. Removals higher than 90% were reached for most of the target analytes. Only a group of ﬁve compounds, gemﬁbrozil, chlorophene, diuron, ibuprofen and mefenamic acid, yielded lower removal efﬁciencies, which were higher than 84% in all cases. The combination of ozone and hydrogen peroxide still enhances oxidizing ability, providing almost total elimination of the contaminants in most cases. Considering the total charge of compounds initially present in the wastewater, a reduction of 97% was observed after O3 treatment and their almost total elimination (99%) was reached by the application of O3 /H2 O2 , as is shown in Fig. 11. With these results, it can be concluded that both treatments can be considered as promising alternatives for pharmaceuticals and related compounds which persist through conventional biological treatments.

Ozone-Based Technologies in Water and Wastewater Treatment

171

3 Conclusions Ozone is an efﬁcient oxidant of organic matter but its production is expensive. To optimize the use of ozone it is coupled with coagulation and ﬁltration processes in surface or ground water treatments. It is used alone or coupled with UV in water disinfection processes or it is coupled with other oxidants, energy forms or catalysts in AOPs based on ozone in industrial wastewater treatment. The process of hydroxyl radical generation from ozone/hydrogen peroxide was modelled in the 1980s, which made it possible to optimize the use of ozone in the elimination of hazardous pollutants, such as pesticides, PAHs, etc. Now the system ozone/hydrogen peroxide is a new choice for water reclamation and potable reuse. The use of the ozone/hydrogen peroxide system as a tertiary treatment of domestic and urban wastewater could provide reclaimed water to use in agriculture or industrial processes. The model of these processes connects the grade of elimination of TOC and ozone doses with the ct-exposure parameter, being the rate of TOC elimination described by a ﬁrst-order kinetic equation with a kinetic parameter R which is obtained by multiplying the hydroxyl-to-ozone ratio, Rct , and the elimination kinetic constant of TOC, kHO· . The efﬁciency of homogeneous catalytic ozonation has been reported for several metals, especially iron and manganese. The reaction mechanism involves the oxidation of a reduced form of the metal by ozone, hydroxyl radicals or hydrogen peroxide followed by interaction with the organic compounds. Heterogeneous catalytic ozonation is a complex process whose underlying chemistry is not well known. Several mechanisms have been proposed for describing it that can be classiﬁed according to the kind of surface interaction proposed. A Langmuir–Hinshelwood rate expression may account for a reaction between adsorbed organics and oxidized catalytic sites, while an Eley–Rideal model can explain the direct oxidation of an adsorbed organic compound by hydroxyl radicals from the bulk. The ozonation on activated carbon seems to be based on the role of the surface as initiator of the radical chain reactions that transform ozone into radicals. Other mechanisms exclude adsorption equilibrium and lead to models in which the rate of the catalytic process is not dependent on the concentration of oxidant. An adsorption-limited kinetics seems to be more realistic considering the difﬁculty of adsorption encountered by organics in aqueous solutions, especially on the surface of oxides. Depending on the pH of the solution, the surface of an oxide may be charged or not. On a neutral surface, the adsorbate must displace water coordination molecules and at basic conditions, Lewis sites would be inhibited by the hydroxide anion. On oxides, such as titanium dioxide, the reaction is probably better described by an interaction between Lewis acid sites and organic molecules, with an optimum mineralization rate obtained in slightly acidic conditions.

172

A. Rodríguez et al.

The ozonation reaction of individual compounds showed that the ozonation starts with a rapid mineralization period followed by a slow decay of the organic carbon associated with the accumulation of refractory compounds. Some other circumstances complicate the modelling and description of a catalytic ozonation process. For example, the distribution of reaction products markedly differs from that encountered in non-catalytic reactions. Understanding the role of catalyst in the inhibition of the ozone decomposition reaction and the determination of values of the hydroxyl-to-ozone ratio, Rct , a parameter that may change during the reaction and that the catalyst can modify, are additional difﬁculties. Acknowledgements The authors wish to express their gratitude to the Ministry of Education of Spain (Contracts CTM2005-03080/TECNO, CTM2004-06265-C03-03 (EVITA) and CONSOLIDER-INGENIO 2010 CSD2006-00044), the Dirección General de Universidades e Investigación de la Comunidad de Madrid under Contract No. PAMB-000395-0505 and the research network from Comunidad de Madrid REMTAVARES Ref. 0505/AMB-0395.

References 1. Andreozzi R, Caprio V, Ermellino I, Insola A, Tufano V (1996) Ind Eng Chem Res 35:1467–1471 2. Rischbieter E, Stein H, Shumpe A (2000) J Chem Eng Data 45:338–340 3. Roth JA, Sullivan DE (1981) Ind Eng Chem Fundam 20:137–140 4. Sotelo JL, Beltran FJ, Benitez FJ, Beltran-Heredia J (1989) Water Res 23:1239–1246 5. Beltran FJ, García-Araya JF, Encinar JM (1997) Ozone Sci Eng 19:281–296 6. Charpentier JC (1981) Mass-transfer rates in gas–liquid absorbers and reactors. In: Drew TB, Cokelet GR, Hoopes HW Jr, Vermeulen T (eds) Advances in chemical engineering, vol 11. Academic, New York, pp 3–133 7. Danckwerts PV (1970) Gas–liquid reactions. McGraw-Hill, New York, p 113 8. Onda K, Sada E, Kobayashi T, Fujine M (1979) Chem Eng Sci 25:761–768 9. Beltran FJ, Encinar JM, García-Araya JF, Alonso MA (1992) Ozone Sci Eng 14:303– 327 10. Poling BE, Prausnitz JM, O’Connell JP (2001) The properties of gases and liquids. McGraw-Hill, New York, p 25 11. Beltrán FJ (2004) Ozone reaction kinetics for water and wastewater systems. CRC, Boca Raton, pp 70–87 12. Chalmet S, Ruiz-López M (2006) J Chem Phys 124:1945 13. Buhler RE, Staehelin S, Hoigné J (1984) J Phys Chem 88:2560–2564 14. Staehelin S, Buhler RE, Hoigné J (1984) J Phys Chem 88:5999–6004 15. Tomiyasu H, Fukutomi H, Gordon G (1985) Inorg Chem 24:2962–2966 16. Hoigné J (1998) Chemistry of aqueous ozone and transformation of pollutants by ozone and advanced oxidation processes. In: Handbook of environmental chemistry, vol 5, part C. Quality and treatment of drinking water II. Springer, Berlin 17. Staehelin S, Hoigné J (1983) Vom Wasser 61:337–348 18. Kasprzyk-Hordern B, Ziolek M, Nawrocki J (2003) Appl Catal B Environ 46:639–669 19. Lin J, Kawai A, Nakajima T (2002) Appl Catal B Environ 39:157–165 20. Logemann FP, Annee JHJ (1997) Water Sci Technol 35:353–360

Ozone-Based Technologies in Water and Wastewater Treatment

173

21. Sánchez-Polo M, von Gunten U, Rivera-Utrilla J (2005) Water Res 39:3189–3198 22. Rivera-Utrilla J, Sánchez-Polo M, Mondaca MA, Zaror CA (2002) J Chem Technol Biotechnol 77:883–890 23. von Gunten U (2003) Water Res 37:1443–1467 24. Elovitz MS, von Gunten U (1999) Ozone Sci Eng 21:239–260 25. Rosal R, Rodríguez A, Zerhouni M (2006) Appl Catal A Gen 305:169–175 26. Beenackers AACM, van Swaaij WPM (1993) Chem Eng Sci 48:3109–3139 27. Acero JL, von Gunten U (2001) J Am Water Works Assoc 93:99–100 28. Bufﬂe MO, Schumacher J, Meylan S, Jekel M, Gunten U (2006) Ozone Sci Eng 28:245– 249 29. Bufﬂe MO, Schumacher J, Salhi E, Jekel M, Gunten U (2006) Water Res 40:1884–1894 30. Langlais B, Reckhow DA, Brink DR (1991) Ozone in water treatment: application and engineering. AWWA Research Foundation and Lewis, Denver 31. Sonntag C (2007) Water Sci Technol 55:19–23 32. Barreto R, Gray KA, Andres K (1995) Water Res 29:1243–1248 33. Albarran G, Schuler RH (2005) J Phys Chem A 109:9363–9370 34. Palmisano G, Addamo M, Augugliaro V, Caronna T, Paola A, García E, Loddo V, Marci G, Palmisano L, Schiavello M (2007) Catal Today 122:118–127 35. Fang X, Schuchmann HP, Sonntag J (2000) J Chem Soc Perkin Trans 2 1391–1398 36. Legube B, Karpel N (1999) Catal Today 53:61–72 37. Sallanko J, Lakso E, Röpelinen J (2006) Ozone Sci Eng 28:269–273 38. Langlais B, Reckhow DA, Brink DR (eds) (1991) Ozone in drinking water treatment: application and engineering. AWWARF and Lewis, Boca Raton 39. Ramírez F (2005) Tratamiento de desinfección de agua potable. Canal de Isabel II, Madrid, p 9 40. Viraraghavan T, Subramanian KS, Aruldoss JA (1999) Water Sci Technol 40:69–76 41. Nishimura T, Umetsu Y (2001) Hydrometallurgy 62:83–92 42. Kim MJ, Nriagu J (2000) Sci Total Environ 247:71–79 43. von Gunten U (2003) Water Res 37:1469–1487 44. USEPA (1989) National primary drinking water regulations: ﬁnal rules 54:27485– 27541 45. von Gunten U, Elovitz MS, Kaiser HP (1999) J Water Supply Aqua 48:250 46. Do-Quang Z, Roustan M, Duguet JP (2000) Ozone Sci Eng 22:99–111 47. Xu P, Janex ML, Savoye P, Cockx A, Lazarova V (2002) Water Res 36:1043–1055 48. Camel V, Bermond A (1998) Water Res 32:3280–3222 49. Rice RP (1999) Ozone Sci Eng 21:99–118 50. Larocque RL (1999) Ozone Sci Eng 21:119–125 51. Matsumoto N, Watanable K (1999) Ozone Sci Eng 21:127–138 52. Kruithof CJ, Masschelin W (1999) Ozone Sci Eng 21:139–152 53. Paulose J, Langlais B (1999) Ozone Sci Eng 21:153–162 54. Böhme A (1999) Ozone Sci Eng 21:163–176 55. Pollo I (1999) Ozone Sci Eng 21:177–186 56. Geering F (1999) Ozone Sci Eng 21:1187–1200 57. Lowndes R (1999) Ozone Sci Eng 21:201–205 58. Haag WR, Hoigné J (1985) Chemosphere 14:1569–1671 59. Hoigné J (1997) Water Sci Technol 35:1–8 60. Yurteri C, Gurol DM (1988) Ozone Sci Eng 10:277–290 61. Glaze WH, Kang J-W (1989) Ind Eng Chem Res 28:1573–1580 62. Glaze WH, Kang J-W (1989) Ind Eng Chem Res 28:1580–1587 63. Beltran FJ, Rivas J, Alvarez PM, Alonso MA (1999) Ind Eng Chem Res 38:4189–4199

174

A. Rodríguez et al.

64. Legube B (2003) Ozonation by-products. In: Handbook of environmental chemistry, vol 5, part G. Springer, Berlin, pp 95–116 65. Andreozzi R, Insola A, Caprio V, D’Amore MG (1992) Water Res 26:917–925 66. Andreozzi R, Marotta R, Sanchirico R (2000) J Chem Technol Biotechnol 75:59–65 67. Beltrán FJ, Rivas FJ, Montero R (2002) Appl Catal B Environ 39:221–231 68. Beltrán FJ, Rivas FJ, Montero R (2003) J Chem Technol Biotechnol 78:1225–1233 69. Beltrán FJ, Rivas FJ, Montero R (2005) Water Res 39:3553–3564 70. Rivera-Utrilla J, Sánchez-Polo M (2002) Appl Catal B Environ 39:319–329 71. Faria PCC, Órfao JJM, Pereira MFR (2005) Water Res 39:1461–1470 72. Andreozzi R, Casale MS, Marotta R, Pinto G, Pollio A (2000) Water Res 34:4419–4429 73. Gracia R, Cortes S, Sarasa J, Ormad P, Ovelleiro JL (2000) Water Res 34:1525–1532 74. Ma J, Graham NJD (1999) Water Res 33:785–793 75. Lin J, Nakajima T, Jomoto T, Hiraiwa K (1999) Ozone Sci Eng 21:241–247 76. Karpel N, Delouane B, Legube B, Luck F (1999) Ozone Sci Eng 21:261–276 77. Fu H, Karpel N, Legube B (2002) New J Chem 26:1662–1666 78. Hewes CG, Davinson RR (1972) Water AIChE Symp Ser 70:69–71 79. Gracia R, Aragües JL, Cortés S, Ovelleiro JL (1996) Ozone Sci Eng 18:195–208 80. Piera E, Calpe JC, Brillas E, Domenech X, Peral J (2000) Appl Catal B Environ 27:169– 177 81. Nowell LH, Hoigné J (1987) Interaction of iron(II) and other transition metals with aqueous ozone, 8th Ozone World Congress, Zurich, September 1987, p E80 82. Neyens E, Baeyens JH (2003) J Hazard Mater B29:33–50 83. Pines DS, Reckhow DA (2002) Environ Sci Technol 36:4046–4051 84. Dhandapani B, Oyama ST (1997) Appl Catal B Environ 11:129–166 85. Bulanin KM, Lavalley JC, Tsyganenko AA (1995) Colloids Surf A 101:153–158 86. Fernandez P, Nieves FJDL, Malato S (2000) J Colloid Interface Sci 227:510–516 87. Zhaobin W, Xienian G, Sham EL, Grange P, Delmon B (1990) Appl Catal 63:305–317 88. Kasprzyk-Hordern B (2004) Adv Colloid Interface Sci 110:19–48 89. Beltrán FJ, Rivas FJ, Monterio R (2004) Appl Catal B Environ 47:101–109 90. Vannice MA (2007) Catal Today 123:18–22 91. Agustina TE, Ang HM, Vareek VK (2005) J Photochem Photobiol C Photochem Rev 6:264–273 92. Moraes SG, Freire RS, Duran N (2000) Chemosphere 40:369–373 93. Kusic H, Koprivanac N, Bozic AL (2006) Chem Eng J 123:127–137 94. Ciardelli G, Ranieri N (2001) Water Res 35:567–572 95. Li L, Zhu W, Zhang P, Zhang Q, Zhang Z (2006) J Hazard Mater 128:145–149 96. Bijan L, Mohsein M (2005) Water Res 39:3763–3772 97. Amat AM, Arques A, Miranda MA, López F (2005) Chemosphere 60:1111–1117 98. Fernández J, Riu J, García-Calvo E, Rodríguez A, Fernández-Alba AR, Barceló D (2004) Talanta 64:69–79 99. Chiron S, Fernández-Alba AR, Rodríguez A, García-Calvo E (2000) Water Res 34:366–377 100. Hernando MD, Petrovic M, Radjenovic J, Fernández-Alba AR, Barceló D (2007) Removal of pharmaceuticals by advanced treatment technologies. In: Petrovic M, Barceló D (eds) Analysis, fate and removal of pharmaceuticals in the water cycle. Elsevier Science Ltd., Oxford, UK 101. Gómez MJ, Martínez Bueno MJ, Lacorte S, Fernández-Alba AR, Agüera A (2007) Chemosphere 66:993–1002 102. Hernando MD, Mezcua M, Fernández-Alba AR, Barceló D (2006) Talanta 69:334–342 103. Zuccato E, Castiglioni S, Fanelli RE (2005) J Hazard Mater 122:205–209

Ozone-Based Technologies in Water and Wastewater Treatment

175

104. Carballa M, Omil F, Lema JM, Llompart M, García-Jares C, Rodríguez I, Gómez M, Ternes TA (2004) Water Res 38:2918–2926 105. Hernando MD, Ferrer I, Agüera A, Fernández-Alba AR (2004) In: Barcelo D (ed) Emerging organic pollutants in waste waters and sludge. Springer, Berlin, pp 53–77 106. Bendz D, Paxeus NA, Ginn TR, Loge FJ (2005) J Hazard Mater 122:195–204 107. Bound JP, Voulvoulis N (2004) Chemosphere 56:1143–1155

Hdb Env Chem Vol. 5, Part S/2 (2008): 177–197 DOI 10.1007/698_5_094 © Springer-Verlag Berlin Heidelberg Published online: 28 September 2007

Removal of Emerging Contaminants in Waste-water Treatment: Removal by Photo-catalytic Processes Sixto Malato Plataforma Solar de Almería (CIEMAT), Carretera Senés, km 4, 04200 Tabernas, Spain [email protected] 1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

178

2 2.1

Heterogeneous Photo-catalysis (TiO2 ) . . . . . . . . . . . . . . . . . . . . . Fundamental Parameters in Heterogeneous Photo-catalysis . . . . . . . . .

179 182

3 3.1

Homogeneous Photo-catalysis (photo-Fenton) . . . . . . . . . . . . . . . . Fundamental Parameters in Homogeneous Photo-catalysis . . . . . . . . .

184 185

4 4.1 4.2 4.3 4.4

Removal of Emerging Contaminants by Photo-catalytic Processes VOC Case Study: MTBE . . . . . . . . . . . . . . . . . . . . . . . . Antibiotic Case Study: Lincomycin . . . . . . . . . . . . . . . . . . NSAIDs Case Study: Diclofenac . . . . . . . . . . . . . . . . . . . NSAIDs Case Study: Dipyrone or 4-Methylaminoantipyrine . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

187 188 190 191 194

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

196

Abstract Although advanced oxidation processes (AOPs) are well known for their capacity to oxidize and mineralize almost any organic contaminant, commercial applications are still scarce. Future applications of these processes could be improved through the use of catalysis and solar energy. Therefore, their investigation increasingly focuses on the two AOPs that can be powered by solar radiation (i.e. light with a wavelength greater than 300 nm), namely homogeneous catalysis by the photo-Fenton reaction and heterogeneous catalysis by the UV/TiO2 process. This work summarizes the main characteristics of both photo-Fenton and heterogeneous photo-catalysis, noting the main advantages and drawbacks, and focuses mainly on applications related to emerging contaminants. The solar photo-catalytic degradation of these new environmental contaminants, many of which had been unknown until recently, is the focus of a great deal of research. The work also comments on different technological issues that should be taken into account and points out the distinctiveness of each treatment, on a case by case basis: MTBE, lincomycin, diclofenac, and dipyrone. Keywords Lymcomycin · MTBE · NSAIDs · Photo-Fenton · Photo-catalysis Abbreviations AOPS Advanced oxidation processes CPC Compound parabolic collectors DOC Dissolved organic carbon

178 EEC L-H 4-MAA MTBE IPPC NSAID PPCP PS PSA TBA TBF TOC VOC WFD

S. Malato European Economic Community Langmuir Hinshelwood Law 4-methylaminoantipyrine Methyl tert-butyl ether Integrated Pollution Prevention and Control Directive Non-steroidal anti-inﬂammatory drug Pharmaceutical or Personal Care Product Priority substances Plataforma Solar de Almería tert-Butyl alcohol tert-Butyl formate Total organic carbon Volatile organic compound Water Framework Directive

1 Introduction Since the ﬁrst European directive in 1976 (76/464/EEC) [1], much progress has been made in tackling point-source contamination of Europe’s waters. However, a great deal of pressure regarding priority substances (PS) has remained. There are also a number of additional waste-water problems related to non-biodegradable compounds, such as pharmaceutical or personal care products (PPCPs), dyes, or certain agricultural waste waters, which have to be addressed. Moreover, human health is threatened by use of antibiotics, analgesics, pesticides, chlorinated hydrocarbons, etc., dissolved in water at low to medium concentration. In this context, the Integrated Pollution Prevention and Control Directive (IPPC, 96/61/EEC) has requested that technologies and management practices for speciﬁc sectors are to be developed to minimize pollution and to develop water- recycling systems. Also, through the Water Framework Directive (WFD, 2000/60/EC) [2], speciﬁc measures must be adopted at the EC level against pollution of water by individual pollutants or groups of pollutants, presenting a signiﬁcant risk to or via the aquatic environment. Such measures are aimed at the progressive reduction and, for PS, at the cessation of phasing out of discharges, emissions, and losses. As a consequence, simple, low-cost, and at-hand technologies are necessary to treat non-biodegradable waste-water. Due to their versatility, oxidation technologies, including advanced oxidation processes (AOPs), are considered to be an interesting option for solving of this problem. Adsorption technologies, air stripping, and extraction technologies are alternative treatments. Nonetheless, they are only phase-transfer technologies, which do not destroy the contaminant. AOPs are generally deﬁned as oxidation processes, generating hydroxyl radicals, which, in turn, are responsible for organic degradation, due to their

Removal of Contaminants in water by photocatalysis

179

strong oxidative power (Eq. 1). Therefore, they are able to oxidize and mineralize almost every organic molecule, yielding CO2 and inorganic ions. Rate constants for most reactions involving hydroxyl radicals in aqueous solution are usually in the order of 106 to 109 M–1 s–1 [3]. · OH + H++ e– → H O; 2

E0 = 2.33 V .

(1)

Most of the systems classiﬁed as AOPs make use of a combination of the following: (1) two oxidants (O3 /H2 O2 ), (2) a catalyst and an oxidant (Fe+2 /H2 O2 ), (3) an oxidant and irradiation (UV/O3 , UV/H2 O2 ), (4) irradiation and a catalyst (UV/TiO2 ), (5) irradiation, a catalyst, and an oxidant (UV/Fe+2 /H2 O2 ), or (6) an oxidant (H2 O2 ) and ultrasonic irradiation [4]. The common drawback of such systems is the high demand of electrical energy for devices such as ozonators, UV lamps, ultrasound generators, etc., making these treatments economically disadvantageous. As a result, commercial applications are still scarce, although AOPs are well known for their capacity to oxidize and mineralize almost any organic contaminant. Furthermore, future applications of these processes could be improved through the use of catalysis and solar energy. Therefore, investigation has been increasingly focused [5] on the two AOPs that can be powered by solar radiation (i.e. light with a wavelength greater than 300 nm), namely homogeneous catalysis by the photo-Fenton reaction and heterogeneous catalysis by the UV/TiO2 process. Although these processes have been studied for at least two decades (see Fig. 1), industrial/commercial applications, engineering systems, and engineering design methodologies have been developed only recently [6]. Photo-catalysis aims to mineralize contaminants into carbon dioxide, water, and inorganics. However, there is no general rule, and each case is completely different. Consequently, preliminary research is required to assess potential treatments of pollutants, and to optimize solution for any speciﬁc problem, on a case-by-case basis. This work summarizes the main characteristics of both photo-Fenton and heterogeneous photo-catalysis, noting the main advantages and drawbacks, and focuses mainly on applications related to emerging contaminants.

2 Heterogeneous Photo-catalysis (TiO2 ) In heterogeneous photo-catalysis, dispersed solid particles efﬁciently absorb large fractions of the UV spectrum, and they generate chemical oxidants from dissolved oxygen or water in situ. These advantages make heterogeneous photo-catalysis particularly attractive for environmental detoxiﬁcation. The most important features of this process, making it applicable to the treat-

180

S. Malato

Fig. 1 Publications (scientiﬁc journals) in heterogeneous and homogeneous photo-catalysis

ment of contaminated aqueous efﬂuents, are the following: (1) The process takes place at ambient temperature; (2) Oxidation of the substances into CO2 is complete; (3) The oxygen necessary for the reaction is obtained from the atmosphere; (4) The catalyst is inexpensive; (5) Innocuous. It can be reused and it can be attached to different types of inert matrices. Whenever different semi-conducting materials were tested under comparable conditions for the degradation of the same compounds, TiO2 was generally demonstrated to be the most active. TiO2 ’s strong resistance to chemical- and photo-corrosion, its safety and low cost, limit the choice of convenient alternatives. This semi-conductor is of special interest, since it can use natural (solar) UV, because it has an appropriate energetic separation between its valence bands and conduction bands, which can be surpassed by the energy content of a solar photon (λ > 300 nm). Other semi-conductor particles, for example, CdS or GaP, absorb larger fractions of the solar spectrum. They can form chemically activated surface-bond intermediates, however, unfortunately, these photo-catalysts are degraded during the repeated catalytic cycles involved in heterogeneous photo-catalysis. Therefore, degradation of the organic pollutants present in waste-water, using irradiated suspensions of TiO2 , is the most promising process, and research and development in heterogeneous photo-catalysis has grown very quickly over the last few years (see Fig. 1).

Removal of Contaminants in water by photocatalysis

181

The ﬁrst effect, after absorption of near-UV radiation, λ < 390 nm, is the generation of electron-hole pairs (see Fig. 2), which are separated between the conduction bands and valence bands. In order to avoid re-combination of generated pairs, if the solvent is oxido-reductively active (water), it acts both as a donor and acceptor of electrons. Thus, on a hydrated and hydroxylated TiO2 surface, the holes trap · OH radicals linked to the surface. In any case, it should be emphasized that even trapped electrons and holes can rapidly re-combine on the surface of a particle. This can be partially avoided through the capture of the electron by pre-adsorbed molecular oxygen, forming a super-oxide radical, or by other electron scavengers added to water as H2 O2 , S2 O8 2– , etc. [7]. Whatever the formation pathway, it is well known that O2 and water are essential for photo-oxidation with TiO2 . There is no degradation in the absence of either. Furthermore, the oxidative species formed (in particular, the hydroxyl radicals) react with the majority of organic substances. For example, in aromatic compounds, the aromatic part is hydroxylated, and successive steps in oxidation/addition lead to opening of the ring. The resulting aldehydes and carboxylic acids are de-carboxylated,

Fig. 2 Effect of UV radiation on a TiO2 particle dispersed in water, and subsequent destruction of the organics dissolved in water. TiO2 absorption spectrum, compared to solar spectrum, is also shown

182

S. Malato

and they ﬁnally produce CO2 . However, the important issue, governing the efﬁciency of photo-catalytic oxidative degradation, is minimizing electronhole re-combination, by maximizing the rate of inter-facial electron transfer, to capture the photo-generated electron and/or hole [8]. 2.1 Fundamental Parameters in Heterogeneous Photo-catalysis As oxidation proceeds, less and less of the surface of the TiO2 particles is covered, as the contaminant is decomposed. Evidently, at total decomposition, the rate of degradation is zero, and a decreased photo-catalytic rate is to be expected with increasing illumination time. Most authors agree that, with minor variations, the expression for the rate of photo-mineralization of organic substrates with irradiated TiO2 could be described by the LangmuirHinshelwood (L-H) law. According to the L-H model, the rate of reaction (r) is proportional to the fraction of surface covered by the substrate, where kr is the reaction-rate constant, K the reactant adsorption constant, and C the concentration at any time (see Fig. 3). The effect of the initial concentration on the degradation rate is also shown in Fig. 3. Due to the saturation

Fig. 3 Behaviour of reaction-rate as function of initial concentration of contaminant. The inset shows reaction-rate dependence on the intensity of illumination

Removal of Contaminants in water by photocatalysis

183

produced on the surface of semi-conductor, as the concentration of the reactant increases, the rate becomes steady. It should be emphasized that photodecomposition gives rise to intermediates, which could also be competitively adsorbed (Ki Ci ) on the surface of the catalyst. The concentration of these intermediates varies throughout the reaction, up to their mineralization. Thus, i is the number of intermediates formed during degradation. An understanding of the rates of reaction and ways in which they are inﬂuenced by different parameters is important for the design and optimization of treatment systems. The L-H reaction-rate constants are useful for comparing the rate of reaction under different experimental conditions [9]. Since 1990, there has been a clariﬁcation regarding the kind of solar technology, which is to be involved in detoxiﬁcation [10]. It was questioned, if it was necessary to concentrate the radiation for photo-catalysis, and if a non-concentrating collectors could be as efﬁcient as concentrating ones. Initially, it was thought that the concentrating collectors were the ideal alternative. In fact, the ﬁrst large pilot plants operated with them. However, their high cost and the fact that they could only operate with direct solar radiation (this implies their location in highly insulated areas) lead one to consider the alternative of static non-concentrating collectors. It has been experimentally shown that, above a certain ﬂux of UV photons (I), the rate of reaction changes from 1 to < 1 order-dependence on intensity (see Fig. 3). Most authors believe that the transition from r = f (I1.0 ) to r = f (I<1 ) occurs due to the excess of photo-generated species (e– , h+ , and · OH). At very high radiation intensities, r = f (I0 ) = k, as mass-transfer limits the photo-catalytic reaction. So, the rate is constant, although the radiation increases. There may be several reasons which limit the mass-transfer, such as the lack of electrons scavengers (i.e. O2 ), or organic molecules in the proximity of TiO2 surface, and/or excess of products occupying active centres of the catalyst. Whether in static-ﬂow photo-reactors, slurry-ﬂow photo-reactors, or dynamic-ﬂow photo-reactors, the initial rates of reaction were also found to be proportional to catalyst mass. This indicates a truly heterogeneous catalytic regime. However, above a certain value, the rate of reaction levels off and becomes independent of catalyst mass. This limit depends on the geometry and working conditions of the photo-reactor, and it is for a deﬁnite amount of TiO2 , in which all the particles (i.e., all the surface exposed) are totally illuminated. When catalyst concentration is very high, after travelling a certain distance on an optical path, turbidity impedes further penetration of light in the reactor. In any given application, this optimal catalyst mass has to be determined, in order to avoid excess catalyst, and to ensure total absorption of efﬁcient photons. In the literature, there are dozens of studies on the inﬂuence of catalyst concentration on process efﬁciency [11]. The results vary, but, from all of them, it can be deduced that radiation incident on the reactor and path-length inside the reactor are fundamental in determining the optimal catalyst concentration. If the lamp is outside, but the path-length is

184

S. Malato

several centimetres long, similar to a reactor illuminated by solar radiation, the appropriate catalyst concentration is several hundreds of milligrams per litre. In this case, it is very clear that the optimal rate is attained at lower catalyst-concentrations, when the diameter of the photo-reactor is increased. Although TiO2 suspensions absorb photons with wavelengths of less than 390–400 nm, there is also a strong light-scattering effect due to particles. Both effects must be considered in determination of optimal catalyst-load as a function of light path-length in the photo-reactor. Because of photonic activation, heterogeneous photo-catalytic systems do not require heating, and they operate at room temperature. The true activation energy is zero, whereas, in the medium temperature range (20–80 ◦ C),the apparent activation energy is often very low (a few kJ mol–1 ). However, at very low temperatures (–40–0 ◦ C), activity decreases and activation energy becomes positive. The decrease in temperature favours adsorption of reactants, which is a spontaneous exothermic phenomenon. However, it also favours adsorption of the ﬁnal reaction-products, desorption of which tends to be the rate-limiting step. By contrast, at “high” temperatures (> 70–80 ◦ C) for various types of photo-catalytic reactions, the activity decreases and the apparent activation energy becomes negative. When temperature increases above 80 ◦ C, nearing the boiling point of water, the exothermic adsorption of reactants is disfavoured, and this tends to become the rate-limiting step.

3 Homogeneous Photo-catalysis (photo-Fenton) Mixtures of ferrous iron and hydrogen peroxide are called Fenton reagents. If ferrous is replaced by ferric iron, this mixture is called Fenton-like reagent. The Fenton reaction (Eq. 2) was ﬁrst reported by H.J.H. Fenton in 1894. If organic substances (quenchers, scavengers, or water pollutants) are present in the system Fe2+ /Fe3+ /H2 O2 , they react in many ways with the generated hydroxyl radicals. Yet, in all cases, the oxidative attack is electrophilic, and the rate constants are close to the diffusion-controlled limit. The possible reactions with organic substrates are hydrogen abstractions from aliphatic carbon atoms, electrophilic additions to double bonds or aromatic rings, and electron- transfer reactions (Eqs. 3–5). The generated organic radicals continue to react, prolonging the chain- reaction. There is one major setback of the Fenton method. Especially when the treatment goal is total mineralization of organic pollutants, carboxylic intermediates cannot be further degraded. Carboxylic and di-carboxylic acids are known to form stable iron complexes, which inhibit the reaction with peroxide. Hence, the catalytic iron cycle reaches a stand-still before total mineralization is accomplished. Pignatello and co-workers [12] recently presented a rather comprehensive review

Removal of Contaminants in water by photocatalysis

185

of the Fenton process, which is as follows: Fe2+ + H2 O2 → Fe3+ + OH– + OH· OH· + RH → R· + H2 O R – CH = CH2 + OH· → R – C· H – CH2 OH OH· + RX → RX· + + OH– .

(2) (3) (4) (5)

Irradiation with light up to 580 nm leads to photo-reduction of dissolved ferric iron to ferrous iron [13], with the regeneration of Fe2+ being rate-limiting for the overall process. The primary step is a ligand-to-metal charge-transfer reaction. Subsequently, intermediate complexes dissociate. The ligand can be any Lewis base, which is able to form a complex with ferric iron (OH– , H2 O, Cl– , R-COO– , R-OH, R-NH2 , etc.). Depending on the reacting ligand, the product may be a hydroxyl radical or another radical derived from the ligand. The direct oxidation of an organic ligand is possible, and it is shown in Eq. 8 for carboxylic acids. The most important features of this process, making it applicable to the treatment of contaminated aqueous efﬂuents, are that, compared to other oxidants, hydrogen peroxide is inexpensive, safe, easy to handle, and it poses no lasting environmental threat, since it readily decomposes to water and oxygen. In addition, iron is also reasonably priced, safe, and environmentally friendly, because iron is the second most abundant metal. Ferric iron/carboxylic acids complexes can also be photolysed with very high quantum yields and with visible wavelengths. Consequently, concerning the solar spectrum, a far greater share of photons can be employed for the photo-Fenton process, compared to heterogeneous photo-catalysis, which generally leads to far higher quantum yields and rates of reaction. These, of course, also always depend on the amount of iron catalyst employed. Another important advantage of the photo-Fenton process is the inexistence of mass-transfer limitations, due to its homogeneous catalytic nature. As it will be later explained, the most important drawback of the photo-Fenton process, is the necessity for pH-adjustment. [Fe3+ L] + hν → [Fe3+ L]∗ → Fe2+ + L· [Fe(H2 O)]3+ + hν → Fe2+ + OH· + H+ [Fe(OOC – R)]2+ + hν → Fe2+ + CO2 + R· .

(6) (7) (8)

3.1 Fundamental Parameters in Homogeneous Photo-catalysis Depending on the ligand, the ferric iron complex has different lightabsorption properties, and reaction 6 takes place with different quantum yields and at different wavelengths. Consequently, the pH plays a crucial role in the efﬁciency of the photo-Fenton reaction, because it strongly inﬂuences which complexes are formed. When pH increases to more than 2.5 to 3.5,

186

S. Malato

depending on the iron concentration and temperature, the ferric ion precipitates in amorphous ferric oxyhydroxides, forming a red-brown sludge that can co-precipitate organic compounds and produce technological challenges. Thus, pH 2.8 has been postulated as an optimal pH for photo-Fenton treatment, also because at this pH Fe precipitation does not yet take place, and the dominant iron species in solution is [Fe(OH)]2+ , the most photo-active ferric iron–water complex. In fact, ferric iron can form complexes with many substances, and it can undergo photo-reduction. Of special importance are carboxylic acids, because they are frequent intermediate products in an oxidative treatment. Such ferric iron–carboxylate complexes can have much higher quantum yields than ferric iron–water complexes. Therefore, it is typically observed that a reaction shows an initial lag- phase until intermediates are formed, which can more efﬁciently regenerate ferrous iron from ferric iron, accelerating the process. Fenton and photo-Fenton reactions are inhibited by inorganic ions. In particular, ferric ion forms complexes with phosphate, which are quite insoluble in neutral or mildly acidic solution [14]. Moreover, sulphate, chloride, and ﬂuoride inhibit the process, because these ions reduce the reactivity of ferric ion through coordination to form less reactive complexes. Different studies have discussed the inﬂuence of iron concentration, its catalytic behaviour and temperature [15]. Usually, an increase of the respective parameter also meant an increase in the rate of reaction. We are unaware of any explicit studies concerning reaction-rate dependency on radiation intensity. Our own experience indicates that photo-Fenton shows zero order-dependence on the radiation intensity. In general, the rate of the pollutant degradation increases with increasing temperature. When waste-water contains high concentrations of organic compounds, the exothermic Fenton and photo-Fenton processes can deliver enough energy to heat the wastewater, yielding higher rates of reaction at elevated temperatures. Since the rate of reaction of the rate-limiting steps of the Fenton process can be considerably enhanced at elevated temperatures, and since light between 200 and 500 nm increases the rate of the overall process, it is an obvious approach to combine thermal enhancement and photo-chemical enhancement in order to reach optimal reaction conditions. It has been demonstrated several times that optimal conditions for the photo-Fenton process are reached at 50 ◦ C. Fe3+ (related species and organic complexes) absorbs solar photons as a function of its absorptivity. This effect must be considered when determining the optimal load as a function of light path-length in the photo-reactor. However, the main factor affecting cost is H2 O2 consumption, as iron salts are applied only in catalytic concentrations, and acids and bases used for pH adjustment are rather inexpensive. Hydrogen peroxide can be rate-limiting, if applied in very low concentrations (approx. 3–6 mM, or less). This is undesirable, due to its detrimental effect on the rate of reaction. On the other hand,

Removal of Contaminants in water by photocatalysis

187

very high concentrations of hydrogen peroxide can successfully compete with contaminants for the hydroxyl radicals generated by Eqs. 2–7. A pathway for the decomposition of hydrogen peroxide to water and oxygen is induced (Eq. 12), which is the sum of Eqs. 9–11, without any involvement of any organic contaminants present in the waste-water. In other words, hydrogen peroxide, which reacts by this pathway, is wasted. The concentration of hydrogen peroxide, at which this reaction pathway is pre-dominant, depends on the concentration and nature of the organic contaminants, against which hydrogen peroxide competes for the hydroxyl radicals. In this sense, the optimal range of hydrogen peroxide concentration during photo-Fenton treatment depends on the type and degree of waste-water pollution. Within this optimal range, there is no rate-limitation due to a lack of hydrogen peroxide, or any useless hydrogen peroxide consumption due to concentrations that are too high. Fe2+ + H2 O2 → Fe3+ + · OH + OH– Fe3+ + H2 O2 → Fe2+ + · HO2 + H+ · OH + H2 O2 → H2 O + · HO2 2H2 O2 → 2H2 O + O2 .

(9) (10) (11) (12)

4 Removal of Emerging Contaminants by Photo-catalytic Processes Solar photo-catalysis aims to mineralize contaminants into carbon dioxide, water, and inorganics, and treatment of industrial waste-water seems to be one of the most promising ﬁelds of application of photo-catalytic processes. However, each case is completely different. Consequently, preliminary research is required to assess potential pollutant treatments, and to optimize the best option for any speciﬁc problem on a case-by-case basis. The following sections covers different case studies, focused on different examples, each of which have their own distinctiveness (volatilization, pH-dependence, hydrolysis, membrane separation, etc.), which allow evaluation of applicability of the process from different points of view. In photodegradation, the parent organic compound is transformed to eliminate its toxicity and persistence. The oxidation of carbon atoms into CO2 is relatively easy. In general, however, it is notably slower than the de-aromatization of the molecule. Until now, the absence of total mineralization has been observed only in s-triazine herbicides, for which the ﬁnal product obtained was essentially 1,3,5-triazine-2,4,6, trihydroxy (cyanuric acid), which is, fortunately, non-toxic [16]. This is because the triazine nucleus is so highly stable that it resists most methods of oxidation. Cl-ions are easily released into the solution from chlorinated molecules. Nitrogen-containing molecules

188

S. Malato

are mineralized mostly into NO3 – and NH4 + . Ammonium ions are relatively stable, and the proportion depends mainly on the oxidation stage of organic nitrogen and irradiation time [17]. Organo-phosphorous contaminants produce phosphate ions. However, in the pH-range used (usually < 4), phosphate ions remain adsorbed on TiO2 . This strong adsorption somewhat inhibits the rate of reaction, though it is still acceptable. In photoFenton, phosphate sequestrates iron, forming the corresponding non-soluble salt and retarding the rate of reaction. Therefore, more iron is necessary when water-containing phosphates is treated by photo-Fenton. Until now, the analyses of fragments, resulting from the degradation of the aromatic ring, have revealed formation of aliphatics (organic acids and other hydroxylated compounds), which explains why total mineralization takes much longer than de-aromatization, as mineralization of aliphatics is the slowest step. Special attention has recently been given the so-called “emerging contaminants”, mostly unregulated compounds, which may be candidates for future regulation, depending on research on their potential effects on health and monitoring data concerning their occurrence [18]. Particularly relevant examples of such emerging compounds are those which do not need to persist in the environment to cause a negative effect, because their high rates of transformation/removal can be compensated by their continuous introduction into the environment [19]. The solar photo-catalytic degradation of these new environmental contaminants, many of which have been unknown until recently, is the focus of much research. The following sections summarizes them and remarks on different technological issues, and distinctiveness of each treatment, which should be taken into account on a case-by-case basis. 4.1 VOC Case Study: MTBE Methyl tert-butyl ether (MTBE) has been used as a gasoline additive since the 1970s, as an octane-enhancer, producing cleaner burning gasoline. Although its use has been promoted because of its low cost, simple production and characteristics that favour blending with conventional gasoline, its stability, high water-solubility, and volatility, as well as its scarcely studied potential effect on human health and aquatic organisms, make it likely to become an important environmental contaminant. In effect, presence of MTBE has been reported in the aquatic [20] environment of regions in the US and Europe, at concentrations that range from environmental background levels to high mg L–1 levels at sites affected by occasional sources. Several studies involving the degradation of MTBE in aqueous solution, using AOPs, have been conducted [20]. See also references included in [21]. Because MTBE is highly volatile (33 kPa at 298 K), loss of MTBE from volatilization may be expected during the treatment. To quantify these possible losses, simultaneous blank

Removal of Contaminants in water by photocatalysis

189

experiments had to be performed, along with the assays of photo-catalysis, in a second identical photo-reactor. For example, photo-catalytic treatment tests performed with TiO2 in a solar pilot plant revealed that MTBE volatilization losses were as important as photo-catalytic degradation, making its application unfeasible. In the case of photo-Fenton, 35 min was already enough to almost completely degrade 100 mg L–1 MTBE. Therefore, volatilization might have been considered irrelevant. However, TOC values measured were still high (41 mg L–1 ), pointing out the presence of a considerable amount of byproduct (see Fig. 4) generated during the process, and attaining more than 97% mineralization (1.6 mg L–1 ) in approximately 155 min of treatment. The photo-Fenton rate allows photo-catalytic degradation of MTBE without signiﬁcant loss from volatilization.

Fig. 4 Pathway for degradation of MTBE in water, during photo-catalytic treatment by photo-Fenton (Fe 0.05 mM) reaction

190

S. Malato

4.2 Antibiotic Case Study: Lincomycin The antibiotics and their degradation products are present as polluting agents in many waste-waters. A common antibiotic, often present in waste-water, is the lincomycin [22]. Puriﬁcation processes, using methodologies that combine photo-catalytic reactions with membrane-separation processes, taking advantage of the synergy of both technologies, have been found to be very interesting [23]. Indeed, the membrane would play both the role of simple barrier for the photo-catalyst and the role of selective barrier for the molecules to be degraded, thus creating a very powerful system. With respect to “traditional batch” photo-reactors, another advantage of membrane photo-reactors is the possibility to make a continuous process with simultaneous product(s) separation from the reaction environment. Lincomycin treatment has been used as model pollutant for such a study [24], using both membrane rejection and photo-degradation tests in a continuous conﬁguration (see Fig. 5). The photoreactivity experiments were carried out using compound parabolic collectors (CPC), installed at the “Plataforma Solar of Almería” (PSA, Spain). The reactor conﬁguration is a common one in heterogeneous photo-catalysis: a plug- ﬂow

Fig. 5 Scheme of the photo-reacting continuous system treating lincomycin (LM): (T) non-reacting tank; (P) pump; (VC) regulation valve; (PI) manometer; (weighted line) line under pressure; (FI) rota-meter. Lincomycin concentrations versus irradiation time, for runs carried out in continuous regimen, using the hybrid system at different initial concentrations of lincomycin. 10 microM: Retentate, •; permeate, ◦. 25 microM: Retentate, ; permeate,

Removal of Contaminants in water by photocatalysis

191

photo-reactor in a total re-circulation loop with a non-reacting tank, whose function is to provide aeration and samples for analyses. Figure 5 shows the hybrid system. Each photo-reactor consisted of three CPC modules in series, placed on ﬁxed supports, inclined 37◦ (latitude of the PSA) with respect to the horizontal plane, and facing south, in order to maximize the daily absorption of solar radiation. Detailed description of the solar system could be found in other sources [25]. The membrane module is equipped with a multi-stage centrifugal pump, and different commercial nano-ﬁltration membrane modules were used at an operative pressure of 4 bar. The rejection measurements were carried out using the following procedure. By operating the hybrid system at total recycle in the dark, samples of the substrate solution were withdrawn both from the retentate and the permeate. When steady-state conditions were reached (i.e. the values of permeate and retentate concentrations reached constant values), the photo-catalyst was added, and the solar collector cover was removed. Samples of the suspension were withdrawn at the starting of irradiation and at ﬁxed time intervals. The hybrid system allowed the separation of the photo-catalyst particles contemporaneously to lincomycin and its degradation products from the permeate ﬂow. The analyses of the data collected using the hybrid system in continuous regime indicated that the presence of the membranes allows reduction of both, the substrate and intermediates, down to very low concentration levels. Rejections of nano-ﬁltration membrane, obtained during operation of the membrane photo-reactor in the degradation of lincomycin, were signiﬁcantly lower than those obtained in the absence of photo-degradation, probably because of the small molecular size of byproducts and intermediate species, generated during the photo-degradation process. This means that in order to select a suitable membrane, rejection should be determined during operation of the photo-reactor. The pressure in the membrane cell, the pH of the polluted water, the molecular size of the pollutants, and, mainly, the photo-generated by-products and intermediate species can inﬂuence the permeate ﬂux of the membrane, and, consequently, its choice. The experience indicates that the choice of a suitable membrane is essential for application of the photo-catalytic membrane process to the treatment of real efﬂuents [26–29]. 4.3 NSAIDs Case Study: Diclofenac Among the groups of pharmaceutical compounds of greatest environmental interest are the non-steroidal anti-inﬂammatory drugs (NSAIDs). Concentrations of NSAIDs, in the range of several hundred ng/L, have been found in European rivers [30, 31]. Diclofenac is a commonly used analgesic, antiarthritic, and antirheumatic NSAID. Although it has been proven that diclofenac is rapidly degraded by direct photolysis, under normal environmental conditions [32, 33], it is still one of the most frequently detected compounds in

192

S. Malato

the water cycle. This suggests a continuous, signiﬁcant input of diclofenac into the environment, which can be explained, in part, by the inefﬁciency of typical biological treatments in removal of this compound. Although the ecotoxicity of diclofenac is relatively low, and acute effects are rather improbable at the concentration levels present in the environment (around 1000 times lower than the effective concentrations), it has been demonstrated that, in combination with other pharmaceuticals present in water samples, the toxic effect can be considerably increased, even at concentrations at which the substances alone showed either no effect at all, or only a very slight one [34]. Diclofenac is a very soluble compound in neutral-alkaline medium (50 g L–1 ) and it is an acidic pharmaceutical (pKa = 4.15) that becomes almost insoluble below pH 4. Thus, below this pH value, diclofenac precipitates. This ﬁnding is very important, because treating it using any AOP as an acid medium is very easily achieved during diclofenac decomposition due to chloride release. Results could be inconsistent, if “precipitation” were not clearly stated and confused with “decomposition”. Photo-Fenton treatment is usually performed in an acid medium, to maintain the iron in solution (it is also well known that ferric iron precipitates at neutral pH). In acid medium, diclofenac becomes practically insoluble. Consequently, oxidizing species (e.g., hydroxyl radicals) have to diffuse into the pollutant precipitate (a heterogeneous process), or react with the small, dissolved share of the pollutant (a homogeneous process with pollutant solubility equilibrium as an inﬂuencing factor). DOC mineralization, diclofenac decomposition, chloride evolution, and ammonium evolution are shown for an experiment with a 0.075 mM iron concentration, in Fig. 6, in which DOC and diclofenac are already seen to decrease after the addition of acid and iron, but before the addition of hydrogen peroxide, indicating the precipitation of diclofenac (section A, in Fig. 6). During the experiment, the following unusual behaviour is provoked by continuous decomposition, precipitation, and re-dissolution processes (section B, in Fig. 6). However, the evolution of inorganic ions can give an indication of the oxidation process, as these ions are soluble, and they can be measured. Chloride release from the aromatic ring is often one of the ﬁrst steps in the oxidation of aromatic compounds, taking place in treatment with AOPs. Evolution of the theoretical amount of chloride, to be released from the known initial diclofenac concentration, indicates that the original compound is fully degraded. The mineralization of the organic nitrogen in the diclofenac molecule, which joins the two aromatic rings, is somewhat slower than chloride evolution (see Fig. 6). Treatment in neutral and acid media were shown to be disadvantageous to iron (neutral media) and diclofenac precipitation (acid media). Therefore, an alternative could be to perform photo-Fenton treatment, starting at natural pH of water (around pH 7), so that the resulting pH would change from 7 to acid during treatment, due to the formation of intermediate carboxylic acids and de-chlorination. This means that diclofenac is more soluble, because the resulting pH is not as low as optimal photo-Fenton pH (2.8), and iron pre-

Removal of Contaminants in water by photocatalysis

193

Fig. 6 Photo-Fenton treatment of diclofenac at pH 2.8 with 0.075 mM iron in standard fresh-water. Diclofenac structure is also shown

cipitation should be slowed down, possibly yielding an overall accelerated decomposition. The results of this experiment are depicted in Fig. 7. In this experiment, diclofenac decomposition and chloride evolution almost coincide, indicating that diclofenac was dissolved during the overall treatment. The DOC-increase after several minutes is due to soluble intermediates that are formed and mineralize as the reaction further proceeds. Diclofenac decomposition takes around 50 min, and this is considerably faster than photo-Fenton performed at pH = 2.8. The solubility equilibrium, dependent on the pH of diclofenac (it is a weak acid, soluble in its de-protonized form, but practically insoluble when protonized) and iron, were shown to be determining for application of solardriven iron/hydrogen peroxide-catalyzed decomposition of diclofenac. Results of photo-Fenton treatment at pH 2.8 indicated that diclofenac decomposition takes place in the homogeneous phase, and the kinetics of this process are governed by the continuous re-dissolution of diclofenac. PhotoFenton treatment, starting at a pH of around 7, caused a decrease of pH during the experiment, and diclofenac and iron precipitation could be partly overcome [35]. Therefore, although diclofenac and the photo-Fenton process initially seem to be incompatible due to the compound’s insolubility at low pH, the rate of reaction is still signiﬁcantly higher in the photo-Fenton pro-

194

S. Malato

Fig. 7 Photo-Fenton treatment of diclofenac at 0.075 mM iron concentration in standard fresh-water

cess at higher pH. From the results obtained, pH-control around pH 4.5–5 seems promising. At the same time, as proposed by several authors, addition of iron-complexing substances could also enhance the process at elevated pH to prevent iron-precipitation [36]. 4.4 NSAIDs Case Study: Dipyrone or 4-Methylaminoantipyrine One of the most popular analgesic, antipyretic drugs is Dipyrone, also known as Metamizole, or Novalgin. Despite its potential side effects, Dipyrone is widely used in both pediatric and adult patients, because it is a strong analgesic, it is inexpensive, and it does not require a prescription. Dipyrone is a pro-drug which, after oral intake, is spontaneously hydrolysed into its main metabolite, 4-methylaminoantipyrine (4-MAA), and afterwards into a variety of compounds by enzymatic reactions. These metabolites are not biodegradable and, although little is known about their behaviour and persistence in the environment, they have already been detected in efﬂuents and surface-water at signiﬁcantly high concentrations [37]. Dipyrone is a special case for being treated by AOPs, as it readily hydrolyses to 4-MAA, after dissolving in water for a few minutes. Once Dipyrone is completely transformed, MAA remains stable for a longer period of time. Figure 8 shows its photo-Fenton degrada-

Removal of Contaminants in water by photocatalysis

195

tion, using only 2 mg L–1 iron concentration. This low concentration of iron is enough for complete disappearance of MAA during the dark Fenton reaction, within 15 min after hydrogen peroxide is added. At this point TOC decreases by around 10%. Under illumination, the decrease in TOC takes place mainly in two stages, dropping rapidly, from 20 to 5 mg L–1 (75% mineralization) in the ﬁrst 30 min, and then, very slowly, from 5 to 2.5 mg L–1 (12.5% mineralization) in 160 min. The ﬁrst step could be explained by the mineralization of the aromatic ring and the three methyl moieties (responsible for 75% of MAA TOC). The slow second stage is related to the opening and mineralization of the pyrazole ring. This could be explained by the formation of carboxylic acids, described as re-calcitrant to the photo-Fenton process [12]. Carboxylic and di-carboxylic acids are known to form stable iron complexes, which inhibit the reaction with peroxide. The main carboxylic acids found are acetate, formiate, and oxalate, at maximum concentrations of 13, 6, and 5 mg L 1, respectively. These compounds are present throughout all degradation, and at a signiﬁcant concentration, at the end of the treatment. The discussion of the degradation pathway, including additional information about this point can be found in other sources [38]. Hydrogen peroxide was measured and its consumption determined throughout MAA (from dipyrone) degradation, as shown in Fig. 8. It can

Fig. 8 Photo-Fenton MAA (from dipyrone) degradation kinetics (2 mg L–1 Fe2+ ) and hydrogen peroxide consumption

196

S. Malato

clearly be seen that the amount of degradation is strongly correlated to the amount of hydrogen peroxide consumed. The theoretical hydrogen peroxide consumption for complete mineralization of 50 mg L–1 MAA is 8.8 mmol L–1 (300 mg L–1 ), according to Eq. 13. It is estimated that 80% of the hydrogen peroxide is required to degrade the last 20% of TOC. This means that the intermediates formed at the end of the treatment are more resistant, and it is uneconomical (due to the heavy hydrogen peroxide consumption, under these conditions) to completely mineralize MAA by photo-Fenton. As mentioned above, the last compounds to form are also the most re-calcitrant to OH radical oxidation. This reinforces the current trend in AOP applications of evaluating the toxicity and biodegradability of partially mineralized contaminants, in order to reduce treatment time and chemical consumption. C12 H16 N3 O + 38.5H2 O2 → 12CO2 + 3HNO3 + 45H2 O .

(13)

Toxicity (by Vibrio ﬁscheri) was assessed in samples collected only after illuminating the photo-reactor. All the samples evaluated during the photoFenton treatment were below the 50% bioluminescent bacteria inhibition threshold. These results demonstrate that in photo-Fenton treatment, solution toxicity does not increase due to the degradation products formed, as the treatment is sufﬁcient to produce an efﬂuent within safe toxicity limits. Samples analysed during the ﬁrst stages of the treatment, showed a slight increase in toxicity, when the main degradation products generated were present at their higher concentration. Indeed, when only MAA was present at t = 0, toxicity was already at the 50% threshold. Therefore, it may be concluded that MAA is a matter of concern, when discharged into the environment. Photo-Fenton at very low iron concentration (Fe = 2 mg L–1 ) is an efﬁcient method of treating these types of compounds with a pyrazole ring moiety. Using such a low iron concentration demonstrates that solar photo-Fenton provides a way to polish off the efﬂuent of sewage treatment plants, without removing iron after the treatment. Toxicity assessment has demonstrated that MAA, the DP of major environmental concern, is degraded by both solar AOPs, allowing safe disposal of the efﬂuent. Acknowledgements The author acknowledges the Spanish Ministry of Education and Science (Programa Consolider Ingenio 2010 CE-CSD2006-004) and the European Commission (INNOWATECH project, contract no. 036882, FP6-2005-Global-4, SUSTDEV-20053.II.3.2).

References 1. Directive 76/464/EEC of 4 May 1976 on pollution caused by certain dangerous substances discharged into the aquatic environment of the Community 2. Directive 2000/60/EC of the European Parliament and of the Council establishing a framework for the Community action in the ﬁeld of water policy

Removal of Contaminants in water by photocatalysis

197

3. Haag WR, Yao CD (1992) Environ Sci Technol 26:1005 4. Pera-Titus M, García-Molina V, Baños MA, Giménez J, Esplugas S (2004) Appl Catal B-Environ 47:219 5. Malato S, Blanco J, Alarcón DC, Maldonado MI, Fernández-Ibáñez P, Gernjak W (2007) Catal Today 122:137 6. Bahnemann D (2004) Solar Energ 77:445 7. Malato S, Blanco J, Richter C, Braun B, Maldonado MI (1998) Appl Catal B-Environ 17:347 8. Herrmann JM (2005) Top Catal 14:48 9. Minero C, Pelizzetti E, Malato S, Blanco J (1996) Solar Energ 56:421 10. Blanco-Galvez J, Fernández-Ibáñez P, Malato-Rodríguez S (2007) J Sol Energy Eng 129:4 11. Cassano AE, Alfano OE (2000) Catal Today 58:167 12. Pignatello JJ, Oliveros E, MacKay A (2006) Crit Rev Environ Sci Technol 36:1 13. Bauer R, Waldner G, Fallmann H, Hager S, Klare M, Krutzler T, Malato S, Maletzky P (1999) Catal Today 53:131 14. De Laat J, Truong Le G, Legube B (2004) Chemosphere 55:715 15. Sagawe G, Lehnard A, Lubber M, Rochendorf G, Bahnemann D (2001) Helv Chim Acta 84:3742 16. Hincapié M, Maldonado MI, Oller I, Gernjak W, Sánchez JA, Ballesteros MM, Malato S (2005) Catal Today 101:203 17. Calza P, Pelizzetti E, Minero C (2005) J Appl Electrochem 35:665 18. Petrovicì M, Gonzalez S, Barceloì D (2003) TRAC-Trend Anal Chem 22:685 19. Andreozzi R, Raffaele M, Nicklas P (2003) Chemosphere 50:1319 20. Deeb RA, Chu KH, Shih T, Linder S, Suffet I, Kavanaugh MC, Alvarez-Cohen L (2003) Environ Eng Sci 20:433 21. Agüera A, Mezcua M, Hernando D, Malato S, Caìceres J, Fernaìndez-Alba A (2004) Int J Environ Anal Chem 84:149 22. Carucci A, Cappai G, Piredda M (2006) J Environ Sci Health A 41:1831 23. Augugliaro V, Litter M, Palmisano L, Soria J (2006) J Photochem Photobiol C 7:127 24. Augugliaro V, Garcia-Lopez E, Loddo V, Malato-Rodriguez S, Maldonado I, Marcì G, Molinari R, Palmisano L (2005) Solar Energ 79:402 25. Kositzi M, Poulios I, Malato S, Cáceres J, Campos A (2004) Water Res 38:1147 26. Molinari R, Borgese M, Drioli E, Palmisano L, Schiavello M (2002) Catal Today 75:77 27. Chin SS, Lim TM, Chiang K, Fane AG (2007) Chem Eng J 130:53 28. Azrague K, Aimar P, Benoit-Marquieì F, Maurette MT (2007) Appl Catal B-Environ 72:197 29. Le-Clech P, Lee EK, Chen V (2006) Water Res 40:323 30. Barceloì D (2007) TRAC-Trend Anal Chem 26:454 31. Petrovicì M, Hernando MD, Diìaz-Cruz MS, Barceloì D (2005) J Chromatogr A 1067:1 32. Tixier C, Singer HP, Oellers S, Müller SR (2003) Environ Sci Technol 37:1061 33. Agüera A, Pérez-Estrada LA, Fernández-Alba AR, Malato S, Ferrer I, Thurman EM (2005) J Mass Spectrom 40:908 34. Cleuvers M (2004) Ecotox Environ Safe 59:309 35. Perez-Estrada LA, Maldonado MI, Gernjak W, Agüera A, Fernández-Alba AR, Ballesteros MM, Malato S (2005) Catal Today 101:219 36. Paciolla MD, Kolla S, Jansen SA (2002) Adv Environ Res 7:169 37. Gomez MJ, Martínez Bueno MJ, Lacorte S, Fernández-Alba A, Agüera A (2007) Chemosphere 66:993 38. Pérez-Estrada LA, Malato S, Agüera A, Fernández-Alba AR. Catal Today (in press). doi: 10.1016/j.cattod.2007.08.008

Hdb Env Chem Vol. 5, Part S/2 (2008): 199–217 DOI 10.1007/698_5_096 © Springer-Verlag Berlin Heidelberg Published online: 20 October 2007

Behavior of Emerging Pollutants in Constructed Wetlands Víctor Matamoros · Josep M. Bayona (u) Environmental Chemistry Department, IIQAB-CSIC, 08034 Barcelona, Spain [email protected] 1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

200

2

Types of Constructed Wetlands . . . . . . . . . . . . . . . . . . . . . . . .

201

3 3.1 3.2 3.3 3.4 3.5 3.6

Behavior of Emerging Contaminants in CWs . . . Non-steroidal Anti-inﬂammatory Drugs (NSAIDs) Lipid Regulator Drugs and Anti-epileptic Agents . Fragrances . . . . . . . . . . . . . . . . . . . . . . Estrogens . . . . . . . . . . . . . . . . . . . . . . . Surfactants . . . . . . . . . . . . . . . . . . . . . . Other Emerging Compounds . . . . . . . . . . . .

. . . . . . .

203 204 207 209 210 212 213

4

Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

214

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

214

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

Abstract Constructed wetlands (CWs) constitute a cost-effective wastewater treatment alternative for small communities due to the low operational cost, reduced energy consumption, and reduced sewage sludge production. Although much information is available about conventional water quality parameters in CWs, few data exist regarding emerging pollutants. In this chapter, following a short introduction on the different wetland conﬁgurations, the removal efﬁciency for anti-inﬂammatory drugs, lipid regulators, anti-epileptic agents, fragrance materials, surfactants, and estrogens in different wetland systems is discussed. Among the parameters affecting wetland performance, it is shown that removal efﬁciency of a variety of emerging pollutants is dependent on the oxygen availability and sorption interactions. For that reason the vertical ﬂow constructed wetlands exhibited the best performance in terms of hydraulic residence time and removal efﬁciency. Keywords Anionic surfactants · Constructed wetlands · Estrogens · Hydraulic loading rate · NSAID · PPCPs · Removal efﬁciency Abbreviations APE Alkylphenol ethoxylates APEC Alkylphenol polyethoxycarboxylate CAPEC Carboxylic acid alkylphenol ethoxylate CW Constructed wetland BOD5 Biological oxygen demand FM Fragrance material HFCW Horizontal ﬂow constructed wetland

200 HRT HLR MLR NP NPEO NSAID PE PPCPs SFCW SSFCW TSS VFCW WWTP

V. Matamoros · J.M. Bayona Hydraulic residence time Hydraulic loading rate Mass loading rate Nonylphenol Nonylphenol ethoxylate Non-steroidal anti-inﬂammatory drug Person equivalent Pharmaceuticals and personal care products Surface ﬂow constructed wetland Subsurface ﬂow constructed wetland Total suspended solids Vertical ﬂow constructed wetland Wastewater treatment plant

1 Introduction Constructed wetlands (CWs) are land-based wastewater treatment systems that consist of shallow ponds, beds, or trenches that contain ﬂoating or emergent-rooted wetland vegetation [1]. CWs have been used in order to treat domestic wastewater from rural areas all over the world since they were ﬁrstly applied in Germany in the 1960s [2]. The potential of CWs for the removal of contaminants occurring in urban wastewater has attracted increasing interest over the past decade, with a view to treating wastewaters from small populations to comply with environmental regulations such as the European Union Directive 91/271 and the US EPA Clean Water Act. Other wetland applications are the attenuation of agricultural contamination runoff to surface waters [3, 4], combined sewer overﬂows [5], urban storm water [6], industrial efﬂuents [7–9], landﬁll leachates [10], and sludge consolidation [11]. In addition, such treatment systems are well suited to treat wastewater from isolated populations because they do not require external energy, there is no need for highly qualiﬁed personnel, the production of sewage sludge is low, and because of the low operational and maintenance costs. Nevertheless, available information on the removal performance of these systems is limited to common contamination parameters, such as total suspended solids (TSS), biological oxygen demand (BOD5 ), nutrients, bacteria, metals, herbicides, and pesticides [2–4, 12]. Nonetheless, limited information is available related to emerging pollutants, so that will be reviewed hereafter. This chapter will focus on the use of CWs for urban wastewater treatment according to different wetland conﬁgurations, and their performance is discussed under diverse operational conditions. Following the experience gathered in different case studies, a positive correlation has been found between the dissolved oxygen concentration in the CW efﬂuent and the removal

Behavior of Emerging Pollutants in Constructed Wetlands

201

of emerging pollutants. Moreover, the CW performance for the emerging pollutant selected will be compared with conventional wastewater treatment plants (WWTPs).

2 Types of Constructed Wetlands The two principal classes of CWs designed for treating wastewater are the surface and subsurface ﬂow systems [13, 14], depending on whether or not the wastewater is ﬂowing on the wetland surface. Surface ﬂow constructed wetlands (SFCWs) consist of shallow basins or channels with planted wetland vegetation where water ﬂows over a compacted low permeability clay liner at relatively shallow depths. These wetlands are typically used to polish efﬂuents from secondary treatment processes and require large surface areas associated with long retention times (i.e., several weeks) (Fig. 1a). Subsurface ﬂow constructed wetlands (SSFCWs) involve shallow basins or channels with planted vegetation overlying a liner where the wastewater is treated as it ﬂows through the gravel media and around the roots and rhizomes of planted vegetation. Attending to the water ﬂow direction, SSFCWs can be classiﬁed as horizontal ﬂow constructed wetlands (HFCW) (Fig. 1b) or vertical ﬂow constructed wetlands (VFCW) (Fig. 1c). If the plant is properly designed, wastewater ﬂows by gravity, so no external energy is needed. According to the oxygen availability, the biodegradation of organic matter in CW occurs through different pathways. In this regard, whereas in the HFCW the organic matter removal is mostly by anaerobic pathways (i.e., denitriﬁcation, sulfate reduction, and methanogenesis), in VFCW the aerobic environment prevails. These different environments, according to the CW conﬁguration, are of primary interest in order to eliminate the emerging pollutants from wastewater because of their high oxygen dependence [15, 16]. Hence, these wetlands are typically used to treat primary efﬂuents (following a sedimentation step) to reach typical secondary treatment standards. In addition, a few European countries have incorporated willow systems [17] to treat wastewater. The main feature of these treatment systems is the absence of efﬂuent (zero-discharge of water). It is accomplished by water evapotranspiration and removal of the organic matter by biodegradation, plant uptake, or/and mineralization. Different types of plants are used in CWs depending on their conﬁguration. In SFCW, the vegetation can be emerged or submerged and ﬁxed or free-ﬂoating. In case of SSFCWs, the common reed (Phagmitis australis) and narrowleaf cattail (Typha angustifolia) are the most widely used since they are resistant under a variety of environmental conditions and facilitate air transport by the stem. Accordingly, a localized oxic environment in the rhi-

202

V. Matamoros · J.M. Bayona

Fig. 1 Longitudinal section of constructed wetlands with three different conﬁgurations: a SFCW, b HFCW, and c VFCW. Arrows indicate the water ﬂow direction; 1 inﬂow; 2 outﬂow; 3 aeration tip

zosphere is generated, which facilitates the biodegradation of contaminants in such oxic zones. The main difference between CWs and trickling ﬁlters or sand ﬁlters, extensively used in wastewater treatment, is the presence of vegetation. The role of plants in CWs is multiple, from biological to physical aspects (i.e., bioﬁlm support, plant uptake, rhizosphere oxygenation, and enhanced ﬁltration) [18]. One of the problems associated with SSFCWs is clogging. It is characterized by a gradual loss in the hydraulic conductivity according to the operational time. It is associated with the accumulation of settled solids in the wetland inlet. Accordingly, clogging induces several processes, which lead to a reduction in the inﬁltration capacity at the substrate surface. The lower inﬁltration rate causes a reduced oxygen supply and as a consequence leads to

Behavior of Emerging Pollutants in Constructed Wetlands

203

Fig. 2 Types of CWs according to the water ﬂow. Main removal pathways for organic matter in each system

a rapid failure of the treatment performance [19]. Causes of substrate clogging include accumulation of suspended solids, large bioﬁlm development, chemical precipitation and deposition in the pores, growth of plant-rhizomes and roots, generation of gas, and compaction of the clogging layer [20]. The removal of pollutants from wastewater is related to a variety of physical, chemical, and biological processes, which depend on the wetland conﬁguration (Fig. 2). The main removal processes of organic contaminants are due to the physical interactions with organic matter, biodegradation, microbial and plant uptake, volatilization, and photodegradation. Obviously, the latter removal process only takes place in surface ﬂow beds where water is directly exposed to sunlight. Although the use of CWs for wastewater has been postulated as a feasible technology, their high area requirements, ca. 3.2 m2 person–1 equivalent (PE) in VFCW [21] and 5 m2 PE–1 in HFCW [22] have limited their use to small communities with less than 2000 inhabitants. Nevertheless, CWs possess several others functions in addition to improve water quality. In fact, they can also function as landscape restoration, natural habitats, recreational areas, hydrological buffers, or as a reservoir. So, this technology is of increasing interest all over the world. In Europe alone, more than 5000 SSFCWs systems are in operation [13].

3 Behavior of Emerging Contaminants in CWs In the following sections, the behavior of the different PPCP categories are reviewed according to the different CW conﬁgurations and their performance compared with conventional WWTPs.

204

V. Matamoros · J.M. Bayona

3.1 Non-steroidal Anti-inflammatory Drugs (NSAIDs) NSAIDs are the pharmaceuticals with the highest widespread human use. They and their metabolites exhibit a high frequency of detection in inﬂuent and efﬂuent from urban WWTPs [23]. Consequently, they can occur at high concentrations, exceeding in some cases the predicted no-effect concentration (PNEC) [24, 25]. Hence, their occurrence attracted the attention of the scientiﬁc community and legislators. As shown in Fig. 3, these compounds are negatively charged at environmental pH (nearly 8) due their low pKa (between 3 and 4), consequently their sorption into sludge was found to be negligible [26]. Then, their removal in CWs as well as in WWTPs is mainly attributable to biodegradation [19, 27].

Fig. 3 Chemical structures of some NSAIDs. Ionized forms are shown. log Dow calculated at pH 8 [64, 73]

Table 1 summarizes the removal efﬁciencies of some analgesic drugs evaluated in different CWs case studies and compares them to those reported for WWTPs. As shown, CWs could be a suitable alternative for elimination of these compounds from wastewater or for improving the quality of WWTP efﬂuents. VFCW and SFCW appear to be the CW conﬁgurations that allows the highest NSAIDs removal, even higher than conventional WWTP. Nevertheless, when hydraulic residence times (HRTs) were taken into account, VFCW appears to be the most efﬁcient system. Indeed, NSAID removal in the SFCW with low HRT (48–96 h) is lower than in the others (Matamoros et al., unpublished results). Following these case studies, factors affecting NSAID removal in CW are listed and discussed (i.e., oxygen, photodegradation, water depth, loading rates, clogging and plant occurrence). As reported earlier for bioﬁlm reactors [28], oxygen is a key factor affecting the removal of some NSAIDs. Therefore, the removal in VFCW is more efﬁcient because different pathways are involved. In vertical beds with unsaturated ﬂow, aerobic biodegradation pathways prevail whereas in saturated HFCWs anaerobic biodegradation is predominant (Matamoros et al., unpublished results).

Behavior of Emerging Pollutants in Constructed Wetlands

205

Table 1 Removal efﬁciency (%) of NSAIDs in CWs and comparison with WWTPs HFCW a b

VFCW SFCW c d

e

f

95

– 47

Salicylic acid Ibuprofen

96 71

87 34

98 99

– 96

OH-ibuprofen CA-ibuprofen Naproxen

62 87 85

26 49 24

99 99 89

– – 92

52

33 – –

Diclofenac

15

6

73

96

73

–

Ketoprofen HRT (h)

38 114

n.r. 155

– 6

99

97 720

– 48–96

WWTP

Refs.

99 60–70 90 95 95 40–55 66 24 17 48–69 12–24

[26] [67] [23] [68] [68] [67] [23] [69] [70] [23]

HFCW case study was carried out in Spain on primary efﬂuent at different water depths: a 0.3 m; b 0.5 m [19]; c VFCW case study was carried out in Denmark on primary efﬂuent SFCW case studies were carried out in Spain on secondary efﬂuent in different seasons: d warm season (high sun light); e cold season (low sun radiation) [31], and f in the EUU in river efﬂuent [40]; WWTP measurements were made in raw wastewater; n.r. no removal

Additionally, water depth has been noted as an important design parameter for HFCWs. When a normally designed bed (0.5 m water depth) was compared with a shallower bed (0.3 m) a difference in the removal of NSAIDs was shown (Table 1). These differences were attributable to a less negative redox potential of the shallower bed compared to the deeper one [29]. Therefore, whereas in the deep bed the anaerobic pathways are predominant, in the shallow one they coexist with some aerobic pathways (i.e., nitriﬁcation). Consequently, the redox status seems to be a key factor for NSAID removal. The mass loading rate (MLR), deﬁned as the mass of pollutant treated for each bed surface area, is an important operational parameter in CWs. The MLR is optimized by using kinetic constants through two different approaches. Whereas in HFCW this was accomplished by increasing the surface treatment area, in VFCW kinetics were calculated by increasing the hydraulic loading rate (Fig. 4). The elimination of emerging pollutants, as well as general parameters (i.e., BOD5 , TSS), usually follows ﬁrst-order kinetics, showing a concentration dependence [30]. In this regard, NSAID removal through SSF ﬁtted ﬁrst-order kinetics, with values of 0.04–0.19 and 0.11–0.40 m d–1 for HFCW and VFCW, respectively. The fact that these values agree with the BOD5 elimination rates reported by SSF (0.06–1.00 m d–1 ) [30] might suggest that the NSAID removal pathway is comparable to the elimination of the unspeciﬁc organic matter. Therefore, the high kinetic values obtained in VFCW

206

V. Matamoros · J.M. Bayona

yield similar elimination in VFCW to elimination in HFCW using a low surface area or high mass loading rate. Figure 4 summarizes the three main behaviors observed in HFCW: the highly efﬁciently removed compounds, when a low surface area is needed to obtain a high removal (i.e., carboxy-ibuprofen and salicylic acid); the moderately removed compounds, when their removal progresses by increasing the surface area (i.e., naproxen, ibuprofen, and OH-ibuprofen), and ﬁnally the recalcitrant compounds (i.e., diclofenac and ketoprofen), for which their removal is independent of the surface area. Nevertheless, small differences in their performance were noted when hydraulic loading was increased in VFCW, showing similar high removal for all compounds with the exception of diclofenac. Another important parameter that affects NSAIDs removal in VFCW and in HFCW is clogging. In addition to decreasing the HRT, clogging in VFCW seems to directly affect the removal pathway. It produces a decrease in the contribution of the aerobic pathways (i.e., decline on ammonium removal) and replaces it by anaerobic pathways (denitriﬁcation, sulfate reduction, and methanogenesis). Therefore, aerobic pathways seem to be more efﬁcient for the elimination of most of the NSAIDs evaluated than the anaerobic pathways.

Fig. 4 PPCP behavior through the HFCW and VFCW according to the surface area and loading rates, respectively. Discontinuous line represents 100% removal. Reprinted with permission from [19]. © (2006), American Chemical Society

Behavior of Emerging Pollutants in Constructed Wetlands

207

The role of plants on NSAIDs removal was evaluated in an intercomparison study between two vertical beds (VFCW vs. sand ﬁlter) (Matamoros et al., unpublished results). Plant enhanced the removal of ibuprofen and OHibuprofen, attributable to the oxygen availability [28]. On the other hand, salicylic acid and diclofenac were only slightly affected by vegetation, similarly to BOD5 and TSS. Finally, removal efﬁciency of CA-ibuprofen and naproxen showed a moderate plant effect. Furthermore, photodegradation has been shown to be a key factor in SFCW. Indeed, the high diclofenac and ketoprofen (i.e., 99%) removal obtained in SFCW treatment of a WWTP efﬂuent [31] was ascribed to their high photodegradation rates [32, 33]. Moreover, whereas a seasonal difference in diclofenac and ketoprofen removal attributable to sunlight has been observed in SFCW (Table 1) as well as in other open systems [32, 34], no differences were observed for the rest of the analgesics. This could be attributable to a fast biodegradation rate instead of the transformation of the compound by photodegradation. In fact, ibuprofen is considered to be a compound recalcitrant to photodegradation [33]. To summarize, NSAID removal in VFCW and SFCW apparently is better than in conventional WWTP, and is similar to emerging wastewater treatment technologies like membrane bioreactors [35]. Indeed, as shown previously in membrane reactors [36], we found that the removal efﬁciencies of drug residues in HFCW were dependent on their molecular structure [19] (e.g., extended aromatic structures as in ketoprofen and diclofenac). 3.2 Lipid Regulator Drugs and Anti-epileptic Agents Carbamazepine is an important drug used in the treatment of epilepsy, as well as for other psychotherapy applications. On the other hand, cloﬁbric acid is the active metabolite in a series of widely used blood lipid regulators (Fig. 5).

Fig. 5 Chemical structures. Ionized forms are shown. log Dow calculated at pH 8 [64, 73]

208

V. Matamoros · J.M. Bayona

Studies in Europe and North America have shown that these two compounds are two of more frequently detected pharmaceuticals in WWTP efﬂuents, in surface and groundwater and even in sea water [23, 37]. Generally, the low removal of carbamazepine and cloﬁbric acid observed in CWs agrees with the high recalcitrance described for these compounds in conventional WWTP as well as in MBR. Accordingly, removal has not been observed for both compounds in HFCWs [38] but it was fair for carbamazepine in VFCW (ca. 25%) and moderate for carbamazepine, cloﬁbric acid, and gem-

Fig. 6 Cumulative percent mass recovery for carbamazepine and cloﬁbric acid in a continuous injection experiment for two HFCWs (a shallow system and b depth system) and comparison with bromide as a tracer. Reprinted with permission from [39]. © (2005), American Chemical Society

Behavior of Emerging Pollutants in Constructed Wetlands

209

ﬁbrozil in SFCW (ca. 30–60%) [39, 40]. The higher removal of cloﬁbric acid and carbamazepine in SFCW is attributable to their high HRT (i.e., 720 h) that allows interactions with the different wetland compartments (organic matter, bioﬁlm, biota, and plants) and the completeness of removal processes. We suggested that the transport of these compounds along CWs [31] is similar to that in groundwater and packed sediment columns [41–43]. Therefore, in a singular study where carbamazepine and cloﬁbric acid were continuously injected into a HFCW together with bromide as tracer (Fig. 6), their retention was shown to be dependent on their hydrophobicity. Hence, whereas cloﬁbric acid behaved similarly to bromide due its low hydrophobicity (Fig. 5), carbamazepine was retarded by interaction with the gravel bed. Consequently, a carbamazepine load on the gravel bed was observed. No plant or clogging effect was found on the removal of these compounds on account of their high recalcitrance. Nevertheless, in vertical beds, the presence of plants led to a moderate carbamazepine attenuation (i.e., 26 vs. 11%) perhaps attributable to some sorption or/and degradation associated with roots. 3.3 Fragrances Fragrances are compounds used frequently in washing powder, fabric softeners, shampoos, and other consumer products. They are semi-volatile compounds with a wide range of water solubilities (103 –10–1 mg L–1 ). For example, whereas, galaxolide and tonalide are strongly bound to suspended solids due their high hydrophobicity (log Kow = 5.7–5.9), methyl dihydrojasmonate and benzyl acetate occurred in the dissolved phase [44] (Fig. 7). Simonich et al. investigated the fate of fragrances in WWTP and reported a mean removal of 87% during wastewater treatment. Since residues of fragrances are discharged via WWTP, these compounds occur at high concentrations in sewage sludge but have also been detected in the surface waters from continental to marine waters and exposed biota.

Fig. 7 Chemical structures and physicochemical properties of two fragrances [64]

210

V. Matamoros · J.M. Bayona

Fig. 8 Fragrance behavior through HFCW, according surface area. Reprinted with permission from [19]. © (2006), American Chemical Society

Because removals achieved in HFCW, VFCW, and in SFCW are always higher than 80% [19, 31], the use of CWs for fragrance removal is a suitable technology close to WWTP. Fragrance behavior along CWs was studied according to the HFCW surface area (Fig. 8). The high removal observed in the ﬁrst section was in concordance with the high organic matter accumulation in this part of the wetland. The moderate concentrations of polycyclic musks (i.e., galaxolide and tonalide) in the gravel bed (up to 824 µg kg–1 ), in contrast to the low concentrations of methyl dihydrojasmonate, suggest their recalcitrance to biodegradation as compared to the high biodegradability of methyl dihydrojasmonate. Therefore, whereas sorption onto organic matter retained in the ﬁrst part of CW was the predominant removal mechanism of the polycyclic musks, biodegradation appears to be prevalent for methyl dihydrojasmonate, a less hydrophobic compound. Current knowledge suggests that the plant contribution to polycyclic musk elimination is negligible. As expected from its high hydrophobicity, its sorption is primarily dependent on the presence organic matter in the gravel bed. 3.4 Estrogens The natural estrogens (e.g., estrone, estriol, 17α-estradiol and 17β-estradiol), and the synthetic 17α-ethynyl estradiol are considered to be the most potent endocrine disruptors. The estrogen hormone concentrations in the efﬂuent of a conventional WWTP typically range from a few nanograms per liter to several micrograms [45]. These hormones have an impact on the reproductive development in some ﬁsh, even at concentrations as low as a few nanograms per liter [46, 47] (Fig. 9). Table 2 summarizes the estrogen removal efﬁciency, reported as estrogen removal or as a decrease in estrogenic activity, depending on the analyti-

Behavior of Emerging Pollutants in Constructed Wetlands

211

Fig. 9 Chemical structures and physicochemical properties of some estrogens [64]

Table 2 Removal efﬁciency (%) of estrogens in CW and its comparison with WWTP

E2 EE2 Estrogens HRT (days) a b

SFCW [49]

SFCW [71]

HFCWVFCW [48]

WWTP [72]

36 41 – 3.5

– – 83–93 a 22–55

– – > 90 b 3

85–99 71–78 60–99 b 0.5–1

Referred to estrogenic activity Overall average estrogens removal

cal procedure used. Masi et al. [48] reported a high estrogen removal (i.e., > 90%) in CWs composed of a ﬁrst stage HFCW and a second stage VFCW (the so-called hybrid system). These authors pointed out that adsorption onto suspended particulate matter was considerable due to its high hydrophobicity (Fig. 9). On the other hand, although SFCW could reach similar removals to those obtained in a hybrid system, SFCWs require retention times of months instead of days. Moreover, large differences in estrogen removal were observed when SSFCWs with HRTs of days and months were compared, attributable to the high interaction of estrogens with organic matter. Therefore, CWs can reach similar performances to those reported for WWTP, at least for VFCW and HFCW. The literature concluded that conventional WWTP is efﬁcient for the removal of E2 (85–99%) but that estrone removal is relatively poor (25–80%) due the metabolism of E2 to estrone [45]. The mobility of estrogens in CWs was evaluated by a discrete injection of E2, EE2 (Fig. 9), and lithium chloride as tracer in a pilot SFCW [49]. The results show that concentration proﬁles of hormone and lithium differ signiﬁcantly.

212

V. Matamoros · J.M. Bayona

The literature seems to indicate that removal mechanism for these compounds in CWs are mainly associated with sorption by organic matter and bioﬁlm interaction, similar to that reported for sewage sludge in WWTP [45] with subsequent biodegradation. Estrogen removal could be improved by an increase in the HRT since it might increase the interaction time, as reported for activated sludge systems [50]. Because the major elimination pathway is by sorption, subsurface ﬂow is considered to give the best elimination performance, as compared to HFCW or VFCW. 3.5 Surfactants About 3–4 million t year–1 of synthetic surfactants are produced in Western Europe, Japan, and the USA and are intentionally released in large amounts into the aquatic environment [51, 52]. Although the behavior of the most relevant classes of surfactants (namely non-ionic and anionic) has been investigated in CWs, only the non-ionic ones will be reported here. Alkylphenol ethoxylates (APEO), namely nonyl (NPEO) and to a lesser extent octylethoxylates (OPEO), are surfactants of major environmental concern due to the estrogenic properties of their degradation intermediates, nonylphenol (NP) and octylphenol (OP) (Fig. 10) [53]. Their behaviors in CWs have been evaluated at scales from the microcosmic to pilot plant.

Fig. 10 Chemical structure and physicochemical properties of some non-ionic surfactants [64]; a m = n = 1

Belmont and Metcalfe [54]observed high removal rates of NPEO (96.6%) in planted and unplanted microcosmic scale HFCW (HRT = 1–2 days). However, the presence of plants did not signiﬁcantly contribute to the removal of the selected contaminants and it was suggested that adsorption to the substrate was the removal mechanism since the shorter alkyl-chain NPEOs are relatively hydrophobic. Moreover, the removal of ethoxylated (1–3 ethoxy groups) nonylphenols was higher (96–98%) than nonylphenols without ethoxy groups (54–57%). This was attributed to the conversion of ethoxylated derivatives to nonylphenol, as reported previously in conventional WWTPs [55].

Behavior of Emerging Pollutants in Constructed Wetlands

213

In a subsequent study, Belmont et al. [56] evaluated CWs for the removal of NP and NPEOs from a polluted domestic wastewater. The CWs consisted of three sedimentation terraces, stabilization pond, and then a hybrid system (HFCW–VFCW) with a HRT of 2.3 days. The NPEO removal was higher than 75%, nevertheless, most of the reduction occurred in the sedimentation terraces and in the stabilization pond. The rest of the system contributed only marginally to the total removal rate. These results showed again that sorption to sediments and solids is a major removal pathway for these compounds from the aqueous phase [57]. In other study, the elimination of APs, APEOs, AP polyethoxycarboxylates (APECs), and carboxylated APECs (CAPECs) by a SFCW, in order to reduce river pollution, resulted in a partial removal of these compounds. In fact, whereas AP and APEOs were removed at an average of 75%, only 8% of the APECs and CAPECs were eliminated. Thus, it appears that SFCW treatment is fairly effective at removing neutral compounds but it does not correctly eliminate the acidic biodegradation intermediates due their low hydrophobicity. 3.6 Other Emerging Compounds Caffeine and triclosan are employed worldwide as a stimulant and antibacterial agent, respectively [58–61]. Although both compounds have been detected in several aquatic compartments (i.e., wastewater, surface water, groundwater, and sea water) [62], caffeine shows a higher frequency of detection. Consequently, it has been used as a tracer of anthropogenic pollution in several studies [63]. High removal of caffeine in subsurface ﬂow constructed wetlands has been reported in VFCW as well as in HFCW (94–99%). These results are comparable to the ones in WWTPs [19] (also, Matamoros et al. unpublished results). In this regard, caffeine is presented as an easily biodegradable compound, according with its high biodegradation index provided by the EPI suite software [64]. Matamoros and Bayona [19] reported that water depth is an important design factor for the removal of caffeine in HFCW, as mentioned before for BOD5 [29], LAS [65], and NSAIDs (see Sect. 3.1). Furthermore, the negative clogging effect on caffeine removal was also reported in VFCW as well as in HFCW. Waltman and colleagues [66] studied the elimination of triclosan in a SFCW that received the efﬂuent of treated wastewater from an activated sludge. They achieved moderate removals (i.e., 50%) instead of the high removals obtained in WWTP (i.e., 97–99%). As reported above for NSAIDs (see Sect. 3.1), the elimination of emerging pollutants often follows ﬁrst-order kinetics. Therefore, removal efﬁciency is concentration dependent: whereas

214

V. Matamoros · J.M. Bayona

high inlet concentrations are conductive to high removals, low inlet concentrations lead to a low removal.

4 Concluding Remarks The elimination of NSAIDs, lipid regulator drugs, anti-epileptic agents, fragrances, surfactants, estrogens, caffeine, and triclosan by CWs could constitute a feasible alternative to WWTP in small communities. However, until now most of the information has been related to small plants or pilot scale studies. Different factors affecting removal of emerging pollutants have been found relevant, of which oxygen concentration was shown to be one of the most important because of the best performance of aerobic pathways. Hence, VFCW has been postulated as the best option for a wide variety of compounds. Water depth in HFCW was observed to be an important design parameter for the removal of compounds that were better removed by aerobic pathways (i.e., ibuprofen). Loading rate was another important design parameter that needs to be optimized. Whereas some compounds require high surface area or low hydraulic loading rate (e.g., naproxen) for their elimination, the removal of others can be achieved using less surface area (i.e., methyl dihydrojasmonate, CA-ibuprofen). On the other hand, bed clogging produces a negative impact on pollutant removal, except for those that were removed by interaction with organic matter (i.e., galaxolide and tonalide) and those that were not eliminated at all (i.e., cloﬁbric acid). Another important factor is sorption into organic matter (roots and bioﬁlm), which was shown to be important for compounds with high or moderate hydrophobicity (musk, estrogens, and some alkyl phenol derivates). Finally, photodegradation, which only occurs in SFCW, is observed to be compound-dependent, allowing the possibility of increasing the removal of compounds that were low or not biodegradable at all in other systems (i.e., diclofenac and ketoprofen). In summary, VFCWs are postulated as the best option for achieving high removals with low HRTs, a few hours instead of the days or weeks typically used for other CWs. Nevertheless, the performance, in terms of efﬂuent quality, is believed to be better for some emerging pollutants (estrogens and AP), as shown in hybrid designs (combination in series of different CW conﬁgurations). Therefore, the possibilities of using CWs rather than high-cost technologies (i.e., MBR, ultraﬁltration, and ozonation) to remove emerging pollutants from contaminated waters are opened to discussion.

References 1. Cole S (1998) Environ Sci Technol 32:218A 2. Kadlec RH, Knight RL (1996) Treatment wetlands. CRC, Boca Raton, FL

Behavior of Emerging Pollutants in Constructed Wetlands

215

3. Schulz R, Hahn C, Bennett ER, Dabrowski JM, Thiere G, Peall SKC (2003) Environ Sci Technol 37:2139 4. Moore MT, Schulz R, Cooper CM, Simith S, Rodgers JH (2002) Chemosphere 46:827 5. Green MB, Martin JR (1994) Constructed red beds clean up storm overﬂows on small works. In: Proceedings of the Water Environment Federation 67th annual conference and exposition, Chicago, IL. WEF, Alexandria, VA 6. Schueler TR (1992) Design of stormwater wetlands systems: guidelines for creating diverse and effective stormwater wetlands in the mid-Atlantic region. Metropolitan Washington Council of Goverments, Washington, DC 7. Noller B, Woods PH, Ross BJ (1994) Water Sci Tech 29:257 8. Morris M, Herbert R (1997) Water Sci Tech 35:197 9. Litchﬁeld DK (1993) Constructed wetlands for wastewater treatment at Amoco Oil Company’s Madan, North Dakota, reﬁnery. In: Moshiri GA (ed) Constructed wetlands for water quality improvement. Lewis, Chelsea, MI, pp 485–488 10. DeBusk WF (1999) Evaluation of a constructed wetland for treatment of leachate at a municipal landﬁll in northwest Florida. In: Raton B (ed) Constructed wetlands for the treatment of landﬁll leachates. Lewis, Boca Raton, FL, pp 175–186 11. Nielsen SM (1994) Biological sludge drying in reed bed systems – six years operations experience. In: Fourth international conference on wetland systems for water pollution control. ICWS, Guangzhou, China 12. Stearman KG, George KB, Carlson K, Lansford S (2003) J Environ Qual 32:1548 13. Kadlec RH, Knight RL, Vymazal J, Brix H, Cooper P, Haberl R (2000) Constructed wetlands for pollution control: processes, performance, design and operation 2000. IWA, London 14. Wallace DS, Knight RL (2006) Small-scale constructed wetland treatment systems: feasibility, design criteria, and O&M requirements (WERF report). IWA, London 15. Zwiener C, Seeger S, Glauner T, Frimmel FH (2002) Anal Bioanal Chem 372:569 16. Andersen H, Siegrist H, Halling-Srensen B, Ternes AT (2003) Environ Sci Technol 37:4021 17. Arias CA, Brix H (2005) Water Sci Technol 51:1 18. Brix H (1997) Water Sci Tech 97:11 19. Matamoros V, Bayona JM (2006) Environ Sci Technol 40:5811 20. Langergraber G, Haberl R, Laber J, Pressl A (2003) Water Sci Tech 48:25 21. Brix H, Arias CA (2005) Ecol Eng 25:491 22. Vymazal J (2005) Ecol Eng 25:478 23. Daugthon CG, Ternes TA (1999) Environ Health Perspect 107:907 24. Santos JL, Aparicio I, Alonso E (2007) Environ Int 33:596 25. Stuer-Lauridsen F, Birkved M, Hansen LP, Holten Lützhoft HC, Halling-Sorensen B (2000) Chemosphere 40:783 26. Ternes AT, Joss A, Siegrist H (2004) Environ Sci Technol 38:393A 27. Ternes TA, Herrmann N, Bonerz N, Knacker T, Siegrist H, Joss A (2004) Water Res 38:4075 28. Zwiener C, Frimmel FH (2003) Sci Total Environ 309:201 29. García J, Aguirre P, Mujeriego R, Huang Y, Ortiz L, Bayona JM (2004) Water Res 38:1669 30. Rousseau DPL, Vanrolleghem PA, de Pauw N (2004) Water Res 38:1484 31. Matamoros V, García J, Bayona JM (2007) Water Res (in press) [doi:10.1016/j.watres. 2007.08.016] 32. Andreozzi R, Raffaele M, Nicklas P (2003) Chemosphere 50:1319 33. Lin A, Reinhard M (2005) Environ Toxicol Chem 24:1303

216

V. Matamoros · J.M. Bayona

34. 35. 36. 37. 38.

Poiger T, Buser HR, Muller MD (2001) Environ Toxicol Chem 20:256 Radjenovic J, Petrovic M, Barceló D (2007) Anal Bioanal Chem 387:1365 Kimura K, Hara H, Watanabe Y (2005) Desalination 178:135 Buser H-R, Müller MD (1998) Environ Sci Technol 32:188 Dordio A, Teimao J, Ramalho I, Carvalho AJP, Candeias AJE (2007) Sci Total Environ 380:237 Matamoros V, García J, Bayona JM (2005) Environ Sci Technol 39:5449 Gross B, Montgomery-Brown J, Naumann A, Reinhard M (2004) Environ Toxicol Chem 23:2074 Clara M, Strenn B, Kreuzinger N (2004) Water Res 38:947 Scheytt TJ, Mersmann P, Heberer T (2006) J Contam Hydrol 83:53 Drillia P, Stamatelatou K, Lyberatos G (2005) Chemosphere 60:1034 Simonich S, Federle TW, Eckhoff WS, Rottiers A, Webb S, Sabaliunas D, Wolf W (2002) Environ Sci Technol 36:2839 Khanal SK, Xie B, Khanal SK, Thompson ML, Sung SW, Ong SK, Van Leeuwen J (2006) Environ Sci Technol 40:6537 Desbrow C, Routeledge EJ, Brighty GC, Sumpter JP, Waldock MJ (1998) Environ Sci Technol 32:1549 Routeledge EJ, Sheahan D, Desbrow C, Brighty GC, Waldock MJ, Stumpter JP (1998) Environ Sci Technol 32:1559 Masi F, Conte G, Lepri L, Martellini T, Del Bubba M (2004) Endocrine disrupting chemicals (EDCs) and pathogens removal in and hybrid CW system for a tourist facility wastewater treatment and reuse. In: Proceedings 9th international conference on wetland systems for water pollution control, Avignon, France. IWA, London, pp 461–468 Gray JL, Sedlak DL (2005) Water Environ Res 77:24 Ternes TA, Stumpf M, Mueller J, Haberer K, Wilken RD, Servos M (1999) Sci Total Environ 225:81 McAvoy J, Giger W (1986) Environ Sci Technol 20:376 Cook AM (1998) Tenside Surfactant Deterg 35:52 White R, Jobling S, Hoare SA, Sumpter JP, Parker MG (1994) Endocrinology 135:175 Belmont MA, Metcalfe C (2003) Ecol Eng 21:233 Giger W, Brunner PH, Schaffer C (1984) Science 225:623 Belmont MA, Ikonomou I, Metcalfe CD (2006) Environ Toxicol Chem 25:29 Bennett ER, Metcalfe CD (2000) Environ Toxicol Chem 19:784 Stavric B, Klassen R, Watkinson B, Karpinski K, Stapley R, Fried P (1988) Found Chem Toxicol 26:180 Gardinalia PR, Zhaob X (2002) Environ Int 28:521 Reiss R, Mackay N, Habig C, Grifﬁn J (2002) Environ Toxicol Chem 21:2483 Orvos DR, Versteeg DJ, Inauen J, Capdevielle M, Rothenstein A, Cunningham V (2002) Environ Toxicol Chem 21:1338 Siegener R, Chen RF (2002) Marine Pollut Bull 44:383 Sankararamakrishnan N, Guo Q (2005) Environ Int 31:1133 US Environmental Protection Agency (2004) Estimation program interface (EPI) suite. Software available from http://www.epa.gov/oppt/exposure/pubs/episuitedl.htm; last visited: 24 Sept 2007 Huang Y, Latorre A, Barceló D, García J, Aguirre J, Mujeriego R, Bayona JM (2004) Environ Sci Technol 38:2657 Waltman E, Venables B, Waller WT (2006) Environ Toxicol Chem 25:367 Carballa M, Omil F, Lema JM, Llompart M, Garcia-Jares C, Rodriguez I, Gomez M, Ternes T (2004) Water Res 38:2918

39. 40. 41. 42. 43. 44. 45. 46. 47. 48.

49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64.

65. 66. 67.

Behavior of Emerging Pollutants in Constructed Wetlands 68. 69. 70. 71.

217

Buser H-R, Poiger T, Müller MD (1999) Environ Sci Technol 33:2529 Bernhard M, Müller J, Knepper TP (2006) Water Res 40:3419 Heberer T (2002) Toxicol Lett 31:5 Shappell NW, Billey LO, Forbes D, Matheny TA, Poach ME, Reddy GB, Hunt PG (2007) Environ Sci Technol 41:444 72. Auriol M, Filali-Meknassi Y, Tyagi RD, Adams CD, Surampalli RY (2006) Proc Biochem 41:525 73. Jones OAH, Voulvolis N, Lester JN (2001) Environ Technol 22:1383

Hdb Env Chem Vol. 5, Part S/2 (2008): 219–238 DOI 10.1007/698_5_099 © Springer-Verlag Berlin Heidelberg Published online: 23 November 2007

Input of Pharmaceuticals, Pesticides and Industrial Chemicals as a Consequence of Using Conventional and Non-conventional Sources of Water for Artificial Groundwater Recharge M. Silvia Díaz-Cruz (u) · D. Barceló Department of Environmental Chemistry, IIQAB-CSIC, c/ Jordi Girona 18–26, 08034 Barcelona, Spain [email protected] 1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

220

2 2.1 2.2

Artiﬁcial Recharge Systems . . . . . . . . . . . . . . . . . . . . . . . . . . Artiﬁcial Recharge Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . Sources and Pretreatments of Recharge Waters . . . . . . . . . . . . . . . .

221 222 222

3 3.1 3.2 3.3

Potential Groundwater Contamination . . . . . Sources of Contamination . . . . . . . . . . . . Organic Chemical Contamination . . . . . . . . Mitigation of Organic Chemical Contamination

. . . .

224 224 227 232

4

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

234

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

236

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

Abstract Population growth and unpredictable climate changes will pose high demands on water resources in the future. Even at present, surface water is certainly not enough to cope with the water requirement for agricultural, industrial, recreational, and drinking purposes. In this context, the usage of groundwater has become essential, therefore its quality and quantity has to be carefully managed. Artiﬁcial recharge of aquifers can guarantee a sustainable level of groundwater, whilst strict quality control of waters intended for recharge will minimize contamination of both the groundwater and aquifer area. However, all water resources on the planet are threatened by multiple sources of contamination coming from the extended use of chemicals worldwide. In this respect, the environmental occurrence of organic micropollutants such as pesticides, pharmaceuticals, industrial chemicals and their metabolites has experienced fast-growing interest. In addition to conventional sources of water for recharge, mainly surface water and drinking water surplus, non-conventional sources are attracting interest. Recently, the exploitation of alternative water sources for recharge including reclaimed municipal wastewater, treated industrial efﬂuents, and storm water has been evaluated. In this chapter an overview of the priority and emerging organic micropollutants found to be present in recharge, inﬁltrated, and recovered water at managed aquifer recharge sites is presented. Reported results indicated that the drug metabolite 1-acetyl-1methyl-2-dimethyl-oxamoyl-2-phenylhydrazide (AMDOPH) was the compound found at the highest mean concentration (> 1000 ngL–1 ) in the drinking water supply in an artiﬁcial groundwater recharge plant in Berlin, replenishing groundwater from the Lake Tegel.

220

M.S. Díaz-Cruz · D. Barceló

A similar mean concentration was measured for the pesticide cyanazine (1150 ngL–1 ). The pharmaceuticals diclofenac, indomethazine, and bezaﬁbrate, and the industrial chemical tris(2-chloroisopropyl)-phosphate experienced a signiﬁcant removal in the inﬁltration process, with elimination rates in the range 60–100%. In contrast, carbamazepine was found to be one of the most persistent contaminants in groundwater environments, with elimination rates below 10%. Different attenuation behaviors can be found depending on the recharge strategy, for instance bank ﬁltration reduced 80% of the recharge water concentration of the antimicrobial sulfamethoxazole, meanwhile only 50% was decreased through inﬁltrations basins. Keywords Artiﬁcial recharge · Aquifer · Groundwater · Organic contaminants · Surface water · Wastewater Abbreviations ABS Branched-chain alkylbenzenesulfonates AMDOPH 1-Acetyl-1-methyl-2-dimethyl-oxamoyl-2-phenylhydrazide APEOS Alkylphenolethoxylates AR Artiﬁcial recharge DATS Dialkyltetralinsulfonates EPA Environmental Protection Agency LAS Linear alkylbenzene sulfonates MTBE Methyl tert-butyl ether PAHs Polycyclic aromatic hydrocarbons STP Sewage treatment plant THMs Trihalomethanes VOCs Volatile organic compounds WWTP Wastewater treatment plant

1 Introduction The provision of safe and high quality drinking water is essential to preserve human health and ensure a high quality of life. Since the world population is steadily increasing, there is an associated increasing demand for water. In addition, potential climatic changes may in the future cause more periods with low rainfall, causing droughts. In many areas of the world, especially the drier ones, groundwater is the main water source [1]. However, when groundwater withdrawal exceeds the groundwater recharge, aquifer water depletion occurs. In coastal regions this can lead to seawater intrusion, and then the aquifer will not render drinkable water. Therefore, there is an urgent need for artiﬁcial storage of water by suitable facilities. Generally it has been achieved by the construction of different surface reservoirs, such as dams. However, proper sites are getting scarce and an adverse environmental impact on the area is produced. In addition, water stored in surface reservoirs is exposed to evaporation losses, especially important in areas of warm climates. Other

Contamination in Sources of Water for Artiﬁcial Groundwater Recharge

221

drawbacks of dams include the potential for structural damage and failure, and the relatively short lifetime due to the accumulation of sediments and other solids [2]. As a suitable alternative, underground storage via different strategies avoids evaporation losses and algae growth, being sustainable in terms of economical cost and environmental impact [1]. Artiﬁcial recharge (AR) is one of the recharge techniques commonly used in which water is moved, through man-made systems, from the surface of the earth to underground water-bearing strata, where it may be stored for future use [3]. In general, AR is used in addition to natural bank ﬁltration in sites where surface soils are permeable [4]. Artiﬁcial recharge is becoming increasingly important in groundwater management. As an example of the interest paid to AR issues in Europe, the European Commission has recently funded several projects dealing with aquifer recharge under the EC 6th Framework Program priority “Global Change and Ecosystems” [5]. These include ARTDEMO (artiﬁcial recharge demonstration project), GABARDINE (groundwater artiﬁcial recharge based on alternative sources of water: advanced integrated technologies and management) and RECLAIM WATER (water reclamation technologies for safe artiﬁcial groundwater recharge). Brieﬂy, ARTDEMO was focused on the development of management tools (on-line sensors, fast ﬁeld analysis kits) and implementation of operational schemes to decrease the risk of contaminated drinking water supply and to ensure a stable water production of artiﬁcially recharged aquifers. The exploitation of non-conventional sources of water for recharge has been the aim of the projects GABARDINE and RECLAIM WATER. The identiﬁcation of alternative water sources for recharge and the assessment of their feasibility, both environmental and economic, was the focus of the GABARDINE project. Sources evaluated included reclaimed municipal wastewater, treated industrial efﬂuents, and storm water runoff. The objective of RECLAIM WATER was to develop hazard mitigation technologies for water reclamation providing safe and cost-effective routes for artiﬁcial groundwater recharge systems.

2 Artificial Recharge Systems Roughly, AR is the process by which groundwater is replenished at a rate much higher than those under natural conditions. The artiﬁcial recharge of aquifers is a technique broadly used in several European countries, such as Germany and The Netherlands, and especially in the USA and Australia. The type of system to be selected for AR for optimum performance depends on several factors: soil conditions, hydrogeology, topography, water availability, and climate. Another important factor in the selection of the strategy is the pretreatment of the water before recharge.

222

M.S. Díaz-Cruz · D. Barceló

2.1 Artificial Recharge Strategies A number of techniques have been developed for AR of aquifers that can be roughly classiﬁed into direct and indirect approaches. Through indirect recharge techniques, the water from the aquifer is hydraulically pumped to the surface water. This pumping lowers the groundwater level and, therefore, induces the surface water to replenish the aquifer. Direct artiﬁcial recharge involves two levels of actions: at surface and at subsurface level. At surface level, different systems are used, all aiming at providing the maximum water contact surface area for recharge (see Fig. 1). Among them, the system based on basins and percolation tanks is the most common and cost-effective, probably because it is applicable in alluvial areas as well as in hard rock formations. In this technique, the water is impounded in a series of basins/tanks with an effective depth of water of ca. 1.25 m for a maximum rate of inﬁltration. In subsurface recharge, the water is introduced into the unsaturated (vadose) zone below the ground surface to facilitate inﬁltration, or directly into the aquifer. Wells, pits, and shafts are used as recharge pathways. The main advantages of these systems are that they do not need large piece of land and that there are no losses of water in the form of soil mixtures and evaporation.

Fig. 1 Artiﬁcial recharge techniques

2.2 Sources and Pretreatments of Recharge Waters The basic requirement for recharging groundwater is source water availability. Quality and quantity are important factors driving the performance of

Contamination in Sources of Water for Artiﬁcial Groundwater Recharge

223

the recharge system. Several sources of water have been used for recharge purposes, including surface water (streams, canals, and lakes), reclaimed wastewater (domestic and industrial), urban and rural storm water runoff, imported water from other areas, and treated drinking water. However, those sources, especially surface river water [6], are exposed to permanent and sudden pollution by a wide variety of chemicals that are potentially adverse for the aquifer system. At present, the main source of water for groundwater recharge is surface water. Urban development increases storm water run-off and contaminant concentrations, meanwhile it decreases natural inﬁltration areas. Because of this, the use of storm water runoff as a recharge water has substantially increased. As well as wastewater, storm water needs to be treated before recharge when it comes from highly contaminated areas, such as airport deicing facilities, vehicle fueling/washing, auto-recyclers, industrial manufacturing facilities, marinas, and other hotspots [7]. When an excess of drinking water is produced in waterworks, part of this potable water can be used for recharging the aquifer, which will improve the yield of the aquifer as well as the quality of the groundwater. However, disinfection byproducts such as trihalomethanes (THMs) and haloacetic acids formed during the chlorination of water for consumption can be introduced. The widespread ﬁnding that many groundwater basins have been contaminated by a huge range of pollutants derived from industrial, municipal, and agricultural activities, coupled with the fact that once a groundwater and aquifer area are polluted it is difﬁcult and expensive to restore the aquifer availability, provides justiﬁcation for concern about the potential impact of transported micropollutants in recharge waters on the aquifer system, especially when the recovered waters are intended for drinking water supply. As a consequence, distinction between potable and non-potable aquifers is essential. Differentiation is also essential between direct recharge using surface spreading and subsurface injection wells. When direct recharge is performed, the quality of the injected water should meet the quality required from the water that will be subsequently withdrawn from the aquifer. The implication of this is that only potable water should be injected into potable aquifers; and that when aquifer water is to be pumped for unrestricted irrigation, the injected water should meet the standards established for the reuse of wastewater for unrestricted irrigation. Most water sources for AR require some type of pretreatment prior to recharge. This includes typical drinking water treatment for further potable water supply, sedimentation when using streams or lake water, or extensive tertiary treatment for wastewater. Wastewater emerged as an attractive alternative source to surface water for AR even for further potable use [8]. However, it presents some important negative factors such as the presence of pathogens and residual chemicals. As a consequence, wastewater requires primary and secondary treatments with disinfection for use in surface recharge systems [9]. In particular, when wastewater is intended for direct injection

224

M.S. Díaz-Cruz · D. Barceló

or subsurface recharge higher levels of treatments are necessary, which may include microﬁltration, chemical clariﬁcation, air stripping, reverse osmosis, and carbon ﬁltration [10–12].

3 Potential Groundwater Contamination Aquifer systems as environmental compartments are substantially different from other aqueous systems. Long residence times, low temperatures, low degrees of dilution, and decreased microbial population are factors favoring the long-term fate of organic microcontaminants in groundwater. When organic contaminants are introduced into an aquifer through artiﬁcial recharge, they will either move with the water or be adsorbed on the solid surfaces. Whenever retained contaminants do not break down they will accumulate in the aquifer. This accumulation may have long-term impacts. An early study on the transport and fate of more than ﬁfty volatile organic compounds (VOCs) evidenced their long persistence (> 50 years) and the long distances reached (> 10 km) [13]. Reclaiming an aquifer that has been contaminated is a difﬁcult, long and expensive process; therefore, a prerequisite of artiﬁcial recharge is that it should not risk the quality of the groundwater resource. It is well known that hydrogeology inﬂuences the fate and transport of organic microcontaminants. Factors such as depth to water table, sediment porosity and permeability, and groundwater ﬂow control how fast and to what extent contaminants make their way from the recharge point to the recovery point. Geochemical and nutrient conditions also drive the fate of organics into the aquifer, with low dissolved oxygen/low nutrient conditions favoring long-term persistence. 3.1 Sources of Contamination Trace constituents of impaired water quality of concern include organic chemicals, metals, pathogens, and suspended solids. Such concerns apply both to potable and non-potable reuses due to direct and indirect human exposure, although the potential risk associated with non-drinking purposes is signiﬁcantly less. The primary concern regarding contaminant input is the removal of pathogenic microorganisms; nevertheless, trace metals and organic compounds are also important issues. Point source pollution can be controlled and prevented by accomplishing the standing legislation and directives. But even so, there are episodes of uncontrolled (fortuitous or not) chemical spills from industry causing serious water contamination. Agriculture is the main cause of diffuse pollution, which is quite difﬁcult to face and prevent

Contamination in Sources of Water for Artiﬁcial Groundwater Recharge

225

since agrochemicals are spread on extensive ﬁelds. Animal farms are important non-point sources of pollution since a wide variety of phramaceuticals (antibiotics, hormones, and other veterinary formulations) are used in the prevention and treatment of microbial infections and as growth promoters. Following medication, the treated organism excretes between 40% and 90% of the dosage applied in the feces and urine as the parent compound or as metabolites, which in some cases can be bioactive residues even more harmful than the parent compound [14]. This manure is commonly used as organic fertilizer and is spread on soils, from where the pharmaceutical residues excreted can contaminate groundwater directly by percolation or indirectly via surface water contamination during runoff. Efﬂuents from sewage treatment plants (STPs) and wastewater treatment plants (WWTPs) are also important sources of contamination. Therefore, compounds such as industrial chemicals, pharmaceuticals, and personal care products [15] (some of them known to be endocrine disrupters) can be present. In areas where there is low-ﬂow surface water, sewage efﬂuents are needed to increase the amount of water to be further treated for drinking water production. This would be the case for Berlin (Germany), where a survey of surface waters revealed that a number of human pharmaceutical residues and metabolites, in addition to polar pesticides, were found at the microgram per liter level. Further, some of these contaminants were transferred to the groundwater when the contaminated surface waters were used for groundwater recharge in drinking water production plants [16] (see Fig. 2). Other sources of contamination can be identiﬁed in areas where storm water and roof runoff are used for recharging activities [17]. In rain water

Fig. 2 Sources and uses of water in artiﬁcial groundwater recharge systems. input of contaminants. Moderate/low input of contaminants

High

226

M.S. Díaz-Cruz · D. Barceló

pesticides, herbicides, polar compounds photochemically formed in the atmosphere (such as halogenated acetic acids), and nitrophenols emitted in fuel combustion are present. Although pesticides and herbicides are known to be present in storm water, at high concentration during and right after their application in agricultural areas, their contribution to groundwater contamination when inﬁltration is done is almost irrelevant in comparison with the contribution due to their direct application in agriculture. In addition, a wide variety of chemicals (e.g., biocides and anticorrosion agents) are employed in roof construction and preservation and can be washed off by the storm water. Runoff from industrial areas has been shown to contain high concentrations of polycyclic aromatic hydrocarbons (PAHs), pentachlorophenol, and bis(2-hethylhexyl)phthalate [18, 19]. In a comprehensive survey at national scale carried out by the US EPA [20], the concentrations of a number of priority pollutants in urban storm water runoff were determined. Table 1 summarizes the range of concentrations and the frequency of detection of the organic microcontaminants monitored. Outcomes evidenced that many of these contaminants largely exceeded the US EPA water quality criteria for

Table 1 Concentration and frequency of detection of organic priority pollutants in urban storm water runoff. Data from US EPA (1983) [20] Pollutants

Frequency of detection (%)

Range of concentration (µgL–1 )

20 19 17 15

0.0027–0.1 0.008–0.2 0.01–10 0.007–0.1

11

5–15

19 14 10

1–115 1–13 1–37

Pesticides α-BHC α-Endosulfan Chlordane Lindane Halogenated aliphatics Methylene chloride Phenols and cresols Pentachlorophenol Phenol 4-Nitrophenol

Phthalate esters Bis(2-ethylhexyl)phthalate 22 Polycyclic aromatic hydrocarbons (PAHs) Fluoranthene 16 Pyrene 15 Phenanthrene 12 Chrysene 10

4–62 0.3–21 0.3–16 0.3–10 0.6–10

Contamination in Sources of Water for Artiﬁcial Groundwater Recharge

227

human health protection. For instance, the typical standard for PAHs in surface waters for drinking water supplies is 0.0028 µgL–1 [21]. In this study, the concentrations of the four PAHs most commonly detected (chrysene, ﬂuoranthene, pyrene, and phenanthrene) were reported to be a hundred to several thousand times higher than the EPA criteria. Disinfected waters (reclaimed wastewater and drinking water) frequently contain halogenated compounds (known as disinfection by-products) that are formed by an oxidation reaction between the organic carbon present in the water and the used disinfectant agents. The by-products formed are a health concern due to their proven carcinogenicity. Nevertheless, it is known that previous aquifer storage and further recovery signiﬁcantly reduces the concentration of such compounds. The reduction achieved may be explained, in addition to dilution, by microbial degradation usually under reduction conditions. In general, brominate disinfection by-products have proven to be more resistant to degradation than chlorinated species. The by-products of other disinfectant agents, such as ozone and chloramines, are not as well characterized, therefore data is not available. Nevertheless, when balancing the risk in using disinfectant agents to eliminate pathogens with those associated with the by-products formed during the process, it has to be pointed out that effective disinfection is mandatory. Bull et al. [22] showed that the calculated probability of mortality induced by improperly disinfected potable water exceeded the carcinogenic risk posed by disinfection by-products related to chlorine by 1000-fold. 3.2 Organic Chemical Contamination Pharmaceuticals, pesticides, and industrial chemicals have been found to be present in inﬁltrated waters [23–25]. According to an early work by Zoeteman et al. [26], compounds of major concern due to their widespread use, mobility, and persistence in ground waters are hydrocarbons, halogenated hydrocarbons, and ketones. Table 2 lists some of the most frequently detected trace organic contaminants in surface waters that are transferred to the aquifer area. The use of treated wastewater efﬂuents for recharge purposes was investigated in a recent study by Cordy et al. [27]. Results evidenced that the concentration of most pharmaceuticals and other organic wastewater compounds signiﬁcantly decreased. Nevertheless, a number of compounds (i.e., carbamazepine, sulfamethoxazole, benzophenone, 5-methyl-1H-benzotriazole, N,N diethyltoluamide, tributylphosphate and tri(2-chloroethyl)phosphate) were detected in all the waters after soil–aquifer passage, indicating that they have the potential to reach groundwater under recharge conditions. Estrogenic steroids are not efﬁciently removed during wastewater treatment [28–30] and potentially could be introduced in groundwaters through

228

M.S. Díaz-Cruz · D. Barceló

Table 2 Mean concentrations (ngL–1 ) of organic contaminants in surface, inﬁltrated, and supply waters when surface water is used for groundwater enhanced recharge in drinking water production A. Pharmaceuticals and metabolites Compound

Conc. in surface water

Conc. in inﬁltrated water

Conc. in supply water

Refs.

Diclofenac

15 135 40 20 230 120 355 455 325 470 55 135 nd 20 20 30 841 737 485 463

NR NR NR NR NR NR NR NR NR NR NR NR NR NR NR NR 151 (BF) 81 (AR) 122 (BF) 118 (AR)

35 10 50 5 240 40 1250 1570 60 210 40 100 nd nd <5 nd NR NR NR NR

[11] [33] [11] [33] [11] [33] [11] [33] [11] [33] [11] [33] [11] [33] [53] [33] [53] [53] [53] [53]

Cloﬁbric acid Propyphenazone AMDOPH Carbamazepine Primidone Indometazine Bezaﬁbrate Iopromide Sulfamethoxazole

NR not reported, BF bank ﬁltration, AR artiﬁcial recharge, nd not detected B. Pesticides and metabolites Compound

Conc. in surface water

Conc. in supply water

Refs.

Bentazone Mecoprop pp -DDA op -DDA Deethylatrazine Deisopropylatrazine Atrazine amide-I Hydroxyatrazine Hydroxydeethylatrazine

15 15 <5 nd 850 480 < 200 610 200

10 15 5 nd 850 410 < 200 410 130

[11] [11] [11] [11] [24] [24] [24] [24] [24]

Contamination in Sources of Water for Artiﬁcial Groundwater Recharge

229

Table 2 (continued) Compound

Hydroxydeisopropylatrazine Didealkylatrazine Cyanazine Deethylcyanazine Cyanazine amide Deethylcyanazine amide Cyanazine acid Deethylcyanazine acid Propazine Simazine

Conc. in surface water

Conc. in supply water

Refs.

120

110

[24]

260 1360 < 10 150 < 10

210 1150 < 10 500 < 10

[24] [24] [24] [24] [24]

130 < 10

270 270

[24] [24]

120 50

110 30

[24] [24]

nd not detected C. Industrial chemicals Compound

Conc. in surface water

Conc. in supply water

Refs.

N-(Phenylsulphonyl)-sarcosine Tris(2-chloroethyl)-phosphate Tris(2-chloroisopropyl)-phosphate 1,5- Naphthalenesulfonic acid 1,7- Naphthalenesulfonic acid 2,7- Naphthalenesulfonic acid

10 525 1800 90–96 350–368 102

95 215 350 77–90 172–191 38–47

[11] [11] [11] [53] [53] [53]

AR activities; however, because of their physical-chemical properties these steroids generally tend to adsorb onto sediments and organic matter, and therefore are unlikely to be transported to the groundwater. The non-ionic surfactants alkylphenol ethoxylates (APEOs) are a group of organic micropollutants used in many household and industrial cleaning products, plastic manufacturing, and spermicidal formulations. APEOs and their degradation products are among the most frequently detected aquatic contaminants in the environment [6]. Their relative concentrations, with respect to other trace organic contaminants, in wastewater are high. During wastewater treatment, these compounds are incompletely degraded and approximately 40% are released into the aquatic environment as breakdown

230

M.S. Díaz-Cruz · D. Barceló

products [31]. Moreover, several of these APEO metabolites are estrogenic and can persist for long periods of time (> 1 year) in groundwater environments. Naphthalendisulfonates are intermediates for different industrial processes that can either be found in industrial and domestic sewage. Some have been found to be quite stable in wastewater treatment, as is the case for 1,5-naphthalenesulfonic acid [32], which has been reported to occur in extracted water from enhanced recharge sites without any attenuation in its concentration. Surface water for recharge is also often contaminated with pharmaceutical residues and their metabolites as consequence of discharges from municipal WWTP efﬂuents, where they are not effectively removed. For example, six pharmaceuticals belonging to different prescription classes (carbamazepine, primidone, bezaﬁbrate, indomethacine, diclofenac, and propyphenazone) and two drug metabolites (cloﬁbric acid and 1-acetyl-1-methyl2-dimethyl-oxamoyl-2-phenylhydrazide, AMDOPH) were detected in waters from recharging ponds and control wells and six of them persisted in the drinking water supply in an artiﬁcial groundwater recharge plant in Berlin (Germany), replenishing groundwater from the Lake Tegel [33]. Concentrations found for both metabolites in the recharge ponds were in the range 20–455 ngL–1 and in the drinking water supplied between 1570 and 5 ngL–1 . The extremely high value corresponding to the metabolite of dimethylaminophenazone seemed to be related to a spill of the parent compound from a close-by industrial site. In a recent study, Heberer et al. [11], in addition to these pharmaceuticals, investigated the presence of herbicides, pesticides, and ﬂame retardant compounds at the transects of Lakes Tegel and Wannsee, whose waters are used for groundwater recharge. Results revealed that the residues of the herbicides mecoprop and bentazone, as well as two metabolites of the banned DDT, were detected. Regarding ﬂame retardants, both compounds analyzed, tris(2-chloroethyl)-phosphate and tris(2-chloroisopropyl-9phosphate, were found to be present, but considerably attenuated. Other pharmaceuticals, such as antimicrobials, have been monitored in managed aquifer recharge sites. A mean concentration of 118 ngL–1 sulfamethoxazole has been reported in drinking water extracted from the production well in the artiﬁcial recharge facility at Lake Tegel [34]. The analgesic diclofenac [35, 36] and the lipid regulator bezaﬁbrate [37] were also detected in groundwater at low concentrations. Heberer et al. [11] recently investigated the behavior of N-(phenylsulphonyl)-sarcosine, the metabolite of the anti-corrosion agent N-(phenylsulfonyl)-N-methyl-caprionic acid, in surface water and in monitoring production wells in a groundwater recharge site in Berlin. Although data on surface water indicated low levels of the metabolite, the concentrations of the production wells were more than ninefold higher [38]. The removal of VOCs is strongly concentration-dependent and, therefore, can eventually represent a problem in artiﬁcial recharge sites located close to a source of such compounds [39]. Especial attention has been given to the oxy-

Contamination in Sources of Water for Artiﬁcial Groundwater Recharge

231

genate methyl tert-butyl ether (MTBE), an additive added to gasoline as an octane enhancer and to reduce emissions. The consumption of MTBE in Europe has increased dramatically by around 25% in 1999 in comparison to its use in 1995 [40]. The main contribution of MTBE in groundwater can be found in the leakage from underground storage tanks; however, another minor contribution comes from the surface waters used in AR, where levels up to the microgram per liter level have been reported [41, 42]. However, it is still not clear whether MTBE is attenuated or not during inﬁltration, since concentrations ranging from 110 ngL–1 to 5.8 µgL–1 were detected in water samples from recovering wells [43, 44], and up to 400 ngL–1 in tap water from several groundwater sources [45]. These concentrations in any case exceeded the drinking water advisory concentrations set up by the US EPA for MTBE (20–40 µgL–1 ) [46]. Hospital wastewater efﬂuents contain, in addition to residues of typical prescription drugs, certain pharmaceutical compounds of hospital-use, such as X-ray contrast media. According to Ternes and Hirsch [47], these contrast media constitute the main fraction of adsorbable organic iodine in waters. Such compounds are very polar, persistent in the aquatic environment, and have been found to be recalcitrant in wastewater treatments. Therefore, compounds such as diatrizoate, iopromide, iopamidol, and amidotrizoic acid were detected up to the microgram per liter concentration level in ground and inﬁltrated waters [36, 47, 48]. Recently, the iodinated X-ray contrast agent iopromide has been looked for in extracted raw water from monitoring and production wells of a bank ﬁltration site and one artiﬁcial recharge site at Lake Tegel. However, in both production wells the concentrations found were below the limit of quantiﬁcation (< 20 ngL–1 ). In storm water, triazines, acetamides, and phenoxy acids have been detected, a few of them quite frequently (atrazine 33 ngL–1 , alachlor 19 ngL–1 , metholachlor 15 ngL–1 , R-mecoprop 10 ngL–1 , R-diclorprop 12 ngL–1 ). In addition, some metabolites (desethylatrazine 29 ngL–1 , deisopropilatrazine 26 ngL–1 ) and enantiomers (S-mecoprop 10 ngL–1 , S-diclorprop 9 ngL–1 ) were found to be present [17]. In the same study, roof runoff used for inﬁltration was also searched for contaminants. In this case, the pesticides detected and their concentrations were in concordance with those found associated with direct agricultural use. Moreover, a major portion of the compounds present in such waters were transferred to the groundwater without any signiﬁcant decrease in their concentrations, especially when the inﬁltration process was carried out in areas with soils of high permeability (rapid inﬁltration through particles of quite large size), and they can then be detected in the inﬁltrate water [49]. It has to be pointed out that the concentration of metabolites can increase in inﬁltrated water in comparison to surface water since the expanded residence time, due to the low velocity in groundwater, allows the formation of metabolites. After ﬁltration, increases in the concentration of certain metabolites of triazine and acetamide herbicides (namely cyanazine amide, cyanazine acid, and deethylcyanazine) have been identiﬁed [49].

232

M.S. Díaz-Cruz · D. Barceló

3.3 Mitigation of Organic Chemical Contamination Several mechanisms including adsorption, dispersion, and biological transformations (within the ﬁrst few meters of inﬁltration) are the main processes driving the attenuation and fate of organic contaminants in groundwater environments during AR [50]. Nevertheless, these processes are not yet sufﬁciently understood. Despite this, it is unquestionable that both aerobic and anoxic underground passage of recharge water can signiﬁcantly lower the concentration of several organic microcontaminants present in the source water; however, the extent of reduction is often difﬁcult to predict since several factors inﬂuence it. Different patterns and kinetics for the removal of the trace organic compounds can be found; however, there is a general agreement that signiﬁcant improvement is achieved in the quality of the recovered waters [33, 51]. According to an early study by Zullei-Seibert [39], the efﬁciency of pesticide residue removal in artiﬁcial inﬁltration recharge is within the range 10% (atrazine) to 100% (lindane). The behavior of organic chemicals during recharge is affected by factors such as geological conditions, permeability of the soil, residence time, Ta , and pH, as well as by the physical-chemical properties of every single compound. In general, the behavior of such contaminants is mainly inﬂuenced by biological processes [52]. Under anoxic conditions and Ta > 5 ◦ C a number of halogenated C1 –C3 aliphatics, nitroand chloro-benzenes are found to be extensively degraded by sequential dehalogenation, while at lower temperatures the degradation rate considerably decreases and the formation of by-products occurs, such as dichloroand bromochloro-methane, and dichloroanilines. On the other hand, aerobic processes appear to be responsible for the high degree of elimination of trihalomethanes and phenoxy pesticides, associated with the unavoidable formation of the hazardous chlorophenols due to the degradation of such pesticides. Red-ox processes in enhanced groundwater enrichment sites are known to affect the elimination of pharmaceutical residues (e.g., sulfamethoxazole, carbamazepine, phenazone, X-ray contrast agents such as iopromide) as well as for industrial chemicals (e.g., the isomers 1,7- and 2,7-naphthalenesulfonic acid) [34]. Sulfamethoxazole can be reduced to 20% of the surface concentration by bank ﬁltration (anoxi/anoxic); however, inﬁltration basins were only able to degrade it by half [34]. For iopromide, results indicate that removal rates at both sites are around 82% and 89%, respectively, which is in concordance with the rapid degradation observed under both aerobic and anoxic conditions. However, it was only partially dehalogenated under anoxic conditions, as observed by Grünheid et al. [53] during bank ﬁltration and AR. Both aromatic sulfonates appeared to be more persistent, with degradation rates between 50% and 63% under aerobic conditions [34].

Contamination in Sources of Water for Artiﬁcial Groundwater Recharge

233

In groundwater, red-ox processes are known to be temperature-dependent, since these processes are in general microbially catalyzed and temperature variations inﬂuence microbial activity [54]. Large temperature variations of the inﬁltrate are known to affect the red-ox related conditions. Under warmer temperatures (> 14 ◦ C), anoxic conditions prevail, while below 14 ◦ C oxic conditions develop. Recent studies investigated the impact of the inﬁltration water temperature on the red-ox degradation of carbamazepine, phenazone, and phenazone-like pharmaceuticals in an AR site in Berlin. Outcomes evidenced that removal during inﬁltration was not observed for carbamazepine, irrespective of its red-ox state. Meanwhile, different degradation rates were shown by the rest of the compounds under aerobic conditions [55, 56]]. These results are in concordance with the poor mitigation reported by several authors of carbamazepine and another antiepileptic drug, primidone, which were found to be present in ground and drinking water where inﬁltration was carried out [29, 37, 57–59]. Some laboratory and ﬁeld experiments on inﬁltration processes evidenced signiﬁcant removal of diclofenac and bezaﬁbrate in the inﬁltration process as result of their tendency to be adsorbed on the soil/sediment [24, 33, 37, 60] and to be transformed during passage [61]. According to Preuss et al. [57], the elimination of bezaﬁbrate along with diclofenac and ibuprofen would be in the range 60–80%. A higher rate of > 90% was, however, reported by Heberer and Adam [33] for diclofenac and bezaﬁbrate, which along with indomethazine were not detected in the drinking water produced (< 1 ngL–1 ). The elimination of indomethazine and bezaﬁbrate may be explained in terms of their high octanol–water partition coefﬁcient (log Kow ), which would be responsible for adsorption processes onto the soil. Other mechanisms, such as hydrogen bonding, would be related to the removal of more polar organic micropollutants. On the basis of the studies of Heberer et al. [37] and of Scheytt et al. [61], littleto-medium attenuation is observed for propyphenazone, which was detected in groundwater, shallow wells, and in production wells. Cloﬁbric acid, according to several authors, exhibited a high persistence since little or no adsorption and only very low degradation was observed during the soil passage [57, 60–64]. Concerning steroids, Zuehlke et al. [65] pointed out that a short path of few centimeters can substantially decrease the concentration of estrone. This removal can be attributed mainly to biodegradation rather than to adsorption, although biodegradation of adsorbed hydrophobic steroids may occur [29, 30]. Data on the behavior of musk fragrances during soil–aquifer passage indicates that these compounds are also extensively removed in the top soil. However, the formation of degradation compounds cannot be excluded, further research being essential to differentiate between sorption and degradation through the identiﬁcation of degradation products [60]. The behavior of chlorinated ﬂame retardants during inﬁltration has also been studied [30]. Different removal extents were reported: for instance, tris-

234

M.S. Díaz-Cruz · D. Barceló

(chloroethyl)-phosphate) (TCEP) was readily biodegraded during long-term passage, whilst for tris-(trichloroisopropyl)-phosphate (TCIPP) a signiﬁcant removal was not observed. High degradation rates (up to 98%) have been reported in an earlier study by Schaffner et al. [66] for the non-polar volatile compounds, nonylphenol, nonylphenol ethoxylate, nonylphenol diethoxylate, nitrilotriacetate ethoxylate, and diamintetracetic acid in bank ﬁltered water in Switzerland. In a more recent study [67], the occurrence, transformation, and elimination of APEOs and their metabolites (mainly alkylphenol polyethoxycarboxylates, APECs, and carboxyalkylphenol polyethoxycarboxylates, CAPECs [31]) during soil aquifer treatment mimicking riverbank ﬁltration in Arizona (USA) was examined. The results of this ﬁeld study evidenced that both metabolite groups are signiﬁcantly (> 90%) and rapidly removed during inﬁltration within the ﬁrst 3 m, under both anoxic and aerobic conditions. At depths > 3 m only alkylphenoxy acetic acids, carboxyalkylphenoxy acetic acids, and alkyl phenols remain, which suggests a relationship between the number of ethoxy units and the depth at which the compound is attenuated. Several attempts to improve understanding of the transport processes and fate of organic microcontaminants in the subsurface environment have been conducted by the US Geological Survey (USGS) [68] through a number of studies on VOCs [12, 69], alkylbenzenesulfonates and dialkyltetralinsulfonates (DATS) [70, 71], atrazine [72], and disinfection byproducts [73, 74]. According to these studies, THMs behaved as conservative contaminants in the aquifer since bacteria were unable to degrade them under the anaerobic conditions present in the aquifer. Adsorption processes were not found to occur, therefore dilution seemed to be the only effective mechanism controlling the concentration of THMs in the extracted water. The herbicide atrazine was found to be present in almost all monitoring wells at the recharge sites investigated along the Equus Beds aquifer in Wichita, (Kansas, USA) in a 1995 - 2000 surveillance study on the effects of AR on water quality [72]. Results revealed that after recharge, the concentration of atrazine was similar to before recharge concentrations, and that was substantially lower than the US EPA standard for drinking water (3.0 µgL–1 ) [46]. As regards sulfonated compounds, concentrations in groundwater after recharge with reclaimed wastewater were < 10–20 µgL–1 for linear alkylbenzenesulfonates (LAS) and DATS, and ca. 2 mgL–1 for branched-chain alkylbenzenesulfonates (ABS).

4 Summary Groundwater recharge is important in order to extend the clean water chain and provides major beneﬁts in terms of improving the way in which water

Contamination in Sources of Water for Artiﬁcial Groundwater Recharge

235

is used as a renewable resource, especially to ensure the supply of drinking water of good quality. However, most of the rivers, lakes, and other sources of water are polluted to some extent. Depending on the operational controls of aquifer recharge activities, the nature of the recharge water, and the hydrogeology of the receiving aquifer, a variety of trace organic contaminants can be introduced into the groundwater. Artiﬁcial recharge using recharge waters of impaired quality (such as reclaimed wastewater and storm water runoff) is a viable alternative to surface water where recharge is intended for non-potable purposes (such as landscape irrigation) after receiving at least a secondary treatment. However, the health risk associated with such sources is high where the recharge water is to be used as drinking water, and should only be used when better quality sources are unavailable. Signiﬁcant behaviors of trace organic contaminants during artiﬁcial aquifer recharge have been found. Outcomes reported underlined the potential of groundwater recharge systems to remove contaminants to some extent. However, at recharge sites the removal of organic pollutants depends on a number of factors, such as the climate, hydrogeological features of the area, and the physical-chemical conditions of the aquifer. In addition, different attenuation behaviors can also be found depending on the recharge strategy: bank ﬁltration reduced 80% of the recharge water concentration of the antimicrobial sulfamethoxazole, but only 50% was decreased through inﬁltration basins. Pharmaceuticals, personal care products, pesticides, and a number of industrial chemicals have been found in inﬁltrated waters and supply waters even at the low microgram per liter level, but usually below advisory concentration set up by the water authorities. Reported results indicated that the drug metabolite 1-acetyl-1-methyl-2-dimethyl-oxamoyl2-phenylhydrazide (AMDOPH) was the compound found at the highest mean concentration (> 1000 ngL–1 ) in the drinking water supply in an artiﬁcial groundwater recharge plant in Berlin, replenishing groundwater from Lake Tegel. A similar mean concentration was measured for the pesticide cyanazine (1150 ngL–1 ). The pharmaceuticals diclofenac, indomethazine, and bezaﬁbrate, and the industrial chemical tris(2-chloroisopropyl)phosphate experienced a signiﬁcant removal in the inﬁltration process, with elimination rates in the range 60–100%. In contrast, carbamazepine was found to be one of the most persistent contaminants in groundwater environments. Acknowledgements This work has been funded by the EC 5th and 6th Framework Programmes (Projects ARTDEMO EVK1-CT2002-00114, NOMIRACLE GOCE-003956, and RISKBASE GOCE-036938), and by the Spanish Ministry of Education and Science (Project BQU2002-10903-E). M.S. Díaz-Cruz acknowledges her Ramon y Cajal contract from the Spanish Ministry of Education and Science.

236

M.S. Díaz-Cruz · D. Barceló

References 1. Bouwer H (2000) Agr Water Manage 45:217 2. Postel S (1999) Pillar of sand. Worldwatch Institute, Washington DC 3. Bouwer H (1999) Artiﬁcial recharge of groundwater: Systems, designs and management. In: Mays LW (ed) Hydraulic design handbook. McGraw-Hill, New York, Chap 24 4. Bouwer H (2002) Artiﬁcial recharge of groundwater: hydrogeology and engineering. Hydrogeol J 10:121 5. Community Research and Development Information Service (2006) Global change and ecosystems. Europa Publications Ofﬁce, Brussels. http://cordis.europa.eu/sustdev/ environment/ Last accessed: 25 Oct 2007 6. Kolpin DW, Furlong ET, Meyer MT, Thurman EM, Zaugg SD, Barber LB, Buxton HT (2002) Environ Sci Technol 36:1202 7. Pitt RE, Field R, Lalor M, Brown M (1995) Water Environ Res 67:260 8. Pinholster G (1995) Environ Sci Technol 29:174A 9. National Research Council (1994) Groundwater recharge using waters of impaired quality. National Academy Press, Washington, DC 10. Asano T, Cotruvo JA (2004) Water Res 38:1941 11. Heberer T, Mechlinski A, Franck B, Knappe A, Massmann G, Fritz B (2004) Ground Water Monitor Remed 24:70 12. Bonné PAC, Hofman JAMH, van der Hoek JP (2002) Water Sci Technol 2:139 13. Barber LB, Thurman EM, Schroeder MP (1988) Environ Sci Technol 22:205 14. Jjemba PK (2002) Agr Ecosyst Environ 93:267 15. Daughton CG, Ternes TA (1999) Environ Health Persp 109:907 16. Heberer T, Schmidt-Bäumler K, Stan HJ (1998) Acta Hydrochim Hydrobiol 26:272 17. Bucheli TD, Müller SR, Heberle S, Schwarzenbach RP (1998) Environ Sci Technol 32:3457 18. Pitt R, McLean J (1986) Toronto area watershed management strategy study: Humber River pilot watershed project. Water Resources Branch, Ministry of the Environment, Toronto, Ontario 19. Pitt R, Field R, Lalor M, Brown M (1995) Water Environ Res 67:260 20. US Environmental Protection Agency (1983) Result of the nationwide urban runoff program. NTIS publication no. PB 84-185552. Water Planning Division, US EPA, Washington, DC 21. US Environmental Protection Agency (1986) Quality criteria for water. EPA 440/5-86001. US EPA, Washington, DC 22. Bull RJ, Gerba C, Trussell RR (1990) Crit Rev Environ Control 20:77 23. Hiemstra P, Kolpa RJ, van Eekhout JM, van Kessel AL, Adamse ED, van Paassen JAM (2003) J Water Supply Res Technol 52:37 24. Verstraeten IM, Heberer T, Scheytt T (2002) Occurrence, characteristics and transport and fate of pesticides, pharmaceutical active compounds, and industrial and personal care products at bank-ﬁltration sites. In: Ray C, Melin R, Linsky RB (eds) Riverbank ﬁltration improving source water quality. Kluwer, Dordrecht 25. Heberer T (2002) Toxicol Lett 131:5 26. Zoeteman BCJ, De Greef E, Brinkmann FJJ (1981) Sci Total Environ 21:187 27. Cordy GE, Duran NL, Bouwer H, Rice RC, Furlong ET, Zaugg SD, Meyer MT, Barber LB, Kolpin DW (2004) Ground Water Monit R 24:58 28. Baronti C, Curini R, D’Ascenzo G, Di Corcia A, Gentili A, Samperi R (2000) Environ Sci Technol 34:50 29. Mansell J, Drewes J (2004) Ground Water Monit R 24:94

Contamination in Sources of Water for Artiﬁcial Groundwater Recharge 30. 31. 32. 33. 34. 35. 36. 37. 38. 39.

40.

41.

42. 43.

44. 45. 46.

47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60.

237

Amy G, Drewes J (2007) Environ Monto Assess 129:19 DiCorcia A, Cavallo R, Crescenzi C, Nazzari M (2000) Environ Sci Technol 34:3914 Stüber M, Reemtsma T, Jekel M (2002) Vom Wasser 98:133 Heberer T, Adam M (2004) Environ Chem 1:22 Grünheid S, Amy G, Jekel M (2005) Water Res 39:3219 Heberer T, Dünnbier U, Reilich C, Stan HJ (1997) Fresen Environ Bull 6:438 Sacher F, Lange FT, Brauch HJ, Blankenhorn I (2001) J Chrom A 938:199 Heberer T, Verstraeten IM, Meyer MT, Mechlinski A, Reddersen K (2001) Water Res Update 120:4–17 Knepper PK, Kirschhöfer F, Lichter I, Maes A, Wilken RD (1999) Environ Sci Technol 33:945 Zullei-Seibert N (1996) In: Kivimäki A-L, Suokko T (eds) Proceedings of an international symposium on artiﬁcial recharge of groundwater (ARG96), Helsinki, 3–5 June 1996. Nordic Hydrological Programme Report no 38, p 247 European Council (2001) Risk assessment of methyl tert-butyl ether (MTBE), EINECS No. 216-653-1, carried out in the framework of European Council Regulation (EEC) 793/93 on the evaluation and control of the risk of existing substances. Ministry of the Environment Finland, Helsinki (Final draft 06/2001) Reiser RG, O’Brien AK (1998) Occurrence and seasonal variability of volatile organic compounds in seven New Jersey streams. Water Resources Investigations Report 984074. US Geological Survey, Denver. http://nj.usgs.gov/publications/WRIR/wrir984074.pdf. Last accessed: 25 Oct 2007 Achten C, Püttmann W (2000) Environ Sci Technol 34:1359 Grady SJ (1998) Volatile organic compounds in groundwater in the Connecticut, Housatonic and Thames river basins, 1993–1995. National Water Quality Assessment Fact Sheet 029-97. US Geological Survey, Denver Achten C, Kolb A, Püttmann W (2002) Environ Sci Technol 36:3662 Piazza F, Barbieri A, Violante FS, Roda A (2001) Chemosphere 44:539 US Environmental Protection Agency (2006) Drinking water standards. US EPA, Washington, DC. http://www.epa.gov/waterscience/criteria/drinking/dwstandards.html. Last accessed: 25 Oct 2007 Ternes TA, Hirsch R (2000) Environ Sci Technol 34:2741 Putschew A, Wischnack S, Jekel M (2000) Sci Total Environ 255:129 Verstraeten IM, Thurman EM, Lindsey ME, Lee EC, Smith RD (2002) J Hydrology 266:190 Roberts PV, Valocchi AJ (1981) Sci Total Environ 21:161 Yun G, Cheng X, Jian X, Wu T, Pi Y (2000) Tsingua Sci Technol 5:333 Hrubec J, den Engeslam G, de Groot AC, den Hartog RS (1995) Int J Environ Anal Chem 58:185 Grünheid S, Amy G, Jekel M (2005) Water Res 39:3219 Prommer H, Stuyfzand PJ (2005) Environ Sci Technol 39:2200 Massmann G, Greskowiak J, Dünnbier U, Zuehlke S, Knappe A, Pekdeger A (2006) J Hydrology 328:141 Greskowiak J, Prommer H, Massmann G, Nützmann G (2006) Environ Sci Technol 40:6615 Preuss G, Willme U, Zullei-Seibert N (2002) Acta Hydrochim Hydrobiol 29:269 Kuehn W, Mueller UJ (2000) J AWWA 92:60 Drewes J, Heberer T, Rauch T, Reddersen K (2003) Ground Water Monit R 23:64 Ternes TA, Bonerz M, Hermann N, Teiser B, Andersen HR (2007) Chemosphere 66:894

238

M.S. Díaz-Cruz · D. Barceló

61. Scheytt TJ, Mersmann P, Rejman-Rasinski E, These A (2007) J Soil Sediment 7:75 62. Buser HR, Müller MD, Theobald N (1998) Env Sci Technol 32:188 63. Scheytt TJ, Grams S, Rejman-Rasinski E, Heberer T, Stan HJ (2001) Pharmaceuticals in groundwater: chloﬁbric acid beneath sewage farms south of Berlin, Germany. In: Daughton CG, Jones-Lepp TL (eds) Pharmaceuticals and personal care products in the environment – scientiﬁc and regulatory issues. ACS Symposium series 791:84 64. Mersmann P, Scheytt TJ, Heberer T (2002) Acta Hydrochim Hydrobiol 30:275 65. Zuehlke S, Heberer T, Duennbier U, Fritz B (2004) Ground Water Monit R 24:78 66. Schaffner C, Ahel M, Giger W (1987) Water Sci Technol 19:1195 67. Montgomery-Brown J, Drewes JE, Foz P, Reinhard M (2003) Water Res 37:3672 68. US Geological Survey (2007) Artiﬁcial recharge. USGS, Denver. http://water.usgs.gov/ ogw/artiﬁcial_recharge.html. Last accessed: 25 Oct 2007 69. Barber LB (1994) Environ Sci Technol 28:890 70. Thurman EM, Willoughby T, Barber LB, Thorn KA (1987) Anal Chem 59:1798 71. Field JA, Barber LB, Thurman EM, Moore BL, Lawrence DL, Peake DA (1992) Environ Sci Technol 26:1140 72. US Geological Survey (2007) Atrazine in source water intended for artiﬁcial groundwater recharge, south-central Kansas. USGS, Denver. http://ks.water.usgs.gov/Kansas/ pubs/fact-sheets/fs.074-98.html. Last accessed: 25 Oct 2007 73. Rostad CE, Leenheer JA, Katz BG, Martin BS, Noyes TI (2000) Characterization and disinfection by-product formation potential of natural organic matter in surface and ground waters from northern Florida. In: Natural organic matter and disinfection byproducts: characterization and control in drinking water. American Chemical Society Symposium Series 761. American Chemical Society, Washington, DC, pp 154–172 74. Leenheer JA, Rostad CE, Barber LB, Schroeder RA, Anders R, Davisson ML (2001) Environ Sci Technol 35:3869

Hdb Env Chem Vol. 5, Part S/2 (2008): 239–264 DOI 10.1007/698_5_097 © Springer-Verlag Berlin Heidelberg Published online: 28 November 2007

Advanced Sorbent Materials for Treatment of Wastewaters Petar Jovanˇci´c (u) · Maja Radeti´c Textile Engineering Department, Faculty of Technology and Metallurgy, University of Belgrade, Karnegijeva 4, 11120 Belgrade, Serbia [email protected] 1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

240

2 2.1 2.1.1 2.1.2 2.1.3 2.2 2.2.1 2.2.2

Textile Wastewaters . . . . . Color Removal . . . . . . . . Activated Carbon . . . . . . Zeolites . . . . . . . . . . . . Low-Cost Sorbents . . . . . Removal of Heavy Metal Ions Low-Cost Sorbents . . . . . Nanosorbents . . . . . . . .

. . . . . . . .

242 246 247 250 252 255 255 259

3

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

260

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

261

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

Abstract Despite the existence of a wide range of wastewater treatment technologies, sorption is still commonly applied for the puriﬁcation of industrial efﬂuents. This is particularly valid for textile industry efﬂuents, where different techniques must be combined to achieve the optimum effect of puriﬁcation. The combination of biological and physicochemical processes results in the removal of most organic and inorganic pollutants from textile wastewater but the resulting efﬂuent is still fairly colored and needs to be additionally treated. Therefore, an overview of conventional and advanced sorbents for the treatment of wastewaters is reported. The use of activated carbon and zeolites in color removal and removal of heavy metal ions from textile efﬂuents is discussed in detail. The potential of low-cost or cost-effective sorbents as an alternative to the conventionally used sorbents is also underlined. Finally, the opportunities and challenges of using nanomaterials in the treatment of industrial wastewaters are highlighted. Keywords Color removal · Heavy metal removal · Low-cost sorbents · Sorbents · Textile efﬂuents Abbreviations AC activated carbon ACs activated carbons ACF activated carbon ﬁber ACFs activated carbons ﬁbers AOX halogenated organic compounds BOD biochemical oxygen demand BV bed volume (Vf /Vr )

240 C/Co COD EDTA CHT GAC H2 O2 HTAB LTP MWCNTs PAC PET RO TSS UF Vf Vr

P. Jovanˇci´c · M. Radeti´c normalized efﬂuent concentration chemical oxygen demand ethylenediaminetetraacetic acid chitosan granulated activated carbon hydrogen peroxide hexadecyl trimethyl ammonium bromide low temperature plasma multiwalled carbon nanotubes powdered activated carbon polyethyleneterephalate reverse osmosis total suspended solids ultraﬁltration total water volume passing column during the adsorption process (m3 ) ﬁxed-bed volume (m3 )

1 Introduction Any treatment of industrial or municipal wastewaters is a process which results in the transformation of a mixture of substances into two or more products that differ from each other in composition. This process, also known as the separation, may be difﬁcult to achieve inducing the high efﬂuent treatment costs in chemical, petrochemical, and pharmaceutical industries. For many separation processes, the separation is caused by a mass separating agent [1]. The mass separating agent for the process of adsorption (adsorptive separation) is the adsorbent or sorbent. Consequently, the performance of any adsorptive separation or puriﬁcation process is directly determined by the quality of the sorbent. Because of the progress made in sorbent and cyclic process developments, adsorption has already become a key separation tool that is used in textile and many other industries. Adsorption is usually performed in columns packed with sorbent particles, or ﬁxed-bed adsorbers. The high separating power of chromatography achieved in a column is a unique advantage of adsorption compared to other separation processes. The high separating potential is provided by the continuous contact and equilibration between the liquid and sorbent phases. Adsorption is ideally suited for puriﬁcation applications such as industrial wastewater treatment as well as difﬁcult separations [2]. In spite of many research efforts and patents on adsorption, there are only a few commercially available sorbents that are used in the current adsorption processes (activated carbon, zeolites, silica gel, and activated alumina). Future applications of adsorption are limited by the availability of new and more efﬁcient sorbents. Ideally, the sorbent should be tailored in such a way to meet the requirements of each speciﬁc application. Development of better sorbents

Advanced Sorbent Materials for Treatment of Wastewaters

241

can also improve the performance of the current industrial processes. The last two decades showed an explosion in the development of new forms of carbon (carbon molecular sieves, super-activated carbon, activated carbon ﬁbers, carbon nanotubes, and graphite nanoﬁbers) and nanoporous materials i.e. nanomaterials (metal-containing nanoparticles, dendrimers etc.). However, the potential use of the adsorption properties of these new materials is in a process of intensive exploration. Advances in nano-scale science and engineering are providing great opportunities to develop more cost effective and environmentally acceptable water puriﬁcation processes [3]. Activated carbon (AC) has been used as an all-purpose sorbent. Its precedent, charcoal, was ﬁrst used in the sugar industry in England in 1794 to decolorize sugar syrup. The major development of activated carbon took place during World War I, for use in ﬁlters to remove chemical agents from air. The commercial activated carbon has taken its present form since the 1930s [4]. Silica gel and activated alumina are used mainly as desiccants, although many modiﬁed forms are available for special puriﬁcation applications. Synthetic zeolites, the youngest type among the four, were invented by Milton in 1959 [5, 6]. The zeolites that are in commercial use today are mainly the types in Milton’s invention, i.e., types A, X, and Y. Zeolites exhibit particularly good adsorption properties due to their unique surface chemistries and crystalline pore structures. It should be noted, however, that a sizable portion of the commercial zeolites is used for ion exchange processes and as catalysts. Polymeric resins are increasingly used in potable water puriﬁcation, because they can reduce some organics to lower concentration levels than activated carbon does. Acid-treated clays and pillared clays are used for treatments of edible and mineral oils. The adsorptive separation is achieved by one of three mechanisms: the steric, kinetic, or equilibrium effect [2]. In the case of the steric effect only small and properly shaped molecules can diffuse into the sorbent, whereas other molecules are totally excluded. Kinetic separation is achieved by virtue of the differences in diffusion rates of different molecules. A large majority of processes operate through the equilibrium adsorption of a mixture and hence are called equilibrium separation processes. Steric separation is characteristic for zeolites and molecular sieves because of the uniform aperture size in the crystalline structure. For equilibrium separation, the starting point for sorbent design/selection is to examine the fundamental properties of the targeted molecule that is to be adsorbed (compared with the other molecules in the mixture): polarizability, magnetic susceptibility, permanent dipole moment, and quadrupole moment. If the targeted molecule has high polarizability and magnetic susceptibility, but no polarity, carbon with a high surface area would be a good candidate. Sorbents with highly polar surfaces (e.g., activated alumina, silica gel, and zeolites) would be desirable for a targeted molecule that has a high dipole moment (and high polarizability). If the targeted molecule has a high quadrupole mo-

242

P. Jovanˇci´c · M. Radeti´c

ment, sorbents with surfaces that have high electric ﬁeld gradients are needed. Zeolites are the only such sorbents, as the cations are dispersed above the negatively charged oxides on their surfaces. Cations with high valences (i.e., charges) and small ionic radii would result in strong interactions. Sorbent design/selection is a complex problem, because the process for which the sorbent is used needs to be considered at the same time. For puriﬁcation and particularly ultrapuriﬁcation, strong adsorption bonds are needed. Sorbents that form weak chemical bonds with the targeted molecule can be particularly useful. For kinetic separation, the pore size needs to be precisely tailored to lie between the kinetic diameters of the two molecules that are to be separated. Many microporous molecular sieves with various pore dimensions have been synthesized which could be used as sorbents [7].

2 Textile Wastewaters The major part of the waste generated by the textile industry corresponds to wastewater deriving from the wet processing stages. Textile plants, particularly those involved in dyeing, printing, and ﬁnishing processes are major water consumers and the source of considerable pollution. The average water consumption in textile processes is 160 kg per kg of ﬁnished product [8]. In addition to the direct environmental impact of the wastewater, the large consumption of water resource is becoming intolerable in countries subjected to real or potential water shortages (as the southern part of Europe is). A rational policy in water management would allocate the purest fresh water sources for potable use while encouraging new efﬂuent recycling solutions for industrial consumption. A ﬁnal important environmental impact can be identiﬁed in the large consumption of energy mainly due to the heating of chemical baths and to the drying of fabrics and yarns, which could be substantially reduced by efﬁcient water recycling [9]. Water and chemical consumption is rather important and consequently large volumes of wastewater are generated. These streams contain a wide range of contaminants which must be removed from the textile efﬂuents before their disposal. Organic and inorganic compounds used in the textile processes are discharged in the wastewater at average levels of 80% and 90%, respectively. Table 1 presents a summary of compounds potentially released during the various textile operations. Scouring, dyeing, printing, and ﬁnishing generate the majority of textile wastewater, as they require many rinsing sequences after each step. As an average, 60 to 90% of the total water consumption is for rinsing purposes [10]. Desizing, in some cases combined with scouring, is one of the industry’s largest sources of wastewater pollutants, contributing to a relevant part of the total organic pollution in wastewater [9].

Advanced Sorbent Materials for Treatment of Wastewaters

243

Table 1 Potential speciﬁc pollutants from textile wet processing operations [11] Process

Compounds

Desizing Scouring

Sizes, enzymes, starch, waxes, ammonia Disinfectant and insecticide residues, NaOH, surfactants, soaps, fats, waxes, pectin, oils, sizes, anti-static agents, spent solvents, enzymes H2 O2 , AOX, sodium silicate or organic stabilizer, high pH High pH, NaOH Color, metals, salts, surfactants, organic processing assistants, sulﬁde, acidity/alkalinity, formaldehyde Urea, solvents, color, metals, Resins, waxes, chlorinated compounds, acetate, stearate, spent solvents, softeners

Bleaching Mercerizing Dyeing Printing Finishing

The speciﬁc polluting load of sizing agents is 1–2 grams of COD per gram of size. Natural sizes, such as those based on starch or proteins, are also characterized by a high BOD and a BOD/COD ratio of 0.6–0.7. Synthetic sizing agents such as polyvinyl alcohol or carboxymethyl cellulose, on the contrary, have almost no BOD. The organic load due to desizing can be roughly calculated on the basis of the amount of size applied in the previous phases (the size applied is usually 5 to 20% on a weight basis of the yarn treated). The content of pollutants in scouring efﬂuents depends on the nature and quantity of the impurities present on the ﬁbers and on the intensity of the process itself. High TSS and high organic loads are common in efﬂuents from scouring of natural ﬁbers, due to the removal of dirt, waxes, vegetable matter, etc. Soaps, detergents, alkali, solvents, as well as pesticides may also be present (Table 1). Synthetic ﬁbers usually require milder scouring operations, to remove sizes or oils previously applied. For these ﬁbers, scouring and desizing are frequently combined in a single process. In the bleaching step the most common agents are: hydrogen peroxide, sodium hypochlorite, sodium chlorite, and sulfur dioxide gas. Hydrogen peroxide is by far the most commonly used bleaching agent for cotton and cotton blends, covering over 90% of the bleach used in textile operations and it is typically dosed in caustic solutions. Mercerizing of cotton is used to enhance the following dye application and the luster of ﬁbers. It is carried out with solutions of caustic soda followed by neutralization and several rinses. Alkali consumption may be about 20% of the weight of goods. In Table 2, the average concentration and speciﬁc loads for main operations is shown. From the pollution prevention and reduction point of view, the pollutants of major concern are recalcitrant or hazardous organics, such as dyes or some surfactants, metals, and salts.

244

P. Jovanˇci´c · M. Radeti´c

Table 2 Average concentration and speciﬁc loads for main operations in textile ﬁnishing [11] Fiber

Process

Desizing Scouring or Kiering Bleaching Mercerizing Dyeing Wool Scouring Dyeing Washing Neutralization Bleaching Nylon Scouring Dyeing Acrylic/ Modacrylic Scouring Dyeing Final scour Polyester Scouring Dyeing Final scour Viscose Scouring and dyeing Salt bath Acetate Scouring and dyeing

pH

Cotton

10–13 8.5–9.6 5.5–9.5 5–10 9–14 4.8–8 7.3–10.3 1.9–9 6 10.4 8.4 9.7 1.5–3.7 7.1

8.5 6.8 9.3

BOD (mg l–1 )

TSS (mg l–1 )

1700–5200 50–2900 0–1700 45–65 11–1800 30 000–40 000 380–2200 4000–11 000 28 390 1400 370 2200 170–2000 670 500–800 480–27 000 650 2800 58 2000

16 000–32 000 7600–17 000 2300–14 000 600–1900 500–14 000 1100–64 000 3900–8300 4800–19 000 1200–4800 900 1900 640 1900 830–2000 1200

3300 4900 1800

In the past, highly colored waste streams were considered objectionable primarily for aesthetic reasons [8]. Even though some dyes considered as potentially toxic have been withdrawn from the market and new, more biodegradable molecules have recently been developed, textile dyes are still a major source of water pollution. Of the 700 000 tons of dyes annually produced world-wide, about 10 to 15% is disposed of in efﬂuents after dyeing operations [9]. Wastewaters from textile dyehouses are usually characterized by high COD and BOD, inadequate pH, often high toxicity, and most of all unacceptable coloration. In general, the COD/BOD5 ratio of a textile industry efﬂuent ranges from 3 to 4, meaning that the efﬂuent is moderately biodegradable. The biodegradability of textile wastewater has been increasing during recent years, thanks to substitutions of the chemicals used in the process [12]. In addition to dye molecules, mostly non-biodegradable in aerobic conditions, the persistent organics are: surfactants or their by-products, dyeing auxiliaries such as polyacrylates, phosphonates, sequestering agents (EDTA), synthetic sizes, anti-static, dispersing or ﬁxing agents, preservatives and a large number of ﬁnishing auxiliaries.

Advanced Sorbent Materials for Treatment of Wastewaters

245

Two different approaches are used to classify dyes, the ﬁrst one is based on the nature of the chromophore (the aromatic group absorbing visible light to impart color) while the second follows the mode of application. According to the former criterion, 12 dye classes are usually deﬁned among which the most important group are azo dyes, because of the great extent in number and tonnage of their application. Azo dyes can be used on natural ﬁbers (cotton, silk, wool) and synthetic ﬁbers (polyesters, polyacrylic, rayon, etc.); their molecules include one or more azo groups (– N = N –). The latter classiﬁcation of dyes includes seven classes: acid, basic, direct, disperse, reactive, sulfur, and vat. Overlaps of the two classiﬁcations are possible e.g. azo dyes may belong to the acid, direct, disperse, basic, reactive, and vat dye classes. Signiﬁcant differences in the degree of ﬁxation are reported for the various dye classes. Reactive dyes, which presently represent 20–30% of the total dyes market, are characterized by a low ﬁxation rate, particularly the mono-reactive dyes that represent the major concern about this dye class. It is obvious that the lower the ﬁxation, the higher the residual color and COD discharged in dyeing and rinsing operations. After reactive dyeing operations, as much as 800 mg l–1 of hydrolyzed reactive dye may remain in the bath water [13]. Dyes are always used in combination with other chemicals (acids, alkali, salts, ﬁxing agents, carriers, dispersing agents, surfactants, etc.) that are partly or almost completely discharged in the efﬂuents together with the numerous additives and impurities present in the commercial dye products. Each dye class produces a speciﬁc wastewater. An overview of possible problems related to speciﬁc types of dye is presented in Table 3. Sometimes, a dye class causing environmental problems can be replaced by a dye from another class but we still may see problems derived from auxiliary chemicals used in the dyeing recipes. The concentration of heavy metals in textile mills has decreased in the last decade, mainly because of the reduction of metal contents in the dyes. Metals include copper, cadmium, chromium, nickel, lead, and zinc. Sources of metals in the efﬂuents may be ﬁbers, supply water, dyes, and chemical im-

Table 3 Overview of environmental problems related to the speciﬁc dye class Dye class

Constituent causing problems

Pigments

Acrylic binders Metals Redox agents (metals) Alkali Salt Low exhaust of dyebath Metals

Vat/sulfur dyes Fiber reactive dyes Direct dyes Mordant dyes

High TSS Toxicity Toxicity High pH Salt concern High color Toxicity

246

P. Jovanˇci´c · M. Radeti´c

purities. Dyes may contain metals such as zinc, nickel, chromium, and cobalt, as a functional part of the dye molecule or as impurities. Heavy metal concentrations in dyebath efﬂuents, typically in the range of 1 to 10 mg L–1 , were reviewed by Correia et al. [11]. Several authors have identiﬁed as a potential problem the presence of salts in textile dyeing wastewater. Many salts are either used as raw materials or produced as by-products of neutralization or other reactions in textile wet processes. Salt concentrations in efﬂuent from cotton dyeing may reach 2000 to 3000 ppm and quantities of salts added in dyeing operations range from 20 to 80% of the weight of the goods. Hazardous organic wastes may also result from the use of solvents in some scouring or printing operations while halogenated organic compounds (AOX) may derive from hypochlorite bleaching operations or from spent liquors following shrink-prooﬁng ﬁnishing treatment by chlorine. AOX concentrations up to 100 mg L–1 in bleaching efﬂuents, including considerable amounts of carcinogenic chloroform have been reported [14]. Today, however, bleaching is mostly performed with hydrogen peroxide. Finally, it is worth noting that some reactive dyes are AOX. 2.1 Color Removal As already mentioned, wastewaters from dyeing and ﬁnishing operations in the textile industry are generally colored with high suspended solids and dissolved organics. Many dyes are resistant to fading on light exposure and to different chemicals, primarily due to complex chemical structure and synthetic origin. Residual color is usually due to insoluble dyes that have low biodegradability. In Table 4 is presented the COD/BOD5 ratio of some dyes with poor biodegradability [15]. Azo dyes based on the azo chromogen (– N = N –) are presently the most important compounds constituting about 20–40% of the total dyes used for dyeing [16–18]. The reactive dyes (water-soluble dyes) cannot be easily removed by chemical coagulation-ﬂocculation because they strongly resist biodegradation in an aerobic environment [17, 19, 20]. Therefore, color removal from textile efﬂuents has been the target of great attention in the last few years, because of its potential toxicity and due to its visibility problems [21, 22]. Recent estimates indicate that approximately 12% of synthetic textile dyes used each year is lost during manufacture and processing operations and 20% of these lost dyes enter the environment through efﬂuents that result from the treatment of industrial wastewaters [23]. Some existing technologies (conventional activated sludge treatment, chemical coagulation-ﬂocculation, membrane technologies, oxidative destruction via UV/ozone treatment, photocatalytic degradation, electrochemical reduction etc.) may have certain efﬁciency in the removal of dyes, but

Advanced Sorbent Materials for Treatment of Wastewaters

247

Table 4 The COD/BOD ratios for some dyes [15] Dye, C.I generic name

COD/BOD5 ratio

Direct Blue 80 Disperse Red 68 Reactive Blue 21 Vat Violet 21

17.7 7.0 59.0 10.8

their initial and operational costs are too high, making them often economically unacceptable in dyeing and ﬁnishing plants [24–26]. A 40 to 50% color removal could be anticipated due to biodegradation and adsorption of the dyes on ﬂocculated sludge. COD removal could reach about 70%. Physicalchemical processes alone do not provide satisfactory results because the color removal of insoluble dyes is insufﬁcient whereas the COD removal is about 50%. A combination of biological and physico-chemical processes will remove more than 85% of the COD. This results in an efﬂuent containing COD refractory to these conventional processes between 150 and 300 mg L–1 , low BOD5 and usually high color [15]. Therefore, despite the existence of a wide range of wastewater treatment techniques, there is no single process capable of adequate treatment for these efﬂuents [27]. Thus, the best solution for textile wastewater treatment appears to be a combination of different techniques. The conventional association of biological and physico-chemical processes results in the removal of most of the organic pollutants from textile wastewaters. However, the resulting efﬂuent is still fairly colored. Ideally, a ﬁnal reﬁning treatment is needed to remove the color [15]. This ﬁnal step in the treatment of textile efﬂuents is indispensable for a potential water reuse, since total color removal and extensive elimination of organic pollutants are required for water recycling in the textile industry [28]. The high afﬁnity of dyes to several sorbent materials makes possible the efﬁcient removal of many dyes and the production of high quality efﬂuents. In this context, adsorption techniques, which have been extensively used in industrial processes for a variety of separation and puriﬁcation purposes, appear as sufﬁciently efﬁcient alternatives for the removal of dyes in textile wastewater. Efﬁciency of color removal by adsorption is inﬂuenced by many physico-chemical factors, such as dye/adsorbent interaction, adsorbent surface area, particle size, temperature, pH, and contact time. 2.1.1 Activated Carbon The most commonly studied and commercially used sorbent for color removal is activated carbon (AC), because of its capability to efﬁciently adsorb

248

P. Jovanˇci´c · M. Radeti´c

a broad range of different types of adsorbates [29–34]. Yet, little attention has been given to the role of the AC surface chemistry in the adsorption process. A thorough knowledge of AC surface chemistry enables the preparation of adsorbents with appropriate characteristics for speciﬁc applications [35]. Karanﬁl et al. [36] evaluated the uptake of two organic contaminants on a set of commercial and treated ACs in order to study the inﬂuence of the surface chemistry on the adsorption of those organic pollutants. Activated carbons have a high degree of porosity and an extensive surface area. The adsorption capacity of an AC is determined not only by the textural properties but also by the chemical nature of its surface, i.e., the amount and nature of oxygen-containing functional groups [35]. The nature of an activated carbon surface can be changed by different treatments, which include liquid-phase oxidations with HNO3 and H2 O2 and gas-phase oxidations with O2 or N2 O, as well as thermal treatments at high temperatures [37]. Generally, the performance depends on the type of carbon used and on the characteristics of the wastewater. It can be used either as granular activated carbon (GAC) or as powdered activated carbon (PAC). The former is more expensive and requires ad hoc reactors, while the latter is cheaper and may be directly added in a biological (activated sludge) aeration tank. Moreover, if the granular bed is packed and back wash is required, a mechanically resistant GAC must be used otherwise losses may be quite important. Some authors began to study the effect of activated carbon surface chemistry on the removal of dyes from textile efﬂuents. Al-Degs et al. [38] studied the adsorption of three reactive dyes on a Filtrasorb 400 activated carbon, and attributed the high adsorption capacity of this adsorbent to its net positive surface charge during the adsorption process. In a previous work, Pereira et al. [39] followed a different approach. Starting from the same material, a set of activated carbons differing in their chemical surface properties was prepared. In the mentioned work, a screening test was carried out in order to evaluate the performance of the treated ACs in the removal of a series of dyes. It was concluded that basic carbons were the most efﬁcient for the removal of both cationic and anionic dyes. In another study, Faria et al. [40] showed that the adsorption of different anionic and cationic dyes (basic, reactive, and acid dyes) strongly depend on the surface chemistry of the activated carbon. Starting material used was a GAC being oxidatively and thermally modiﬁed. It was pointed out that it is possible to tailor the surface of activated carbon by introducing new surface groups, in order to optimize its adsorption capacity towards different dye molecules. Thermal treatment with H2 at 700 ◦ C of activated carbon is proven to be the most efﬁcient for removal of both anionic and cationic dyes investigated. Marmagne et al. [15] used PAC at a single concentration in a laboratory ﬂocculator in an attempt to remove the color of a variety of dyes as possible constituents of textile efﬂuents. The PAC is left to act for a few minutes in

Advanced Sorbent Materials for Treatment of Wastewaters

249

an agitated reactor, and the supernatant is analyzed following settling and ﬁltration. The results obtained are presented in Table 5. High removal rates (over 90%) are achieved using PAC for cationic, mordant, and acid dyes. For direct, dispersed, and reactive dyes, efﬁciency is moderate (over 40%), or can be improved using massive dosages of activated carbon. For vat dyes, color removal is very low (under 20%). The decisive parameter seems to be the class of dye rather than its chemical structure. Therefore, the treatment with activated carbon could be combined with a physico-chemical or biological treatment of wastewater containing speciﬁc dyes. The pore size distributions of the activated carbons signiﬁcantly inﬂuence the adsorption capacity of natural organic materials subject to their molecular sizes [41]. Several researchers studying dye adsorption on activated carbon [42, 43] indicated that the presence of mesopores together with micropores in the activated carbon enhances their adsorption capacities especially for large adsorbates [44, 45]. These reports suggest that the pore size distributions of activated carbons determine their potential applications. However, it should be noted that not only porosity but also surface chemistry of an activated carbon affect adsorption [38]. Production of activated carbons from solid wastes is one of the most environmentally friendly solutions by transforming negative-valued wastes to valuable chemicals. There are several reports on the production of activated Table 5 Efﬁciency of PAC treatment for color removal [15] Dye, C.I generic name

Structure

COD removal (%)

Color removal (%)

Acid Blue 142 Acid Blue 113 Acid Blue 260 Basic Blue 411 Basic Yellow 13 Basic Blue 3 Vat Blue 4 Vat Green 1 Direct Blue 199 Direct Red 89 Disperse Blue 56 Disperse Yellow 235 Acid Brown 298 Acid Black 142 Reactive Blue 204 Reactive Blue 209 Reactive Red 184 Reactive Blue 41 Reactive Blue 49

Triphenylmethane Azoic Anthraquinone Azoic Methine Oxazine Antraquinone Antraquinone Phtalocyanique Azoic Antraquinone Azoic Azoic Azoic Oxazine Formazan Azoic Phtalocyanine Anthraquinone

71.4 70.9 58.8 84.7 81.5 67.8

93.7 95.6 86.8 100.0 99.7 98.5 10.1 12.9 71.7 59.0 30.2 83.4 97.1 98.1 69.0 78.5 77.6 57.4 94.6

25.6 45.5 59.3 91.5 82.1 70.6 89.8 69.4 74.6 19.2

250

P. Jovanˇci´c · M. Radeti´c

carbons from various city wastes [46–48]. These wastes have been considered to be unfavorable raw materials for activated carbons because their composition is variable and inhomogeneous. However, these city wastes relatively rich in carbon are useful for making good activated carbons that have sufﬁcient surface area. For instance, it is reported that activated carbons prepared from biosludge were effective as adsorbents for wastewater treatment [49]. The authors have not only demonstrated that activated carbons having high surface area could be produced from various waste materials but also conﬁrmed that the pre-treatment method (i.e., mixture of raw materials with a metal salt, carbonization and acid treatment prior to steam-activation) could be adopted to effectively enhance the formation of mesopores on some waste-derived activated carbons [50–55]. The activated carbons derived from solid wastes (PET bottles, tires, refuse derived fuel and waste generated during lactic acid fermentation from garbage) depending on the preparation procedures possess different pore size distributions. Highly mesoporous activated carbons are obtained from waste PET bottles and waste tires exhibiting large adsorption capacities for the reactive dye investigated [56]. Moreover, it was concluded that activated carbons obtained from solid waste have adsorption capacities for phenol similar to commercial activated carbon. 2.1.2 Zeolites Natural zeolites are low-cost aluminosilicates, with a cage-like structure suitable for ion exchange due to isomorphous replacement of Al3+ with Si4+ in the structure, giving rise to a deﬁciency of positive charge in the framework. This is balanced by mono and divalent exchangeable cations such as Na+ , Ca2+ , K+ and Mg2+ [57]. These cations are coordinated with the deﬁned number of water molecules, and located on speciﬁc sites in framework channels. These cations can be exchanged with organic and inorganic cations [58, 59]. Such zeolite properties have been utilized for a variety of purposes. Sorption on zeolitic particles is a complex process because of their porous structure, inner and outer charged surfaces, mineralogical heterogeneity, existence of crystal edges, broken bonds, and other imperfections on the surface [60]. The typical unit cell formula of natural zeolite mineral, clinoptilolite, is Na6 [(AlO2 )6 (SiO2 )30 ]·24H2 O [61]. Clinoptilolite is probably the most abundant zeolite in nature because of its wide geographic distribution and large size of deposits. The presence of 4.5 million tons of natural zeolites of high quality, mainly those of clinoptilolite in Turkey, created an impetus for the utilization of clinoptilolite in wastewater treatment. Natural zeolites with 70–80% clinoptilolite content are often used in technological applications. Natural deposits of zeolite with a relatively low content of clinoptilolite (50%) indicate higher heavy metal uptake capacity due to the participation of the adsorption in the overall process.

Advanced Sorbent Materials for Treatment of Wastewaters

251

Still, there are only a few studies on color removal in textile efﬂuents with zeolites [34, 62]. Benkli et al. [63] investigated the modiﬁcation of the zeolite (clinoptilolite) surface with a quaternary amine, hexadecyl trimethyl ammonium bromide (HTAB), in a ﬁxed-bed reactor, to improve the removal efﬁciency of three reactive azo dyes. The reactive dyes investigated have negatively charged sulfonate groups which are repelled by the negatively charged zeolite surface. As a result, a relatively low adsorption capacity of natural zeolite for these dyes could be anticipated. To increase the adsorption capacity, the surface of natural zeolite is modiﬁed with HTAB. In Fig. 1 is shown a comparative analysis of adsorption of three reactive dyes onto the zeolite modiﬁed under optimum conditions (3 g L–1 of HTAB dosage and a ﬂow rate of 0.025 L min–1 ). As can be seen, the order of dye removal by modiﬁed zeolite is: C.I. Reactive Black 5 > C.I. Reactive Yellow 176 > C.I. Reactive Red 239. The possible mechanism of dye adsorption is schematically illustrated in Fig. 2. Natural zeolite has a negative surface charge in the entire range of pH thus, preventing the adsorption of negatively charged dye molecules (sulfonate groups). However, the modiﬁed zeolite exhibits a positive surface charge due to the adsorption of cationic HTAB molecules. Consequently, the adsorption of negatively charged dye molecules on HTAB-modiﬁed zeolite occurs via electrostatic attraction. The uptake of dye molecules largely depends on the packing and conﬁguration of HTAB molecules. The surface coverage for HTAB was calculated and it was proposed that HTAB forms a bi-

Fig. 1 Comparative analysis of the adsorption of reactive dyes onto modiﬁed zeolite [63]

252

P. Jovanˇci´c · M. Radeti´c

Fig. 2 A schematic illustration of the adsorption of reactive anionic dyes onto the HTAB modiﬁed zeolite surface [63]

layer on the zeolite surface in the case of C.I. Reactive Yellow 176 (Fig. 2). For the other two reactive dyes investigated the forming of a monolayer of HTAB is more probable. 2.1.3 Low-Cost Sorbents As already commented, the use of activated carbon is limited, because of its high cost. Effective and alternative sorbents, more economically attractive than GAC, have been investigated for color removal on synthetic wastewater [64]. Several researchers have been studying the use of alternative materials, which, although less efﬁcient, are less expensive. Among the studied alternative materials are agricultural, forest, animal and several low-cost industrial by-products such as peat, wood, tree barks, chitin, silica gel, bauxite, bentonite clay, certain synthetic polymeric adsorbents—cucurbituril, etc. [65–72]. There is also a series of similar studies of color removal from wastewaters by the use of a wide range of low-cost sorbents derived from biological matter—biomass [73] such as shale oil ash [74], chitosan [75–77], sunﬂower stalk [78], and natural clay [69, 79]. E.Voudrias et al. [80] compared the efﬁciency of three low-cost sorbents (bentonite, ﬂy ash, and bleaching earth) against the powdered activated carbon for the removal of color from aqueous solutions containing various reactive dyes and real wastewaters. The results of removal of color of real wastewater by low-cost sorbent and activated carbon are shown in Fig. 3.

Advanced Sorbent Materials for Treatment of Wastewaters

253

Fig. 3 The removal of color of the real wastewater by using different low-cost sorbents [80]

Although the activated carbon was conﬁrmed as the most effective sorbent in this study, the low-cost sorbents could be used for color removal. The proper sorbent selection is not only the matter of sorption capacity but also it is necessary to take into account the reusability, waste management, and total cost of the treatment. Therefore, the low-cost sorbent with low sorption capacity could be, in special circumstances, the sorbent of choice compared to the expensive activated carbon. It has been known for decades that wool ﬁbers can efﬁciently bind different dyes [81–87]. However, raw wool is too expensive for such purposes, but recent studies indicated that wool as a recycled material can be used for removal of a wide range of textile dyes from water [88–99]. Recycled wool-based non-woven material can be used as an efﬁcient sorbent for color removal of dye solutions. It has shown good sorption properties for acid dyes [92–94]. Its treatment with natural biopolymer chitosan (CHT) and low-temperature plasma (LTP) treatments caused signiﬁcant increase in dye uptake, which was more pronounced at higher temperatures and lower pH due to increased protonation of amino groups. Sorption kinetics for direct, basic, reactive, and metal complex dyes is shown in Fig. 4. Obviously, after the rapid sorption in the ﬁrst 3 hours, the process slowed down and equilibrium was reached within 24 hours only in some cases [97]. Treatment with hydrogen peroxide (H2 O2 ) and particularly CHT treatment signiﬁcantly improved the uptake of direct dye, which can be attributed to the existence of new amino groups on the wool ﬁber surface originating from chitosan, thereby increasing the positive zeta-potential of the ﬁber surface [91, 99]. Amino groups in acidic conditions are protonated and ionic interaction between sulfonate groups of direct dyes and protonated amino groups of wool are expected. Higher basic dye uptake on H2 O2 treated material can be due to wool oxidation and formation of appropriate functional

254

P. Jovanˇci´c · M. Radeti´c

Fig. 4 Sorption kinetics for a direct dye Tubantin blue GLL; b basic dye Bezacryl gold yellow GL; c reactive dye Lanaset yellow 4GN; d metal complex dye Lanaset red 2B, (C0 = 100 g L–1 , T = 20 ◦ C, pH 3.30) [97]

groups [100, 101]. It is well established that binding of basic dye to wool is carried out via carboxylic groups. Electrostatic attraction between deprotonated negatively charged carboxylic groups of wool and positively charged cationic basic dyes could be established at operated pH [102]. Each modiﬁcation of sorbent brought about an increase in the uptake of reactive dye. In the ﬁrst few hours, the CHT-treated sample showed higher uptakes in comparison with H2 O2 -treated material, but the sorption decreased within the last 12 hours, causing almost the same ﬁnal uptake for both samples. Metal complex dyes are conventionally used for dyeing of wool and good sorption of these dyes was anticipated. Uptake of metal complex dyes on CHT- and H2 O2 -treated material was higher compared to untreated sample. Again, CHT treatment contributed to an increase of amino groups on the ﬁber surface, thus providing the potential sites for ionic interaction with the sulfonate groups of dyes.

Advanced Sorbent Materials for Treatment of Wastewaters

255

2.2 Removal of Heavy Metal Ions 2.2.1 Low-Cost Sorbents In addition to many harmful and toxic substances present in industrial efﬂuents, high concentrations of metal cations is of great concern from an environmental point of view. A wide range of industries (mining, metal processing, electroplating, electronics, chemical processing, textile etc.) release into the environment various metals in amounts that can pose a risk to human health. It is known that the heavy metal ions can accumulate in living species with a permanent toxic and carcinogenic effect [103–105]. Therefore, before being released in the environment, the treatment of the contaminated wastewaters is of great importance. The complexity of efﬂuents makes the process of heavy metals removal more difﬁcult, as well as strict limitations that have been imposed to wastewater discharge everywhere in aquatic recipients. The most common treatment processes used include chemical precipitation, oxidation/reduction, ultraﬁltration (UF), reverse osmosis (RO) and electrodialysis, solvent extraction, ion exchange etc. Each process presents its own advantages and disadvantages, but in the past few years, extensive research has been undertaken to develop alternative and economic adsorbents. The optimization of wastewater puriﬁcation processes is required as a consequence of development of new adsorbents based on lowcost raw materials with high pollutant-removal efﬁciency. Among the mentioned low-cost raw materials, natural zeolites seem to be an attractive adsorbent for removal of heavy metals [106, 107]. Natural zeolites have been intensively studied recently because of their applicability in removing trace quantities of heavy metal ions from aqueous solutions using the ion exchange phenomenon [108, 109]. Attempts have been made to develop models for adsorption of metal ions on heterogeneous zeolitic surfaces, particularly in the case of competitive adsorption. The Langmuir and Freundlich adsorption models are widely used because they are convenient to describe experimental results in a wide range of concentrations. The mathematical correlation of the Langmuir isotherm has provided a basis for development of other models, which include similarity of adsorption and ion exchange, Gaussian energy distribution, degree of surface heterogeneity, and high solute concentrations range [60]. Since heavy metals such as zinc, lead, and copper are prior toxic pollutants in industrial wastewater, and they also become common groundwater contaminants, J.Peri´c et al. [110] investigated their uptake from aqueous solutions by ion exchange on natural zeolite originating from Croatia. The uptake of each of the heavy metal ions as a function of the initial concentration is presented in Fig. 5.

256

P. Jovanˇci´c · M. Radeti´c

Fig. 5 The removal efﬁciency of Zn2+ , Cu2+ , and Pb2+ ions depending on their initial concentration [110]

The removal efﬁciency follows the sequence Pb2+ ≥ Cu2+ > Zn2+ and the uptake of these heavy metal ions from aqueous solutions on natural zeolite is a complex process consisting mainly of ion exchange and adsorption although precipitation at a higher initial concentration of metal ions could occur [111]. A similar sequence of removal efﬁciency for lead, copper, and zinc was obtained by M.Radeti´c et al. [112] using a recycled wool-based non-woven material as a low-cost adsorbent for heavy metal ions present in textile efﬂuents. Metal cation uptakes obtained after 24 h of sorption using this type of sorbent are presented in Table 6.

Table 6 Metal cation uptake after 24 h of sorption (T = 20 ◦ C, pH0 = 4.50, C0 = 100 mg L–1 ) [99]

untreated LTP CHT A∗ LTP + CHT A CHT B∗∗ LTP + CHT B H2 O2 H2 O2 + CHT B ∗ , ∗∗

q, mg/g Pb2+

Cu2+

Zn2+

Co2+

4.39 4.33 4.78 4.88 4.34 4.29 5.00 5.00

2.04 2.15 2.93 3.00 2.31 2.29 4.80 2.70

1.27 1.01 1.19 1.05 1.58 1.62 3.84 1.56

0.96 0.93 1.95 1.87 0.88 0.89 3.21 1.47

Different procedures of sorbent treatment with biopolymer chitosan [91, 112]

Advanced Sorbent Materials for Treatment of Wastewaters

257

The sorbent exhibited satisfactory sorption behavior and good selectivity even as an untreated material. In order to improve its sorption properties, the material was modiﬁed by low-temperature air plasma (LTP), biopolymer chitosan (CHT), and hydrogen peroxide (H2 O2 ). LTP-treated samples exhibited similar behavior as untreated samples whereas chitosan and particularly hydrogen peroxide treatments led to a signiﬁcant increase in uptake of all investigated metal cations [86, 88, 89]. The results of sorption kinetics demonstrated that high amounts of cations had already been removed within 3 h, depending on the sorbent treatment. Although the increase in initial concentration brought about a rise of metal cation uptake, the percentage of metal cations adsorbed by untreated and differently treated materials decreased, indicating that higher amounts of metal cations are left in the solution [87]. Temperature showed a strong inﬂuence on the sorption of all investigated metal cations: the higher the temperature, the higher the metal cation uptake [91, 112]. The inﬂuence of ﬁnal pH value (after 1 hour of sorption) on Pb2+ , Cu2+ , Zn2+ , and Co2+ ion uptake for differently treated samples is presented in Fig. 6 [112]. At lower pH the material is positively charged, i.e. end and side amino groups of the wool protein chains are protonated. However, at higher pH carboxylate groups that are the main sites for binding of metal cations [81, 84] existed as anions and the wool became negatively charged, causing the enhanced metal cation sorption. The afﬁnity of recycled wool to bind metal cations is in good correlation with literature data for raw wool at pH 5.00. Metal cation uptake in both cases increases in the following order: Pb2+ > Cu2+ > Zn2+ > Co2+ [85, 86]. The rise of pH during the sorption of metal cations is suggested to be due to the sorption of metal cations in the form of Me2+ , but the parallel release and uptake of protons should not be neglected. In other words, two competitive processes govern pH changes: the sorption of metal cations and release and/or uptake of protons [112]. Recently, biosorption has emerged as a potential and cost-effective option for heavy metal removal from aqueous solutions. Agricultural by-products as biosorbents for heavy metals also offer a potential alternative to existing techniques and as yet, have not been the subject of extensive studies. Agricultural biosorbents including soybean hulls, peanut hulls, almond hulls, cottonseed hulls, and corncobs have been demonstrated to remove heavy metal ions [113–116]. The use of algae has been reviewed by Volesky [117]. However, more efforts have to be made in the application of bacteria and/or engineered bioﬁlm as low-cost adsorbents in metal-removal processes. Among the different types of process conﬁgurations, batch reactor or ﬁxed-bed reactors have been widely investigated [118].

258

P. Jovanˇci´c · M. Radeti´c

Advanced Sorbent Materials for Treatment of Wastewaters

259

Fig. 6 Final pH vs. uptake of: a Pb2+ ions, b Cu2+ ions, c Zn2+ ions, d Co2+ ions (T = 20 ◦ C, C0 = 100 mg L–1 , t = 1 h) untreated; LTP; ◦ CHT A; • LTP + CHT A; CHT B; LTP +CHT B; ♦ H2 O2 ; H2 O2 + CHT B [13]

2.2.2 Nanosorbents Advances in nanoscale science and engineering suggest that many of the current problems involving water quality could be resolved or greatly minimized using nanosorbents and other materials, products, and processes resulting from the development of nanotechnology [3]. Nanomaterials have a number of key physico-chemical properties that make them particularly attractive as separation media for water puriﬁcation. Nanomaterials have much larger surface area than bulk particles and can be functionalized with various chemical groups to increase their afﬁnity towards target compounds. Several research groups are exploiting the unique properties of nanoparticles to develop high capacity and selective sorbents for metal ions and anions. Li et al. [119] have investigated the sorption of Pb2+ , Cu2+ , and Cd2+ onto multiwalled carbon nanotubes. The maximum sorption capacities of 97.08 mg g–1 for Pb2+ , 24.49 mg g–1 for Cu2+ , and 10.86 mg g–1 for Cd2+ were obtained at room temperature. The metal ion sorption capacities of the multiwalled carbon nanotubes were 3–4 times larger than those of powder activated carbon and granular activated carbon, two commonly used sorbents in water puriﬁcation. Qi et al. [120] have evaluated the sorption of Pb2+ onto chitosan nanoparticles (40–100 nm) prepared by ionic gelation of chitosan and tripolyphosphate [120]. The phosphate-functionalized chitosan nanoparticles have a maximum Pb2+ sorption capacity of 398 mg g–1 . Peng et al. [121] have recently developed a novel sorbent with high surface area (189 m2 g–1 ) consisting of cerium oxide supported on carbon nanotubes as an effective sorbent for As(V). Mangun et al. [122] have synthesized nanoporous activated carbon ﬁbers with an average pore size of 1.16 nm and surface area ranking from 171 to 483 m2 g–1 thus serving as a high capacity and selective sorbent for organic solutes in aqueous solutions. This sorbent had much higher organic sorption capacity than granular activated carbon [122]. Dentritic polymers (random hyper-branched polymers, dendrigraft polymers, dendrons, and dendrimers) exhibit many features that make them particularly attractive as functional materials for water puriﬁcation. These nanoparticles with sizes in the range of 1–20 nm can be used as recyclable water-soluble ligands for toxic metal ions, radionuclides, and inorganic ions [123–125]. Also, dendritic polymers can be used as unimolecular micelles for recovering organic solutes from water [126] and as delivery vehicles for antimicrobial agents such as Ag1+ and quaternary ammonium chlorides [127, 128].

260

P. Jovanˇci´c · M. Radeti´c

Diallo et al. [129] developed a dendrimer-enhanced ultraﬁltration process for recovery of metal ions from aqueous solutions. It was documented that dentritic polymers improve the performance of RO and NF membranes thus permitting better efﬁcacy in removing of dissolved organic and inorganic solute with small molar mass. Therefore, with dentritic polymers it will be possible to develop effective UF processes for puriﬁcation of water contaminated by toxic metal ions, radionuclides, organic and inorganic solutes, bacteria, and viruses [3].

3 Conclusions In spite of a wide range of other wastewater treatment technologies, adsorption (adsorptive separation) is still a key separation tool that is used in many industries. Future application of adsorption is limited by the availability of new and more efﬁcient sorbents. The sorbents should be tailored in such a way to meet the requirements of each speciﬁc application. Despite many research efforts and patents on adsorption, there are only a few commercially available sorbents that are used in the current adsorption processes (activated carbon, zeolites, silica gel, and activated alumina). Sorption processes are ideally suited for puriﬁcation of textile wastewaters which are characterized by high COD and BOD, inadequate pH, often high toxicity and most of all unacceptable coloration. In addition to harmful and toxic substances present in industrial efﬂuents, high concentration of metal cations is of great concern from an environmental point of view. The potential of low-cost or cost-effective sorbents for color removal and removal of heavy metal ions from textile efﬂuents compared to the conventionally used sorbents such as activated carbon and zeolites is clearly documented. Advances in nano-scale science and engineering are providing great opportunities to develop more cost effective and environmentally acceptable water puriﬁcation processes. Nanosorbents will certainly become critical components of industrial wastewater and public water puriﬁcation systems as more progress is made towards the synthesis of cost-effective and environmentally acceptable functional materials. The potential use of the adsorption properties of new forms of carbon (carbon molecular sieves, super-activated carbon, activated carbon ﬁbers, carbon nanotubes, and graphite nanoﬁbers) and nanoporous materials, i.e. nanomaterials (metalcontaining nanoparticles, dendrimers etc.) is in a process of intensive exploration. Acknowledgements We greatly acknowledge the support from European Community FP6 programme through ﬁnancing the EMCO project INCO CT 2004/509188 and Ministry of Science and Environmental Protection of the Republic of Serbia for project TD-7017B.

Advanced Sorbent Materials for Treatment of Wastewaters

261

References 1. King CJ (ed) (1980) Separation Processes, 2nd edn. McGraw-Hill, New York 2. Yang RT (ed) (2003) Adsorbents: Fundamentals and Applications. Wiley, Hoboken, New Jersey 3. Sovage N, Diallo MS (2005) J Nanopart Res 7:331 4. Jankowska H, Swiatkowski A, Choma J (1991) Active Carbon. Ellis Harwood, New York 5. Milton RM (1959) US Patent 2 882 243 6. Milton RM (1959) US Patent 2 882 244 7. Hartman M, Kevan L (1999) Chem Rev 99:935 8. EPA (1996) Best Management Practices for Pollution Prevention in the Textile Industry. Manual EPA/625/R-96/004, Cincinnati 9. Freeman HM (ed) (1995) Industrial Pollution Prevention Handbook. McGraw-Hill Inc., New York 10. Groves GR, Buckley CA, Turnbull RH (1979) J Water Pollution Control Federation, March 1979:499 11. Correia VM, Stephenson T, Judd SJ (1994) Environ Technol 15:917 12. Vandevivere PC, Bianchi R, Verstaete W (1998) J Chem Technol Biotechnol 72:289 13. Lens P, Hulshoff PL, Wilderer P, Asano T (eds) (2002) Water Recycling and Resource Recovery in Industry. IWA Publishing, London 14. Steiner N (1995) Text Chem Col 27:29 15. Marmagne O, Coste C (1996) Am Dyestuff Reporter 85:15 16. Zollinger H (1991) Color Chemistry, 2nd edn. VCH Publishers, New York 17. Wu J, Eitman MA, Law SE (1998) J Environ Eng 12:272 18. Beidilli MI, Pavlostathis SG, Tincher WC (2000) Water Environ Res 72:698 19. Pagga U, Brown D (1986) Chemosphere 15:479 20. Panswad T, Luangdilok W (2000) Water Res 17:4177 21. Yu-Li Y-R, Thomas A (1995) J Chem Tech Biotechnol 63:55 22. Morais L, Freitas O, Gancolves E, Vaskancelos L, Gonzales C (1999) Water Res 33:979 23. Hwang M, Chen K (1993) J Appl Polym Sci 49:975 24. Kang S, Liao C, Po S (2000) Chemosphere 41:1287 25. Churchley J (1998) Ozone-Sci Eng 20:111 26. Ogutveren U, Kaparal S (1994) J Environ Sci Health A 29:1 27. Cooper P (1993) J Soc Dyers Col 109:97 28. Rott U, Minke R (1999) Water Sci Technol 40:137 29. Walker GM, Weatherley LR (1997) Water Res 31:2093 30. Walker GM, Weatherley LR (2001) Chem Eng J 84:125 31. McKay G (1981) J Chem Technol Biotechnol 31:717 32. McKay G (1982) J Chem Technol Biotechnol 32:731 33. Blum DJW, Suffet IH, Duguet JP (1993) Crit Rev Environ Sci Technol 23:121 34. Meshko V, Markovska L, Mincheva M, Rodrigues AE (2001) Water Res 35:3357 35. Bansal RC, Donnet JB, Stoeckli F (1988) Active Carbon. Marcel Decker, New York 36. Karanﬁl T, Kilduff JE (1999) Environ Sci Technol 33:3217 37. Figueiredo JL, Pereira MFR, Freitas MMA, Órfão JJM (1999) Carbon 37:1379 38. Al-Degs Y, Khraisheh MAM, Allen SJ, Ahmad MN (2000) Water Res 34:927 39. Pereira MFR, Soares SF, Órfão JJM, Figueiredo JL (2003) Carbon 41:811 40. Faria PCC, Órfão JJM, Pereira MFR (2004) Water Res 38:2043 41. Newcombe G, Drikas M, Hayes R (1997) Water Res 31:1065 42. Lin SHJ (1993) J Chem Technol Biotechnol 57:387

262

P. Jovanˇci´c · M. Radeti´c

43. Sankar M, Sekaran G, Sadulla S, Ramasami T (1999) J Chem Technol Biotechnol 74:337 44. Hsieh C, Teng H (2000) Carbon 38:863 45. Tamai H, Kakii T, Hirota Y, Kumamoto T, Yasuda H (1996) Chem Mater 8:454 46. Bota A, Laszlo K, Nazy LG, Subklew G, Schlimper H, Schwuger MJ (1996) Adsorption 3:81 47. Cunliffe AM, Williams PT (1998) Environ Technol 19:1177 48. Shimada M, Hamabe H, Iida T, Kawarada K, Okayama T (1999) J Porous Mater 6:191 49. Pikkov L, Kallas J, Ruutman T, Rikmann E (2001) Environ Technol 22:229 50. Ariyadejwanich P, Tanthapanichakoon W, Nakagawa K, Mukai SR, Tamon H (2002) Carbon 41:157 51. Nagano S, Tamon H, Azumi T, Nakagawa K, Suzuki T (2000) Carbon 38:915 52. Nakagawa K, Fuke A, Tamon H, Suzuki T, Nagano S (2001) Jpn J Food Eng 2:141 53. Nakagawa K, Tamon H, Suzuki T, Nagano S (2002) J Porous Mater 9:25 54. Tamon H, Nakagawa K, Suzuki T, Nagano S (1999) Carbon 37:1643 55. Tamon H, Nakagawa K, Mukai SR (2002) In: Proc 6th Int Symp Separation Technology between Japan and Korea, Tokyo, Japan, p 481 56. Nakagawa K, Namba A, Mukai SR, Tamon H, Ariyadejwanich P, Tanthapanichakoon W (2004) Water Res 38:1791 57. Haggerty GM, Bowman RS (1994) Environ Sci Technol 28:452 58. Ackley MW, Yang RT (1991) AIChE J 37:1645 59. Ersoy B, C ¸ elik MS (2004) Environ Technol 25:341 60. Altin O, Oybelge Ho, Dogu T (1998) J Colloid Interface Sci 198:130 61. Breck DW (1974) Zeolite Molecular Sieves. Wiley, New York 62. Armaˇ gan B, Özdemir O, Turan M, C ¸ elik MS (2003) J Chem Technol Biotechnol 78:725 63. Benkli YE, Can MF, Turan M, C ¸ elik MS (2005) Water Res 39:487 64. Ramakrishna KR, Viraraghavan T (1997) Wat Sci Tech 36:189 65. McKay G, Otterburn M, Aga J (1985) Mater Air Soil Pollut 24:307 66. Fytianos K, Vasilikiotis G, Agelidis T, Kofos P, Stefanidis K (1985) Chemosphere 14:411 67. Allen S, Khader K, McKay G (1989) J Chem Tech Biotechnol 45:291 68. Poots V, Mckay G, Healy J (1976) Water Res 10:1067 69. El Geundi M (1991) Water Res 25:271 70. McKay G, Blair HS, Gardner J (1982) J Appl Polym Sci 27:4251 71. Robinson T, McMullan G, Marchant R, Nigam P (2001) Biores Technol 77:24 72. Bousher A, Shen X, Edyvean RGJ (1997) Water Res 31:2087 73. Joseph AL (1994) Am Dyest Rep 83:17 74. AI-Qodah Z (2000) Water Res 34:4295 75. Juang RS, Tseng RL, Wu FC, Lee SH (1997) J Chem Technol Biotechnol 70:391 76. Bhavani KD, Dutta PK (1999) Am Dyest Rep 88:53 77. Smith B, Koonce T, Hudson S (1993) Am Dyest Rep 82:18 78. Sun G, Xu X (1997) Ind Eng Chem Res 36:808 79. Rytwo G, Nir S, Crepsin M, Margulies L (2000) J Colloid Interface Sci 222:12 80. Voudrias E, Fytianos K, Bozani E (2002) Global Nest: the Int J 4:75 81. Hemrajani SN, Narwani CS (1967) J Indian Chem Soc 44:704 82. Hartley FR (1968) Aust J Chem 21:1013 83. Guthrie RE, Laurie SH (1968) Aust J Chem 21:2437 84. Kokot S, Feughelman M (1972) Textile Res J 42:704 85. Masri MS, Reuter FW, Friedman M (1974) J Appl Polym Sci 18:675

Advanced Sorbent Materials for Treatment of Wastewaters

263

86. Weltrowski M, Patry J, Beaudoin B (1995) Wool based ﬁlter for adsorption of heavy metals and textile dyes. In: Bona M (ed) Proc 9th Int Wool Text Res Conf, vol 4. Città degli Studi, Biella, p 343 87. Shefﬁeld A, Doyle M, Wool M (2005) Textile Res J 75:203 88. Radeti´c M, Joci´c D, Jovanˇci´c P, Rajakovi´c Lj, Potkonjak B, Petrovi´c ZLj (2001) 40 Serbian Chem Soc Meeting. Novi Sad, Serbia 89. Radeti´c M, Joci´c D, Jovanˇci´c P, Rajakovi´c Lj, Potkonjak B, Petrovi´c ZLj (2001) 1st AUTEX Conference on Technical Textiles: Designing Textiles for Technical Applications. Povoa de Varzim, Portugal 90. Radeti´c M, Joci´c D, Jovanˇci´c P, Rajakovi´c Lj, Thomas H, Petrovi´c ZLj (2002) 21st Summer School and International Symposium on the Physics of Ionized Gases. Soko Banja, Yugoslavia 91. Radeti´c M, Joci´c D, Jovanˇci´c P, Rajakovi´c Lj, Thomas H, Petrovi´c ZLj (2003) J Appl Polym Sci 90:379 92. Radeti´c M, Joci´c D, Jovanˇci´c P, Petrovi´c ZLj, Thomas H (2003) 3rd AUTEX Conference. Gdansk, Poland 93. Radeti´c M, Joci´c D, Jovanˇci´c P, Rajakovi´c Lj, Petrovi´c ZLj, Thomas H, Phan KH (2003) DWI Reports, 126:274 94. Radeti´c M, Joci´c D, Jovanˇci´c P, Petrovi´c ZLj, Thomas H (2005) Ind J Fibre Textile Res 29:82 95. Ðordevi´c N, Radeti´c M, Joci´c D Potkonjak B, Puaˇc N, Jovanˇci´c P (2005) 1st EMCO Workshop on Analysis and Removal of Contaminants from Wastewaters for the Implementation of the Water Framework Directive. Dubrovnik, Croatia 96. Ðordevi´c N, Aleksi´c M, Radeti´c M, Joci´c D, Jovanˇci´c P (2005) 1st South East European Congress of Chemical Engineering. Belgrade, Serbia and Montenegro 97. Radeti´c M, Jovanˇci´c P, Radojevi´c D, Ili´c V, Miladinovi´c R, Joci´c D, Puaˇc N, Petrovi´c ZLj (2007) 2nd EMCO Workshop on Emerging Contaminants in Wastewaters: Monitoring Tools and Treatment Technologies. Belgrade, Serbia 98. Jovanˇci´c P, Molina R, Bertran E, Joci´c D, Julià MR, Erra P (2007) Wool surface modiﬁcation and its inﬂuence on related functional properties. In: Mittal KL (ed) Polymer Surface Modiﬁcation: Relevance to Adhesion, vol 4. VSP, Netherlands, p 139 99. Radeti´c M, Radojevi´c D, Ili´c V, Joci´c D, Povrenovi´c D, Potkonjak B, Puaˇc N, Jovanˇci´c P (2007) J Serbian Chem Soc 72:606 100. Trotman ER (2001) Dyeing and Chemical Technology of Textile Fibres, 4th edn. Grifﬁn, London 101. Jovanˇci´c P, Joci´c D, Molina R, Julià MR, Erra P (2001) Textile Res J 71:948 102. Julià MR, Erra P, Jocic D, Canal JM (1998) Text Chem Color 30:78 103. Sitting M (1981) Handbook of Toxic and Hazardous Chemicals. Noyes Publications, Park Ridge, USA 104. Liu DHF, Liptack BG, Bouis PA (1997) Environmental Engineer‘s Handbook, 2nd edn. Lewis, Boca Raton, FL, USA 105. Manahan SE (1997) Environmental Science and Technology. Lewis, Boca Raton, FL, USA 106. Blanchard G, Maunaye M, Martin G (1984) Water Res 18:1501 107. Vaca-Mier M, Callegas RL, Gehr R, Jimenez CB, Alvarez PJ (2001) Water Res 35:373 108. Shanableh A, Kharabsheh A (1996) J Hazard Mater 45:207 109. Misaelides P, Godelitsas A, Charistos V, Ioannou D, Charistos D (1994) J Radioanal Nucl Chem 183:159 110. Peri´c J, Trgo M, Vukojevi´c Medvidovi´c N (2004) Water Res 38:1893 111. Manceau A, Schlegel M, Nagy KL, Charlet L (1999) J Colloid Interface Sci 220:181

264

P. Jovanˇci´c · M. Radeti´c

112. Radeti´c M, Radojevi´c D, Ili´c V, Joci´c D, Povrenovi´c D, Puaˇc N, Petrovi´c ZLj, Jovanˇci´c P (2006) Trends in Applied Sci Res 1:564 113. Brown P, Atly Ji, Parrish D, Gill S, Graham E (2000) Adv Environ Res 4:19 114. Marshall WE, Wartelle LH, Boler DE, Toles CA (2000) Environ Technol 21:601 115. Gardea-Torresdey J, Hejazi M, Yiemann K, Parsons JG, Duarte-Gardea M, Henning J (2001) J Hazard Mater 2784:1 116. Reddad Z, Gérente C, Andrès Y, Le Cloirec P (2002) Environ Sci Technol 36:2067 117. Volesky B (1990) Removal and Recovery of Heavy Metals by Biosorption. In: Volesky B (ed) Biosorption of heavy metals. CRC Press, Boca Raton, FL, USA, p 7 118. Atkinson BW, Bux F, Kasan HC (1998) Water SA 24:129 119. Li Y, Ding J, Luan Z, Di Z, Zhu YF, Xu CL, Wu DH, Wei BQ (2003) Carbon 41:2787 120. Qi L, Xu Z (2004) Colloid Surf A 251:183 121. Peng X, Luan Z, Ding J, Di Z, Li Y, Tian B (2005) Mater Lett 59:399 122. Mangun CL, Yue ZR, Economy J, Maloney S, Kemme P, Cropek D (2001) Chem Mater 13:2356 123. Birnbaum ER, Rau KC, Sauer NN (2003) Sep Sci Technol 32:389 124. Ottaviani MF, Favuzza M, Bigazzi M, Turro NJ, Jockusch S, Tomkalia DA (2000) Langmuir 16:7368 125. Diallo MS, Christie S, Swaminathan P, Balogh L, Shi X, Um W, Papelis C, Goddard WA, Johnson JH Jr (2004) Langmuir 20:2640 126. Arkas M, Tsiourvas D, Paleou CM (2003) Chem Mater 15:2844 127. Balogh L, Swanson DR, Tomalia DA, Hagnauer Gl, McManus AT (2001) Nano Lett 1:18 128. Chen CZS, Cooper S (2002) Biomaterials 23:3359 129. Diallo MS, Christie S, Swaminathan P, Johnson JH Jr, Goddard WA (2005) Environ Sci Technol 39:1366

Hdb Env Chem Vol. 5, Part S/2 (2008): 265–274 DOI 10.1007/698_5_109 © Springer-Verlag Berlin Heidelberg Published online: 29 April 2008

Conclusions and Future Research Needs Damià Barceló1 (u) · Mira Petrovic1,2 (u) 1 Department

of Environmental Chemistry, IIQAB-CSIC, c/Jordi Girona 18–26, 08034 Barcelona, Spain [email protected], [email protected]

2 Institució

Catalana de Reserca i Estudis Avançats (ICREA), Passeig Lluis Companys 23, 08010 Barcelona, Spain

1

General Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

266

2

Chemical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

266

3

Occurrence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

268

4

Toxicology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

269

5 5.1 5.2 5.3 5.4 5.5 5.6

Treatment Technologies . . . . . . . . . . Conventional Treatment . . . . . . . . . . Membrane Bioreactors . . . . . . . . . . Ozonation and Photocatalytic Processes . Constructed Wetlands . . . . . . . . . . . Advanced Sorbents and Nanotechnology Artiﬁcial Recharge . . . . . . . . . . . . .

. . . . . . .

270 270 271 272 272 273 273

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

274

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

Abstract This chapter gives the main concluding remarks in relation to the chemical analysis, occurrence, toxicological impact, and different treatment technologies presented in this book, highlighting their possibilities and summarizing future research needs in the area of wastewater treatment of emerging contaminants. Keywords Chemical analysis · Ecotoxicology · Occurrence · Research needs · Treatment technologies

Abbreviations AOP CAS CW EDA GCxGC HPLC IT-(MS) MBR MS/MS

Advanced oxidation processes Conventional activated sludge Constructed wetlands Effect direct assay Two-dimensional gas chromatography High performance liquid chromatography Ion trap (mass spectrometry) Membrane bioreactor Tandem mass spectrometry

266 MWCO NSAID QqLIT-(MS) QqTOF-(MS) TIE TOF-(MS) UPLC WWTP

D. Barceló · M. Petrovic Molecular weight cut-off Non-steroidal antiinﬂammatory drugs Quadrupole linear ion trap (mass spectrometry) Quadrupole time-of-ﬂight (mass spectrometry) Toxicity identiﬁcation evaluation Time-of-ﬂight (mass spectrometry) Ultra performance liquid chromatography Wastewater treatment plant

1 General Remarks We have recently published a book on the analysis, fate, and removal of pharmaceuticals in the water cycle [1], which discussed many aspects related to the trace analysis, fate, and removal of pharmaceuticals in wastewater treatment plants (WWTP). This book is in a way the follow up to the previous one and covers complementary aspects. It covers a broad range of emerging contaminants, not only coming from urban wastewaters, like some of the pharmaceuticals, but also from industrial wastes. Although the ﬁrst book contained several chapters on removal, such as those on advanced oxidation processes and on membrane technologies, this book offers a more comprehensive variety of removal technologies for emerging contaminants, In this chapter, the main concluding remarks and the possibilities of these treatment technologies will be highlighted. General remarks on the progress of the analytical techniques and on toxicological issues in this ﬁeld will also be made since the number of publications in this particular area is continuously growing.

2 Chemical Analysis The issue of emerging contaminants is closely tied to analytical capabilities. Increased sensitivity in mass spectrometry, as a result of more efﬁcient ionization techniques and better detectors, has allowed detection of virtually any new and potentially harmful contaminant at a very low level. The introduction of new chromatographic techniques, such as fast liquid chromatography (LC), fast gas chromatography (GC), and two-dimensional gas chromatography (GCxGC) and advanced tandem and hybrid mass spectrometry (MS) detection systems (i.e., quadrupole time-of-ﬂight (QqTOF), quadrupole linear ion trap (QqLIT), and Orbitrap), has improved analysis of complex mixtures. However, the main drawback of the conventional approach is target compound monitoring, which is often insufﬁcient to assess

Conclusions and Future Research Needs

267

the environmental relevance of emerging contaminants. The real challenge is to decide on the signiﬁcance of the chemical data, and the discussion on the risks posed is still open. The key issue is the quality of analytical data, which should be tested and conﬁrmed through the performance of interlaboratory tests, combined with the use of reference materials. Additionally, current analytical methods only focus their attention on parent target compounds and rarely include metabolites and transformation products. But the question is whether chemical analysis of target compounds alone is sufﬁcient to assess contaminants present in the environment. Recent developments in the ﬁeld of mass spectrometry, such as the introduction of QqTOF and QqLIT instruments, allow the simultaneous determination of both parent and transformation products. Exact mass measurements provided by QqTOF and the capability of combining several scan functions are powerful tools for providing more accurate identiﬁcation of target compounds in complex samples, as well as for enabling structural elucidation of unknown compounds [2, 3]. However, general screening for unknown substances is time-consuming and expensive, and is often shattered by problems such as lack of standards and mass spectral libraries. In this respect, we should differentiate two types of “unknowns”: the “suspected unknowns”, like pesticide and/or pharmaceutical degradation and phototransformation products in the environment; and the “real unknowns” or “unknown unknowns”. A very powerful tool is the combined approach of TOF-MS (time-ofﬂight) and IT-MS (ion trap) instruments and hybrids such as QqTOF-MS and QqLIT-MS, which enable either accurate mass determination or fragmentation patterning based on MSn experiments, thus providing complementary structural information. Combination of ultra performance liquid chromatography (UPLC)-QqTOF-MS and high performance liquid chromatography (HPLC)-IT-MSn on LC-QqLIT-MS in the same basic format can be applied as a universal generic strategy to ﬁnd, characterize, and conﬁrm unknown compounds in environmental and food analyses. The availability of the precursor or product full mass spectra throughout each HPLC or UPLC chromatogram and either accurate mass measurements or multiple stages of mass spectrometry will provide qualitative information that can be used to ascertain whether metabolites or any other compound are present in the sample [4, 5]. The knowledge of exact analyte masses allows determination of the possible molecular formula and chemical structure of suspected metabolites. The sequential fragmentations will provide a better understanding of the fragmentation patterns. It seems feasible that this approach could be useful in providing information on metabolites while simultaneously achieving their quantiﬁcation. Examples of the “suspected unknowns” by the combined use of both techniques have been recently reported in the literature, such as the identiﬁcation of four new phototransformation products of enalapril [6]. Accurate mass

268

D. Barceló · M. Petrovic

measurements recorded on a hybrid quadrupole time-of-ﬂight (QqToF) instrument in MS/MS mode allowed the proposal of elemental compositions for the molecular ions of the transformation products (346 Da C19 H26 N2 O4 ; 207 Da C12 H17 NO2 ; 304 Da C17 H24 N2 O3 ) as well as of their fragment ions. Based on these complementary data sets from the two distinct mass spectrometric instruments, plausible structures could be postulated. The second class of unknowns is the “real unknowns” and/or “unknown unknowns”. Basically these compounds are detected by the accurate mass measurements of QqTOF when analyzing other target compounds, since TOF analyzers offer full mass spectrum with accurate mass, and consequently the possibility to look for other analytes in a speciﬁc target analysis. Two examples are reported here. In the ﬁrst case, when selected estrogens like estrone, estradiol, and ethynil estradiol were analyzed in the QqTOF-MS mode, QqTOF allowed the identiﬁcation of non-target and/or suspected compounds like the phytoestrogens daidzein, genistein, and biochanin [7]. Published literature on the subject, chemical databases, and websites, combined with the QqTOF results, are very useful tools for unambiguously identifying compounds. The main problem with this modus operandis is the lack of libraries that allow the search for a possible structure for a given elemental composition within the equipment software.

3 Occurrence Numerous papers have reported on the occurrence of a wide range of emerging contaminants in the aquatic environment, wastewater and treated wastewater (WWTP efﬂuents) being the principle sources and routes of their entry into the environment (see chapter by M Petrovic et al., in the ﬁrst volume). However, more monitoring studies are needed, not only to conﬁrm the presence of emerging substances in the aquatic environment, but also to allow the reﬁnement of risk assessments in combination with relevant ecotoxicological data. In relation to the emergence of new pollutants in the environment, the integration of physical/chemical techniques, effect monitoring techniques (e.g., bioassays, functional monitoring, etc.) and ecological monitoring/assessment (community surveys) techniques plays a crucial role. The use of effect direct assay (EDA) or toxicity identiﬁcation evaluation (TIE) seems to be a good choice and the way to proceed in such cases. Effect-directed analysis combining biotesting, fractionation and chemical analysis helps to identify hazardous compounds in complex environmental mixtures. Conﬁrmation of tentatively identiﬁed toxicants can help to avoid artifacts and to establish reliable cause–effect relationships. A tiered approach of conﬁrmation is suggested. The ﬁrst tier focuses on analytical conﬁrma-

Conclusions and Future Research Needs

269

tion of tentatively identiﬁed structures. Conﬁrmation is a crucial element in EDA. As a tiered approach it provides evidence on identiﬁed chemical structures and quantitative estimates of cause–effect relationships. Conﬁrmation on all tiers is impeded by multiple restrictions including small sample amounts, limited availability of standards, limited availability of spectra and retention data, a lack of information on concentration–effect relationships, a lack of clarity in joint effect prediction, and very limited tools for effect conﬁrmation in situ. Thus, consistent guidance for the conﬁrmation of toxicants on different levels of complexity is urgently needed. Important elements could be automated structure generation according to the analytical information gained during EDA, improved tools for structure elucidation, and conﬁrmation by LC-MS/MS and its different forms listed in the previous section [8].

4 Toxicology (Eco)toxicologists often seems to know too little too late, and are far too slow to respond to the numerous chemicals that enter the market every day. Moreover, most (eco)toxicological testing is done using traditional acute toxicity test protocols. As was reliably demonstrated with pharmaceuticals, acute toxicity cannot always serve as a reliable proxy for chronic toxicity effects encountered in real environmental situations. Certain substances may elicit adverse effects far after exposure has ceased; weeks, months or years after exposure. Carcinogenicity is a classic example – an ultimate adverse outcome where the causal connections are difﬁcult to characterize. Consequently, chronic exposure assessments cannot be avoided and proper toxicological characterization will probably continue to be a time-consuming process. The key points regarding the ecotoxicology of emerging contaminants are: 1. The threat posed by numerous emerging contaminants present in industrial and municipal waste is serious, poorly characterized, and should not be underestimated. 2. The research capacity of (eco)toxicology is at the moment far beyond capacity of analytical chemistry to detect new, emerging contaminants, and even more distant from the capacity of the industrial sector to design and introduce new chemical entities, the likely “emerging” contaminants of the future. 3. Chronic, low-level exposure assessments do not have any scientiﬁcally sound alternative and should represent an obligatory part of (eco)toxicity characterization of single chemicals and complex environmental mixtures.

270

D. Barceló · M. Petrovic

4. The necessary improvements in the ﬁeld of (eco)toxicology will not be possible without a major shift in the regulatory arena, including signiﬁcant changes in environmental funding schemes. Countries that adopt proactive approaches, such as the EU REACH initiative, will be afforded distinct environmental, economic, and scientiﬁc advantages because they will be better serving human and nonhuman populations and ecosystems, with tangible savings to the costs of health-care and environmental protection. Without the adoption of proactive plans to identify contaminants before they emerge, regulatory communities that remain in reactionary modes will be unable to fully serve the needs of the populations they represent.

5 Treatment Technologies A variety of treatment technologies are described in the literature and during the last few years signiﬁcant improvements have been achieved in this area. However, the overall conclusion on the different technologies presented in this book is that we certainly need an improvement of wastewater treatment options for the removal of the comprehensive variety of emerging contaminants present. It is obvious that there is not a unique treatment option for all the cases. It is also clear that in order to verify that a certain treatment is adequate, appropriate analytical methodologies need to be used that will show if the removal of certain compound during the wastewater treatment process took place or not. 5.1 Conventional Treatment The capacity to eliminate micropollutants depends on various factors, including the chemical and biological persistence of the individual compound, its sorption behavior, its volatility, and the technology and operation conditions (e.g., temperature, hydraulic retention time, and sludge retention time). Sorption to suspended solids in the wastewater and to sludge particles, and subsequent removal by sedimentation, is relevant for lipophilic compounds (log Pow > 3 at ambient pH) and for some hydrophilic compounds that can interact in a speciﬁc way (e.g., surfactants). For high polar substances (such as most pharmaceuticals) and the corresponding metabolites the most important removal process is the biological transformation or mineralization by microorganisms, which strongly depends upon the treatment technology and operation conditions. Removal rates showed a great variability according to sewage treatment plants and types of treatments.

Conclusions and Future Research Needs

271

The mere occurrence of a contaminant in conventional activated sludge (CAS) efﬂuents does not imply that it will spread in the aquatic environment. Only when this polar pollutant is stable would this risk have to be considered. Therefore, a moderate efﬂuent concentration of a poorly degradable compound that did not experience any concentration decrease in CAS should be considered more problematic than the same efﬂuent concentration of a compound that was degraded extensively from a much higher inﬂuent concentration. However, in order to understand the process taking place in the WWTP and to increase the knowledge of biodegradation of contaminants in WWTP, biodegradation studies of selected emerging contaminants of interest, under laboratory controlled conditions simulating WWTPs, should be conducted. A few studies have investigated biodegradation pathways in various environmental compartments and reported the identities of biotransformation products during primary biodegradation. The identiﬁcation of degradates in environmental samples is a challenging task because not only are they present in very low concentrations but they are also mixed with complex matrices that interfere with detection. There is a need to increase our knowledge about the fate of emerging contaminants during sewage treatment for the implementation of better removal technologies. Future work on WWTP will show to what extent emerging contaminants can be removed from wastewater and to what extent the implementation of an improved technology is feasible, taking into account other macro- and micropollutants as well as the broad variety of complex matrices. 5.2 Membrane Bioreactors A need for the development of membrane bioreactor (MBR) technology arose mainly from the limitations of the CAS process. The potential of MBR to efﬁciently remove hazardous substances from wastewater is often highlighted due to several important factors: (1) there is physical retention of all molecules larger than the molecular weight cut-off (MWCO) of the membrane, (2) hydrophobic substances also tend to accumulate onto the sludge and are therefore removed from the efﬂuent, and (3) as all the bacteria are held back, they are better adapted to mineralizing of micropollutants present in the reactor. However, the efﬁciencies of diverse microbial populations in the elimination of selected emerging contaminants and the optimization of design and operating parameters are the subjects of future research needed in this area. However, there are certain drawbacks to the wider implementation of MBR technology. MBR is widely viewed as being a state-of-the-art technology but is also sometimes seen as high-risk and still costly compared to CAS and other more established technologies. MBRs were historically perceived as

272

D. Barceló · M. Petrovic

suitable only for small-scale plants with high operator skill requirements, and the key operating expenditure parameters such as membrane life were unknown. Many of these drawbacks are no longer true. Perhaps the biggest challenge to companies active in the market is to persuade decision-makers of the capability of MBRs and of the beneﬁts they will undoubtedly bring to the customer. In the past, there were an insufﬁcient number of full-scale MBR treatment plants to convince decision-makers of the reliability of this advanced treatment. Scale-up from pilot MBR to real-world WWTP should also be investigated in order to assess if the processes and elimination in the pilot pant are still valid in a large-scale plant. Presently, there are a number of examples of successful implementation of MBRs across the range of applications, and there is certainly less reason to be suspicious of this technology. 5.3 Ozonation and Photocatalytic Processes In recent years, new technologies have been studied for the treatment of wastewaters and among them advanced oxidation processes (AOPs) are frequently considered to be one of the most promising technologies. For example, the use of an ozone/hydrogen peroxide system as a tertiary treatment of domestic and urban wastewater could provide water reclamation for use in agriculture or industrial processes. Additionally, a signiﬁcant reduction of the discharge of emerging contaminants into the environment could be expected through the use of catalysis and solar energy. Recent investigations increasingly focus on the two AOPs that can be powered by solar radiation, namely homogeneous catalysis by the photo-Fenton reaction and heterogeneous catalysis by the UV/TiO2 process; however, commercial applications are still scarce. 5.4 Constructed Wetlands The elimination of emerging contaminants such as non-steroidal antiinﬂammatory drugs (NSAIDs), lipid regulator drugs, antiepileptic agents, fragrances, surfactants, estrogens, caffeine, and triclosan by constructed wetlands (CWs) could constitute a feasible alternative to WWTP in small communities. However, until now most of the information has been related to small plants or pilot-scale studies. Different factors affecting removal of emerging pollutants have been found relevant, of which oxygen concentration was shown to be one of the most important because of the best performance of aerobic pathways. Nevertheless, the performance, in terms of efﬂuent quality, is believed to be better for some emerging pollutants (estrogens and alkylphenols), as shown in hybrid designs (combination in series of differ-

Conclusions and Future Research Needs

273

ent CW conﬁgurations). Therefore, the possibilities of using CWs rather than high-cost technologies (i.e., MBR, ultraﬁltration, and ozonation) to remove emerging pollutants from contaminated waters are open to discussion. This technology has been successfully applied for small communities but it will not be possible to use it for big cites since, basically, there is no space to perform the wastewater treatment processes. 5.5 Advanced Sorbents and Nanotechnology In spite of a wide range of other wastewater treatment technologies, adsorption (adsorptive separation) is still a key separation tool and is used in many industries. Future application of adsorption is limited by the availability of new and more efﬁcient sorbents. The sorbents should be tailored in such a way as to meet the requirements of each speciﬁc application. Despite many research efforts and patents on adsorption, there are only a few commercially available sorbents that are used in current adsorption processes (activated carbon, zeolites, silica gel, and activated alumina). Advances in nanoscale science and engineering are providing great opportunities to develop more cost-effective and environmentally acceptable water puriﬁcation processes. Nanosorbents will certainly become critical components of industrial wastewater and public water puriﬁcation systems as more progress is made towards the synthesis of cost-effective and environmentally acceptable functional materials. The potential use of the adsorption properties of new forms of carbon (carbon molecular sieves, superactivated carbon, activated carbon ﬁbers, carbon nanotubes, and graphite nanoﬁbers) and nanoporous materials, i.e., nanomaterials, (metal-containing nanoparticles, dendrimers etc.) is in a process of intensive exploration. 5.6 Artificial Recharge Artiﬁcial recharge using recharge waters of impaired quality (such as reclaimed wastewater and storm water runoff) is a viable alternative to surface water where recharge is intended for non-potable purposes (such as landscape irrigation) after receiving at least a secondary treatment. However, the health risk associated with such sources is high where the recharge water is to be used as drinking water, and should only be used when better quality sources are unavailable. Signiﬁcant behaviors of trace organic contaminants during artiﬁcial aquifer recharge have been found. Outcomes reported underlined the potential of groundwater recharge systems to remove contaminants to some extent. However, at recharge sites the removal of organic pollutants depends on a number of factors, such as the climate, hydrogeological features of the area, and the physical–chemical conditions of the aquifer. In addition,

274

D. Barceló · M. Petrovic

different attenuation behaviors can also be found depending on the recharge strategy (bank ﬁltration vs. inﬁltration basins).

References 1. Petrovic M, Barceló D (Eds) (2007) Analysis, fate and removal of pharmaceuticals in the water cycle. Wilson & Wilson’s comprehensive analytical chemistry, vol 50. Elsevier Science, Amsterdam, p 564 2. Ibañez M, Sancho JV, Pozo OJ, Niessen W, Hernandez F (2005) Rapid Commun Mass Spectrom 19:169 3. Hernandez F, Pozo OJ, Sancho JV, Lopez FJ, Marin JM, Ibañez M (2005) Trends Anal Chem 24:596 4. Petrovic M, Gros M, Barceló D (2006) J Chromatogr A 1124:68 5. Petrovic M, Barcelo D (2006) J Mass Spectrom 41:1259 6. Pérez S, Eichhorn P, Barceló D (2007) Anal Chem 79:8293 7. Farré M, Kuster M, Brix R, Rubio F, López de Alda MJ, Barceló D (2007) J Chromatogr A 1160:166 8. Brack W, Schmitt-Jansen M, Machala M, Brix R, Barceló D, Schymanski E, Steck G, Schulze T (2008) Anal Bioanal Chem (in press) doi: 10.1007/s00216-007-1808-8

Hdb Env Chem Vol. 5, Part S/2 (2008): 275–280 DOI 10.1007/698_5_104 © Springer-Verlag Berlin Heidelberg Published online: 12 January 2008

Erratum to Membrane Bioreactor (MBR) as an Advanced Wastewater Treatment Technology Jelena Radjenovi´c1 (u) · Marin Matoˇsi´c2 · Ivan Mijatovi´c2 · Mira Petrovi´c1,3 · Damià Barceló1 1 Department

of Environmental Chemistry, IIQAB-CSIC, c/ Jordi Girona 18–26, 08034 Barcelona, Spain [email protected] 2 Faculty of Food Technology and Biotechnology (PBF), Pierottijeva 6, Zagreb, Croatia 3 Institució

Catalana de Reserca i Estudis Avanzats (ICREA), Barcelona, Spain

Unfortunately, some ﬁgures were published without giving credit to the original source of the ﬁgure. The corrected ﬁgure citations are provided here:

Fig. 11 Net observed yield · · · and biomass concentration as a function of sludge age in a MBR, HRT=2.7 h, Y=0.4, k = 0.07 d–1 , kd = 0.06 d–1 . Reprinted from [48], copyright (2003), with permission from Wiley-Blackwell Publishing Ltd.

276

J. Radjenovi´c et al.

Fig. 12 Death kinetics of P. Fluorescens at different temperatures. Reprinted from [69], copyright (1987), with permission from Springer-Verlag

Fig. 13 Sludge particle size distributions at different SRTs. Reprinted from [56], copyright (2001), with permission from Elsevier

Erratum to Membrane Bioreactor (MBR)

277

Fig. 14 Growth of MLSS concentrations in submerged MBR at different HRTs. Reprinted from [113], copyright (2006), with permission from Elsevier

Fig. 15 Schematic illustration of virus structure with electrokinetic double layers. Reprinted from [129], copyright (2001), with permission from Elsevier

278

J. Radjenovi´c et al.

Fig. 16 Elimination rates of PhACs in MBR and CAS treatment. Reprinted from [156], copyright (2007), with permission from Elsevier

Fig. 18 Relative abundance of NPEOs and their metabolites in primary and secondary efﬂuents (weight-based average value of 11 WWTPs in the Glatt Valley, Switzerland). Reprinted from [237], copyright (1995), with permission from Elsevier

Erratum to Membrane Bioreactor (MBR)

279

Fig. 19 Boxplot (calculated on a molar basis) and average composition of NP compounds in inﬂuent, CAS efﬂuent and MBR efﬂuent. Reprinted from [227], copyright (2007), with permission from Elsevier

Fig. 22 Aeration demand for biodegradation of organic matters as a function of target MLSS and HRT. Flow rate and COD of inﬂuent were 1000 m3 day–1 and 400 mg L–1 , respectively. Reprinted from [272], copyright (2004), with permission from Elsevier

280

J. Radjenovi´c et al.

Additionally, in the second paragraph of Sect. 5.4, the sentence that begins, “Hence, given their low infective dose ...” should be deleted since this information appears in the previous paragraph. Also, reference 156 should be updated to: Radjenovi´c J, Petrovi´c M, Barceló D (2007) Trends Anal Chem 26:1132 Finally, reference 157 was published incorrectly. The correct reference is: Clara M, Strenn N, Kreuzinger N (2004) Water Res 38:947

The authors regret these errors.

Subject Index

Acetamides 231 Acetaminophen 72 1-Acetyl-1-methyl-2-dimethyl-oxamoyl-2phenylhydrazide (AMDOPH) 219 Activated carbon 247 Advanced oxidation processes 127, 178 Advanced sorbents 273 Aeration 52, 59 – demand 93 – ﬂow 65 Alcohol ethoxylates (AEO) 28, 29, 80 Aldrin 13 Alkyl ether sulfates (AES) 26 Alkyl sulfates (AS) 26 Alkylphenol ethoxylates (APEOs) 28, 80, 212, 229 Alkylphenols (APs) 81 Aminobenzothiazole (ABT) 83 Analgesics 71 Anionic surfactants 199 Antibiotics 71, 74, 190 – removal 118 Anti-epileptic agents 207 Anti-inﬂammatory drugs 71 Aquifer 219 Arsenic, removal by oxidation 143 Artiﬁcial recharge 219, 221, 273 Atenolol 19, 71, 167 Atrazine 13, 121, 231 Azithromycin 74, 75 Azo dyes 246 Bacillus subtilis spores, ozone Bacteria removal 65 Bacteriophages 66 Bayrepel 24 Bentazone 83, 230 Benzothiazoles 83 Benzotriazoles 84

144

Benzoylecgonine 30 Benzyl acetate 209 Benzylpenicillin 119 Bezaﬁbrate 19, 71–75, 233 Biological oxygen demand (BOD) 5, 61 Bis(2-hethylhexyl)phthalate 226 Bisphenol A 30, 78 β-Blockers 20, 71, 166 Blue baby disease 59 Bulking sludge 86 Caffeine 213 Carbadox 118 Carbamazepine (anticonvulsant) 73, 120, 128, 158, 207 – ozonation 161 Carboxylic acids, chlorinated ozonation 154 CAS 1 Catalysis, metals 156 Catalytic ozonation 153 Charge exclusion 111 Chemical oxygen demand (COD) 5, 61, 147 Chitosan 252 Chlorophene 170 Cholesterol lowering statin drugs 71 Ciproﬂoxacin 167 Clarithromycin 74, 75 Clinoptilolite 250 Cloﬁbric acid 19, 71, 207 Cocaine 30 Coliphages 66 Color removal 239, 246 – PAC 249 Common reed (Phagmitis australis) 201 Constructed wetlands 199, 272 Conventional activated sludge (CAS) 40 – treatment 3

282 Copper 255 Critical ﬂux 50 Cryptosporidium parvum oocysts Cucurbituril 252

Subject Index

143

DDT 13 DEET 24, 84 Denitriﬁcation 58 Dentritic polymers 259 Dialkyltetralinsulfonates (DATS) 234 Dichlorobenzoic acid (2,4-D) 15, 83 Diclofenac 19, 71–74, 191, 207, 233 Dieldrin 13 Diethylenetriamino pentaacetate (DTPA) 29 Dimethylaminophenazone 230 1,7-Dimethylxanthine (paraxanthine) 167 Dipyrone 167, 194 Disinfection 127 – drinking water 143 Disinfection by-products (DBPs) 66, 227 Diuron 170 Dyes 245 E. coli, ozone 144 EDDP 32 Electro deionization (EDI) 41 Electrodialysis (ED) 41 Endocrine disrupting compounds (EDCs) 77, 104, 210 Energy usage, MBR 92 Enhanced biological phosphorus removal (EBPR) 60 Enroﬂoxacin 119 Erythromycin 71, 74 Estrogens 77, 119, 199, 210, 227 17α-Ethinyl estradiole (EE2) 77, 210 Ethylenediamino tetraacetate (EDTA) 29, 84 Extracellular polymeric substances (EPS) 49 Feminization 77 Fenoﬁbric acid 167 Fenton reaction 184 Filtration resistance 48 Flame retardants 30, 84 Flocs 54 Food-to-microorganism (F/M) ratio 55

Fouling, control 52 – membranes 121 Fragrances 209 Galaxolide 22, 84, 209 Gemﬁbrozil 19, 71, 170, 207 Giardia lambia cysts, ozone 144 Giardia muris cysts, ozone 144 Glycogen-accumulating organisms (GAOs) 61 Groundwater 219 – contamination 224 Haloacetic acids 223 HCHs 13 Heavy metal ions, removal 239, 255 Heptachlor 13 Herbicides 12, 83 Hexadecyl trimethyl ammonium bromide (HTAB) 241 Hollow-ﬁber (HF) membrane module 43 Horizontal ﬂow constructed wetlands (HFCW) 201 Hormones, removal 77 Hydraulic loading rate 199 Hydraulic retention time (HRT) 5 Hydrochlorothiazide 19, 167 Hydrogen peroxide 252 Hydroxybenzothiazole (OHBT) 83 Ibuprofen 19, 72, 120, 170, 204, 233 Indomethacine 72, 74 Industrial wastewater 163 Insect repellents 23, 84 Iron, removal by oxidation 141 Iron complexes 184 Isoproturon 15 Ketoprofen 19, 71, 204 Kinetic models 127 Lead 255 Legislation, Europe 6 Lincomycin 190 Lindane 232 Linear alkylbenzene sulfates/sulfonates (LAS) 26, 80, 165, 234 Lipid regulators 71, 207 Lymcomycin 177

Subject Index Macrolide antibiotics 74 Manganese, removal by oxidation 141 Mass loading rate (MLR) 205 Mecoprop 13, 230 Mefenamic acid 71, 170 Membrane bioreactors (MBR) 37, 271 – application/cost analysis 90 – biological performance 37, 54 – conﬁgurations 41 – hydraulics 45 – side-stream 42 – submerged 42 – trace organic compounds, removal 69 Membrane classiﬁcation 40 Membrane ﬂux dependency, pressure 122 Membrane fouling 47, 121 Membrane pore size 110 Metal oxides 156 Methadone 32 Methyl dihydrojasmonate 209 Methyl tert-butyl ether (MTBE) 177, 188, 231 4-Methylaminoantipyrine 167, 194 Methylthiobenzothiazole (MTBT) 83 Metoprolol 19, 71 Microﬁltration (MF) 41, 105 Micropollutants, removal 82 – CAS 11 Mixed-liquor suspended solids (MLSS) 55 Mixed-liquor volatile suspended solids (MLVSS) 56 Morphine 32 MTBE 177, 188, 231 Musk fragrances, removal 22, 82, 233 Nanoﬁltration 41, 103, 105 Nanosorbents 259, 273 Nanotechnology 273 Naphthalene disulfonates (NDSAs) 83 Naphthalene monosulfonates (NSAs) 83 Naproxen 19, 72, 128, 158, 207 – ozonation 161 Narrowleaf cattail (Typha angustifolia) 201 Natural water treatment 145 NF/RO, organics removal 119 – membrane, solute rejection 108 Nitrate-oxidizing bacteria (NOB) 60 Nitriﬁcation 58

283 Nitrite/nitrate 59 Non-steroidal anti-inﬂammatory drugs (NSAIDs) 204 Nonylphenol 78, 234 Nonylphenol ethoxylates (NPEOs) 28, 80, 212, 234 Nonylphenoxy carboxylates (NPECs) 28 NSAIDs 177, 199, 204 Octylphenol (OP) 212 Octylphenol ethoxylates (OPEOs) 28, 80, 212 Oﬂoxacin 71 Organic chemicals, groundwater contamination 227 Organic contaminants 219 Organic matter, removal 62 Organochlorine pesticides 13 Oxalic acid, ozonation 155 Oxides, precipitation 141 Oxytetracycline waste liquor 118 Ozonation 127, 272 – homogeneous catalytic 155 Ozonation processes 129 Ozone, decomposition in water 134 – reactions with organic compounds 138 – solubility in water 130 – water treatment 140 Ozone mass transfer 131 Pentachlorophenol 226 Permeate 45 Personal care products 16, 178 Pesticides 1, 11 – removal 82 Pharmaceutically active compounds, removal 69, 70 Pharmaceuticals 1, 16 – wastewater, removal efﬁciency 166 Phenoxy acids 83, 231 Phosphate accumulating microorganisms (PAOs) 60 Phosphates 59 Phosphorus removal 58 Photocatalysis 177 – heterogeneous 179 – homogeneous (photo-Fenton) 184 Photocatalytic processes 272 Photo-Fenton 177

284 Phthalate esters (PAEs) 164 Phytoestrogens 77 Piracetam 75 Polycyclic aromatic hydrocarbons (PAHs) 84, 226 Polyethylene glycols (PEG) 29 Pore size distribution 110 Powdered activated carbon (PAC) 53 PPCPs 199 Pravastatin 71 Praziquantel 119 Predominant ammonium bacteria (AOB) 60 Progesterone 78, 119 Propyphenazone 71, 75 Pulp mill wastewaters 164 Pyrogens 67 Radicals 184 Ranitidine 71 Recharge waters, sources/pretreatments 222 Rejection mechanisms 103 Removal efﬁciency 199 Reverse osmosis 41, 103, 105 Rotavirus, ozone 144 Roxythromycin 74 Sewage treatment, pathogenic organisms 66 Simazine 13, 121 Size exclusion 108 Sludge particle size distributions 63 Sludge retention time (SRT) 5 Sludge treatment, activated 3 Sludge volume index (SVI) 65 Solid catalysts 127 Solid retention time (SRT) 50 Soluble microbial products (SMP) 49 Solution-diffusion models 107 Sorbents, low-cost 239, 252, 255 Sorption 78 Starch degradation, ozone 165 Submicrogram concentrations 116 Sulfachlorpyridazine 118 Sulfadimethoxine 118 Sulfamerazine 118 Sulfamethazine 118 Sulfamethoxazole 74, 120, 230 Sulfathiazole 118

Subject Index Sulfonated organic compounds, removal 82 Sulfophenyl carboxylates (SPC) 80 Surface ﬂow constructed wetlands (SFCWs) 201 Surface water 219 Surfactants 1, 212 – metabolites 25 – removal 80 Suspended solids, removal 62 TCEP/TCPP 30, 234 Temazepan 32 Testosterone 119 Textile efﬂuents/wastewaters 239, 242 Titanium oxide 154, 157, 180 Tolyltriazole 32 Tonalide 22, 84, 209 Total coliform (TC), drinking water 66 Total Kjeldahl nitrogen (TKN) 59 Total suspended solids (TSS) 5 Trace organic pollutants 37 Triazines 187, 231 Triclosan 213 Trihalomethanes (THMs) 223 Trimethoprim 74, 118 Turbidity removal 64 Ultraﬁltration (UF) 41, 105 UV ﬁlters 23 Vertical ﬂow constructed wetlands (VFCW) 201 Virus removal 37, 65 Virus–virus coagulation 68 VOCs 224, 234 – MTBE 188 Volatile suspended solids (VSS) 5, 63 Wastewater treatment 37, 145, 219 Wastewater treatment plants (WWTPs) 40 Water Framework Directive (WFD) 6 Water treatment 103 Wetlands, constructed 201 Xenoestrogens Zeolites 250 Zinc 255

78