No title

WATER RESEARCH A Journal of the International Water Association Editor-in-Chief Mogens Henze Institute of Environment &...

Author: Mark van Loosdrecht

302 downloads 3515 Views 14MB 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

WATER RESEARCH A Journal of the International Water Association

Editor-in-Chief Mogens Henze Institute of Environment & Resources Technical University of Denmark Bygningstorvet DK-2800 KGS Lyngby Denmark Tel: 45 4525 1477 Fax: 45 4593 2850 E-mail: [email protected]

Editors J. Block Université H. Poincaré, Nancy I France David Dixon University of Melbourne Australia Hiroaki Furumai The University of Tokyo Japan Gregory Korshin University of Washington USA Anna Ledin Technical University of Denmark Denmark Eberhard Morgenroth University of Illinois Urbana-Champaign USA W. Rauch University Innsbruck Austria Maria Reis Universidade Nova de Lisboa/FCT Portugal Hang-Shik Shin Korea Advanced Institute of Science and Technology Korea Mark van Loosdrecht Delft University of Technology The Netherlands Thomas Ternes Bundesanstalt für Gewässerkunde Germany Stefan Wuertz Univ. of California, Davis USA Hanqing Yu University of Science & Technology of China China

Associate Editors Andrew Baker The University of Birmingham UK

Damien Batstone The University of Queensland Australia

S-L. Lo National Taiwan University Taiwan

G-H. Chen The Hong Kong University of Science & Technology Hong Kong China

Y. Matsui Hokkaido University Japan

Tom Curtis Univ. of Newcastle upon Tyne UK Ana Deletic Monash University USA Francis de los Reyes III North Carolina State University USA Rob Eldridge The University of Melbourne Australia Rosina Girones University of Barcelona Spain Stephen Gray Victoria University Australia Kate Grudpan Chiang Mai University Thailand Xiaodi Hao Beijing University of Civil Engineering and Architecture China E.E. Herricks University of Illinois - Urbana USA H-Y. Hu Tsinghua University China P.M. Huck University of Waterloo Canada Bruce Jefferson Cranfield University UK Sergey Kalyuzhnyi Moscow State University Russian Federation

Max Maurer EAWAG Switzerland How Yong Ng National University of Singapore Singapore Satoshi Okabe Hokkaido University Japan S.L. Ong National University of Singapore Singapore Jong M. Park Pohang University of Science & Technology Korea Miguel Salgot University of Barcelona Spain David Sedlak University of California, Berkeley USA Jean-Philippe Steyer LBE-INRA France M. Takahashi Hokkaido University Japan Kai Udert EAWAG Switzerland V.P. Venugopalan BARC Facilities India E. von Sperling Federal University of Minas Gerais Brazil A.I. Zouboulis Aristotle University Greece

Jaehong Kim Georgia Institute of Technology USA

Editorial Office

Jes La Cour Jansen Lund Institute of Technology Sweden

Water Research Radarweg 29 1043 NX Amsterdam The Netherlands Tel: 31 20 485 2701 Fax: 31 20 485 2521; 31 20 485 3264 E-mail: [email protected]

G. Langergraber BOKU/Univ. of Natural Res. and Applied Life Scs. Austria

Mailing address

Publication information: Water Research (ISSN 0043-1354). For 2010, volume 44 is scheduled for publication. Subscription prices are available upon request from the publisher or from the Elsevier Customer Service Department nearest you or from this journal’s website (http://www.elsevier.com/locate/watres). Further information is available on this journal and other Elsevier products through Elsevier’s website (http://www.elsevier.com). Subscriptions are accepted on a prepaid basis only and are entered on a calendar year basis. Issues are sent by standard mail (surface within Europe, air delivery outside Europe). Priority rates are available upon request. Claims for missing issues should be made within six months of the date of dispatch. Author Enquiries: Please visit this journal’s homepage at http://www.elsevier.com/locate/watres. You can track accepted articles at http://www.elsevier.com/ trackarticle and set up e-mail alerts to inform you of when an article’s status has changed. Also accessible from here is information on copyright, frequently asked questions and more. Contact details for questions arising after acceptance of an article, especially those relating to proofs, will be provided by the publisher. Orders, claims, and journal enquiries: Please contact the Elsevier Customer Service Department nearest you: St. Louis: Elsevier Customer Service Department, 11830 Westline Industrial Drive, St. Louis, MO 63146, USA; phone: (877) 8397126 [toll free within the USA]; (+1) (314) 4537076 [outside the USA]; fax: (+1) (314) 5235153; e-mail: [email protected] Oxford: Elsevier Customer Service Department, The Boulevard, Langford Lane, Kidlington OX5 1GB, UK; phone: (+44) (1865) 843434; fax: (+44) (1865) 843970; e-mail: [email protected] Tokyo: Elsevier Customer Service Department, 4F Higashi-Azabu, 1-Chome Bldg, 1-9-15 Higashi-Azabu, Minato-ku, Tokyo 106-0044, Japan; phone: (+81) (3) 5561 5037; fax: (+81) (3) 5561 5047; e-mail: [email protected] Singapore: Elsevier Customer Service Department, 3 Killiney Road, #08-01 Winsland House I, Singapore 239519; phone: (+65) 63490222; fax: (+65) 67331510; e-mail: [email protected] Application for membership of International Water Association should be made to: Executive Director, IWA, Alliance House, 12 Caxton Street, London SW1H 0QS, U.K. (Tel.: +44 207 654 5500; Fax: +44 207 654 5555; e-mail: [email protected]; website: http://www.IWAhq.org.uk). Registered Charity (England) No. 289269. Individual membership is available from £30 upwards. For details contact IWA.

water research 44 (2010) 351

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/watres

Editorial

Editorial to special issue in Water Research Emerging contaminants in water The chemical pollution of natural waters is one of the big challenges of the 21st century. Based on the rapid evolution in analytical chemistry, with the possibility to detect more polar compounds a whole new suite of ‘‘emerging contaminants’’ such as pharmaceuticals, hormones and perfluorinated compounds has been identified in various compartments of the water cycle including both natural and technical aquatic systems. The discovery of these micropollutants in the aquatic environment has triggered research efforts to investigate sources and mitigation strategies with conventional and novel treatment processes and assess their importance with regard to ecotoxicology and human health. In this special issue on emerging contaminants, we have compiled 20 research articles and 2 reviews, which deal with these timely issues. As mentioned above, quantification of micropollutants in aquatic systems is a key requirement to assess their fate. Several studies in this issue address the determination of pharmaceuticals in archived biosolids (Halden et al.), in wastewater treatment (Lindberg et al.) and lagoon treatment (Wong et al.). Another study uses analytical data to evaluate the fraction of pharmaceutical residues in wastewater originating from hospitals (Ort et al.). Finally, a review covers the occurrence and fate of phytoestrogens in the environment (Liu et al.). To further elucidate the relevance of the micropollutants detected in various aquatic compartments, their (eco)toxicological potential has to be assessed. Several papers address related issues for individual compounds (lipid regulators, Fernandez-Pinas et al.) or classes of compounds (ionic liquids, Yun et al.) and for pharmaceuticals in advanced wastewater treatment systems such as powdered activated carbon and ozonation with in vivo and in vitro tests (Escher et al., Stalter et al.). Furthermore, one study is focused on toxicity nanotube suspensions (Tarabara et al.), which have been on the radar of emerging contaminants recently. Finally, one paper focuses on the toxicological relevance of emerging contaminants for drinking water (Schriks et al.). In recent years, municipal wastewater has been recognized as an important source of micropollutants to the

receiving water bodies. Therefore, mitigation strategies for the minimization of the discharge of these compounds play an increasingly important role in the urban water management. In this special issue, the removal of benzotriazoles (Reemtsma et al.) and pharmaceuticals, caffeine and DEET (Yu et al.) and other emerging contaminants (Rosal et al.) has been investigated for conventional wastewater treatment. Numerous papers address the oxidative removal of micropollutants from wastewater with chlorine, chlorine dioxide, ferrate, ozone, advanced oxidation processes, such as UV/ H2O2, (solar) photo-Fenton and non-thermal plasma (Lee et al., Malato et al., Reungoat et al., Gerrity et al., MendezArriaga). Other options for removal of micropollutants include membrane processes such as reverse osmosis (Hu et al.) and nanofiltration (Yangali-Quintanilla et al.) as well as sorption on sludge (Carrere et al.). The challenges caused by harmful algae producing toxins for desalination operations were reviewed by David Caron and co-workers. Finally, mitigation of micropollutants may also occur during managed aquifer recharge (Drewes et al.) or in biological Fenton-like processes (Vicent et al.). In short, we are very happy to provide you this Theme Issue on Emerging Contaminants. We appreciate the contributions by the authors, reviewers and editorial staff of Water Research to this project.

Thomas Ternes* Federal Institute of Hydrology (BFG), Am Mainzer Tor 1, 56068 Koblenz, Germany *Corresponding author. Tel.: þ49 261 1306 5560; fax: þ49 261 1306 5363. Urs von Gunten EAWAG, Ueberlandstrasse 133, Duebendorf CH-8600, Switzerland 0043-1354/$ – see front matter ª 2010 Published by Elsevier Ltd. doi:10.1016/j.watres.2010.01.015

water research 44 (2010) 352–372



Review

Environmental fate and toxicity of ionic liquids: A review Thi Phuong Thuy Pham a, Chul-Woong Cho a, Yeoung-Sang Yun a,b,* a

Department of Bioprocess Engineering, Chonbuk National University, Jeonju, Chonbuk 561-756, Republic of Korea Division of Semiconductor and Chemical Engineering and Research Institute of Industrial Technology, Chonbuk National University, Chonbuk 561-756, Republic of Korea b

article info

abstract

Article history:

Ionic liquids (ILs) are organic salts with low melting point that are being considered as

Received 31 May 2009

green replacements for industrial volatile organic compounds. The reputation of these

Received in revised form

solvents as ‘‘environmental friendly’’ chemicals is based primarily on their negligible vapor

27 August 2009

pressure. Nonetheless, the solubility of ILs in water and a number of literature

Accepted 12 September 2009

documenting toxicity of ILs to aquatic organisms highlight a real cause for concern. The

Available online 24 September 2009

knowledge of ILs behavior in the terrestrial environment, which includes microbial degradation, sorption and desorption, is equally important since both soil and aquatic

Keywords:

milieu are possible recipients of IL contamination. This article reviews the achievements

Ionic liquids

and current status of environmental risk assessment of ILs, and hopefully provides

Toxicity

insights into this research frontier.

Degradation

ª 2009 Elsevier Ltd. All rights reserved.

Biodegradation Environmental fate Sorption

Contents 1. 2.

3.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Toxicological aspect of ILs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Effects of ILs in an enzyme level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Antibacterial activity of ILs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Toxicity of ILs to algae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Cytotoxicity of ILs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5. Phytotoxicity of ILs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6. Toxicity of ILs to invertebrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.7. Inhibitory effects of ILs on vertebrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Environmental fate of ILs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Chemical degradation of ILs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Biodegradability of ILs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Sorption of ILs in environmental systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

353 354 354 356 361 361 363 363 364 364 364 365 367

* Corresponding author. Division of Semiconductor and Chemical Engineering, Chonbuk National University, Jeonju, Chonbuk 561-756, Republic of Korea. Tel.: þ82 63 270 2308; fax: þ82 63 270 2306. E-mail address: [email protected] (Y.-S. Yun). 0043-1354/$ – see front matter ª 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2009.09.030


4.

1.

353

Concluding remarks and future directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369

Introduction

Most of volatile organic compounds (VOCs) commonly used in industrial applications cause a major concern in the current chemical processing industry. The main problems are the toxicity of the organic solvents to both the process operators and the environment as well as the volatile and flammable nature of these solvents which make them a potential explosion hazard (Schmid et al., 1998). Recently, the deleterious effects of many solvents combined with serious environmental issues, such as atmospheric emissions and contamination of aqueous effluents are making their use prohibitive. Thus, many researchers have focused on the development of ‘‘green engineering’’ which represents research aimed at finding environmentally benign alternatives to harmful chemicals. Among the neoteric solvents applicable in ‘‘green technologies’’ ionic liquids (ILs) have garnered increasing attention over the others such as supercritical CO2 (Blanchard et al., 1999; Blanchard and Brennecke, 2001; Kazarian et al., 2000) and aqueous biphasic systems (Myasoedov et al., 1995; Rogers et al., 1995; Willauer et al., 1999). Ionic liquids, formerly known as molten salts, constitute one of the hottest areas in chemistry these days. Basically, they have melting points below 100 C, which can be achieved by incorporating a bulky asymmetric cation into the structure together with a weakly-coordinating anion (Ranke et al., 2004). The unique, highly solvating, yet non-coordinating environment of ILs provides an attractive medium for various types of chemical processes. Also, the physical properties of ILs can be tailored by a judicious variation in the length and branching of the alkyl chain and the anionic precursor (Fuler et al., 1997; Huddleston et al., 2001). In this way, ILs can be made taskspecific for a certain application. The almost limitless structural possibilities of ILs, as opposed to limited structural variations within molecular solvents, make them ‘‘designer

solvents’’ (Marsh et al., 2004; McFarlane et al., 2005; Sheldon, 2005). Some independent reports (Hagiwara and Ito, 2000; Olivier, 1999; Welton, 1999) and many reviews (Earle and Seddon, 2000; Rooney and Seddon, 2001) have highlighted ILs as representing a state-of-the-art, innovative approach to sustainable chemistry, with the argument that their vapor pressure is immeasurably low and they are not flammable. Recently, the application of these liquids as reaction media for organic synthesis, catalysis, or biocatalysis has been well documented (Earle and Seddon, 2000; Wasserscheid and Keim, 2000; Welton, 1999) (Fig. 1). Gordon (2001) pointed out that there is an obvious advantage in performing many reactions in ILs due to the improvement in process economics, reaction activity, selectivity and yield. Although ILs can lessen the risk of air pollution due to their insignificant vapor pressure, they do have significant solubility in water (Anthony et al., 2001; McFarlane et al., 2005; Wong et al., 2002). As a result, this is the most likely medium through which ILs will be released into the environment. Ionic liquids currently are not widely used in industrial applications; nonetheless, continued development and further use of these solvents may lead to accidental discharge and contamination. The properties that make them be the target of industrial interest (i.e. high chemical, thermal stability and non-volatility) suggest potential problems with degradation or persistence in the environment. In general, the deficiency of information and uncertainty surrounding the environmental impact of ILs is a major barrier to the utilization of these compounds by industry. Initial efforts have been made to overcome this drawback and offer a preliminary insight into the behavior of ILs in the aqueous environments. These studies provided extensive data sets, e.g. on (eco)toxicity, biodegradability, bioaccumulation and distribution of ILs in different environmental compartments. Therefore, it is necessary to consolidate all the available data in a single

Fig. 1 – Applications of ionic liquids.

354


review to lay the groundwork for more comprehensive community and ecosystem investigations. The overall objective of this review is to systematically gather and interpret existing information about the fate, removal options and (eco)toxicological assessment strategies of ILs.

2.

Toxicological aspect of ILs

The current literature represents a number of studies addressing the biological effects of ILs evaluated on the basis of toxicological test systems. The ILs toxicities towards these systems of different levels of biological complexity as well as several environmental compartments (Fig. 2) are successively discussed in the following subsections. All the structures of IL compounds discussed in this review were listed in Table 1. The acronyms used for these substances were adapted from Ranke et al. (2007a). In this way, the cation head groups were abbreviated as ‘‘IM’’ for imidazolium, ‘‘Py’’ for pyridinium, ‘‘Pyr’’ for pyrrolidinium, ‘‘Mor’’ for morpholinium, ‘‘Pip’’ for piperidinium, ‘‘Quin’’ for quinolinium, ‘‘N’’ for quaternary ammonium and ‘‘P’’ for quaternary phosphonium. The alkyl chains attached to the head group were given as numbers corresponding to the number of carbon in the alkyl residues. For example, the 1-butyl-3-methylimidazolium moiety was denoted as IM14. In case the carbon chain length equals or exceeds 10, the numbers were separated by a hyphen (e.g. IM1-10 indicated 1-decyl-3-methylimidazolium). Particularly, for pyridinium

entities, the carbon-bound alkyl chains were appended to the head group at different positions and the abbreviation was made by noting the position of attachment and a symbol for the attached group (e.g. Py4-2Me for 1-butyl-2-methylpyridinium). The anionic components were shortened as they are in the periodic table for the halides. For tetrafluoroborate, hexafluorophosphate, bis(trifluoromethylsulfonyl)imide, dicyanamide and hydrogen sulfate the abbreviations were BF4, PF6, (CF3SO2)2N, CN(N)2 and HSO4 in respective to their structural formula.

2.1.

Effects of ILs in an enzyme level

Enzyme inhibition data by ILs include those of the acetylcholinesterase from electric eel (Electrophorus electricus) (Arning et al., 2008; Jastorff et al., 2005; Matzke et al., 2007; Ranke et al., 2007b; Stasiewicz et al., 2008; Stock et al., 2004; Torrecilla et al., 2009; Zhang and Malhotra, 2005), the AMP deaminase (Sk1adanowski et al., 2005) and the antioxidant enzyme system of mouse liver (Yu et al., 2009a). The enzyme acetylcholinesterase plays the most important role in nerve response and function. Also, acetylcholinesterase catalyzes the hydrolysis of acetylcholinesters with a relative specificity for acetylcholine, which is a neurotransmitter common to many synapses throughout mammalian nervous systems (Fulton and Key, 2001; Massoulie´ et al., 1993). Thus, an inhibition of acetylcholinesterase leads to various adverse effects in neuronal processes, such as heart diseases or myasthenia

Fig. 2 – The flexible (eco)toxicological test battery considering aquatic and terrestrial compartments as well as different trophic levels including enzymes, luminescent marine bacteria, freshwater green algae, mammalian cells, duckweed, freshwater crustacean and zebrafish (Adapted from Matzke et al. (2007) by permission of the Royal Society of Chemistry).

355


Table 1 – Selection of cationic and anionic structures of commonly used ionic liquids. Head group

R2 N

N

+

Side chain R1 ¼ -C2H5, -C3H7, -C4H9, -C5H11, -C6H13, -C7H15, -C8H17, -C9H19, -C10H21, -C14H29, -C16H33, -C18H37, -C19H39 R2 ¼ -CH3, -C2H5

R1

Imidazolium (IM)

CH3 CH3

R1

+ N

R1 ¼ -C2H5, -C3H7, -C4H9, -C5H11, -C6H13, -C8H17

CH3

Pyridinium (Py)

CH3 N

+

R1 ¼ -C4H9, -C6H13, -C8H17

R1 Pyrrolidinium (Pyr)

O

N

CH3

+

R1 Morpholinium (Mor)

R1 ¼ -C4H9

Cation

R1

+

N

CH3

R1 ¼ -C4H9

Piperidinium (Pip)

+

N

R1 ¼ -C4H9, -C6H13, -C8H17

R1 Quinolinium (Quin)

R1

R2 +

N R4

R1-4 ¼ -CH3, -C2H5, -C3H7, -C4H9, -C6H13

R3

Quaternary ammonium (N)

R1

R2 +

P

R1-4 ¼ -C4H9, -C6H13, -C14H29

R3 (

R4

Quaternary phosphonium (P

(continued on next page)

356


Table 1 (continued) Head group Anion

Side chain Cl Br

Chloride Bromide

F Tetrafluoroborate [BF4]

B

F

-

F

F F

F

F

-

P

Hexafluorophosphate [PF6]

F

F F

F3 C

S

Bis(trifluoromethylsulfonyl)imide [(CF3SO2)2N]

O N

Dicyanamide [(CN)2N]

in humans (Chemnitius et al., 1999; Pope et al., 2005). Ranke et al. (2007b) published a comprehensive collection of acetylcholinesterase inhibition values for 292 compounds covering a large variety of ILs and closely related salts. Among these data, only those of the commonly tested ILs are summarized in Table 2 for the ease of comparison of ILs toxicity from molecular up to organism levels of biological complexity. It was found that all observed inhibitory effects on the enzyme could be exclusively accounted for the cationic moiety (Arning et al., 2008). In particular, the IL with pyridinium as cationic core structure inhibited the enzyme slightly stronger than the imidazolium analogue whereas the compounds based on phosphonium was less inhibitory. All anion species exerted no effect on the enzyme activity with only exception of the fluoride anion and the fluoride containing [SbF6] and [PF6] species. Both species are known to readily undergo hydrolysis in contact with moisture and thus the fluoride seems to be the active compound. The non-inhibiting effects of anion might be explained by their limited interactions with the active site of this enzyme (Matzke et al., 2007). In addition, a correlation between an increasing chain length of the side chains connected to the cationic head groups and an enhanced inhibitory potential of the ILs was found. It is believed that the mechanism involves the similarity of the positively charged imidazolium or pyridinium to the choline part that binds to the anionic site of the enzyme, such that the longer alkyl chain results in an improved fit (Stock et al., 2004). Sk1adanowski et al. (2005) discussed the usefulness of in vitro AMP deaminase inhibition assay as a potential molecular method in prospective risk analysis of imidazolium-based ILs. The results revealed that IM14 salts associated with [PF6], [BF4], p-tosylate and [Cl] demonstrated a dose-dependent inhibition of AMP deaminase activity. The IC50 values (concentration of ILs inhibiting 50% of enzyme activity) for those containing a fluorine compartment [PF6] and [BF4] are lower (5 mM) than those for [Cl] and p-tosylate (10 mM), which indicated the adverse effect of these fluoride-containing anions. The other study on enzyme inhibition assay dealt with

O

O -

CF3 S O

N -

N

N

the effects of acute exposure of intraperitoneal injection of aqueous IM18 Br on the antioxidant enzymes of the treated mouse liver (Yu et al., 2009a). The antioxidant enzymes tested included superoxide dismutase, catalase, glutathione peroxidase and glutathione-S-transferase. The results showed that administration of IM18 Br modified activities of these defense enzymes in mouse liver, and caused damage to livers of treated mice at median lethal dose (LD50) of 35.7 mg/kg. Though data published by these authors did not cover a large variety of ILs, the enzyme inhibition assays suggest the trend in which cationic moiety is the dominating factor influencing the toxicity of ILs, especially when substituted with a long alkyl side chain. Regarding the anion types, perfluoronated ions are of toxicological interest due to hydrolysis resulting in HF formation, while the others cause less prominent effect.

2.2.

Antibacterial activity of ILs

Bacteria serve as an ideal starting point for ILs toxicity estimations as they have short generation times. Preliminary toxicological investigations have shown quaternary ammonium and pyridinium compounds have critical inhibitory effects on a variety of bacteria and fungi (Babalola, 1998; Kelman et al., 2001; Li et al., 1998). In the studies of Pernak’s group (Cieniecka-Ros1onkiewicz et al., 2005; Pernak et al., 2001a; Pernak et al., 2001b; Pernak and Chwa1a, 2003; Pernak et al., 2003; Pernak et al., 2004a), they observed a trend of increasing toxicity with an increase in the alkyl chain length substituent in the pyridinium, imidazolium and quaternary ammonium salts to various bacteria including rods, cocci and fungi. As a measure of microbial activity of imidazolium and pyridinium ILs with varying alkyl chain lengths, Docherty and Kulpa (2005) also used a group of microorganisms possessing a variety of physiological and respiratory activities. It was found that imidazolium and pyridinium bromides incorporated hexyl- and octyl-chain had considerable antimicrobial effect to pure cultures of Escherichia coli,

Table 2 – Toxicity of ILs to different levels of biological complexity including enzyme, bacteria, algae, rat cell line, human cell lines, duckweed and invertebrate. Log10EC50 (mM)a

Compound Acetylcholin esterase IM12 Cl

2.0614

IM12 IM12 IM12 IM13 IM13 IM13 IM14

2.0514 2.0514 2.0314 2.2714 2.28 0.0318 2.2214 1.91 0.0411

BF4 PF6 (CF3SO2)2N Cl BF4 PF6 Cl

1.90 0.0218

IM14 BF4

1.98 0.01811

IM14 PF6

2.15 0.0518

IM14 (CF3SO2)2N IM14 (CN)2N

1.96 0.02111 1.95 0.0718

IM15 IM15 IM15 IM16

1.9614 1.8614 1.8514 1.9214

Cl BF4 PF6 Cl

IM16 Br

N.A.

IM16 IM16 IM16 IM17 IM17 IM17 IM18

1.8814 1.8814 2.1514 2.0714 2.1214 1.9114 1.6014

BF4 PF6 (CF3SO2)2N Cl BF4 PF6 Cl

IM18 Br

N.A.

4.5510 4.33 0.1119 N.A. N.A. N.A. N.A. 3.94 0.0613 N.A. 3.71 0.144 2.955 3.34 0.136 3.47 0.0419 4.01 0.054 3.07 0.0313 3.355 3.27 0.096 3.55 0.0413 3.10 0.176 3.12 0.3516 3.07 0.296 3.39 0.084 3.67 0.104 2.995 N.A. 3.14 0.0213 N.A. 1.9410 2.32 0.166 2.91 0.0913 1.42 0.124 0.815 3.18 0.0313 2.17 0.066 N.A. N.A. 2.44 0.0613 N.A. 1.19 0.116 1.01 0.0619 0.63 0.074 0.075

MCF7b

Lemna minor

N.A.

N.A.

N.A.

3.4414 3.9214 N.A. >4.3014 3.4714 >3.0014 3.5514

4.00 0.0420 N.A. 3.26 0.0420 N.A. N.A. N.A. N.A.

N.A. N.A. N.A. N.A. N.A. N.A. N.A.

N.A. N.A. N.A. N.A. N.A. N.A. 2.8211

N.A. N.A. N.A. N.A. N.A. N.A. 1.934 1.93 0.061

N.A.

3.4314

3.44 0.1120

N.A.

N.A.

1.574 1.56 0.071 1.85 0.0622

N.A.

2.1111

3.1214

3.72 0.0517 3.66 0.0820

N.A.

2.497

1.684 1.67 0.111

4.15 0.069

2.20 0.0421

N.A.

3.1014

4.14 0.2217

N.A.

N.A.

2.55 0.159 N.A.

1.80 0.0712 N.A.

1.81 0.1511 N.A.

2.6814 3.1514

3.07 0.0820 N.A.

N.A. N.A.

2.45 0.0811 N.A.

1.854 1.85 0.101 N.A. N.A.

N.A. N.A. N.A. N.A.

N.A. N.A. N.A. 1.92 0.0921

N.A. N.A. N.A. 0.0819

>3.0014 >3.0014 >3.0014 2.8514

N.A. N.A. N.A. N.A.

N.A. N.A. N.A. N.A.

N.A. N.A. N.A. N.A.

N.A. N.A. N.A. N.A.

N.A.

2.57 0.152

N.A.

N.A.

N.A.

N.A.

N.A.

N.A. 3.25 0.679 2.53 0.159 N.A. N.A. N.A. N.A.

N.A. N.A. N.A. N.A. N.A. N.A. 1.4621

N.A. N.A. N.A. N.A. N.A. N.A. 2.67 0.3719

2.9814 2.9114 2.2414 2.5314 2.5814 2.3014 2.0114


N.A. N.A. 2.818 N.A. N.A. N.A. N.A.


0.784 1.06 0.0422 N.A. N.A. N.A. N.A. N.A. N.A. N.A.

N.A.

1.65 0.252

N.A.

2.48 0.0420

N.A.

N.A.

Escherichia coli

Pseudo kirchneriella subcapitata

Scenedesmus vacuolatus

IPC-81

N.A.

N.A.

2.78 0.0619

N.A.

5.25 0.069 N.A. N.A. N.A. N.A. N.A. N.A.

N.A. N.A. N.A. N.A. N.A. N.A. 2.34 0.0121

N.A. N.A. N.A. N.A. N.A. N.A. 2.26 0.0811

N.A.

3.46 0.0623

4.60 0.029

N.A.

HeLa

Daphnia magnac N.A.


IM14 Br

Vibrio fischeri

1.334 0.54 0.1222

357


358

Table 2 (continued) Log10EC50 (mM)a

Compound Acetylcholin esterase 1.53 0.02511 2.0314 2.0314 N.A. 1.0914

IM1-10 BF4 IM1-10 PF6 IM1-14 Cl IM1-16 Cl IM1-18 Cl IM1-19 Cl IM1-19 BF4 IM1-19 PF6 IM22 Br IM23 Br IM24 BF4 IM25 BF4 IM26 Br IM26 BF4 IM2-10 Br Py Cl Py2 Cl Py3 Br Py3 (CF3SO2)2N Py4 Cl

1.10 0.0418 1.6814 0.5414 0.6814 0.9614 1.3614 1.4314 1.6214 2.0814 2.2114 2.03 0.0118 N.A. 1.7714 1.8414 0.9214 >3.0014 2.1014 2.2214 2.2114 1.7014

Py4 Br

1.7714

Py4 BF4 Py4 PF6 Py4 (CN)2N

1.8014 1.8414 N.A.

Py5 Br Py5 (CF3SO2)2N Py6 Cl Py6 Br Py6 PF6 Py6 (CF3SO2)2N Py8 Cl Py8 (CF3SO2)2N Py4-2Me Cl Py4-2Me BF4

1.5214 1.5514 1.7214 N.A. 1.7614 1.8514 1.6014 1.4014 0.7014 0.8214

Escherichia coli

Pseudo kirchneriella subcapitata

Scenedesmus vacuolatus

IPC-81

HeLa

MCF7b

Lemna minor

1.41 0.0713 0.95 0.126 N.A. 0.72 0.0413 0.50 0.0713 0.23 0.0619 0.18 0.0613 N.A. 0.15 0.0719 0.23 0.0819 1.45 0.0519 N.A. N.A. N.A. N.A. N.A. 2.8 0.0413 3.1413 N.A. 2.15 0.0513 N.A. N.A. N.A. N.A. N.A. 3.41 0.084 2.645 3.18 0.0619 3.40 0.014 2.735 N.A. N.A. 3.31 0.104 2.615 N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A.

N.A. 2.64 0.159 N.A. N.A. N.A.

N.A. N.A. N.A. N.A. N.A.

2.3011 N.A. N.A. N.A. 3.57 0.0619

1.5914 1.9614 1.6414 N.A. 1.3414

2.48 0.0220 N.A. 2.28 0.0220 N.A. N.A.

2.848 N.A. N.A. N.A. N.A.

0.907 N.A. N.A. N.A. N.A.

N.A. N.A. N.A. N.A. N.A.

N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A.

N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. 2.57 0.0621

N.A. N.A. 2.48 0.219 >2.0019 >2.0019 N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. 2.59 0.1119

0.7714 1.5014 0.4214 0.1914 0.0114 1.4014 1.6514 1.8514 >3.0014 >3.3014 3.2614 N.A. 2.0114 2.2614 0.5314 N.A. N.A. N.A. N.A. N.A.

N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. 4.36 0.0917 N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A.


N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. 2.32 0.1819


N.A.

N.A.

N.A.

3.9014

3.50 0.0720

N.A.

N.A.

N.A.

N.A. N.A. N.A.

N.A. N.A. N.A.

N.A. N.A. N.A.

3.1814 N.A. N.A.

N.A. N.A. N.A.

N.A. N.A. N.A.

N.A. N.A. N.A.

N.A. N.A. N.A.

N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A.



N.A. N.A. N.A. N.A. N.A. N.A. 1.2714 N.A. N.A. 3.2514




N.A. N.A. N.A. 1.074 N.A. N.A. N.A. N.A. N.A. N.A.

Daphnia magnac


IM18 BF4 IM18 PF6 IM18 (CF3SO2)2N IM19 BF4 IM1-10 Cl

Vibrio fischeri

1.1514 N.A.

Py4-3Me BF4 Py4-3Me PF6 Py4-3Me (CN)2N

1.53 0.0218 1.45 0.0218 1.2214

Py6-3Me Cl Py6-3Me Br

1.0614 N.A.

Py6-4Me Py6-4Me Py8-3Me Py8-3Me

Cl BF4 Cl Br

1.4414 1.4814 0.6414 N.A.

Py8-4Me Cl Py8-4Me BF4 Pyr14 Cl Pyr14 Br

N.A. 3.46 0.0623

N.A. N.A.

N.A. N.A.

N.A. N.A.

N.A. N.A.

N.A. N.A.

N.A. 1.764

N.A. N.A. N.A.

N.A. N.A. N.A.

N.A. N.A. N.A.

3.3014 N.A. N.A.

N.A. N.A. N.A.

N.A. N.A. N.A.

N.A. N.A. N.A.

N.A. N.A. N.A.

N.A. N.A.

N.A. N.A.

N.A. N.A.

N.A. N.A.

N.A. N.A.

N.A. N.A.

N.A. N.A.

N.A. 0.594

N.A. N.A. N.A. N.A.

N.A. N.A. N.A. N.A.

N.A. N.A. N.A. N.A.

N.A. 2.1714 N.A. N.A.

N.A. N.A. N.A. N.A.

N.A. N.A. N.A. 1.008

N.A. N.A. N.A. N.A.

N.A. N.A. N.A. 0.404

N.A. N.A. N.A. N.A.

N.A. N.A. N.A. 3.67 0.283

N.A. N.A. 3.37 0.1019 N.A.

1.6314 1.4914 >4.3014 3.7714

N.A. N.A. N.A. N.A.

N.A. N.A. 2.16 0.2519 N.A.

N.A. N.A. N.A. N.A.

Pyr14 BF4 Pyr14 (CF3SO2)2N Pyr14 (CN)2N Pyr16 Cl Pyr16 (CF3SO2)2N Pyr18 Cl Pyr18 BF4 Pyr66 Mor14 Cl Mor14 Br Mor14 (CF3SO2)2N Pip14 Br

1.9114 2.1314 1.9814 2.4814 2.6014 2.3614 2.0214 2.0814 N.A. 2.7114 2.7814 1.8314

N.A. N.A. N.A. 2.9910 N.A. N.A. N.A. N.A. >4.3019 N.A. N.A. 4.27 0.0919

N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A.

N.A. >2.3812 N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A.

N.A. 2.5319 N.A. N.A. N.A. N.A. N.A. N.A. N.A. >4.0019 2.0019 3.27 0.1219

2.9014 3.0114 4.2314 2.9114 N.A. 2.5914 1.8214 1.2314 N.A. >4.3014 3.4314 4.0314


N.A. 2.98 0.3219 N.A. N.A. N.A. N.A. N.A. N.A. N.A. 3.11 0.1319 3.15 0.1319 0.4719


Pip14 (CF3SO2)2N Quin4 Br Quin4 BF4 Quin6 BF4 Quin8 Br Quin8 BF4 N1111 Br N1114 (CF3SO2)2N N1123 (CF3SO2)2N N1124 Cl N1124 (CF3SO2)2N N2222 Cl N2222 Br N2226 Br N4444 Br P4444 Br P666-14 Br

1.7814 0.7914 0.6214 0.4814 N.A 0.3014 N.A. 2.6014 2.3414 2.0614 2.0314 2.8014 N.A. N.A. 2.3014 2.6114 2.8514

N.A. N.A. N.A. N.A. N.A. N.A. >5.004 N.A. N.A. N.A. N.A. N.A. >5.004 2.46 0.164 3.27 0.074 2.714 3.41 0.024

N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A.

N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A.

2.0819 N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. >4.0019 1.78 0.1719 N.A. N.A. N.A. N.A. N.A. N.A.

3.4114 2.3214 2.1614 1.0714 0.0314 0.1714 N.A. 3.6114 N.A. >4.3014 3.4313 >3.4814 N.A. N.A. 2.2514 1.6614 N.A.

N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. 4.26 0.0420 N.A. N.A. N.A. N.A.

N.A. N.A. N.A. 4.64 0.0215 4.258 N.A. 3.148 N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. 4.15 0.0415 4.158 2.938 N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A.

2.85 0.0719 N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. 0.83 0.6719 N.A. N.A. N.A. N.A. N.A. N.A. N.A.

N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. 1.474 0.954 N.A.


359

N.A. N.A.

1.1114 1.2214 1.9214 1.9314

N.A. 2.75 0.134 2.125 N.A. N.A. 2.66 0.054 1.995 1.4410 2.06 0.164 1.485 N.A. N.A. N.A. 0.79 0.054 0.255 N.A. N.A. >4.3019 N.A.


Py4-3Me Cl Py4-3Me Br

360

References: 1Bernot et al. (2005a); 2Cho et al. (2007); 3Cho et al. (2008a,b); 4Couling et al. (2006); 5Docherty and Kulpa (2005); 6Garcia et al. (2005); 7Jastorff et al. (2005); 8Kumar et al. (2009); 9Lee et al. (2005); 10 Luis et al. (2007); 11Matzke et al. (2007); 12Pretti et al. (2008); 13Ranke et al. (2004); 14Ranke et al. (2007b); 15Salminen et al. (2007); 16Samorı` et al. (2007); 17Stepnowski et al. (2004); 18Stock et al. (2004); 19 Stolte et al. (2007a); 20Wang et al. (2007); 21Wells and Coombe (2006); 22Yu et al. (2009). a N.A. means not available (not determined). b Toxicity of ILs is expressed as log10IC50 (mM) in case of MCF7 cell line. c Toxicity of ILs is expressed as log10LC50 (mM) in case of D. magna.

N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. 1.90 0.0520 N.A. 0.4814 N.A. 0.2414 N.A. N.A. N.A. N.A. N.A.

Scenedesmus vacuolatus Pseudo kirchneriella subcapitata

N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. 3.47 0.0818 >3.3017 >3.4814 3.40 0.218 BF4 PF6 (CF3SO2)2N (CN)2N P666-14 P666-14 P666-14 P666-14

Compound

Table 2 (continued)

Acetylcholin esterase

Vibrio fischeri

Escherichia coli

Log10EC50 (mM)a

IPC-81

HeLa

MCF7b

Lemna minor

Daphnia magnac


Staphylococcus aureus, Bacillus subtilis, Pseudomonas fluorescens and Saccharomyces cerevisiae. The anion performed nearly no effect on antimicrobial activity in the case of imidazolium analogues (Docherty and Kulpa, 2005; Garcia et al., 2005; Lee et al., 2005; Pernak et al., 2003; Pernak et al., 2004a) whereas this was not the case for phosphonium salts. Within the group of alkyltrihexylphosphonium ILs in the study of CienieckaRos1onkiewicz et al. (2005), both cation structure and the type of anion had effects on the biological activity. The antibacterial activity of ILs not only involves in hampering the growth rate of microbes but also interferes with their productivity. Matsumoto et al. (2004a) tested the toxicity of imidazolium-based ILs to lactic acid producing bacterium Lactobacillus rhamnosus to examine whether these compounds can replace conventional organic solvents in the extractive fermentation of lactate. The results showed that the bacterium L. rhamnosus grew, consumed glucose, and produced lactate in the presence of imidazolium-based ILs. A change of alkyl length in the imidazolium cation had little difference on the survival of the cells. In a similar study (Matsumoto et al., 2004b), they focused on hiochii bacteria, Lactobacillus homochiochii and Lactobacillus fructivorans and also found that the bacteria could produce lactic acid in the presence of ILs. Nonetheless, the lactic acid producing activities of these bacteria generally decreased with the extension of alkyl chain length in the imidazolium cation moiety. Water miscible ILs had various effects on the physiology of Clostridium sporogenes when tested as additives in culture media or reaction media for reduction of nitrobenzene (Dipeolu et al., 2008). In their study, 2-hydroxyethyltrimethylammonium dimethylphosphate and N,N-dimethylethanolammonium acetate increased the growth rate of C. sporogenes; by contrast, IM14 BF4 and IM12 EtSO4 inhibited growth. Although IM12 EtSO4 inhibited growth, it was sufficiently non-toxic to allow efficient reduction of nitrobenzene using harvested cells. Thus, it is recommended that both noninhibitory and partially inhibitory ILs should be screened for use in biotransformation. Nonetheless, Ganske and Bornscheuer (2006) referred that ILs could have substantial inhibitory effects on the growth of microorganisms when they explored the effects of the two most commonly used ILs IM14 BF4 and IM14 PF6 on the growth of E. coli, Pichia pastoris and Bacillus cereus. Regarding inhibition assays used in assessment of environmental potential risk of a compound in aquatic milieu, the bioluminescence assay using Vibrio fischeri (formerly known as Photobacterium phosphoreum) is one of the most applied (Kaiser and Palabrica, 1991; Steinberg et al., 1995). This is a rapid, costeffective, and well-established method for toxicity determination focusing on environmental issues, and also a standard ecotoxicological bioassay in Europe (DIN EN ISO 11348). The published data on ILs toxicity towards V. fischeri were listed on Table 2 and were comprehensively interpreted in the study of Peraccini et al. (2007). Although it has been claimed that modifications of the anion lead to changes in chemical and physical properties of ILs (Sheldon, 2001), no clear increase in toxicity caused by the anion could be observed, and toxicity seemed to be determined mainly by the cationic component (Ranke et al., 2004). This is likely explained by the fact that lipophilic part of the molecules can be intercalated into the


membrane, whereas their ionic head group is at least partially solvated in the aqueous solution, as suggested by Austin et al. (1998). The ILs toxicity was also observed to correlate directly with the length of the n-alkyl residues in the methylimidazolium cation (Romero et al., 2008). Interestingly, Ranke et al. (2004) noted a slight hormetic effect at concentrations below inhibitory concentrations. Concerning the anionic influence, compounds with [PF6] were found to be slightly more toxic than compounds with other anions in their study (Ranke et al., 2004). The anion [(CF3SO2)2N] showed no intrinsic toxicity to V. fischeri in the report of Matzke et al. (2007); in contrast, an increased in toxicity was found for all tested compounds combined with [(CF3SO2)2N] for V. fischeri (Stolte et al., 2007a). Couling et al. (2006) extended the bioluminescence inhibition assay to pyridinium derivatives and it was noted that the quaternary ammonium compounds seemed to be less toxic to V. fischeri than the pyridinium and imidazolium analogues. Also, the quantitative structureproperty relationship (QSPR) modeling suggested that imidazolium cations, with two nitrogen atoms, are predicted to be more toxic than pyridinium moieties, which only have one nitrogen atom in the structure. In addition, the QSPR correlation predicted that quaternary ammonium cations are less toxic than those with cations containing nitrogen-bearing rings, which was in agreement with the experimental results (Couling et al., 2006). However, in contrast to the cases of aromatic ILs and ammonium compounds, the authors were unsuccessful in modeling the behavior of phosphonium salts using the developed correlation.

2.3.

Toxicity of ILs to algae

As algae are primary producers, either directly or indirectly, of organic matter required by animals in freshwater food chains, their ecology is crucial in providing the energy for sustaining other higher trophic levels. The ubiquity of algae makes these organisms ideal for toxicological studies and, because they have a short life cycle they can respond quickly to environmental change (Blaise, 1993; Lewis, 1995). To date, several groups have focused their attention on the use of algal primary producers to assess the effects of ILs to aquatic environments (Cho et al., 2007; Cho et al., 2008a,b,c; Grabinska-Sota and Kalka, 2006; Kulacki and Lamberti, 2008; Lata1a et al., 2005; Matzke et al., 2007; Matzke et al., 2008; Pham et al., 2008a,b; Pretti et al., 2009; Stolte et al., 2007a; Wells and Coombe, 2006). Cho and co-workers used Pseudokirchneriella subcapitata (formerly known as Selenastrum capricornutum) to study the effect of different head groups, side chains and anions of ILs on algal growth rate and photosynthetic activity. The data revealed that the toxic influence of ILs on growth rates were more significant than those of photosynthetic performance (Pham et al., 2008b). Once again, the trend of increasing toxicity with increasing alkyl chain length was observed in their reports (Cho et al., 2007; Pham et al., 2008b). Regarding the anionic effects, P. subcapitata was sensitive to the anion moieties in the order: [SbF6] > [PF6] > [BF4] > [CF3SO3] > [C8H17OSO3] > [Br] z [Cl]. In particularly, it was found that with respect to IL incorporating perfluorinated anion (i.e. IM14 BF4), EC50 values (concentrations which lead to a 50% reduction of the exposed organisms

361

relative to control) of the previously prepared stock solution (6 months prior to experiment) were significantly lower compared to those of the freshly made one (Pham et al., 2008a). This might be due to hydrolytic effects of IM14 BF4 leading to fluoride formation, as confirmed by ion chromatography analysis. This implies that after ILs are released into the aqueous system; they can become more hazardous than expected by laboratory data with fresh ILs. In a detailed study on hydrolysis of fluoride-containing anions, Cho et al. (2008a) showed that IM14 SbF6 generated a greater amount of fluoride compared to IM14 BF4, but no fluoride formation occurred with the hexafluorophosphate. When only small amounts of fluoride ions were formed from IM14 SbF6 and IM14 BF4 within 96 h, the formed fluoride ion did not affect the algal growth rate. Nevertheless, the fluoride ion formation from IM14 BF4 increased with incubating time of the stock solution; thus, the toxicity might significantly increase according to the further formed fluoride ions. In view of cationic effect, Pyr14 Br was found to be the least toxic of all the ILs tested to P. subcapitata (Cho et al., 2008b). For the limnic green alga Scenedesmus vacuolatus, a severe toxicity was found for 1-butyl-4-(dimethylamino)pyridinium, whereas the quaternary ammonium and morpholinium compounds exhibited no toxicity (Stolte et al., 2007a). Despite the extensive studies on the toxicological impact of ILs towards freshwater phytoplankton, inhibition mechanism of both the growth rate and photosynthetic activity by ILs has not been described by the authors. Lata1a et al. (2005), who selected two marine algae Oocystis submarina (green algae) and Cyclotella meneghiniana (diatom) as testing organisms, found that the two species differed dramatically in their ability to recover from IL exposure. Additionally, it was discovered that IL toxicity declined with increasing salinity. The lower toxicity of IL in this case is probably due to the reduced permeability of IL cations through the algal cell walls. High amounts of chloride provide a good ion-pairing environment for imidazolium cations, which consequently compete with hydroxyl or silanol functional groups in the cell-wall structure of green alga and diatom, respectively. Though no information on EC50 values was described, the facts emerged from this work provide useful information in the further fate assessment of ILs in marine environments.

2.4.

Cytotoxicity of ILs

As a cellular test system, promyelotic leukemia rat cell line IPC-81 has been frequently used in cytotoxicity assays of ILs, with the reduction of the WST-1 dye as an indicator of cell viability (Matzke et al., 2007; Ranke et al., 2004; Ranke et al., 2007a; Stasiewicz et al., 2008; Stolte et al., 2006; Stolte et al., 2007b; Torrecilla et al., 2009). It was observed that ILs with polar ether, hydroxyl and nitrile functional groups within the side chains exhibited low cytotoxicity compared to those incorporated with ‘‘simple’’ alkyl side chains (Kumar et al., 2009; Stasiewicz et al., 2008; Stolte et al., 2007b). Those functional groups were thought to impede cellular uptake by membrane diffusion and reduce lipophilicity based interactions with the cell membrane (Stolte et al., 2007b). Taking a closer look at the effects of sub-structural elements of ILs, [(CF3SO2)2N] anion and 4-(dimethylamino)pyridinium cation

362


were described to have intrinsic effects of anion and head group on cytotoxicity, respectively (Stolte et al., 2007b). The well known side chain length effect (decrease in EC50 values with elongation of the alkyl side chain) could also be confirmed in these studies. To date many studies have analyzed the toxicity of ILs on human cell lines (Frade et al., 2007; Garcıá-Lorenzo et al., 2008; Hassoun et al., 2002; Kumar et al., 2009; Salminen et al., 2007; Stepnowski et al., 2004; Wang et al., 2007). These in vitro systems have been extremely beneficial in studying the molecular basis of chemical’s biological activity, including its toxic mode of action (Blaauboer et al., 1998) and could facilitate extrapolation of in vitro data with regard to possible effects on humans (Malich et al., 1997). Most of studies dealt with HeLa cells exemplifying prototypical cells of the human epithelium which is normally the site of first contact of an organism with toxicants. According to Stepnowski et al. (2004), the cytotoxicity data implied that effects of IM14 cation coupled with chloride, tetrafluoroborate or hexafluorophosphate were probably dependent on the anionic moieties. The lowest effect concentrations for tetrafluoroborate species were found to be 0.63 mM, whereas hexafluorophosphate and chloride inhibited HeLa cell growth at comparably high concentrations of >10 mM. Surprisingly, when the anion effect was compared, the strongest inhibition was found for [PF6]. This might be due to hydrolysis affecting fluoride formation, thus causing serious toxicological consequences through the decomposition product. A similar phenomenon was observed by Ranke et al. (2004) in IPC-81 leukemia cells, where the lower toxicity of 1-n-butyl-3methylimidazolium hexafluorophosphate in comparison to the hexafluorophosphate anion alone was explained by reduced anion uptake due to the formation of an ion pair. The anion in this ion pair can, however, also be partially decomposed. This was shown in recent work by Swatloski et al. (2003), who identified traces of 1-n-butyl-3-methylimidazolium fluoride hydrate as a decomposition product formed during the purification of the 1-n-butyl-3-methylimidazolium hexafluorophosphate. As shown by Wang et al. (2007) the phosphonium bis(trifluoromethylsulfonyl)imide salts performed the highest inhibitory to HeLa cells, followed by alkylimidazolium, alkylpyridinium, alkyltriethylammonium and N-alkyl-N,Ndimethyl-N-(2-hydroxylethyl)ammonium salts, in decreasing order. For each cation class the toxicity increased with increasing chain length of the alkyl substituent for a given anion: 1-ethyl-3-methylimidazolium bromide yielded an EC50 of 8.4 mM, substituting the ethyl moiety for a butyl group led to an EC50 of 2.8 mM, and for an octyl moiety an EC50 of 0.3 mM. This result was consistent with what has been observed in other studies. Salts containing the tetrafluoroborate anion showed the highest EC50, followed closely by bromide and chloride. Bis(trifluoromethylsulfonyl)imide salts were significantly more toxic than their halide counterparts. However, the effect of changing the anion was smaller than that of changing the alkyl substituent, e.g. while 1-ethyl-3-methylimidazolium tetrafluoroborate was observed to have an EC50 of 9.9 mM, the corresponding bromide and bis(trifluoromethylsulfonyl)imide salts had EC50 of 8.4 and 1.8 mM, respectively – these all considerably less toxic than 1-octyl-3-methylimidazolium bromide.

The CaCo-2 cells were used in the study of Garcıá-Lorenzo et al. (2008) with the aim of a convenient screening method for obtaining first rough estimates for the toxic potential of ILs. The obtained data showed that in general, ILs with longer alkyl chains were more lipophilic than those with shorter alkyl chains. The former can be presumed to have a tendency to be incorporated into the phospholipid bilayers of biological membranes. In this respect, some authors have indicated that the increased toxicity of longer ILs can be accounted for enhanced membrane permeability altering the physical properties of the lipid bilayer (Lata1a et al., 2005; Ranke et al., 2004; Stepnowski et al., 2004). Additionally, it has been proposed that the mode of toxic action for ILs takes place through membrane disruption because of the structural similarity of imidazolium-based ILs to detergent, pesticides and antibiotics able to cause membrane-bound protein disturbance (Docherty and Kulpa, 2005). Recently, Ranke et al. (2007a,b) have demonstrated that lipophilicity of ILs dominates their in vitro cytotoxicity over a wide range of structural variations. The contribution of the anionic part of the ILs to the observed biological effect was evaluated by comparing the EC50 values obtained for the cations IM16 and IM18, combined with two different anions [Cl] and [PF6]. For both cations, a stronger toxic effect was found for chloride derivatives, but not for fluoride containing hexafluorophosphate. A similar result was reported by Stock et al. (2004) where the inhibitory effects of IM14 Cl and IM14 PF6 on the acetylcholinesterase activity were compared. In addition, slightly higher cytotoxicity for the chloride derivative has also been observed when the cytotoxicity of IM14 Cl and IM14 PF6 on HeLa cells was tested (Stepnowski et al., 2004). This implies the effect of perfluorinated ions is not drastic to all but vary according to species of organisms tested. Several authors have pointed out that altering the anion has only minimal effects on the toxicity of several imidazolium compounds (Bernot et al., 2005a; Garcia et al., 2005; Ranke et al., 2004). This indicates that ILs toxicity seems to be related to the alkyl chain branching and to the hydrophobicity of the imidazolium cation but not to the various anions. In this respect, a recent study using the IPC-81 rat leukemia cell line with a large pool of anions demonstrated that most of the commercially available anions showed no or only marginal cytotoxic effects. However, anionic compartments with lipophilic and hydrolysable structural elements are likely to be of considerable relevance with respect to the toxicity of ILs (Stolte et al., 2006). In a recent study (Frade et al., 2007), the human cell lines such as HT-29 and CaCo-2 cells were utilized to estimate the inhibitory effect of ILs with several types of cations and anions. In both cells, IM14, IM12OH (1-(2-hydroxyethyl)-3methylimidazolium), IM12O2O1 (1-(2-(2-methoxyethoxy)ethyl)-3-methylimidazolium) and cholines were the least toxic cations independently of the anion. Within the studied combinations, it can be noted that IM14 PF6, IM14 acesulfame, IM12OH BF4/PF6, IM12O2O1 BF4/PF6, IM12OH acesulfame and IM12OH saccharine are not toxic and present good alternatives to organic solvents. Meanwhile, increasing the length of the substituent chain may contribute to a significant increasing of imidazolium toxicity. It was also noted that [(CF3SO2)2N] anion decreased the toxicity to a large extent,


independently of the cation and for both cell types, which was in accordance with Salminen et al. (2007).

2.5.

Phytotoxicity of ILs

The studies on phytotoxic activity of ILs were conducted mostly on the duckweed, Lemna minor, a common aquatic vascular plant (Jastorff et al., 2005; Larson et al., 2008; Matzke et al., 2007; Stolte et al., 2007a). In general, 1-alkyl-3methylimidazolium compounds with longer alkyl chains were more toxic to L. minor than those with short alkyl chain lengths. Imidazolium and pyridinium cations with butyl groups had similar EC50s (the concentrations that produced a 50% reduction in root growth) (39.07 and 32.54 mM, respectively); while the equivalent ammonium cation had a much higher EC50 (101.48 mM; i.e., less toxic) (Larson et al., 2008). In consideration of anionic effect, [(CF3SO2)2N] was found to cause moderate toxicity to this duckweed (EC50 ¼ 6300 mM) (Matzke et al., 2007). On the other hand, this anion had no or even a positive influence on the observed effects on L. minor (Stolte et al., 2007a). Focusing on the terrestrial environment, Matzke et al. (2009a) investigated the influence of differently composed soils, with varying contents of the clay minerals smectite and kaolinite, on the toxicity of different anion species of imidazolium-based ILs towards the wheat Triticum aestivum. The data showed that IM14 (CF3SO2)2N appeared the most toxic, independently of the type and concentration of added clay. This is totally in contrast to the findings of Stolte et al. (2007a), who reported that [(CF3SO2)2N] caused no harm to L. minor, indicating the toxic effect of this anion is different between certain plants. The toxicity of 1-butyl-3-methylimidazolium incorporated chloride, tetrafluoroborate and hydrogen sulfate was mainly controlled by the cationic moiety. The observed effects varied according to the added clay type and clay concentration. An increase of clay content resulted in less inhibitory effects of these substances. On the contrary, for IM14 combined with bis(trifluoromethylsulfonyl)imide the addition of clay minerals led to higher toxicity compared to the reference soil. Since results are contradictious further study is necessary to unravel the underlying mechanism. Moreover, a detailed study on the effect of IM14 BF4 on the wheat T. aestivum seedlings (Wang et al., 2009) showed that IM14 BF4 was hazardous to the early development of wheat and had varying effects on different organs. At low concentrations, IM14 BF4 did not inhibit, and even promoted, wheat seedling growth. Nonetheless, at high concentrations, this IL inhibited wheat seedling growth significantly and decreased chlorophyll content, thereby reducing photosynthesis and plant growth. Therefore, the authors suggested that dilution could decrease the toxicity of IM14 BF4 to plants and would be a good method for remediating IL-polluted environments. In another research, the phytotoxicity tests of chiral ILs containing (-)-nopyl derivatives were carried out in a plant house using spring barley (Hordeum vulgare) which is a monocotyledonous plant, and a common radish (Raphanus sativus L. subvar. radicula Pers.) which is a dicotyledonous plant (Ba1czewski et al., 2007). According to the data obtained, increasing the concentration of ILs resulted in a systematic decrease in the crop fresh weight of total sprouts and the crop fresh

363

weight per plant, both for spring barley and for common radish. It could also be noted that common barley was a more resistant plant which fairly well tolerates test IL concentrations up to 200 mg kg1 of soil; whereas, for radish, the growth and development inhibiting concentration is 100 mg kg1 of soil. Using the same target plant (H. vulgare), Pernak et al. (2004b) reported that the 1,3-dialkoxymethylimidazolium tetrafluoroborate salts introduced to the soil at concentration of 1,000 mg kg1, or 100 mg kg1 dry mass of soil, were found to exert a phytotoxic effect on monocotyledonous plants. On the other hand, at a concentration of 10 mg kg1 no such effect on the growth of the roots was notified. Concerning phytotoxicity of ILs to garden cress (Lepidium sativum L.) in soil environment, Studzin´ska and Buszewski (2009) have proved that hazardous effects of imidazolium ILs are closely connected with organic matter content in soil. Soil with more organic carbon was observed to sorb IL cations more extensively than soil with little or no organic matter; hence, the more fertile in soil, the lower probability of hazardous effect of ILs to plants. On the other hand, the hazardous character of analyzed ILs was strongly connected with their hydrophobicity, indicating that the more hydrophobic IL, the higher decrease of seed germination. Although intensive work has not been conducted on phytotoxic influence of ILs, the available data offer initial hints for environmental scientists dealing with the potential impact of ILs towards aqueous and terrestrial plants.

2.6.

Toxicity of ILs to invertebrates

Ecotoxicological literature of ILs to invertebrates mainly focus on the use of Daphnia magna as a test organism (Bernot et al., 2005a; Couling et al., 2006; Garcia et al., 2005; Grabinska-Sota and Kalka, 2006; Luo et al., 2008; Nockemann et al., 2007; Pretti et al., 2009; Samorı` et al., 2007; Wells and Coombe, 2006; Yu et al., 2009b). Daphnia is an important link between microbial and higher trophic levels (McQueen et al., 1986), and has been the subject of hundreds of intensive ecological studies. The results of all studies again observed the well-established link between toxicity and alkyl chain length of the tested ILs containing imidazolium, pyridinium or quaternary ammonium as counter cations. The most toxic compound towards D. magna was found to be IM18 Br whereas the least toxic one was IM14 Cl with log10EC50 values of 1.33 and 1.93, respectively (Table 2). Also, the nature of the anion was suggested to have smaller effects compared to those of the cation. In a recent study, Luo et al. (2008) investigated the developmental toxicity of IM18 Br on D. magna. It was found that this compound exhibited toxicity on the development of three generation of D. magna with the decrease of number of offspring and average brood size correlated to increasing IM18 Br concentrations. This indicated that IM18 Br could cause deleterious effect to the population of Daphnia and indirectly disturb freshwater food webs. Couling et al. (2006) used experimental data to determine which part of the IL molecule is responsible for the observed toxic effects through a quantitative structure-property relationship (QSPR) modeling. In this respect, correlative and predictive equations were generated and proved that there was a distinct influence of the length of alkyl residues attached to the aromatic nitrogen atoms

364


towards D. magna. Moreover, the models predicted that the toxicity increased slightly with increasing number of aromatic nitrogen atoms in cation ring. This implies that ammonium salts are less toxic than pyridinium salts, which in turn are less toxic than imidazolium moieties. Interestingly, it was noted that methylating the aromatic carbons could be effective in reducing toxicity to D. magna, indicating that 1-nbutylpyridinium bromide can be more toxic than 1-n-butyl-3methylpyridinium bromide, which is more toxic than 1-nbutyl-3,5-dimethylpyridinium analogue. The QSPR, though is still in its infancy, has contributed initial guidelines for a rationale design of a new category of ILs with an acceptable environmental profile. Other studies include data on the snail Physa acuta (Bernot et al., 2005b), the spring tail Folsomia candida (a soil invertebrate) (Matzke et al., 2007), Caenorhabditis elegans (a soil roundworm) (Swatloski et al., 2004) and Dreissena polymorpha (zebra mussel) (Costello et al., 2009). It was also demonstrated a positive relationship between alkyl chain length and toxicity in these reports. In the research of Bernot et al. (2005b), the estimated LC50 (median lethal concentration) ranged from 3.50 to 1799.8 mM (0.54 to 3.26 in the logarithmic form), which implied that P. acuta are less sensitive to ILs than are D. magna (log10LC50 (mM) ranging from 1.33 to 1.93 (Table 2)). Also, the authors observed that at low concentrations, the IL may suppress snail movement, but concentrations above this threshold level trigger an escape response, causing the organism to move faster. Grazing patterns, nonetheless, showed that snails grazed less at higher IL concentrations. Physa spp. are key components of freshwater food webs, because they graze algae and are themselves important prey for fish and invertebrate predators (Bernot and Turner, 2001; Osenberg and Mittelbach, 1989). Thus, nonlethal IL concentrations affected P. acuta behaviors, potentially influencing individual fitness and good web interactions.

2.7.

being 42.4, 43.4 and 85.1 mg/L, respectively, indicating that the developmental toxicity of IM18 Br in the frog was stagesensitive. The number of dead embryos was also found to increase with the increasing concentrations of the IL IM18 Br. The developmental impact of IM18 Br was claimed not only in this finding but also in the work of Luo et al. (2008), who investigated on D. magna. Other work in the literature has focused on the acute toxicity of ILs on rats and mice (Bailey et al., 2008; Cheng et al., 2009; Landry et al., 2005; Pernak and Czepukowicz, 2001; Sipes et al., 2008). The values of acute toxicity of 3-hexyloxymethyl1-methylimidazolium tetrafluoroborate were found to be LC50 ¼ 1400 and 1370 mg kg1 for female and male Wistar rats, respectively (Pernak and Czepukowicz, 2001). Bailey et al. (2008) studied the effects of prenatal exposure of mice to IM14 Cl due to the potential for human exposure as a result of water or soil contamination from industrial effluent or accidental spills. As shown in the experimental data, after being contacted to the IL, fetal weight was considerably reduced at the two highest concentrations (169 and 225 mg kg1 d1). Malformations were also somewhat more numerous at the highest dosage, suggesting that IM14 Cl may be teratogenic. Maternal toxicity was also present, indicating that IM14 Cl appeared to be developmentally toxic at maternally toxic dosages. Also, IM14 Cl has been shown to cause thermal irritation when applied topically to rats, but produced only minimal contact sensitization when evaluated in the mouse local lymph node assay (Landry et al., 2005). Additionally, in this report, it is worth noting that the transdermal toxicity of IM14 Cl was influenced by the vehicle of administration. Use of the organic solvent, dimethylformamide, accentuated the acute toxicity. Very high concentrations of IM14 Cl (up to 95% IM14 Cl in water) applied to the rat skin were markedly less acutely toxic. This result may have a practical guideline that to reduce the acute toxicity, ILs can be handled in pure form with water as a co-solvent.

Inhibitory effects of ILs on vertebrates

Zebrafish (Danio rerio) plays an important role in ecotoxicology as a prominent model vertebrate. Concerning toxicity of ILs to the zebrafish, Pretti et al. (2006) revealed that ILs may cause a completely different effect on fish according to their chemical structures. As imidazolium, pyridinium and pyrrolidinium showed a LC50 (lethal effect) >100 mg L1, they could be regarded as non-highly lethal towards zebrafish. On the other hand, the ammonium salts showed LC50 remarkably lower than that reported for organic solvents and tertiary amines. In general, these data referred that fish are less sensitive to ILs toxicity compared to other species belonging to lower trophic levels. In a recent report, Li et al. (2009) used the frog Rana nigromaculata as an amphibian model for toxicity testing. Amphibians are often the main vertebrate group prone to contaminant exposure in aquatic systems mostly because their larvae live in water (Lahr, 1997; Mann and Bidwell, 2000). In their study, they evaluated the toxic effects of IM18 Br on the early embryonic development of the frog R. nigromaculata. The results demonstrated that the highest embryonic mortality occurred in the neural plate stage, followed by the early gastrula and early cleavage stages with the LC50 values

3.

Environmental fate of ILs

3.1.

Chemical degradation of ILs

Ionic liquids possess excellent chemical and thermal stability, which gives, unfortunately, a negative aspect for their treatment after usage prior to disposal. To assess the persistence of ILs in the environment as well as verify possibilities of their cleanup by chemical methods, several groups have focused their attention on oxidative and thermal degradation of ILs in aqueous media (Awad et al., 2004; Baranyai et al., 2004; Berthon et al., 2006; Itakura et al., 2008; Li et al., 2007; Morawski et al., 2005; Siedlecka and Stepnowski, 2009; Siedlecka et al., 2008a,b; Stepnowski and Zaleska, 2005). Pioneering work in the field of oxidative degradation was done by Stepnowski and Zaleska (2005) and Morawski et al. (2005) who showed that the greatest degradation efficiency for imidazolium ILs was achieved with a combination of UV light and a catalytic oxidant such as hydrogen peroxide or titanium dioxide. Subsequently, Li et al. (2007) studied the oxidative degradation of 1,3-dialkylimidazolium ILs in hydrogen peroxide/acetic acid medium assisted by ultrasonic chemical irradiation. It was


observed that 99% of tested compounds was degraded after 72 h. In addition, advanced oxidative degradation in the presence of reactive peroxides generated by Fenton reagent has been applied for the removal of ILs from water (Siedlecka and Stepnowski, 2009; Siedlecka et al., 2008a,b). According to the results, in a Fenton system with 1 mM of Fe(III) and 100 mM of H2O2, more than 97% of IM14 Cl was observed to degrade after 90 min. For Pyr14 Cl, IM16 Cl and IM18 Cl, the levels of degradation were 92%, 88% and 68%, respectively. Investigations of the degradation mechanisms indicated IM18 Cl was more resistant to oxidation by OH radicals cleaved from H2O2, suggesting that the oxidation rates of imidazolium ILs by OH are structure-dependent (Siedlecka and Stepnowski, 2009). The level of degradation was dependent on the alkyl chain length, consistent with Stepnowski and Zaleska (2005), who indicated that lengthening the alkyl chain lowered the rate of IL degradation. On contrast, the different length of the side chains and the type of anions did not affect the degradation process (Li et al., 2007). Regarding the thermal degradation studies of alkylimidazolium salts (Awad et al., 2004), extension of the alkyl chain enhanced the thermo-oxidative degradation of imidazolium salts. Interestingly, methyl substitution in the 2-position (i.e. between the two N atoms) was observed to decrease the oxidative decomposition of imidazolium ILs. The longer alkyl chain was also observed to induce an enhancement in photocatalytic decomposition of ILs (Morawski et al., 2005), which was not in the case of oxidative degradation (Stepnowski and Zaleska, 2005; Siedlecka and Stepnowski, 2009). Nonetheless, detailed account on the degradation of ILs by photocatalysis is required to verify this phenomenon.

3.2.

Biodegradability of ILs

In contrast to chemical degradation, which requires the assistance of a certain oxidant for catalysis, biodegradation is the microbial breakdown of chemical compounds. Biodegradation seems to be more environmentally friendly compared to chemical decomposition process. The initial attempt to examine the degradation potential of different IM14 cations combined with [Br], [BF4], [PF6], [N(CN)2], [(CF3SO2)2N] and octylsulfate as the counter ion was done using the Sturm and Closed-Bottle test protocols by the group of Scammells (Garcia et al., 2005; Gathergood and Scammells, 2002; Gathergood et al., 2004; Gathergood et al., 2006). Nonetheless, no compound showed significant degree of biodegradation with the exception of the octylsulfate-containing IL. The next step study on the biodegradation of ILs involved the design of ILs containing biodegradable side chains (Gathergood and Scammells, 2002). The design was done according to the principles of Boethling (Boethling, 1994, 1996; Howard et al., 1991) who identified three important parameters including the potential sites of enzymatic hydrolysis (for example, esters and amides) and oxygen in form of hydroxyl, aldehyde or carboxylic acid groups as well as unsubstituted linear alkyl chains (especially 4 carbons) and phenyl rings, which represent possible sites for attack by oxygenases. However, for a balance between chemical properties and biodegradability, not all of these factors were suitable for ILs. The

365

addition of oxygen containing functional groups such as alcohols, aldehyde and carboxylic acids was reported to restrict the ILs performance as reaction media whereas the incorporation of phenyl rings was known to increase the melting points of IL solvents (McGuinness and Cavell, 2000). Therefore, ester or amide group was selected to be coupled in alkyl side chain of ILs. The introduction of ester groups derived from a C2 acid and C4 or higher alcohol in the 3-N-substitutent was demonstrated to increase the biodegradation of imidazolium-based ILs (Gathergood et al., 2004). This can be explained by the fact that introduction of ester moiety probably provides a site susceptible to enzymatic attack (Gathergood et al., 2004; Gathergood et al., 2006) and hence, improves the biodegradation level. Though the addition of amide group is informed to improve the biodegradation of organic compounds (Boethling, 1994, 1996; Howard et al., 1991), no critical enhancement of biological degradation was noted when this group was appended into the imidazoliumbased ILs (Gathergood et al., 2004). However, no compound could be classified as ‘‘readily biodegradable’’ corresponding to Organization for Economic Cooperation and Development (OECD) standards (U.S. EPA, 1998), for which 60–70% or greater biodegradation by activated sludge microbial inoculate is required within a 10-day window in a 28-day period. Finally, the combination of the octylsulfate anion and imidazolium cation containing ester side chains resulted in readily biodegradable IL (Gathergood et al., 2006). Recently, Stolte et al. (2008) also paid their attention on investigation of functional groups incorporating alkyl chain ILs. Nonetheless, the introductions of terminal hydroxyl, carboxyl, ether and nitrile groups did not improve the biological degradation as expected. Kumar et al. (2006) investigated the fate of IM14 BF4 when in contact with soil-microorganisms, wastewater microorganisms, Pseudomonas putida and E. coli. Although IM14 BF4 was indicated to be recalcitrant in Sturm and Closed-Bottle test assays as mentioned above, it was observed in this study that P. putida was able to break down IM14 BF4 after 15 days of incubation. The breakdown products were monitored using GC-MS and identified to be 1-H-methylimidazole and 1-H-butylimidazole, which were in consistent with the theoretical metabolism scheme proposed by Jastorff et al. (2003). In case of bacteria from soil and wastewater, the metabolic intermediates appeared on the 12th day. It was also noted that different intermediate peaks were observed at different retention time with different microbes, indicating that the degradation mechanism of IM14 BF4 may vary in correspondence to certain microbes and metabolic pathways. In another study, the biodegradation pathway of IM18 moiety (Fig. 3) was proposed based on intermediate products via HPLC-MS analysis after 24-day period of incubation with activated sludge (Stolte et al., 2008). The metabolism of IM18 cation appeared to undergo oxidation reactions catalyzed probably by mono-oxygenases, e.g. the cytochrome P450 system on the terminal methyl group (u-oxidation). The alcohol formed was subsequently oxidized and converted into aldehydes, and then into carboxylic acids by dehydrogenases. The resulting carboxylic acids then might undergo b-oxidation and finally generated two carbon fragments that can enter the tricarboxylic acid cycle as acetyl Co-A (Fig. 3).

366


retention time in min.

m/z+

intensity

8.9

195

4*105

13.5 / 14.4

211

3*105 / 2*105

m/z+

N

N

+

N

+

N

OH

OH

12.2 / 12.7

209

1*106 / 2*106

N

+

N

N

+

N

211

O

10.0 – 12.5

225

2*104 – 6*104

N

H

+

N

O

O

19.5

183

1*105

OH

N

N

+

N

+

N

209

OH

16.2

197

3*10

5

N

+

N

OH

26.5

155

0.5*105

OH

N

N

+

N OH

24.2

169

4*106

O +

O

N

N

+

225

+

N

OH O

26.7

141

1*105

N

N

+

O OH

Fig. 3 – Biodegradation pathways of 1-octyl-3-methylimidazolium by activated sludge microbial community (Reproduced from Stolte et al. (2008) by permission of the Royal Society of Chemistry).

The proposed pathways provide basic information to both environmental scientists and chemical engineers; however, no studies have sought to examine the toxicity of metabolic products after degradation of ILs. This issue is of paramount importance since metabolism might not always end in less toxic products. Subsequently, Wells and Coombe (2006) extended the microbial degradation study with ammonium, imidazolium, phosphonium and pyridinium compounds by measuring the biological oxygen demand. The authors observed no biodegradability of cations incorporated short chains (C 4) within this test series, which was in agreement with Docherty et al. (2007) and Stolte et al. (2008). For longer alkyl chains (C12, C16 and C18) containing ILs, a strong inhibitory effect of these compounds on the inoculum used was found, indicating the active microbial consortium was significantly impacted by ILs toxicity. In recent studies (Docherty et al., 2007; GrabinskaSota and Kalka, 2004; Harjani et al., 2008; Stasiewicz et al., 2008), pyridinium-based ILs were reported to be fully catabolized by microbial community in activated sludge. This can be inferred from the fact that degradation pathways for pyridine –

the precursor of pyridinium-based compounds – under aerobic and anaerobic conditions were intensively investigated in the work of Kaiser et al. (1996). With respect to the common 1,3-dialkylpyridinium ILs, Pham et al. (2009) reported that after 21 days of incubation, microorganisms from activated sludge were able to break down Py4-3Me Br. Analyses of HPLC and MS/MS demonstrated that this biodegradation led to the formation of 1-hydroxybutyl-3-methylpyridinium, 1-(2-hydroxybutal)-3-methylpyridinium, 1-(2-hydroxyethyl)3-methylpyridinium and methylpyridine. Based on these intermediate products, biodegradation pathways were also suggested (Fig. 4), thereby providing the basic information which might be useful for assessing the factors related to the environmental fate and behavior of this commonly used pyridinium IL. Although this is the first report on biodegradation intermediates and pathway of pyridinium ILs, the authors have failed to systematically screen a single microorganism or a microbial consortium responsible for biodegradation of ILs. Therefore, it is needed to further investigate which type of microorganism is adaptable to ILs and which is responsible for degradation process. Also, the broken

367


II H3C

+

N

H3C

CH3

+

N

CH3 OH

I H3C

OH

+

N

H3C

H3C

H3 C

OH

+

N

+

N

+

H

+

OH

+

O

+

N

H3 C

OH

H3C

+

N

+

N

CH3 H3C

H

+

H3C OH

H3C O

H3C OH

Fig. 4 – Biodegradation pathways of 1-butyl-3-methylpyridinium entity by microorganisms from activated sludge (Reprinted with permission from Pham et al. (2009). Copyright 2009 American Chemical Society).

structures much further than methylpyridinium were not measured in the study. In particular, the possibility of cleavage of heterocyclic ring in ILs molecular structure (both pyridinium in this work and imidazolium in the study of Stolte et al. (2008)) has not been indentified. This issue should be clarified through further studies. Interestingly, Py4-3Me Br was not found to be metabolized by the activated sludge community (Docherty et al., 2007). This was attributed likely to the high IL concentration used in the study of Docherty’s group, which consequently inhibited the microbial consortium. Nonetheless, it was demonstrated that the structural manipulation of the pyridinium skeleton may lead to ILs with greater biodegradable extent compared to imidazolium-based compouds (Harjani et al., 2008). Concerning the anionic effect, ILs with halide counter ions were postulated to be more stable to degradation than perfluoronated ions (Awad et al., 2004; Gathergood and Scammells, 2002). In a preliminary study, Gathergood and Scammells (2002) confirmed this assumption and showed that the biodegradation efficiency decreased in the order [PF6] > [BF4] > [Br] with 60%, 59% and 48% of CO2 evolution values, respectively. In a later study (Gathergood et al., 2006), it was found that the octylsulfate anion is considerably more biodegradable than the other commonly used anions. The alkyl chain with C4, C6 or C8 was found to increase the rate of degradation (Docherty et al., 2007; Stolte et al., 2008); nonetheless, further increasing the chain length to C12, C16 or C18 was noted to cause toxic effect towards inoculum (Wells and Coombe, 2006). However, it was stated that the long octyl side chain was not a compulsory factor for biodegradation, but more important is a certain overall lipophilicity of the compound (Stolte et al., 2008).

3.3.

Sorption of ILs in environmental systems

Because the aquatic and terrestrial environments are possible recipients for contaminants, the distribution and behavior of ILs in soil are also extremely important. The retention and mobility of ILs in soils and sediments are strongly influenced by its tendency to be sorbed onto the various components of the soil matrix (Stepnowski, 2005). Since imidazolium-based ILs possess high electron acceptor potential of delocalized aromatic systems and hydrophobic components (e.g. the alkyl chain) (Stepnowski, 2005), they could be sorbed onto soils and sediments via several mechanisms. In a preliminary study, Stepnowski (2005) proved that electrostatic interactions contributed to the sorption of the imidazolium cations. Moreover, totally contrast trends were also observed demonstrating an extremely strong and practically irreversible sorption onto fine-textured marine sediments and a relatively weakly and reversibly binding to peaty soil (with the highest organic carbon content) (Stepnowski, 2005). This indicates the importance of the mineral component of the soil (sediment) in the sorption mechanism of ILs. Also, in this work, the author pointed out that compounds with longer alkyl chains were irreversibly bound to the soil component, which was in agreement with Stepnowski et al. (2007) and Matzke et al. (2009b). Interestingly, Beaulieu et al. (2008) did not find a positive effect of alkyl chain length on the sorption of alkylimidazolium-based ILs to aquatic sediments and suggested that hydrophobic interactions were not the most important sorption mechanism. The contrast results between these groups imply that sorption mechanisms of ILs may vary according to properties and composition of the environmental systems. The studies suggest that ILs may be retained by aquatic sediments;

368


nonetheless, the toxic action of these sorbed materials towards aquatic organisms has not been addressed. Also, further efforts should be continued to elucidate the reversibility of IL sorption on these sediments to give a better understanding of the ILs fate after being released into aqueous environment. Matzke et al. (2009b) investigated the influences of the two different clay minerals kaolinite and smectite as well as of organic matter on the cation sorption and desorption behaviors of three imidazolium based ILs including IM14 BF4, IM18 BF4 and IM14 (CF3SO2)2N in soil. The addition of organic matter and clays was observed to increase the sorption/decrease the desorption of all ILs tested, and in particular smectite had striking effects on the sorption efficiency of all substances. It is worth noting that not only the cationic moiety with different alkyl side chain lengths but the anionic compartments can also play an importance role in the sorption/ desorption processes. Imidazolium compounds with BF 4 as a counter anion showed higher sorption capacity compared to that of [(CF3SO2)2N], indicating the high potential of this type of IL to form ionic pairs in the soil matrix (Matzke et al., 2009b). However, further work with a variety of ILs incorporated different cationic and anionic moieties should be carried out to verify these phenomena. Gorman-Lewis and Fein (2004) examined the sorption behavior of IM14 Cl onto a range of surfaces which are commonly found in the near-surface environment. The results suggested that IM14 Cl could be minimally retarded by noninterlayer clay system and might lead to unimpeded transport through subsurface groundwater. Also, the adsorption capacity of this IL onto bacterial surfaces was not high, which might be due to the low hydrophobicity of IM14 Cl. Additionally, investigations of the adsorption of IM14 Cl towards different media carried out by our group (Vijayaraghavan et al., 2009) have shown that retardation of this compound was possible only by an ion-exchange resin and activated carbon, which was in consistency with the work of Anthony et al. (2001). However, no significant adsorption of IM14 Cl in the media of a fermentation waste (Corynebacterium glutamicum) and dried activated sludge was observed in our study. Conclusively, the data currently available demonstrated that ILs incorporated imidazolium cation can be sorbed to organic matter, whether found in aquatic sediments or terrestrial soils, and the presence of clays significantly enhanced the sorption capacity.

4.

Concluding remarks and future directions

Ionic liquids, of which the most often cited attribute is their negligible vapor pressure, have been suggested as a green alternative to traditional organic solvents with the desire to minimize diffusion to the atmosphere. Low volatility, however, does not completely eliminate potential environmental hazards and might pose serious threats to aquatic and terrestrial ecosystems. The studies of environmental fate and toxicity of ILs have shown that the ILs commonly used to date are toxic in nature and their toxicities vary considerably across organisms and trophic levels. In general, the effect of anionic moieties is not drastic as the alkyl length effect except for the case of [(CF3SO2)2N], which shows a clear (eco)toxicological

hazard potential. The other perfluorinated anions have been also proved to be hazardous due to hydrolytically unstable properties. In addition, the introduction of functional polar groups to the alkyl chain has been shown to reduce the toxicity of ILs and increase the biodegradation efficiency to some extent. This indicates the possibility of tailoring ILs by coupling suitable functional groups to their structure, which in turn leads to a more environmental friendly compound. The side chain length effect has been found to be consistent in all levels of biological complexity as well as different environmental compartments. Also, an increase in alkyl-chain length, or lipophilicity, was observed to be related to an increase in the rate of degradation as well as an increase in toxicity. This indicates a conflict of aims between minimizing the toxicity and maximizing the biodegradability of these neoteric solvents. Regarding the cationic compartment, pyridinium has been found to be more environmental friendly than imidazolium from both viewpoints of toxicology and microbial degradation. It can therefore be suggested that the structural manipulation of the pyridinium skeleton should be considered in design of a sustainable IL. From the currently available data, it is clear that some commonly used ILs are very far away from the image of green chemicals that are often cited in the literature. The uncertainties in their sustainable development hinder the applications of ILs under real conditions. Although some attempts have been made to give important hints in the prospective design and synthesis of inherently safer ILs, comprehensive studies dealing with the behaviors of ILs in aqueous media still await to be conducted. The important features required for the thorough insight into environmental fate of ILs include, but are not limited to: Providing more fundamental understanding into the mechanism for IL-induced toxicity to different levels of biological complexity. The underlying mechanisms of IL toxicity have rarely been studied. - Assessing the biodegradability of cationic and anionic compartments and toxicity of their degradation intermediates. This may provide useful information in consciously designing safer chemicals. - Investigating the aerobic and anaerobic biodegradation of ILs, which would suggest initial guidelines for the treatment of ILs waste by using the existing aerobic and anaerobic wastewater treatment facilities. Especially, anaerobic degradation awaits to be investigated. - Defining which organisms or enzymes may promote degradation pathways and determining specific microbial consortium or cultivatable communities capable of biotransformation of ILs. - Performing the ecotoxicity and biodegradation tests in real environmental conditions instead of controlled conditions of laboratory experiments, which would be advantageous in understanding the fate and behavior of ILs under real conditions. For this, the potential toxicological effects at population level and community level should be addressed. It must be encouraged to use tools such as experimental mesocosms to study the effects of ILs at higher levels of organization. - Creating database of environmentally benign structure moieties of ILs based upon their toxicological and -


biodegradation information, which would be practically useful as a reference for manufacturers and regulators to properly develop and regulate the use of ILs.

Acknowledgements This work was supported by NRF Grant funded by the Korean Government (KRF-2007-521-D00106, NRL 2009-0083194, and in part WCU R31-2008-000-20029-0).

references

Anthony, J.L., Maginn, E.J., Brennecke, J.F., 2001. Solution thermodynamics of imidazolium-based ionic liquids and water. J. Phys. Chem. B 105, 10942–10949. Arning, J., Stolte, S., Bo¨schen, A., Stock, F., Pitner, W.-R., WelzBiermann, U., Jastorff, B., Ranke, J., 2008. Qualitative and quantitative structure activity relationships for the inhibitory effects of cationic head groups, functionalized side chains and anions of ionic liquids on acetylcholinesterase. Green Chem. 10, 47–58. Austin, R.P., Barton, P., Davis, A.M., Manners, C.N., Stansfield, M.C., 1998. The effect of ionic strength on liposome-buffer and 1-octanol-buffer distribution coefficients. J. Pharm. Sci. 87, 599–607. Awad, W.H., Gilman, J.W., Nyden, M., Harris, R.H., Sutto, T.E., Callahan, J., Trulove, P.C., DeLong, H.C., Fox, D.M., 2004. Thermal degradation studies of alkyl-imidazolium salts and their application in nanocomposites. Thermochim. Acta 409, 3–11. Babalola, G.O., 1998. Anti-bacterial activity of synthetic N-heterocyclic oxidizing compounds. Lett. Appl. Microbiol. 26, 43–46. Bailey, M.M., Townsend, M.B., Jernigan, P.L., Sturdivant, J., HoughTroutman, W.L., Rasco, J.F., Swatloski, R.P., Rogers, R.D., Hood, R.D., 2008. Developmental toxicity assessment of the ionic liquid 1-butyl-3-methylimidazolium chloride in CD-1 mice. Green Chem. 10, 1213–1217. Ba1czewski, P., Bachowska, B., Bia1as, T., Biczak, R., Wieczorek, W. M., Balin´ska, A., 2007. Synthesis and phytotoxicity of new ionic liquids incorporating chiral cations and/or chiral anions. J. Agric. Food Chem. 55, 1881–1892. Baranyai, K.J., Deacon, G.B., MacFarlane, D.R., Pringle, J.M., Scott, J.L. , 2004. Thermal degradation of ionic liquids at elevated temperatures. Aust. J. Chem. 57, 145–147. Beaulieu, J.J., Tank, J.L., Kopacz, M., 2008. Sorption of imidazoliumbased ionic liquids to aquatic sediments. Chemosphere 70, 1320–1328. Bernot, R.J., Turner, A.M., 2001. Predator identity and traitmediated indirect effects in a littoral food web. Oecologia 129, 139–146. Bernot, R.J., Brueseke, M.A., Evans-White, M.A., Lamberti, G.A., 2005a. Acute and chronic toxicity of imidazolium-based ionic liquids on Daphnia magna. Environ. Toxicol. Chem. 24, 87–92. Bernot, R.J., Kennedy, E.E., Lamberti, G.A., 2005b. Effects of ionic liquids on the survival, movement, and feeding behavior of the freshwater snail, Physa acuta. Environ. Toxicol. Chem. 24, 1759–1765. Berthon, L., Nikitenko, S.I., Bisel, I., Berthon, C., Faucon, M., Saucerotte, B., Zorz, N., Moisy, Ph, 2006. Influence of gamma

369

irradiation on hydrophobic room-temperature ionic liquids [BuMeIm] PF6 and [BuMeIm] (CF3SO2)2N. Dalton Trans, 2526–2534. Blaauboer, B.J., Balls, M., Barratt, M., Casati, S., Coecke, S., Mohamed, M.K., Moore, J., Rall, D., Smith, K.R., Tennant, R., Schwetz, B.A., Stokes, W.S., Younes, M., 1998. Alternative testing methodologies and conceptual issues. Environ. Health Perspect 106, 413–418. Blaise, C.R., 1993. In: Richardson, M.L. (Ed.), Ecotoxicology Monitoring. VCH, Weinheim, pp. 83–108. Blanchard, L.A., Hancu, D., Beckman, E.J., Brennecke, J.F., 1999. Green processing using ionic liquids and CO2. Nature 399, 28–29. Blanchard, L.A., Brennecke, J.F., 2001. Recovery of organic products from ionic liquids using supercritical carbon dioxide. Ind. Eng. Chem. Res. 40, 287–292. Boethling, R.S., 1994. Cationic Surfactants, Surfactant Science Series, vol. 53. Marcel Dekker, New York, pp. 95–135. Boethling, R.S., 1996. Designing safer chemicals. ACS Symp. Ser. 640, 156. Chemnitius, J.-M., Sadowski, R., Winkel, H., Zech, R., 1999. Organophosphate inhibition of human heart muscle cholinesterase isoenzymes. Chem. Biol. Interact 120, 183–192. Cheng, Y., Wright, S.H., Hooth, M.J., Sipes, I.G., 2009. Characterization of the disposition and toxicokinetics of N-butylpyridinium chloride in male F-344 rats and female B6C3F1 mice and its transport by organic cation transporter 2. Drug Metab. Dispos. 37, 909–916. Cho, C.-W., Pham, T.P.T., Jeon, Y.-C., Vijayaraghavan, K., Choe, W.-S., Yun, Y.-S., 2007. Toxicity of imidazolium salt with anion bromide to a phytoplankton Selenastrum capricornutum: effect of alkylchain length. Chemosphere 69, 1003–1007. Cho, C.-W., Pham, T.P.T., Jeon, Y.-C., Yun, Y.-S., 2008a. Influence of anions on the toxic effects of ionic liquids to a phytoplankton Selenastrum capricornutum. Green Chem. 10, 67–72. Cho, C.-W., Jeon, Y.-C., Pham, T.P.T., Vijayaraghavan, K., Yun, Y.-S., 2008b. The ecotoxicity of ionic liquids and traditional organic solvents on microalga Selenastrum capricornutum. Ecotoxicol. Environ. Saf. 71, 166–171. Cho, C.-W., Pham, T.P.T., Jeon, Y.-C., Min, J., Jung, H.Y., Lee, D.S., Yun, Y.-S., 2008c. Microalgal photosynthetic activity measurement system for rapid toxicity assessment. Ecotoxicology 17, 455–463. Cieniecka-Ros1onkiewicz, A., Pernak, J., Kubis-Feder, J., Ramani, A., Robertson, A.J., Seddon, K.R., 2005. Synthesis, anti-microbial activities and anti-electrostatic properties of phosphoniumbased ionic liquids. Green Chem. 7, 855–862. Costello, D.M., Brown, L.M., Lamberti, G.A., 2009. Acute toxic effects of ionic liquids on zebra mussel (Dreissena polymorpha) surviving and feeding. Green Chem. 11, 548–553. Couling, D.J., Bernot, R.J., Docherty, K.M., Dixon, J.K., Maginn, E.J., 2006. Assessing the factors responsible for ionic liquid toxicity to aquatic organisms via quantitative structure-property relationship modeling. Green Chem. 8, 82–90. Dipeolu, O., Green, E., Stephens, G., 2008. Effects of watermiscible ionic liquids on cell growth and nitro reduction using Clostridium sporogenes. Green Chem. 11, 397–401. Docherty, K.M., Kulpa, C.F., 2005. Toxicity and antimicrobial activity of imidazolium and pyridinium ionic liquids. Green Chem. 7, 185–189. Docherty, K.M., Dixon, J.K., Kulpa, C.F., 2007. Biodegradability of imidazolium and pyridinium ionic liquids by an activated sludge microbial community. Biodegradation 18, 481–493. Earle, M.J., Seddon, K.R., 2000. Ionic liquids. Green solvents for the future. Pure Appl. Chem. 72, 1391–1398. Frade, R.F.M., Matias, A., Branco, L.C., Afonso, C.A.M., Duarte, C. M.M., 2007. Effect of ionic liquids on human colon carcinoma HT-29 and CaCo-2 cell lines. Green Chem. 9, 873–877. Fuler, J., Carlin, R.T., Osteryoung, R.A., 1997. The room temperature ionic liquids 1-ethyl-3-methylimidazolium tetrafluoroborate:

370


electrochemical couples and physical properties. J. Electrochem. Soc. 144, 3881–3886. Fulton, M.H., Key, P.B., 2001. Acetylcholinesterase inhibition in estuarine fish and invertebrates as an indicator of organophosphorus insecticide exposure and effects. Environ. Toxicol. Chem. 20, 37–45. Ganske, F., Bornscheuer, U.T., 2006. Growth of Escherichia coli, Pichia pastoris and Bacillus cereus in the presence of the ionic liquids [BMIM] [BF4] and [BMIM] [PF6] and organic solvents. Biotechnol. Lett. 28, 465–469. Garcia, M.T., Gathergood, N., Scammells, P.J., 2005. Biodegradable ionic liquids. part II: effect of the anion and toxicology. Green Chem. 7, 9–14. Garcıá-Lorenzo, A., Tojo, E., Tojo, J., Teijeira, M., Rodrı´guezBerrocal, F.J., Gonza´lez, M.P., Martıńez-Zorzano, V.S., 2008. Cytotoxicity of selected imidazolium-derived ionic liquids in the human Caco-2 cell line. Sub-structural toxicological interpretation through a QSAR study. Green Chem. 10, 508–516. Gathergood, N., Scammells, P.J., 2002. Design and preparation of room-temperature ionic liquids containing biodegradable side chains. Aust. J. Chem. 55, 557–560. Gathergood, N., Garcia, M.T., Scammells, P.J., 2004. Biodegradable ionic liquids: part I. Concept, preliminary targets and evaluation. Green Chem. 6, 166–175. Gathergood, N., Scammells, P.J., Garcia, M.T., 2006. Biodegradable ionic liquids. Part III: the first readily biodegradable ionic liquids. Green Chem. 8, 156–160. Gordon, C.M., 2001. New developments in catalysis using ionic liquids. Appl. Catal. A 222, 101–117. Gorman-Lewis, D.J., Fein, J.B., 2004. Experimental study of the adsorption of an ionic liquid onto bacterial and mineral surfaces. Environ. Sci. Technol 38, 2491–2495. Grabinska-Sota, E., Kalka, J., 2004. An assessment of the toxicity of pyridinium chlorides and their biodegradation intermediates. Environ. Int 28, 687–690. Grabinska-Sota, E., Kalka, J., 2006. Toxicity of imidazolium chlorides to aquatic organisms. Polish J. Environ. Stud. 15, 405–409. Hagiwara, R., Ito, Y., 2000. Room temperature ionic liquids of alkylimidazolium cations and fluoroanions. J. Fluorine Chem. 105, 221–227. Harjani, J.R., Singer, R.D., Garcia, M.T., Scammells, P.J., 2008. The design and synthesis of biodegradable pyridinium ionic liquids. Green Chem. 10, 436–438. Hassoun, E.A., Abraham, M., Kini, V., Al-Ghafri, M., Abushaban, A., 2002. Cytotoxicity of the ionic liquid, 1- N-butyl-3methylimidazolium chloride. Res. Commun, Pharmacol. Toxicol 7, 23–31. Howard, P.H., Boethling, R.S., Stiteler, W., Meylan, W., Beauman, J., 1991. Development of a predictive model for biodegradability based on BIODEG, the evaluated biodegradation data base. Sci. Total Environ. 110, 635–641. Huddleston, J.G., Visser, A.E., Reichert, W.M., Willauer, H.D., Broker, G.A., Rogers, R.D., 2001. Characterization and comparison of hydrophilic and hydrophobic room temperature ionic liquids incorporating the imidazolium cation. Green Chem. 3, 156–164. Itakura, T., Hirata, K., Aoki, M., Sasai, R., Yoshida, H., Itoh, H., 2008. Decompostition and removal of ionic liquid in aqueous solution by hydrothermal and photocatalytic treatment. Environ. Chem. Lett. doi:10.1007/s10311-008-0177-7. Jastorff, B., Sto¨rmann, R., Ranke, J., Mo¨lter, K., Stock, F., Oberheitmann, B., Hoffmann, W., Hoffmann, J., Nu¨chter, M., Ondruschka, B., Filser, J., 2003. How hazardous are ionic liquids? Structure-activity relationships and biological testing as important elements for sustainability evaluation. Green Chem. 5, 136–142. Jastorff, B., Mo¨lter, K., Behrend, P., Bottin-Weber, U., Filser, J., Heimers, A., Ondruschka, B., Ranke, J., Schaefer, M.,

Schro¨der, H., Stark, A., Stepnowski, P., Stock, F., Sto¨rmann, R., Stolte, S., Welz-Biermann, U., Ziegert, S., Tho¨ming, J., 2005. Progress in evaluation of risk potential of ionic liquids-basis for an eco-design of sustainable products. Green Chem. 7, 362–372. Kaiser, K.L.E., Palabrica, V.S., 1991. Photobacterium phosphoreum, toxicity data index. Water Poll. Res. J. Can. 26, 361–431. Kaiser, J.P., Feng, Y.C., Bollag, J.M., 1996. Microbial metabolism of pyridine, quinoline, acridine, and their derivatives under aerobic and anaerobic conditions. Microbiol. Rev. 60, 483–498. Kazarian, S.G., Briscoe, B.J., Welton, T., 2000. Combining ionic liquids and supercritical fluids: in situ ATR-IR study of CO2 dissolved in two ionic liquids at high pressures. Chem. Commun., 2047–2048. Kelman, D., Kashman, Y., Rosenberg, E., Ilan, M., Ifrach, I., Loya, Y., 2001. Antimicrobial activity of the reef sponge Amphimedon viridis from the Red Sea: evidence for selective toxicity. Aquat. Microb. Ecol 24, 9–16. Kulacki, K.J., Lamberti, G.A., 2008. Toxicity of imidazolium ionic liquids to freshwater algae. Green Chem. 10, 104–110. Kumar, R.A., Papaı¨conomou, N., Lee, J.-M., Salminen, J., Clark, D.S., Prausnitz, J.M., 2009. In vitro cytotoxicities of ionic liquids: effect of cation rings, functional groups, and anions. Environ. Toxicol. 24, 388–395. Kumar, S., Ruth, W., Sprenger, B., Kragl, U., 2006. On the biodegradation of ionic liquid 1-butyl-3-methylimidazolium tetrafluoroborate. Chim. Oggi 24, 24–26. Lahr, J., 1997. Ecotoxicology of organisms adapted to life in temporary freshwater ponds in arid and semi-arid regions. Arch. Environ. Contam. Toxicol 32, 50–57. Landry, T.D., Brooks, K., Poche, D., Woolhiser, M., 2005. Acute toxicity profile of 1-butyl-3-methylimidazolium chloride. Bull. Environ. Contam. Toxicol 74, 559–565. Larson, J.H., Frost, P.C., Lamberti, G.A., 2008. Variable toxicity of ionic liquid-forming chemicals to Lemna minor and the influence of dissolved organic matter. Environ. Toxicol. Chem. 27, 676–681. Lata1a, A., Stepnowski, P., Nedzi, ˛ M., Mrozik, W., 2005. Marine toxicity assessment of imidazolium ionic liquids: acute effects on the Baltic algae Oocystis submarina and Cyclotella meneghiniana. Aqua. Toxicol 73, 91–98. Lee, S.-M., Chang, W.-J., Choi, A.-R., Koo, Y.-M., 2005. Influence of ionic liquids on the growth of Escherichia coli. Korean. J. Chem. Eng 22, 687–690. Lewis, M.A., 1995. In: Rand, G.M. (Ed.), Fundamentals of Aquatic Toxicology: Effects, Environment Fate, and Risk Assessment, second ed. Taylor and Francis, Washington, DC, USA, pp. 135–170. Li, G., Shen, J., Zhu, Y., 1998. Study of pyridinium-type functional polymers. II. Antibacterial activity of soluble pyridinium-type polymers. J. Appl. Polym. Sci. 67, 1761–1768. Li, X., Zhao, J., Li, Q., Wang, L., Tsang, S.C., 2007. Ultrasonic chemical oxidative degradations of 1,3-dialkylimidazolium ionic liquids and their mechanistic elucidations. Dalton Trans, 1875–1880. Li, X.-Y., Zhou, J., Yu, M., Wang, J.-J., Pei, Y.C., 2009. Toxic effects of 1-methyl-3-octylimidazolium bromide on the early embryonic development of the frog Rana nigromaculata. Ecotox. Environ. Saf. 72, 552–556. Luis, P., Ortiz, I., Alcado, R., Irabien, A., 2007. A novel group contribution method in the development of a QSAR for predicting the toxicity (Vibrio fischeri EC50) of ionic liquids. Ecotoxicol. Environ. Saf 67, 423–429. Luo, Y.-R., Li, X.-Y., Chen, X.-X., Zhang, B.-J., Sun, Z.-J., Wang, J.-J., 2008. The developmental toxicity of 1-butyl-3-octylimidazolium bromide on Daphnia magna. Environ. Toxicol 23, 736–744. Malich, G., Markovic, B., Winder, C., 1997. The sensitivity and specificity of the MTS tetrazolium assay for detecting the in


vitro cytotoxicity of 20 chemicals using human cell lines. Toxicology 124, 179–192. Mann, R.M., Bidwell, J.R., 2000. Application of the FETAX protocol to assess the development toxicity of nonylphenol ethoxylate to Xenopus laevis and two Australian frogs. Aquat. Toxicol 51, 19–29. Marsh, K.N., Boxall, J.A., Lichtenthaler, R., 2004. Room temperature ionic liquids and their mixtures–a review. Fluid Phase Equilibr 219, 93–98. Massoulie´, J., Pezzementi, L., Bon, S., Krejci, E., Vallette, F.-M., 1993. Molecular and cellular biology of cholinesterases. Prog. Neurobiol. 41, 31–91. Matsumoto, M., Mochiduki, K., Fukunishi, K., Kondo, K., 2004a. Extraction of organic acids using imidazolium-based ionic liquids and their toxicity to Lactobacillus rhamnosus. Sep.Purif. Technol. 40, 97–101. Matsumoto, M., Mochiduki, K., Kondo, K., 2004b. Toxicity of ionic liquids and organic solvents to lactic acid-producing bacteria. J. Biosci. Bioeng 98, 344–347. Matzke, M., Stolte, S., Thiele, K., Juffernholz, T., Arning, J., Ranke, J., Welz-Biermann, U., Jastorff, B., 2007. The influence of anion species on the toxicity of 1-alkyl-3-methylimidazolium ionic liquids observed in an (eco)toxicological test battery. Green Chem. 9, 1198–1207. Matzke, M., Stolte, S., Bo¨schen, A., Filser, J., 2008. Mixture effects and predictability of combination effects of imidazolium based ionic liquids as well as imidazolium based ionic liquids and cadmium on terrestrial plants (Triticum aestivum) and limnic green algae (Scenedesmus vacuolatus). Green Chem. 10, 784–792. Matzke, M., Stolte, S., Arning, J., Uebers, U., Filser, J., 2009a. Ionic liquids in soils: effects of different anion species of imidazolium based ionic liquids on wheat (Triticum aestivum) as affected by different clay minerals and clay concentrations. Ecotoxicology 18, 197–203. Matzke, M., Thiele, K., Mu¨ller, A., Filser, J., 2009b. Sorption and desorption of imidazolium based ILs in different soil types. Chemosphere 74, 568–574. McFarlane, J., Ridenour, W.B., Luo, H., Hunt, R.D., Depaoli, D.W., Ren, R.X., 2005. Room temperature ionic liquids for separating organics from produced water. Sep. Sci. Technol 40, 1245–1265. McGuinness, D.S., Cavell, K.J., 2000. Donor-functionalized heterocyclic carbine complexes of palladium(II): efficient catalysts for C-C coupling reactions. Organometallics 19, 741–748. McQueen, D.J., Post, J.R., Mills, E.L., 1986. Trophic relationships in freshwater pelagic ecosystems. Can. J. Fish. Aqua. Sci. 43, 1571–1581. Morawski, A.W., Janus, M., Goc-Maciejewska, I., Syguda, A., Pernak, J., 2005. Decomposition of ionic liquids by photocatalysis. Pol. J. Chem. 79, 1929–1935. Myasoedov, B.F., Molochnikova, N.P., Shkinev, V.M., Spivakov, B.Y., 1995. The Behavior of Actinides in Two-Phase Aqueous Systems Based on Polyethylene Glycol. Plenum, New York, pp. 91. Nockemann, P., Thijs, B., Driesen, K., Janssen, C.R., Hecke, K.V., Meervelt, L.V., Kossmann, S., Kirchner, B., Binnemans, K., 2007. Choline saccharinate and choline acesulfamate: ionic liquids with low toxicities. J. Phys. Chem. B 111, 5254–5263. Olivier, H., 1999. Recent developments in the use of non-aqueous ionic liquids for two-phase catalysis. J. Mol. Cat. A: Chem. 146, 285–289. Osenberg, C.W., Mittelbach, G.G., 1989. Effects of body size on the predator–prey interaction between pumpkinseed sunfish and gastropods. Ecol. Monogr. 59, 405–432. Pernak, J., Czepukowicz, A., 2001. New ionic liquids and their antielectrostatic properties. Ind. Eng. Chem. Res. 40, 2379– 2383.

371

_ J., Mirska, I., 2001a. Synthesis and Pernak, J., Rogoza, antimicrobial activities of new pyridinium and benzimidazolium chlorides. Eur. J. Med. Chem. 36, 313–320. Pernak, J., Kalewska, J., Ksycin´ska, H., Cybulski, J., 2001b. Synthesis and anti-microbial activities of some pyridinium salts with alkoxymethyl hydrophobic group. Eur. J. Med. Chem. 36, 899–907. Pernak, J., Chwa1a, P., 2003. Synthesis and anti-microbial activities of choline-like quaternary ammonium chlorides. Eur. J. Med. Chem. 38, 1035–1042. Pernak, J., Sobaszkiewicz, K., Mirska, I., 2003. Anti-microbial activities of ionic liquids. Green Chem. 5, 52–56. Pernak, J., Goc, I., Mirska, I., 2004a. Anti-microbial activities of protic ionic liquids with lactate anion. Green Chem. 6, 323–329. Pernak, J., Sobaszkiewicz, K., Foksowicz-Flaczyk, J., 2004b. Ionic liquids with symmetrical dialkoxymethyl-substituted imidazolium cations. Chem. Eur. J 10, 3479–3485. Pham, T.P.T., Cho, C.-W., Vijayaraghavan, K., Min, J., Yun, Y.-S., 2008a. Effect of imidazolium-based ionic liquids on the photosynthetic activity and growth rate of Selenastrum capricornutum. Environ. Toxicol. Chem. 27, 1583–1589. Pham, T.P.T., Cho, C.-W., Min, J., Yun, Y.-S., 2008b. Alkyl-chain length effects of imidazolium and pyridinium ionic liquids on photosynthetic response of Pseudokirchneriella subcapitata. J. Biosci. Bioeng 105, 425–428. Pham, T.P.T., Cho, C.-W., Jeon, C.-O., Chung, Y.-J., Lee, M.-W., Yun, Y.-S., 2009. Identification of metabolites involved in the biodegradation of the ionic liquid 1-butyl-3-methylpyridinium bromide by activated sludge microorganisms. Environ. Sci. Technol 43, 516–521. Peraccini, D., Chiappe, C., Intorre, L., Pretti, C., 2007. In: Letcher, T. M. (Ed.), Thermodynamics, Solubility and Environmental Issues. Elsevier, UK, pp. 259–278. Pope, C., Karanth, S., Liu, J., 2005. Pharmacology and toxicology of cholinesterase inhibitors: uses and misuses of a common mechanism of action. Environ. Toxicol. Pharmacol 19, 433–446. Pretti, C., Chiappe, C., Pieraccini, D., Gregori, M., Abramo, F., Monni, G., Intorre, L., 2006. Acute toxicity of ionic liquids to the zebrafish (Danio rerio). Green Chem. 8, 238–240. Pretti, C., Chiappe, C., Baldetti, I., Brunini, S., Monni, G., Intorre, L., 2009. Acute toxicity of ionic liquids for three freshwater organisms: Pseudokirchneriella subcapitata, Daphnia magna and Danio rerio. Ecotoxicol. Environ. Saf. 72, 1170–1176. Ranke, J., Mo¨lter, K., Stock, F., Bottin-Weber, U., Poczobutt, J., Hoffmann, J., Ondruschka, B., Filser, J., Jastorff, B., 2004. Biological effects of imidazolium ionic liquids with varying chain length in acute Vibrio fischeri and WST-1 cell viability assays. Ecotox. Environ. Saf 58, 396–404. Ranke, J., Mu¨ller, A., Bottin-Weber, U., Stock, F., Stolte, S., Arning, J., Sto¨rmann, R., Jastorff, B., 2007a. Lipophilicity parameters for ionic liquid cations and their correlation to in vitro cytotoxicity. Ecotox. Environ. Saf 67, 430–438. Ranke, J., Stolte, S., Sto¨rmann, R., Arning, J., Jastorff, B., 2007b. Design of sustainable chemical products – the example of ionic liquids. Chem. Rev. 107, 2183–2206. Rogers, R.D., Bond, A.H., Bauer, C.B., Zhang, J., Rein, S.D., Chomko, R.R., Roden, D.M., 1995. Partitioning behavior of 99 Tc and, 129 I from simulated Handford tank wastes using polyethylene-glycol based aqueous biphasic systems. Solvent Extr. Ion Exch. 13, 689–713. Romero, A., Santos, A., Tojo, J., Rodrı´guez, A., 2008. Toxicity and biodegradability of imidazolium ionic liquids. J. Hazard. Mater 151, 268–273. Rooney, D.W., Seddon, K.R., 2001. In: Wypych, G. (Ed.), Handbook of Solvents. ChemTech Publishing, Toronto, pp. 1459–1484. Salminen, J., Papaiconomou, N., Kumar, R.A., Lee, J.-M., Kerr, J., Newman, J., Prausnitz, J.M., 2007. Physicochemical properties and toxicities of hydrophobic piperidinium and pyrrolidinium ionic liquids. Fluid Phase Equilibr 261, 421–426.

372


Samorı`, C., Pasteris, A., Galletti, P., Tagliavini, E., 2007. Acute toxicity of oxygenated and nonoxygenated imidazoliumbased ionic liquids to Daphnia magna and Vibrio fischeri. Environ. Toxicol. Chem. 26, 2379–2382. Schmid, A., Kollmer, A., Mathys, R.G., Withot, B., 1998. Development toward large-scale bacterial bioprocesses in the presence of bulk amounts of organic solvents. Extremophiles 2, 249–256. Sheldon, R., 2001. Catalytic reactions in ionic liquids. Chem. Commun, 2399–2407. Sheldon, R.A., 2005. Green solvents for sustainable organic synthesis: state of the art. Green Chem. 7, 267–278. Siedlecka, E.M., Stepnowski, P., 2009. The effect of alkyl chain length on the degradation of alkylimidazolium- and pyridinium-type ionic liquids in a Fenton-like system. Environ. Sci. Pollut. Res. 16, 453–458. Siedlecka, E.M., Mrozik, W., Kaczyn´ski, Z., Stepnowski, P., 2008a. Degradation of 1-butyl-3-methylimidazolium chloride ionic liquid in a Fenton-like system. J. Hazard. Mater 154, 893–900. Siedlecka, E.M., Go1ebiowski, ˛ M., Kumirska, J., Stepnowski, P., 2008b. Identification of 1-butyl-3-methylimidazolium chloride degradation products formed in Fe(III)/H2O2 oxidation system. Chem. Anal. (Warsaw) 53, 943–951. Sipes, I.G., Knudsen, G.A., Kuester, R.K., 2008. The effects of dose and route on the toxicokinetics and disposition of 1-butyl3-methylimidazolium chloride in male F-344 rats and female B6C3F1 mice. Drug Metab. Dispos. 36, 284–293. Sk1adanowski, A.C., Stepnowski, P., Kleszczyn´ski, K., Dmochowska, B., 2005. AMP deaminase in vitro inhibition by xenobiotics. A potential molecular method for risk assessment of synthetic nitro- and polycyclic musks, imidazolium ionic liquids and N-glucopyranosyl ammonium salts. Environ. Toxicol. Phar. 19, 291–296. Stasiewicz, M., Mulkiewicz, E., Tomczak-Wandzel, R., Kurmirska, J., Siedlecka, E.M., Go1ebiowski, M., Gajdus, J., Czerwicka, M., Stepnowski, P., 2008. Assessing toxicity and biodegradation of novel, environmentally benign ionic liquids (1-alkoxymethyl-3-hydroxypyridinium chloride, saccharine and acesulfamates) on cellular and molecular level. Ecotox. Environ. Saf 71, 157–165. Steinberg, S.M., Poziomek, E.J., Engelmann, W.H., Rogers, K.R., 1995. A review of environmental applications of bioluminescence measurements. Chemosphere 30, 2155–2197. Stepnowski, P., Sk1adanowski, A.C., Ludwiczak, A., qaczyn´ska, E., 2004. Evaluating the cytotoxicity of ionic liquids using human cell line HeLa. Hum. Exp. Toxicol 23, 513–517. Stepnowski, P., 2005. Preliminary assessment of the sorption of some alkyl imidazolium cations as used in ionic liquids to soils and sediments. Aust. J. Chem. 58, 170–173. Stepnowski, P., Zaleska, A., 2005. Comparison of different advanced oxidation processes for the degradation of room temperature ionic liquids. J. Photoch. Photobio. A 170, 45–50. Stepnowski, P., Mrozik, W., Nichthauser, J., 2007. Adsorption of alkylimidazolium and alkylpyridinium ionic liquids onto natural soils. Environ. Sci. Technol 41, 511–516. Stock, F., Hoffmann, J., Ranke, J., Sto¨rmann, R., Ondruschka, B., Jastorff, B., 2004. Effects of ionic liquids on the acetycholinesterase–a structure-activity relationship consideration. Green Chem. 6, 286–290. Stolte, S., Arning, J., Bottin-Weber, U., Matzke, M., Stock, F., Thiele, K., Uerdingen, M., Welz-Biermann, U., Jastorff, B., Ranke, J., 2006. Anion effects on the cytotoxicity of ionic liquids. Green Chem. 8, 621–629. Stolte, S., Matzke, M., Arning, J., Bo¨schen, A., Pitner, W.-R., Welz-Biermann, U., Jastorff, B., Ranke, J., 2007a. Effects of

different head groups and functionalized side chains on the aquatic toxicity of ionic liquids. Green Chem. 9, 1170–1179. Stolte, S., Arning, J., Bottin-Weber, U., Mu¨ller, A., Pitner, W.-R., Welz-Biermann, U., Jastorff, B., Ranke, J., 2007b. Effects of different head groups and functionalized side chains on the cytotoxicity of ionic liquids. Green Chem. 9, 760–767. Stolte, S., Abdulkarim, S., Arning, J., Blomeyer-Nienstedt, A.-K., Bottin-Weber, U., Matzke, M., Ranke, J., Jastorff, B., Tho¨ming, J., 2008. Primary biodegradation of ionic liquid cations, identification of degradation products of 1-methyl-3-octylimidazolium chloride and electrochemical wastewater treatment of poorly biodegradable compounds. Green Chem. 10, 214–224. ´ ska, S., Buszewski, B., 2009. Study of toxicity of imidazolium Studzin ionic liquids to watercress (Lepidium sativum L.). Anal. Bioanal. Chem. 393, 983–990. Swatloski, R.P., Holbrey, J.D., Rogers, R.D., 2003. Ionic liquids are not always green: hydrolysis of 1-butyl-3-methylimidazolium hexafluorophosphate. Green Chem. 5, 361–363. Swatloski, R.P., Holbrey, J.D., Memon, S.B., Caldwell, G.A., Caldwell, K.A., Rogers, R.D., 2004. Using Caenorhabditis elegans to probe toxicity of 1-alkyl-3-methylimidazolium chloride based ionic liquids. Chem. Commun, 668–669. Torrecilla, J.S., Garcıá, J., Rojo, E., Rodrı´guez, R., 2009. Estimation of toxicity of ionic liquids in leukemia rat cell line and acetylcholinesterase enzyme by principal component analysis, neural networks and multiple linear regressions. J. Hazard. Mater 164, 182–194. U.S. Environmental Protection Agency 1998. Fate, transport, and transformation test guidelines. OPPTS 835.3110. Readily Biodegradability. Vijayaraghavan, K., Pham, T.P.T., Cho, C.-W., Won, S.W., Choi, S. B., Mao, J., Kim, S., Kim, Y.-R., Chung, B.W., Yun, Y.-S., 2009. An assessment on the interaction of a hydrophilic ionic liquid with different sorbents. Ind. Eng. Chem. Res. 48, 7283–7288. Wang, L.-S., Wang, L., Wang, L., Wang, G., Li, Z.-H., Wang, J.-J., 2009. Effect of 1-butyl-3-methylimidazolium tetrafluoroborate on the wheat (Triticum aestivum L.) seedlings. Environ. Toxicol 24, 296–303. Wang, X., Ohlin, C.A., Lu, Q., Fei, Z., Hu, J., Dyson, P.J., 2007. Cytotoxicity of ionic liquids and precursor compounds towards human cell line HeLa. Green Chem. 9, 1191–1197. Wasserscheid, P., Keim, W., 2000. Ionic liquids–New ‘‘solutions’’ for transition metal catalysis. Angew. Chem. Int. Ed. 39, 3772–3789. Wells, A.S., Coombe, V.T., 2006. On the freshwater ecotoxicity and biodegradation properties of some common ionic liquids. Org. Pro. Res. Dev. 10, 794–798. Welton, T., 1999. Room-temperature ionic liquids, solvents for synthesis and catalysis. Chem. Rev. 99, 2071–2083. Willauer, H.D., Huddleston, J.G., Griffin, S.T., Rogers, R.D., 1999. Partitioning of aromatic molecules in aqueous biphasic systems. Sep. Sci. Technol 34, 1069–1090. Wong, D.S.H., Chen, J.P., Chang, J.M., Chou, C.H., 2002. Phase equilibria of water and ionic liquids [emim][PF6] and [bmim][PF6]. Fluid Phase Equilibr 194–197, 1089–1095. Yu, M., Li, S.-M., Li, X.-Y., Zhang, B.-J., Wang, J.-J., 2009a. Acute effects of 1-octyl-3-methylimidazolium bromide ionic liquid on the antioxidant enzyme system of mouse liver. Ecotox. Environ. Saf. 71, 903–906. Yu, M., Wang, S.-H., Luo, Y.-R., Han, Y.-W., Li, X.-Y., Zhang, B.-J., Wang, J.-J., 2009b. Effects of the 1-alkyl-3-methylimidazolium bromide ionic liquids on the antioxidant defense system of Daphnia magna. Ecotox. Environ. Saf. 72, 1798–1804. Zhang, C., Malhotra, S.V., 2005. Increased paraoxon detection by acetylcholinesterase inactivation with ionic liquid additives. Tatanta 67, 560–563.




A QSAR model for predicting rejection of emerging contaminants (pharmaceuticals, endocrine disruptors) by nanofiltration membranes Victor Yangali-Quintanillaa,b,*, Anwar Sadmania, Megan McConvillea,b, Maria Kennedya, Gary Amya,b a

UNESCO-IHE, Institute for Water Education, Westvest 7, 2611AX Delft, The Netherlands Delft University of Technology, Stevinweg 1, 2628CN Delft, The Netherlands

b

article info

abstract

Article history:

A quantitative structure activity relationship (QSAR) model has been produced for pre-

Received 25 February 2009

dicting rejection of emerging contaminants (pharmaceuticals, endocrine disruptors,


pesticides and other organic compounds) by polyamide nanofiltration (NF) membranes.

4 June 2009

Principal component analysis, partial least square regression and multiple linear regres-

Accepted 26 June 2009

sions were used to find a general QSAR equation that combines interactions between

Available online 3 July 2009

membrane characteristics, filtration operating conditions and compound properties for predicting rejection. Membrane characteristics related to hydrophobicity (contact angle),

Keywords:

salt rejection, and surface charge (zeta potential); compound properties describing

Pharmaceuticals

hydrophobicity (log Kow, log D), polarity (dipole moment), and size (molar volume, molec-

Endocrine disruptors

ular length, molecular depth, equivalent width, molecular weight); and operating condi-

Nanofiltration

tions namely flux, pressure, cross flow velocity, back diffusion mass transfer coefficient,

Modelling

hydrodynamic ratio (Jo/k), and recovery were identified as candidate variables for rejection

QSAR

prediction. An experimental database produced by the authors that accounts for 106 rejection cases of emerging contaminants by NF membranes as result of eight experiments with clean and fouled membranes (NF-90, NF-200) was used to produce the QSAR model. Subsequently, using the QSAR model, rejection predictions were made for external experimental databases. Actual rejections were compared against predicted rejections and acceptable R2 correlation coefficients were found (0.75 and 0.84) for the best models. Additionally, leave-one-out cross-validation of the models achieved a Q2 of 0.72 for internal validation. In conclusion, a unified general QSAR equation was able to predict rejections of emerging contaminants during nanofiltration; moreover the present approach is a basis to continue investigation using multivariate analysis techniques for understanding membrane rejection of organic compounds. ª 2009 Elsevier Ltd. All rights reserved.

* Corresponding author: UNESCO-IHE, Institute for Water Education, Westvest 7, 2611AX Delft, The Netherlands. Tel.: þ31 15 215 1745; fax: þ31 15 215 2921. E-mail addresses: [email protected], [email protected] (V. Yangali-Quintanilla). 0043-1354/$ – see front matter ª 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2009.06.054

374

1.


Introduction

Nanofiltration (NF) and reverse osmosis (RO) are technologies that provide medium to high rejections of organic compounds present as emerging contaminants (micropollutants) in water, namely endocrine disrupting compounds (EDCs), pharmaceutically active compounds (PhACs), and pesticides (Kiso et al., 2001; Scha¨fer et al., 2003; Kimura et al., 2003, 2004; Nghiem et al., 2004). The presence of micropollutants has been identified in surface water bodies, sewage treatment plant effluents, and stages of drinking water treatment plants and even at trace-levels in finished drinking water (Kolpin et al., 2002; Heberer, 2002; Castiglioni et al., 2006). The possible effects on aquatic organisms and human health, associated with the consumption of water containing low concentrations of single compounds, have been presented in toxicology studies (Pomati et al., 2006; Escher et al., 2005; Vosges et al., 2008). The studies demonstrate that researchers do not yet understand the exact risks from decades of persistent exposure to random combinations of low levels of pharmaceuticals, EDCs, and other organic compounds; hence, the long-term effects of consumption of water containing low concentrations of micropollutants will remain as an unanswered question for the foreseeable future. Meanwhile, water treatment facilities are implementing monitoring programs, research organizations dealing with water reuse have published reports, and studies have addressed the topic (Drewes et al., 2006; Verliefde et al., 2007). An important aspect to deal with the problem has been the identification of compound physicochemical properties and membrane characteristics to explain transport, adsorption and removal of micropollutants by different mechanisms, explicitly size/ steric exclusion, hydrophobic adsorption and partitioning, and electrostatic repulsion (Kiso et al., 2001; Scha¨fer et al., 2003; Kimura et al., 2003; Nghiem et al., 2004, Ozaki and Li, 2002; Van der Bruggen and Vandecasteele, 2002; Bellona and Drewes, 2005; Xu et al., 2005). A number of articles have proposed a mechanistic understanding of the interaction between membranes and organic compounds; others have tried to apply fitting parameter models to model rejection (Cornelissen et al., 2005; Kim et al., 2007; Verliefde et al., 2008). However, there have been few models to ‘‘predict’’ the rejection of compounds. To overcome that status, our objective was to create a general QSAR model to predict rejections based on an integral approach that considers membrane characteristics, filtration operating conditions and physicochemical compound properties. A quantitative structure activity relationship (QSAR) is a method that relates an activity of a set of compounds quantitatively to chemicals descriptors (structure or property) of those compounds (Sawyer et al., 2003). QSAR has the objective of prediction but maintaining a relationship to mechanistic interpretation. Applications of QSAR for the development of models to find relationships between membranes and organic compounds have been presented in journals related to drug discovery and medicinal chemistry for analysis of permeability of membranes to organic compounds (Ren et al., 1996; Fujikawa et al., 2007). The study of reverse

osmosis membranes has also experienced the application of QSAR principles. Campbell et al. (1999) performed a QSAR analysis of surfactants influencing attachment of a mycobacterium to cellulose and aromatic polyamide reverse osmosis membranes; their objective was to understand the relationship between surfactant molecular properties and activity on the membrane surface to inhibit bacterial attachment to the membrane in order to reduce biofilm formation and to increase permeate production. More recently, Libotean et al. (2008) developed an artificial neural network model based on quantitative structure-property relations, the model claim to predict organic solute passage through reverse osmosis membranes considering simultaneous correlation of organic solutes (molecular descriptors) and membrane properties. Alike, our study uses the concept of QSAR analysis to quantify an activity, compound rejection by a membrane, in terms of organic compound physicochemical properties, membrane characteristics (salt rejection, pure water permeability, molecular weight cut-off, charge, hydrophobicity) and operating conditions (pressure, flux, cross flow velocity, back diffusion mass transfer coefficient, recovery). In this work a QSAR model was constructed with internal experimental data used for training. The model was internally validated using measures of goodness of fit and prediction. Subsequently, after identification of a relationship in form of an equation, estimation of rejections for an external dataset for different compounds and membranes were used to externally validate the model. Similarly, rejections of more emerging organic contaminants can be predicted in advance, before nanofiltration or reverse osmosis applications. Nevertheless, the QSAR model is applicable in the range of boundary experimental conditions that will be defined in the experimental section of this publication (Section 3).

2.

Theory

2.1.

Principal component analysis

Principal component analysis (PCA) is a method that allows simplification of many variables into a group of a few variables that might be measuring the same principles of a system. It may occur that a system considers an abundance of variables to explain a process; in this case principal component analysis reduces the redundancy of information. The general objectives of PCA are data reduction and interpretation. Although p variables (components) are required to reproduce the total system variability, often much of this variability can be accounted for by a small number k of principal components. The k is the number of components (reduced) that represent the initial p variables. In general, PCA is concerned with whether the covariances or correlations between a set of observed variables x1, x2, ., xp can be explained in terms of a smaller number of unobservable components, c1, c2, ., ck, where k < p. Thus, there is as much information in the k components as there is in the original p variables. Comprehensive details about the theory of PCA can be found elsewhere (Jolliffe, 2002; Johnson and Wichern, 2007; Everitt and Dunn, 2001).


2.2.

Multiple linear regression

Multiple linear regression is a method of analysis for assessing the strength of the relationship between a set of explanatory variables known as independent variables, and a single response or dependent variable. Applying multiple regression analysis to a set of data results in what are known as regression coefficients, one for each explanatory variable. The multiple regression model for a response variable, y, with observed values, y1, y2, ., yn (where n is the sample size) and q explanatory variables, x1, x2, ., xq with observed values, x1i, x2i, ., xqi for i ¼ 1, ., n, is yi ¼ b0 þ b1 x1i þ b2 x2i þ / þ bq xqi þ 3i

(1)

The regression coefficients, b0, b1, ., bq, are generally estimated by least squares. The term 3i is the residual or error for individual i and represents the deviation of the observed value of the response for this individual from that expected by the model. These error terms are assumed to have a normal distribution with variance s2. The fit of a multiple regression model can be judged with calculation of the multiple correlation coefficient, R, defined as the correlation between the observed values of the response variable and the values predicted by the model. The squared value of R (R2) gives the proportion of the variability of the response variable accounted for by the explanatory variables. Analysis of variance (ANOVA) will provide an F-test of the null hypothesis that each of b0, b1, ., bq, is equal to zero, or in other words that R2 is zero (Landau and Everitt, 2004).

2.3. Principal component regression and partial least squares regression Principal component regression (PCR) is a method in which the components from the principal component method are used for regression. Hence, the principal components of the matrix X are used as regressors of a dependent Y. The orthogonality of the principal components eliminates the multicollinearity problem. But, the problem of choosing an optimum subset of predictors remains. A possible strategy is to keep only a few of the first components. But they are chosen to explain X rather than Y, and therefore, nothing guarantees that the principal components, which ‘‘explain’’ X, are relevant for Y. Problems may arise, however, if there is a lot of variation in X. PCR finds, somewhat uncritically, those latent variables that describe as much as possible of the variation in X. But sometimes the variable itself gives rise to only small variations in X, and if the interferences vary a lot, then the latent variables found by PCR may not be particularly good at describing Y. In the worst case important information may be hidden in directions in the X-space that PCR interprets as disturbance, and therefore leaves out. Partial least squares regression (PLS) is able to cope better with this problem, by forming variables that are relevant for describing Y. By contrast, PLS regression finds components from X that are also relevant for Y. Specifically, PLS regression searches for a set of components (called latent vectors) that performs a simultaneous decomposition of X and Y with the constraint that these components explain as much as possible of the

375

covariance between X and Y. This step generalizes PCA. The goal of PLS regression is to predict Y from X and to describe their common structure (Abdi, 2003; Jørgensen and Goegebeur, 2006).

3.

Experimental

3.1. Chemicals, membranes, materials and experimental conditions The list of organic compounds representing emerging contaminants is presented in Table 1. The compounds were selected considering: their occurrence in surface water and drinking water, their identification as priority emerging contaminants, the availability of analytical methods, their quantitative and qualitative representation of physicochemical properties. The pharmaceutical compounds (caffeine, sulfamethoxazole, acetaminophen, phenacetin, phenazone, carbamazepine, naproxen, ibuprofen, metronidazole), endocrine disrupting compounds (17b-estradiol, estrone, bisphenol A, nonylphenol, atrazine) and sodium alginate were purchased from Sigma–Aldrich (Sigma–Aldrich, Schnelldorf, Germany). Potassium chloride, sodium hydroxide, hydrochloric acid and magnesium sulphate anhydrous were purchased from J.T.Baker (J.T.Baker, Deventer, Netherlands). Two thin film composite NF membranes were selected for this study (NF-200 and NF-90, Dow-Filmtec, Dow Chemical Co., Midland, MI). The experimental setup consisted of two filtration SEPA CF II (GE Osmonics, Minnetonka, MN) cells and cell holders in parallel, in order to increase permeate production and achieve hydrodynamic conditions, two hydraulic pumps (Power Team, Bega Int. BV, Netherlands), a 60 litres stainless steel tank (Tummers, Netherlands), a positive displacement pump (Hydra-Cell pump, Wanner Eng. Inc., Minneapolis, MN), a frequency converter (VLT microdrive, Danfoss, SA), a chiller/heater (Julabo, Germany), control needle valves, pressure gauges, flow meters, a proportional pressure relief valve and stainless steel tubings (Swagelok BV, Netherlands), a digital balance (Sartorius, Germany) and, a computer for flow rate data acquisition. A piece of membrane was compacted with deionised water for 6 h at a pressure of 276 kPa before performing an experiment with the membrane. The experiments were conducted in a recycle mode in which permeate and concentrate were recirculated into the feed tank for the first 72 h (a pre-equilibration period); then, permeate was collected within the next 24 h. The feed solution of all the experiments contained a cocktail of 14 compounds (concentration ranging from 6.5 to 65 mg/L). The main reason for conducting experiments at concentrations of mg/L was to accelerate steady state (after membrane adsorption) conditions in a limited time (3 days). At very low concentrations more time would be needed to achieve steady state conditions and at short time tests low concentrations may lead to over-estimation of rejections. A specific flux decline of 15% of initial flux was targeted to foul the membranes using a feed solution containing w10 mg/L DOC of sodium alginate. The study of Lee et al. (2004) concluded that polysaccharides were important membrane foulants. Sodium alginate was used as

376


Table 1 – List of compounds and physicochemical properties. Name

Acetaminophen Phenacetine Caffeine Metronidazole Phenazone Sulfamethoxazole Naproxen Ibuprofen Carbamazepine Atrazine 17b-estradiol Estrone Nonylphenol Bisphenol A a b c d e f

Molecular Acid Log Kowb Log Da (pH 7) weight pKa 20 Ca (g/mol) 151 179 194 171 188 253 230 206 236 216 272 270 220 228

10.2 N/A N/A N/A N/A 5.7 4.3 4.3 N/A N/A 10.3 10.3 10.3 10.3

0.46 1.58 0.07 0.02 0.38 0.89 3.18 3.97 2.45 2.61 4.01 3.13 5.71 3.32

0.23 1.68 0.45 0.27 0.54 0.45 0.34 0.77 2.58 2.52 3.94 3.46 5.88 3.86

Dipole moment (debye)c

Molar volumed (cm3/ mol)

Molec. length (nm)e

Molec. width (nm)e

Molec. depth (nm)e

4.55 4.05 3.71 6.30 4.44 7.34 2.55 4.95 3.66 3.43 1.56 3.45 1.02 2.13

120.90 163.00 133.30 117.80 162.70 173.10 192.20 200.30 186.50 169.80 232.60 232.10 236.20 199.50

1.14 1.35 0.98 0.93 1.17 1.33 1.37 1.39 1.20 1.26 1.39 1.39 1.79 1.25

0.68 0.69 0.87 0.90 0.78 0.71 0.78 0.73 0.92 1.00 0.85 0.85 0.75 0.83

0.42 0.42 0.56 0.48 0.56 0.58 0.75 0.55 0.58 0.55 0.65 0.67 0.59 0.75

Equiv. Classificationf width (nm)e 0.53 0.54 0.70 0.66 0.66 0.64 0.76 0.64 0.73 0.74 0.74 0.76 0.66 0.79

HL-neutral HL-neutral HL-neutral HL-neutral HL-neutral HL-ionic HB-ionic HB-ionic HB-neutral HB-neutral HB-neutral HB-neutral HB-neutral HB-neutral

ADME/Tox Web Software. Experimental database: SRC PhysProp Database. Chem3D Ultra 7.0. ACD/ChemSketch Properties Batch. Molecular Modeling Pro. HL ¼ hydrophilic, HB ¼ hydrophobic.

surrogate of polysaccharides. Alginate is frequently used as a model for organic matter of algae origin (Henderson et al., 2008). Experiments were carried out for both clean and fouled NF-90 and NF-200 membranes. The selection of membranes was based on a qualitative rejection assessment of emerging contaminants with molecular weight of more than 150 Da by membranes with a MWCO between 200 and 300 Da. A total of eight experiments were performed; at hydrodynamic ratio of pure water permeation flux to back diffusion mass transfer coefficient (J0/k) of w1 and recovery 3% (NF-90 clean, NF-200 clean), at J0/k of w1 and recovery 8% (NF-200 clean), at J0/k of w2 and recovery 3 (NF-90 clean), and at J0/k of w2 and recovery 8 (NF-90 clean and fouled, NF-200 clean and fouled). The calculation method of k (back diffusion mass transfer coefficient) was presented in a previous publication (YangaliQuintanilla et al., 2009). All the experiments were carried out at a controlled temperature of 20 C, an ionic strength of 10 mM as KCl and a pH of 7. The transmembrane pressures were in the range of 276–483 kPa. The fluxes were between 4.3 and 30.2 L/m2 per day; cross flow velocities were between 0.5 and 4.5 cm/s. The experiments produced a total internal dataset of 106 rejection cases; the dataset can be accessed as supplementary data. The boundary experimental conditions of the internal dataset are presented in Table 2. The internal dataset was used to develop the model. An external dataset that gathered three different datasets was used for validation of the model. The external dataset is presented as supplementary data. Experimental conditions for the first part of the external dataset can be obtained from Kim et al. (2007) and Yangali-Quintanilla et al. (2008). Experimental conditions for the second and third parts can be found in Verliefde et al. (2008); the data correspond to filtration experiments using synthetic water solutions.

3.2. Analytical equipment, analyses of compounds and membranes The pharmaceuticals and endocrine disruptors (with the exception of atrazine) were analyzed by Technologiezentrum Wasser (TZW, Karlsruhe, Germany). The detection limit was 10 ng/L per compound. The uncertainty of estimates was of 15% according to a validation method of the analysis protocol and due to high concentrations of the samples (mg/L). The analyses of pharmaceuticals were performed according to

Table 2 – Data range of membrane characteristics, operating conditions and rejections. Variable Molecular weight cut-off (MWCO) Pure water permeability (PWP) Salt rejection (SR)a Zeta potential (ZP) Contact angle (CA) Pressure (P) Cross flow velocity (v) Back diffusion mass transfer coefficient (k) Flux (J ) Hydrodynamic ratio (J0/k) Recovery (recov) Rejection (rejection)

Unit

Min. value Max. value

Da

200

300

L/m2 per day/kPa – MV

0.86

2.23

kPa cm/s cm/s

0.96 48.04 39.3 276 0.5 2.70E-04

0.98 10.78 58.0 483 4.5 5.99E-04

L/m2 per day –

4.3 1

30.2 2

% %

3 17.7

8 99.0

a 2000 mg/L MgSO4, 25 C, recovery 15%, pressure 1034 kPa, pH 8.


the protocols described by Sacher et al. (2001, 2008). Analyses of 17b-estradiol, estrone, nonylphenol and bisphenol A were done by gas chromatography/mass spectrometry (GC/MS) after automated solid-phase extraction onto a polymeric material and subsequent silylation of the analytes. First, 10 ml of a 50 ng/ml solution of 4-n-nonylphenol in acetone which was used as internal standard for the overall procedure were added to an aliquot of the water sample (1000 ml). Automated solid-phase extraction (Tekmar AutoTrace, Germany) was done on plastic cartridges filled with 200 mg of bondelut material (Fa. Varian, Darmstadt, Germany). After the enrichment step the solid-phase material was dried in a gentle stream of nitrogen. Elution was done with 4 ml of acetone. The acetone was evaporated to 100 ml in a stream of nitrogen and to dryness in a drying oven at 80 C. The dry residue was reconstituted with 100 ml of a silylation reagent mixture (N-methyl-N-trimethylsilyltrifluoro acetamide (MSTFA)/2% trimethyliodo silane). After a reaction time of 20 min at 80 C (drying oven), determination of the derivatives was done by GC/MS using a 6890 GC/MS system from Agilent Technologies (Waldbronn, Germany). Concentrations of atrazine were determined using microplate enzyme-linked immunoabsorbent assay (ELISA) kits (Abraxis LLC, Warminster, PA). Atrazine was determined with a detection limit of 0.04 mg/L, and uncertainty of 15%. To determine the hydrophobicity of membranes, contact angles of clean and fouled membrane surfaces were measured with CAM200 optical contact angle and surface tension meter (KSV Instruments, Finland) at Delft University of Technology; to measure contact angle, the sessile drop method was used. Surface charge, in terms of zeta potential, of clean and fouled membranes was quantified using ELS-8000 zeta potential analyzer (Otsuka Electronics, Japan). The zeta potential analyses were determined using a Milli-Q water solution at pH 7 and ionic strength of 10 mM KCl. The zeta potential was determined using the electrophoresis method using a cell consisting of membrane and quartz cells. The zeta potential was calculated from the electrophoretic mobility using the Smoluchowski formula, a detailed explanation of calculation was provided in a previous publication (Shim et al., 2002). The pH of the solutions was measured using a calibrated Metrohm 691 pH-meter (Metrohm AG, Herisau, Switzerland); the electrical conductivity and temperature were measured with a WTW Cond 330i (WTW GmbH, Weilheim, Germany) portable conductivity meter. Clean and fouled membranes were characterized to determine magnesium sulphate salt rejection at standard conditions specified by manufacturers, a pure water solution containing 2000 mg/L of magnesium sulphate at 25 C and pH 8 was filtrated at pressure of 1034 kPa and recovery of 15%.

3.3.

Characterization and classification of compounds

The acid dissociation constant as log Ka (pKa) was used to determine the speciation of the organic compound in ionic species at pH 7. For hydrophobicity determination, log Kow and log D were used; log Kow is the octanol–water partition coefficient and log D is the ratio of the equilibrium concentrations of all species (unionized and ionized) of a molecule in octanol to the same species in the water phase. Values of pKa and log D

377

were calculated by ADME/Tox web software. Solubility and log Kow values were obtained from SRC Physprop experimental database. The value of the molecular dipole moment was equal to the vector sum of the individual bond dipole moments. Dipole moment was calculated by Chem3D Ultra 7.0, Cambridgesoft. Size descriptors included molar volume (MV), molecular length, molecular width, molecular depth and equivalent molecular width. The molecular length is defined as the distance between the two most distant atoms. The molecular width and molecular depth (width > depth) are measured by projecting the molecule on the plane perpendicular to the length axis and the equivalent molecular width is defined as the geometric mean of width and depth (Santos et al., 2006). Molar volumes were calculated using the program ACD/ChemSketch Properties Batch, ACD/Labs; and Molecular Modeling Pro, ChemSW, was used to compute size descriptors after optimization geometry of a molecule from the interaction of conformational analysis and energy minimization with a semi-empiric method MOPAC-PM3. Based on pKa and log Kow values, the compounds were classified as hydrophilic neutral, hydrophilic ionic, hydrophobic ionic and hydrophobic neutral (see Table 1). Compounds with log Kow 2 were referred to as hydrophobic; therefore those with log Kow < 2 were hydrophilic. The classification was based on an early reference (Connell, 1990). Although the value of 2 may seem low to consider hydrophobicity, the classification was not used in constructing the models and therefore the magnitude of log Kow or log D became more important. Table 1 shows the calculated values of molecular weight, pKa, log Kow, log D, dipole moment, molar volume, molecular length, molecular width, molecular depth and equivalent width.

4.

Results and discussion

4.1.

QSAR methodology

The procedure to find a general QSAR equation to describe rejection was performed in four phases. The first phase was the organization of data from the experimental part. The data comprised of 106 rejection cases. The database showing the rejection cases is presented as supplementary data. A total of 21 initial variables were used. The variables considered as compound descriptors were molecular weight (MW), solubility, log Kow, log D, dipole moment, molar volume, molecular length, molecular width, molecular depth and equivalent width; variables describing membrane characteristics were molecular weight cut-off (MWCO), pure water permeability (PWP), magnesium sulphate salt rejection (SR), charge of the membrane as zeta potential (ZP), and hydrophobicity as contact angle (CA); variables describing operating conditions were operating pressure (P), cross flow velocity (v), back diffusion mass transfer coefficient (k), flux (J ), ratio of pure water permeation flux J0 and back diffusion mass transfer coefficient (J0/k) and recovery. The range of values for membrane characteristics, operating conditions and rejections is presented in Table 2. The second phase was dedicated to the process of variables reduction using a correlation matrix and factor analysis with principal component analysis. The third phase corresponded to the regression analysis. In

378


the third phase three methodologies were implemented; the first was principal component analysis (PCA), with sequential application of multiple linear regressions (MLR). The second method was the use of partial least squares (PLS) regression and MLR; and the third method was the use of MLR only. The last phase was the validation process. The model was internally validated using measures of goodness of fit (regression coefficients) and prediction (leave-one-out cross-validation); Section 4.5 gives details about the validation process. External validation of the general QSAR model was implemented by predicting rejections for an external dataset of experiments performed with different compounds and membranes, and with comparable operating conditions. PCA and PLS were performed using the research and statistical package SPSS Statistics 16.0. Leave-one-out cross-validations of the models were performed with MobyDigs (Talete, Milano, Italy).

4.2. Variables reduction with principal component analysis and QSAR model The correlation matrix of the initial 21 variables was scrutinized in order to obtain a not positive definite matrix as a requisite of PCA; the matrix is accompanied as supplementary data. A matrix is called not positive definite when there are both positive and negative eigenvalues. In the case of symmetric matrices, such as a correlation matrix, positive definiteness will only hold if the matrix and every principal submatrix have a positive determinant. A non-positive definite input matrix may signal a perfect linear dependency of one variable on another, known as collinearity. This was the case for MWCO and salt rejection (SR) that were perfectly linearly correlated. Therefore application of PCA considering independently one variable or the other will give the same results of variables reduction and number of components. In other words, MWCO will not be excluded with PCA, the variable will be separated in advance and the results obtained for SR may be replaced by the variable MWCO, or vice versa. Once an appropriate matrix was defined, the variables were analyzed in terms of how significant their correlations with rejection were; those correlations are also shown as an additional row and column of the 21 21 variables matrix. Rejection is only a reference variable to evaluate correlation with the rest of variables. After a sequential implementation of PCA, three components were extracted; they defined the initial database of 21 variables with 11 variables describing three relations namely membrane/operating-conditions (comp. 1: flux, pure water permeability, salt rejection, zeta potential, mass transfer coefficient, cross flow velocity), hydrophobicity/size (comp. 2: length, log Kow, log D) and size (comp. 3: equivalent width, depth). The final three components accounted for 89.3% explanation of total variance. It is important to mention that these results were produced for the experimental dataset. The next step was the implementation of multiple linear regression (MLR) using the new set of variables. The use of MLR after PCA presents the advantage of a more simplified modelling approach. Moreover, the analysis of data before MLR may help to identify variables that are similar in response, which was the case of SR and MWCO. The dependent variable for all regression analyses was rejection. Two

methods of linear regression were used, the first method is called enter (forced) method; which performs a regression with the contribution of all variables entered to model the dependent variable. The second method is stepwise regression; which is a more sophisticated method. Each variable is entered in sequence and its contribution is assessed according to an F-test. In the present study an F-test with a statistical significance >0.10 implied removal of the variable, and F-test with a significance 0.9 as excellent (Eriksson et al., 2003). For the developed QSAR models, the model with SR (Eq. (2)) presented a Q2 leave-one-out of 0.72, and the model with MWCO (Eq. (3)) presented a Q2 leave-oneout of 0.72. After internal cross-validation it was demonstrated that Eqs. (2) and (3) were valid to model rejection, however, an adjustment must be made to the equation before using it to compare measured vs. predicted rejections for external databases. This adjustment was necessary to overcome the mathematical structure of the equation. Using a physical interpretation, it was evident that size parameters referring to variables length and equivalent width may be large enough to cause rejection predictions over 100%, which can be explained after observing positive coefficients for equivalent width and length. This situation may also be detrimental for rejection predictions of ionic compounds of medium to large size (0.6–1.2 nm as equivalent width) that mostly are rejected due to electrostatic repulsion and less steric hindrance. Therefore Eqs. (2) and (3) can be transformed to the following conditional equation rejection ¼

100 if QSAR model 100 QSAR model

(4)

An external dataset (that gathered three different datasets) was selected for external validation of the QSAR model. The external dataset is presented as supplementary data. The first part of the external dataset corresponds to membrane Filmtec NF-90. The second part corresponds to NF membrane Trisep TS-80 and the third part corresponds to NF membrane Desal HL. Desal HL membrane has a main difference with NF-90 and Trisep TS-80 membranes, viz. the MWCO of Desal HL is in the range of 150–300 Da, while NF-90 and Trisep TS-80 were reported to have a MWCO of 200 Da. Therefore an average MWCO of 225 Da was assumed for Desal HL during the application of Eq. (3). It is worthwhile to mention that the

second and third parts of the external dataset were generated using spiral wound membrane elements instead of flat sheet membranes. Fig. 2a show plotted results of measured rejections vs. predicted rejections after calculations with Eq. (4), for QSAR model with Eq. (2) (SR). According to Fig. 2a an R2 of 0.75 was obtained after considering all compounds of the external dataset. However, after observing the rejection cases by NF-90 for bromoform (BF) and trichloroethene (TCE), it appeared that BF and TCE may be considered atypical results because their rejections did not correspond to their size and hydrophobicity when compared to other compounds with comparable molecular descriptors. According to Table 3 BF has approximately the same hydrophobicity and polarity as CF, but BF is bigger than CF; therefore, measured rejection for BF was expected to be higher than rejection of CF (0%) due to size exclusion. The measured rejection of TCE (3%) was not comparable to the rejection of perchloroethene (39%) although they have the same length but a very small difference in equivalent width. In an experiment conducted by Kim et al. (2007) rejections of BF and TCE were of 50 and 33%, respectively, for a low pressure reverse osmosis membrane. It was also observed that the measured rejection (53%) of linuron (LNU) may be considered as atypical observation because a higher rejection was expected. According to Table 3, monolinuron is close to LNU in size and polarity; thus measured rejection of monolinuron was of 77%, higher than 53%. Also according to Table 3, carbamazepine is close to LNU in size and hydrophobicity; thus measured rejection of carbamazepine was of 94%, higher than 53%. A similar explanation was given for the low rejection (70%) of N-acetylL-tyrosine (NAT) by Desal HL compared to 94% rejection by TS-80 as can be seen in Table 3. Other non-expected rejection was that of 2-(1H)-quinoline (QNL), with a rejection of 22% when compared to 2-methoxyethanol and perchloroethene with rejections of 32 and 39%, respectively; even though the length of QNL is greater than the length of 2-methoxyethanol and perchloroethene, a lower rejection of QNL was observed. However, rejections of methacetin (MTC) and NDPA may be influenced by their small equivalent width as can be seen in

381


b 100

100

80

80

measured rejection (%)


a

NAT

MTBE

60 LNU NDPA

40 ETH

MTC QNL

20

60

40

20

NDPA

R2 = 0.84

2

BF

0

0

20

MTC

TCE

R = 0.75 0

40

60

80

100

0

20

predicted rejection (%)

NF-90

TS-80

HL

40

60

80

100


Linear ( )

NF-90

TS-80

HL

Linear ( )

Fig. 2 – Predicted rejections for external dataset using magnesium sulphate SR: (a) all external data; and (b) selected external data.

Table 3. The response of the model for rejection of those particular compounds (with 2*equivalent width < or w length) was of over prediction, but the model presented better response for rejections of metribuzin, atrazin and N-acetyl-Ltyrosine (all with length >2*equivalent width) as observed in Table 3. Ethanol (ETH) and MTBE with rejections of 38 and 60%, respectively, were expected to be lower for Desal HL membrane because it was observed that rejection of ETH was of 9% for TS-80; and MTBE was expected to have a rejection compared to that of 2-methoxyethanol (32%) or perchloroethene (39%) due to proximities in size. After selection

and justified separation of the mentioned rejection cases, Fig. 2b presents the predictions of the external dataset with an R2 of 0.84. In a similar explanatory scenario, Figs. 3a,b show measured rejections vs. predicted rejections after calculations with Eq. (4), for QSAR model with equation 3 (MWCO). The main difference between Figs. 2b and 3b was that the model with MWCO (Fig. 3b) showed a lower R2 (0.80) than the model with SR (R2 ¼ 0.84), meaning that the latter had a better goodness of fit for external prediction response. Moreover, the characterization of magnesium sulphate salt rejection for a membrane may be preferred instead of MWCO, particularly

Table 3 – Partial list of rejections and compound properties of external dataset. Compound Chloroform Ethanol Ethanol Carbontetrachloride Bromoform MTBE Trichloroethene Perchloroethene 2-methoxyethanol 2-(1H)-quinoline NDPA NDPA Metribuzin Carbamazepine Linuron Monolinuron Atrazin Methacetin Methacetin N-acetyl-L-tyrosine N-acetyl-L-tyrosine

Abb.

ETH BF MTBE TCE

QNL NDPA NDPA

LNU

MTC MTC NAT NAT

Length

Eqwidth

Depth

Log D

Dipole

Measured rejection

Predicted rejection

0.53 0.64 0.64 0.64 0.69 0.77 0.78 0.78 0.87 1.00 1.16 1.16 1.17 1.20 1.21 1.22 1.26 1.28 1.28 1.33 1.33

0.42 0.52 0.52 0.6 0.56 0.63 0.49 0.59 0.52 0.52 0.60 0.60 0.74 0.73 0.69 0.69 0.74 0.52 0.52 0.71 0.71

0.35 0.51 0.51 0.57 0.48 0.59 0.36 0.45 0.51 0.36 0.53 0.53 0.64 0.58 0.53 0.65 0.55 0.42 0.42 0.60 0.60

1.97 0.31 0.31 2.83 2.40 0.94 2.29 3.40 0.77 1.26 1.36 1.36 0.47 2.45 3.20 2.30 2.61 1.03 1.03 2.18 2.18

1.12 1.55 1.55 0.30 1.00 1.37 0.95 0.11 0.25 3.38 3.40 3.40 0.52 3.66 2.11 2.02 3.43 2.20 2.20 3.45 3.45

0 9 38 35 0 60 3 39 32 22 45 19 97 88 53 79 91 38 5 94 70

0 14 0 26 33 25 36 46 34 53 69 55 99 94 84 77 100 72 58 100 100

Membrane NF-90 TS-80 Desal HL NF-90 NF-90 Desal HL NF-90 NF-90 TS-80 TS-80 TS-80 Desal HL TS-80 TS-80 TS-80 TS-80 TS-80 TS-80 Desal HL TS-80 Desal HL

382


a

b 100

80 NAT

MTBE

60 LNU NDPA ETH

40

MTC QNL NDPA

20



100

80

60

40

20

R2 = 0.80

2

MTC BF

0 0

R = 0.74

TCE

20

0 40

60

80

100

0

20


NF-90

TS-80

HL

40

60

80

100


Linear ( )

NF-90

TS-80

HL

Linear ( )

Fig. 3 – Predicted rejections for external dataset using MWCO: (a) all external data; and (b) selected external data.

for nanofiltration and low pressure reverse osmosis membranes; besides, the effect of fouling in membranes can also be quantified by salt rejection experiments. We can state that the QSAR model with SR demonstrated to be acceptable for the external dataset of NF-90, Trisep TS-80 and Desal HL with an R2 of 0.75 and 0.84 for the total external dataset and justified selected external dataset, respectively. Although the model can be valid with limitations related to boundary experimental conditions mentioned in Section 3, its applicability and approach can be of value for the construction of a model with combined datasets organized in training and testing groups.

5.

Acknowledgements The authors acknowledge Delft Cluster and EU Techneau Project for funding this project. The authors also acknowledge Tae-Uk Kim and Arne Verliefde for providing data and details of their research. The authors thank Filmtec (Dow Chemical Co.) for donating the membranes, Dr. Jaeweon Cho of GIST (Korea), Dr. Frank Sacher of TZW (Germany) and Steven Mookhoek of TU Delft (Netherlands) for contributing with analytical results and facilities.

Conclusions – A general QSAR model equation was developed to merge information about interaction of membrane characteristics, filtration operating conditions and solute properties to predict rejections of emerging contaminants during nanofiltration. – The QSAR model identified that the most important variables that influence rejection of organic solutes were log D, salt rejection, equivalent width, depth and length. – Rejection increased by size/steric hindrance effects, solute hydrophobicity decreased rejection due to adsorption and partitioning mechanisms. – Salt rejection incorporated steric hindrance and electrostatic repulsion effects that were related to the membrane structure and operating conditions. – The use of MWCO was acceptable for modelling purposes; however NF membranes with a broad range of MWCO (pore size and distribution) may difficult estimation of rejection of contaminants, thus magnesium sulphate salt rejection may be more appropriate.

Appendix. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.watres.2009.06.054

references

Abdi, H., 2003. Partial least squares (PLS) regression. In: LewisBeck, M., Bryman, A., Futing, T. (Eds.), Encyclopaedia of Social Sciences Research Methods. Sage, Thousand Oaks, CA. Bellona, C., Drewes, J.E., 2005. The role of membrane surface charge and solute physico-chemical properties in the rejection of organic acids by NF. Journal of Membrane Science 249 (1–2), 227–234. Campbell, P., Srinivasan, R., Knoell, T., Phipps, D., Ishida, K., Safarik, J., Cormack, T., Ridgway, H., 1999. Quantitative structure-activity relationship (QSAR) analysis of surfactants influencing attachment of a Mycobacterium sp. to cellulose


acetate and aromatic polyamide reverse osmosis membranes. Biotechnology and Bioengineering 64 (5), 527–544. Castiglioni, S., Bagnati, R., Fanelli, R., Pomati, F., Calamari, D., Zuccato, E., 2006. Removal of pharmaceuticals in sewage treatment plants in Italy. Environmental Science & Technology 40 (1), 357–363. Connell, D.W., 1990. Bioaccumulation of Xenobiotic Compounds. CRC Press, Boca Raton, FL, pp. 75-110. Cornelissen, E.R., Verdouw, J., Gijsbertsen-Abrahamse, A.J., Hofman, J.A.M.H., 2005. A nanofiltration retention model for trace contaminants in drinking water sources. Desalination 178 (1–3), 179–192. Drewes, J.E., Amy, G., Kim, T.-U., Xu, P., Bellona, C., Oedekoven, M., Macalady, D., 2006. Rejection of WastewaterDerived Micropollutants in High-Pressure Membrane Applications Leading to Indirect Potable Reuse Effect of Membranes and Micropollutant Properties. WateReuse, Alexandria, VA. Eriksson, L., Jaworska, J., Worth, A.P., Cronin, M.T.D., McDowell, R.M., Gramatica, P., 2003. Methods for reliability and uncertainty assessment and for applicability evaluations of classification and regression-based QSARs. Environmental Health Perspectives 111 (10), 1361–1375. Escher, B.I., Bramaz, N., Eggen, R.I.L., Richter, M., 2005. In vitro assessment of modes of toxic action of pharmaceuticals in aquatic life. Environmental Science & Technology 39 (9), 3090–3100. Everitt, B.S., Dunn, G., 2001. Applied Multivariate Data Analysis. Arnold, London. Fujikawa, M., Nakao, K., Shimizub, R., Akamatsu, M., 2007. QSAR study on permeability of hydrophobic compounds with artificial membranes. Bioorganic & Medicinal Chemistry 15 (11), 3753–3767. Gramatica, P., 2007. Principles of QSAR models validation: internal and external. QSAR & Combinatorial Science 26 (5), 694–701. Heberer, T., 2002. Occurrence, fate, and removal of pharmaceutical residues in the aquatic environment: a review of recent research data. Toxicology Letters 131 (1–2), 5–17. Henderson, R.K., Baker, A., Parsons, S.A., Jefferson, B., 2008. Characterisation of algogenic organic matter extracted from cyanobacteria, green algae and diatoms. Water Research 42 (13), 3435–3445. Hoek, E.M.V., Elimelech, M., 2003. Cake-enhanced concentration polarization: a new fouling mechanism for salt-rejecting membranes. Environmental Science & Technology 37 (24), 5581–5588. Johnson, R.A., Wichern, D.W., 2007. Applied Multivariate Statistical Analysis. Pearson Prentice Hall, Upper Saddle River, NJ. Jolliffe, I.T., 2002. Principal Component Analysis. Springer-Verlag, New York. Jørgensen, B., Goegebeur, Y., 2006. Multivariate Data Analysis and Chemometrics. Department of Statistics, University of Southern Denmark. Kim, T.-U., Drewes, J.E., Summer, R.S., Amy, G., 2007. Solute transport model for trace organic neutral and charged compounds through nanofiltration and reverse osmosis membranes. Water Research 41 (17), 3977–3988. Kimura, K., Amy, G., Drewes, J., Heberer, T., Kim, T., Watanabe, Y., 2003. Rejection of organic micropollutants (disinfection byproducts, endocrine disrupting compounds, and pharmaceutically active compounds) by NF/RO membranes. Journal of Membrane Science 227 (1–2), 113–121. Kimura, K., Toshima, S., Amy, G., Watanabe, Y., 2004. Rejection of neutral endocrine disrupting compounds (EDCs) and pharmaceutical active compounds (PhACs) by RO membranes. Journal of Membrane Science 245 (1–2), 71–78.

383

Kiso, Y., Sugiura, T., Kitao, T., Nishimura, K., 2001. Effects of hydrophobicity and molecular size on rejection of aromatic pesticides with nanofiltration membranes. Journal of Membrane Science 192 (1–2), 1–10. Kolpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M., Zaugg, S.D., Barber, L.B., Buxton, H.T., 2002. Pharmaceuticals, hormones, and other organic wastewater contaminants in U.S. Streams, 1999–2000: a National Reconnaissance. Environmental Science & Technology 36 (6), 1202–1211. Landau, S., Everitt, B., 2004. A Handbook of Statistical Analysis using SPSS. Chapman & Hall/CRC, Boca Raton, FL. Lee, N., Amy, G., Croue, J.-P., Buisson, H., 2004. Identification and understanding of fouling in low-pressure membrane (MF/UF) filtration by natural organic matter (NOM). Water Research 38 (20), 4511–4523. Libotean, D., Giralt, J., Rallo, R., Cohen, Y., Giralt, F., Ridgway, H., Rodriguez, G., Phipps, D., 2008. Organic compounds passage through RO membranes. Journal of Membrane Science 313 (1-2), 23–43. Nghiem, L.D., Scha¨fer, A.I., Elimelech, M., 2004. Removal of natural hormones by nanofiltration membranes: measurement, modelling, and mechanisms. Environmental Science & Technology 38 (6), 1888–1896. Nghiem, L.D., Scha¨fer, A.I., Elimelech, M., 2005. Pharmaceutical retention mechanisms by nanofiltration membranes. Environmental Science & Technology 39 (19), 7698–7705. Ozaki, H., Li, H., 2002. Rejection of organic compounds by ultralow pressure reverse osmosis membrane. Water Research 36 (1), 123–130. Pomati, F., Castiglioni, S., Zuccato, E., Fanelli, R., Vigetti, D., Rossetti, C., Calamari, C., 2006. Effects of a complex mixture of therapeutic drugs at environmental levels on human embryonic cells. Environmental Science & Technology 40 (7), 2442–2447. Ren, S., Das, A., Lien, E.J., 1996. QSAR analysis of membrane permeability to organic compounds. Journal of Drug Targeting 4 (2), 103–107. Sacher, F., Lange, F.T., Brauch, H.-J., Blankenhorn, I., 2001. Pharmaceuticals in groundwaters: analytical methods and results of a monitoring program in Baden-Wu¨rttemberg, Germany. Journal of Chromatography A 938 (1–2), 199–210. Sacher, F., Ehmann, M., Gabriel, S., Graf, C., Brauch, H.-J., 2008. Pharmaceutical residues in the river Rhinedresults of a onedecade monitoring programme. Journal of Environmental Monitoring 10 (5), 664–670. Santos, J.L.C., de Beukelaar, P., Vankelecom, I.F.J., Velizarov, S., Crespo, J.G., 2006. Effect of solute geometry and orientation on the rejection of uncharged compounds by nanofiltration. Separation and Purification Technology 50 (1), 122–131. Sawyer, C.N., McCarty, P.L., Parkin, G.F., 2003. Chemistry for Environmental Engineering and Science. McGraw-Hill, New York, pp. 479-520. Scha¨fer, A.I., Nghiem, L.D., Waite, T.D., 2003. Removal of the natural hormone estrone from aqueous solutions using nanofiltration and reverse osmosis. Environmental Science & Technology 37 (1), 182–188. Shim, Y., Lee, H.-J., Lee, S., Moon, S.-H., Cho, J., 2002. Effects of natural organic matter and ionic species on membrane surface charge. Environmental Science & Technology 36 (17), 3864–3871. Van der Bruggen, B., Vandecasteele, C., 2002. Modeling of the retention of uncharged molecules with nanofiltration. Water Research 36 (5), 1360–1368. Van der Bruggen, B., Schaep, J., Wilms, D., Vandecasteele, C., 1999. Influence of molecular size, polarity and charge on the retention of organic molecules by nanofiltration. Journal of Membrane Science 156 (1), 29–41. Verliefde, A., Cornelissen, E., Amy, G., Van der Bruggen, B., van Dijk, H., 2007. Priority organic micropollutants in water

384


sources in Flanders and the Netherlands and assessment of removal possibilities with nanofiltration. Environmental Pollution 146 (1), 281–289. Verliefde, A.R.D., Cornelissen, E.R., Heijman, S.G.J., Verberk, J.Q.J.C., Amy, G.L., Van der Bruggen, B., van Dijk, J.C., 2008. The role of electrostatic interactions on the rejection of organic solutes in aqueous solutions with nanofiltration. Journal of Membrane Science 322 (1), 52–66. Vosges, M., Braguer, J.-C., Combarnous, Y., 2008. Long-term exposure of male rats to low-dose ethinylestradiol (EE2) in drinking water: effects on ponderal growth and on litter size of their progeny. Reproductive Toxicology 25 (2), 161–168.

Xu, P., Drewes, J.E., Bellona, C., Amy, G., Kim, T.-U., Adam, M., Heberer, T., 2005. Rejection of emerging organic micropollutants in nanofiltration-reverse osmosis membrane applications. Water Environment Research 77 (1), 40–48. Yangali-Quintanilla, V., Kim, T.-U., Kennedy, M., Amy, G., 2008. Modeling of RO/NF membrane rejections of PhACs and organic compounds: a statistical analysis. Drinking Water Engineering and Science 1 (1), 7–15. Yangali-Quintanilla, V., Sadmani, A., McConville, M., Kennedy, M., Amy, G., 2009. Rejection of pharmaceutically active compounds and endocrine disrupting compounds by clean and fouled nanofiltration membranes. Water Research 43 (9), 2349–2362.




Harmful algae and their potential impacts on desalination operations off southern California David A. Carona,*, Marie-E`ve Garneaua, Erica Seuberta, Meredith D.A. Howarda,b, Lindsay Darjanya, Astrid Schnetzera, Ivona Cetinicá, Gerry Filteauc, Phil Laurid, Burton Jonesa, Shane Trusselle a

Department of Biological Sciences, University of Southern California, 3616 Trousdale Parkway, Los Angeles, CA 90089-0371, USA Southern California Coastal Water Research Project, 3535 Harbor Blvd., Suite 110, Costa Mesa, CA 92626, USA c Separation Processes, Inc., 3156 Lionshead Avenue, Suite 2, Carlsbad, CA 92010, USA d West Basin Municipal Water District, 17140 Avalon Blvd., Suite 210, Carson, CA 90746, USA e Trussell Technologies, Inc., 6540 Lusk Boulevard, Suite C175, San Diego, CA 92121, USA b

article info

abstract

Article history:

Seawater desalination by reverse osmosis (RO) is a reliable method for augmenting

Received 2 April 2009

drinking water supplies. In recent years, the number and size of these water projects have


increased dramatically. As freshwater resources become limited due to global climate

12 June 2009

change, rising demand, and exhausted local water supplies, seawater desalination will play

Accepted 23 June 2009

an important role in the world’s future water supply, reaching far beyond its deep roots in

Available online 30 June 2009

the Middle East. Emerging contaminants have been widely discussed with respect to wastewater and freshwater sources, but also must be considered for seawater desalination

Keywords:

facilities to ensure the long-term safety and suitability of this emerging water supply.

Harmful algal blooms

Harmful algal blooms, frequently referred to as ‘red tides’ due to their vibrant colors, are

Desalination

a concern for desalination plants due to the high biomass of microalgae present in ocean

Red tides

waters during these events, and a variety of substances that some of these algae produce.

Phytoplankton

These compounds range from noxious substances to powerful neurotoxins that constitute

Phytotoxins

significant public health risks if they are not effectively and completely removed by the RO membranes. Algal blooms can cause significant operational issues that result in increased chemical consumption, increased membrane fouling rates, and in extreme cases, a plant to be taken off-line. Early algal bloom detection by desalination facilities is essential so that operational adjustments can be made to ensure that production capacity remains unaffected. This review identifies the toxic substances, their known producers, and our present state of knowledge regarding the causes of toxic episodes, with a special focus on the Southern California Bight. ª 2009 Elsevier Ltd. All rights reserved.

* Corresponding author. Tel.: þ1 213 740 0203; fax: þ1 213 740 8123. E-mail address: [email protected] (D.A. Caron). 0043-1354/$ – see front matter ª 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2009.06.051

386

1.


Introduction

1.1. General overview of harmful algal blooms: a growing global concern Microscopic algae constitute an essential component of all aquatic food webs. Photosynthetic production of organic material by this diverse group of species comprises the primary source of nutrition for all heterotrophic forms of life in much of the world’s ocean and freshwater ecosystems. Microalgae can reach high abundances in the plankton during periods of optimal growth and reduced grazing pressure by herbivores. Such localized mass proliferations are known as algal (or phytoplankton) blooms. In addition, a small proportion of microalgal species are capable of producing a number of noxious or toxic compounds that cause a variety of adverse effects on ecosystem structure and function. These substances pose the potential for ecosystem damage, food web disruption and marine animal mortality, and present a significant human health risk through the consumption of contaminated seafood and, in at least one case, direct exposure to water or aerosols containing these toxic compounds. Additionally, the algal biomass and the associated organic load cause significant desalination operational issues, impacting the pretreatment system and possibly forcing the treatment plant to be taken off-line (Petry et al., 2007). Countless human deaths resulting from the consumption of seafood contaminated with algal toxins have been avoided through rigorous monitoring programs, but sea life has not been so fortunate. Approximately one half of all unusual marine mammal mortality incidents are now attributable to the ingestion of food or prey contaminated by harmful algal blooms (Ramsdell et al., 2005). Losses in revenue due to the direct contamination of seafood products and indirect effects on tourism and other uses of coastal areas have been estimated in the tens of millions of dollars annually in the U.S. states along the Pacific coast (Trainer et al., 2002). There is now convincing evidence that harmful algal bloom (HAB) events are increasing at local, regional and global scales worldwide (Smayda, 1990; Hallegraeff, 1993, 2003; Anderson et al., 2002; Glibert et al., 2005a) and along the North American west coast in particular (Horner et al., 1997; Trainer et al., 2003). This increased occurrence may be due in part to better detection of HAB episodes in recent years or the global dispersal of toxic algal species via the transport of resting spores in ships’ ballast waters (Hallegraeff and Bolch, 1992; Burkholder et al., 2007), but another very likely cause is the increasing impact of anthropogenic activities on coastal ecosystems (Smayda, 1990; Anderson et al., 2002; Glibert et al., 2005b, 2006; Howard et al., 2007; Cochlan et al., 2008; Kudela et al., 2008a). Recent reports reveal extensive and, in some cases, newly emerging occurrences of HABs along the coasts of the U.S. (Fig. 1). These incidents engender a variety of noxious impacts on ecosystems and public health, including direct effects on organisms due to the production of acutely toxic substances, and indirect effects such as reduced availability of dissolved oxygen in the water column resulting from the decomposition of the extensive amounts of organic substances usually produced during such blooms. The

Fig. 1 – Distribution of some well-known regional HAB issues along U.S. shores, including (a) Alaska and (b) Hawaii. Causes and impacts of these poisoning events are defined in Tables 1–3. Summarized from information presented on the Harmful Algae webpage (http://www. whoi.edu/redtide/).

dramatic increases in biomass and organic load that accompany these events pose a significant threat to seawater desalination facilities (Gaid and Treal, 2007).

1.2.

Regional HAB issues along U.S. coastlines

Harmful algae are present throughout U.S. coastal waters, but not all species are of equal concern in all regions (Fig. 1). For example, toxic species of the dinoflagellate genus Alexandrium are common over vast stretches of the U.S. coastline, but coastal regions of the northeastern and northwestern U.S. appear to experience particularly high rates of occurrence of toxic ‘red tides’ caused by these species. The neurotoxins produced by Alexandrium, called saxitoxins, cause paralytic shellfish poisoning (PSP) in humans when ingested through contaminated seafood (particularly filter-feeding shellfish). Similarly, several toxic species of the diatom genus Pseudonitzschia occur along the entire U.S. coastline but significant concentrations of the neurotoxin, domoic acid, produced by these species have historically constituted a health threat primarily in the northeastern and northwestern U.S. (Bates et al., 1989) where it has been documented as the cause of amnesic shellfish poisoning (ASP) in humans. However, high concentrations of domoic acid in the plankton and in diverse planktivorous organisms have been recently documented along the entire Pacific coast of the U.S. (Scholin et al., 2000; Trainer et al., 2002; Schnetzer et al., 2007), as well as in the Gulf of Mexico (Pan et al., 1998). Domoic acid has been attributed to numerous marine animal mortalities along the U.S. west coast. In the Gulf of Mexico, primarily along the west coast of Florida, extensive and recurrent blooms of the dinoflagellate Karenia brevis produce a suite of toxins, known as brevetoxins, that can be aerosolized by breaking waves and


induce neurotoxic shellfish poisoning (NSP) in people inhaling the aerosols (see review of Kirkpatrick et al., 2004). The Tampa Bay seawater desalination facility is the only operating seawater desalination treatment plant of significant size in the United States. It is located along the west coast of Florida and is likely to encounter algal blooms that contain brevetoxin. Less toxic blooms also take place with regional specificity. The pelagophyte Aureococcus anophagefferens causes ‘brown tides’ in coastal waters of Rhode Island, near Long Island (NY) and southward along the mid-Atlantic coast of the U.S. since 1985. No specific toxins have been identified from A. anophagefferens, and no human fatalities have been directly attributed to these blooms. Nevertheless, this species appears to be unpalatable or inhibitory to many filter-feeding mollusks and has caused substantial mortality among these populations, including commercially valuable species (Bricelj and Lonsdale, 1997). Other microalgal species can disrupt food webs or cause reductions in water quality without producing acutely toxic conditions. Among these are the ‘colorful’ red tides of the dinoflagellate Lingulodinium polyedrum, a yessotoxin producer, that have occurred periodically throughout several decades along the south and central Californian coasts (Horner et al., 1997; Gregorio and Pieper, 2000). These blooms have so far been found to be relatively innocuous in these waters but massive accumulations of these cells could have significant impact on desalination plants because of increased turbidity, high suspended solids and organic loading of influent water. Furthermore, accumulations of cells in protected harbors can cause fish mortality by depleting oxygen dissolved in the water, further challenging influent screening and pretreatment systems at desalination plants. Other taxa, such as species of the prymnesiophyte genus Phaeocystis, produce substances that can lead to enormous buildups of sea foam along coasts (Armonies, 1989).

1.3.

Desalination, plankton and water quality issues

Large research programs have developed within different geographic areas throughout the U.S. to address regional HAB issues. These programs are designed to study the species, toxins and environmental causes of HAB outbreaks. These efforts, as well as local, county, state and federal monitoring programs provide basic information for marine resource use and have focused almost exclusively on threats to human health via the consumption of contaminated seafood. Unfortunately, few if any of these programs provide sufficient information on appropriate temporal and spatial resolution for thoroughly assessing the potential impact of HAB events on reverse osmosis desalination operations. Moreover, toxin analyses have primarily examined the presence of these substances in particulate material (plankton or animal tissue, particularly shellfish and finfish), and therefore may be poor predictors for the amount of toxins that might occur in seawater in the dissolved state during algal blooms, which would be most likely to be loaded onto reverse osmosis membranes during desalination. There are two potential impacts that HABs may have on seawater desalination facilities: (1) algal toxins in ocean water

387

pose a significant treatment challenge for the reverse osmosis system to ensure that these molecules are effectively removed and (2) increased turbidity, total suspended solids and total organic content resulting from algal biomass and growth challenge the entire desalination facility’s treatment train. The significance of these issues will depend on the specific algae forming a bloom and the toxin(s) or other substances that they produce, the magnitude and duration of the bloom, and the specific desalination process conducted. For example, multi-effect distillation and multi-flash distillation might be susceptible to (2) but would be much less affected by toxins in the water (1). Desalination using reverse osmosis presumably would be vulnerable to both issues. Therefore, for the latter desalination approach, a thorough understanding of HAB episodes in terms of incidence and seasonality, vertical and horizontal spatial distribution, as well as biological aspects such as algal composition within a geographical region could help optimize the design and operational efficiency of desalination plants employing reverse osmosis. This paper provides an overview of HABs occurring along the continental U.S. coastline with special emphasis on the southwestern U.S., and provides some insight on the potential impacts that these events may have on the seawater desalination process. In recent years, this geographical area has become a focal point of discussions regarding desalination (Cooley et al., 2006) because of its sizable population and the particularly tenuous nature of the water supply to this region. Although numerous issues involving the desalination process are now being examined (Separation Processes Inc., 2005; Gaid and Treal, 2007; Pankratz, 2008, 2009), very limited information exists on the risks that algal blooms pose to seawater desalination facilities. A review of the major species producing harmful blooms, the substances they produce, and information on the spatial and temporal distributions of blooms are presented along with some conclusions on their potential impacts. This paper also provides some general guidelines on how early detection may help prevent or minimize the impact of HABs on a desalination facility’s production capacity or its water quality.

2. Toxin producers and toxin concentrations of the west coast A variety of toxins including several powerful neurotoxins are produced by microalgae, and a number of these toxins and potentially toxic algal species have been detected on the U.S. west coast (Table 1). The ability to rapidly detect and quantify toxic algae in natural water samples is problematic at this time. Many of these species are difficult to identify using light microscopy. For this reason, new genetic and immunological methods for species identification and enumeration have been appearing rapidly in the literature (Miller and Scholin, 1998; Bowers et al., 2000, 2006; Coyne et al., 2001; Caron et al., 2003; Galluzzi et al., 2004; Anderson et al., 2005; Mikulski et al., 2005, 2008; Ahn et al., 2006; Handy et al., 2006; Moorthi et al., 2006; Iwataki et al., 2007, 2008; Demir et al., 2008; Matsuoka et al., 2008). Moreover, many toxin-producing algal species exhibit variable toxin production in response to environmental conditions, and among different strains of the same species

388


Table 1 – Planktonic species occurring along the west coast of the U.S. that are potential concerns for reverse osmosis operations. Microalgae

Toxin(s)

Poisoning Event

References

Amnesic Shellfish Poisoning (ASP)

Subba Rao et al. (1988), Bates et al. (1989), Martin et al. (1990), Buck et al. (1992), Garrison et al. (1992), Rhodes et al. (1996), Horner et al. (1997), Lundholm et al. (1997), Rhodes et al. (1998), Trainer et al. (2000, 2001), Baugh et al. (2006)

Diatoms Pseudo-nitzschia spp. P. australisb P. cuspidatab P. delicatissimab P. fraudulentab P. multiseriesb P. pungensb P. pseudodelicatissimab P. seriataa

Domoic acid (DA)

Dinoflagellates Alexandrium spp. A. acatenellaa A. catenellab A. fundyensea A. hiranoia A. ostenfeldiia A. tamarensea

Saxitoxins (STXs)

Dinoflagellates Lingulodinium polyedrumb Gonyaulax spiniferaa Protoceratium reticulatuma,c

Yessotoxins (YTXs)

Human and ecosystem effects None reported

Holmes et al. (1967), Satake et al. (1997, 1999), Draisci et al. (1999a), Paz et al. (2004, 2007), Armstrong and Kudela (2006), Rhodes et al. (2006), Howard et al. (2007)

Dinoflagellates Dinophysis spp. D. acuminataa D. acutaa D. caudate D. fortiia D. norvegicaa D. rotundataa D. triposa

Okadaic acid (OA) Dinophysistoxins (DTXs) Pectenotoxins (PTXs)

Diarrhetic Shellfish Poisoning (DSP)

Holmes et al. (1967), Yasumoto et al. (1980), Murata et al. (1982), Yasumoto et al., (1985), Cembella (1989), Lee et al. (1989), Horner et al. (1997), Cembella (2003), Miles et al. (2004), Shipe et al. (2008), Sutherland (2008)

Human effects Gastro-intestinal symptoms Neurologic symptoms Death Ecosystem effects Marine mammal mortalities Bird mortalities

Paralytic Shellfish Poisoning (PSP) Human effects Gastro-intestinal symptoms Paralysis Death Ecosystem effects Marine mammal mortalities

Human effects Gastro-intestinal symptoms Ecosystem effects None reported

Sommer and Meyer (1937), Gaines and Taylor (1985), Steidinger (1993), Scholin et al. (1994), Taylor and Horner (1994), Jester (2008)

Prorocentrum spp. P. micans P. minimuma,d Raphidophytes Chattonella marinaa Fibrocapsa japonicaa Heterosigma akashiwoa

a b c d

Brevetoxins (PbTxs)

Neurotoxic Shellfish Poisonining (NSP) Human effects Gastroenteritis Neurologic symptoms Respiratory irritation and/or failure Ecosystem effects Marine mammal mortalities Fish mortality events

Loeblich and Fine (1977), Hershberger et al. (1997), Gregorio and Connell (2000), Hard et al. (2000), Tyrell et al. (2002), O’Halloran et al. (2006)

Reported to produce toxin. Reported to produce toxin on the west coast of the United States. Conflicting reports on toxicity of P. reticulatum cultures isolated from California, Washington and Florida. Reported to be present on the west coast of Mexico.

even when isolated from the same geographic region (Smith et al., 2001; Trainer et al., 2001; Kudela et al., 2004). Laboratory experiments have revealed a wide range of physico-chemical factors that increase or decrease toxin

production by harmful species of algae, and which appear to be species-specific (see review of Grane´li and Flynn, 2006). Reports of factors inducing toxin production have sometimes been conflicting, presumably indicating that multiple factors, or


perhaps generally stressful conditions, may stimulate toxin production. Factors affecting toxin production include: (1) temperature (Ono et al., 2000); (2) light intensity (Ono et al., 2000); (3) salinity (Haque and Onoue, 2002a,b); (4) trace metal availability, especially iron (Ladizinsky and Smith, 2000; Rue and Bruland, 2001; Maldonado et al., 2002; Wells et al., 2005; Sunda, 2006) but also copper (Maldonado et al., 2002) and selenium (Mitrovic et al., 2004, 2005); (5) macronutrient availability including silicate (Pan et al., 1996b; Fehling et al., 2004; Kudela et al., 2004), phosphate (Pan et al., 1996a, 1998; Fehling et al., 2004), nitrogen (Bates et al., 1991; Pan et al., 1998; Kudela et al., 2004) and combinations of nutrient limitation (Anderson et al., 1990; Flynn et al., 1994; John and Flynn, 2000); (6) cellular elemental ratios of nutrients and physiological stress (Grane´li and Flynn, 2006; Schnetzer et al., 2007); (7) growth phase (Anderson et al., 1990; Bates et al., 1991; Flynn et al., 1994; Johansson et al., 1996; Maldonado et al., 2002; Mitrovic et al., 2004). The precise combination(s) of environmental factors that select for population growth of particular algal species within diverse natural assemblages, and the specific conditions that induce toxin production, are poorly understood for most harmful algae. This present state of knowledge makes it difficult to predict the timing, duration or spatial extent of the vast majority of HAB events and the toxic events resulting from them. Our ability to thoroughly characterize HABs is also complicated by the complex array of toxins produced by algae. Marine algal species produce a suite of toxic components (Yasumoto and Murata, 1993), and unidentified toxins undoubtedly remain to be described. Additionally, most toxins are actually composed of families of closely related compounds. Slightly different forms of a toxin can exhibit very different levels of toxicity, or may be characterized differently by some detection methods and analytical approaches (Garthwaite et al., 2001; Lefebvre et al., 2008). Such complexity and variability can sometimes yield vague or contradictory conclusions regarding the exact source of toxicity in a natural sample (Bates et al., 1978; Paz et al., 2004, 2007). Finally, characterization of HAB events is complicated by inherent difficulties associated with linking specific toxins measured in natural water samples to a specific algal species in a complex, natural phytoplankton assemblage and, as noted above, the presence of toxic species in a water sample does not necessarily indicate the presence of toxins. Despite these shortcomings, there is considerable knowledge of many of the major algal toxins and their producers in U.S. coastal waters that constitute the most important potential concerns for desalination activities because they are the most likely to be encountered in ocean water intakes.

2.1.

Domoic acid

2.1.1.

Toxin description and activity

Domoic acid (Fig. 2; Table 3) is an amino acid derivative belonging to the kainoid class of compounds containing three carboxyl groups and one secondary amino group (Wright et al., 1990; Jeffery et al., 2004). All four groups are charged at neutral pH, and the carboxyl groups become successively protonated as pH decreases, yielding five possible protonated forms of domoic acid (Quilliam, 2003; Jeffery et al., 2004). There

389

are currently ten known isomers of domoic acid, including the isodomoic acids A through H and the domoic acid 5’ diestereomer (Jeffery et al., 2004). Domoic acid and other members of the kainoid class are glutamate analogues that interfere with neurochemical pathways by binding to glutamate receptors of brain neurons (Wright et al., 1990; Quilliam, 2003). The resulting effect of these neuroexcitants, or excitotoxins, is a continuous stimulation of the neurons, which can lead to rupture and/or eventual formation of lesions (Wright et al., 1990). Depolarized neurons result in short-term memory loss (Clayden et al., 2005), which has led to the common name for the illness related to the consumption of seafood contaminated with domoic acid: amnesic shellfish poisoning (ASP). Symptoms of ASP include gastroenteritis (vomiting, diarrhea, abdominal cramps) that can be experienced in humans within 24 h after ingestion, and neurological symptoms of confusion, memory loss, disorientation, seizures, coma and/or cranial nerve palsies that are typically experienced within 48 h (Perl et al., 1990; Wright et al., 1990). The number of human illnesses resulting from domoic acid poisoning has been few (Horner et al., 1997), likely due to active monitoring of fisheries. However, cultured blue mussels (Mytilus edilus) contaminated with domoic acid poisoned 107 people and killed three during the first major documented ASP outbreak in 1987 on Prince Edward Island, Canada (Perl et al., 1990). ASP poses a serious threat to marine wildlife, and the deaths of thousands of marine mammals and sea birds have been attributed to domoic acid intoxication (Bates et al., 1989; Scholin et al., 2000; Gulland et al., 2002; Caron et al., unpublished data). The first documented poisoning episode of marine animals related to domoic acid on the U.S. west coast was attributed to Pseudo-nitzschia australis and occurred in September 1991 off central California (Table 2; Buck et al., 1992; Fritz et al., 1992). High concentrations of domoic acid were also detected in Washington and Oregon in the 1990s (Wekell et al., 1994; Adams et al., 2000; Trainer et al., 2002), and a decade later in coastal waters off southern California (Schnetzer et al., 2007). The frequency and severity of these toxic events appears to be increasing (Trainer et al., 2007).

2.1.2.

Producers

The production of domoic acid and its isomers is confined to approximately a dozen chain-forming marine pennate diatom species within the genus Pseudo-nitzschia (Bates and Trainer, 2006), a genus containing species that form long chains of cells attached at their ends (Fig. 3a and b). The main toxin producing species that have been documented on the U.S. west coast include: P. australis, P. delicatissima, P. fraudulenta, P. multiseries, P. pungens, P. pseudodelicatissima, P. seriata and P. cuspidata (Tables 1 and 2). These species are distinguished based on fine morphological features of their silica frustules (Fig. 3a and b). These distinctions are subtle and require careful electron microscopical analysis and elaborate taxonomic training. As a consequence, historical misidentifications are not unusual and debates regarding some species descriptions are still unresolved. It is surprising that the first reports of ASP on the west coast of the U.S. were not recorded until the 1990s, even though Pseudo-nitzschia species have been recorded in surveys of

390


Fig. 2 – Chemical structures of commonly encountered toxins produced by microalgae in U.S. coastal waters.

phytoplankton species in the Southern California Bight since 1917 (Allen, 1922, 1924, 1928, 1936, 1940, 1941; Reid et al., 1970, 1985; Lange et al., 1994; Fryxell et al., 1997; Thomas et al., 2001). Given that these species generally comprise a significant portion of the total diatom assemblage in these waters, it can be surmised that either toxin production has increased in these west coast species, or that poisoning events prior to the 1990s have occurred but have not been attributed to these diatoms. Historical accounts of ‘unusual animal mortality events’ along the U.S. west coast tend to support the latter hypothesis. There have been increasing numbers of toxic events recorded along the U.S. west coast (Table 2), notably in Puget Sound (Trainer et al., 2003, 2007), Monterey Bay (Vigilant and Silver, 2007; R. Kudela, unpubl. data), Santa Barbara Channel (Trainer et al., 2000; Anderson et al., 2006; Mengelt, 2006), San Pedro Channel (Busse et al., 2006; Schnetzer et al., 2007), Newport Beach (Busse et al., 2006) and San Diego (Lange et al., 1994; Busse et al., 2006). Most recently, toxic blooms of Pseudonitzschia in the Long Beach-Los Angeles Harbor and San Pedro Channel have been particularly toxic, with some of the highest domoic acid concentrations recorded for the U.S. west coast (Caron et al., unpublished data). The increased incidence and severity of these toxic episodes off the western U.S. coast parallels the increase in frequency and intensity of harmful

algal blooms observed globally (Smayda, 1990; Hallegraeff, 1993, 2003; Anderson et al., 2002; Glibert et al., 2005b).

2.2.

Saxitoxins

2.2.1.

Toxin description and activity

Saxitoxin is a complex guanidine-based alkaloid that exists as more than 30 identified analogues in nature (Llewellyn, 2006). It is the most powerful marine toxin currently known and among the most dangerous poisons on Earth, except for some venoms and bacterial toxins (Schantz et al., 1957). Due to its acute toxicity, saxitoxin is currently listed as a chemical weapon in Schedule 1 of the Chemical Weapons Convention (Llewellyn, 2006). Saxitoxins display a rigid tricyclic core (Fig. 2; Table 3) and are very stable in biological and physiological solutions (Rogers and Rapoport, 1980). This nitrogenrich molecule and its chemical relatives are polar and have a positive charge at pH 7.7 (Shimizu et al., 1981). Consequently, they are soluble in water and alcohols, and insoluble in organic solvents (Schantz et al., 1957). Saxitoxins are known to disrupt the flow of ions through voltage gate sodium channels (Catterall, 1992; Cestele and Catterall, 2000). It has also been recently discovered that they have the ability to bind to calcium (Su et al., 2004) and

Table 2 – Distribution and concentrations of marine toxins in plankton of confirmed toxin producers in U.S. west coast waters. Toxin(s) Domoic acid

Location and year

Causative species

Particulate mg L1 (nmol L1)

Cellular pg cell1 b.d.–4.6

Washington coast and Juan de Fuca Eddy, WA (1997, 1998)

P. pseudodelicatissima Pseudo-nitzschia spp.

b.d.–2.7 3.6–8.7

Penn Cove, WA (1997)

P. pungens P. multiseries P. australis P. pseudodelicatissima

b.d.–0.8

Washington coast, WA (2001)

P. australis

b.d.–0.03

Pseudo-nitzschia spp.

(0.4–15)

Puget Sound, WA (2005)

P. pseudodelicatissima Pseudo-nitzschia spp.

b.d.–14

Central Oregon coast, OR (1998)

P. australis

Pt. Anõ Nuevo, San Francisco, CA (1998)

References Adams et al. (2000), Trainer et al. (2001, 2002) Trainer et al. (1998)

Marchetti et al. (2004) 4

2 10 –0.3 0.1–94.4

a

(b.d.–4.3) (1–5)

a

Baugh et al. (2006)

Trainer et al. (2007)

0.5

35


P. pungens P. multiseries

0.1–0.7

0.3–6.3


Bolinas Bay, San Francisco, CA (2003)

P. australis

0.15–9.4

Monterey Bay, CA (1991, 1998)

P. australis

b.d.–12.3 0.1–6.7

3–37 7.2–75

Monterey Bay, CA (1998)

P. pseudodelicatissima P. multiseries

0.1–0.4 0.67

0.8–1.2 6

Monterey Bay, CA (2000)

Pseudo-nitzschia spp. P. australis

Monterey Bay, CA (2002–2003)

Pseudo-nitzschia spp.

Morro Bay, CA (1998)

P. australis

1.3–7.4

37–78

Trainer et al. (2000, 2001)

San Luis Obispo, CA (2003–2005)

P. australis P. multiseries

1.5–7.6

9–38

Mengelt (2006)

Point Conception, CA (1998)

P. australis

2.2–6.3

15–22


Santa Barbara, CA (1998)

P. australis P. pungens P. pseudodelicatissima

0.5–1.2

0.1–0.9


Howard et al. (2007)

b.d.–24

24

Buck et al. (1992), Garrison et al. (1992), Walz et al. (1994), Scholin et al. (2000) Trainer et al. (2000, 2001) b.d.–8491


Washington coast, WA (2003)

Dissolved pg mL1 (nmol L1)

Bargu et al. (2002, 2008)

Vigilant and Silver (2007)

391


392

Table 2 (continued ) Toxin(s)

Saxitoxins

Location and year

Causative species

Particulate mg L1 (nmol L1)

Cellular pg cell1

0.03–1.7

0.14–2.1

Anderson et al. (2006)

6–12

b.d.–80

Mengelt (2006)

b.d.–117

Schnetzer et al. (2007)

P. australis

Santa Barbara (Santa Rosa Island and north San Miguel) (2004)


Southern California Bight, CA (2003, 2004)

Pseudo-nitzschia spp. P. australis P. cuspidata

5.6–12.7

San Diego and Orange counties, CA (2004)


b.d.–2.33

Sequim Bay, WA (2004–2007)

Alexandrium spp.

0.02–0.5

Oregon coast, OR (2004)

Alexandrium spp.

0.004–0.028

Humboldt Bay, CA (2004)

A. catenella

San Mateo County coast, CA (2004) Monterey Bay, CA (2004)

A. catenella

Monterey Bay, CA (2003–2005)

A. catenella

Morro Bay, CA (2004)

A. catenella

Yessotoxin

La Jolla, CA (1993)

Lingulodinium polyedrum

Brevetoxins

Indian Inlet, Bald Eagle Creek and Torque Canal, DE (2000)

Chattonella cf. verruculosa

b.d.: Below detection limit. a Toxin concentration from cells in culture.

Busse et al. (2006)

150–800

Jester (2008)

2.1–62.6

a

Jester (2008)

0.6–31.3

a

Jester (2008)

1.4–16.6

a

b.d.–0.962

0.008–90%), and purchased from Sigma–Aldrich (Steinheim, Germany). Isotopically labeled compounds, used as internal standards, were 13C-phenacetin obtained from Sigma– Aldrich, and 3D-mecoprop from Dr. Ehrenstorfer (Augsburg, Germany). HPLC grade methanol, acetone, dichloromethane, hexane, as well as formic acid were provided by Dikma (USA), and ultra-pure water was produced by a Milli-Q unit (Millipore, USA). Stock solutions of individual compound were prepared in methanol and mixture standards with different concentrations were prepared by diluting the stock solutions before each analytical run. All the solutions were stored at 4 C in the dark.

2.2.

Sample collection

Four full-scale municipal WWTPs, referred as A, B, C and D, were selected in our study. These WWTPs employ similar conventional treatment processes: primary treatment to remove particles coupled with secondary biological treatment. For the secondary biological treatment processes, WWTPs A and D employ anaerobic/anoxic/oxic (A2/O) activated sludge process, anoxic/oxic (A/O) activated sludge process is adopted in WWTP B, and WWTP C employs oxidation ditch (OD). Other detailed information on each WWTP, such as inhabitants served, daily flow, HRT and SRT are shown in Table 1. Part of the secondary effluents was further treated in WWTPs A, B and D, by the processes of ultrafiltration (UF)/ozone, sand filtration (SF) and microfiltration/reverse osmosis (MF/RO), respectively. In WWTP A, a dead-end ultrafiltration system (Zenon GE) is used. The whole system has 6 trains of Zee-Weed 1000 membrane. Each train contains 9 cassettes of 57–60 modules per cassette. The membrane, with the pore size of 0.02 mm, is made by PVDF. The module is operated in an outside/in configuration at a constant flow of 23 L (m2 h)1 and the total treatment capacity reaches 80,000 m3 d1. The membrane is hydraulically backwashed at a constant flow rate of 34 (m2 h)1, and 29 times per day. The backwash phase lasts for 1 min. Maintenance cleaning is conducted once per day. Membranes are soaked in the sodium hypochlorite solution (50 mg L1) for 25 min. For the ozonation process, gaseous ozone is generated from an ozone generator (Mitsubishi Electric). The ozone dosage and contact time in the reaction tank is 5 mg L1 and 15 min, respectively. The pH of the wastewater before ozonation ranges 6.5–8.0 and shows no significant change after ozonation. As the heart of the advanced treatment in WWTP D, a spiral-wound crossflow module is employed for the reverse osmosis (RO) membrane filtration. The RO membrane (Filmtec, DOW) is made from a thin-film composite polyamide material. Each module is designed to operate at a water flux of 1.3 m3 h1, and a product water recovery of 75–80%. The trans-membrane pressure is between 0.04 and 0.06 MPa, and the salt rejection remained at the level of 99%. Every 3–6 months, normally when the transmembrane pressure reaches above 0.06 MPa, the membrane is cleaned with 0.1%(w) sodium hydroxide solution (for organic foulants), 2%(w) citric acid (for inorganic foulants) and 0.5%(w) formaldehyde (as biocide). Schematic diagram of treatment processes in the four WWTPs is shown in Fig. 1.

419


Table 1 – Information of the WWTPs investigated. WWTP A B C D

Inhabitants served 103

Daily flow (103 m3)

HRT (h)

SRT (d)

Secondary treatment

Tertiary treatment

814 2400 480 2415

400 1000 200 600

11 11 15 10.67

12–15 20 12–16 15

A2/O A/O OD A2/O

UF/ozone SF – MF/RO

The samples were collected once from the four WWTPs during June and July 2008, with no compensation for HRT. All of them were collected as grab samples in duplicate (500 mL for influents and 1000 mL for the others) in prewashed amber glass bottles, kept in the cooler and transported to the laboratory. Immediately after delivery to the laboratory, they were filtered through prebaked (400 C, >4 h) glass microfiber filters (GF/F, Whatman) to remove particles and stored at 4 C before extraction.

2.3.

conditioned, wastewater samples, added with internal standards and adjusted to pH ¼ 7, were introduced to the cartridge via a PTFE tube, at a flow rate of 5–10 mL min1. After washing by 5 mL of 5.0% methanol solution, the cartridge was dried under vacuum for 2 h and eluted with 5 mL of methanol. The extract was then concentrated to 0.4 mL under a gentle nitrogen stream and stored at 4 C for analysis. Concentrations of the target compounds were analyzed using ultraperformance liquid chromatography coupled with tandem mass spectrometry (UPLC–MS/MS). Analytes were separated using Waters Acquity UPLC system (Waters Corporation, USA) equipped with Acquity UPLC BEH C18 column (50 2.1 mm, particle size of 1.7 mm), and detected by Quattro Premier XE tandem quadrupole mass spectrometry (Waters Corp., USA) equipped with an electrospray ionization source. The analysis was carried out in multiple reaction monitoring (MRM) mode,

Sample extraction and analysis

The method for the extraction and analysis of pharmaceuticals and consumer products is presented elsewhere (Sui et al., in press) and briefly described here. After the solid-phase extraction (SPE) cartridges (Oasis, HLB, 200 mg, 6 mL) were

a

Ozonation

Ultrafiltration

WWTP A

Tertiary Effluent Influent Grit Removal

A/O treatment

Secondary Clarifier

Secondary Effluent

Screen

b

Sand Filtration

WWTP B

Tertiary Effluent Influent Grit Removal

A2/O treatment

Primary Clarifier

Secondary Clarifier Secondary Effluent

Screen

c

WWTP C

Influent

Grit Removal

Oxidation Ditch

Screen

Secondary Clarifier

Secondary Effluent

MF

d

WWTP D

RO

Tertiary Effluent

`

Influent Grit Removal

Primary Clarifier

A2/O treatment

Secondary Clarifier

Secondary Effluent

Screen

Fig. 1 – Schematic diagram of the treatment processes in the four WWTPs and sampling site location (C).

420


and in general, two precursor ion/product ion transitions were monitored for one compound with the purpose of quantification and confirmation.

concentrations of some target compounds (i.e. caffeine, DEET, carbamazepine) resulted in slightly higher deviations.

2.4.

3.

Result and discussion

3.1.

Influents

Quality control

For each sampling, 500 mL Milli-Q water in an amber glass bottle as a field blank was brought to the WWTPs, exposed to the environment where the samples were taken from, and then delivered back to the laboratory with samples. For each set of samples (normally 10 samples), at least one procedural blank was prepared from ultra-pure water in the laboratory. Both the field blanks and procedural blanks were run identically to the wastewater samples, and the concentrations of target compounds were below the limit of quantification (LOQ). The absolute recoveries, calculated by comparing the concentrations of target compounds in spiked and unspiked wastewaters, were proved to be 73–102% and 50–95% in the effluent and influent for most compounds, respectively. While for several compounds (i.e. sulpiride, gemfibrozil, mefenamic acid), the absolute recoveries were not satisfactory. However, 13 C-phenacetin and 3D-mecoprop, the surrogate standards used for positive and negative ion mode respectively were able to compensate for the loss of most analytes, and relative recoveries were 67–130% for all the analytes in the effluent and 79–140% in the influents except mefenamic acid (251%) and nalidixic acid (178%). Therefore, the concentrations of these two compounds in the wastewater influents were not quantitatively determined and reported. The LOQs were 0.3–5.5 ng L1 and 0.7–20 ng L1 in the effluent and influent, respectively. Detailed information about the calibration, recoveries, LOQ, matrix effects, etc. were described in Sui et al. (in press), and briefly listed in Table 2. As duplicate samples were collected at each sampling site, mean concentrations were adopted. In most cases, deviations of duplicate samples were less than 20%. For some tertiary effluent samples, low

As shown in Fig. 2, 12 target compounds were detected in all the influent samples from the four WWTPs, while ketoprofen was below LOQ in all wastewater samples. The most abundant compounds detected were the consumer products, caffeine (3.4–6.6 mg L1) and N,N-diethyl-meta-toluamide (0.6–1.2 mg L1), probably due to the large consumption of drinks containing caffeine (i.e. coffee, tea, etc.) and wide application of insect repellent during the summer time when we sampled. Diclofenac, trimethoprim, sulpiride, carbamazepine, indometacin and metoprolol showed relatively high concentrations (Fig. 2). A similar composition distribution was observed among all the influents of the four WWTPs. The concentrations of target pharmaceuticals except diclofenac and trimethoprim, were much lower than those reported in the European and North American countries (Thomas and Foster, 2005; Lishman et al., 2006; Vanderford and Snyder, 2006; Santos et al., 2007; Gomez et al., 2007; Vieno et al., 2007b; Huerta-Fontela et al., 2008). For instance, the concentrations of ketoprofen in the wastewater influents were recorded to be 2.0 0.6 mg L1 in Finland (Lindqvist et al., 2005), 200 ng L1 in Australia (Al-Rifai et al., 2007), and 300–1360 ng L1 in Spain (Santos et al., 2007), while in the influents of four WWTPs in Beijing, it could not be detected. Concerning gemfibrozil, which is used to lower cholesterol and triglyceride levels in the blood, the contamination level found in the present study was 24–140 ng L1, even 1 or 2 order of magnitude lower than those in the USA (4770 ng L1, Vanderford and Snyder, 2006) and Canada (418 ng L1,

Table 2 – Instrumental quantification limit (IQL), limit of quantification (LOQ), absolute recovery (AR), relative recovery (RR) and matrix suppression of target compounds. Compounds

BF CA CBZ CF CP DEET DF GF IM KP MA MTP NA SP TP

IQL (pg)

0.5 10 2.5 2.5 2.5 2.5 10 10 2.5 10 10 2.5 2.5 0.5 2.5

LOQ (ng L1) Effluent

Influent

0.3 5.4 1.0 1.7 1.0 1.3 4.7 4.2 1.3 5.5 5.5 1.1 1.1 0.4 1.0

0.7 16 2.8 3.3 2.3 3.2 9.4 20 2.8 18 5.2 3.3 2.1 1.0 2.7

AR (n ¼ 6, %) Effluent

a Value in the brackets refers to the deviation of the recovery.

74 74 100 60 98 76 86 95 80 69 73 88 91 51 102

(3) (4) (4) (3) (7) (3) (12) (12) (12) (9) (11) (3) (9) (5) (7)

a

RR (n ¼ 6, %) a

Influent 58 50 71 61 86 63 85 40 73 44 154 60 95 42 74

(5) (4) (12) (37) (9) (6) (19) (11) (13) (3) (24) (3) (9) (3) (16)

a

Effluent 94 94 130 77 124 99 109 120 101 87 92 115 118 67 132

(4) (6) (6) (4) (10) (5) (16) (16) (16) (12) (15) (5) (12) (7) (9)

Matrix effect (%) a

Influent 94 82 133 114 140 118 139 65 119 72 251 113 178 79 139

(10) (8) (23) (70) (17) (14) (32) (18) (22) (7) (41) (9) (20) (7) (31)

27 30 6 49 39 24 4 28 1 46 13 15 3 63 5

421


a

10000

Concentration (ng L-1)

Influents

1000

100

10

Cpred ¼ SP

CBZ

MTP

IM

DF

GF

CA

BF

TP

CP

CF

DEET

1

10000

Concentration (ng L-1)

Pharmaceuticals & Consumer Products

b

1000

Lishman et al., 2006). The low levels of target pharmaceuticals were probably due to the lower per capita consumption in China than in the countries with higher socioeconomic statuses, where medical care is more prevalent (Thomas and Foster, 2005). The per capita consumption rate of gemfibrozil in China is estimated to be 0.036 mg person1 d1 (Table 3), lower than those in Germany (0.2 mg person1 d1, Ternes, 1998) and Canada (0.2 mg person1 d1, Lishman et al., 2006). Since the levels of target pharmaceuticals were somewhat different from those of European and North American countries, we theoretically calculate the concentration of pharmaceuticals in the wastewater influent by the following equation (Lindqvist et al., 2005; Nakada et al., 2006)

Secondary effluents

100

10

SP

CBZ

MTP

MA

IM

DF

GF

CA

BF

TP

CF

CP

DEET

1

Pharmaceuticals & Consumer Products Fig. 2 – Concentrations of target pharmaceuticals in wastewater influents (a) and secondary effluents (b) of four WWTPs in Beijing.

T e% I 1012 365 P Q

(1)

where Cpred is the predicted concentration of the pharmaceutical in wastewater influent (ng L1); T is the total production of a pharmaceutical both for human and animal use in China per year (ton year1), P is the population of China, e% is the amount of the pharmaceutical excreted unchanged, I is the number of inhabitants served and Q is the influent flow (m3 d1). The predicted concentrations of gemfibrozil, diclofenac, indometacin, ketoprofen, carbamazepine, and sulpiride were comparable to those measured in the influents (Table 3). Much lower measured concentration than predicted concentration of chloramphenicol was probably because it had been forbidden for use in food and aquaculture in China since 2005, and the available data about the production of pharmaceuticals were based on the year of 2004. It should be noticed that since there is no available data on the total consumption of any pharmaceutical, we used figures for total production of individual pharmaceutical instead. Therefore, the differences between the amounts actually produced and applied as well as the amount used in human and veterinary medicine could not be distinguished, which might result in overestimation of the theoretical concentration. Nevertheless, the comparability between the predicted concentrations and measured concentrations illustrates the overall reasonability of the approach.

Table 3 – Outputs, per capita consumption, predicted concentrations (PECs) and measured concentrations (MECs) of some pharmaceuticals in the wastewater influents of WWTPs investigated. Compound

CP TP GF DF IM KP CBZ SP

Outputa (tons year1)

Per capita consumption (mg person1 d1)

Excreted unchangedb (%)

PEC (ng L1)

MEC (mean, ng L1)

1929 2352 17 328 277 92 395 98

4.051 4.939 0.036 0.689 0.582 0.193 0.830 0.206

5–10 45 76 15 10–20 2.7 2–3 15–70

844 6173 76 287 242 15 58 243

31 400 60 318 129 n.d.c 113 157

a From CMEIN (2005). b From Bolton and Null (1981), Ternes (1998), Khan and Ongerth (2004), Niwa et al. (2005), Nakada et al. (2006), Jjemba (2006). c n.d. ¼ Not detected.

422


Compound

MECeff (ng L1)

PNECa (ng L1)

MEC/PNEC

15 140 4.7 13 204 n.d. 79 108

182,000 16,000 6000 42,000 100 306,000 31,000 420

0.0001 0.0087 0.0007 0.0003 2.0440 – 0.0025 0.2574

CF TP BF CA DF KP MTP CBZ

a From Santos et al. (2007), Lindqvist et al. (2005), Grung et al. (2008), Ferrari et al. (2003), Huschek et al. (2004).

the dominant contributor in the wastewater effluent, the risk quotient was higher than 1, implicating a risk to the aquatic environment.

3.3.

Removal efficiency of conventional treatment

The removal efficiency during the primary treatment was low, indicating no significant adsorption of target compounds to the particles removed in this stage (Fig. 4). Most of the pharmaceuticals and consumer products have log Kow values of less than 3.0, so they are not expected to adsorb significantly to the particles. Other pharmaceuticals with higher Kow values, such as gemfibrozil, have much lower pKa values than the pH of wastewater. Therefore, they are dissociated and expected to be >98% in the aqueous phase (Thomas and Foster, 2005), and not bound to the particles. During the secondary treatment, the average removal rate for different compounds ranged from 12% to 100% (Fig. 4). Caffeine, bezafibrate, trimethoprim and DEET were effectively removed, with the average efficiency of 100 0%, 88 12%, 76 24% and 69 21%, respectively. These results were comparable with those found in the previous studies (Ternes, 1998; Okuda et al., 2008; Thomas and Foster, 2005; Castiglioni et al., 2006). Caffeine was proved to be readily biodegradable (Okuda et al., 2008; Thomas and Foster, 2005; Huerta-Fontela

120

Primary treatment Secondary treatment

100

A

40 20

-40 0

20

40

60

80

100

Pharmaceutical Compostion (%) Fig. 3 – Composition profiles of target pharmaceuticals in secondary effluent samples from four WWTPs in Beijing.

-60

Pharmaceutical Fig. 4 – Removal efficiencies of target pharmaceuticals during the conventional treatment.

SP

CBZ

GF

CA

-20

BF

0 TP

B

60

CF

C

80

DEET

Others GF MTP IM CBZ SP TP DF DEET

Removal efficiency (%)

WWTP

D

MTP

Similar to the influent samples, ketoprofen was below the LOQ in all the secondary effluent samples. Nalidixic acid and chloramphenicol were detected only in one WWTP, with the concentration of 8.1 and 19 ng L1, respectively. The mean concentrations of the other 12 compounds ranged from 5 to 200 ng L1 (Fig. 2). Diclofenac, N,N-diethyl-meta-toluamide, trimethoprim, sulpiride and indometacin showed high concentrations in the secondary effluents. Carbamazepine and metoprolol followed, with the concentrations ranging from 69 to 120 ng L1, and 60 to 108 ng L1, respectively. Other compounds, such as caffeine, gemfibrozil and mefenamic acid, occurred at the lowest levels. Despite of a wide variation of trimethoprim from different WWTPs, the composition profiles of target pharmaceuticals in secondary effluents from the four WWTPs were quite similar (Fig. 3). The concentration levels of most pharmaceuticals and consumer products detected in the secondary effluent were also lower than those reported in the Europe. They were over 100 ng L1, in some cases even up to 500 ng L1 in the wastewater effluents of the European countries (Santos et al., 2007; Gomez et al., 2007; Ternes, 1998; Vieno et al., 2007b). While in the present study, 10 out of 15 compounds were less than 100 ng L1, and none of them exceeded 400 ng L1 in any effluent samples (Fig. 2). Our results were in agreement with those in Japan (Nakada et al., 2006), Korea (Kim et al., 2007) and some other cities of China (Xu et al., 2007; Gulkowska et al., 2008; Chen et al., 2008). For instance, the concentrations of chloramphenicol in the effluents of 4 WWTPs in Guangzhou were 10 d) may contribute to an increased removal rate of pharmaceuticals (Jones et al., 2007; Vieno et al., 2007b). In the present study, the WWTP C, in which the HRT was higher than the others, was the best in removing these compounds, due to increased contact time of target compounds and the microorganisms. On the other hand, the different SRTs did not have significant effects on the removal efficiency, probably because the SRTs in all the four WWTPs were relatively high (>10 d), and without large differences. In addition, it is noteworthy that the WWTP C employed oxidation ditches, which showed better removal of natural estrogens and estrogenic activity than A/O (Hashimoto et al., 2007). It also could be the reason for the higher removal efficiencies in the WWTP C. Further investigation for different types of WWTPs is necessary to confirm the results mentioned above.

3.4.

Removal efficiency in advanced treatment processes

The removal efficiencies of the pharmaceuticals during the SF, UF/ozonation, as well as MF/RO treatment in three corresponding WWTPs are listed in Table 5. Generally, sand filtration was not effective for these compounds. Only trimethoprim, DEET and gemfibrozil were removed slightly during this treatment process. It should be noticed that these compounds were efficiently removed in the secondary treatment, indicating that the biodegradation on

Table 5 – Removal efficiencies (%) of target pharmaceuticals and consumer products by advanced treatment processes in studied WWTPs. Compound

DEET CF TP BF CA GF DF IM MA MTP CBZ SP

WWTP A

WWTP B

WWTP D

UF

Ozone

SF

MF/RO

0–50 90 80–90 80–90 >90 >90

0–50 90 0–50 >90 >90 >90

the biofilm present on the sand particle, rather than the removal with particles, may be the main reason for their elimination (Gobel et al., 2007). The results showed that ozonation is effective in removing most of the target compounds, probably due to the operation conditions employed in WWTP A (ozone dosage: 5 mg L1, contact time: 15 min). Carbamazepine, diclofenac, indomethacin, sulpiride and trimethoprim were significantly eliminated, with the removal rates of above 95%. The double bond in the azepine ring of carbamazepine and pyrrole ring of indomethacin, and the non-protonated amine of diclofenac and trimethoprim were susceptible to ozone attack (Vieno et al., 2007a; Nakada et al., 2007; Westerhoff et al., 2005). The removal efficiencies of DEET and metoprolol were modest. The amide group, which is not reactive with ozone, could be the reason for the modest removal of DEET (Nakada et al., 2007). Low removal efficiencies were found for bezafibrate, clofibric acid, as well as caffeine. Only 14% of bezafibrate disappeared in the ozone process, consistent with its low rate constants with ozone (590 50 M1 S1, Huber et al., 2003). The reaction site of bezafibrate is the R-oxysubstituent (–O–C(CH3)2COOH) on one of the aromatic rings. However, as the pKa of bezafibrate is 3.6, the R-oxysubstituent cannot be deprotonated and consequently the overall rate constant at pH > 4 is much lower (Huber et al., 2003). It should be noticed that during the ozonation, most of the pharmaceuticals were not mineralized but transformed to the oxidation products. For instance, three oxidation products containing quinazoline-based functional groups were identified during the ozonation of CBZ (Mcdowell et al., 2005). The good performance of ozonation in the present study was consistent with Ternes et al. (2003), Huber et al. (2005) and Okuda et al. (2008). When 5 mg L1 ozone was applied to the effluent of a municipal WWTP in Germany (contact time: 18 min), target compounds, such as trimethoprim, carbamazepine, indomethacin, clofibric acid, were removed by more than 50% (Ternes et al., 2003). Huber et al. (2005) conducted a pilot study on the oxidation of pharmaceuticals during ozonation of conventional activated sludge (CAS) and membrane bio-reactor (MBR) effluents with various ozone dosages, and found that macrolide and sulfonamide antibiotics, estrogens, and acidic pharmaceuticals diclofenac,


a

Tertiary treatment Conventional treatment

120 100 80 60 40 20

SP

CBZ

MTP

IM

DF

GF

CA

BF

TP

CP

-20

CF

0 DEET

Removal Contribution (%)

Pharmaceutical

b 180

Tertiary treatment Secondary treatment Primary treatment

160 140 120 100 80 60 40 20

-40

SP

CBZ

MTP

IM

DF

GF

CA

BF

TP

CF

0 -20

DEET

naproxen and indomethacin were oxidized by more than 90–99% for ozone doses 2 mg L1 in all effluents. The elimination by ultrafiltration in the WWTP A was low for all the investigated compounds. The molecular weight cutoff (MWCO) of UF membranes was much higher than 1000 Da, thus UF membranes showed poor retention of all the investigated pharmaceuticals, of which the molecular weight are less than 400 Da. The removal of individual target compound was less than 50%, and might be due to the adsorption onto the membrane. It has been also demonstrated that UF membrane typically had less than 40% retention of 27 PPCPs, and the mass balances calculated based on the concentration of each compound in feed, permeate and retentate showed the observed retention was significantly governed by adsorption (Yoon et al., 2006). In contrast, MF/RO employed in WWTP D was very effective. In the effluent of MF/RO, all the target compounds except caffeine were not detected. Generally, one or combination of three basic mechanisms could be involved during the rejection of solute by NF/RO membrane: steric effect, charge exclusion and adsorption (Radjenovic et al., 2008). For most pharmaceuticals, the rejections were considered to be dominated by steric interaction in ‘‘tight’’ NF or RO membrane filtration (Nghiem et al., 2005; Radjenovic et al., 2008). As most investigated compounds have molecular weights about 200–400 Da, smaller than MWCO of RO membrane applied, excellent rejection of most pharmaceuticals by RO membrane was observed in this study as well as in previous studies (Kimura et al., 2004; Al-Rifai et al., 2007; Radjenovic et al., 2008). Besides, membrane fouling and the presence of organic matter in the wastewater effluents likely contributed to higher rejections of pharmaceuticals, especially for some hydrophobic ionogenic compound (Nghiem and Coleman, 2008; Comerton et al., 2008). Nevertheless, the rejections of two compounds, caffeine and mefenamic acid were slightly lower (i.e. 50–80% and 0–50%, respectively). The concentration of mefenamic acid in feed wastewaters of MF/RO membrane process was very low, only a bit higher than its LOQ in the wastewater effluent, which could be the reason for the low rejection rate. The low retention of caffeine in the present study was inaccordance with Drewes et al. (2005). They found that in two full-scale RO facilities, target EDCs and PPCPs were efficiently rejected to below detection limit except for caffeine, still detected in the permeates. The physiochemical properties might explain the low rejection rate of caffeine. As a representative of hydrophilic and non-ionic compounds, the rejection driven by charge exclusion and adsorption is negligible, and steric exclusion is solely responsible for the retention of caffeine (Nghiem et al., 2005). However, the molecular weight of caffeine is 195 Da, smaller than other target compounds, and might result in the decreased removal efficiency during the RO membrane filtration process. Compared to the other two, the WWTPs employing ozone and RO membrane filtration as advanced treatment were more efficient in removing pharmaceuticals. For these WWTPs, the advanced treatment made a significant contribution to the total elimination of most pharmaceuticals (Fig. 5). Therefore, the utility of efficient advanced treatment could be considered as a tool to reduce pharmaceuticals in the

Contribution to removal efficiency (%)

424

-60

Pharmaceutical Fig. 5 – Contributions of primary, secondary (or conventional treatment) and tertiary treatment to the total elimination of selected pharmaceuticals in WWTP A (a) and WWTP D (b).

municipal wastewater treatment plants. However, the problems of membrane fouling and further treatment or disposal of retentate challenge the application of RO membrane filtration (Van der Bruggen et al., 2008). For ozonation, as most of the pharmaceuticals could not be mineralized, and oxidation products are formed from parent pharmaceutical compounds (Mcdowell et al., 2005), more research is required to identify the oxidation products and their potential toxicity during the partial oxidation process (Nakada et al., 2007). Besides, economic feasibility should be evaluated by estimating the energy consumption and investment and operation costs for both advanced treatment processes (Joss et al., 2008).

4.

Conclusion

13 out of 15 pharmaceuticals and consumer products from eight classes were detected at four WWTPs in Beijing, China. The concentrations of most compounds in the influent and secondary effluent were lower than those reported in the USA and Europe, but consistent with the production profile of the


pharmaceuticals in China. According to the result of risk assessment for the secondary effluent, only diclofenac might pose a risk to the aquatic environment. The removal efficiencies by the conventional treatment varied for different compounds, depending on their chemical structures, physiochemical properties, as well as the specific treatment processes utilized at each WWTP. Further removal could be achieved by adopting some advanced treatment processes, such as ozonation and MF/RO. However, others, such as sand filtration, showed low efficiency in removing these compounds from secondary effluent.

Acknowledgement This study was supported by the National Science Fund for Distinguished Young Scholars (No. 50625823).

Appendix. Supplementary data Supplementary information related to this article can be found at doi:10.1016/j.watres.2009.07.010.

references

Al-Rifai, J.H., Gabelish, C.L., Schafer, A.I., 2007. Occurrence of pharmaceutically active and non-steroidal estrogenic compounds in three different wastewater recycling schemes in Australia. Chemosphere 69 (5), 803–815. Bolton, S., Null, G., 1981. Caffeine: psychological effects, use and abuse. Orthomolecular Psychiatry 10 (3), 202–211. Castiglioni, S., Bagnati, R., Fanelli, R., Pomati, F., Calamari, D., Zuccato, E., 2006. Removal of pharmaceuticals in sewage treatment plants in Italy. Environmental Science and Technology 40 (1), 357–363. Chen, H.C., Wang, P.L., Ding, W.H., 2008. Using liquid chromatography–ion trap mass spectrometry to determine pharmaceutical residues in Taiwanese rivers and wastewaters. Chemosphere 72 (6), 863–869. China Medicine Economic Information Net (CMEIN), 2005. Chinese Medical Statistical Yearbook. Products section. China Medicine Economic Information Net (CMEIN), Beijing, China. Comerton, A.M., Andrews, R.C., Bagley, D.M., Hao, C.Y., 2008. The rejection of endocrine disrupting and pharmaceutically active compounds by NF and RO membranes as a function of compound and water matrix properties. Journal of Membrane Science 313 (1–2), 323–335. Daughton, C.G., Ternes, T.A., 1999. Pharmaceuticals and personal care products in the environment: agents of subtle change? Environmental Health Perspectives 107 (S6), 907–938. Drewes, J.E., Bellona, C., Oekekoven, M., Xu, P., Kim, T.U., Amy, G., 2005. Rejection of wastewater-derived micropollutants in high-pressure membrane applications leading to indirect potable reuse. Environmental Progress 24 (4), 400–409. European Environment Agency, 1998. Environmental Risk Assessment – Approaches, Experiences and Information Source. European Environment Agency, London, pp. 68–86. Ferrari, B., Paxeus, N., Giudice, R.L., Pollio, A., Grric, J., 2003. Ecotoxicological impact of pharmaceuticals found in treated

425

wastewaters: study of carbamazepine, clofibric acid, and diclofenac. Ecotoxicology and Environmental Safety 55 (3), 359–370. Gobel, A., Mcardell, C.S., Joss, A., Siegrist, H., Giger, W., 2007. Fate of sulfonamides, macrolides, and trimethoprim in different wastewater treatment technologies. Science of the Total Environment 372 (2–3), 361–371. Gomez, M.J., Bueno, M.J.M., Lacorte, S., Fernandez-Alba, A.R., Aguera, A., 2007. Pilot survey monitoring pharmaceuticals and related compounds in a sewage treatment plant located on the Mediterranean coast. Chemosphere 66 (6), 993–1002. Grung, M., Kallqvist, T., Sakshaug, S., Skurtveit, S., Thomas, K.V., 2008. Environmental assessment of Norwegian priority pharmaceuticals based on the EMEA guideline. Ecotoxicology and Environmental Safety 71 (2), 328–340. Gulkowska, A., Leung, H.W., So, M.K., Taniyasu, S., Yamashita, N., Yeung, L.W.Y., Richardson, B.J., Lei, A.P., Giesy, J.P., Lam, P.K.S., 2008. Removal of antibiotics from wastewater by sewage treatment facilities in Hongkong and Shenzhen, China. Water Research 42 (1–2), 395–403. Hashimoto, T., Onda, K., Nakamura, Y., Tada, K., Miya, A., Murakami, T., 2007. Comparison of natural estrogen removal efficiency in the conventional activated sludge process and the oxidation ditch process. Water Research 41 (10), 2117–2126. Huber, M.M., Canonica, S., Park, G.Y., Gunten, U.V., 2003. Oxidation of pharmaceuticals during ozonation and advanced oxidation processes. Environmental Science and Technology 37 (5), 1016–1024. Huber, M.M., Gobel, A., Joss, A., Hermann, N., Loffler, D., Mcardell, C.S., Ried, A., Siegrist, H., Ternes, T.A., Gunten, U.V., 2005. Oxidation of pharmaceuticals during ozonation of municipal wastewater effluents: a pilot study. Environmental Science and Technology 39 (11), 4290–4299. Huerta-Fontela, M., Galceran, M.T., Martin-Alonso, J., Ventura, F., 2008. Occurrence of psychoactive stimulatory drugs in wastewaters in north-eastern Spain. Science of the Total Environment 397 (1–3), 31–40. Huschek, G., Hansen, P.D., Maurer, H.H., Krengel, D., Kayser, A., 2004. Environmental risk assessment of medicinal products for human use according to European Commission recommendations. Environmental Toxicology 19 (3), 226–240. Jjemba, P.K., 2006. Excretion and ecotoxicity of pharmaceutical and personal care products in the environment. Ecotoxicology and Environmental Safety 63 (1), 113–130. Jones, O.A.H., Voulvoulis, N., Lester, J.N., 2005. Human pharmaceuticals in wastewater treatment processes. Critical Reviews in Environmental Science and Technology 35 (4), 401–427. Jones, O.A.H., Voulvoulis, N., Lester, J.N., 2007. The occurrence and removal of selected pharmaceutical compounds in a sewage treatment works utilising activated sludge treatment. Environmental Pollution 145 (3), 738–744. Joss, A., Siegrist, H., Ternes, T.A., 2008. Are we about to upgrade wastewater treatment for removing organic micropollutants. Water Science and Technology 57 (2), 251–255. Kasprzyk-Hordern, B., Dinsdale, R.M., Guwy, A.J., 2009. The removal of pharmaceuticals, personal care products, endocrine disruptors and illicit drugs during wastewater treatment and its impact on the quality of receiving waters. Water Research 43 (2), 363–380. Khan, S.J., Wintgens, T., Sherman, P., Zaricky, J., Schafer, A.I., 2004. Removal of hormones and pharmaceuticals in the advanced water recycling demonstration plant in Queensland, Australia. Water Science and Technology 50 (5), 15–22. Khan, S.J., Ongerth, J.E., 2004. Modelling of pharmaceutical residues in Australian sewage by quantities of use and fugacity calculations. Chemosphere 54 (3), 355–367.

426


Kim, S.D., Cho, J., Kim, I.S., Vanderford, B.J., Snyder, S.A., 2007. Occurrence and removal of pharmaceuticals and endocrine disruptors in South Korean surface, drinking, and waste waters. Water Research 41 (5), 1013–1021. Kimura, K., Toshima, S., Amy, G., Watanabe, Y., 2004. Rejection of neutral endocrine disrupting compounds (EDCs) and pharmaceutical active compounds (PhACs) by RO membranes. Journal of Membrane Science 245 (1–2), 71–78. Lishman, L., Smyth, S.A., Sarafin, K., Kleywegt, S., Toito, J., Peart, T., Lee, B., Servos, M., Beland, M., Seto, P., 2006. Occurrence and reductions of pharmaceuticals and personal care products and estrogens by municipal wastewater treatment plants in Ontario, Canada. Science of the Total Environment 367 (2–3), 544–558. Lindqvist, N., Tuhkanen, T., Kronberg, L., 2005. Occurrence of acidic pharmaceuticals in raw and treated sewages and in receiving waters. Water Research 39 (11), 2219–2228. Mcdowell, D.C., Huber, M.M., Wagner, M., Gunten, U.V., Ternes, T.A., 2005. Ozonation of carbamazepine in drinking water: identification and kinetic study of major oxidation products. Environmental Science and Technology 39 (20), 8014–8022. Nakada, N., Tanishima, T., Shinohara, H., Kiri, K., Takada, H., 2006. Pharmaceutical chemicals and endocrine disrupters in municipal wastewater in Tokyo and their removal during activated sludge treatment. Water Research 40 (17), 3297–3303. Nakada, N., Shinohara, H., Murata, A., Kiri, K., Managaki, S., Sato, N., Takada, H., 2007. Removal of selected pharmaceuticals and personal care products (PPCPs) and endocrine-disrupting chemicals (EDCs) during sand filtration and ozonation at a municipal sewage treatment plant. Water Research 41 (19), 4372–4382. Nghiem, L.D., Schafer, A.I., Elimelech, M., 2005. Pharmaceutical retention mechanisms by nanofiltration membrane. Environmental Science and Technology 39 (19), 7698–7705. Nghiem, L.D., Coleman, P.J., 2008. NF/RO filtration of the hydrophobic ionogenic compound triclosan: transport mechanisms and the influence of membrane fouling. Separation and Purification Technology 62 (3), 709–716. Niwa, T., Inoue, S., Shiraga, T., Takagi, A., 2005. No inhibition of cytochrome P450 activities in human liver microsomes by sulpiride, an antipsychotic drug. Biological and Pharmaceutical Bulletin 28 (1), 188–191. Okuda, T., Kobayashi, Y., Nagao, R., Yamashita, N., Tanaka, H., Tanaka, S., Fujii, S., Konishi, C., Houwa, I., 2008. Removal efficiency of 66 pharmaceuticals during wastewater treatment process in Japan. Water Science and Technology 57 (1), 65–71. Paxeus, N., 2004. Removal of selected non-steroidal antiinflammatory drugs (NSAIDs), gemfibrozil, carbamazepine, beta-blockers, trimethoprim and triclosan in conventional wastewater treatment plants in five EU countries and their discharge to the aquatic environment. Water Science and Technology 50 (5), 253–260. Quinn, B., Gagne, F., Blaise, C., 2008. An investigation into the acute and chronic toxicity of eleven pharmaceuticals (and their solvents) found in wastewater effluent on the cnidarian, Hydra attenuata. Science of the Total Environment 389 (2–3), 306–314. Radjenovic, J., Petrovic, M., Ventura, F., Barcelo, D., 2008. Rejection of pharmaceuticals in nanofiltration and reverse osmosis

membrane drinking water treatment. Water Research 42 (14), 3601–3610. Santos, J.L., Aparicio, I., Alonso, E., 2007. Occurrence and risk assessment of pharmaceutically active compounds in wastewater treatment plants. A case study: Seville city (Spain). Environment International 33 (4), 596–601. Sui, Q., Huang, J., Deng, S. B., Yu, G. Rapid determination of pharmaceuticals from multiple therapeutic classes in wastewater by solid-phase extraction and ultra-performance liquid chromatography tandem mass spectrometry. Chinese Science Bulletin, in press. Ternes, T.A., 1998. Occurrence of drugs in German sewage treatment plants and rivers. Water Research 32 (11), 3245–3260. Ternes, T.A., Meisenheimer, M., Mcdowell, D., Sacher, F., Brauch, H.J., Haist-gulde, B., Preuss, G., Wilme, U., Zulei-seibert, N., 2002. Removal of pharmaceuticals during drinking water treatment. Environmental Science and Technology 36 (17), 3855–3863. Ternes, T.A., Stuber, J., Herrmann, N., Mcdowell, D., Ried, A., Kampmann, M., Teiser, B., 2003. Ozonation: a tool for removal of pharmaceuticals, contrast media and musk fragrances from wastewater? Water Research 37 (8), 1976–1982. Thomas, P.M., Foster, G.D., 2005. Tracking acidic pharmaceuticals, caffeine, and triclosan through the wastewater treatment process. Environmental Toxicology and Chemistry 24 (1), 25–30. Vanderford, B.J., Snyder, S.A., 2006. Analysis of pharmaceuticals in water by isotope dilution liquid chromatography/tandem mass spectrometry. Environmental Science and Technology 40 (23), 7312–7320. Van der Bruggen, B., Manttari, M., Nystrom, M., 2008. Drawbacks of applying nanofiltration and how to avoid them: a review. Separation and Purification Technology 63 (2), 251–263. Vieno, N.M., Harkki, H., Tuhkanen, T., Kronberg, L., 2007a. Occurrence of pharmaceuticals in river water and their elimination in a pilot-scale drinking water treatment plant. Environmental Science and Technology 41 (14), 5077–5084. Vieno, N., Tuhkanen, T., Kronberg, L., 2007b. Elimination of pharmaceuticals in sewage treatment plants in Finland. Water Research 41 (5), 1001–1012. Watkinson, A.J., Murby, E.J., Costanzo, S.D., 2007. Removal of antibiotics in conventional and advanced wastewater treatment: implications for environmental discharge and wastewater recycling. Water Research 41 (18), 4164–4176. Westerhoff, P., Yoon, Y., Snyder, S., Wert, E., 2005. Fate of endocrine-disruptor, pharmaceutical, and personal care product chemicals during simulated drinking water treatment processes. Environmental Science and Technology 39 (17), 6649–6663. Xu, W.H., Zhang, G., Li, X.D., Zou, S.C., Li, P., Hu, Z.H., Li, J., 2007. Occurrence and elimination of antibiotics at four sewage treatment plants in the Pearl River Delta, South China. Water Research 41 (19), 4526–4534. Yoon, Y., Westerhoff, P., Snyder, S.A., Wert, E.C., 2006. Nanofiltration and ultrafiltration of endocrine disrupting compounds, pharmaceuticals and personal care products. Journal of Membrane Science 270 (1–2), 88–100.




Application of the combination index (CI)-isobologram equation to study the toxicological interactions of lipid regulators in two aquatic bioluminescent organisms Ismael Rodea-Palomaresa,1, Alice L. Petreb,1, Karina Boltesb, Francisco Legane´sa, Jose´ Antonio Perdigoń-Melońb, Roberto Rosalb, Francisca Fernańdez-Pinãsa,* a

Departamento de Biologıá, Facultad de Ciencias, Universidad Autońoma de Madrid, 2 Darwin Street, Cantoblanco, 28049 Madrid, Spain Departamento de Ingenierıá Quı´mica, Universidad de Alcala´, Alcala´ de Henares, E-28871 Madrid, Spain

b

article info

abstract

Article history:

Pharmaceuticals in the aquatic environment do not appear singly and usually occur as

Received 26 March 2009

complex mixtures, whose combined effect may exhibit toxicity to the aquatic biota. We


report an environmental application of the combination index (CI)-isobologram equation,

13 July 2009

a method widely used in pharmacology to study drug interactions, to determine the nature

Accepted 18 July 2009

of toxicological interactions of three fibrates toward two aquatic bioluminescent organ-


isms, Vibrio fischeri and the self-luminescent cyanobacterial recombinant strain Anabaena CPB4337. The combination index-isobologram equation method allows computerized

Keywords:

quantitation of synergism, additive effect and antagonism. In the Vibrio test, the fibrate

Antagonism

combinations showed antagonism at low effect levels that turned into an additive effect or

Combination index-isobologram

synergism at higher effect levels; by contrast, in the Anabaena test, the fibrate combinations

equation

showed a strong synergism at the lowest effect levels and a very strong antagonism at high

Cyanobacterium

effect levels. We also evaluated the nature of the interactions of the three fibrates with a real

Fibrates

wastewater sample in the cyanobacterial test. We propose that the combination index-

Synergism

isobologram equation method can serve as a useful tool in ecotoxicological assessment. ª 2009 Elsevier Ltd. All rights reserved.

Vibrio fischeri

1.

Introduction

Fibrates and statins (HMG-CoA reductase inhibitors) are the main lipid-lowering drugs prescribed either alone or in combination therapy in order to decrease plasma cholesterol levels and reduce the incidence of coronary heart disease. Although partially displaced by statins, the total number of fibrate prescriptions is in constant increase in the United States (Holoshitz et al., 2008). Fibric acids are the active forms

of fibrates and belong to the nuclear receptor superfamily of ligand-activated transcription factors. Gemfibrozil and fenofibrate are the fibrates currently marketed in the US, whereas bezafibrate is also available in Europe and other developed countries (Lambropoulou et al., 2008). Fenofibric acid, 2-[4-(4chlorobenzoyl)phenoxy]-2-methylpropanoic acid, is the active metabolite of fenofibrate, the inactive prodrug marketed and dispensed. Gemfibrozil, 5-(2,5-dimethylphenoxy)-2,2-dimethylpentanoic acid and bezafibrate, p-[4-[chlorobenzoylamino-

* Corresponding author. Tel.: þ34 9 1497 8176; fax: þ34 9 1497 8344. E-mail address: [email protected] (F. Fernańdez-Pinãs). 1 Both authors contributed equally to this work. 0043-1354/$ – see front matter ª 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2009.07.026

428


ethyl]-phenoxy]-b-methylpropionic acid, are also fibric acid derivatives with similar pharmacokinetic behaviour (Miller and Spence, 1998). The occurrence of lipid regulators in the discharge of treated urban and municipal wastewater has been relatively well documented. Bezafibrate has been detected in effluents of two British STP with averages up to 230 ng/L (KasprzykHordern et al., 2009). Metcalfe et al. (2003) found around 1 mg/L of gemfibrozil in effluents of Canadian STP, whereas fenofibrate has been reported in concentrations up to 0.5 mg/L in the influent of several Brazilian STP (Stumpf et al., 1999). Andreozzi et al. (2003) found lipid regulators in the effluent of several European STP at concentrations up to 4.76 mg/L (gemfibrozil), 1.07 mg/L (bezafibrate) and 0.16 mg/L (fenofibrate). Rosal et al. (2008), reported the occurrence of bezafibrate and gemfibrozil at levels of 139 and 608 ng/L respectively in the effluent of a Spanish STP. In the same plant Rodrı´guez et al. (2008) found 165 ng/L of fenofibric acid, 61 ng/L of bezafibrate and 143 ng/L of gemfibrozil. It is also significant that removal efficiencies observed in current STP are not always high. Fent et al. (2006) reported maximum removal rates of 50–75% for fenofibric acid and gemfibrozil and somewhat higher for bezafibrate, although for the later, efficiencies below 15% have also been reported. Stumpf et al. (1999) reported a 45% removal of fenofibric acid by an activated sludge conventional treatment. KasprzykHordern et al. (2009) encountered an average degradation of bezafibrate not higher than 67%. On the other hand, Castiglioni et al. (2006) reported that the removal efficiency of bezafibrate during an activated sludge treatment greatly varied from 15% in winter to 87% in summer. At measured environmental concentrations as those reported above (mostly in the ng/L and mg/L range), many studies have shown that the risk of acute toxicity is unlikely (Fent et al., 2006; Han et al., 2006; Borgmann et al., 2007); however, there is a lack of data on chronic toxicity effects. Moreover, pharmaceuticals in the aquatic environments occur as complex mixtures from different classes, not as single contaminants (Gros et al., 2007); thus, although the concentration of individual pharmaceuticals is low, their mixture could prove ecotoxicologically significant (Brain et al., 2004). Current methods of risk assessment usually focus on the assessment of single chemicals, which may underestimate the risk associated with toxic action of mixtures; probably for this reason, in the last years there is an increasing number of studies dealing with complex mixtures of pharmaceuticals (Cleuvers, 2003, 2004; Crane et al., 2006; Han et al., 2006; Borgmann et al., 2007; Christensen et al., 2007; Pomati et al., 2008; Quinn et al., 2009). However, assessment of combined toxicities is not an easy issue. Basically, two different models are in use for the prediction of mixture toxicity, i.e., concentration addition, when pharmaceuticals have a similar mode of toxic action, and response addition or independent action, when pharmaceuticals have different modes of toxic action (Cleuvers, 2003; Teuschler, 2007). However, toxicological interactions, synergisms or antagonisms, between the pharmaceuticals and their effects can occur independently of mode of action; moreover, in most cases, the pharmacological mechanisms of action is known but the toxic mode of action may remain unknown (Cleuvers, 2003; Chou, 2006). In an effort

to overcome this limitation, we report an environmental application of a method widely used in pharmacology to interpret drug interactions; this method, termed as the median-effect/combination index (CI)-isobologram equation (Chou, 2006) allows quantitative determinations of chemical interactions where CI 1 indicate synergism, additive effect and antagonism, respectively. One important property of the method is that previous knowledge of the mechanisms of action of each chemical is not required. Besides, the method takes into account both the potency and the shapes of the dose-effect curve of each chemical. The method has been computerized allowing an automated simulation of synergism and antagonism at different concentrations and at different effect levels of the chemicals in a mixture. The aim of our study was to assess the nature of the toxicological interactions of three fibrates, gemfibrozil, bezafibrate and fenofibric acid, by the method of combination index (CI)-isobologram equation. The three pharmaceuticals were used singly or in two- and three-drug combinations. As toxicity endpoint we have chosen the bioluminescent response of two prokaryotes, the naturally luminescent Vibrio fischeri and the recombinant bioluminescent cyanobacterium Anabaena sp. PCC 7120 CPB4337 (hereinafter, Anabaena CPB4337), both bioluminescent organisms have proved very useful in evaluating the toxicity of individual fibrates in a previous study (Rosal et al., 2009). For Anabaena CPB4337, we also evaluated the nature of the interactions of the three fibrates with a wastewater sample from a local STP, which already proved very toxic to the cyanobacterium (Rosal et al., 2009).

2.

Materials and methods

2.1.

Materials

Gemfibrozil (þ99%) and bezafibrate (þ98%) were purchased from Sigma–Aldrich. Fenofibric acid was produced from fenofibrate (Sigma–Aldrich, þ99% purity) by hydrolysis. A suspension of fenofibrate in isopropanol (30 wt.%, 400 mL) was refluxed during 4 h with an aqueous sodium hydroxide solution (2.0 M, 200 mL). After cooling to less than 70 C, a solution of hydrochloric acid (1.0 M, 325 mL) was slowly added while keeping the temperature over 60 C. The product crystallized after cooling and keeping at room temperature during 4 or more h. The product was filtered and rinsed with water and dried overnight at 60 C under nitrogen. The purity of the product was over 97% checked by HPLC. Solubility of acidic drugs in water is strongly pH dependent with few data considering this variable. Comerton et al. (2007) reported a solubility of 10.9 mg/L of gemfibrozil in water, but we could solve over 125 mg/L in 2 mM MOPS (3-[N-morpholino] propanesulfonic acid) at pH 6 and higher quantities for the pH at which V. fischeri bioassays were performed. In all cases, we avoided the use of solvents and the upper limit for the concentrations of the studied compounds was their solubility in pure water or wastewater at the pH of the bioassay. Wastewater samples were collected from the secondary clarifier of a STP located in Alcala´ de Henares (Madrid) that

429


receives domestic wastewater with a minor contribution of industrial effluents from facilities located near the city. This STP used a conventional activated sludge treatment and has been designed for a total capacity of 375,000 equivalent inhabitants with a maximum flow rate of 3000 m3/h. In a recent previous study (Rosal et al., 2009), we found that this wastewater was very toxic to Anabaena cells with a wastewater dilution as low as 0.11 causing 50% luminescence inhibition (wastewater EC50).

2.2.

Toxicity tests

Bioassays with the photo-luminescent bacteria Vibrio fischeri were carried out according to ISO 11348-3 standard protocol (ISO, 2007). This bioassay measures, during the prescribed incubation period, the decrease in bioluminescence induced in the cell metabolism due to the presence of a toxic substance. The bacterial assay used the commercially available Biofix Lumi test (Macherey-Nagel, Germany). The bacterial reagent is supplied freeze-dried (Vibrio fischeri NRRL-B 11177) and was reconstituted and incubated at 3 C for 5 min before use. The desired pH was set by using NaOH or HCl. The analysis media was 0.34 M NaCl (2% w/v) and tests were performed at 15 C and the measurements of light were made using a luminometer (Optocomp I). The effect of toxicants or toxicant mixtures (i.e., fibrates or fibrate combinations) was measured as percent inhibition with respect to the light emitted under test conditions in the absence of any toxic influence. Toxicity values were routinely obtained after 30 min exposure. Phenol and ZnSO4 7 H2O have been used as toxicity standards and all tests have been replicated to ensure reproducibility. The bioassays using the recombinant bioluminescent cyanobacterium Anabaena CPB4337 were based on the inhibition of constitutive luminescence caused by the presence of any toxic substance (Rodea-Palomares et al., 2009; Rosal et al., 2009). Anabaena CPB4337 was routinely grown at 28 C in the light, ca. 65 mmol photons m2 s1 on a rotary shaker in 50 mL AA/8 (Allen and Arnon, 1955) supplemented with nitrate (5 mM) in 125 ml Erlenmeyer flasks and 10 mg/mL of neomycin sulphate (Nm). Luminescence inhibition-based toxicity assays were performed as follows: 160 mL from five to seven serial dilutions of each tested toxicant or toxicant mixture (i.e.; fibrates or fibrate combinations) plus a control (ddH2O buffered with MOPS at pH 5.8) were disposed in an opaque white 96-well microtiter plates. 40 mL cells, grown as described, were washed twice and resuspended in ddH2O buffered with MOPS at pH 5.8 and were added to the microtiter plate wells to reach a final cell density at OD750 nm of 0.5. The luminescence of each sample was recorded every 5 min for up to 1 h in the Centro LB 960 luminometer. Three independent experiments with duplicate samples were carried out for all Anabaena toxicity assays. CuSO4 has been used as toxicity standard and all tests have been replicated to ensure reproducibility.

2.3.

Experimental design of fibrate combinations

Solutions of gemfibrozil (Gm), bezafibrate (Bz) and fenofibric acid (Fn) prepared as described above were used singly and in two (Bz þ Gm; Fn þ Gm; Fn þ Bz) and three (Fn þ Gm þ Bz)

combinations. Anabaena and Vibrio fischeri cells were treated with serial dilutions of each fibrate individually and with a fixed constant ratio (1:1), based on the individual EC50 values, in their binary and ternary combinations. Five dilutions (serial dilution factor ¼ 2) of each fibrate and combination plus a control were tested in three independent experiments with replicate samples. For evaluating the nature of the interaction of fibrates with wastewater, binary combinations of each fibrate plus wastewater (Fn þ WW; Gm þ WW; Bz þ WW) and a quaternary combination of the three fibrates plus wastewater (Fn þ Gm þ Bz þ WW) were also prepared and tested for Anabaena CPB4337. Anabaena cells were treated with serial dilutions of each fibrate and wastewater individually and with a fixed constant ratio (1:1), based on the individual EC50 values, in their binary and quaternary combinations. Five dilutions (serial dilution factor ¼ 2) of each fibrate and wastewater and their combinations plus a control were tested in three independent experiments with replicate samples. The experimental design is shown in Table 1. All individual fibrate, wastewater and their combination assays were carried out at the same time as recommended by Chou (2006) to maximize computational analysis of data.

2.4. Median-effect and combination index (CI)isobologram equations for determining combined fibrate interactions The results were analyzed using the median-effect/combination index (CI)-isobologram equation by Chou (2006) and Chou and Talalay (1984) which is based on the median-effect principle (mass-action law) (Chou, 1976) that demonstrates that there is an univocal relationship between dose and effect independently of the number of substrates or products and of the mechanism of action or inhibition. This method involved plotting the dose-effect curves for each compound and their combinations in multiple diluted concentrations by using the median-effect equation: fa ¼ fu

m

D Dm

(1)

Where D is the dose, Dm is the dose for 50% effect (e.g., 50% inhibition of bioluminescence or EC50), fa is the fraction affected by dose D (e.g., 0.75 if cell bioluminescence is inhibited by 75%), fu is the unaffected fraction (therefore, fa ¼ 1 fu), and m is the coefficient of the sigmoidicity of the dose-effect curve: m ¼ 1, m > 1, and m < 1 indicate hyperbolic, sigmoidal, and negative sigmoidal dose-effect curve, respectively. Therefore, the method takes into account both the potency (Dm) and shape (m) parameters. If Eq. (1) is rearranged, then: D ¼ Dm½fa=ð1 faÞ

1=m

(2)

The Dm and m values for each fibrate are easily determined by the median-effect plot: x ¼ log (D) versus y ¼ log ( fa/fu) which is based on the logarithmic form of Eq. (1). In the medianeffect plot, m is the slope and log (Dm) is the x-intercept. The conformity of the data to the median-effect principle can be ready manifested by the linear correlation coefficient (r) of the data to the logarithmic form of Eq. (1) (Chou, 2006).

Pure fibrate experiments

Fibrates plus wastewater experiments

Vibrio fischeri Dilutions

1

⁄4 (EC50) 1 ⁄2 (EC50) 1 (EC50) 2 (EC50) 4 (EC50)

Gm

(D)2 (D)1 0.4 8.75 0.8 17.5 1.6 35 3.2 70 6.4 140 Two toxicant combo (D)1 þ (D)2 (1.6:35)

Single toxicant Bz

Fn

(D)3 37.5 75 150 300 600

(D)1 (D)2 2.5 2.5 5 5 10 10 20 20 40 40 Two toxicant combo (D)1 þ (D)2 (1:1)

⁄4 (EC50) ⁄2 (EC50) 1 (EC50) 2 (EC50) 4 (EC50)

0.4 8.75 0.8 17.5 1.6 35 3.2 70 6.4 140 (D)1 þ (D)3 (1.6:150)

1

⁄4 (EC50) ⁄2 (EC50) 1 (EC50) 2 (EC50) 4 (EC50)

0.4 0.8 1.6 3.2 6.4 (D)2 þ (D)3 (35:150)

37.5 75 150 300 600

2.5 5 10 20 30* (D)2 þ (D)3 (1:5)

1

8.75 17.5 35 70 140 Three toxicant combo (D)1 þ (D)2 þ (D)3 (1.6:35:150)

37.5 75 150 300 600

0.4 0.8 1.6 3.2 6.4

37.5 75 150 300 600

1

1

⁄4 (EC50) ⁄2 (EC50) 1 (EC50) 2 (EC50) 4 (EC50) 1

1

⁄4 (EC50) ⁄2 (EC50) 1 (EC50) 2 (EC50) 4 (EC50) 1

8.75 17.5 35 70 140

Anabaena CPB4337

2.5 5 10 20 25* (D)1 þ (D)3 (1:5)

Gm

Single toxicant Bz (D)3 12.5 25 50 100 200

Fn

1

⁄4 (EC50) 1 ⁄2 (EC50) 1 (EC50) 2 (EC50) 4 (EC50)

1

2.5 5 10 20 25*

Gm

(D)1 (D)2 2.5 2.5 5 5 10 10 20 20 40 40 Two toxicant combo (D)1 þ (D)4 (1:0.01)

Bz

WW**

(D)3 12.5 25 50 100 200

(D)4 0.025 0.05 0.1 0.2 0.4

⁄4 (EC50) ⁄2 (EC50) 1 (EC50) 2 (EC50) 4 (EC50)

2.5 5 10 20 40 (D)2 þ (D)4 (1:0.01)

0.025 0.05 0.1 0.2 0.4

12.5 25 50 100 150*

1

2.5 5 10 20 40 (D)3 þ (D)4 (1:0.002)

0.025 0.05 0.1 0.2 0.4

2.5 5 10 20 40 Three toxicant combo (D)1 þ (D)2 þ (D)3 (1:1:5)

12.5 25 50 100 200

1

2.5 5 10 20 40

12.5 25 50 100 200

2.5 5 10 20 40

1

⁄4 (EC50) ⁄2 (EC50) 1 (EC50) 2 (EC50) 4 (EC50)

1

⁄4 (EC50) ⁄2 (EC50) 1 (EC50) 2 (EC50) 4 (EC50)

12.5 25 50 100 200

1

0.025 0.05 0.1 0.2 0.4

Four toxicant combo (D)1 þ (D)2 þ (D)3 þ (D)4 (1:1:5:0.01) 1/8 (EC50) ⁄4 (EC50) 1 ⁄2 (EC50) 1 (EC50) 2 (EC50) 1

1.25 2.5 5 10 20

1.25 2.5 5 10 20

6.25 12.5 25 50 100

0.0125 0.025 0.05 0.1 0.2

For the Anabaena test, the design for the experiment with the wastewater [WW (D4)] sample is also included. The experimental design is based on EC50 ratios as proposed by Chou and Talalay (1984). EC50 is the effective concentration of a toxicant which caused a 50% bioluminescence inhibition. The combination ratio was approximately equal to the EC50 ratio of the combination components (i.e., close to their equipotency ratio). *Upper maximal possible dose due to the solubility limit of fibrates in pure water. **EC50 for wastewater is the dilution which caused 50% luminescence inhibition. (D)1, (D)2 and (D)3 in mg/L, (D)4 is the dilution of wastewater in ddH2O.


1

Anabaena CPB4337

Single toxicant Fn

430

Table 1 – Experimental design for determining toxicological interactions of fenofibric acid [Fn (D)1], gemfibrozil [Gm (D)2], bezafibrate [Bz (D)3] and their binary and ternary combinations for Vibrio fischeri and Anabaena CPB4337 bioluminescence tests.

431


These parameters were then used to calculate doses of the fibrates and their combinations required to produce various effect levels according to Eq. (1); for each effect level, combination index (CI) values were then calculated according to the general combination index equation for n chemical combination at x% inhibition (Chou, 2006):

n

ðCIÞx ¼

n X

ðDÞj

j¼1

ðDx Þj

¼

P ðDx Þ1n ½Dj = n1 ½D

n X j¼1

ðDm Þj

(3)

n h io1=mj fax j = 1 fax j

where n(CI)x is the combination index for n chemicals (e.g., fibrates) at x% inhibition (e.g., bioluminescence inhibition); (Dx)1n is the sum of the dose of n chemicals that exerts x% P inhibition in combination, {[Dj]/ n1 ½D} is the proportionality of the dose of each of n chemicals that exerts x% inhibition in combination; and (Dm)j {( fax)j/[1 ( fax)j]}1/mj is the dose of each drug alone that exerts x% inhibition. From Eq. (3), CI 1 indicates synergism, additive effect and antagonism, respectively.

2.5.

versus fa, the fraction affected by a particular dose; see Eq. (1)) and polygonograms (a polygonal graphic representation depicting synergism, additive effect and antagonism for three or more drug combinations). Linear regression analyses were computed using MINITAB Release 14 for Windows (Minitab Inc; USA).

Analysis of results

Computer program CompuSyn (Chou and Martin, 2005, Compusyn Inc, USA) was used for calculation of dose-effect curve parameters, CI values, fa-CI plot (plot representing CI

3.

Results

3.1. Toxicological interactions of fibrate combinations in Vibrio fischeri and Anabaena CPB4337 bioluminescence tests Applying the combination index-isobologram method, we evaluated the nature of gemfibrozil (Gm), fenofibric acid (Fn) and bezafibrate (Bz) interactions both in Vibrio fischeri and Anabaena CPB4337 bioluminescence tests. Table 2 shows the dose-effect curve parameters (Dm, m and r) of the three fibrates singly and their binary and ternary combinations, as well as mean combination index (CI) values of fibrate combinations. Dm was the dose required to produce the medianeffect (analogous to the EC50); Dm values for Fn were the lowest both, in Vibrio and Anabaena tests, Dm values for Gm were in the same range for both Vibrio and Anabaena while Bz

Table 2 – Dose-effect relationship parameters and mean combination index (CI) values (as a function of fractional inhibition of luminescence) of gemfibrozil (Gm), fenofibric acid (Fn), and bezafibrate (Bz) individually and of their binary and ternary combinations on Vibrio fischeri and Anabaena CPB4337 bioluminescence tests. Drug combo

Vibrio fischeri Dose-effect parameters Dm

Fn Gm Bz Gm þ Bz Fn þ Bz Fn þ Gm Fn þ Gm þ Bz

mg/L

(mM)

1.45 20.58 252.07 78.20 153.79 9.84 55.69

(4.01) (82.11) (696.46) (234.20) (424.93) (38.74) (166.69)

CI values

m

r

0.78 1.53 1.15 1.54 1.09 1.15 1.23

0.989 0.966 0.975 0.991 0.981 0.973 0.993

EC10

– – – 1.13 0.13 2.98 0.15 0.99 0.17 1.46 0.06

EC50

Add Ant Add Ant

– – – 0.97 0.04 1.71 0.03 0.75 0.05 1.01 0.02

EC90

Add Ant Syn Add

– – – 0.86 0.05 1.17 0.06 0.86 0.08 0.99 0.03

Syn Ant Syn Add

Anabaena CPB4337 Dose-effect parameters Dm mg/L Fn Gm Bz Gm þ Bz Fn þ Bz Fn þ Gm Fn þ Gm þ Bz

8.53 10.69 12.56 19.17 13.92 12.26 6.62

CI values

m

r

0.96 0.81 1.08 0.84 0.76 0.46 0.53

0.971 0.959 0.990 0.972 0.965 0.955 0.960

EC10

EC50

EC90

(mM) (23.62) (42.67) (34.70) (56.88) (38.49) (41.45) (19.45)

– – – 1.06 0.15 0.55 0.06 0.13 0.02 0.09 0.01

Add Syn Syn Syn

– – – 1.57 0.06 1.19 0.04 1.29 0.05 0.57 0.02

Ant Ant Ant Syn

– – – 2.5 0.22 2.59 0.14 12.9 2.33 3.92 0.19

Ant Ant Ant Ant

The parameters m, Dm and r are the antilog of x-intercept, the slope and the linear correlation coefficient of the median-effect plot, which signifies the shape of the dose-effect curve, the potency (EC50), and conformity of the data to the mass-action law, respectively (Chou, 1976; Chou and Talalay, 1984; Chou, 2006). Dm and m values are used for calculating the CI values (Eq. (3)); CI 1 indicate synergism (Syn), additive effect (Add), and antagonism (Ant), respectively. EC10, EC50 and EC90, are the doses required to inhibit bioluminescence 10, 50 and 90%, respectively. Computer software CompuSyn was used for automated calculation and simulation.

432


Dm values were an order of magnitude higher in the Vibrio test (Rosal et al., 2009); m was the Hill coefficient used to determine the shape of the dose-response curve, hyperbolic (m ¼ 1), sigmoidal (m > 1) or negative sigmoidal (m < 1); also shown in the table, linear regression correlation coefficients (r-values) of the median-effect plots were >0.95 in all cases, indicating the conformity of the data to the median-effect principle which qualifies for further studies using this method. The Dm and m values for single fibrates and for their combination mixtures were used for calculating synergism or antagonism based on the CI Eq. (3) (Chou, 2006). Fig. 1 shows the fa-CI plot of fibrate interactions both for Vibrio (Fig. 1a) and Anabaena tests (Fig. 1b); the fa-CI plot depicts the CI value

a

3.0

Combination index, CI

2.5

2.0

1.5

1.0

0.5

0.0 0.0

0.2

0.4

0.6

0.8

1.0

Fraction affected, fa

b

3.0


2.5

2.0

1.5

versus fa (effect level or fraction of luminescence inhibited by a fibrate singly or in combination with respect to the control) for two (Fn þ Bz; Fn þ Gm and Bz þ Gm) and three fibrate (Fn þ Gm þ Bz) combinations. The fa-CI plot is an effectoriented plot that shows the evolution of the kind of interaction (synergism, antagonism, additive effect) as a function of the level of the effect ( fa) of a particular toxicant on the reference organism ( fa, where ECa ¼ fa 100; i.e., EC10 ¼ f10 100). In the Vibrio test (Fig. 1a), the Bz þ Gm and Fn þ Gm binary combination showed a slight antagonism at very low fa values and slight synergism (Fn þ Gm) or nearly additive effects (Bz þ Gm) at the highest fa values, the Fn þ Bz combination showed a strong antagonism at low effect levels but the antagonism decreased and approached an additive kind of interaction at the highest fa levels; the ternary combination (Fn þ Gm þ Bz) showed a moderate antagonism at low fa values that also turned into a nearly additive effect at fa values above 0.4. Correlation analyses were made between CI values of the fibrate ternary combination and CI values of each of the fibrate binary combinations to determine which binary combination interaction was predominant in the ternary mixture (Table 3); the highest correlation coefficient was found for the Fn þ Bz combination (r ¼ 0.91), suggesting that this combination interaction predominated in the three fibrate mixture. The fa-CI plot of the Anabaena test (Fig. 1b) showed the opposite pattern of interactions as the three binary and the ternary combinations showed from slight to strong synergism at the lowest fa values that turned into a very strong antagonism at fa values over 0.5; the ternary combination (Fn þ Gm þ Bz) closely followed the interaction pattern of the binary Fn þ Gm combination, this is confirmed by the highest correlation coefficient found between the CI values of the ternary combination and the CI values of the Fn þ Gm combination (r ¼ 0.996) which suggests that in the Anabaena test, this particular combination seemed to be the predominant in the ternary toxicological interaction. Selected average CI values for both Vibrio fischeri and Anabaena CPB4337 tests at three representative dose levels (EC10, EC50 and EC90) and the combined effects are summarized in Table 2.

1.0

3.2. Toxicological interactions of wastewater and fibrate combinations in the Anabaena CPB4337 bioluminescence test

0.5

0.0 0.0

0.2

0.4

0.6

0.8

1.0

Fraction affected, fa

Fig. 1 – Combination index plot ( fa-CI plot) for a set of three fibrate combinations: Fn D Bz (–6–), Bz D Gm (–B–), Fn D Gm (–,–) and Fn D Gm D Bz (–;–) for Vibrio fischeri test (a) and Anabaena CPB4337 test (b). CI values are plotted as a function of the fractional inhibition of bioluminescence ( fa) by computer simulation (CompuSyn) from fa [ 0.10 to 0.95. CI 1 indicates synergism, additive effect and antagonism, respectively. At least three independent experiments with two replicates were used. The vertical bars indicate 95% confidence intervals for CI values based on sequential deletion analysis (SDA) (Chou and Martin, 2005). Fn [ fenofibric acid, Bz [ bezafibrate and Gm [ gemfibrozil.

In a recent previous study (Rosal et al., 2009), we found that a wastewater sample collected from a local STP was very toxic to Anabaena cells with a wastewater dilution of 0.11 causing 50% luminescence inhibition (wastewater EC50). The observed toxicity was attributed to the combined toxicities of over thirty micropollutants, which included fibrates as well as other pharmaceuticals (Rosal et al., 2008). We sought to investigate the nature of the interaction between the wastewater (WW) and the three fibrates in binary (Fn þ WW; Bz þ WW and Gm þ WW) and quaternary (Fn þ Gm þ Bz þ WW) combinations; for these experiments, the wastewater itself was regarded as a toxicant; the experimental design was analogous to the one for the three fibrate interactions and is also shown in Table 1. The r-values of the median-effect plots were >0.95 in all cases, indicating that the data conformed to the medianeffect principle (not shown). Fig. 2 shows the fa-CI plot for each

433


Table 3 – Correlation analyses between CI values of fibrate ternary and fibrate D wastewater quaternary combinations ( y) and their binary combinations (x) for Vibrio fischeri and Anabaena CPB4337 tests. Test organism

Combinations

V. fischeri

Fn þ Gm þ Bz

versus

Anabaena CPB4337

Fn þ Gm þ Bz

versus

WW þ Fn þ Gm þ Bz

versus

Regression parameters

Gm þ Bz Fn þ Bz Fn þ Gm Gm þ Bz Fn þ Bz Fn þ Gm Gm þ Bz Fn þ Bz Fn þ Gm Fn þ WW Gm þ WW Bz þ WW

xo

m

r

0.614 0.594 0.067 5.876 2.716 0.282 0.079 0.199 0.464 0.253 2.131 0.003

1.77 0.281 1.40 4.39 3.00 0.247 0.372 0.246 0.017 1.31 2.41 0.865

0.83 0.91 0.81 0.91 0.941 0.996 0.999 0.998 0.897 0.999 0.89 0.999

Fn ¼ fenofibric acid, Bz ¼ bezafibrate, Gm ¼ gemfibrozil, WW ¼ wastewater. The parameters of linear regression equations: x0 (value of y when x ¼ 0); m (slope) and r (correlation coefficient) with all p-values of 0.001. Analyses were computed using MINITAB Release 14 for Windows.

of the binary fibrate-wastewater combination and the quaternary combination; as can be observed, in a broad range of fa values, the binary combinations showed a strong synergism; however, at fa values above 0.8, the binary Fn þ WW and Bz þ WW combinations approached an additive effect and at fa values above 0.95, these two combinations yielded antagonism; by contrast, the Gm þ WW combination became even more synergistic. The quaternary combination interaction showed a strong synergism through a broad range of fa values but also turned into slight antagonism at fa values above 0.95,


2.0

1.5

1.0

0.5

0.0 0.0

0.2

0.4

0.6

0.8

1.0

Fractions affected, fa

Fig. 2 – Combination index plot ( fa-CI plot) for a set of three fibrates and toxic wastewater sample in their binary and quaternary combinations: Gn D WW (–6–), Fn D WW (–B–), Bz D WW (–,–) and Fn D Gm D Bz D WW (–;–) for the Anabaena CPB4337 test. CI values are plotted as a function of the fractional inhibition of bioluminescence ( fa) by computer simulation (CompuSyn) from fa [ 0.10 to 0.95. CI 1 indicates synergism, additive effect and antagonism, respectively. At least three independent experiments with two replicates were used. The vertical bars indicate 95% confidence intervals for CI values based on sequential deletion analysis (SDA) (Chou and Martin, 2005). Fn [ fenofibric acid, Bz [ bezafibrate, Gm [ gemfibrozil, WW [ wastewater.

closely resembling the pattern of the Fn þ WW and Bz þ WW interactions which is confirmed by the highest r value (r ¼ 0.999) in the correlation analyses (Table 3), which suggests a predominant effect of Fn and Bz in the quaternary interaction. The computer software CompuSyn (Chou and Martin, 2005) displays a type of graphic termed polygonogram, which is a semiquantitative method of representing interactions between three or more compounds at a determined fa value. This graphic allows a simplified visual presentation of the overall results. Fig. 3 shows the polygonogram for the three fibrates and the wastewater at four fa values; synergism is indicated by solid lines and antagonism by broken ones; the thickness of the lines indicates the strength of the interaction. The polygonogram clearly shows the synergistic interaction of wastewater in combination with each of the three fibrates at low fa values and the antagonistic interaction that appeared at the highest fa value, 0.99, for the Fn þ WW and the Bz þ WW combinations. The same wastewater sample collected from a local STP was proved as responsible of stimulation of the bioluminescence activity of Vibrio fischeri to 110–120% of that of the control. Moreover, the EC50 values for the fibrates in the wastewater were higher than those for fibrates in pure water (Rosal et al., 2009). The same trend was observed comparing the dose-effect curve parameters (Dm, m and r) for the ternary combination (Fn þ Gm þ Bz) of fibrates in ddH2O and wastewater. The dose required to produce the median-effect (Dm) in Vibrio fischeri test when (Fn þ Gm þ Bz) were solved in wastewater was 131.936 compared to 55.6951 mg/L required when ddH2O was employed. CI values could not be calculated for Vibrio fischeri due to the fact that the wastewater itself was not toxic to this bacterium; synergism or antagonism could not be properly estimated (Chou, 2006).

4.

Discussion

The three fibrates that we have used in our study are lipid modifying agents that are effective in lowering elevated

434


Fig. 3 – Polygonograms showing the toxicological interactions of three fibrates and a toxic wastewater sample in their binary combinations (Fn D Bz, Bz D Gm, Fn D Gm, Gm D WW, Fn D WW, Bz D WW) as calculated by CompuSyn (Chou and Martin, 2005) for Anabaena CPB4337 test at four effect levels: fa [ 0.1 (a), fa [ 0.5 (b), fa [ 0.9 (c) and fa [ 0.99 (d). Solid lines indicate synergism, broken lines indicate antagonism. The thickness of the line represents the strength of synergism or antagonism. Figure generated by CompuSyn (Chou and Martin, 2005).

plasma triglycerides and cholesterol in humans (Staels et al., 1998). These pharmaceuticals are highly used, ubiquitous and persistent (Daughton and Ternes, 1999), they are found at ng/L to mg/L levels in many STP effluents, surface waters, estuaries of rivers and even in sea water (for a review, see Hernando et al., 2007). Although non-target organisms; the continuous release of these substances into the environment may cause acute or chronic toxicity to the aquatic biota. Regarding fibrates, in the recent literature there are many reports dealing with individual toxicity of different fibrates in a range of aquatic organisms from primary producers to consumers; a great variability has been found in the sensitivity of the different test organisms toward these pharmaceuticals (Hernando et al., 2007). However, pharmaceuticals such as fibrates do not occur singly in a polluted environment and are usually found as mixtures, therefore, for risk assessment strategies it is important to know the combined effects of pharmaceuticals in non-target organisms (Teuschler, 2007). There are two concepts widely used for the prediction of mixture toxicity: concentration addition (CA) and independent action (IA) (Backhaus et al., 2003; Vighi et al., 2003; Backhaus et al., 2004; Junghans et al., 2006). CA is used for mixtures whose components act in a similar mode of action while IA is based on the idea of dissimilar action, meaning that the compounds have different mechanisms of action; however, as discussed by Cleuvers (2003) the terms similar/

dissimilar action may be misleading. Pharmaceuticals such as fibrates may have the same pharmacological mechanism of action [i.e., interaction with the binding peroxisome proliferator-activated receptor a (PPARa)] in their target organism, humans; however, if fibrates released in the aquatic environments prove toxic to different non-target organisms, the exact mechanism of toxicity (probably different to the pharmacological mode of action) should be investigated in depth before choosing which approach, CA or IA, to use. In fact, only if toxicity is regarded as non-specific at all, the concept of CA may be used although it may also have limitations. Cleuvers (2003) found that two totally different pharmaceuticals, a fibrate and an anti-epileptic drug, followed the concept of CA in the Daphnia toxicity test and the concept of IA in an algal test; both pharmaceuticals apparently shared the same nonspecific toxic mode of action for both organisms; so it appeared that the concept of CA or IA did not depend on a similar/dissimilar mode of action but on the tested organism. The author also discussed that by definition, when using CA, substances applied below their individual noneffect concentration (NOEC) will contribute to the total effect of the mixture while when using IA, substances applied below their NOEC will not contribute to the total effect of the mixture, meaning that any combination effect will probably be higher if the substances follow the concept of CA and this may be misleading when considering the terms synergism or


antagonism because as also discussed by Chou (2006), synergism or antagonism may occur independently of a similar or dissimilar mode of action. In this context, Fent et al. (2006) tested mixtures of different kinds of pharmaceuticals (including fibrates) that might have estrogenic activity in a yeast reporter system; they applied the CA model and found that it had severe limitations when the dose-response curves of the individual pharmaceuticals were not identical or at low effect concentrations. As pharmaceuticals released in the environment may have such diverse dose-effect relationships, the lack of appropriate prediction suggests limitation of the CA mixtures concept. To study the nature of the combined fibrate interactions (synergism, additive effect, antagonism) for the Vibrio fischeri and Anabaena CPB4337 bioluminescence tests, we have followed the combination index (CI)-isobologram equation method of Chou (2006) and Chou and Talalay (1984); a method widely used to study drug interactions in pharmacology. This method may be considered a fractional analysis technique for drug interactions (Berenbaum, 1981; Bovill, 1998) that is independent of the mode of action and considers both the potency (EC50, Dm) and the shape (m) of the dose-effect curve for each drug. The method allows prediction of synergism/ antagonism at all effect levels ( fa) for a combination of n drugs; in contrast with the classical graphical isobologram method (Berenbaum, 1981; Bovill, 1998) that cannot be used for more than three compounds and have also graphical limitations to show all effect levels. By using this method, we have been able to determine the nature of interactions for a wide range of effect levels of three fibrates in binary and ternary combinations in two different bioluminescent organisms. However, the nature of these interactions was not uniform along the fa levels range in any of the two organisms. In Vibrio fischeri, antagonism predominated at low and intermediate fa levels but at the highest effect levels, interactions became additive or slightly synergistic. In Anabaena, a dual synergistic/antagonistic behaviour was observed with synergism predominating at fa levels below 0.4–0.5 and strong antagonism above these fa values. It is difficult to give an explanation to this phenomenon because the combination index method only allows quantitative determination of synergism or antagonism and the elucidation of the mechanism by which synergism or antagonism occurs is a separate issue that needs a different kind of approach. However, tentatively, antagonism, which could be considered the predominant interaction in Vibrio fischeri and Anabaena, might be explained by the structural similarity of fibrates which are related pharmaceuticals that share a common structural motif, a cyclic head and a hydrophobic tail (Rosal et al., 2009); at the fa levels where antagonism is found in both organisms, fibrates may compete with one another for the same target/ receptor sites. The slight synergism found at very high levels in Vibrio fischeri could perhaps be explained by the fact that at very high concentrations, fibrates may somehow combine to increase toxicity by an unspecific way of action that is probably not related to their pharmacological mechanism. Perhaps, the most puzzling interaction is the observed high synergism at very low fa levels in Anabaena; the mechanism of such synergistic interaction is not readily apparent. One could speculate that these fibrates at very low concentrations could

435

involve what Jia et al. (2009) in their extensive review of mechanisms of drug combinations call ‘‘facilitating actions’’ that means that secondary actions of one drug enhances the activity or level of another drug in the mixture or alternatively ‘‘complementary actions’’ when drugs act at the same target at different sites, at overlapping sites or at different targets of the same pathway. However, in the literature there are very few reports on possible targets of fibrates on the prokaryotic cell; English et al. (1994) reported that peroxisome proliferators such as fibrates have been shown to induce cytochrome P450BM-3 which catalyzes the hydroxylation of fatty acids, in Bacillus megaterium. Garbe (2004) reported that fibrates induced methyltransferase Rv0560c with a function in the biosynthesis of isoprenoid compounds in Mycobacterium tuberculosis; Garbe (2004) suggested that both effects may act on the plasma membrane, modulating its properties. In mitochondria, which have significant features that resemble those of prokaryotes, fibrates have been found to inhibit respiratory complex I (NDH-1 complex) and to interfere with mitochondrial fatty acid oxidation (Scatena et al., 2007). Whether fibrates may exert similar effects in Vibrio fischeri and Anabaena to those observed in Bacillus or mitochondria needs further research. In this context, we have found that, as the faCI plots show, fibrate interactions do not follow the same pattern in both bacteria, this may be due to the different origin and position in the food web of Vibrio fischeri, a heterotrophic marine prokaryote and Anabaena CPB4337, a recombinant strain of an obligate phototrophic freshwater prokaryote; in fact, Anabaena presents intracellular photosynthetic membranes called thylakoids where several functionally distinct NDH-1 complexes have been found with roles both in respiration and photosynthesis (Battchikova and Aro, 2007). If fibrates are also affecting NDH-1 complexes in Anabaena, their effects might be very different to those in Vibrio fischeri; so, although we have measured the same toxicity endpoint in both bacteria, i.e., luminescence inhibition, the combined effects of fibrates seem to depend on the test organism. Ince et al. (1999) assessed toxic interactions of heavy metals in binary mixtures on Vibrio fischeri and the freshwater aquatic plant Lemna minor and found that most binary metal mixtures exhibited only antagonistic interactions in the plant opposed to fewer antagonistic and some synergistic interactions in the heterotrophic bacterium. These authors also found that in the bacterium, the nature of the interaction (synergism or antagonism) also changed with the effect level of the binary metal combinations, although the authors did not provide a mechanistic explanation for this variability. Cheng and Lu (2002) made a comparison of joint interactions of organic toxicants in binary mixtures in Escherichia coli and Vibrio fischeri and found that toxicants with the same mechanisms of toxicity displayed mostly additive or antagonistic interactions in E. coli and Vibrio fischeri; however a synergistic interaction was found between glutardialdehyde and butyraldehyde in Vibrio. Synergistic effects in both bacteria were mostly associated with toxicants with different mechanisms of toxicity, although antagonism clearly predominated. They also found that for a total of 44 organic binary mixtures, only six mixtures resulted in identical type of interaction in both bacteria. From our results and those of other authors’ (Ince et al., 1999; Cheng and Lu, 2002; Cleuvers, 2003) one may

436


conclude that previous knowledge of the mechanism of toxic action of a compound is not useful enough to predict which kind of interactions it will display when combined with other toxicants with the same or different toxic mechanism; also, as we have shown, the nature of the interaction may depend on the effect level of the mixture. In addition, different types of organisms will show completely different responses to mixtures of potential toxicants. We previously found that a local wastewater was very toxic for the Anabaena CPB4337 test but non-toxic at all for the Vibrio fischeri or Daphnia magna tests. This wastewater is a mixture of over thirty micropollutants, mostly pharmaceuticals of different therapeutics groups that, besides the fibrates used in this study, included antibiotics, analgesics/anti-inflammatories, b-blockers, antidepressants, anti-epileptics/psychiatrics, ulcer healing compounds, diuretics and bronchodilators; personal care products and some priority organic pollutants are also present (Rosal et al., 2008). The method of Chou allows to combine one drug mixture with another drug mixture and determine their interactions; therefore, we studied the nature of the interaction of fibrates and wastewater in the Anabaena bioluminescence test; interestingly, we found that in a wide range of effect levels, the interaction of wastewater and the three fibrate combination was synergistic; particularly, at very low fa values which means that fibrates are at low concentrations and the wastewater is diluted several-fold, the method predicted a strong synergism; this may be due, as discussed above, to the observed synergistic interactions of fibrates with one another as well as interactions with some of the detected micropollutants when present at very low concentrations. This observed synergism may be environmentally relevant since most pharmaceuticals such as fibrates do not usually show acute toxicity on non-target organisms when tested at real environmental concentrations (Hernando et al., 2007) but in a mixture, if they act synergistically, they could prove toxic for a test organism even at low concentrations; these results agree with those found by Hernando et al. (2004) who reported synergistic toxic effects for Daphnia magna test when wastewater was spiked with environmental concentrations of several pharmaceuticals including fibrates. By contrast, our results show that at high fa values ( fa > 0.8), the combined interaction of the quaternary fibrates þ wastewater combination, the binary Fn þ WW and Bz þ WW combinations approached an additive effect and eventually became antagonistic; in our previous study, the wastewater itself decreased Anabaena bioluminescence by 84% with a lower confidence limit of 76% and an upper confidence limit of 91%; when the wastewater was spiked with increasing concentrations of each fibrate we found that, with the exception of gemfibrozil, the EC50 values for the fibrates in the wastewater were higher than those for fibrates in pure water; this was attributed either to reduced bioavailability or to antagonistic effects of fibrates with other chemicals present in the wastewater; although we did not use the method of Chou, we obtained similar results to the ones we report in this study; that is, at high effect levels (>84% luminescence inhibition) the interaction of fibrates with wastewater, except the Gm þ WW combination, showed antagonism. Based on our results, we propose that the combination index (CI)-isobologram equation, a method widely used in

pharmacology both for in vitro and in vivo bioassays, may also be applied in environmental toxicology as a general method to define interactions of potential toxicants in mixtures in any test organism and/or toxicological endpoint of interest and could be especially useful for risk assessment strategies that take into account the toxicological interactions of substances in a mixture.

5.

Conclusions

We report an environmental application of the combination index (CI)-isobologram equation to study the nature of the interactions of fibrate combinations in two bioluminescent aquatic organisms. The method allowed calculating synergism or antagonism of binary and ternary fibrate combinations at all effect levels simultaneously; we could also test the method with a real wastewater sample in binary and quaternary combination with the fibrates, finding that at very low effect levels, the fibrates acted synergistically with the wastewater in the Anabaena test. The proposed method may be used with other test organisms and/or toxicological endpoints and could be particularly useful for risk assessment approaches to toxicity of complex mixtures.

Acknowledgements The research was funded by the Spanish Ministry of Education through grants CTM2005-03080/TECNO and CSD2006-00044 and the Comunidad de Madrid, grants 0505/AMB-0395 and 0505/MB/0321.

references

Allen, M.B., Arnon, D.I., 1955. Studies on nitrogen-fixing blue green algae. I Growth and nitrogen fixation by Anabaena cylindrica Lemm. Plant Physiol 30 (4), 366–372. Andreozzi, R., Raffaele, M., Nicklas, P., 2003. Pharmaceuticals in STP effluents and their solar photodegradation in aquatic environment. Chemosphere 50 (10), 1319–1330. ˚ ., Blanck, H., Faust, M., Backhaus, T., Altenburger, R., Arrhenius, A Finizio, A., Gramatica, P., Grote, M., Junghans, M., Meyer, W., Pavan, M., Porsbring, T., Scholze, M., Todeschini, R., Vighi, M., Walter, H., Horst Grimme, L., 2003. The BEAM-project: prediction and assessment of mixture toxicities in the aquatic environment. Continental Shelf Res. 23 (17–19), 1757–1769. Backhaus, T., Arrhenius, A., Blanck, H., 2004. Toxicity of a mixture of dissimilarly acting substances to natural algal communities: predictive power and limitations of independent action and concentration addition. Environ. Sci. Technol 38 (23), 6363–6370. Battchikova, N., Aro, E.-M., 2007. Cyanobacterial NDH-1 complexes: multiplicity in function and subunit composition. Physiol. Plant 131 (1), 22–32. Berenbaum, M.C., 1981. Criteria for analyzing interactions between biologically active agents. Adv. Cancer Res. 35, 269–335. Borgmann, U., Bennie, D.T., Ball, A.L., Palabrica, V., 2007. Effect of a mixture of seven pharmaceuticals on Hyalella azteca over multiple generations. Chemosphere 66 (7), 1278–1283.


Bovill, J.G., 1998. Analysis of drug interactions. Baillie`re’s Clin. Anaesthesiol 12 (2), 135–168. Brain, R.A., Johnson, D.J., Richards, S.M., Hanson, M.L., Sanderson, H., Lam, M.W., Young, C., Mabury, S.A., Sibley, P.K., Solomon, K.R., 2004. Microcosm evaluation of the effects of an eight pharmaceutical mixture to the aquatic macrophytes Lemna gibba and Myriophyllum sibiricum. Aquat. Toxicol 70 (1), 23–40. Castiglioni, S., Bagnati, R., Fanelli, R., Pomati, F., Calamari, D., Zuccato, E., 2006. Removal of pharmaceuticals in sewage treatment plants in Italy. Environ. Sci. Technol 40 (1), 357–363. Chen, C.-Y., Lu, C.-L., 2002. An analysis of the combined effects of organic toxicants. Sci. Total Environ 289 (1–3), 123–132. Chou, T.C., 1976. Derivation and properties of Michaelis–Menten type and Hill type equations for reference ligands. J. Theor. Biol. 59 (2), 253–276. Chou, T.C., 2006. Theoretical basis, experimental design, and computerized simulation of synergism and antagonism in drug combination studies. Pharmacol. Rev. 58 (3), 621–681. Chou, T.C., Martin, N., 2005. CompuSyn for Drug Combinations: PC Software and User’s Guide: A Computer Program for Quantification of Synergism and Antagonism in Drug Combinations and the Determination of IC50 and ED50 and LD50 Values. ComboSyn, Inc., Paramus, NJ. Chou, T.C., Talalay, P., 1984. Quantitative analysis of dose-effect relationships: the combined effects of multiple drugs or enzyme inhibitors. Adv. Enzyme Regul 22, 27–55. Christensen, A.M., Faaborg-Andersen, S., Ingerslev, F., Baun, A., 2007. Mixture and single-substance toxicity of selective serotonin reuptake inhibitors toward algae and crustaceans. Environ. Toxicol. Chem. 26 (1), 85–91. Cleuvers, M., 2003. Aquatic ecotoxicity of pharmaceuticals including the assessment of combination effects. Toxicol. Lett. 142 (3), 185–194. Cleuvers, M., 2004. Mixture toxicity of the anti-inflammatory drugs diclofenac, ibuprofen, naproxen, and acetylsalicylic acid. Ecotoxicol. Environ. Saf 59 (3), 309–315. Comerton, A.M., Andrews, R.C., Bagley, D.M., Yang, P., 2007. Membrane adsorption of endocrine disrupting compounds and pharmaceutically active compounds. J. Memb. Sci. 303 (1–2), 267–277. Crane, M., Watts, C., Boucard, T., 2006. Chronic aquatic environmental risks from exposure to human pharmaceuticals. Sci. Total Environ 367 (1), 23–41. Daughton, C.G., Ternes, T.A., 1999. Pharmaceuticals and personal care products in the environment: agents of subtle change? Environ. Health Perspect 107 (6), 907–938. English, N., Hughes, V., Wolf, C.R., 1994. Common pathways of cytochrome P450 gene regulation by peroxisome proliferators and barbiturates in Bacillus megaterium ATCC14581. J. Biol. Chem. 269 (43), 26836–26841. Fent, K., Escher, C., Caminada, D., 2006. Estrogenic activity of pharmaceuticals and pharmaceutical mixtures in a yeast reporter gene system. Reprod. Toxicol 22 (2), 175–185. Garbe, T.R., 2004. Co-induction of methyltransferase Rv0560c by naphthoquinones and fibric acids suggests attenuation of isoprenoid quinone action in Mycobacterium tuberculosis. Can. J. Microbiol. 50 (10), 771–778. Gros, M., Petrovic, M., Barcelo, D., 2007. Wastewater treatment plants as a pathway for aquatic contamination by pharmaceuticals in the Ebro river basin (northeast Spain). Environ. Toxicol. Chem. 26 (8), 1553–1562. Han, G.H., Hur, H.G., Kim, S.D., 2006. Ecotoxicological risk of pharmaceuticals from wastewater treatment plants in Korea: occurrence and toxicity to Daphnia magna. Environ. Toxicol. Chem. 25 (1), 265–271. Hernando, M.D., Petrovic, M., Fernańdez-Alba, A.R., Barcelo´, D., 2004. Analysis by liquid chromatography-electrospray

437

ionization tandem mass spectrometry and acute toxicity evaluation for beta-blockers and lipid-regulating agents in wastewater samples. J. Chromatogr. A 1046 (1–2), 133–140. Hernando, M.D., Aguëra, A., Fernańdez-Alba, A.R., 2007. LC-MS analysis and environmental risk of lipid regulators. Anal. Bioanal. Chem. 387 (4), 1269–1285. Holoshitz, N., Alsheikh-Ali, A.A., Karas, R.H., 2008. Relative safety of gemfibrozil and fenofibrate in the absence of concomitant cerivastatin use. Am. J. Cardiol 101 (1), 95–97. ¨ stu¨n, B., 1999. Ince, N.H., Dirilgen, N., Apikyan, I.G., Tezcanli, G., U Assessment of toxic interactions of heavy metals in binary mixtures: a statistical approach. Arch. Environ. Contam. Toxicol 36, 365–372. International Organization for Standardization, 2007. Water quality - Determination of the inhibitory effect of water samples on the light emission of Vibrio fischeri (Luminescent bacteria test), ISO 11348-3 Revised version, Geneva, Switzerland. Jia, J., Zhu, F., Ma, X., Cao, Z.W., Li, Y.X., Chen, Y.Z., 2009. Mechanisms of drug combinations: interaction and network perspectives. Nat. Rev. Drug Discov 8 (2), 111–128. Junghans, M., Backhaus, T., Faust, M., Scholze, M., Grimme, L.H., 2006. Application and validation of approaches for the predictive hazard assessment of realistic pesticide mixtures. Aquat. Toxicol 76 (2), 93–110. Kasprzyk-Hordern, B., Dinsdale, R.M., Guwy, A.J., 2009. The removal of pharmaceuticals, personal care products, endocrine disruptors and illicit drugs during wastewater treatment and its impact on the quality of receiving waters. Water Res. 43 (2), 363–380. Lambropoulou, D.A., Hernando, M.D., Konstantinou, I.K., Thurman, E.M., Ferrer, I., Albanis, T.A., Fernandez-Alba, A.R., 2008. Identification of photocatalytic degradation products of bezafibrate in TiO2 aqueous suspensions by liquid and gas chromatography. J. Chromatogr. A 1183 (1–2), 38–48. Metcalfe, C.D., Miao, X.S., Koenig, B.G., Struger, J., 2003. Distribution of acidic and neutral drugs in surface waters near sewage treatment plants in the lower Great Lakes, Canada. Environ. Toxicol. Chem. 22 (12), 2881–2889. Miller, D.B., Spence, J.D., 1998. Clinical pharmacokinetics of fibric acid derivatives (fibrates). Clin. Pharmacokinet 34 (2), 155–162. Pomati, F., Orlandi, C., Clerici, M., Luciani, F., Zuccato, E., 2008. Effects and interactions in an environmentally relevant mixture of pharmaceuticals. Toxicol. Sci. 102 (1), 129–137. Quinn, B., Gagne, F., Blaise, C., 2009. Evaluation of the acute, chronic and teratogenic effects of a mixture of eleven pharmaceuticals on the cnidarian, Hydra attenuata. Sci. Total Environ 407 (3), 1072–1079. Rodea-Palomares, I., Gonza´lez-Garcıá, C., Legane´s, F., FernańdezPinãs, F., 2009. Effect of pH, EDTA, and anions on heavy metal toxicity toward a bioluminescent cyanobacterial bioreporter. Arch. Environ. Contam. Toxicol. doi:10.1007/ s00244-008-9280-9. Rodrı´guez, A., Rosal, R., Perdigoń, J.A., Mezcua, M., Aguëra, A., Hernando, M.D., Letoń, P., Fernańdez-Alba, A.R., GarcıáCalvo, E., 2008. In: Barcelo´, D., Petrovic, M. (Eds.), Emerging Contaminants from Industrial and Municipal Waste. SpringerVerlag, Berlin, pp. 127–175. Rosal, R., Rodrı´guez, A., Perdigoń-Meloń, J.A., Mezcua, M., Hernando, M.D., Letoń, P., Garcıá-Calvo, E., Aguëra, A., Fernańdez-Alba, A.R., 2008. Removal of pharmaceuticals and kinetics of mineralization by O3/H2O2 in a biotreated municipal wastewater. Water Res. 42 (14), 3719–3728. Rosal, R., Rodea-Palomares, I., Boltes, K., Fernańdez-Pinãs, F., Legane´s, F., Gonzalo, S., Petre, A., 2009. Ecotoxicity assessment of lipid regulators in water and biologically treated wastewater using three aquatic organism. Environ. Sci. Pollut. Res. doi:10.1007/s11356-009-0137-1.

438


Scatena, R., Bottoni, P., Botta, G., Martorana, G.E., Giardina, B., 2007. The role of mitochondria in pharmacotoxicology: a reevaluation of an old, newly emerging topic. Am. J. Physiol. Cell Physiol 293 (1), C12–C21. Staels, B., Dallongeville, J., Auwerx, J., Schoonjans, K., Leitersdorf, E., Fruchart, J.C., 1998. Mechanism of action of fibrates on lipid and lipoprotein metabolism. Circulation 98 (19), 2088–2093. Stumpf, M., Ternes, T.A., Wilken, R.D., Rodrigues, S.V., Baumann, W., 1999. Polar drug residues in sewage and natural waters in the state of Rio de Janeiro. Brazil. Sci. Total Environ 225 (1–2), 135–141.

Teuschler, L.K., 2007. Deciding which chemical mixtures risk assessment methods work best for what mixtures. Toxicol. Appl. Pharmacol 223 (2), 139–147. Vighi, M., Altenburger, R., Arrhenius, A., Backhaus, T., Bodeker, W., Blanck, H., Consolaro, F., Faust, M., Finizio, A., Froehner, K., Gramatica, P., Grimme, L.H., Gronvall, F., Hamer, V., Scholze, M., Walter, H., 2003. Water quality objectives for mixtures of toxic chemicals: problems and perspectives. Ecotoxicol. Environ. Saf 54 (2), 139–150.




Toxication or detoxication? In vivo toxicity assessment of ozonation as advanced wastewater treatment with the rainbow trout Daniel Staltera,*, Axel Magdeburga, Mirco Weilb, Thomas Knackerb, Jo¨rg Oehlmanna a

Goethe University Frankfurt am Main, Biological Sciences Division, Department of Aquatic Ecotoxicology, Siesmayerstrasse 70, 60054 Frankfurt, Hessen, Germany b ECT Oekotoxikologie GmbH, Bo¨ttgerstrasse 2–4, 65439 Flo¨rsheim, Germany

article info

abstract

Article history:

Ozonation as advanced wastewater treatment method is an effective technique for

Received 23 March 2009

micropollutant removal. However, the application of this method carries the inherent


danger to produce toxic oxidation byproducts. For an ecotoxicological assessment

16 July 2009

conventionally treated wastewater, wastewater after ozonation and ozonated wastewater

Accepted 18 July 2009

after sand filtration were evaluated in parallel at an operating treatment plant via the fish


early life stage toxicity test (FELST) using rainbow trout (Oncorhynchus mykiss).

Keywords:

exposed to ozonated WW. This was accompanied by a significant decrease in body weight

Sewage treatment

and length compared to reference water, to the conventionally treated WW and to the

Pharmaceuticals

ozonated water after sand filtration. Hence sand filtration obviously prevents from adverse

Oxidation byproducts

ecotoxicological effects of ozonation.

The FELST revealed a considerable developmental retardation of test organisms

Vitellogenin

An additional test with yolk-sac larvae resulted in a significant reduction of vitellogenin

Emerging contaminants

levels in fish exposed to ozonated wastewater compared to fish reared in conventionally

Advanced oxidation process

treated wastewater. This demonstrates the effective removal of estrogenic activity by

Rainbow trout

ozonation.

Fish early life stage toxicity test

Adverse ozonation effects may have been a result of the conversion of chemicals into more toxic metabolites. However, sand filtration reduced toxication effects indicating that these oxidation byproducts are readily degradable or adsorbable. The results indicate that in any case ozonation should not be applied without subsequent post treatment appropriate for oxidation byproducts removal (e.g. sand filtration). ª 2009 Elsevier Ltd. All rights reserved.

1.

Introduction

Wastewater (WW) is one of the major sources of micropollutants introduced into the aquatic environment (Schwarzenbach et al., 2006). The large spectrum of pharmaceuticals and personal care products (PPCPs) occurring in

surface waters (Daughton and Ternes, 1999), particularly with regard to mixture toxicity, may pose a potential threat to aquatic wildlife. But also single substances may have the capability to endanger the ecosystem as for example diclofenac can lead to an impairment of the general health condition of rainbow trout at environmentally relevant concentrations

* Corresponding author. Tel.: þ49 69 7982 4882; fax: þ49 69 7982 4748. E-mail address: [email protected] (D. Stalter). 0043-1354/$ – see front matter ª 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2009.07.025

440


Nomenclature AOC C COD DOC EE2 ELISA EQS FELST FS GJIC MS222 NH4–N

assimilable organic carbon control chemical oxygen demand dissolved organic carbon 17a-ethinylestradiol enzyme linked immunosorbent assay environmental quality standard fish early life stage toxicity test wastewater after final sedimentation gap junctional intercellular communication tricaine methanesulfonate ammonium nitrogen

(Schwaiger et al., 2004; Triebskorn et al., 2004). 17a-ethinylestradiol (EE2) is considered to be the most potent estrogenic active compound for fish with the lowest observed effect concentration of 0.1 ng/L for vitellogenesis in rainbow trout (Purdom et al., 1994). Moreover it leads to a reduced egg fertilisation success and skewed sex ratio toward females of fat head minnows at a LOEC of 0.32 ng/L (Parrott and Blunt, 2005). Furthermore, industrial surfactants like alkylphenolic ethoxylates are likely to be responsible for feminization effects in rainbow trout (Jobling et al., 1996) and complex mixtures of estrogenic chemicals present in the environment have been shown to act additively (Brian et al., 2005). The concentrations of such compounds in WW often exceed effect concentrations (Ternes et al., 1999; Ying et al., 2002; Tixier et al., 2003) and thus even treated WW contains high amounts of estrogenic substances that impair the endocrine system of exposed fish (Larsson et al., 1999; Rodgers-Gray et al., 2001) resulting in feminization effects of males and consequently reduced fertility of field populations (Jobling et al., 2002). Besides, in the European Union the reduction of the contamination of surface waters with hazardous substances is defined by the Water Framework Directive requiring a ‘‘good status’’ for all coastal and inland waters until 2015 (European Commission, 2000). One tool is the implementation of environmental quality standards (EQSs) for mono substance pollutants exhibiting a significant risk to the aquatic ecosystem. The discharge of such compounds, classified as priority substances, is envisaged to be progressively reduced by 2025 or 5 years after inclusion in the list for priority substances, respectively. However, till now so called ‘‘emerging contaminants’’ are only marginally addressed and PPCPs are not included in this list. But the latter has to be reviewed at least every 4 years and provisions have been made to add several hazardous emerging contaminants. Amongst others the inclusion of diclofenac, EE2 and carbamazepine is discussed and additionally recommended by the European Parliament since 2007 (Jensen, 2007). Moreover EQSs meanwhile have been proposed for several PPCPs based on reliable ecotoxicity data (e.g. diclofenac: 0.1 mg/L, EE2: 0.03 ng/L, carbamazepine: 0.5 mg/L; Jahnel et al., 2006; Moltmann et al., 2007). Unfortunately, some PPCPs exceed the EQSs in surface waters of some urbanised

NO3–N O OECD OS P PAH PAC PPCPs rcf SPM VTG WW WWTP

nitrate nitrogen wastewater after ozonation Organisation for Economic Cooperation and Development wastewater after ozonation and sand filtration phosphate polycyclic aromatic hydrocarbons powdered activated carbon pharmaceuticals and personal care products relative centrifugal force suspended particulate matter vitellogenin wastewater wastewater treatment plant

regions. That is, inter alia, a result of the fact that conventional WW treatment plants are not designed for appropriate removal of PPCPs (Daughton and Ternes, 1999). Nevertheless in many countries surface waters serve as raw water resources for drinking water supply. Due to the public’s demand for safe drinking water and ecosystem health advanced WW treatment methods are required to allow for sufficient removal of micropollutants from sewage treatment effluents. Therefore, end of pipe techniques like activated carbon filtration and ozonation of WW might be crucial to achieve regulative requirements in a medium-term perspective. Huber et al. (2005) demonstrated that ozonation of secondary effluent is an effective tool for the removal of a wide range of pharmaceuticals among them diclofenac, carbamazepine and the estrogens estradiol, estrone and EE2, with degradation rates of more than 90% for ozone doses of 2 mg/L. Merely iodinated X-ray contrast media and a few acidic pharmaceuticals were oxidized only partially. Nakada et al. (2007) confirmed the effective elimination of 22 pharmaceuticals during ozonation in an operating treatment plant. A further advantage of the ozonation is the sanitizing property of the method (Tyrrell et al., 1995). The effective removal of micropollutants and pathogens indicates that ozonation is a suitable technique for advanced WW treatment to reduce the contamination of the aquatic environment. Actually a reduced toxicity of a mixture of six pharmaceuticals treated by ozonation towards algae and rotifer was demonstrated by Andreozzi et al. (2004). Reduced estrogenic activity of an EE2-containing WW after ozonation was shown by Huber et al. (2004) and ozone treatment of WW significantly decreased the toxic potency of urban pollutants to the freshwater mussel Elliptio complanata (Gagne et al., 2007). However, ozone treatment typically transforms chemical compounds but does not mineralize them entirely. Consequently, not only the concentrations of mother compounds but also the inclusion of transformation products and their mixtures should be assessed for toxicity, too. For a risk-benefit analysis an extensive ecotoxicological evaluation of ozonated WW is essential. In vitro bioassays covering different modes of toxic action (e.g. estrogenicity, acetylcholine esterase activity or non-specific baseline toxicity) and performed with enriched water samples are


2.2.

Table 1 – Wastewater quality summary (mg/L). SPM

COD

P

NH4–N

NO3–N

pH

0.1 0.04 0.27

5.4 10.1

8.3 8.12 8.4

Median 10% percentile 90% percentile

Final sedimentation 4.8 17 0.19 3.6 15 0.17 8.96 23.8 0.22

n

37

Median 10% percentile 90% percentile

Ozonation and 2 15 1.6 12.2 2.4 19.8

sand filtration 0.17 0.04 0.14 0.04 0.19 0.15

9.8 7.28 12.48

n

37

37

37

23

23

37

37

37

2

23

8.3 8.02 8.4 23

suitable tools for toxicity characterization of WW as shown by Escher et al. (2008). These methods were shown to be highly sensitive even for identification and characterization of low toxic water samples. Nevertheless, as typical screening tools they are not designed to replace chronic in vivo tests of whole effluent samples. To allow for a more comprehensive and integrative assessment of the potentially hazardous impact of WW ozonation we applied the fish early life stage toxicity test (FELST) with rainbow trout (Oncorhynchus mykiss).

2.

Materials and methods

Ozonated and conventionally treated WW were tested in parallel on site at the wastewater treatment plant (WWTP) Wuëri (Regensdorf, Switzerland) with the fish early life stage toxicity test (FELST) in a flow-through system.

2.1.

Characterization of the wastewater treatment plant

The municipal WWTP Wuëri operates experimentally with a full scale ozonation reactor after final sedimentation and with a sand filtration step after the ozone reactor. Table 1 shows the WW quality parameters. The water parameters after the final sedimentation and after the ozone reactor are on the same level and therefore exemplarily shown for final sedimentation water. Low ammonium and phosphate concentrations indicate that the treatment plant is already working well. The dissolved organic carbon (DOC) ranged from 5.4 to 5.9 mg/L and the pH was nearly constant in all test waters. The WWTP serves for a population equivalent of 25,000. The median discharge in the experimental period was 6190 m3/day (10th percentile: 4430 m3/day, 90th percentile: 10,500, n ¼ 109) and the applied ozone concentration was in a range between 0.4 and 1 mg O3/mg DOC.

441

Experimental setup

WW from three different sampling points of serial treatment steps was tested (Fig. 1): after the final sedimentation (FS), after the ozone reactor (O) and after additional sand filtration (OS). Test waters were passively transported through high-grade steel pipes to aerated high-grade steel reservoir tanks. The retention time in the pipes was adjusted to be at least 45 min to avoid ozone residuals reaching the exposure vessels. Indeed no ozone was detected by indigo blue method (Bader and Hoigne, 1981) in the reservoir tank during maximum required flow-through and during maximum applied ozone concentration (1 mg O3/mg DOC). From reservoir tanks test waters were transported through polytetrafluoroethylene tubes via a peristaltic pump (IPC24, Ismatec, Wertheim– Mondfeld, Germany) to the exposure vessels each equipped with a passive discharge device and tempered using a temperature-controlled water bath. Constant temperature conditions in the water bath were achieved with a flowthrough cooling system (Van der Heijden, Do¨rentrup, Germany). The flow-through rates in the exposure vessels ranged from 11 mL per minute up to 44 mL per minute (2–8 fold water exchange per day in the exposure vessels) depending on the fish size to match the loading rate criteria (OECD, 1992b). Reconstituted water according to OECD guideline 203 (1992a) was used as control water (C). The FELST was performed with the rainbow trout (Oncorhynchus mykiss) according to OECD guideline 210 (1992b) with a constant water temperature of 10 2 C as well as darkness for embryo development and 12 2 C with 12/12 h light/dark photoperiod post hatch. Sixty newly fertilized eggs per replicate were exposed to the test waters for 65 days in high-grade steel 10 L tanks. One test series was performed with unfiltered WW and a second with membrane filtered WW (pore size: 0.4 mm, Kubota Corp., Osaka, Japan) to minimize microbial impacts. Macromolecules and organic compounds were not retained. However, it has to be considered that membrane filtration additionally removes suspended particulate matter and consequently all particle bound pollutants. The filter membranes were placed in the reservoir vessels. A third test was performed with yolk-sac fry (5 days post hatch, 30 larvae per exposure vessel) and non-filtered test waters because we postulated a reduced sensitivity to pathogen contamination of the larvae compared to the egg stage. The test duration was 64 days. All tests were performed with undiluted WW to increase the probability to detect differences between the treatment groups. With the beginning of swim up (the swim up process marks a developmental transition from larval stage to juvenile fish stage and is characterized with the

Fig. 1 – Sampling points at the wastewater treatment plant. Abbreviations: FS, after final sedimentation; O, after the ozone reactor; OS, after sand filtration.

442


beginning of exogenous ingestion) the fish were fed four times per day (trout starter, 4% body weight per day). The tests were run with four (C) and three (FS, O, OS) replicates in the first FELST with unfiltered WW and two (C, OS) and three replicates (FS, O) per test water in the remaining two tests. The latter were performed in parallel and therefore it was not possible to use consistently three replicates per WW-group due to limited space capacities. In each test replicates were placed randomized in the water bath. Observations on egg coagulation, hatching, mortality, swim up, malformation and abnormal behaviour were recorded daily. Fish were humanely killed by MS222 (tricaine methanesulfonate, Sigma– Aldrich, St. Louis, USA) overdose. Individual fish were blotted dry, weight and body length were determined. Afterwards fish were frozen in liquid nitrogen and stored at 80 C until vitellogenin detection in whole body homogenates.

2.3.

Vitellogenin detection

Whole body homogenates of 11 fish (from the third test with yolk-sac fry) per replicate were prepared as described by Holbech et al. (2006) with slight modifications. Aliquots of 0.3 g frozen fish, excised between head and pectoral fin, were mixed with 10 fold of the body weight of homogenisation buffer (50 mM Tris–HCl pH 7.4; 1% protease inhibitor cocktail (P 8340, Sigma–Aldrich, St. Louis, USA)) and homogenated with a dispersing apparatus (T18 basic Ultra–Turrax, IKA, Staufen, Germany). The homogenate was centrifuged for 30 min at 20,000 rcf and the supernatant was used for vitellogenin analysis. Vitellogenin (VTG) was detected with a rainbow trout vitellogenin ELISA test kit (Biosense, Bergen, Norway) using a 1:20 dilution.

2.4.

Statistical analysis

Complete statistical analyses (Kruskal–Wallis with Dunn’s post test, Fisher’s exact test, non-linear regression with variable slope) were performed using GraphPad Prism version 5.0 for windows (GraphPad software, San Diego, CA, USA). All given error values indicate the standard error (SE) and in all figures error bars display the SE. Kruskal–Wallis with Dunn’s post test was chosen to test on significant differences because data were not normally distributed in all cases. For quantal data Fisher’s exact test was applied (mortality, swim up, coagulation). When toxicity endpoints could be analysed individually for each test animal (biomass, body length, vitellogenin), the collected data are presented and statistically evaluated on a per specimen basis.

3.

Results

3.1.

FELST with unfiltered wastewater

Unfiltered WW caused an increased coagulation rate of the exposed eggs in the FELST (Fig. 2). The eggs exposed to WW after final sedimentation (FS) were completely coagulated after 18 days with a 50% coagulation time of 12.1 days. The coagulation of the eggs exposed to WW after the ozone reactor (O) was considerably delayed compared to FS. However the

Fig. 2 – Cumulative coagulation of Oncorhynchus mykiss eggs (mean values ± SE) exposed to differently treated wastewaters. Abbreviations: C, control water; FS, final sedimentation; O, ozonation; OS, ozonation and sand filtration; SE, standard error. Time scale is logarithmized. After 40 days all treatment groups differ significantly from each other (Fisher’s exact test: p < 0.001; n [ 240 (C), 180 (FS, O, OS)).

coagulation after 40 days achieves still an average rate of 87.2 4.8% with a 50% coagulation time of 17.6 days. The lowest coagulation rate in the WW treatments occurred after sand filtration (OS; 64.4 4.0% after 40 days; 50% coagulation time: 25.8 days). After 10–15 days exposure, fungus mycelia (first appearing in the FS vessels) were observed in all WW exposure vessels. Mycelia were found on and between eggs as well as vorticellas on the eggs whereas the reference water (C) remained observably free from mycelia and vorticellas. Hatching success in the control group (53.4 5.2%) did not meet validity criteria (>66%; OECD guideline 210).

3.2.

FELST with membrane filtered wastewater

To exclude microbial impairment the second FELST was performed with membrane filtered WW and eggs were obtained from another fish hatchery. All test vessels remained free from microbial contamination throughout the test duration. The egg coagulation rates in the WW treatment groups of this experimental series were reduced compared to the first FELST with a maximum of 25% (O, OS) and only 20.1 1.1% after FS (Fig. 3A). Egg coagulation was significantly increased and hatching success significantly decreased in the WW treatment groups compared to the control ( p < 0.05, Fisher’s exact). The coagulation rates were slightly but not significantly increased in O and OS compared to FS (Fig. 3A). The hatching progress was slightly delayed in O compared to FS and OS but hatching success achieved at least 75% in all treatments (Fig. 3B). The control group met the validity criteria according to OECD guideline 210 (egg coagulation: 10.0 1.7%, hatching success: 90.0 1.7%).


443

Fig. 3 – Cumulative coagulation of eggs (A) and cumulative hatching of larvae (B) from Oncorhynchus mykiss (mean values ± SE) exposed to differently treated and membrane filtered wastewaters. Abbreviations: C, control water; FS, final sedimentation; O, ozonation; OS, ozonation and sand filtration; SE, standard error. Time scale is logarithmized. After 40 and 36 days, respectively all wastewater treatment groups are significantly different from the control (Fisher’s exact test: p < 0.01–0.05; n [ 120 (C, OS), 180 (FS, O)).

Fig. 4 shows the cumulative swim up of hatched fish beginning after 45 days of exposure. The swim up is considerably delayed in all WW treatment groups compared to the control. This effect is most notable in O, even if compared to

FS and OS. At the end of the experiment only 83.3 1.6% of the fish swam up in O while 100% swam up in C and FS and 97.6 2.4% in OS. Hereby the swim up success after 64 days is significantly decreased in O compared to the other treatment groups ( p < 0.01, Fisher’s exact). The 50% swim up time in the control is 47.4 1.0 days, which is only slightly increased in FS and OS (50.2 1.0; 49.3 1.0) but obviously increased in O (57.5 1.04). The biomass as well as the body length of fish is significantly decreased ( p < 0.001, Kruskal–Wallis with Dunn’s post test) in all WW treatments compared to the control (Fig. 5). Both endpoints in fish exposed to O are furthermore significantly decreased compared to the FS group ( p < 0.001) and significantly decreased compared to OS ( p < 0.05). Generally the mortality is comparatively low in all WW treatment groups (Fig. 6), as they are fulfilling the validity criteria for controls according to OECD guideline 210 (survival after hatch 70%). Nevertheless the mortality in the WW treatment groups is slightly increased compared to C. The highest and significantly increased mortality rate was detected in the ozonated water (24.0 4.0%, p < 0.05, Fisher’s exact).

3.3. Fig. 4 – Cumulative swim up of Oncorhynchus mykiss larvae (mean values ± SE) exposed to differently treated and membrane filtered wastewaters. Abbreviations: C, control water; FS, final sedimentation; O, ozonation; OS, ozonation and sand filtration; SE, standard error. Time scale is logarithmized. After 64 days the O group differs significantly from the other treatment groups (Fisher’s exact test: p < 0.001–0.01; n [ 95 (C), 116 (FS), 103 (O), 75 (OS)).

Fish test starting with yolk-sac fry & vitellogenesis

The fish test starting with the yolk-sac stage revealed no statistically significant differences in development between WW treatments and the control. The biomass exhibited no major deviations between exposure groups (66%) according to OECD guideline (1992b) indicating that egg quality was not sufficient to run a valid test. The disinfectant effect of ozonation (Tyrrell et al., 1995) is likely to be responsible for the delayed egg

Unfiltered WW led to an increased coagulation rate of the exposed eggs in the FELST (Fig. 2). High coagulation rates

Fig. 6 – Mortality of Oncorhynchus mykiss post hatch (mean values ± SE) exposed to differently treated and membrane filtered wastewaters. Abbreviations: C, control water; FS, final sedimentation; O, ozonation; OS, ozonation and sand filtration; SE, standard error. Significant differences to control are indicated with white asterisks (Fisher’s exact test: +, p < 0.05; n [ 108 (C), 143 (FS), 135 (O), 90 (OS)).

Fig. 7 – Whole body vitellogenin concentration of Oncorhynchus mykiss (mean values ± SE) after 60 days exposure to differently treated wastewaters starting with the yolk-sac stage. Abbreviations: FS, final sedimentation; O, ozonation; OS, ozonation and sand filtration; C, control water; SE, standard error. Significant differences to control are indicated with white asterisks, between treatments with black asterisks (Kruskal–Wallis with Dunn’s post test: +, p < 0.05; ++, p < 0.01; n [ 22 (C, OS) – 33 (FS, O)).


coagulation in the O group compared to FS. Ozonation is also known to produce high amounts of assimilable organic carbons (AOC) which may have allowed a fast regrowth of microorganisms (Huang et al., 2005), resulting in still high coagulation rates in the O group. Sand filtration reduces the amount of suspended particulate matter (SPM) and of AOC (Wang and Summers, 1996). This may have reduced microbial development and resulting coagulation rates.

4.2.

FELST with membrane filtered wastewater

Exposure to membrane filtered WW revealed a distinct delay of the swim up process in the O group compared to FS and OS (Fig. 4) accompanied by a significant decrease in body weight and body length (Fig. 5). The reduced biomass and body length after ozonation is most likely a result of the delayed development because fish start exogenous ingestion at the onset of the swim up process. Consequently trout that swim up later may suffer from developmental disadvantages. Oxidative byproducts in ozonated WW may have impeded embryonic and/or larval development of fish. Possibly because sand filtration is able to remove ozonation metabolites (e.g. aldehydes, glyoxal, AOC; Wang and Summers, 1996) the effect in OS was reduced. However, no single compound could be clearly identified for the retardation effect. Possibly the sum of aldehydes (e.g. formaldehyde, glyoxal, acetaldehyde), carboxylic acids (e.g. formate), ketones and brominated organic compounds formed due to ozonation (Huang et al., 2005; Wert et al., 2007) caused these effects. Unfortunately, no chronic toxicity data for juvenile rainbow trout are available in literature for these compounds. Nevertheless, Wang and Summers (1996) were able to demonstrate that sand filtration efficiently removes aldehydes produced during ozone application, supporting this assumption. Besides, the formation of more complex organic metabolites evoked by ozonation may result in an increased toxicity compared to chemical precursors as it has already been documented for polycyclic aromatic hydrocarbons (PAHs). In the experiments of Luster-Teasley et al. (2002, 2005) chrysene and pyrene and byproducts as a result of ozonation were examined for their ability to disrupt gap junctional intercellular communication (GJIC), an indicator for tumor promoting properties. Among the transformation products, aldehydic compounds exhibited an increased toxicity compared to the precursor substance while their carboxylic structure analogue showed no GJIC disrupting activity. For the antiepileptic drug carbamazepine McDowell et al. (2005) identified three new oxidation products of unknown toxicity after ozone treatment. Furthermore, a possible increase in mutagenicity after ozonation of WW, as observed by Monarca et al. (2000) with the Ames test, verifies the potential of ozonation to produce toxic oxidation byproducts. In this context conversion of the widely used fungicide tolylfluanide into the carcinogen N-nitrosodimethylamine during ozonation was discovered by Schmidt and Brauch (2008). Based on these results it is likely that ozonation of WW leads to an increased number of unknown and potentially toxic metabolites depending on the composition of WW at the outset and the post treatment (e.g. sand filtration). Petala et al. (2006) demonstrated with the Vibrio fischeri bioluminescence test,

445

that toxicity originating from ozonated WW decreases with increasing storage time, indicating a rapid decomposition of toxic metabolites. In spite of the developmental retardation and the reduced biomass no significant differences in mortality rates were observed between the WW treatment groups (Fig. 6). Mortality was highest directly after ozonation (O, 24.0 4.0%). Considering the developmental retardation and the decreased biomass of fish exposed to ozonated water the increased mortality of O group specimens compared to FS and OS is noticeable and, although not significant, potentially a result of a reduced fitness of the fish in the ozonated water. It might be assumed that the retarded development is a consequence of an unspecific and general impairment of the fish’s health condition. This may increase their sensitivity towards environmental and anthropogenic stressors, thus leading to an increased mortality. Furthermore, under field conditions the delayed development possibly increases the risk for the fish to fall prey to predators since before swim up the larvae rest on the bottom, not capable to abscond effectively. Based on the assumption that the compounds causing adverse effects after ozonation are readily degradable it is possible that the detoxication after the ozone reactor will occur in the river as well. However, the discharge of ozonated WW without sand filtration would bear the risk to endanger fish populations in a considerable range of the receiving river, depending on WW load and flow velocity. Therefore, ozonation should always be followed by a sand filtration and further studies should focus on whether sand filtration is capable to remove oxidation byproducts sufficiently.

4.3.

Fish test starting with yolk-sac fry & vitellogenesis

The fish test with non-filtered WW starting with the yolk-sac stage resulted in negligible and insignificant deviations of development and biomass between treatment groups. Ozonation had no effect on these toxicity endpoints. These results indicate that yolk-sac fry 5 days post hatch are less sensitive to ozonation metabolites, compared to fish embryos and newly hatched fish, respectively. Nevertheless, the amount of suspended particulate matter in the exposure vessels and in the Teflon tubes was accompanied by an increased biofilm formation. This may have contributed to an increased detoxication of the ozonated WW due to (bio-) degradation of oxidation byproducts. The significant increase of vitellogenin content in fish exposed to FS compared to the reference water indicates an environmentally relevant contamination of the WW with estrogenic active compounds (Fig. 7). After ozonation the VTG content decreased even below the control level. The formation of antiestrogenic compounds is not likely because the VTG decrease compared to control level is not significant and phenols, as an important functional group interacting with the estrogen receptor (Nishihara et al., 2000), are particularly susceptible to ozone attack (von Gunten, 2003). These results confirm the high efficiency of ozonation to eliminate estrogenic contamination in WW as already shown by Huber et al. (2004) with an in vitro test system. Based on these results it can be assumed that ozonation is well suited to reduce the estrogenic burden of WW below environmental relevance.

446


Regarding ecosystem health it should be considered that estrogen and xenoestrogen concentrations detected in the environment have been shown to impair the sustainability of wild fish populations (Kidd et al., 2007), arguing for the establishment of an end of pipe technique for estrogen removal to protect fish populations in surface waters receiving high amounts of WW.

4.4.

Ozonation vs. alternatives

Currently there is a controversy regarding the appropriate advanced WW treatment techniques to reduce contamination of the aquatic ecosystem with micropollutants. Alternative processes other than ozonation also deliver promising results with regard to the removal of contaminants. Wastewater treatment with powdered activated carbon (PAC) achieves reduction rates of 75–90% for pharmaceuticals (including X-ray contrast media) with PAC dosages of 10–20 mg/L (Pu¨ttmann et al., 2008). The main advantage of PAC treatment is that a reduced chemical concentration is equivalent to removal of chemicals whereas ozonation leads basically to a transformation of the compounds. According to cost estimations PAC treatment is expected to be about 30% more expensive than ozonation (Joss et al., 2008). However, the main disadvantage of PAC treatment might be the need for the disposal of the used and contaminated carbon. Membrane filtration is suitable for micropollutants retention but as a result of considerably higher requirement for energy and technical equipment economically not competitive with ozonation or activated carbon treatment (Joss et al., 2008). Overall, end of pipe techniques are presumably a suitable solution to reduce toxicity of hazardous WW in a mediumterm perspective. Nevertheless in the long term source control strategies such as wastewater separation (e.g. urine separation), ecologically correct disposal of drugs by end users, reuse or recycling by the pharmaceutical industry or alternative medical treatments to drug therapies (Daughton, 2003; Joss et al., 2006) could offer environmentally friendly and affordable options.

5.

Conclusions

The sanitizing effect of ozonation was confirmed. Coagulation rates of eggs exposed to ozonated wastewater were on a lower level compared to those exposed to conventionally treated wastewater. Disinfection is presumably only efficient when linked to sand filtration because of rapid recovery of microorganisms due to increased assimilable organic carbon formation as a result of ozonation. Membrane filtered wastewater (for removal of microorganisms) reveals developmental retardation directly after ozonation. After sand filtration this adverse effect disappears. The reduced biomass and body length in fish exposed to ozonated wastewater is most probably a result of the formation of toxic oxidation byproducts. The mortality in the ozonated wastewater was significantly increased as a result of retarded development. Impairment of the fish’s health condition may increase the sensitivity towards environmental and anthropogenic stressors.

Developmental retardation might increase the risk for the fish to fall prey to predators because swim up stage is delayed. Ozonation should not be applied without appropriate barrier for oxidation byproducts. Effectiveness of sand filtration for removal of oxidation byproducts should be further evaluated. Reduction of vitellogenin content in fish exposed to ozonated wastewater on control level confirms the suitability of this technique to reduce estrogenic activity, possibly below environmental relevance.

Acknowledgements The authors should like to express their gratitude to the staff from WWTP Wuëri for their technical cooperation and Adriano Joss from the EAWAG for helpful suggestions and technical support. Ulrike Schulte-Oehlmann is acknowledged for critical comments and helpful suggestions on the manuscript. This study was part of the EU project Neptune (contract no 036845, SUSTDEV-2005-3.II.3.2) within the Energy, Global Change and Ecosystems Programme of the Sixth Framework (FP6-2005-Global-4) and co-funded by Bundesamt fu¨r Umwelt (BAFU), Bern (CH) within the Strategy MicroPoll Programme (contract no 05.0013.PJ/F471-0916).

references

Andreozzi, R., Campanella, L., Fraysse, B., Garric, J., Gonnella, A., Lo Giudice, R., Marotta, R., Pinto, G., Pollio, A., 2004. Effects of advanced oxidation processes (AOPs) on the toxicity of a mixture of pharmaceuticals. Water Science and Technology 50 (5), 23–28. Bader, H., Hoigne, J., 1981. Determination of ozone in water by the indigo method. Water Research 15 (4), 449–456. Brian, J.V., Harris, C.A., Scholze, M., Backhaus, T., Booy, P., Lamoree, M., Pojana, G., Jonkers, N., Runnalls, T., Bonfa, A., Marcomini, A., Sumpter, J.P., 2005. Accurate prediction of the response of freshwater fish to a mixture of estrogenic chemicals. Environmental Health Perspectives 113 (6), 721–728. Daughton, C.G., 2003. Cradle-to-cradle stewardship of drugs for minimizing their environmental disposition while promoting human health. II. Drug disposal, waste reduction, and future directions. Environmental Health Perspectives 111 (5), 775–785. Daughton, C.G., Ternes, T.A., 1999. Pharmaceuticals and personal care products in the environment: agents of subtle change? Environmental Health Perspectives 107, 907–938. Escher, B.I., Bramaz, N., Quayle, P., Rutishauser, S., Vermeirssen, E.L.M., 2008. Monitoring of the ecotoxicological hazard potential by polar organic micropollutants in sewage treatment plants and surface waters using a mode-of-action based test battery. Journal of Environmental Monitoring 10 (5), 622–631. European Commission, 2000. Directive 2000/60/EC of the European Parliament and of the Council of 23 October 2000 establishing a framework for Community action in the field of water policy. Official Journal of the European Communities L327. Retrieved from: http://eur-lex.europa.eu/


LexUriServ/LexUriServ.do?uri¼OJ: L:2000:327:0001:0072:EN: PDF [12.02.2009]. Gagne, F., Andre, C., Cejka, P., Gagnon, C., Blaise, C., 2007. Toxicological effects of primary-treated urban wastewaters, before and after ozone treatment, on freshwater mussels (Elliptio complanata). Comparative Biochemistry and Physiology C – Toxicology & Pharmacology 145 (4), 542–552. Holbech, H., Kinnberg, K., Petersen, G.I., Jackson, P., Hylland, K., Norrgren, L., Bjerregaard, P., 2006. Detection of endocrine disrupters: evaluation of a fish sexual development test (FSDT). Comparative Biochemistry and Physiology C – Toxicology & Pharmacology 144 (1), 57–66. Huang, W.J., Fang, G.C., Wang, C.C., 2005. The determination and fate of disinfection by-products from ozonation of polluted raw water. Science of the Total Environment 345 (1-3), 261–272. Huber, M.M., Ternes, T.A., von Gunten, U., 2004. Removal of estrogenic activity and formation of oxidation products during ozonation of 17a-ethinylestradiol. Environmental Science & Technology 38 (19), 5177–5186. Huber, M.M., Gobel, A., Joss, A., Hermann, N., Loffler, D., McArdell, C.S., Ried, A., Siegrist, H., Ternes, T.A., von Gunten, U., 2005. Oxidation of pharmaceuticals during ozonation of municipal wastewater effluents: a pilot study. Environmental Science & Technology 39 (11), 4290–4299. Jahnel, J., Neamtu, M., Schudoma, D., Frimmel, F.H., 2006. Scientific risk assessment of considered water relevant substances (Bestimmung von Umweltqualitaetsnormen fuer potenziell gewaesserrelevante Stoffe). Acta Hydrochimica et Hydrobiologica 34 (4), 389–397. Jensen, G.K., 2007. European Parliament increases number of chemicals earmarked for priority action by European Commission. The Lancet Oncology 8 (5), 382–383. Jobling, S., Sheahan, D., Osborne, J.A., Matthiessen, P., Sumpter, J.P., 1996. Inhibition of testicular growth in rainbow trout (Oncorhynchus mykiss) exposed to estrogenic alkylphenolic chemicals. Environmental Toxicology and Chemistry 15 (2), 194–202. Jobling, S., Coey, S., Whitmore, J.G., Kime, D.E., Van Look, K.J.W., McAllister, B.G., Beresford, N., Henshaw, A.C., Brighty, G., Tyler, C.R., Sumpter, J.P., 2002. Wild intersex roach (Rutilus rutilus) have reduced fertility. Biology of Reproduction 67 (2), 515–524. Joss, A., Klaschka, U., Knacker, T., Liebig, M., Lienert, J., Ternes, T.A., Wennmalm, A., 2006. Source control, source separation. In: Ternes, T.A., Joss, A. (Eds.), Human Pharmaceuticals, Hormones and Fragrances. The Challenge of Micropollutants in Urban Water Management. ISBN 1843390930, pp. 353–384. Joss, A., Siegrist, H., Ternes, T.A., 2008. Are we about to upgrade wastewater treatment for removing organic micropollutants? Water Science and Technology 57 (2), 251–255. Kidd, K.A., Blanchfield, P.J., Mills, K.H., Palace, V.P., Evans, R.E., Lazorchak, J.M., Flick, R.W. (2007). Collapse of a fish population after exposure to a synthetic estrogen. Proceedings of the National Academy of Sciences of the United States of America 104(21), 8897–8901. Larsson, D.G.J., Adolfsson-Erici, M., Parkkonen, J., Pettersson, M., Berg, A.H., Olsson, P.E., Forlin, L., 1999. Ethinylestradiol – an undesired fish contraceptive? Aquatic Toxicology 45 (2-3), 91–97. Luster-Teasley, S.L., Yao, J.J., Herner, H.H., Trosko, J.E., Masten, S.J., 2002. Ozonation of chrysene: evaluation of byproduct mixtures and identification of toxic constituent. Environmental Science & Technology 36 (5), 869–876. Luster-Teasley, S.L., Ganey, P.E., DiOrio, M., Ward, J.S., Maleczka, R.E., Trosko, J.E., Masten, S.J., 2005. Effect of byproducts from the ozonation of pyrene: biphenyl-2,20 ,6,60 tetracarbaldehyde and biphenyl-2,20 ,6,60 -tetracarboxylic acid

447

on gap junction intercellular communication and neutrophil function. Environmental Toxicology and Chemistry 24 (3), 733–740. McDowell, D.C., Huber, M.M., Wagner, M., Von Gunten, U., Ternes, T.A., 2005. Ozonation of carbamazepine in drinking water: identification and kinetic study of major oxidation products. Environmental Science & Technology 39 (20), 8014–8022. Moltmann, J.F., Liebig, M., Knacker, T., Keller, M., Scheurer, M., Ternes, T., 2007. Relevance of Endocrine Disrupting Substances and Pharmaceuticals in Surface Waters (Gewa¨sserelevanz endokriner Stoffe und Arzneimittel). Final report. Federal Environment Agency. Funding code 20524205. Monarca, S., Feretti, D., Collivignarelli, C., Guzzella, L., Zerbini, I., Bertanza, G., Pedrazzani, R., 2000. The influence of different disinfectants on mutagenicity and toxicity of urban wastewater. Water Research 34 (17), 4261–4269. Nakada, N., Shinohara, H., Murata, A., Kiri, K., Managaki, S., Sato, N., Takada, H., 2007. Removal of selected pharmaceuticals and personal care products (PPCPs) and endocrine-disrupting chemicals (EDCs) during sand filtration and ozonation at a municipal sewage treatment plant. Water Research 41 (19), 4373–4382. Nishihara, T., Nishikawa, J., Kanayama, T., Dakeyama, F., Saito, K., Imagawa, M., Takatori, S., Kitagawa, Y., Hori, S., Utsumi, H., 2000. Estrogenic activities of 517 chemicals by yeast two-hybrid assay. Journal of Health Science 46 (4), 282–298. OECD, 1992a. Fish Acute Toxicity Test. OECD Guidelines for the Testing of Chemicals. No. 203. Organisation for Economic Cooperation and Development, Paris. OECD, 1992b. Fish Early Life Stage Toxicity Test. OECD Guidelines for the Testing of Chemicals. No. 210. Organisation for Economic Cooperation and Development, Paris. Parrott, J.L., Blunt, B.R., 2005. Life-cycle exposure of fathead minnows (Pimephales promelas) to an ethinylestradiol concentration below 1 ng/L reduces egg fertilization success and demasculinizes males. Environmental Toxicology 20 (2), 131–141. Petala, M., Samaras, P., Zouboulis, A., Kungolos, A., Salkellaropoulos, G., 2006. Ecotoxicological properties of wastewater treated using tertiary methods. Environmental Toxicology 21 (4), 417–424. Pu¨ttmann, W., Keil, F., Oehlmann, J., Schulte-Oehlmann, U., 2008. Strategy to reduce pharmaceuticals in drinking water – technical approach (Wassertechnische Strategien zur Reduzierung der Trinkwasserbelastung durch Arzneimittelwirkstoffe). Umweltwissenschaften und Schadstoff-Forschung 20 (3), 209–226. Purdom, C.E., Hardiman, P.A., Bye, V.V.J., Eno, N.C., Tyler, C.R., Sumpter, J.P., 1994. Estrogenic effects of effluents from sewage treatment works. Chemical Ecology 8 (4), 275–285. Rodgers-Gray, T.P., Jobling, S., Kelly, C., Morris, S., Brighty, G., Waldock, M.J., Sumpter, J.P., Tyler, C.R., 2001. Exposure of juvenile roach (Rutilus rutilus) to treated sewage effluent induces dose-dependent and persistent disruption in gonadal duct development. Environmental Science & Technology 35 (3), 462–470. Schmidt, C.K., Brauch, H.J., 2008. N, N-dimethosulfamide as precursor for N-nitrosodimethylamine (NDMA) formation upon ozonation and its fate during drinking water treatment. Environmental Science & Technology 42 (17), 6340–6346. Schwaiger, J., Ferling, H., Mallow, U., Wintermayr, H., Negele, R.D., 2004. Toxic effects of the non-steroidal antiinflammatory drug diclofenac Part 1: histopathological alterations and bioaccumulation in rainbow trout. Aquatic Toxicology 68 (2), 141–150. Schwarzenbach, R.P., Escher, B.I., Fenner, K., Hofstetter, T.B., Johnson, C.A., von Gunten, U., Wehrli, B., 2006. The challenge

448


of micropollutants in aquatic systems. Science 313 (5790), 1072–1077. Ternes, T.A., Stumpf, M., Mu¨ller, J., Haberer, K., Wilken, R.D., Servos, M., 1999. Behavior and occurrence of estrogens in municipal sewage treatment plants – I. Investigations in Germany, Canada and Brazil. Science of the Total Environment 225 (1-2), 81–90. Tixier, C., Singer, H.P., Oellers, S., Mu¨ller, S.R., 2003. Occurrence and fate of carbamazepine, clofibric acid, diclofenac, ibuprofen, ketoprofen, and naproxen in surface waters. Environmental Science & Technology 37 (6), 1061–1068. Triebskorn, R., Casper, H., Heyd, A., Eikemper, R., Ko¨hler, H.R., Schwaiger, J., 2004. Toxic effects of the non-steroidal antiinflammatory drug diclofenac Part II. Cytological effects in liver, kidney, gills and intestine of rainbow trout (Oncorhynchus mykiss). Aquatic Toxicology 68 (2), 151–166.

Tyrrell, S.A., Rippey, S.R., Watkins, W.D., 1995. Inactivation of bacterial and viral indicators in secondary sewage effluents, using chlorine and ozone. Water Research 29 (11), 2483–2490. von Gunten, U., 2003. Ozonation of drinking water: Part I. Oxidation kinetics and product formation. Water Research 37 (7), 1443–1467. Wang, J.Z., Summers, R.S., 1996. Biodegradation behavior of ozonated natural organic matter in sand filters. Revue des Sciences de l’Eau 9 (1), 3–16. Wert, E.C., Rosario-Ortiz, F.L., Drury, D.D., Snyder, S.A., 2007. Formation of oxidation byproducts from ozonation of wastewater. Water Research 41 (7), 1481–1490. Ying, G.G., Williams, B., Kookana, R., 2002. Environmental fate of alkylphenols and alkylphenol ethoxylates – a review. Environment International 28 (3), 215–226.




The role of organic matter in the removal of emerging trace organic chemicals during managed aquifer recharge T. Rauch-Williams, C. Hoppe-Jones, J.E. Drewes* Advanced Water Technology Center (AQWATEC), Colorado School of Mines, Environmental Science and Engineering Division, Golden, CO 80401-1887, United States

article info

abstract

Article history:

This study explored the effect of different bulk organic carbon matrices on the fate of trace

Received 20 April 2009

organic chemicals (TOrC) during managed aquifer recharge (MAR). Infiltration through


porous media was simulated in biologically active column experiments under aerobic and

11 August 2009

anoxic recharge conditions. Wastewater effluent derived organic carbon types, differing in

Accepted 20 August 2009

hydrophobicity and biodegradability (i. e., hydrophobic acids, hydrophilic carbon, organic

Available online 27 August 2009

colloids), were used as feed substrates in the column experiments. These carbon substrates while fed at the same concentration differed in their ability to support soil biomass growth

Keywords:

during porous media infiltration. Removal of degradable TOrC (with the exception of

Trace organic chemicals (TOrC)

diclofenac and propyphenazone) was equal or better under aerobic versus anoxic porous

Groundwater recharge

media infiltration conditions. During the initial phase of infiltration, the presence of

Effluent organic matter

biodegradable organic carbon (BDOC) enhanced the decay of degradable TOrC by

Managed aquifer recharge

promoting soil biomass growth, suggesting that BDOC served as a co-substrate in a co-

Riverbank filtration

metabolic transformation of these contaminants. However, unexpected high removal

Biotransformation

efficiencies were observed for all degradable TOrC in the presence of low BDOC concen-

Co-metabolism

trations under well adopted oligotrophic conditions. It is hypothesized that removal under

Primary substrate

these conditions is caused by a specialized microbial community growing on refractory carbon substrates such as hydrophobic acids. Findings of this study reveal that the concentration and character of bulk organic carbon present in effluents affect the degradation efficiency for TOrC during recharge operation. Specifically aerobic, oligotrophic microbiological soil environments present favorable conditions for the transformation of TOrC, including rather recalcitrant compounds such as chlorinated flame retardants. ª 2009 Elsevier Ltd. All rights reserved.

1.

Introduction

Managed aquifer recharge (MAR) systems, such as riverbank filtration (RBF) and soil aquifer treatment (SAT), are widely used natural processes for drinking water augmentation projects using source water that might be impaired by wastewater discharge. Previous studies have demonstrated that MAR systems are effective in dampening and reducing

the concentrations of dissolved organic carbon (DOC) as well as various trace organic contaminants (TOrC) that might be present in impaired source waters (Drewes and Fox, 1999; Brauch et al., 2000; Gru¨nheid et al., 2005). The presence of TOrC has become a key concern for drinking water augmentation projects during the past decade (Kolpin et al., 2002; Heberer, 2002; Focazio et al., 2008). Although adverse human health effects caused by these compounds at concentrations

* Corresponding author. E-mail address: [email protected] (J.E. Drewes). 0043-1354/$ – see front matter ª 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2009.08.027

450


commonly observed in impaired water resources are very unlikely (Schwab et al., 2005), minimizing exposure of wastewater derived contaminants in these projects is desired. Previous research on the fate of TOrC during MAR has primarily focused on collecting anecdotal and site specific information on their occurrence and removal (Drewes et al., 2002; Montgomery-Brown et al., 2003; Gru¨nheid et al., 2005; Dillon et al., 2008). Studies delineating the mechanisms and boundary conditions for the transformation of wastewater derived TOrC during MAR are lacking. Previous studies demonstrated that the type and bioavailability of effluent organic matter (EfOM) controls the extent of soil biomass growth in MAR systems (Rauch and Drewes, 2004; Rauch and Drewes, 2005). EfOM may consequently effect the degradation of TOrC by serving as a co-substrate in microbiologically facilitated transformations (Stratton et al., 1983). The diversity and expression of the soil microbial community also depends on composition and concentration of the organic carbon substrate controlling trophic cycles in the subsurface (Preuß and Nehrkorn, 1996; Szewzyk et al., 1998). The composition of EfOM (i.e. in terms of its bioavailability) is primarily determined by the degree of wastewater treatment employed (Drewes and Fox, 1999), which can vary widely from primary to conventional to advanced wastewater treatment. As a result, effluent qualities fed to MAR systems can vary in biodegradable dissolved organic carbon (BDOC) concentrations from less than 1 up to 15 mg/L or more. As a consequence, soil microbial communities growing on different levels of BDOC can differ widely in total biomass and diversity. The objectives of this research were to investigate the role of 1) abiotic vs. biotic conditions, 2) BDOC and 3) the type of organic carbon matrices on the removal of select TOrC, such as pharmaceutical residues, personal care products, and household chemicals, during MAR.

2.

Methodology

2.1.

Target organic contaminants

Compounds selected for this study represent small molecular weight organic chemicals (180 to 360 Dalton) that are hydrophilic at neutral pH regimes as indicated by an octanol/water partition coefficient at pH 7 (log DpH¼7 of less than 2.6). TOrC with these properties have a high potential to migrate into groundwater and are not expected to be adsorbed onto porous media. The molecular structures and physicochemical properties of the target compounds chosen for this study are presented in Table 1. These compounds cover a wide range of biodegradability as previously reported for soil/water systems. The anticonvulsants carbamazepine and primidone have been classified as recalcitrant during wastewater treatment and MAR in earlier studies (Heberer, 2002; Drewes et al., 2003; Clara et al., 2004). The chlorinated phosphate esters tris (1-chloroisopropyl)-phosphate (TCPP) and tris(2-chloroethyl)phosphate (TCEP) are two widely used flame retardants and are persistent in the aquatic environment (Heberer et al., 2001; Fries and Pu¨ttmann, 2003). During bank filtration in Germany, however, TCPP and TCEP exhibited a significant reduction, which was attributed to biotransformation in the aquifer

(Heberer et al., 2003). Amy and Drewes (2006) also reported removal of TCEP to concentrations below the detection limit after two years of subsurface travel in an MAR facility supporting that chlorinated flame retardants can be biotransformed under anoxic conditions. Propyphenazone is a poorly biodegradable analgesic that persists during RBF (Heberer et al., 2003) and SAT (Drewes et al., 2003). For diclofenac, a popular analgesic drug, low removal due to biodegradation or adsorption was reported (Buser et al., 1998; Mo¨hle et al., 1999) unless soils contain a high organic carbon content (Drillia et al., 2003). Several studies report a faster degradation of diclofenac under anoxic conditions as compared to aerobic conditions (Zwiener and Frimmel, 2003; Hua et al., 2003). Opposing results were reported by Schmidt et al. (2004) in that diclofenac was almost completely removed during aerobic bank filtration but recalcitrant during anaerobic recharge. Ibuprofen, ketoprofen, and naproxen are common analgesics that are well degradable during wastewater treatment (Buser et al., 1999; Zwiener and Frimmel, 2003; Carballa et al., 2004) and during soil infiltration (Sedlak and Pinkston, 2001; Drewes et al., 2003). Gemfibrozil is a commonly prescribed blood lipid regulator found in unconfined shallow aquifers impacted by wastewater infiltration in the low to moderate ng/L-concentration range (Heberer and Stan, 1997; Drewes and Shore, 2001; Heberer, 2002). Gemfibrozil was removed below the limit of detection within a few weeks during groundwater recharge using SAT (Drewes et al., 2003).

2.2.

Analytical methods

2.2.1.

GC/MS analysis

The following TOrC were analyzed by gas chromatography coupled with mass spectrometry (GC/MS) using a HP 6890 gas chromatograph and a HP 5973 quadrupole mass spectrometer from Agilent Technologies (Waldbronn, Germany) adopting a method published by Reddersen and Heberer (2003): gemfibrozil, primidone, diclofenac sodium salt, carbamazepine, ketoprofen, naproxen, phenacetine, tris (2-chloroethyl)phosphate (TCEP) (Sigma Aldrich Chemicals), tris(chloropropyl)phosphate (TCPP), and propyphenazone (Pfaltz & Bauer, Inc.). Stock solutions were prepared by dissolving the compounds in milli-Q water adjusted to a pH of 10, during sonification and in the dark. A volume of 500 mL of sample was collected and filtered (0.45 mm, Whatman) prior to solid phase extraction. 10,11-dihydrocarbamazepine (Sigma Aldrich Chemicals) and 2-(m-chlorophenoxy) propionic acid (Sigma Aldrich Chemicals) were used as surrogate standards. The limits of detection (LofD) and limit of quantification (LoQ) for the target compounds ranged from 5 to 10 ng/L and from 10 to 50 ng/L, respectively. Only parent target compounds were investigated in this study, metabolites and conjugates were not analyzed.

2.2.2.

HPLC analysis

A high performance liquid chromatography (HPLC) HP 1100 (Agilent Technologies) combined with a UV diode array detector (DAD) was used to quantify concentrations of primidone, phenacetine, carbamazepine, naproxen, diclofenac, and ibuprofen in the lower mg/L range during adsorption breakthrough tests. The samples were directly injected without filtration or other

451


Table 1 – Physical–chemical properties of target organic compounds. Compound

Carbamazepine (anticonvulsant)

Structure

pKa

log P (log DpH7)

Expected degradability

References

13.9

2.67

Persistent

Mersmann et al., 2002; Schmidt et al., 2004

Diclofenac sodium (analgesic/antiinflammatory)

4.0

4.06 (1.06)

Degradable (anoxic)

Buser et al., 1998; Mo¨hle et al., 1999; Drillia et al., 2003; Zwiener and Frimmel, 2003, Hua et al., 2003; Schmidt et al., 2004

Gemfibrozil (blood lipid regulator)

4.8

4.39 (2.19)

Relatively persistent

Heberer and Stan, 1997; Drewes and Shore, 2001; Heberer, 2002; Drewes et al., 2003

3.72 (1.12)

Degradable (better aerobic)

Buser et al., 1999; Zwiener and Frimmel, 2003; Carballa et al., 2004; Sedlak and Pinkston, 2001; Drewes et al., 2003

Degradable


Degradable




Ibuprofen (analgesic/antiinflammatory)

Ketoprofen (analgesic/antiinflammatory)

Naproxen (analgesic/antiinflammatory)

Phenacetine (antipyretic)

4.4

4.2

4.0

n/a

2.81 (0.01)

3.00 (0.00)

1.63 (1.63)

Primidone (anticonvulsant)

12.3

0.4 (0.4)

Persistent


Propyphenazone (analgesic)

n/a

1.74 (1.74)

Poor/ persistent

Heberer et al., 2003; Drewes et al., 2003


452


Table 1 (continued ) Compound

Structure

Tris(2chloroethyl)phosphate (TCEP) (flame retardant)

pKa

n/a

log P (log DpH7)

Expected degradability

0.48 (0.48)


References

Heberer et al., 2001; Heberer et al., 2003, Amy and Drewes, 2006; Fries and Pu¨ttmann, 2003

n/a not applicable. pKa and log P values calculated by software ACD. log DpH¼7 values calculated using equations proposed by Scherrer and Howard, (1977).

sample preparation. The mobile phase consisted of two solvents: 1) type 1 water (adjusted to pH 2.5 with o-phosphoric acid (HPLC grade, Fisher Scientific) and buffered with 25 mmol potassium phosphate monobasic (Fisher Scientific)), and 2) acetonitrile (HPLC grade, Mallinckrodt ChromAR HPLC) with the same concentration of o-phosphoric acid as solvent 1). Sample injection (100 mL) occurred in triplicates for each sample at a flow rate of 2 mL/min. A mobile phase gradient was applied during the run for solvent 2 (acetonitrile) of 30% at t ¼ 0 min., 80% at t ¼ 12 min., and 30% at t ¼ 15 min. with a post-run time of 7 min. for column cleaning (30% of solvent 2). The detection wavelengths (band width 10 nm) were set to 205 nm at t ¼ 0 min. for quantification of primidone, phenacetine, and carbamazepine, 230 nm at t ¼ 6.8 min. (naproxen), and 205 nm at 7.8 min. for detection of diclofenac and ibuprofen with a reference wavelength of 330 nm (band width 60 nm) for all compounds. Standards were run in the range of 5–500 mg/L. The detection limits for each compound were 5 mg/L.

2.2.3.

DOC/UV absorbance/nitrate/ammonium

A Sievers 800 Total Organic Carbon Analyzer (GE, Boulder) was used for DOC quantification after microfiltration (0.45 mm, Whatman) (Standard Method 5310C). UV absorbance (UVA) measurements were conducted at 254 nm using a Beckman Coulter DU 800 Spectrophotometer after 0.45 mm filtration (Standard method 5910B). The specific UV absorbance (SUVA) was calculated as the ratio of UVA and DOC. Ammonium and nitrate concentrations were measured using the Nessler and the chromotropic acid method (HACH) with a detection range of 0.02–2.0 mg/l and 0.2–30 mg/L, respectively.

2.2.4.

Biomass

Soil biomass was determined as total viable biomass (i.e. viable, not necessarily active bacteria) using phospholipid extraction (PLE) as described in Rauch and Drewes (2005). Analyses were conducted in triplicates from the top-soil (0–2 cm, infiltration zone) of the columns.

2.2.5.

Wastewater effluent

Secondary treated wastewater effluent served as the feed to column systems and as source for subsequent organic carbon fractionation. The secondary treated effluent was collected from a local wastewater treatment plant employing nitrification and partial denitrification. The average DOC concentration of the effluent was 8.74 1.44 mg/L.

2.2.6.

Organic carbon fractionation

Organic matter (less than 1 mm in size as defined for this study) of the secondary effluent sample was isolated into three organic fractions: colloidal organic carbon, hydrophobic acids (HPO-A), and hydrophilic carbon (HPI) following the procedure described in Rauch and Drewes (2005) with some modifications. The sample was concentrated by vacuum rotary evaporation at 45 C (concentration factor: 40). The concentrate was separated by dialysis (Spectra/Por, Spectrum, 6000– 8000 Dalton) at pH 4–5 into organic colloids and DOC. The permeate of the dialysis (containing DOC) was collected and further separated into HPI and HPO-A by XAD-8 fractionation according to Leenheer et al. (2000) using a capacity factor of k0 ¼ 4. A carbon mass balance was performed for each carbon fractionation based on UVA and DOC measurements for quality control. In average, the secondary treated effluent contained 16 percent colloidal organic carbon, 38 percent HPO-A, 37 percent HPI, and 8 percent hydrophobic neutrals (HPO-N) (Rauch and Drewes, 2004). Organic carbon isolates were diluted to 3 mg/L DOC using milli-Q water, adjusted for ion strength, macro nutrient (nitrogen, phosphate) and micro nutrient concentrations (Rauch and Drewes, 2004) and promptly utilized as column feed waters.

2.3.

Experimental set-up

2.3.1.

Anoxic column system

The columns (PC system) were operated to study TOrC removal during simulated MAR under anoxic conditions and up to 3–4 weeks of retention time in the subsurface. The anoxic columns consisted of four 1-m plexiglass columns (15 cm i.d.) in series filled with aquifer material (d50 ¼ 0.8 mm, foc ¼ 0.003%) (Fig. SI-1, Supplemental Information). The column system was operated in flow-through mode at a loading rate of 0.065 m/d under saturated, anoxic (denitrifying) flow conditions (Table 2). The hydraulic retention time of the system using four columns in series was previously determined as 25 days. The columns had been continuously fed with secondary or tertiary treated effluents for over 6 years and for 5 months with the secondary treated effluent quality employed in this study prior to spiking of TOrC. The column influent was regularly purged with nitrogen gas to keep dissolved oxygen concentrations below 0.5 mg/L. Samples were collected once to twice a week from column influent and the four column effluents, respectively, and analyzed for DOC, UV absorbance, pH, conductivity and TOrC concentrations.

453


Table 2 – Operational parameters and water quality changes of columns used in this study.a Label

PC C1 C2 C3 C4 Abiotic

Feed water matrix

Length (m)

Anoxic columns (EfOM) Hydrophobic acids (HPO-A) Hydrophilic carbon (HPI) Effluent organic matter (EfOM) Organic colloids (Org. Colloids) Abiotic column, Type I water

41m¼ 4m 0.3

Organic Carbon

Feed water (mg/L) 6.9

Effluent (mg/L)

BOC (mg/L)

Predominant Redox Condition

Total Viable Soil Biomassb (nmol PO3 4 /g d.w soil)

Anoxic

17.3 1.3

3.1 0.9

See Fig. 1 for DOC column profile 2.7 0.02 0.4 0.9

Aerobic

3.3 0.4

0.3

2.8 0.4

2.6 0.1

0.3 0.4

Aerobic

10.5 5.3

0.3

3.0 0.4

2.7 0.4

0.4 0.4

Aerobic

20.6 2.5

0.3

3.1 0.5

2.2 0.3

0.8 0.5

Aerobic

27.2 1.8

0.3

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

Recommend Documents