Analysis of pharmaceutical residues in wastewaters, surface and drinking waters - study of the removal efficiency through conventional and advanced treatment processes
Full text
Maria Augusta Dionísio de Sousa Analysis of pharmaceutical residues in wastewaters, surface and drinking waters – Study of the removal efficiency through conventional and advanced treatment processes Thesis presented to obtain the Ph.D. degree in Pharmaceutical Sciences, Hydrology Specialty FACULTY OF PHARMACY OF THE UNIVERSITY OF PORTO Work conducted under the supervision of: Prof. Dr. Maria de Fátima Pinto Soares Pereira de Magalhães Alpendurada and co-supervision of: Prof. Dr. Jana Hajšlová Porto January 2013
Maria Augusta Dionísio de Sousa Análise de resíduos farmacêuticos em águas residuais, superficiais e de consumo humano – Estudo da eficiência da remoção por processos de tratamento convencionais e processos avançados Tese do Terceiro Ciclo de Estudos Conducente ao Grau de Doutor em Ciências Farmacêuticas, Especialidade Hidrologia FACULDADE DE FARMÁCIA DA UNIVERSIDADE DO PORTO Trabalho realizado sob orientação de: Prof. Dr. Maria de Fátima Pinto Soares Pereira de Magalhães Alpendurada e co-orientação de: Prof. Dr. Jana Hajšlová Porto Janeiro 2013
iv The full replication of this thesis is only authorised for research purposes, upon written declaration of the person concerned, that such pledges. É autorizada a reprodução integral desta tese apenas para efeitos de investigação, mediante declaração escrita do interessado, que a tal se compromete.
v The work presented in this thesis took place at the following laboratories: IAREN – Instituto da Água da Região Norte (Water Institute of the Northern Region), Matosinhos, Portugal LSRE – Laboratory of Separation and Reaction Engineering, Associate Laboratory LSRE/LCM - Department of Chemical Engineering, Faculty of Engineering of the University of Porto, Porto, Portugal Department of Food Analysis and Nutrition, Faculty of Food and Biochemical Technology, Institute of Chemical Technology (ICT), Praha, Czech Republic Water Institute of the Northern Region I A R E N Water Institute of the Northern Region I A R E N This work was supported by the following institutions and projects: FCT – Fundação para a Ciência e Tecnologia (Portuguese Foundation for Science and Technology), through the Ph.D. grant SFRH/BD/44509/2008, co-financed by FSE/POPH – Fundo Social Europeu / Programa Operacional de Potencial Humano (European Union) and by funds of the Ministry of Science and Higher Education IAREN-CIPE Project – Centro de Investigação de Poluentes Emergentes (Research Centre of Emerging Pollutants), co-funded by QREN ON.2 – O Novo Norte, through the research project NORTE-07-0162-FEDER-000022 (2010-2012) Projects PTDC/AAC-AMB/113091/2009 and PEst-C/EQB/LA0020/2011 financed by FEDER through COMPETE - Programa Operacional Factores de Competitividade and by FCT ICT Praha - Specific University Research MSMT no. 21/2012
vi
vii Acknowledgements The work herein presented would have not been achieved without the teaching, technical and financial support, encouragement, help and friendship of several persons and entities, to whom I address the following acknowledgements: I am deeply grateful to Prof. Dr. Maria de Fátima Alpendurada, scientific supervisor of this thesis, for her valuable advices, comments, suggestions and encouragement that allowed me to accomplish this Ph.D. work. I further acknowledge her as Director of IAREN – Instituto da Água da Região Norte (Water Institute of the Northern Region), for all the advanced instrumentation, technical and financial means at my disposal, which definitely made the difference. My thanks also belong to Prof. Dr. Jana Hajšlová, Head of the Department of Food Analysis and Nutrition, Institute of Chemical Technology (ICT) Praha, for the warm welcome, proactivity, advices and encouragement, as well as for the opportunity to work with such end-of-pipe instrumentation. I express my gratitude as well to Prof. Dr. Rui Boaventura and Dr. Vítor Vilar, from LSRE (Laboratory of Separation and Reaction Engineering) – Faculty of Engineering of the University of Porto, for the teaching and availability, which enabled me to work with sophisticated technology in the field of advanced oxidation processes. To the Faculty of Pharmacy of the University of Porto, I thank for the support of more than ten years of my learning process. I also acknowledge all the Professors, namely Prof. Dr. José Fernandes, and all the colleagues I recently met, for the voting trust and strong encouragement. To FCT – Fundação para a Ciência e Tecnologia (Portuguese Foundation for Science and Technology), I thank the financial support provided by the Ph.D. grant SFRH/BD/44509/2008, as well as the prompt handling of every related process. To all the colleagues and friends from IAREN: Dra. Edite, Andreia, Ana, Eng.º André, Marta L., Marta T., Vera R., Vera P., Cláudia, Sara, Dª Maria José, Dª Lurdes, Luís and Dr. José João, I thank the everyday companionship, kindness and encouragement. I further address very special thanks to Dr. Carlos, for the character and competence, but, above all, for the friendship, complicity and the never ending support and patience.
xiv 2.5. References …………………………………………………………………………….… 57 3. MONITORING OF PHARMACEUTICALS IN MUNICIPAL WASTEWATERS AND SURFACE WATERS ……………………………………………………………………… 61 3.1. Water Sampling ………………………………………………………………….…….. 62 3.1.1. Sampling Protocol ………………………………………………………………….…………… 62 3.1.2. Materials and Equipments ………………………………………………………….……… 64 3.2. A Portuguese Case Study …………………………………………………….……… 66 3.2.1. Monitoring Sites ………………………………………………………………………………… 66 3.2.1.1. Febros WWTP ………………………………………………………………………… 66 3.2.1.2. Febros River and Douro River ………………………………………….……… 67 3.2.2. Methodology ……………………………………………………………………………………… 69 3.2.2.1. Sampling ……………………………………………………………………….……… 69 3.2.2.2. Analytical Method – SPE-LC-MS/MS ………………………………….…… 69 3.2.3. Results and Discussion ……………………………………………………………………… 70 3.2.3.1. Wastewaters from Febros WWTP ………………………………………….… 70 3.2.3.2. Surface Waters from Febros River and Douro River ………….……… 73 3.2.4. Conclusions ……………………………………………………………………………………..… 74 3.3. Paper 2 – Effects of treated domestic effluents and tributaries on the contamination of Douro river with pharmaceutical compounds – Mitigation processes ………………………………………………………………… 75 3.4.Inter-laboratory Exercises: Method Validation and Quality Control . 86 3.5. References …………………………………………………………………………….…. 95 4. PHOTO-REMEDIATION OF CONTAMINATED WATERS USING ADVANCED OXIDATION PROCESSES (AOPS) ……………………………………..…….……… 93 4.1. Advanced Oxidation Processes ………………………………….………………… 94 4.1.1. Definition ………………………………………………………………………………………….. 94 4.1.2. Classification and Applications to Water Treatment …………………….………... 96 4.1.2.1. Heterogeneous Photocatalysis …………………………….…………………… 98 4.1.2.1.1. Fundamentals …………………………………………………….…………. 98 4.1.2.1.2. Catalysts ……………………………………………………………….……. 100 4.1.2.1.3. Variables affecting Photocatalysis ………………………….…….. 101 4.1.2.1.4. Photochemical Reactors ………………………………………………. 105 4.1.2.1.5. Analytical Determinations …………………………………….…….. 113
xv 4.2. Paper 3 – Photolytic and TiO2-Assisted Photocatalytic Oxidation of the Anxiolytic Drug Lorazepam (Lorenin® pills) under Artificial UV Light and Natural Sunlight: A Comparative and Comprehensive Study ….. 115 4.3.Paper 4 – Suspended TiO2-Assisted Photocatalytic Degradation of Emerging Contaminants in a Municipal WWTP Effluent using a solar Pilot Plant with CPCs ………………………………………………………………… 127 4.4. References ……………………………………………………………………………… 138 5. PHOTOTREATMENT MECHANISMS AND BY-PRODUCTS STRUCTURAL ELUCIDATION ………………………………………………………………………….. 141 5.1. Paper 5 – Lorazepam Photofate under Photolysis and TiO2-assisted Photocatalysis: By-products’ Identification and Evolution Profiles during Phototreatment of a Contaminated WWTP Effluent ………….. 143 6. OCCURRENCE OF EMERGING POLLUTANTS’ BY-PRODUCTS IN WATER RESOURCES ………………………………………………………………………….... 183 6.1. Brief Introduction on the Sources of Transformation Products of Emerging Contaminants ………………………………………….………………. 184 6.2.Transformation Products in Natural Waters: from Contamination Sources to Drinking Water Production ………………………………………. 185 6.3.Wastewaters as Major Source of Transformation Products ….…….... 189 6.4. Origin and Presence of Transformation Products in Drinking Water …………………………………………………………………………………………….…. 196 6.5. Ubiquity and Regio-Specificity of Transformation Products …………. 197 6.6. Transformation Products of Emerging Contaminants: Fate and Behaviour …………………………………………………………………………….…. 199 6.7. Conclusions …………………………………………………………………….……… 201 6.8. References …………………………………………………………………………….. 203 7. FINAL CONCLUSIONS AND REMARKS: WORK NOVELTY, KNOWLEDGE GAPS AND FUTURE RESEARCH …………………………………………………………………. Annex I – NORMAN list of emerging substances (most frequently discussed) (latest update approved March 2011) ……………………………………………………………………………….. 211 Annex II – Analysis of acidic, basic and neutral pharmaceuticals in river waters: clean-up by 1º,2º amino anion exchange and enrichment using an hydrophilic adsorbent (associated paper) ……………………………………………………………………………………………… 231 Annex III – Report of the intercalibration exercise under PHARMAS Project: “Ecological and human health risk assessments of antibiotics and anti-cancer drugs found in the environment” …………………………………………………………………………………………………….. 255 Annex IV – Treatment of a pesticide-containing wastewater using combined biological and solar-driven AOPs at pilot scale (associated paper) ………………………………………….. 282
xvi
xvii Figure Index Figure 1.1. Map presenting the national and international river basin districts, as designated by Member States ………………………………………………………………..… 3 Figure 1.2. Representative sources and routes of pharmaceutical compounds in the environment ……………….………………………………………………………………………… 11 Figure 1.3. Therapeutic classes of pharmaceuticals detected in the environment (expressed in relative percentage). Data collected from 134 articles published between 1997 and 2009 ……………………………………………………………………………………. 13 Figure 1.4. Graphical representation of EMEA guidelines ………………………………………… 17 Figure 1.5. FDA tiered approach to fate and effects testing ………………………………………… 18 Figure 2.1. Illustration of the SPE process: - interferences; - target analytes. 1) Retention of the analyte: analyte molecules are enriched on the adsorbent > interfering components and solvent molecules (matrix) are not retained > remaining interfering components are washed from the adsorbent > the analyte is removed from the adsorbent by elution with a suitable solvent (e.g., methanol, tetrahydrofuran (THF), -propanol); 2) Retention of interfering components: analyte molecules show no interaction with the adsorbent > interfering components and solvent molecules (matrix) are retained > analyte molecules are “washed” from the adsorbent > the solid phase is simply used to “filter” the sample …………….… 30 Figure 2.2. SPME components: on the left – scheme of a SPME set-up; in the centre – detail of a SPME fibre; on the right – operating modes of SPME: A – immersion, B – headspace …………………………………………………………………… 34 Figure 2.3. Analytical methods applied for the most common pharmaceutical compounds present in water and wastewater …………………………………………………………… 36 Figure 2.4. Overview of mass spectrometric techniques …………………………………………… 37 Figure 2.5. Time-of-flight mass analyser. (1) Pulse of ions from the orthogonal accelerator (spatial focusing); (2) separation of ions according to their flight times; (3) focusing of the kinetic energy of ions; (4) separation of focused ions according to their flight times, dependent on their weight ……………………………………… 38
xviii Figure 3.1. Sampling devices. (a) multiparameter field probe; (b) sampling flask with remotely triggered closure; (c) automatic sampler used to collect and refrigerate grab or composite samples over a 24-h period ……………………….. 64 Figure 3.2. Image of Febros WWTP, which discharges to Febros river (flowing on the right-hand side). In the top-right part of the image one can identify Douro rver, dividing the city of Vila Nova de Gaia (left bank) from the city of Porto (right bank) …………………………………………………………………………………………. 66 Figure 3.3. Geographical location of the case study – Portuguese hydrographical region no. 3: Douro ……………………………………………………………………………………………… 67 Figure 3.4. Image of Douro river ………………………………………………………………………........ 68 Figure 3.5. Image of Febros river, close to the river mouth ……………………………………... 68 Figure 3.6. Grab samples of WWTP effluent, collected hourly for a 21-h period, analysed individually. Graphics represent concentration variations for each pharmaceutical over the sampling period. AVG – theoretical average value calculated for a composite sample; MIN – minimum value; MAX – maximum value; LOQ – limit of quantification ……………………………………………………… 72 Figure 3.7. Z-score absolute values (outliers excluded). Sample A – spiked tap water; sample B – spiked river water ………………………………………………………………. 89 Figure 4.1. Reactions involved in the processes of direct photolysis, indirect photolysis, heterogeneous photocatalysis and formation of oxygen radical reactive species …………………………………………………………………………………………………………… 95 Figure 4.2. AOPs classification, possible combinations and some relevant references …. 96 Figure 4.3. Number of publications relative to photodegradation treatment processes of waters contaminated with pharmaceutical compounds, between 2000 and 2012 …………………………………………………………………………………………………… 97 Figure 4.4. Schematic diagram illustrating the principle of heterogeneous photocatalysis ……………………………………………………………………………………………………………. 99 Figure 4.5. Lab-scale photochemical reactor equipped with a medium pressure mercury lamp Heraeus TQ150W (LsTQUV) ……………………………………………………… 106 Figure 4.6. Lab-scale photochemical reactor equipped with a LYNX S11W blacklight blue lamp (LsBLBUV) ………………………………………………………………………………. 106
xix Figure 4.7. Solar Pilot Plant with CPCs (SPP-CPCs) …………………………………………….... 107 Figure 4.8. Transmittance spectra of quartz and borosilicate (Schott-Duran type 3.3, Germany, cut-off at 280 nm) ……………………………………………………………… 108 Figure 4.9. Chemical structure of the complex Fe2+ -1,10-phenanthroline (ratio 1:3) … 109 Figure 4.10. Lab-scale prototype LsTQUV: graphic representation of the accumulated UV energy per litre of solution, as a function of the time after turning the UV lamp on [obs.: insert picture represents the emission spectrum of the medium pressure mercury lamp Heraeus TQ 150W] …………………………………….……… 111 Figure 4.11. Design of photocatalytic reactors used for water treatment: (a) one-sun system; (b) parabolic-trough concentrator (PTC); (c) compound parabolic collector (CPC) …………………………………………………………………………………… 112 Figure 4.12. Ray tracing example for a CPC collector ……………………………………………… 112
xx
xxi Table Index Table 1.1. Therapeutic groups according to the Portuguese Therapeutic Handbook …....... 6 Table 1.2. Pharmaceutical emerging pollutants, according to the NORMAN Network ……. 7 Table 1.3. Pharmaceutical compounds ordered from the highest to the lowest sales amount in Portugal, over the years 2005/2007/2009, according to INFARMED …..… 10 Table 2.1. Commercial brands of SPE extraction columns constituted by polymeric adsorbents, frequently utilized in the extraction of pharmaceutical compounds from water samples ………………………………………………………………………………… 31 Table 3.1. Error sources during sampling procedures …………………………………………….…. 63 Table 3.2. Passive sampling devices for organic contaminants …………………………………. 65 Table 3.3. GPS coordinates of the sampling spots in Febros river, identified in Fig. 3.3 .. 67 Table 3.4. Target pharmaceuticals quantified in Febros WWTP influent and effluent ..…. 71 Table 3.5. Antibiotics included in the ILE-PHARMAS exercise …………………………………. 88 Table 3.6. Concentration (ng L-1) of antibiotics determined in the three replicates of sample series A (spiked tap water). Each extract was injected twice for LCMS/MS analysis ……………………………………………………………………………………. 88 Table 3.7. Concentration (ng L-1) of antibiotics determined in the three replicates of sample series B (spiked river water). Each extract was injected twice for LCMS/MS analysis ……………………………………………………………………………………. 89 Table 4.1. Properties of some semiconductor materials ………………………………………...... 100
xxii
xxiii Publications, Oral and Poster Communications • List of Publications Paper 1 (In English language) Sousa, M.A., Gonçalves, C., Cunha, E., Hajšlová, J. and Alpendurada, M.F., Cleanup strategies and advantages in the determination of several therapeutic classes of pharmaceuticals in wastewater samples by SPE–LC–MS/MS. Analytical and Bioanalytical Chemistry, 2011. 399: 807-822. Paper 2 (In Portuguese language) Sousa, M.A., Gonçalves, C., Vilar, V.J.P., Boaventura, R.A.R. e Alpendurada, M.F., Efeitos dos efluentes domésticos tratados e rios tributários no pool de resíduos farmacêuticos do Rio Douro - Processos de Mitigação [Effects of treated domestic effluents and tributaries on the contamination of Douro river with pharmaceutical compounds – Mitigation processes]. Submitted for publication in Recursos Hídricos (APRH - Associação Portuguesa de Recursos Hídricos), June 2011. Paper 3 (In English language) Sousa, M.A., Gonçalves, C., Pereira, J.H.O.S., Vilar, V.J.P., Boaventura, R.A.R. and Alpendurada, M.F., Photolytic and TiO2-Assisted Photocatalytic Oxidation of the Anxiolytic Drug Lorazepam (Lorenin® pills) under Artificial UV Light and Natural Sunlight: A Comparative and Comprehensive Study. Solar Energy, 2013. 87: 219-228. Paper 4 (In English language) Sousa, M.A., Gonçalves, C., Vilar, V.J.P., Boaventura, R.A.R. and Alpendurada, M.F., Suspended TiO2-Assisted Photocatalytic Degradation of Emerging Contaminants in a Municipal WWTP Effluent using a Solar Pilot Plant with CPCs. Chemical Engineering Journal, 2012. 198-199: 301-309. Paper 5 (In English language) Sousa, M.A., Lacina, O., Hrádková, P., Pulkrabová, J., Vilar, V.J.P., Gonçalves, C., Boaventura, R.A.R., Hajšlová, J. and Alpendurada, M.F., Lorazepam Photofate under Photolysis and TiO2-assisted Photocatalysis: By-products’ Identification and Evolution Profiles during Phototreatment of a Contaminated WWTP Effluent. Submitted for publication in Environmental Science and Technology, September 25, 2012.
xxx ETC 4-Epitetracycline EU European Union FDA Food and Drug Administration FEDESA European Federation of Animal Health GCB Graphitised carbon black GC-MS Gas chromatography-mass spectrometry HLB Hydrophilic-lipophilic-balanced HPLC-MS High performance liquid chromatography-mass spectrometry HRT Hydraulic retention time HS Humic substances HTLC High-temperature liquid chromatography IAC Immuno-affinity chromatography ICTC Isochlortetracycline ILE Inter-laboratory exercise INFARMED Portuguese National Authority of Medicines and Health Products, IP JRC Joint Research Centre LC Liquid chromatography LC50 Half-maximal lethal concentration LC-MS Liquid chromatography-mass spectrometry LLE Liquid-liquid extraction LOD Limit of detection LOEC Lowest observed effect concentration LOQ Limit of quantification LPME Liquid-phase microextraction LsBLBUV Lab-scale blacklight blue-UV system LsTQUV Lab-scale medium pressure mercury lamp Heraeus TQ 150W LZP Lorazepam MAX Mixed-mode anion exchange sorbent MCX Mixed-mode cation exchange sorbent MDA 4-Methylenedioxyamphetamine MDEA 3,4-Methylenedioxyethamphetamine MDMA 3,4-Methylenedioxymethamphetamine MESCO Membrane-enclosed sorptive coating MIP Molecularly imprinted polymer MISPE Molecularly imprinted solid-phase extraction MS Mass spectrometry
xxxi NMR Nuclear magnetic resonance NOEC No-observed effect concentration NPEOs Nonylphenol polyethoxylates NRZ Nordazepam NSAIDs Non-steroidal anti-inflammatory drugs OA Oxanilic acid OFL Ofloxacin OTC Over-the-counter OXA Oxazepam PAHs Polycyclic aromatic hydrocarbons PBDEs Polybrominated diphenyl ethers PCBs Polychlorinated biphenyls PDB Passive diffusion bag PEC Predicted environmental concentration PFAs Poly- and perfluoroalkyl substances PFOS Perfluorooctanosulfonates PFCA Perfluoroalkylcarboxylates PISCES Passive in situ concentration-extraction sampler PNEC Predicted no-effect concentration POCIS Polar organic integrative sampler PoD Points of departure PS-DVB Poly(styrene-divinylbenzene) PTC parabolic-trough concentrator QAPP Quality assurance project plan (previously known as QA/QC) QA/QC Quality assurance/quality control QTRAP Hybrid triple-quadrupole linear ion trap mass spectrometer QUV Accumulated UV energy RAMs Restricted access materials REACH Registration, evaluation and authorisation of chemicals SC Semiconductor SM-LLME Stir membrane liquid-liquid microextraction SMX Sulfamethoxazole SMZ Sulfamethazine SPATT Solid-phase adsorption toxin tracking SPE Solid-phase extraction SPMD Semi-permeable membrane device
xxxii SPME Solid-phase microextraction SPP-CPCs Solar pilot plant with compound parabolic collectors SPY Sulfapyridine STPs Sewage treatment plants TC Tetracyclines THF Tetrahydrofuran TMP Trimethoprim TMZ Temazepam TOF-MS Time-of-flight mass spectrometry TP Transformation product UHPLC Ultra high-performance liquid chromatography USA United States of America UV Ultraviolet UWWD Urban Waste Water Directive VB Valence band VOCs Volatile organic compounds WAX Weak-anion exchange WFD Water Framework Directive WWTPs Wastewater treatment plants
General Introduction 1 Chapter 1 General Introduction 1.1. Water Quality Policy and Legislation Nowadays, one of the most relevant topics in the environmental field is water quality. It has been evident for a considerable time now the increasing demand by citizens and environmental organizations for cleaner rivers, lakes, groundwater and coastal beaches. This demand was precisely one of the main reasons why the European Commission (EC) has made water protection one of its top work priorities. Whilst some actions taken in the past by the European Union (EU), such as the Drinking Water Directive (DWD) and the Urban Waste Water Directive (UWWD), can properly be considered milestones, a renewed European Water Policy was mandatory to address the increasing awareness of citizens and other involved parties. Consequently, and as the outcome of a consultation process involving all interested parties, the Commission presented a Proposal for a Water Framework Directive, with the following key aims (1): i) expanding the scope of water protection to all waters, surface waters and groundwater; ii) achieving “good status” for all waters by a set deadline; iii) water management based on river basins; iv) “combined approach” of emission limit values and quality standards; v) getting the prices right; vi) getting the citizens involved more closely; vii) streamlining legislation. Hence, the Water Framework Directive (WFD, Directive 2000/60/EC) was created and adopted as an operational tool, setting the objectives for water protection. It presented a huge breakthrough in the EU water policy, aiming to achieve a good ecological and chemical status for all surface waters at the latest 15 years from the date of entry into force, i.e., 22 December 2000. Furthermore, the WFD was later amended by Decision no. 2455/2001/EC, that established a list of 33 priority substances given their significant risk
Chapter 1 2 to the aquatic environment, and even more recently by Directive 2008/105/EC, which laid down environmental quality standards (EQS) for priority substances and other certain pollutants, as provided for in Article 16 of the WFD (1, 2). In order to accomplish the co-ordination of the objectives and achieve a good status for all waters by the set deadline, it became necessary to define the best model for a single system of water management. As a result, the management by river basin – the natural geographical and hydrological unit – was adopted, instead of according to administrative or political boundaries. Meanwhile, though several Member States are already undertaking “the river basin approach”, some national and international river basin districts are still not established (Fig. 1.1). Furthermore, all river basin districts must be updated every six years, in order to provide the context for all co-ordination requirements aforementioned (1). 1.2. Emerging Environmental Pollutants Over the last few decades, the focus of environmental research has been extended from the conventional “priority pollutants”, such as polycyclic aromatic hydrocarbons (PAHs), polychlorinated biphenyls (PCBs), dioxins and pesticides, to the so called “emerging pollutants” (EPs), many of which are ubiquitously distributed in the aquatic compartment. EPs are not necessarily new chemicals; they also include substances that have long been present in the environment but whose presence and significance are only now being elucidated. Therefore, the rising interest on such organic micropollutants is not only concomitant with their increased diversity, widespread and growing use, but also with the development of new highly sensitive analytical tools, namely liquid chromatography (LC) coupled with mass spectrometry (MS). Such analytical hyphenated technique led to a “revolution” in environmental analysis, enabling the identification and quantification of polar organic pollutants down to relatively low concentrations, in all kinds of waters (wastewater, surface water, ground water, drinking water) and in solid matrices (sewage sludge, manure, sediment) (3-5). However, emerging pollutants correspond in most cases to unregulated contaminants and are generally not included in the monitoring campaigns enforced by EU legislation. Consequently, there is the need to build knowledge on their occurrence, levels and fate in the environment, as well as to elucidate their long-term risks, ecotoxicity and human
General Introduction 3 Fig. 1.1. Map presenting the national and international river basin districts, as designated by Member States (1).
Chapter 1 4 health impact, which might be used as evidence for future regulation (2, 6). Emerging pollutants are considered “pseudo-persistent” compounds, since their transformation/removal rates can be compensated by their continuous introduction into the environment. Moreover, their recalcitrant character, together with their polarity, favour their spread along the aquatic compartment (7). The NORMAN Network, established since 2009 as a permanent self-sustaining network of reference laboratories, research centres and related organizations for the monitoring and biomonitoring of emerging environmental substances, started its activities in September 2005, with the financial support of the European Commission (NORMAN project – 6th Framework Programme – Priority 6.3). Currently, it proposes a preliminary list of emerging substances, built upon the results collected by Member States, after several European surveys. Herein, different classes of compounds are included, from algal toxins to anticorrosives, antifoaming agents, antifouling compounds, antioxidants, biocides, bioterrorism/sabotage agents, complexing agents, detergents, disinfection by-products (drinking water), drugs of abuse, flame retardants, food additives, fragrances, gasoline additives, industrial chemicals, nanoparticles, perfluoroalkylated substances and their transformation products, personal care products, pesticides, pharmaceuticals, plasticizers, trace metals and their compounds, wood preservatives, among others (for individual substances, please refer to Annex I) (8). Furthermore, the NORMAN Association has also created an interactive database – EMPODAT, designed to assemble the geo-referenced monitoring and bio-monitoring data collected by its members on emerging substances in air, water and soil. Each contact point across the Member States is requested to provide information concerning data sources, sample matrices and analytical methods employed to obtaining the results. In conclusion, the EMPODAT database allows access to the latest information on emerging pollutants, with an overview of benchmark values on the occurrence of emerging substances across Europe, as well as the identification of gaps in data relating to time, geographical areas and/or environmental matrices. Among all emerging substances in water, special attention has been given to pharmaceutical products and residues. This rising awareness is attributed to their growing variety and consumption, recognised (eco)toxicity, as well as their unpredictable environmental impact, even at low concentration levels (9, 10).
General Introduction 5 1.2.1. Pharmaceutical Compounds 1.2.1.1. Definition and Classification by Therapeutic Group The European Medicines Agency (EMEA) is responsible for establishing the rules governing human and veterinary medicinal products in the EU. According to Directive 2001/83/EC, of November 6, on the Community code relating to medicinal products for human use, later amended by Directive 2004/27/EC, of March 31, a medicinal product is (11): i) any substance or combination of substances presented as having properties for treating or preventing disease in human beings; or ii) any substance or combination of substances which may be used in or administered to human beings either with a view to restoring, correcting or modifying physiological functions by exerting a pharmacological, immunological or metabolic action, or to making a medical diagnosis. In Portugal, there is an annual edition of the Therapeutic Handbook, mainly addressed to medical doctors, pharmacists and other health care professionals, which sets the guidelines for the therapeutic use of drugs and constitutes an essential tool for the rational use of medicines, providing guidance during prescription. It contains the monographs of all medicines marketed in Portugal and available in pharmacies, as well as some that, although available only in hospitals, are of particular interest for patients continuing the treatment process at the clinic. In the Therapeutic Handbook, medicines are grouped by pharmaceutically active substance, according to the respective therapeutic group. The first 10 therapeutic groups and respective sub-groups included in this Handbook are displayed in Table 1.1, since they include the pharmaceutical compounds most widely studied as environmental pollutants (12). According to NORMAN, Table 1.2 shows a list of pharmaceutical pollutants which are object of study by EU Member States, as well as all over the world. Moreover, at the NORMAN Association website, further information including the systematic name (IUPAC), CAS number and therapeutic sub-class of every pharmaceutical compound can be found (8).
Chapter 1 6 Table 1.1. Therapeutic groups according to the Portuguese Therapeutic Handbook (12) THERAPEUTIC HANDBOOK Therapeutic Groups Edition 2011 GROUP 1 - Anti-infective Medicines 1.1. Antibacterials 1.2. Antifungals 1.3. Antivirals 1.4. Antiparasitics GROUP 2 - Central Nervous System (CNS) 2.1. General Anesthetics 2.2. Local Anesthetics 2.3. Muscle Relaxing Agents 2.4. Antimyasthenia Agents 2.5. Antiparkinsonian Agents 2.6. Antiepileptics and Anticonvulsivants 2.7. Antiemetics and Antivertigo Agents 2.8. Unspecific Stimulants of CNS 2.9. Psychopharmaceuticals 2.10. Analgesics and Antipyretics 2.11. Medicines for Headache 2.12. Narcotics 2.13. Others GROUP 3 - Cardiovascular System 3.1. Cardiotonics 3.2. Antiarrhythmics 3.3. Sympathomimetics 3.4. Antihypertensives 3.5. Vasodilators 3.6. Venotropics 3.7. Lipid Regulators GROUP 4 - Blood 4.1. Antianemic Agents 4.2. Hematopoietic Growth Factors 4.3. Anticoagulant and Antithrombotic Agents 4.4. Antihemorrhagic Agents GROUP 5 - Respiratory System 5.1. Antiasthmatics and Bronchodilators 5.2. Antitussives and Expectorants 5.3. Pulmonary Tensoactive Agents (Surfactants) GROUP 6 - Digestive System 6.1. Medicines acting on Mouth and Orofaringe 6.2. Antacids and Anti-ulcer Agents 6.3. Modifiers of Gastrointestinal Motility 6.4. Antispasmodics 6.5. Enzyme Inhibitors 6.6. Enzyme Suplements and Analogs 6.7. Antihemorrhoid Agents 6.8. Intestinal Anti-inflammatory Agents 6.9. Medicines acting on Liver and Biliary GROUP 7 - Genitourinary System 7.1. Vaginal Medicines 7.2. Medicines acting on Uterus 7.3. Urinary Anti-infective and Antiseptic Agents 7.4. Others GROUP 8 - Hormones and Medicines for Endocrine Diseases 8.1. Hypothalamus and Pituitary Hormones, their Analogs and Antagonists 8.2. Corticosteroids 8.3. Thyroid Hormones and Antagonists 8.4. Insulins, Oral Antidiabetics and Glucagon 8.5. Sex Hormones 8.6. Ovulation Stimulants and Gonadotropins 8.7. Anti-hormones GROUP 9 - Locomotor System 9.1. Non-Steroidal Anti-inflammatory Drugs 9.2. Rheumatism Modifiers 9.3. Medicines used for Arthritis 9.4. Medicines used for Arthrosis 9.5. Anti-inflammatory Enzymes 9.6. Medicines acting on Bones and Calcium Metabolism GROUP 10 - Antialergics 10.1. Antihistamines 10.2. Corticosteroids 10.3. Sympathomimetics
General Introduction 7 Table 1.2. Pharmaceutical emerging pollutants, according to the NORMAN Network (8) (R)-O-Desmethyl Naproxen 17-alpha-Estradiol 17-alpha-Ethinylestradiol 17-beta-Estradiol Acebutolol Acecarbromal Aceclofenac Acemetacin Acetaminophen (paracetamol) Acetazolamide Acetylsalicylic acid (aspirin) Acyclovir Albuterol Albuterol sulfate Alclofenac Allobarbital Alprazolam Amitryptiline Amobarbital Amoxicillin Ampicillin Anthracen-1,4-dione Apramycin Aprobarbital Atenolol Azithromycin Baclofen Baquiloprim Betamethasone Beta-sitosterol Betaxolol Bezafibrate Bisoprolol Bromazepam Butalbital Caffeine Carazolol Carbamazepine Cefacetrile Cefalexin Cefalonium Cefapirin Cefazolin Cefoperazone Chloral hydrate Chloramphenicol Chlorobutanol Chlortetracycline Cholesterol Ciprofloxacin Citalopram Clarithromycin Clenbuterol Clofibric acid Clotrimazole Cloxacillin Codeine Cotrimoxazole Crotamiton Cyclophosphamide Cyclophosphamide (anhydrous) Danofloxacin Dantrolene Dapsone Daunorubicin Dexamethasone Diatrizoate Diazepam Diclofenac Dicloxacillin Diethylstilbestrol Difloxacin Diphenhydramine Domperidone Doxepine Doxorubicin Doxycycline (anhydrous) Doxycycline (monohydrate) Enoxacin Enrofloxacin Epirubicin Erythromycin Escitalopram Esomeprazole Estriol Estrone Estrone sulphate Ethosuximide Etofibrate Famotidine Fenfluramine Fenofibrate Fenofibric acid Fenoprofen Fenoprofen calcium salt dihydrate Fenoterol Flucloxacillin Flumequine Fluorouracil Fluoxetine Fluvoxamine Furosemide Gemfibrozil Gentamicin Glyburide (glibenclamid; glybenzcyclamide) Hexobarbital Hydrochlorothiazide Hydrocodone Ibuprofen Ibuprofen 1-hydroxy Ibuprofen 2-hydroxy Ifosfamide Imapramine Iminostilbene Indomethacin Iohexol Iomeprol Iopamidol Iopromide Ivermectin Josamycin Kanamycin sulfate Ketoprofen Lamotrigine Lansoprazole Levetiracetam Lidocaine Lincomycin Lithium carbonate Loratadine Lorazepam Lovastatin Marbofloxacin Mebeverine Meclofenamic acid Medazepam Mefenamic acid Meprobamate Mestranol Metformin Methicillin Methylphenobarbital Metoprolol Mevastatin Minocycline Nadolol Nafcillin Nandrolone Naproxen Neomycin B N-Methylphenacetine Nordiazepam Norfloxacin Novobiocin Ofloxacin Oleandomycin Omeprazole Oxacillin Oxalinic acid Oxazepam Oxprenolol Oxytetracycline Paroxetine Penicillin G Penicillin V Pentobarbital Pentoxifylline Phenazone Phenobarbital Phenylbutazone Phenytoine Pindolol Pipamperon Pravastatin Prednisolone Primidone Propranolol Propyphenazone Ranitidine Roxithromycin Salbutamol Sarafloxacin Secobarbital Secobarbital sodium Sertraline Simvastatin Sotalol Spectinomycin Spiramycin Streptomycin Sulfadiazine Sulfadimethoxine Sulfadoxine Sulfamerazine Sulfamethazine Sulfamethoxazole Sulfapyridine Taloxa Temazepam Terbutaline Tetracycline Tiamulin Tilmicosin Timolol Tolfenamic acid Tramadol Trimethoprime Tylosin Valnemulin Valproic acid Verapamil Zolpidem
Chapter 1 14 In the Netherlands, however, antibiotics, antiepileptics and β-blockers have been detected at concentrations up to 100 ng L-1 in the drinking water supply (39). It is estimated that, to date, there have been detected between 15 and 25 pharmaceutical compounds in drinking water all over the world (37). As previously stated, pharmaceutical compounds can be metabolized in humans or animals organism, then both native compounds and metabolites can reach the environment and originate different transformation products, under various physicochemical and biological processes, as in WWTPs or waterworks. Thus, the designation of “by-products” (BPs) was created to comprise both the excreted metabolites and the environmental transformation products (10). The fate and behaviour of pharmaceuticals and their BPs in the environment still requires further elucidation. Nevertheless, it is understandable that their concentration levels in receiving waters can be attenuated first by dilution and then adsorption on suspended solids and sediments, colloids or other organic matter. They may also undergo biotic and/or physico-chemical transformations, namely direct and indirect photodegradation reactions, besides, hydrolysis, oxidation/reduction, isomerisation, etc. (40, 41). These phototransformation processes will be further discussed on chapter 4 of the present thesis. 1.2.1.4. Risk Assessment and (Eco)Toxicity Results of toxicology studies have revealed that some pharmaceuticals are suspected to have direct toxicity towards certain aquatic organisms. Moreover, their continual but undetectable effects can slowly accumulate, finally leading to irreversible changes on both wildlife and human beings (42, 43). It is essential to foster the study on ecotoxicological hazards, to go through a comprehensive evaluation of the toxicity effects on non-target organisms, using tests for both acute effects (e.g., measuring mortality rate) and chronic effects (e.g., exposing to different concentrations of a chemical over a prolonged period of time and measuring growth index or reproduction rate) (44). These latter are particularly scarce, possibly owing to their complex experimental procedure (45). There is a huge concern associated to the environmental contamination with antibiotics and the possible development of resistance mechanisms by bacteria. This could subsequently compromise public health in terms of treatment effectiveness. Chemical risk assessment methods for substances found in water involve establishing different points of
General Introduction 15 departure (PoD) or toxicological endpoints such as the predicted environmental concentration (PEC), predicted no-effect concentration (PNEC), no-observed effect concentration (NOEC), lowest observed effect concentration (LOEC), half-maximal lethal concentration (LC50) and/or half-maximal effective concentration (EC50) (9, 37). For instance, Yamashita et al. (46) evaluated the inhibition of the algae P. subcapitata by levofloxacin and clarithromycin and the results obtained showed that the latter presented higher toxicity, with an EC50 of 11 µg L-1 and a LOEC and a NOEC of 6.3 and 3.1 µg L-1, respectively. On the other hand, Isidori et al. (47) tested five antibiotics (oxytetracyclin, sulfamethoxazole, ofloxacin, lincomycin and clarithromycin) on aquatic organisms from different trophic levels (bacteria, algae, rotifers, crustaceans and fish). Results showed that the antibiotics were less active against rotifers, crustaceans and fish, and that while acute toxicity levels were in the order of mg L-1, chronic toxicity became visible at concentrations in the order of µg L-1, mainly for algae. Another therapeutic class raising great awareness is sex hormones, extremely active biological compounds capable of inducing intense therapeutic effect at very low doses. Within this group, estrogens are the most commonly found in the environment, existing either as natural or synthetic substances and acting as endocrine-disrupting compounds (48). For example, Kidd et al. (49) developed a 7-year experiment which demonstrated that the chronic exposure of fathead minnow to 5-6 ng L-1 of ethinylestradiol (EE2) led to feminization of males fish and altered oogenesis in females. EE2 is a synthetic estrogen present in oral contraceptive pills with proved estrogenic effects in fish. In the group of NSAIDs, 28-days chronic toxicity tests conducted on the rainbow trout (Oncorhynchus mykiss) by exposure to only 1 µg L-1 of diclofenac resulted in cytological changes in the liver, kidneys and gills (50). Regarding the toxicity of lipid regulators such as statins, larval and adult grass shrimp (Palaemonetes pugio) 96-hours exposure to simvastatin resulted in a LC50 of 1.18 and 10 mg L-1, respectively (51). Also the widely used antiepileptic carbamazepine showed to be carcinogenic to rats, though not presenting mutagenic properties in mammals. Furthermore, it proved to be lethal to zebrafish at 43 µg L-1 and induce sub-lethal changes in Daphnia sp. at 92 µg L-1 (52). Finally, it must be highlighted that pharmaceutical compounds do not occur alone in the environment, but in mixtures of different active substances, their metabolites and transformation products, which may display potentially additive, antagonistic and even synergistic effects among them and with other naturally occurring compounds (53).
Chapter 1 16 1.2.1.5. Legislation Nowadays, there is still a big hiatus in legislation concerning the environmental contamination by pharmaceutical compounds. This gap is probably due to the scarcity of studies and available data which could allow a coherent profiling and a precise approach to this issue. Up until 1992, there were no requirements imposed by the EU legislation towards the pharmaceutical industry, demanding the inclusion of environmental risk assessment data to obtain the Marketing Authorization of a new pharmaceutical product. This was first achieved through Directive 92/18/EEC (54), but only regarding veterinary products. Consequently, EMEA published a Note for Guidance (55), establishing the guidelines for the accurate risk assessment of veterinary medicinal products. The extension of this procedure to pharmaceutical products for human use was performed by the EC through Directive 2001/83/EC, later amended by Directive 2004/27/EC (11). These documents required the Marketing Authorization of a new pharmaceutical product for human use to be accompanied by a proper environmental risk assessment study, according to the guidelines set out by EMEA. Evaluation of the environmental impact of both human and veterinary medicines embraced two phases: Phase I, concerning the environmental exposure assessment to pharmaceutical compounds and/or their metabolites; Phase II, comprising the study of their effects and fate in the environment. This second phase was still sub-divided into two parts: Tier A, regarding the evaluation of potential effects and fate of the respective pharmaceuticals and/or metabolites, and Tier B, focusing on their potential effects on fauna and flora within specific environmental compartments. However, only pharmaceutical products for human use presenting a PEC in surface waters equal to or above 0.01 µg L-1 are requested to undergo Phase II studies (56) (Fig. 1.4). In the USA, concerns over the environmental presence of pharmaceutical compounds have long been raised. At the moment, regulation established by the USA Food and Drug Administration (FDA) requires the inclusion of the environmental risk assessment study in order to obtain the Marketing Authorization, according to the Guidance for Industry - Environmental Assessment of Human Drug and Biologic Applications (57). Nonetheless, this study is only mandatory for pharmaceuticals with PEC superior to 1 µg L-1 (Fig. 1.5).
General Introduction 17 Fig. 1.4. Graphical representation of EMEA guidelines (58). * Committee for Proprietary Medicinal Products (CPMP) was the former designation for the EMEA Committee.
Chapter 1 18 Fig. 1.5. FDA tiered approach to fate and effects testing (58). * Expected Introductory Concentration (EIC); ** Maximum Expected Environmental Concentration (MEEC); *** Centre for Drug Evaluation and Research (CDER)/ Centre for Biologics Evaluation and Research (CBER).
General Introduction 19 1.3. Thesis 1.3.1. Main Objectives Considering the state-of-the-art a priori, as well as the main knowledge gaps and analytical obstacles, briefly exposed in the preceding sections, the ultimate purpose of this Ph.D. work was to significantly contribute to the enlightening of the problematic concerning the aquatic compartment contamination by pharmaceutical residues. This implies the development of novel analytical and experimental approaches to streamlining the detection, quantification and removal of pharmaceuticals from water samples, consequently providing new data on the impact and fate of these compounds in the environment. Research holds an important role in the support of new legal decisions, helping to reach a balance between all the actors involved, including official authorities, environmental managers and industry officers. The restrictive legislation that will be adopted at EU level will oblige to spend huge financial budgets and technical resources. Bearing it in mind, a multidisciplinary strategy was adopted with the aim of help finding answers to the key-topics in environmental studies: occurrence, behaviour and toxicity. Hence, the following tasks were addressed within this Ph.D. project: 1) Development of a multiresidue method for the analysis of pharmaceutical compounds in water, using an adequate sample preparation/extraction procedure, followed by the analysis through liquid chromatography–tandem mass spectrometry (LC-tandem MS or LC-MS/MS or LC-MS2). Sample preparation processes are crucial for a good separation of target analytes from the remaining sample compounds. Basically, they comprise the following steps: i) extraction; ii) concentration; iii) cleansing; iv) extract injection. Nowadays, a wide variety of adsorbents/coatings is commercially available, allowing the efficient and selective extraction of compounds, within a broad range of polarities. These solidphase adsorbents will be thoroughly studied during this work. One of the biggest limitations in the analysis of emerging pollutants relies on the lack of analytical methodologies able to detect and quantify contaminants present in low concentrations. One prerequisite for the quality assessment and monitoring of wastewaters, surface and drinking waters, is the capability to develop multiresidue methods, capable of providing rapid and accurate results. Even though there has been an increase of such methodologies reported in literature over the past few years, there
Chapter 1 20 is still the need to improve both sensitivity and selectivity in the case of more complex matrices as wastewaters. 2) Participation in inter-laboratory exercises (ILEs) promoted by the Joint Research Centre (JRC) and organized under the scope of European projects, with the purpose of establishing a valid system for quality control and methods’ harmonization. 3) Monitoring campaigns encompassing WWTP samples and receiving surface water bodies, impact assessment and further evaluation of the constraints influencing a representative sampling plan. During the last few decades, the scientific community has started reporting the presence and concentrations of pharmaceutical residues in WWTP effluents, surface waters, its fauna and flora. In the case of WWTPs, the great variability of results achieved is usually attributed to the type of treatment process employed, as well as its location, social/economic factors, most commercialized medicines and other demographic parameters. Regarding sampling processes, either grab or composite samples can be collected using automatic samplers, according to the results obtained after preliminary exploratory examinations. Recently, passive sampling is considered a very promising technique. However, it is still in a preliminary experimental phase. 4) Study of the photodegradation kinetic of several pharmaceutical compounds, at labscale and pre-industrial pilot scale. The influence of several factors must be taken into consideration, such as temperature, pH, presence of dissolved and particulate material and radiation source. The results attained may be of great relevance for the conception and design of future WTPs and WWTPs. 5) Identification of phototreatment by-products and elucidation of the involved photo(cata)lytic degradation mechanisms. Herein, there is the need to take profit of advanced high-resolution techniques, such as ultra-performance liquid chromatography coupled to quadrupole time-of-flight mass spectrometry (UPLC-QToF-MS), which enable an accurate highly sensitive nontarget screening approach.
General Introduction 21 1.3.2. Organization The presented Ph.D. thesis is organized in chapters, each one dealing with a different topic and including its own bibliographic references. In Chapter 1, “General Introduction”, a broad overview is presented concerning water quality policy and the recent problematic of emerging water micropollutants. Special emphasis is given to pharmaceutical compounds, including their therapeutic classification and identification of main environmental contamination sources. Moreover, a brief review on pharmaceuticals’ environmental distribution, occurrence and fate, as well as their consequent (eco)toxicological effects, already demonstrated in different species, is also displayed. A final note regarding the still incomplete legislation on this issue is presented in the end. In Chapter 2, “Analytical Method Development for Detection and Quantification of Pharmaceuticals in Waters”, the different stages of a multiresidue analytical method for pharmaceuticals’ quantification are approached. From the optimization of the cleanup and extraction procedure, to the selection of the most suitable separation and detection techniques, different aspects are taken into consideration, according to the specific target analytes and the nature and complexity of samples’ matrix. This chapter also includes the publication “Cleanup strategies and advantages in the determination of several therapeutic classes of pharmaceuticals in wastewater samples by SPE-LC-MS/MS”, which reports on the work conducted during the development, optimization and validation of the analytical method, applied to the detection and quantification of diverse pharmaceutical compounds in wastewater samples. In Chapter 3, “Monitoring of Pharmaceuticals in Municipal Wastewaters and Surface Waters”, the monitoring results achieved with the developed method on Febros WWTP, and its impact on Febros river and later on Douro river, are assessed regarding the pharmaceutical content. Furthermore, different sampling methodologies are comparatively evaluated. This chapter also includes the scientific paper in Portuguese language “Effects of treated domestic effluents and tributaries on the contamination of Douro river with pharmaceutical compounds – Mitigation processes”, concerning the results aforementioned. Finally, the participation in two inter-laboratory exercises, achieved results and withdrawn conclusions regarding the method reliability are also reported.
Chapter 1 22 In Chapter 4, “Photo-Remediation of Contaminated Waters using Advanced Oxidation Processes (AOPs)”, various advanced oxidation processes for pharmaceuticals’ mitigation in water samples are described. Special attention is given to heterogeneous photocatalytic processes, while the influence of different experimental variables on the overall treatment is also studied. This chapter also includes two published papers “Photolytic and TiO2- Assisted Photocatalytic Oxidation of the Anxiolytic Drug Lorazepam (Lorenin® pills) under Artificial UV Light and Natural Sunlight: A Comparative and Comprehensive Study” and “Suspended TiO2-Assisted Photocatalytic Degradation of Emerging Contaminants in a Municipal WWTP Effluent using a solar Pilot Plant with CPCs”, concerning the optimization of TiO2-assisted photocatalytic degradation processes of lorazepam, a recalcitrant anxiolytic drug, and later application to the treatment of a municipal WWTP effluent, contaminated with diverse pharmaceuticals. In Chapter 5, “Phototreatment Mechanisms and By-Products Structural Elucidation”, lorazepam photo(cata)lytic degradation mechanisms are enlightened, while the resulting by-products are also identified and structurally elucidated. This chapter is mostly composed by the publication “Lorazepam Photofate under Photolysis and TiO2-assisted Photocatalysis: By-products’ Identification and Evolution Profiles during Phototreatment of a Contaminated WWTP Effluent”. In Chapter 6, “Occurrence of Emerging Pollutants’ By-Products in Water Resources”, a résumé of a broad review regarding emerging pollutants’ by-products environmental occurrence in different water sources is presented. In Chapter 7, “Final Conclusions and Remarks: Work Novelty, Knowledge Gaps and Future Research”, the key achievements of the work are presented, as well as the main knowledge lacunas and missing scientific evidences which may be filled in and enlightened during some future experiments’ proposals.
General Introduction 23 1.4. References 1. The EU Water Framework Directive - integrated river basin management for Europe. Online at: http://ec.europa.eu/environment/water/water-framework/index_en.html (last accessed on December 12, 2012). 2. Barceló D. Emerging pollutants in water analysis. Trends in Analytical Chemistry. 2003;22(10):xiv-xvi. 3. Richardson SD. Environmental mass spectrometry: Emerging contaminants and current issues. Analytical Chemistry. 2002;74(12):2719-41. 4. Richardson SD. Water analysis: Emerging contaminants and current issues. Analytical Chemistry. 2007;79(12):4295-323. 5. Ternes TA, Joss A. Human Pharmaceuticals, Hormones and Fragrances - The challenge of micropollutants in urban water management. London, U.K.2006. 6. Kot-Wasik A, Debska J, Namieśnik J. Analytical techniques in studies of the environmental fate of pharmaceuticals and personal-care products. Trends in Analytical Chemistry. 2007;26(6):557-68. 7. Rodil R, Quintana JB, López-Mahía P, Muniategui-Lorenzo S, Prada-Rodríguez D. Multiresidue analytical method for the determination of emerging pollutants in water by solid-phase extraction and liquid chromatography–tandem mass spectrometry. Journal of Chromatography A. 2009;1216(14):2958-69. 8. Why do we need to address emerging substances? Online at: http://www.normannetwork.net/index_php.php?module=public/about_us/comment_substances&menu2=public/ about_us/about_us (last accessed on December 12, 2012). 9. Santos LHMLM, Araújo AN, Fachini A, Pena A, Delerue-Matos C, Montenegro MCBSM. Ecotoxicological aspects related to the presence of pharmaceuticals in the aquatic environment. Journal of Hazardous Materials. 2010;175(1-3):45-95. 10. Mompelat S, Le Bot B, Thomas O. Occurrence and fate of pharmaceutical products and byproducts, from resource to drinking water. Environ Int. 2009;35(5):803-14. 11. EudraLex – Pharmaceutical Legislation Medicinal Products for Human Use, Online at: http://ec.europa.eu/health/human-use/legal-framework/index_en.htm (last accessed on December 12, 2012). 12. Prontuário terapêutico - 10, Março 2011. Online at: http://www.infarmed.pt/portal/page/portal/INFARMED/PUBLICACOES/PRONTUARIO/pt1 0_web.pdf (last accessed on December 12, 2012). 13. Kummerer K, editor. Pharmaceuticals in the environment - Sources, fate, effects and risks. Berlin, Heidelberg, New York: Springer; 2004. 14. Sedlak DL, Pinkston K, Huang CH. Occurence survey of pharmaceutically active compounds. Denver, CO: Awwa Research Foundation. 2005. 15. Stockholm County Council 2005. Environmentally Classified Pharmaceuticals, Department of the Environment, Stockholm, Sweden. 16. Sedlak DL, Huang CH, Pinkston K. Strategies for selecting pharmaceuticals to assess attenuation during indirect potable water reuse. In: Kummerer K, editor. Pharmaceuticals in the environment - Sources, fate, effects and risks. Berlin, Heidelberg, New York: Springer; 2004. p. 107-20. 17. Jørgensen SE, Halling-Sørensen B. Drugs in the environment. Chemosphere. 2000;40(7):691-9. 18. Khan SJ, Ongerth JE. Estimation of pharmaceutical residues in primary and secondary sewage sludge based on quantities of use and fugacity modelling. Water Science and Technology. 2002;46:105-13. 19. McArdell CS, Molnar E, Suter MJF, Giger W. Occurence and Fate of Macrolide Antibiotics in Wastewater Treatment Plants and in the Glatt Valley Watershed, Switzerland. Environmental Science and Technology. 2003;37:5479-86.
Chapter 2 30 SPE also has several drawbacks, such as the loss of more polar compounds during sample percolation, the co-extraction of matrix interferences, filters clogging, a noticeable time consumption for the extraction of large water volume, a significant consumption of organic solvents, an incomplete desorption of target analytes, the loss of volatile compounds and the incomplete recovery of the dry extract (1, 18). Fig. 2.1. Illustration of the SPE process: - interferences; - target analytes. 1) Retention of the analyte: analyte molecules are enriched on the adsorbent > interfering components and solvent molecules (matrix) are not retained > remaining interfering components are washed from the adsorbent > the analyte is removed from the adsorbent by elution with a suitable solvent (e.g., methanol, tetrahydrofuran (THF), -propanol); 2) Retention of interfering components: analyte molecules show no interaction with the adsorbent > interfering components and solvent molecules (matrix) are retained > analyte molecules are “washed” from the adsorbent > the solid phase is simply used to “filter” the sample (19). 2.2.1.2. Adsorbents Several SPE adsorbents have been developed. Among them, the most adequate for the extraction of organic compounds are the reverse phase sorbents, from the conventional
Analytical Method Development for Detection and Quantification of Pharmaceuticals in Waters 31 alkyl-modified silica materials (C8 and C18 non-polar phase), to the more recent polymeric sorbents, which improve the retention of polar compounds due to the high surface area and cross-links (e.g., poly(styrene-divinylbenzene) (PS-DVB)), and carbon materials (e.g., graphitised carbon black (GCB)) (18, 20). Polymeric sorbents present different functional groups combined with a polymeric skeleton, allowing more than one kind of interaction between analytes and the sorbent (hydrophilic, hydrophobic and ππ interactions) (18). Some examples of these sorbents are displayed in Table 2.1. Table 2.1. Commercial brands of SPE extraction columns constituted by polymeric adsorbents, frequently utilized in the extraction of pharmaceutical compounds from water samples Cartridge commercial designation Adsorbent composition Application Reference Strata-X Polydivinylbenzene resin chemically modified with piperidone groups Extraction of non-steroidal antiinflammatory drugs (NSAIDs) from surface and groundwaters (21) Evolute ABN Surface modified polystyrene divinylbenzene polymer Extraction of human pharmaceuticals from river waters (22) Oasis HLB (Hydrophiliclipophilic-balanced reversed-phase sorbent for acids, bases and neutrals) Hydrophobic-lipophilic copolymer of divynilbenzene and vinylpirrolidone Extraction of veterinary pharmaceuticals from wastewaters (23) Oasis MCX (Mixed-mode cation exchange sorbent for bases) Poly(divynilbenzeneco-N- vinylpirrolidone) copolymer on which strong-cation exchanging sulfonic groups are convalently bonded Extraction of basic/neutral pharmaceuticals and illicit drugs from surface waters (15) Oasis MAX (Mixed-mode anion exchange sorbent for acids) Poly(divynilbenzeneco-N- vinylpirrolidone) copolymer on which strong-anion exchanging quaternary ammonium groups are convalently bonded Extraction of human pharmaceuticals from wastewaters (24) (Paper 1)
Chapter 2 32 The recent trends in the development of SPE (and cleanup) adsorbents contemplate the extraction of increasingly polar compounds and wider polarity ranges, thus extending the scope of SPE and further supporting the improvement of multiresidue methods. Some recent developments in this field will now be presented. Monolithic (particle free) SPE is a recent strategy to deal with selectivity and sensitivity issues regarding liquid chromatography-mass spectrometry (LC-MS) bio-analytical methods. In contrast to SPE cartridges, monolithic stationary phases are presented in the format of SPE discs (available in 15 different sorbent chemistries, such as C2,C8 and C18) (25) and 96-wells plates, more easily automated. These materials present several advantages: i) low cost, ii) mechanical robustness and high stability, iii) no void volumes, iv) easy control of the porous properties of adsorbents, v) high hydraulic permeability and vi) dominance of the convection over the diffusion mechanism of mass-exchange under dynamic conditions that allow the separation at extremely high flow rates (26). Extraction selectivity may also be increased using molecularly imprinted SPE (MISPE). Molecularly imprinted polymers (MIPs) are synthetic polymers with a predetermined selectivity for a given analyte or group of structurally related compounds. They enclose specific recognition sites which are complementary in shape, size and functional group, to the analyte or analytes of interest (18, 26). MIPs have already been synthesised for several emerging pollutants, including pharmaceuticals and endocrine disrupting compounds (EDCs) (27). Immuno-affinity chromatography (IAC) is a technique based on the affinity between antibodies and antigens, mainly applied as a cleanup technique in food and biological analyses. Although presenting high selectivity and enabling the better cleanup attained for polar compounds in complex matrices, its main drawbacks include small breakthrough volumes and low stability against solvents, strong pH and temperature (26, 28). Restricted access materials (RAMs) are frequently used as pre-columns in columnswitching LC systems. They consist on a support type which allows direct injection of complex samples by limiting the accessibility of interaction sites within the pores only to small molecules (26). Most recently, new nanomaterials have been exploited as well to integrate new extraction and cleanup technologies (29).
Analytical Method Development for Detection and Quantification of Pharmaceuticals in Waters 33 Besides being used for sample extraction (Fig. 2.1 -1)), SPE can also be applied to sample cleanup (Fig.2.1-2)); in this case, normal phase adsorbents such as silica, alumina and florisil are usually preferred (10) (Paper 1). It must also be noticed that pH control plays an important role in SPE, since it governs the dissociation of ionisable compounds, therefore affecting their hydrophobicity and interaction with the sorbent. For instance, when using a reverse phase sorbent, the retention of a compound increases for pH values favouring their non-ionised form (21). Lately, the development of on-line SPE coupled to LC-MS/MS methodologies has further increased the overall method robustness. Besides saving time on the analysis, it offered an alternative as a more environmentally friendly method for pharmaceutical determinations in waters. In fact, on-line SPE requires low sample and solvent volumes (14). 2.2.2. Solid-Phase Microextraction Solid-phase microextraction (SPME) was introduced by Pawliszyn and co-workers (30), in 1989, representing a significant step forward in terms of simplification and miniaturization of the analytic process. SPME represents a solvent-free sampling technique, often used in the extraction of organic micropollutants from water samples. It employs a fused-silica fibre coated on the outside with an appropriate stationary phase. SPME fibres are commercially available in a “syringe-like” format, which allows its protection during septum puncture (30). Volatile analytes can be emitted from the sample and isolated from the headspace or adsorbed by direct immersion into the liquid sample and concentrated in the fibre coating. SPME association with gas chromatography-mass spectrometry (GC-MS) is overwhelming comparatively to LC-MS (31, 32). Therefore, after the extraction, thermal desorption in the hot GC injection port usually follows (33). Fig. 2.2 shows a schematic representation of a SPME set-up and operating modes, as well as the composition of a SPME fibre.
Chapter 2 34 Fig. 2.2. SPME components: on the left – scheme of a SPME set-up; in the centre – detail of a SPME fibre; on the right – operating modes of SPME: A – immersion, B – headspace. SPME is associated with strong matrix effects and, consequently, complications in quantification. Moreover, the variability of detection limits for several analytes in dependent upon the equilibrium between the fibre coating and the matrix (33). Different factors can influence the efficiency of the SPME process: i) the type of adsorbent polymer (34), ii) sample stirring (35), iii) sample temperature (9), iv) sample pH (34), v) sample ionic strength (36) and vi) the duration of the extraction (37). On the other hand, SPME presents as main advantages its simplicity, low cost and low time consumption, high selectivity and sensitivity, when coupled to adequate detection methods (37). It integrates, in only one continuous process, sampling, extraction, concentration and introduction of target analytes in the chromatographic system for separation and detection. All these pros, along with its “solventless” character, make of SPME a rather promising extraction technique (37, 38). SPME successful application on the extraction of various emerging contaminants, including pharmaceuticals, from water samples has also been reported in several studies (39-42).
Analytical Method Development for Detection and Quantification of Pharmaceuticals in Waters 35 2.2.3. Other Extraction Techniques Apart from SPE, which is the most used technique in pharmaceutical analyses, SPME has also earned great acknowledgment in this field, presenting further increasing development potential. On the contrary, more conventional techniques are now less employed. Such is the case of liquid-liquid extraction (LLE), which presents several drawbacks, including the emulsion formation, the use of large volumes of solvents, environmentally unfriendly, high time consumption and high cost. Liquid-phase microextraction (LPME) is a solventminimised sample pre-treatment modality of LLE, requiring some µL of organic solvent to concentrate analytes (43). Nevertheless, LPME shows a low precision (44). Other more recent extraction techniques include dispersive liquid-liquid microextraction (DLLME) and stir membrane liquid-liquid microextraction (SM-LLME), both presenting high extraction efficiencies (18). 2.3. Separation and Detection of Pharmaceutical Compounds In order to determine very low concentration levels of pharmaceutical compounds in waters, the appropriate sample pre-treatment is fundamental. Afterwards, separation, detection, quantification and confirmation of analytes follow. The separation of organic compounds is usually achieved either by liquid or gas chromatography, according to analytes’ polarity, volatility and decomposition at high temperatures. While volatile or semi-volatile compounds may be analysed by GC, more polar or thermolable non-polar compounds are analysed by LC, with no need of prior derivatization. Generally, every compound is amenable to LC, although not all are directly and easily detected by the available detectors (10). Fast and high resolution LC systems have been developed, without compromising resolution and separation efficiency. They include liquid chromatography at ultra high pressures, using sub-2-µm particle packed columns (UHPLC) (Paper 5), the use of monolith columns, the use of fused core columns and high-temperature liquid chromatography (HTLC) (26).
Chapter 2 36 Nowadays, the detection at sub-ppt concentrations is practically routine for many organic contaminants, including pharmaceutical compounds in water and wastewater, whereas some methods can even detect few hundred fentograms of some analytes (45). These impressive improvements in method detection limits are mostly attributed to the use of hyphenated chromatography-mass spectrometry techniques, which currently constitute the methods of election for the analysis of trace organic contaminants in environmental samples (26). Fig. 2.3. presents some examples of pharmaceutical compounds usually detected in water and wastewater, using these advanced analytical methods (46). Fig. 2.3. Analytical methods applied for the most common pharmaceutical compounds present in water and wastewater (from Fatta et al. (46)).
Analytical Method Development for Detection and Quantification of Pharmaceuticals in Waters 37 Regarding mass spectrometry (MS), it is an instrumental technique based on the separation of ions according to their mass-to-charge ratios (m/z), in vacuum, in the gas phase. After ionization, the (pseudo)molecular ions formed can undergo further fragmentation and all these formed ions are then separated in the mass spectrometer according to their m/z and are detected. A mass spectrometer is composed by three fundamental parts: an ion source, a mass analyser and a detector (Fig. 2.4) (33). Several ionisation methods are available. Hence, the fitness of the ion source will depend on the polarity, molecular weight and thermal lability of the analytes, as well as the complexity of the sample to be analysed. Concerning mass analysers, they can vary from quadrupole, quadrupole ion trap, linear quadrupole ion trap and magnetic sector, representing scanning instruments, to other type of mass analysers, such as time-of-flight (ToF) (Fig. 2.5), Fourier-transform ion cyclotron resonance analyser and orbitrap, which belong to the group of non-scanning instruments. Afterwards, the detector allows to monitor the ion current and finally records the data as mass spectra (33). Fig. 2.4. Overview of mass spectrometric techniques (from Cajka et al. (33)).
Chapter 2 38 Fig. 2.5. Time-of-flight mass analyser. (1) Pulse of ions from the orthogonal accelerator (spatial focusing); (2) separation of ions according to their flight times; (3) focusing of the kinetic energy of ions; (4) separation of focused ions according to their flight times, dependent on their weight (adapted from Cajka et al. (33)). Full-scan sensitivity, a low-femtomole level of sensitivity, high-mass resolution and mass accuracy (mass errors below 2 mDa or 5 ppm) can be achieved with QqToF-MS (47). Consequently, these instruments are often the first choice in studies devoted to the elucidation of degradation mechanisms of organic emerging pollutants, under environmental conditions. Several strategies can be employed for the identification of degradation products, taking profit of exact mass measurements, tandem mass fragmentation and typical “diagnostic ions”. This subject will be further addressed in Chapter 5, applied to the photo(cata)lytic degradation of the anxiolytic drug lorazepam (Paper 5). Herein follows Paper 1, which reports on the analytical approaches and obtained results during the optimization of a multiresidue method for the determination of several pharmaceutical compounds in wastewaters, giving particular emphasis to the cleanup step. Another work on this topic was also published by our group, although this one more specifically oriented to cleaner river waters (Annex II).
Analytical Method Development for Detection and Quantification of Pharmaceuticals in Waters 39 2.4. Paper 1 – Cleanup strategies and advantages in the determination of several therapeutic classes of pharmaceuticals in wastewater samples by SPE-LC- MS/MS Published Scientific Paper: Sousa, M.A., Gonçalves, C., Cunha, E., Hajšlová, J. and Alpendurada, M.F., Cleanup strategies and advantages in the determination of several therapeutic classes of pharmaceuticals in wastewater samples by SPE–LC–MS/MS. Analytical and Bioanalytical Chemistry, 2011. 399: 807-822.
__________________________________ Chapter 2 ________________________________________ 46
____ Analytical Method Development for Detection and Quantification of Pharmaceuticals in Water___ 47
__________________________________ Chapter 2 ________________________________________ 48
____ Analytical Method Development for Detection and Quantification of Pharmaceuticals in Water___ 49
__________________________________ Chapter 2 ________________________________________ 50
____ Analytical Method Development for Detection and Quantification of Pharmaceuticals in Water___ 51
__________________________________ Chapter 2 ________________________________________ 52
____ Analytical Method Development for Detection and Quantification of Pharmaceuticals in Water___ 53
__________________________________ Chapter 2 ________________________________________ 54
____ Analytical Method Development for Detection and Quantification of Pharmaceuticals in Water___ 55
Chapter 3 62 3.1. Water Sampling Sampling could be defined as the process of selecting a representative water sample with the appropriate volume to be conveniently transported and handled in the laboratory. It is an integral part of the analytical process, often representing the main contribution to the error of the analytical result (2). Proper sampling and analytical techniques are of fundamental importance in the characterization of water samples. Firstly, it is essential to ensure that the analysed sample is truly representative of the environment sampled. Once this requirement is met, it is then necessary to adopt the appropriate analytical procedures, specifically developed for water or wastewater analyses (3). The data collected during the sampling plan must be useful, thus meeting the objectives of the monitoring plan (4), reproducible and defensible, i.e., have associated documentation which validates the sampling procedure and provides information concerning the degree of accuracy and precision of the acquired data. While accuracy measures the closeness of results to the true value, precision measures the reproducibility of the results when the same sample is repeatedly analysed (3, 5). 3.1.1. Sampling Protocol A detailed sampling protocol must be developed, along with a quality assurance project plan (QAPP) (previously known as quality assurance/quality control - QA/QC), encompassing the following items (4, 5): i) sampling plan: including the number of sampling locations, number and type of samples, frequency and time intervals (e.g., real-time and/or time-delayed samples); ii) sample types and size: catch or grab samples, composite samples, integrated samples, separate samples for different analyses, sample volume, etc.; iii) sample labelling and chain of custody; iv) sampling methods: indicating the techniques and equipment to be used (e.g., manual, automatic, passive sampling);
Monitoring of Pharmaceuticals in Municipal Wastewaters and Surface Waters 63 v) sampling storage and preservation: referring the type of containers and preservation methods; vi) sample constituents: consisting of a list of parameters to be measured; vii) analytical methods: consisting of a list of field and laboratory test methods and procedures to be used. When designing a sampling plan, it is essential to clearly specify the objective of the exercise. For instance, one may aim to estimate maximum or mean concentrations, to detect changes or trends or even to provide a basis for industrial effluent charges. Nevertheless, to attain an accurate assessment of the scenario, it is usually necessary to generate composite samples. Integrated composite samples can be obtained by bulking individual samples collected at known time intervals and in proportion to the appropriate flow of the stream, along a certain period of time. Likewise, in cases when sampling rivers with large channel sections, it is desirable to collect samples at different points across the section and at several depths. Furthermore, it is always necessary to bear in mind the resources available for both sampling and analysis procedures. Consequently, a realistic level of uncertainty of the results must be set, based upon the established goals (3). Whether high levels of uncertainty lead to unacceptable levels of reliability in the decisions based upon them depends on a rigorous evaluation of the fitness for purpose (6-8). The potential sources of error during sampling procedures are of diverse origin. Table 3.1 displays some error sources which should be taken into consideration, according to a review study by Madrid and Zayas (2): Table 3.1. Error sources during sampling procedures Sampling stage (decision making) Possible sources of error Definition and sub-division of the field Heterogeneity of the sample; spatial and/or temporal change of pollutants: “hot spots” Sampling method No representative statistics; skew distribution; contamination or analyte loss Number of samples Few replicates; absence of representativeness Sample mass Absence of representativeness Moment of sampling Seasonal changes; climatic conditions Experimental conditions Matrix effects; lixiviation or irreproducible deposits Bottling Contamination or extraction by the equipment or container material; volatilization Storage during sampling Contamination or losses by volatilization; chemical reactions (change of species)
Chapter 3 64 3.1.2. Materials and Equipments Ideally, all analyses should be conducted immediately after sample collection, so that the obtained results are a true assessment of the nature of the water sample in situ. In the real world, only some analytical determinations are viable on site, such as temperature, pH, conductivity, dissolved oxygen (DO), salinity, among others. These parameters cannot be adequately determined after transportation to the laboratory (9), thus multiparameter probes are usually taken to the field (Fig. 3.1 (a)). In the case of surface waters, samples are often collected by direct filling of the sample bottle, at ca. 0.5 m depth. For deep water bodies or in places of difficult access, special sampling flasks can be used: these are lowered in a closed state, on a rope or steel cable, and then remotely triggered to open (Fig. 3.1 (b)). Various automatic devices are also available to collect composite samples, being operated on either a time basis or on a flowproportional basis (Fig. 3.1 (c)). These instruments are of particular interest during the sampling of WWTP effluent discharges, since these are often intermittent in nature. In such circumstances, it is fundamental to fully understand the nature of the operations producing the discharge, in order to draw an appropriate sampling programme and consequently obtain a true picture of the discharge. Fig. 3.1. Sampling devices. (a) multiparameter field probe; (b) sampling flask with remotely triggered opening; (c) automatic sampler used to collect and refrigerate grab or composite samples over a 24-h period. In the last two decades, passive sampling methods have shown much promise as tools for measuring aqueous concentrations of a wide variety of pollutants. In its broadest sense, passive sampling includes any sampling technique based on free flow of analyte molecules
Monitoring of Pharmaceuticals in Municipal Wastewaters and Surface Waters 65 from the sampled medium to a receiving phase in a sampling device, due to a difference between the chemical potentials of the analyte in the two media. Thus, there is no need of any energy source other than this chemical potential difference. The net flow of analyte molecules between both media continues until reaching the equilibrium state or until the sampling period is stopped (10). The main advantages of passive sampling compared to active water sampling, regarding the collection of organic pollutants, include the following aspects: i) it is based on diffusion processes (no power supply), ii) it occurs in situ (selective analyte isolation and pre-concentration (11)) and it is often time-integrative, iii) it shows low variability, due to a standardised configuration of the synthetic samplers, and finally iv) it mimics the respiratory exposure for aquatic organisms, since, traditionally, it includes a mass transfer and rate limiting membrane(s) in their design (12). Summing up, passive samplers simplify the operations performed at the sampling site. Table 3.2 presents an overview of some of the most used passive samplers for monitoring of organic pollutants, as reviewed by Namiesnik et al. (13). Table 3.2. Passive sampling devices for organic contaminants Sampler Full Name Analytes References Chemcatcher Universal passive sampler using Empore disk Polar and non-polar organics (14) Ecoscope A sampler based on solventfilled dialysis membranes and chelating sorbent discs Non-polar organics (15) MESCO Membrane-enclosed sorptive coating PAHs, PCBs, organochlorine pesticides (16) PDB Passive diffusion bag samplers Polar organic compounds, VOCs, metals, trace elements (17) PISCES Passive in situ concentrationextraction sampler PCBs (18) POCIS Polar organic chemical integrative sampler Herbicides and pharmaceuticals with log KOW < 3 (19) SPATT Solid-phase adsorption toxin tracking Polar phytotoxins (20) SPMD Semi-permeable membrane devices Hydrophobic semi-volatile organic compounds (21) Lastly, it must be highlighted that the physical, chemical and biological integrity of samples during interim periods between sampling and analysis shall be preserved. Changes in sample composition can be retarded by storage at low temperatures (4 ºC), in the dark. The more polluted a sample is, the shorter the time which can be allowed between sampling and analysis, in order to avoid significant errors (3).
Chapter 3 66 3.2. A Portuguese Case Study 3.2.1. Monitoring Sites Monitoring water quality is carried out for various purposes and its design will depend on its later uses. Obviously the cost of monitoring has to be related to the ultimate benefit. In our study, both wastewater and surface water samples were analysed in terms of pharmaceutical content. While the first were collected at Febros WWTP, the latter were sampled in Febros river and Douro river. 3.2.1.1. Febros WWTP Febros WWTP was built in 2003, in the city of Vila Nova de Gaia. It was dimensioned for 80 000 inhabitants, designed to treat a wastewater flow of 39 605 m3 per day and a load corresponding to 5 334 kg per day in terms of biochemical oxygen demand (BOD5). Febros WWTP (Fig. 3.2) comprises: an harrowing/sieving, desanding and degreasing system; 2 aeration tanks with a capacity of 6 053 m3 each; 3 secondary clarifiers with a 25 m diameter; 1 thickener with a diameter of 14 m; sludge mechanical dehydration with 2 centrifuges; sludge storage silo of 50 m3 capacity; and, finally, an odour deodorizing treatment step using a column of activated carbon for 6 380 m3 h-1. Fig. 3.2. Image of Febros WWTP, which discharges to Febros river, flowing on the righthand side. In the top-right part of the image one can identify Douro river, dividing the city of Vila Nova de Gaia (left bank) from the city of Porto (right bank) (22).
Monitoring of Pharmaceuticals in Municipal Wastewaters and Surface Waters 67 3.2.1.2. Febros River and Douro River Fig. 3.3. Geographical location of the case study – Portuguese hydrographical region no. 3: Douro. Table 3.3. GPS coordinates of the sampling spots in Febros river, identified in Fig. 3.3 Sampling Spots Latitude Longitude Upstream WWTP F1 41° 7'2.92" N 8°34'11.48" W Downstream WWTP F2 41° 7'8.75" N 8°34'14.42" W F3 41° 7'15.52" N 8°34'13.24" W F4 41° 7'21.04" N 8°34'9.41" W F5 41° 7'24.65" N 8°34'9.72" W * Sampling Spot H (latitude 41º08’00.89’’ N; longitude 8º34’20.77’’ W): located in Douro River, after the debouching site of Febros River. The remaining sampling spots in Douro River were not part of the present work, but were included in the monitoring campaigns addressed in the study by Gonçalves et al. (Annex II).
Chapter 3 68 The study area is situated in the hydrographical region no. 3 – Douro (Fig. 3.3). Douro river (Fig. 3.4) is the third longest river in the Iberian Peninsula, with a total length of 927 km. Douro riverhead is located in Sierra de Urbión, Sória (Spain), at 2 080 m of altitude, the Portuguese river basin is 18 643 km2 (total 97 603 km2) and has an average flow of 450 m3 s-1 (ranging between 200 and 710 m3 s-1). It crosses the northern region of Portugal, flowing towards the Atlantic Ocean, with the river mouth close to the cities of Vila Nova de Gaia and Porto. Fig. 3.4. Image of Douro river. In contrast, Febros river (Fig. 3.5) is a small tributary on the left bank of Douro river. Febros riverhead is located in Seixezelo and the river mouth is in the Esteiro de Avintes. This river also crosses the Natural Park of Vila Nova de Gaia. Fig. 3.5. Image of Febros river, close to the river mouth.
Monitoring of Pharmaceuticals in Municipal Wastewaters and Surface Waters 69 Hence, due to Febros river low flow and small basin area, a high impact caused by the discharge of the WWTP effluents was predictable from the beginning. Nonetheless, the construction of Febros WWTP proved to be quite advantageous in the remediation of the river. The removal of contaminants of both domestic and industrial origin, in collaboration with the natural city park, allowed Febros river repopulation with a mammal species long disappeared in this river - the otter. On the other hand, a strong dilution effect occurs during Febros river debouching into Douro river. Therefore, its potential impact on Douro river, in terms of pharmaceutical content, was expected to be negligible. 3.2.2. Methodology 3.2.2.1. Sampling At Febros WWTP, both grab and composite (400 mL per hour, over 24-h) influent and treated effluent samples were collected from May 2009 to March 2011. Concerning surface waters, grab samples were collected in 5 sampling spots in Febros river (1 upstream Febros WWTP and 4 downstream Febros WWTP) and 12 sampling spots in Douro river (7 upstream the debouching point of Febros river and 5 others downstream this point), during two sampling campaigns conducted in 2010 (spring and winter) (please refer to Fig. 3.3 and Table 3.3). The sampling plan included the collection of 2.5 L sample for wastewaters and 1 L for surface waters (in this case, collected at 1 m of depth), transport to the laboratory in icepacked containers and subsequent preservation at 4 ºC until analysis (max. 3 days after sampling). Moreover, influent and effluent wastewater samples were filtered using glass fibre filters (Ø 1.2 µm), immediately after arrival to the laboratory. 3.2.2.2. Analytical Method – SPE-LC-MS/MS The analytical methods developed and optimised for the determination of 23 pharmaceutical compounds potentially present in the collected samples were described in the previous chapter. Briefly, they consisted on an initial cleanup step (optional in the case of surface water samples), followed by a solid-phase extraction (SPE) procedure, in order
Chapter 3 70 to remove most interferences and concentrate the target analytes, and subsequent analysis/quantification by high performance liquid chromatography coupled to mass spectrometry (HPLC-MS). For wastewater samples, the cleanup and SPE processes were simultaneously carried out using Oasis MAX columns. Oasis MAX is a mixed-mode sorbent consisting of a strong anion exchanger (quaternary amine) and a non-polar N-vinylpyrrolidone/divinylbenzene polymer. It has been optimised to achieve higher selectivity and sensitivity for extracting acidic compounds with anion exchange groups, while basic and neutral compounds are retained by the polymeric sorbent. In the case of surface water samples, whenever necessary, cleanup was achieved using Bakerbond 1º,2º-Amino (weak-anion exchange, WAX) cartridges, while SPE procedure was conducted with JTBaker H2Ophilic extraction columns (23). Besides the sorbent, other extraction parameters were as well optimised, including the pH (pH 2), percolation flow (10 mL min-1) and the breakthrough volume of 50 mL for WWTP influents, 100 mL for WWTP effluents and 200 mL for river waters. The separation, identification and quantification of pharmaceuticals was performed by HPLC-ESI-MS/MS (ion-trap detector), assuring the unequivocal confirmation of each compound through the comparison with reference spectra obtained for the respective standards. The final validated analytical method proved to be sensitive, selective, precise, accurate and robust for the determination of the 23 pharmaceuticals in water and wastewater samples. In specific cases, definitive identification was achieved by MS/MS/MS (MS3). 3.2.3. Results and Discussion 3.2.3.1. Wastewaters from Febros WWTP Table 3.4 summarises the results obtained after analysis of Febros WWTP composite influent and effluent samples, using the developed pharmaceutical multiresidue method. Pharmaceuticals like paracetamol and hydrochlorothiazide highlighted in significantly high concentrations in the WWTP influent (> 5 µg L-1). Nonetheless, their removal along the treatment processes showed to be quite efficient, as one can confirm by the concentration values in the corresponding WWTP effluent (below the respective method quantification limit). On the contrary, the removal percentage of other pharmaceutical
Monitoring of Pharmaceuticals in Municipal Wastewaters and Surface Waters 71 compounds, such as azithromycin, bisoprolol, furosemide and ibuprofen, was rather low. Moreover, another very important aspect relates to the cases of lorazepam, ketoprofen, naproxen, bezafibrate, diclofenac and gemfibrozil, whose concentrations detected in influent samples were below those obtained in the corresponding WWTP effluents. This apparent incongruence can be justified considering: i) the difficulty to establish an accurate correspondence between the sample collected at the intake of the WWTP (influent) and that collected after the treatment, at the exit of the WWTP (effluent) – related to the WWTP typical hydraulic retention time (HRT)); ii) the conversion of conjugated metabolites (present in influent samples) to the original substances, over the treatment processes (24). Table 3.4. Target pharmaceuticals quantified in Febros WWTP influent and effluent * Concentration below the method quantification limit. With the purpose of better assessing the concentration levels of pharmaceuticals over the WWTP daily cycle, and possibly identifying potentially critical discharge periods, grab effluent samples were collected each hour, for 21-h, using the aforementioned automatic sampler device (Fig. 3.1 (c)), and further analysed. Pharmaceuticals Concentration (ng L-1) Influent Effluent Hydrochlorothiazide 6 022 <58 * Paracetamol 20 074 <22 Ciprofloxacin <6 <4 Zolpidem <3 <2 Azithromycin 571 530 Bisoprolol 239 187 Omeprazole - - Furosemide 2 675 2 430 Indapamide <65 <57 Paroxetine 105 60 Bromazepam <7 <6 Fluoxetine <5 <4 Alprazolam <5 <4 Lorazepam <7 240 Nimesulide <4 <3 Ketoprofen 122 562 Naproxen 2 583 2 748 Diazepam <3 <3 Bezafibrate 1 348 1 797 Diclofenac 431 1 107 Ibuprofen 1 992 1 527 Gemfibrozil 307 753 Simvastatin 976 197
M.A. Sousa, C. Gonçalves, V. Vilar, R. Boaventura & M.F. Alpendurada 2 potenciais efeitos sobre a saúde e dados de monitorização da sua ocorrência em diversos compartimentos ambientais. Entre a grande variedade de compostos abrangidos sob esta designação incluem-se os Compostos Farmacêuticos (CFs), cujas diversas acções terapêuticas podem originar desde o desenvolvimento de resistências bacterianas (ex. antibióticos), a interferências com a reprodução, fisiologia e crescimento de espécies aquáticas (ex. desreguladores endócrinos – esteróides, hormonas). Além do mais, trata-se de compostos que podem apresentar baixa toxicidade aguda mas elevada toxicidade crónica, com potencial bioacumulação em diferentes tecidos animais. Estes poluentes são considerados emergentes não pela emergência do seu aparecimento mas da sua detecção, dada a cada vez maior capacidade analítica, viabilizada por associações de métodos cromatográficos a processos de detecção avançados, nomeadamente a Espectrometria de Massa, possibilitando a obtenção de limites de detecção na ordem dos µg-ng L-1 [Sousa et al., 2011]. 1.2. Processos Avançados de Oxidação (PAOs) O uso crescente de xenobióticos como fármacos, fragrâncias sintéticas, pesticidas, drogas de abuso e outros contaminantes está a conduzir ao aumento das concentrações destas substâncias nas águas residuais e superficiais. Os CFs são considerados poluentes “pseudopersistentes” dado que as suas taxas de transformação/eliminação (0-100%, dependendo do composto e do processo de remoção natural/artificial) são compensadas pela sua re-introdução contínua no meio ambiente [Petrović et al., 2003]. A ocorrência e destino destes contaminantes têm sido objecto de vários estudos ao longo dos últimos anos (Fig. 1). Em Portugal começaram já a surgir registos da presença ambiental aquática de CFs. Desde efluentes hospitalares contendo, por vezes, elevadas concentrações de antibióticos (quinolonas) como ciprofloxacina, enrofloxacina, ofloxacina e norfloxacina, a efluentes domésticos tratados [Seifrtová et al., 2008 e Pena et al., 2010], bem como cursos de água superficiais, incluindo os rios Douro, Leça e Febros, tem sido reportada a presença de diversos CFs, pertencentes a vários grupos terapêuticos, entre eles analgésicos, antipiréticos e anti-inflamatórios, reguladores lipídicos, antibióticos, anti-epilépticos, etc. [Sousa et al., 2011]. Figura 1. Percurso cíclico dos compostos farmacêuticos nos diferentes compartimentos aquáticos [adaptado de Mompelat et al., 2009]. Dado que a reutilização de água tratada poderá, no futuro, ser a melhor solução para uma gestão sustentada em países mais desprovidos de água, e que uma das preocupações básicas no seu reuso é a presença de microcontaminantes (muitas vezes provenientes das descargas de ETARs municipais), são requeridos processos alternativos e/ou complementares aos tratamentos convencionais, capazes de permitir o aumento da qualidade dos efluentes eliminados e prevenir a contaminação ambiental com CFs. Neste contexto, os PAOs são considerados processos de tratamento bastante competitivos para a remoção destes compostos, mais ou menos recalcitrantes dada a sua elevada estabilidade química e/ou baixa biodegradabilidade. Os PAOs têm em comum a formação do radical hidroxilo (•OH), altamente reactivo e responsável pela degradação da componente orgânica. Dado o seu poder oxidante não-selectivo, estes radicais são capazes de oxidar e mineralizar quase todas as moléculas orgânicas, originando CO2, H2O e iões inorgânicos como produtos finais. Todavia, alguns deles podem apresentar também alguns inconvenientes, tais como altos consumos energéticos (ex. lâmpadas UV, geradores de ozono), quantidades elevadas de oxidantes e catalisadores (ex. H2O2, TiO2), ou ainda a necessidade de ajuste do pH de grandes volumes de amostra (ex. Fenton e foto-Fenton) [Miranda-Garcia et al., 2011 e Oller et al., in press]. 2. Fármacos no Rio Douro 2.1. Compostos Farmacêuticos Os fármacos englobados no presente estudo pertencem a doze classes terapêuticas diferentes, incluindo analgésicos e antipiréticos (paracetamol), anti-inflamatórios não esteróides (cetoprofeno, diclofenac, ibuprofeno, naproxeno, nimesulida), reguladores lipídicos (bezafibrato, gemfibrosil, sinvastatina), antibióticos (azitromicina, ciprofloxacina), ansiolíticos (alprazolam, bromazepam, diazepam, lorazepam, zolpidem), antidepressivos (fluoxetina, paroxetina), diuréticos (furosemida, hidroclorotiazida, indapamida), cardiotónicos (bisoprolol) e anti-ulcerosos (omeprazol), entre outros. Estes CFs encontram-se entre os mais comercializados em Portugal ao longo dos últimos anos, segundo dados fornecidos pelo INFARMED – Instituto Nacional da Farmácia e do Medicamento, e as correspondentes classes terapêuticas fazem parte do grupo das mais estudadas, reportadas e consumidas mundialmente [Sousa et al., 2011]. 2.2. Possíveis Fontes de Contaminação – ETAR Municipal e Rio Tributário O local em estudo situa-se na Região Hidrográfica 3 – Rio Douro (RH3 – Douro, Fig. 2), sendo um dos principais objectivos a avaliação do impacto, no que se refere apenas ao conteúdo em fármacos, da descarga de efluentes tratados pela ETAR de Febros (concelho de V. N. Gaia) no Rio Febros e deste, posteriormente, no Rio Douro, onde desagua. Enquanto que o Rio Douro é o segundo rio mais extenso da Península Ibérica (com 927 km no total, nasce em Espanha na província de Sória, nos picos da Sierra de Urbión, a 2.080 _______________________________________ Chapter 3 __________________________________________ 78
Efeitos dos Efluentes Domésticos Tratados e Rios Tributários no Pool de Resíduos Farmacêuticos do Rio Douro – Processos de Mitigação 3 m de altitude, apresentando uma área da bacia de 97.603 km2 e um débito médio de 450 m3 s-1, atravessa o norte de Portugal, sendo a foz no Oceano Atlântico, junto das cidades do Porto e V. N. Gaia), já o Rio Febros é um pequeno afluente da margem esquerda do Rio Douro (nasce em Seixezelo e desagua no Esteiro de Avintes, passando também pelo Parque Biológico de Gaia). Assim, é expectável que o possível impacto causado pela ETAR seja mais notório no Rio Febros do que no Rio Douro, onde se prevê um grande efeito de diluição. Todavia, a construção da ETAR mostrou-se desde logo vantajosa na despoluição do Rio Febros de resíduos de origem doméstica e industrial, permitindo inclusivamente a sua repovoação, em colaboração com o Parque Biológico, com uma espécie de mamíferos marinhos há já 20 anos desaparecida deste rio – a lontra [http://www.dn.pt/inicio/ciencia/interior.aspx? content_id=1576557&seccao=Biosfera]. Assim, e como segundo principal objectivo deste estudo, pretende-se também aumentar a taxa de remoção de CFs em efluentes aplicando uma etapa adicional com PAO. Como modelo experimental, usou-se um efluente da ETAR de Febros e uma Instalação Solar com Colectores Parabólicos Compostos (CPCs), à escala piloto, para desenvolver um tratamento de Fotocatálise Solar com TiO2. Figura 2. Localização geográfica do local/caso em estudo – Região Hidrográfica 3: Douro. 3. Metodologia Experimental 3.1. Amostragem e Pré-Tratamento das Amostras O plano de monitorização da ocorrência de CFs em amostras ambientais incluíu, por um lado, afluentes/efluentes da ETAR de Febros (amostras compostas), e por outro lado, águas superficiais dos rios Febros e Douro (amostras simples). Efectuou-se a análise de diversas amostras de água residual à entrada e saída da referida ETAR, desde Maio de 2009 a Março de 2011, bem como de amostras colhidas em 5 pontos do Rio Febros (1 a montante e 4 a jusante do ponto de descarga da ETAR) e 12 pontos do Rio Douro (7 a montante e 5 a jusante da confluência com o Rio Febros), ao longo de 2 campanhas de amostragem realizadas em 2010 (Fig. 2). O plano de amostragem incluíu a colheita de 2,5 L para águas residuais e 1 L para águas superficiais (neste caso colhida a 1 m de profundidade), transporte até ao laboratório em malas térmicas e posterior conservação a 4 ºC até realização da análise (máx. 3 dias após a colheita). No caso das amostras de afluente/efluente da ETAR de Febros, estas foram filtradas por filtros de fibra de vidro (Ø 1,2 µm) imediatamente após chegada ao laboratório, previamente à conservação a 4 ºC. 3.2. Método Analítico – SPE-LC-MS/MS Com vista à análise e quantificação dos 23 CFs seleccionados, que pudessem estar presentes nas amostras de água incluídas no programa de monitorização, havia já sido desenvolvido e optimizado pelos autores um método analítico, cf. descrito por Sousa et al. (2011). Sumariamente, o método consiste numa primeira etapa de cleanup (facultativo no caso das águas superficiais), seguido de extracção em fase sólida (SPE – Solid-Phase Extraction), para remoção de interferências e concentração dos analitos em estudo, e posterior análise/quantificação por cromatografia líquida associada a espectrometria de massa (LC-MS – Liquid Chromatography-Mass Spectrometry). Para as águas residuais, ambos os passos de cleanup e SPE foram efectuados simultaneamente com o uso de um adsorvente sólido misto (colunas de extracção Oasis MAX – Mixed-Anion Exchange), constituído por uma mistura entre uma amina quaternária (SAX - Strong-Anion Exchange) e uma fase apolar (polímero divinilbenzeno/N- vinilpirrolidona). Este adsorvente misto foi optimizado de forma a possibilitar uma elevada selectividade e sensibilidade na extracção de compostos ácidos pelos grupos de troca aniónica, enquanto que os compostos básicos e neutros são retidos na fracção polimérica. No caso das águas superficiais, o cleanup é realizado, sempre que necessário, recorrendo a colunas Bakerbond 1º, 2º - Amino (WAX – Weak-Anion Exchange), tendo o SPE sido optimizado com uso de um adsorvente hidrofílico (JTBaker H2Ophilic) [Gonçalves et al., 2011]. A par da optimização do adsorvente de extracção, foram igualmente optimizados outros parâmetros como pH (pH 2), fluxo de percolação (10 mL min-1) e volume de rotura de 50 mL para afluentes de ETAR, 100 mL efluentes de ETAR e 200 mL para águas superficiais. A separação, identificação e quantificação dos CFs foi conseguida com recurso à cromatografia líquida de alta performance, ionização por Electrospray (ESI) e detecção por espectrometria de massa (detector ion-trap), em modo de operação MS-MS, garantido assim a confirmação inequívoca de cada composto por comparação com os espectros de referência obtidos para os respectivos padrões. Uma vez optimizadas todas as variáveis experimentais, o método analítico obtido foi validado em termos de sensibilidade, selectividade, precisão, exactidão e ___________ Monitoring of Pharmaceuticals in Municipal Wastewaters and Surface Waters _______________ 79
M.A. Sousa, C. Gonçalves, V. Vilar, R. Boaventura & M.F. Alpendurada 4 calibração, tendo-se revelado robusto para a determinação de CFs em amostras de água [Sousa et al., 2011]. 3.3. Fotocatálise Solar Heterogénea (TiO2) como Tratamento Terciário de Efluentes Municipais A versatilidade dos processos avançados de oxidação devese à existência de múltiplas reacções capazes de gerar radicais hidroxilo, tais como TiO2/UV, H2O2/UV, Fenton (Fe2+/H2O2)/foto-Fenton (Fe2+/H2O2/UV)/electro-Fenton /electro-foto-Fenton e ozono (O3, O3/UV e O3/H2O2). Todas estas metodologias se tornam economicamente mais atractivas quando existe a possibilidade de utilização da radiação solar como fonte de fotões UV. Entre os diferentes PAOs, a fotocatálise heterogénea com UV/TiO2 tem-se revelado bastante promissora na destoxificação de águas residuais. A fotocatálise heterogénea é um processo que ocorre na presença de radiação UV e de um catalisador na fase sólida (fixo ou em suspensão). Na maioria dos métodos, o catalisador utilizado é um semicondutor. A estrutura electrónica destes semicondutores é caracterizada por uma banda de valência completa, de energia mais baixa, e uma banda de condução vazia, de energia mais alta. Estas duas bandas são separadas por uma quantidade de energia que se denomina gradiente energético. Ao absorver um fotão com uma quantidade de energia igual ou superior à do gradiente energético, um electrão da banda de valência ganha energia suficiente para passar ao nível da banda de condução, gerando um par electrão/lacuna (“electron/hole pair” – e- /h+). Contudo, por este ser um estado instável, em que o electrão se encontra excitado, o par electrão/lacuna tem tendência a anular-se, recombinando-se para o estado inicial numa questão de nanosegundos. No entanto, na presença de H2O e O2 na superfície do semicondutor, estes podem reagir com o par electrão/lacuna, ocorrendo reacções de oxidação/redução em vez da recombinação, dando origem à formação dos radicais hidroxilo. O interesse particular no TiO2 (BGE – Band Gap Energy = 3 eV) deve-se às propriedades fotocatalíticas, elevada hidrofilia e reactividade, baixa toxicidade, baixo custo e estabilidade química [Miranda-García et al., 2011 e Oller et al., 2011]. Neste estudo foi utilizada uma Instalação Solar Piloto com Colectores Parabólicos Compostos (CPCs) (Fig. 3) para o tratamento fotocatalítico de um efluente proveniente da ETAR de Febros, usando como catalisador o TiO2. A planta encontra-se instalada no terraço do Departamento de Engenharia Química da Faculdade de Engenharia, Universidade do Porto (FEUP). O colector solar é constituído por uma unidade CPC (0,91 m2) de 4 tubos de vidro borossilicatado (Schott-Duran type 3.3, Alemanha, com cut-off a 280 nm, diâmetro externo de 50 mm, comprimento de 1500 mm e espessura de 1,8 mm), ligados em série por junções em polipropileno, com reflectores parabólicos em alumínio anodizado, suportados por uma estrutura em alumínio e com uma inclinação de 41º (correspodente à latitude local). A planta solar possui ainda 2 tanques de recirculação (10 L e 20 L), 2 bombas de recirculação (máx. 20 L min-1) e 2 rotâmetros, 5 válvulas em polipropileno e um quadro eléctrico para controlo. A instalação pode ser usada de dois modos: utilizando a área total dos CPCs (0,91 m2) ou 0,455 m2 de cada par de tubos individualmente, permitindo a execução em simultâneo de duas experiências independentes, sob as mesmas condições de irradiação solar [Pintor et al., 2011]. A intensidade da radiação UV solar é medida por um radiómetro (ACADUS 85-PLS), instalado com o mesmo Figura 3. Plantas de Desinfecção Solar à Escala Piloto A – Planta Solar Piloto com radiómetro: (a) Vista Frente, (b) Vista Trás; 1- Radiómetro, 2- Painel Fotovoltáico, 3- Bateria; B – Planta Solar Piloto utilizada na experiência: (c) CPCs com efluente da ETAR de Febros, (d) CPCs com efluente da ETAR de Febros e 200 mg L-1 TiO2, (e) Vista Trás; 4- Colectores Parabólicos Compostos (CPCs), 5- Tanques de Recirculação, 6- Bombas de Recirculação, 7- Quadro Eléctrico, 8- Rotâmetros. _______________________________________ Chapter 3 __________________________________________ 80
Efeitos dos Efluentes Domésticos Tratados e Rios Tributários no Pool de Resíduos Farmacêuticos do Rio Douro – Processos de Mitigação 5 ângulo de inclinação. As medições da radiação ultravioleta instantânea (W m-2) são obtidas com intervalos de 1 min. e usadas no cálculo da energia UV acumulada (QUV,n , kJ L-1), durante um determinado intervalo de tempo (∆t), através da equação: 11 −− −=+= nnn t r n,G nn,UVn,UV ttt; V A UVtQQ ∆∆ onde tn é o tempo correspondente a n-alíquota (s), Vt,n o volume total de efluente em n-alíquota (L), Ar a superfície iluminada do colector e nG UV , o valor médio de radiação UV solar (W m-2) medida no intervalo de tempo ∆tn [Pintor et al., 2011]. A amostra de efluente da ETAR de Febros foi caracterizada antes do início do tratamento por fotocatálise solar. Os resultados obtidos são apresentados na Tabela 2 (secção 4.3.). A concentração óptima de TiO2 determinada foi de 200 mg L-1. As várias alíquotas colhidas ao longo da experiência foram analisadas segundo o método descrito na secção 3.2., com o objectivo de obter as cinéticas de degradação dos diversos fármacos presentes na amostra ao longo das várias horas de exposição solar. Podem ser encontrados detalhes adicionais sobre o modo de funcionamento e utilização da Instalação Solar Piloto com CPCs, bem como sobre as determinações analíticas seleccionadas/efectuadas para caracterização da amostra em Pintor et al., 2011. 4. Resultados e Discussão 4.1. Águas Residuais – Afluente e Efluente da ETAR de Febros O método analítico multi-resíduo desenvolvido foi utilizado na análise rigorosa de amostras compostas de afluente e efluente da ETAR de Febros, cujos resultados obtidos se encontram sumariados na Tabela 1. Compostos como o paracetamol e a hidroclorotiazida destacam-se em concentrações bastante significativas no afluente da ETAR (>5 µg L-1). Todavia, a eficácia dos processos de tratamento utilizados é bastante elevada para ambos os compostos, como se pode comprovar pelos valores obtidos à saída da ETAR (concentrações inferiores aos respectivos limites de quantificação). Já para outros compostos a eficiência de remoção na ETAR não é tão elevada, enquanto que em alguns casos se verifica mesmo um aumento da concentração no efluente, comparativamente ao afluente. Tal pode dever-se, sobretudo, à dificuldade de correlação entre a amostra colhida antes e após o tratamento na ETAR (relacionado com tempo de retenção hidráulica), mas também à possibilidade de não quantificação de alguns compostos à entrada da ETAR pelo facto de se encontrarem na forma conjugada (metabolitos), sendo que à saída se poderão já encontrar na forma original não-conjugada, reconvertida ao longo das etapas de tratamento [Gros et al., 2010]. Tabela 1. Concentrações dos compostos farmacêuticos-alvo detectadas no afluente e efluente da ETAR de Febros * Concentração abaixo do Limite de Quantificação do Método No sentido de conhecer a evolução dos teores de fármacos ao longo do ciclo diário e identificar momentos críticos de descarga destes poluentes, foram realizadas análises, em separado de amostras simples de efluente da ETAR, colhidas de hora em hora, ao longo de 21h consecutivas. Da análise dos gráficos (Fig. 4) podemos concluir que, para a generalidade dos compostos detectados nas diferentes amostras de efluente da ETAR, não existem variações significativas nas concentrações determinadas, daí que a amostragem composta seja considerada adequada e representativa. O tempo de retenção hidraúlica da ETAR permite uma homogenização do afluente e consequente atenuação de eventuais picos de concentração dos fármacos. Apenas o anti-inflamatório naproxeno e o regulador lipídico bezafibrato são indicados como <LOQ por análise da amostra composta, enquanto que foram obtidos valores quantificáveis para algumas amostras simples (máximos de 288 ng L-1 e 28 ng L-1, respectivamente). Todavia, além destas concentrações serem bastante baixas, tratar-se-ão de casos pontuais, sem consequências significativas no que respeita à representatividade da amostragem composta. 4.2. Águas Superficiais – rios Febros e Douro Um resultado típico obtido nas campanhas de monitorização do Rio Febros encontra-se representado na Fig. 5. Pode constatar-se que no rio Febros a jusante do ponto de descarga da ETAR (ponto F2, Fig. 2) se verifica um aumento significativo dos níveis de bisoprolol (ca. dobro) e furosemida (ca. 6 vezes superior), enquanto que a hidroclorotiazida, azitromicina, diclofenac, gemfibrosil e Concentração (ng L-1) Composto Farmacêutico Afluente Efluente Hidroclorotiazida 6.022 <58* Paracetamol 20.074 <22 Ciprofloxacina <6 <4 Zolpidem <3 <2 Azitromicina 571 530 Bisoprolol 239 187 Omeprazol - - Furosemida 2.675 2.430 Indapamida <65 <57 Paroxetina 105 60 Bromazepam <7 <6 Fluoxetina <5 <4 Alprazolam <5 <4 Lorazepam <7 240 Nimesulida <4 <3 Cetoprofeno 122 562 Naproxeno 2.583 2.748 Diazepam <3 <3 Bezafibrato 1.348 1.797 Diclofenac 431 1.107 Ibuprofeno 1.992 1.527 Gemfibrosil 307 753 Sinvastatina 976 197 ___________ Monitoring of Pharmaceuticals in Municipal Wastewaters and Surface Waters _______________ 81
M.A. Sousa, C. Gonçalves, V. Vilar, R. Boaventura & M.F. Alpendurada 6 Figura 4. Amostras simples de efluente da ETAR de Febros, colhidas de hora em hora durante um período de 21h, analisadas individualmente. Os gráficos representam a variação da concentração de cada composto durante o período de colheitas. O valor médio indicado corresponde ao valor teoricamente obtido na amostra composta. LOQ – Limit of Quantification _______________________________________ Chapter 3 __________________________________________ 82
Efeitos dos Efluentes Domésticos Tratados e Rios Tributários no Pool de Resíduos Farmacêuticos do Rio Douro – Processos de Mitigação 7 sinvastatina se tornaram quantificáveis, em concentrações entre os 40 e 360 ng L-1. Tais resultados, em complemento com os já obtidos para a análise de efluente da ETAR (Tabela 1), apontam para a contribuição da ETAR na contaminação do Rio Febros com resíduos farmacêuticos. De notar que no caso do paracetamol os níveis detectados no Rio Febros a montante e a jusante da ETAR são semelhantes, indiciando a existência de outra fonte de contaminação antropogénica. È sabido que o paracetamol se degrada durante o tratmento nas ETARs. Já relativamente ao Rio Douro, o possível impacto do Rio Febros (seu afluente, margem esquerda) no aporte de compostos farmacêuticos ao seu estuário não é visível. De facto, os resultados não revelaram a presença de qualquer um dos fármacos pesquisados acima do respectivo LOQ, provavelmente devido ao grande efeito de diluição verificado, dado o elevado caudal do Rio Douro. Concluindo, dada a reduzida eficácia de remoção de alguns fármacos por parte da ETAR em estudo, e de outras ETARs em geral, é urgente a aplicação de processos de tratamento mais eficientes na remoção de poluentes emergentes nestas Estações de Tratamento de Água. 4.3. Tratamento do Efluente da ETAR de Febros por Fotocatálise Solar numa Planta Piloto com CPCs Com o propósito de avaliar a aplicabilidade dos PAOs no tratamento terciário de águas residuais numa ETAR, usouse como amostra um efluente da ETAR de Febros (Tabela 2) e seleccionou-se a fotocatálise solar com dióxido de titânio como processo de fototratamento. Nas condições experimentais descritas (secção 3.3.), os resultados obtidos são traduzidos nos gráficos da Fig. 6. Da análise desta Figura pode concluir-se que o tratamento por fotocatálise com TiO2 foi eficaz na remoção completa da maioria dos compostos farmacêuticos quantificados no efluente original, recolhido à saída da ETAR de Febros. Exceptuam-se os seguintes compostos: ciprofloxacina, cetoprofeno, losartan1 e carvedilol1, cujas concentrações no 1NOTA: Alguns dos compostos farmacêuticos representados na Fig. 6 não estão incluídos no método descrito na secção experimental. Todavia, a sua inclusão no método foi relativamente simples, seguindo a metodologia de optimização usada para os restantes fármacos. final do tratamento, embora não inferiores aos respectivos LOQ, eram já significativamente inferiores às concentrações correspondentes iniciais. Nestes casos, o processo fotocatalítico teria de ser ainda optimizado, a nível da concentração do catalisador e/ou da quantidade de energia UV acumulada total necessária. Tabela 2. Caracterização físico-química da amostra de efluente da ETAR de Febros usada na experiência de fotocatálise solar Parâmetro (unidades) Efluente da ETAR Febros (18.03.2011) Côr n.d.* (dil. 1/20) Cheiro n.d. (dil. 1/20) pH 7,3 Temperatura (º C) 20,0 Turvação (NTU) 115 Condutividade (µS cm-1) 555 Oxigénio Dissolvido (mg L-1) 2,2 Oxidabilidade (mg L-1) 56,2 Carbono Total Dissolvido (mg L-1) 37,5 Absorvância a 254 nm (UA) 0,07 Sólidos Suspensos Totais (mg L-1) 363 Sólidos Suspensos Voláteis (mg L-1) 223 Azoto amoniacal - NH4+ (mg L-1) 1,2 Nitrato - NO3- (mg L-1) 23,3 Nitrito - NO2- (mg L-1) <0,5 Brometo - Br- (mg L-1) <0,5 Cloreto - Cl- (mg L-1) 78,5 Fluoreto - F- (mg L-1) <0,5 Fosfato - PO43- (mg L-1) 7 Sulfato - SO42- (mg L-1) 59,8 Fósforo - 31P (mg L-1) Metais Sódio - 23Na (mg L-1) Potássio - 39K (mg L-1) Cálcio - 44Ca (mg L-1) 3,0 78,8 21,1 274,8 * n.d. – não detectável Figura 5. Representação gráfica comparativa dos fármacos quantificados no Rio Febros a montante e a jusante do ponto de descarga da ETAR de Febros. ___________ Monitoring of Pharmaceuticals in Municipal Wastewaters and Surface Waters _______________ 83
M.A. Sousa, C. Gonçalves, V. Vilar, R. Boaventura & M.F. Alpendurada 8 5. Conclusões Do presente trabalho podemos concluir que os compostos farmacêuticos se encontram bastante disseminados no compartimento ambiental aquático, sendo as ETARs uma importante fonte de contaminação. No caso particular da ETAR de Febros, a sua provável implicação no input dum conjunto de fármacos no Rio Febros foi posta em evidência através da comparação dos níveis dos vários compostos em amostras colhidas a montante e a jusante do ponto de descarga da ETAR, sendo estes superiores no segundo caso. Já no que respeita à contribuição do Rio Febros para os níveis de compostos farmacêuticos no Rio Douro, tal não foi considerada significativa, provavelmente devido ao grande efeito de diluição resultante do elevado caudal deste rio. Em última análise, este trabalho demonstra ainda a aplicabilidade dos PAOs, em particular da fotocatálise solar com TiO2, como processo de tratamento terciário com potencial aplicação em ETARs (Fig. 7), surgindo como um método promissor no aumento das taxas de remoção de poluentes emergentes antropogénicos como os compostos farmacêuticos. Figura 6. Representação gráfica do decréscimo de concentração dos diferentes fármacos quantificados na amostra de efluente da ETAR de Febros, sujeita ao tratamento fotocatalítico. Figura 7. Representação esquemática da possível instalação de uma planta solar com CPCs numa ETAR, como processo de tratamento terciário, precedente à descarga do efluente. Agradecimentos O trabalho desenvolvido foi co-financiado por fundos do QREN – ON.2, O Novo Norte através do projecto NORTE- 01-0162-FEDER-000023. Os autores agradecem ao IAREN e ao LSRE pelo suporte técnico e instrumental, bem como à Fundação para a Ciência e Tecnologia pela Bolsa de Doutoramento SFRH/BD/44509/2008. _______________________________________ Chapter 3 __________________________________________ 84
Efeitos dos Efluentes Domésticos Tratados e Rios Tributários no Pool de Resíduos Farmacêuticos do Rio Douro – Processos de Mitigação 9 Referências Diário de Notícias Ciência (DNCiência): http://www.dn.pt/inicio/ciencia/interior.aspx?conte nt id=1576557&seccao=Biosfera (última actualização do site a 23-05-2010) – “Lontras estão de volta ao Rio Febros”. Gonçalves, C. et al. “Analysis of acidic, basic and neutral pharmaceuticals in river waters: clean-up by 1º,2º amino anion exchange and enrichment using an hydrophilic adsorbent”. Submetido para publicação na revista IJEAC, Setembro de 2011. Gros, M. et al. (2010). “Removal of pharmaceutical during wastewater treatment and environmental risk assessment using hazard indexes”. Environment International 36(1):15-26. Instituto da Água (INAG): http://dqa.inag.pt/(última actualização do site a 01/14/2011) – Directiva Quadro da Água (DQA, Directiva 2000/60/CE). Miranda-García, N. et al. (2011). “Photocatalytic degradation of emerging contaminants in municipal wastewater treatment plant effluents using immobilized TiO2 in a solar pilot plant”. Applied Catalysis B: Environmental 103(3-4):294-301. Mompelat, S. et al. (2009). “Occurrence and fate of pharmaceutical products and by-products, from resource to drinking water”. Environment International 35(5):803-814. Oller, I. et al. (2011). “Combination of Advanced Oxidation Processes and biological treatments for wastewater decontamination – A review”. Science of the Total Environment 409(20):4141-4166. Pena, A. et al. (2010). “Fluoroquinolone antibiotics determination in piggeries environmental waters”. Journal of Environmental Monitoring 12(3):642-646. Petrović, M. et al. (2003). “Analysis and removal of emerging contaminants in wastewater and drinking water”. Trends in Analytical Chemistry 22(10):685-696. Pintor, A.M.A. et al. (2011). “Decontamination of cork wastewaters by solar-photo-Fenton process using cork bleaching wastewater as H2O2 source”. Solar Energy 85(3):579-587. Seiftrová, M. et al. (2008). “Determination of fluoroquinolone antibiotics in hospital and municipal wastewaters in Coimbra by liquid chromatography with a monolithic column and fluorescence detection”. Analytical and Bioanalytical Chemistry 391(3):799-805. Sousa, M.A. et al. (2011). “Cleanup strategies and advantages in the determination of several therapeutic classes of pharmaceuticals in wastewater samples by SPE-LC-MS/MS”. Analytical and Bioanalytical Chemistry 399(2):807-822. ___________ Monitoring of Pharmaceuticals in Municipal Wastewaters and Surface Waters _______________ 85
Chapter 3 86 3.4. Inter-laboratory Exercises: Method Validation and Quality Control There is currently no standardised practice or protocol for the sampling and analytical determination of pharmaceuticals in water, or any other environmental media, to ensure the comparability and quality of the data generated. The participation in inter-laboratory exercises (ILEs) allows withdrawing information concerning the overall cycle of the sample, including its reception, analysis and data treatment, ultimately leading to a quality improvement of the results. ILEs can be carried out with three purposes: i) to assess the performance of an analytical method, ii) to evaluate the competence of a laboratory, or iii) to enable the certification of reference materials (25). Nowadays, European level ILEs, regarding the analysis of frequently used pharmaceuticals in environmental waters, are still scarce. Nevertheless, some joint efforts have already been reported (26, 27). During this Ph.D. work, we participated in two European inter-laboratory studies, the first promoted by the Joint Research Centre (JRC), a Directorate-General of the European Commission for research, innovation and science (herein addressed as ILE-FATE_SEES), and a second one organized under the PHARMAS EU project, by the Advanced School of Public Health (EHESP) (ILE-PHARMAS). ILE-FATE_SEES presented as main goal the evaluation of the reproducibility of analytical methods employed in the determination of a list of emerging pollutants (pharmaceuticals included) in WWTP effluents. After 1 L sampling of municipal effluents from two WWTP located in the northern region of Portugal (WWTP from Viana do Castelo (VC) and WWTP from Parada, Maia (PM)), a certain volume of each sample was shipped to the organizing laboratory in Italy, while the remaining was analysed in our facilities, using the aforedescribed method conceived for the determination of pharmaceutical compounds in wastewaters. Only 12 out of the 23 pharmaceuticals included in our method were present in the list proposed by ILE-FATE_SEES organization: diclofenac, ketoprofen, naproxen, ibuprofen, bezafibrate, gemfibrozil, ciprofloxacin, bisoprolol, fluoxetine, paroxetine, alprazolam and zolpidem. Afterwards, the attained results by both laboratories were
Monitoring of Pharmaceuticals in Municipal Wastewaters and Surface Waters 87 statistically evaluated in terms of z-scores. Z values were calculated according to the following expression: σ Χ− = lab X Z (1) where: Xlab = result for a laboratory; X = mean values between all laboratories; σ = standard deviation in the corresponding population. The most commonly accepted classification establishes that: │z-score│< 2 – satisfactory result; 2 < │z-score│< 3 – questionable result; │z-score│> 3 – unsatisfactory result (28). In a total of 24 analytical determinations (12 pharmaceuticals for WWTP-VC and for WWTP-PM), our results included: 11 determinations with │z-score│< 2: satisfactory results; 4 determinations with 2 < │z-score│< 3: dubious results; 3 determinations with │z-score│> 3: unsatisfactory results; 6 results with pharmaceutical below the respective LOQ. Even though, the attained results relied on the comparison between two laboratories only. Consequently, more accurate conclusions on the performance of the employed methodologies could only be withdrawn after enlarging the comprehensiveness of the study. In the ILE-PHARMAS exercise, the goal was to evaluate the reproducibility of analytical methods for the measurement of antibiotics in waters (Table 3.5). It was the first initiative organised in Europe in this type of exercises, using different kind of analytical methods and devices. Fourteen laboratories from five countries (Canada, France, Italy, the Netherland and Portugal) participated and a total number of 78 samples were distributed. During the exercise, 2 testing samples (3 bottles of each) prepared from tap water (series A) and river water (series B) were spiked with antibiotics and sent to the participants to be analysed along one month.
Chapter 4 94 However, total mineralization is far from being achieved. Thus, there is an absolute need to develop effective methods for handling these contaminants and potentially toxic compounds, widely dispersed in the environment. Conventional drinking water decontamination processes involve the combination of flocculation, filtration, sterilization and conservation, with the addition of a limited number of chemicals. On the contrary, normal human sewage water is usually treated in conventional biological processing plants. Despite the fact that biological treatment techniques are well established and relatively cheap, they are often susceptible to toxic compounds which inactivate the degrading microorganisms, thus becoming unable to remove the contaminants. In order to solve this issue, besides reducing emissions, a chemical treatment of contaminated drinking, surface and groundwaters should be pursued, as well as a chemical treatment of wastewaters containing biocides or nonbiodegradable compounds (3). Addressing this problem has recently recalled a tremendous amount of research in the field of robust new water treatment technologies, leading to the complete mineralization of organic pollutants at lower cost and with less energy, while at the same time minimizing the use of chemicals and their subsequent impact on the environment (4). Special attention has been given to advanced oxidation treatment processes. 4.1. Advanced Oxidation Processes 4.1.1. Definition Advanced oxidation processes (AOPs) are defined as oxidation methods occurring in the aqueous phase and based upon the generation of highly reactive species such as hydroxyl radicals (●OH). These radicals present a high oxidation potential (E0 = 2.80 V relatively to the normal hydrogen electrode) and are responsible for the oxidation and mineralization of almost any organic molecule, yielding CO2, H2O and inorganic ions as final products, due to their strong, unselective oxidative power. Thus, AOPs can be used to chemically decompose pollutants into harmless end-products, otherwise not removed by conventional treatment processes (3, 5). Moreover, these oxidation processes have the increased interest of taking place at room temperature.
Photo-Remediation of Contaminated Waters using Advanced Oxidation Processes (AOPs) 95 Nevertheless, there are species such as CO32-, HCO3-, SO42-, Cl- and humic acids, which can act as free radicals’ inhibitors and significantly reduce the degradation process efficiency, depending on their concentration in the medium. For that reason, some pre-treatment may be required to reduce these species concentration or even eliminate them. In case the treatment process is occurring under acidic conditions, the problem concerning the presence of carbonates or bicarbonates is automatically solved (6). Fig. 4.1. Reactions involved in the processes of direct photolysis, indirect photolysis, heterogeneous photocatalysis and formation of oxygen radical reactive species (adapted from Burrows et al. (2)).
Chapter 4 96 4.1.2. Classification and Applications to Water Treatment According to the number of phases of the reaction system, AOPs can be divided in homogeneous (single-phase system) or heterogeneous (more than one phase system) processes. Moreover, they can also be classified in compliance with the method used to generate the reactive radical species as chemical, photochemical, electrochemical, sonochemical or thermochemical (Fig. 4.2). HOMOGENEOUS CHEMICAL Ref. Ozone in basic medium (O3/OH–) (7) Ozone and hydrogen peroxide (O3/H2O2) and (O3/H2O2/OH–) (8) Homogeneous catalytic ozonation (O3/catalyst) (9) Fenton (H2O2/Fe2+) (10) PHOTOCHEMICAL Ref. Ozone and ultraviolet radiation (O3/UV) (11) Hydrogen peroxide and ultraviolet radiation (H2O2/UV) (12, 13) Ozone, hydrogen peroxide and ultraviolet radiation (O3/H2O2/UV) (14) Photo-Fenton (H2O2/ Fe2+/UV) (15) ELECTROCHEMICAL Ref. Electrochemical oxidation (16) Electro-Fenton and photo-electro-Fenton (17, 18) SONOCHEMICAL Ref. Ultrasounds (US) (19) Ozone and ultrasounds (O3/US) (20) Hydrogen peroxide and ultrasounds (H2O2/US) (21) THERMOCHEMICAL Ref. Wet air oxidation (WAO) (22) Supercritical wet air oxidation (SWAO) (23) Catalytic wet air oxidation (CWAO) (24, 25) HETEROGENEOUS CHEMICAL Ref. Heterogeneous catalytic ozonation (O3/catalyst) (26, 27) PHOTOCHEMICAL Ref. Heterogeneous photocatalysis (TiO2/UV), (TiO2/UV/H2O2) and (O3/TiO2/UV) (12, 28, 29) Fig. 4.2. AOPs classification, possible combinations and some relevant references.
Photo-Remediation of Contaminated Waters using Advanced Oxidation Processes (AOPs) 97 Many AOPs have recently earned great interest in the field of water decontamination processes, each one presenting its own advantages and disadvantages. Polluted waters vary from industrial effluents with different origins (e.g., agriculture, textile, tannery, etc.), to hazardous effluents from hospitals or pharmaceuticals production facilities, or even to municipal WWTP effluents contaminated with pathogenic agents and recalcitrant pollutants, such as pharmaceuticals and endocrine disrupting compounds (7, 30, 31). In particular, over the last few decades there has been a paramount increase of studies concerning the use of photochemical AOPs for the effective and sustainable treatment of waters contaminated with pharmaceutical compounds, as one can observe in Fig. 4.3. Indeed, owing to some technological/engineering developments, this growth may also have been propelled by the possibility of using solar radiation as the photon source. 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 0100 200 300 400 500 No. Publications Year Fig. 4.3. Number of publications relative to photodegradation treatment processes of waters contaminated with pharmaceutical compounds, between 2000 and 2012 (source: www.sciencedirect.com, using “photodegradation” and “pharmaceuticals” and “water treatment” as keywords). Amongst all photochemical oxidation processes that could be chosen, in the herein described Ph.D. work we opted for the heterogeneous photocatalysis, using titanium dioxide (TiO2) as catalyst (Paper 3 and Paper 4).
Chapter 4 98 4.1.2.1. Heterogeneous Photocatalysis 4.1.2.1.1. Fundamentals In a photochemical process, light is always a reagent, never a catalyst. The term photocatalysis implies the combination of photochemical process with the use of a catalyst, i.e., it is the “acceleration of a photoreaction by the presence of a catalyst” (3). In the case of heterogeneous processes, the catalyst employed is a semiconductor material like metal oxides. When a semiconductor (SC) is exposed to radiation with higher energy than the band gap between its valence band (VB – filled electronic band with the highest energy) and its conduction band (CB - empty electronic band with the lowest energy), there is an electron transfer from VB to CB, resulting in the formation of a hole in the valence band (h+): SC + hν → SC (h+ VB + e- CB) (1) If the semiconductor is in contact with a liquid electrolyte solution containing redox couple, the formed electron/hole (e- and h+) pairs migrate to the surface of the semiconductor particle and charge transfer occurs across the interface, in order to balance the potentials of the two phases (3, 32). Hydroxyl radicals can be formed by two ways: the valence band hole (h+ VB) can either react with the adsorbed water or the surface OH- groups on the semiconductor particle: SC (h+ VB) + H2Oads → SC + HO● ads + H+ (2) SC (h+ VB) + OH- ads → SC + HO● ads (3) The semiconductor is recovered when acceptor molecules (A) such as O2 are adsorbed and react with an electron in the conduction band, while donor molecules (D) such as H2O are adsorbed and react with a hole in the valence band: SC (e- CB) + Aads → SC + A- ads (4) SC (h+ VB) + Dads → SC + D+ ads (5)
Photo-Remediation of Contaminated Waters using Advanced Oxidation Processes (AOPs) 99 It is widely accepted that O2 plays an important role in these reactions. O2 can trap conduction band electrons to originate superoxide radicals (O2● -) (Eq. 6). These radicals can react as well with hydrogen ions (H+) from water splitting process to form HO2● -. SC (e- CB) + O2 ads + H+ → SC + HO2● - → O2● - + H+ (6) H2O2 can also be formed from HO2● – species: SC (e- CB) + HO2● - + H+ → H2O2 (7) and hydroxyl radicals may be originated by cleavage of H2O2 via one of the following reactions: H2O2 + hν → 2 HO● (8) H2O2 + O2● - → HO● + O2 + HO- (9) H2O2 + SC (e- CB) → HO● + HO- + SC (10) Finally, highly reactive hydroxyl radicals will act as oxidants in the degradation of organic compounds (33, 34). A resume of the photocatalytic process detailed by Eqs. (1)-(10) is illustrated in Fig. 4.4. Fig. 4.4. Schematic diagram illustrating the principle of heterogeneous photocatalysis (adapted from Ahmed et al. (33)).
Chapter 4 100 4.1.2.1.2. Catalysts By definition, a catalyst remains unaltered and can be recovered unchanged after many turnovers of the reaction process. In the case of photocatalytic treatment processes, metal oxides and sulphides constitute a large class of suitable semiconductor materials (Table 4.1.). Table 4.1. Properties of some semiconductor materials (from Galvez et al. (3)) Material Band gap (eV) Wavelength corresponding to band gap (nm) BaTiO 3 3.3 375 CdO 2.1 590 CdS 2.5 497 CdSe 1.7 730 Fe2O3 2.2 565 GaAs 1.4 887 GaP 2.3 540 SnO2 3.9 318 SrTiO3 3.4 365 TiO2 3.0 390 WO3 2.8 443 ZnO 3.2 390 ZnS 3.7 336 Titanium dioxide (TiO2; white powder; BGE = 3.0 eV) is one of the most commonly used catalysts. As aforementioned, it was also the catalyst chosen for our experimental work (Paper 3 and Paper 4) due to its high reactivity, low toxicity, low price, chemical stability and, above all, its capability of using natural and renewable solar UV energy (λ ≤ 390 nm). TiO2 has an adequate energetic separation between the valence and the conduction bands, which can be surpassed by the energy content of a solar photon, thus reducing the cost associated with UV radiation production (3, 35). More specifically, we opted for TiO2 P25-Degussa (80% anatase and 20% rutile), which is as well the most frequently employed on photocatalytic tests. Anatase and rutile are two distinct crystalline forms of TiO2, presenting a density of 3.9 and 4.2 g cm-3, respectively.
Photo-Remediation of Contaminated Waters using Advanced Oxidation Processes (AOPs) 101 Titanium dioxide can be used in suspension or, alternatively, immobilized in different supports, such as glass spheres, Teflon or glass fibre. Though in our work TiO2 was employed in suspension, the immobilized form has proved to be quite efficient, mainly at the industrial scale, where suspended TiO2 has revealed some problems concerning its separation and regeneration. Due to the small particle dimensions (usually < 0.5 µm), the filtration process becomes rather expensive, thus hindering its application (36). Studies on the use of this catalyst have also showed that its efficiency can be improved by either increasing the active surface area, redimensioning it to the nanoscopic scale or through the addition of some co-adsorbent materials (e.g., silica, alumina, zeolites or activated carbon) (37-39). 4.1.2.1.3. Variables affecting Photocatalysis The kinetic behaviour of the photocatalytic degradation reactions of organic pollutants in general, and pharmaceutical compounds in particular, depends on several experimental parameters, which shall be optimized in order to maximize the process efficiency. Such parameters include: i) contaminant initial concentration; ii) concentration of the catalyst; iii) radiation source; iv) pH of the reaction medium; v) temperature; vi) dissolved organic matter (DOM); vii) dissolved O2; viii) presence of sensitizers and radical scavengers (3, 40). i) Influence of Contaminant Initial Concentration Most authors agree that, for the majority of the contaminants, the influence of the initial concentration of substrate on the photocatalytic degradation rate follows a pseudo-first order kinetic, according to the Langmuir-Hinshelwood (L-H) law. The photocatalytic reaction rate (r) is proportional to the catalyst covered surface fraction (θ), according to Eq. (11): + == KC KC kkr 1 θ (11)
Chapter 4 102 where k represents the pseudo-first order kinetic constant, C the initial concentration of the substrate and K the constant of the adsorption balance of substrate at the surface of the catalyst (41). As oxidation proceeds, less and less surface of TiO2 particles is covered, as the contaminant is degraded. At total decomposition, the degradation rate equals zero and a decreased photocatalytic rate is to be expected with the increase of accumulated UV energy. ii) Influence of Catalyst Concentration The concentration of TiO2 which maximizes the reaction rate is that corresponding to a maximal particle surface exposed to radiation. When the catalyst concentration is too high, turbidity also rises and hampers the further penetration of light in the reactor. On the other hand, by increasing the number of suspended particles, the probability they agglomerate is higher, leading to a decrease of surface area available for the reaction (3, 41, 42). In every practical application, the optimum catalyst load must be determined to prevent its excess and ensure the maximal absorption of efficient photons. Therefore, in our experimental work (Paper 3), some preliminary tests were conducted comprising a wide range of TiO2 concentrations (from absence, corresponding to photolysis, to 400 mg L-1), according to the set of different experimental systems employed. iii) Influence of Radiation Source The radiation source can affect photocatalysis kinetic according to three factors: the radiation photon flux, the radiation wavelength and the quantum yield. Regarding the photon flux, an increase of radiation intensity generally results in the improvement of the photocatalytic reaction rate, i.e., in the proportional increase of the substrate degradation rate. However, above a certain threshold, the reaction rate becomes independent of the radiation flux, which can be attributed to the recombination of the electron/hole pairs at high radiation intensities (3, 32).
Photo-Remediation of Contaminated Waters using Advanced Oxidation Processes (AOPs) 103 Concerning the radiation wavelength, and according to what was previously explained, the adsorption spectrum of the catalyst will condition the useful wavelength range for a certain photocatalytic reaction. In our case-study, TiO2 just absorbs radiation with λ ≤ 390 nm (please refer to Table 4.1), thus being photo-activated only in the UV spectral region. As regard to the quantum yield, its influence on the photocatalytic process will be later assessed on Section 4.2.2.1.4. In our initial work (Paper 3), two different radiation sources were comparatively tested and optimized: an UV medium pressure mercury lamp (Heraeus TQ 150W) vs. natural solar radiation. iv) Influence of pH It is well known that TiO2 surface is amphoteric and, consequently, its charge is pH- dependent. For TiO2-P25, used in our work, the zero charge point (pHzcp) is located in the range 5.6 < pH < 6.4 (pH ≈ 6.3) (42). For that reason, at lower pH values TiO2 surface is positively charged, as represented in the following equation: TiOH + H+ ↔ TiOH2 + (12) On the contrary, if the medium pH is above TiO2 pHzcp, then the catalyst surface will be negatively charged: TiOH + HO ↔ TiO‒ + H2O (13) In conclusion, pH will affect both the ionization state of the substrate and TiO2 surface, thus promoting or hampering the adsorption of organic compounds to the catalyst surface in case they present different or equal charges, respectively. Moreover, it must be highlighted that, under acidic conditions, TiO2 particles are prone to agglomerate, which will lead to a reduction of active surface area available for the absorption of radiation and adsorption of contaminants.
Chapter 4 110 However, it should be noticed that the lamp photon flux (Einstein s-1) corresponds to its total emission spectrum, whereas only radiation with λ ≤ 390 nm is capable of photoactivating TiO2. Thus, bearing in mind the main emission line intensities up until 390 nm, partial photon fluxes were calculated for each wavelength and the corresponding radiant energy (E≤390 nm, J) was determined according to Eq. (18): λ c hNE A = (18) where E is the energy of electromagnetic radiation, NA is the Avogadro constant (6.022E23 mol-1), h is the Planck constant (6.626E-34 J s), c is the speed of light in the vacuum (3E8 m s-1) and λ is the respective wavelength. The radiant flux (also known as radiant power) obtained was 15 J s-1, i.e., 15 W. Afterwards, calculations regarding the accumulated UV energy (QUV,n) took into account the lab-scale photochemical reactor geometry/dimensions and the volume of solution/suspension (surface contact area of ~233 cm2 and 850 mL of sample, as aforementioned), as well as the time duration of each experiment, according to the following equation: 1 , 1,, ; −− −=∆∆+= nnn t r nG nnUVnUV ttt V A UVtQQ (19) where tn is the time corresponding to n-water sample(s), Vt is the total reactor volume, Ar is the illuminated surface area and nG UV , is the average UV radiation measured in the time interval Δtn. Results showed that the UV mercury lamp provided a constant QUV(total) dose of approximately 3.5 kJ L-1 min-1 (QUV(300–400 nm) ≈ 1 kJ L-1 min-1), except in the first ca. 2 min of lamp warming up (Fig. 4.10). Similar calculations were performed for the LYNX S11W blacklight blue lamp. However, instead of the actinometrical determination of the photon flux, the average UV radiation intensity was measured with an UV portable radiometer (Kipp & Zonen, model CUV5), which provided the result of 10 WUV m-2.
Photo-Remediation of Contaminated Waters using Advanced Oxidation Processes (AOPs) 111 Fig. 4.10. Lab-scale prototype LsTQUV: graphic representation of the accumulated UV energy per litre of solution, as a function of the time after turning the UV lamp on [obs.: insert picture represents the emission spectrum of the medium pressure mercury lamp Heraeus TQ 150W]. Ultimately, some considerations shall be made regarding the CPC solar pilot plant. Besides the composition of the material used for its construction (the abovementioned borosilicate glass, with a cut-off at 280 nm), the design of solar photoreactors conceived for photochemical applications is of no less importance. The original photochemical reactors were based on line-focus parabolic-trough concentrators (PTCs) (Fig. 4.11 (b)). PTCs made an efficient use of direct solar radiation and, as an additional advantage, the thermal energy collected from the concentrated radiation could simultaneously be used with other purposes. Nevertheless, they presented three main disadvantages: i) the use of only direct radiation; ii) the high cost; iii) the low optical and quantum efficiencies for TiO2 applications (49). On the other hand, one-sun (non-concentrating) collectors (Fig. 4.11 (a)) were cheaper than PTCs, since they were
Chapter 4 112 static and had no solar tracking devices. Moreover, they did not concentrate radiation, thus the efficiency of the process was not reduced by factors associated with concentration and solar tracking (50). Finally, compound parabolic concentrators (CPCs) (Fig. 4.11 (c)) appeared as an interesting cross between trough concentrators and one-sun systems. They are considered by many the choice for optimal possibilities (51). CPCs are static collectors with an involute reflective surface around a cylindrical reactor tube. The reflector design enables almost all UV radiation arriving at the CPC aperture to be collected and available for the photochemical process taking place inside the reactor. These static collectors provide the best optics for low concentration systems and present the advantages of both PTCs and one-sun collectors, since they can capture both direct and diffuse UV sunlight (Fig. 4.12) (52, 53). Fig. 4.11. Design of photocatalytic reactors used for water treatment: (a) one-sun system; (b) parabolic-trough concentrator (PTC); (c) compound parabolic collector (CPC). Fig. 4.12. Ray tracing example for a CPC collector (from Ajona et al. (51)).
Photo-Remediation of Contaminated Waters using Advanced Oxidation Processes (AOPs) 113 The intensity of solar UV radiation is measured by radiometers. In our work, a global UV radiometer (ACADUS 85-PLS) was mounted on the pilot plant, tilted at the same angle (41º), providing data in terms of incident WUV m-2. 4.1.2.1.5. Analytical Determinations In order to accompany the degradation of target and non-target emerging compounds over the phototreatment experiments (Paper 3, Paper 4 and Paper 5), several analytical determinations were carried out, two of which deserve special attention: LC- tandem MS and the determination of dissolved organic carbon (DOC). While the utilized LC-tandem MS methodology preceded by a cleanup/SPE step, in order to purify and concentrate the samples on target analytes, was part of our initial work (described in Chapter 2), the identification of transformation products (TPs) and the elucidation of the mechanisms involved in the attenuation of a particular xenobiotic – lorazepam, required a completely different approach. A lot of effort was devoted to this issue and the adopted strategy and achieved results are presented in the following chapter (Chapter 5, including Paper 5). The mineralization state of a sample can be determined by its DOC content. Thus, DOC was also followed throughout photodegradation experiments, showing, as expected, a considerable decrease along the reactions. Notwithstanding, one main conclusion which can be withdrawn from this work relates to the fact that both natural and artificial remediation processes are potential sources of TPs for the environment, should total mineralization be not attained. Herein follow the two papers (Paper 3 and Paper 4) encompassing the experimental work and achieved results in the field of AOPs applications for contaminated water photoremediation. Another publication, involving the use of combined biological treatment and AOPs for the decontamination of pesticide-containing wastewaters, resulted from a collaboration work with the Faculty of Engineering of the University of Porto, and is included in the present thesis as Annex IV.
Chapter 4 114
Photo-Remediation of Contaminated Waters using Advanced Oxidation Processes (AOPs) 115 4.2. Paper 3 – Photolytic and TiO2-Assisted Photocatalytic Oxidation of the Anxiolytic Drug Lorazepam (Lorenin® pills) under Artificial UV Light and Natural Sunlight: A Comparative and Comprehensive Study Published Scientific Paper: Sousa, M.A., Gonçalves, C., Pereira, J.H.O.S., Vilar, V.J.P., Boaventura, R.A.R. and Alpendurada, M.F., Photolytic and TiO2-Assisted Photocatalytic Oxidation of the Anxiolytic Drug Lorazepam (Lorenin® pills) under Artificial UV Light and Natural Sunlight: A Comparative and Comprehensive Study. Solar Energy, 2013. 87: 219-228.
Chapter 4 116
________ Photo-Remediation of Contaminated Waters using Advanced Oxidation Processes (AOPs) _________ 117
_________________________________________ Chapter 4 __________________________________________ 118
________ Photo-Remediation of Contaminated Waters using Advanced Oxidation Processes (AOPs) _________ 119
Annexes 222 Pesticides 86-74-8 Carbazole 10605-21-7 Carbendazim 5234-68-4 Carboxin 1698-60-8 Chloridazon 1897-45-6 Chlorothalonil 15545-48-9 Chlorotoluron 1982-47-4 Chloroxuron 101-21-3 Chlorpropham 5598-13-0 Chlorpyriphos methyl 1861-32-1 Chlorthal-dimethyl 1702-17-6 Clopyralid 21725-46-2 Cyanazine 36576-43-9 Cyanazine acid 52315-07-8 Cypermethrin 90-98-2 DBP 62-73-7 d-Dichlorvos 52918-63-5 Deltamethrin 6190-65-4 Desethylatrazine 30125-63-4 Desethylterbutylazine 1007-28-9 Desisopropylatrazine 13684-56-5 Desmedipham 1014-69-3 Desmetryn 333-41-5 Diazinon 1918-00-9 Dicamba 1194-65-6 Diclobenil 115-32-2 Dicofol 83164-33-4 Diflufenican 108-18-9 Diisopropylamine 87674-68-8 Dimethenamid 60-51-5 Dimethoat 1420-07-1 Dinoterb 563-12-2 Echio 1031-07-8 Endosulfan-sulfate 26225-79-6 Ethofumesate 13194-48-4 Ethoprophos 60168-88-9 Fenarimol 55-38-9 Fenthion 142459-58-3 Flufenacet 69377-81-7 Fluroxypyr 85509-19-9 Flusilazole
Annexes 223 Pesticides 76674-21-0 Flutriafol 65907-30-4 Furathiocarb 1071-83-6 Glyphosate 76-44-8 Heptachlor 1024-57-3 Heptachlor epoxide 23560-59-0 Heptenophos 51235-04-2 Hexazinone 138261-41-3 Imidacloprid 18181-70-9 Iodofenphos 330-55-2 Linuron 121-75-5 Malathion 94-74-6 MCPA 94-81-5 MCPB 93-65-2 MCPP (Mecoprop) 16484-77-8 Mecoprop-P 57837-19-1 Metalaxyl 41394-05-2 Metamitron 67129-08-2 Metazachlor 2032-65-7 Methiocarb 01-10-2635 Methiocarb sulfoxide 16752-77-5 Methomyl 72-43-5 Methoxychlor 51218-45-2 Metolachlor 19937-59-8 Metoxuron 7786-34-7 Mevinphos 96180-79-9 Microcystin-LA / Cyanoginosin-LA 2212-67-1 Molinate 825629-31-0 N-Ethyl-2-tolylsulfonamide 53-19-0 o,p'-Dichlorodiphenyldichloroethane (o,p'-DDD) 1113-02-6 Omethoate 34622-58-7 Orbencarb 19666-30-9 Oxadiazon 77732-09-3 Oxadixyl 76738-62-0 Paclobutrazol 56-38-2 Parathion ethyl 298-00-0 Parathion methyl 40487-42-1 Pendimethalin 52645-53-1 Permethrin 13684-63-4 Phenmedipham
Annexes 224 Pesticides 14816-18-3 Phoxim 29232-93-7 Pirimiphos methyl 67747-09-5 Prochloraz 1610-18-0 Prometon 7287-19-6 Prometryn 1918-16-7 Propachlor 709-98-8 Propanil 139-40-2 Propazine 60207-90-1 Propiconazole 23950-58-5 Propyzamide 90717-03-6 Quinmerac 124495-18-7 Quinoxyfen 26259-45-0 Secbumeton 35507-37-0 Sulfonyl urea 886-50-0 Terbutryn 5915-41-3 Terbutylazine 148-79-8 Thiabendazole 59669-26-0 Thiodicarb 57018-04-9 Tolclofos methyl 731-27-1 Tolylfluanid 43121-43-3 Triadimefon 2303-17-5 Tri-allate 52-68-6 Trichlorfon 101-20-2 Triclocarban Pharmaceuticals 123050-98-6 (R)-O-Desmethyl Naproxen 57-91-0 17-alpha-Estradiol 57-63-6 17-alpha-Ethinylestradiol 50-28-2 17-beta-Estradiol 37517-30-9 Acebutolol 77-66-7 Acecarbromal 89796-99-6 Aceclofenac 53164-05-9 Acemetacin 103-90-2 Acetaminophen (paracetamol) 59-66-5 Acetazolamide 50-78-2 Acetylsalicylic acid (aspirin) 59277-89-3 Acyclovir 18559-94-9 Albuterol 51022-70-9 Albuterol sulfate 22131-79-9 Alclofenac 52-43-7 Allobarbital
Annexes 225 Pharmaceuticals 28981-97-7 Alprazolam 50-48-6 Amitryptiline 57-43-2 Amobarbital 26787-78-0 Amoxicillin 69-53-4 Ampicillin 635-12-1 Anthracen-1,4-dione 37321-09-8 Apramycin 77-02-1 Aprobarbital 29122-68-7 Atenolol 83905-01-5 Azithromycin 1134-47-0 Baclofen 102280-35-3 Baquiloprim 378-44-9 Betamethasone 83-46-5 Beta-sitosterol 63659-18-7 Betaxolol 41859-67-0 Bezafibrate 66722-44-9 Bisoprolol 1812-30-2 Bromazepam 77-26-9 Butalbital 58-08-2 Caffeine 57775-29-8 Carazolol 298-46-4 Carbamazepine 10206-21-0 Cefacetrile 15686-71-2 Cefalexin 5575-21-3 Cefalonium 21593-23-7 Cefapirin 25953-19-9 Cefazolin 62893-19-0 Cefoperazone 302-17-0 Chloral hydrate 56-75-7 Chloramphenicol 57-15-8 Chlorobutanol 57-62-5 Chlortetracycline 57-88-5 Cholesterol 85721-33-1 Ciprofloxacin 59729-32-7 Citalopram 81103-11-9 Clarithromycin 37148-27-9 Clenbuterol 882-09-7 Clofibric acid 23593-75-1 Clotrimazole 7081-44-9 Cloxacillin
Annexes 226 Pharmaceuticals 76-57-3 Codeine 8064-90-2 Cotrimoxazole 483-63-6 Crotamiton 50-18-0 Cyclophosphamide 75526-90-8 Cyclophosphamide (anhydrous form) 112398-08-0 Danofloxacin 7261-97-4 Dantrolene 80-08-0 Dapsone 20830-81-3 Daunorubicin 50-02-2 Dexamethasone 117-96-4 Diatrizoate 439-14-5 Diazepam 15307-86-5 Diclofenac 3116-76-5 Dicloxacillin 56-53-1 Diethylstilbestrol 98106-17-3 Difloxacin 88637-37-0 Diphenhydramine 57808-66-9 Domperidone 1668-19-5 Doxepine 25316-40-9 Doxorubicin 94088-85-4 Doxycycline (anhydrous) 564-25-0 Doxycycline (monohydrate) 74011-58-8 Enoxacin 93106-60-6 Enrofloxacin 56420-45-2 Epirubicin 114-07-8 Erythromycin 128196-01-0 Escitalopram 119141-88-7 Esomeprazole 50-27-1 Estriol 53-16-7 Estrone 481-97-0 Estrone sulphate 77-67-8 Ethosuximide 56775-91-8 Etofibrate 76824-35-6 Famotidine 458-24-2 Fenfluramine 49562-28-9 Fenofibrate 26129-32-8 Fenofibric acid 31879-05-7 Fenoprofen 53746-45-5 Fenoprofen calcium salt dihydrate 03-12-1944 Fenoterol
Annexes 227 Pharmaceuticals 5250-39-5 Flucloxacillin 42835-25-6 Flumequine 51-21-8 Fluorouracil 54910-89-3 Fluoxetine 54739-18-3 Fluvoxamine 54-31-9 Furosemide 25812-30-0 Gemfibrozil 1403-66-3 Gentamicin 10238-21-8 Glyburide (glibenclamid; glybenzcyclamide) 56-29-1 Hexobarbital 58-93-5 Hydrochlorothiazide 125-29-1 Hydrocodone 15687-27-1 Ibuprofen 53949-53-4 Ibuprofen 1-hydroxy 51146-55-5 Ibuprofen 2-hydroxy 3778-73-2 Ifosfamide 50-49-7 Imapramine 256-96-2 Iminostilbene 53-86-1 Indomethacin 66108-95-0 Iohexol 78649-41-9 Iomeprol 60166-93-0 Iopamidol 73334-07-3 Iopromide 70288-86-7 Ivermectin 16846-24-5 Josamycin 08-07-8063 Kanamycin sulfate 22071-15-4 Ketoprofen 84057-84-1 Lamotrigine 103577-45-3 Lansoprazole 102767-28-2 Levetiracetam 137-58-6 Lidocaine 859-18-7 Lincomycin 554-13-2 Lithium carbonate 79794-75-5 Loratadine 846-49-1 Lorazepam 75330-75-5 Lovastatin 115550-35-1 Marbofloxacin 07-06-3625 Mebeverine 644-62-2 Meclofenamic acid
Annexes 228 Pharmaceuticals 06-12-2898 Medazepam 61-68-7 Mefenamic acid 57-53-4 Meprobamate 72-33-3 Mestranol 657-24-9 Metformin 61-32-5 Methicillin 115-38-8 Methylphenobarbital 37350-58-6 Metoprolol 73573-88-3 Mevastatin 13614-98-7 Minocycline 42200-33-9 Nadolol 985-16-0 Nafcillin 434-22-0 Nandrolone 22204-53-1 Naproxen 1404-04-2 Neomycin B 7298-73-9 N-Methylphenacetine 1088-11-5 Nordiazepam 70458-96-7 Norfloxacin 1476-53-5 Novobiocin 82419-36-1 Ofloxacin 3922-90-5 Oleandomycin 73590-58-6 Omeprazole 66-79-5 Oxacillin 14698-29-4 Oxalinic acid 604-75-1 Oxazepam 6452-71-7 Oxprenolol 79-57-2 Oxytetracycline 61869-08-7 Paroxetine 61-33-6 Penicillin G 87-08-1 Penicillin V 76-74-4 Pentobarbital 06-05-6493 Pentoxifylline 60-80-0 Phenazone 50-06-6 Phenobarbital 50-33-9 Phenylbutazone 57-41-0 Phenytoine 13523-86-9 Pindolol 1893-33-0 Pipamperon 81093-37-0 Pravastatin 50-24-8 Prednisolone
Annexes 229 Pharmaceuticals 125-33-7 Primidone 525-66-6 Propranolol 479-92-5 Propyphenazone 66357-35-5 Ranitidine 80214-83-1 Roxithromycin 35763-26-9 Salbutamol 98105-99-8 Sarafloxacin 76-73-3 Secobarbital 309-43-3 Secobarbital sodium 79617-96-2 Sertraline 79902-63-9 Simvastatin 3930-20-9 Sotalol 1695-77-8 Spectinomycin 8025-81-8 Spiramycin 57-92-1 Streptomycin 68-35-9 Sulfadiazine 122-11-2 Sulfadimethoxine 2447-57-6 Sulfadoxine 127-79-7 Sulfamerazine 57-68-1 Sulfamethazine 723-46-6 Sulfamethoxazole 144-83-2 Sulfapyridine 25451-15-4 Taloxa 846-50-4 Temazepam 23031-25-6 Terbutaline 60-54-8 Tetracycline 55297-95-5 Tiamulin 108050-54-0 Tilmicosin 91524-16-2 Timolol 13710-19-5 Tolfenamic acid 27203-92-5 Tramadol 738-70-5 Trimethoprime 1401-69-0 Tylosin 101312-92-9 Valnemulin 99-66-1 Valproic acid 52-53-9 Verapamil 82626-48-0 Zolpidem Plasticizers 131-56-6 2,4-Dihydroxybenzophenone 7425-14-1 2-Ethylhexanoic acid 2-ethylhexyl ester 85-68-7 Benzylbutylphthalate (BBP)
Annexes 230 Plasticizers 80-05-7 Bisphenol A 84-66-2 Diethylphthalate (DEP) 68515-49-1, 26761-40-0 Di-isodecyl-phthalate (DIDP) 68515-48-0, 28553-12-0 Di-isononyl-phthalate (DINP) 131-11-3 Dimetylphthalate (DMP) 84-74-2 Di-n-butylphthalate (DBP) 117-84-0 Di-n-octylphthalate (DOP) 38640-62-9 DIPN 947-19-3 Methanone, Irgacure 184 3622-84-2 NBBS 872-50-4 NMP 77-90-7 Tributylacetylcitrate 126-71-6 Tri-iso-butylphosphate (TIBP) 6846-50-0 TXIB Trace metals and their compounds 7440-38-2 Arsenic 7440-47-3 Chromium 7440-50-8 Copper 03-05-7440 Palladium 12595-26-5 Silver 74-00-2 Tetraethyllead 75-74-1 Tetramethyllead 7440-66-6 Zinc Wood preservatives 51-28-5 2,4-Dinitrophenol (DNP) 106-44-5 para-Cresol
Annexes 231 Annex II – Analysis of acidic, basic and neutral pharmaceuticals in river waters: clean-up by 1º,2º amino anion exchange and enrichment using an hydrophilic adsorbent Published Scientific Paper: Gonçalves, C., Sousa, M.A. and Alpendurada, M.F., Analysis of acidic, basic and neutral pharmaceuticals in river waters: clean-up by 1º,2º amino anion exchange and enrichment using an hydrophilic adsorbent. International Journal of Environmental Analytical Chemistry, 2013. 93: 1-22.
Table 2. Instrumental conditions employed in LC-MS/MS analysis of pharmaceuticals. Pharmaceuticals Retention time (min.) Acquisition segment (min.) Drying Gas Temperature Needle voltage (V) Ionization mode Precursor ion (m/z) Product ions (m/z) Capillary Voltage (V) RF Loading (%) CID amplitude (V) Amoxicillin 0.398 0.0–2.3 300C 5000–4630 ESIþ366 (349), 208, 160 47 93 0.86 Hydrochlorothiazide 0.871 ESIÿ296 (269), 205, 271 84 128 1.10 Paracetamol 0.949 ESIþ152 (110) 41 60 0.81 Paracetamol D4 0.97 ESIþ156 (114) 60 54 0.65 Ciprofloxacin 1.319 ESIþ332 (314), 288, 315 65 88 1.30 Zolpidem 2.64 2.3–5.7 350C 3450 ESIþ308 (263), 235, 236 102 87 1.35 Azithromycin 3.724 ESIþ375 (591.5) 42 90 1.25 Bisoprolol 3.776 ESIþ326 (116) 80 87 0.86 Omeprazole 5.26 ESIþ346 (198), 180, 289 48 86 0.93 Furosemide 5.869 5.7–6.7 350C 4000–4600 ESIÿ329 (285) 53 117 1.10 Indapamide 5.972 ESIÿ364 (189), 233, 190 53 161 1.10 Paroxetine 6.117 ESIþ330 (192), 151 78 85 0.90 Bromazepam 6.361 5.7–7.5 ESIþ316 (288), 263, 209 87 87 1.10 Fluoxetine D5 6.52 6.7–7.5 350C 5300 ESIþ315 (153) 54 85 0.65 Fluoxetine 6.936 ESIþ310 (148) 52 86 3.00 Alprazolam 7.887 7.5–8.85 250C 5000–5000 ESIþ309 (281), 274, 241 92 87 1.28 Lorazepam 8.023 ESIþ321 (303), 305, 275 65 88 0.90 Nimesulide 8.26 ESIÿ307 (229) 53 103 0.90 Ketoprofen 8.589 7.5–10.0 ESIþ255 (209) 75 76 0.75 Diazepam 8.971 8.85–10.0 250C 3700–5000 ESIþ285 (257), 154, 222 87 81 1.10 Naproxen 8.932 ESIþ231 (185), 174 42 77 0.65 Bezafibrate 9.306 ESIÿ360 (274) 63 100 0.60 Diclofenac 10.537 10.0–11.0 300Cÿ5000 ESIÿ294 (250) 36 111 0.75 Ibuprofen 10.712 ESIÿ205 205* 33 57 0.00 Gemfibrozil 11.491 11.0–13.0 250C 3530–5540 ESIÿ249 (121), 127 48 67 0.80 Simvastatin 12.188 11.7–13.0 350C 3530–5540 ESIþ441 (325), 315, 310 126 66 1.00 m/z window is 3 m/z except: amoxicilin, bromazepam, alprazolam, naproxen is 4 m/z; fluoxetine, lorazepam is 5 m/z; ibuprofen, gemfibrozil is 6 m/z High mass eject factor is 100% except: paracetamol, paracetamol D4, fluoxetine, fluoxeine D5, ibuprofen is 50%; lorazepam, gemfibrozil is 80% Fluoxetine: CID frequency ÿ4; The product ion between parenthesis was used for quantitation. *M – H was used for quantitation 6C.M.O. Gonc¸alves et al. Downloaded by [b-on: Biblioteca do conhecimento online UP] at 04:55 22 January 2013
analytical procedure. Unless otherwise stated, (absolute) recoveries means the ratio expressed in percentage of the signal obtained in a sample (subjected to SPE and eventually clean-up) relative to the signal expected without any interference of any sort (direct injection of a standard). Matrix effect observed when analyzing real samples is probably the major source of inaccuracy of the results, therefore the extension of this phenomenon was assessed in two ways: effect on recoveries and ionization suppression effect. To estimate this effect four types of solutions were analyzed: standard solution (StdSol), SPE extract of spiked Milli Q water (SPESol), SPE extract of spiked real samples (SampSol) and spiked extract (after SPE) of real sample (ExtrSol). Spiking with pharmaceuticals was always performed at 500 ng L ÿ1 concentration level. Absolute recoveries in Milli Q water and real samples were determined as: Re c¼SPESol ðorSamplSol Þ 100 StdSol ð1Þ Ionization effect in the electrospray source (ESI) was determined as: IonizEffect ¼ðExtrSol ÿStdSol Þ 100 StdSol ð2Þ Matrix effect on recoveries was determined as: EfectRe c¼SampSol ExtrSol ÿSPESol StdSol 100 ð3Þ Additionally the loss of analyte in the clean-up step was determined comparing the response of Milli Q water solutions submitted to clean-up þSPE and just SPE and; samples spiked before and after clean-up. These trials also allowed investigating if acidification ought to be performed before or after clean-up. The physico-chemical characteristics of four river water matrices tested (1 Ave river, 2 Lec¸ a river and 1 Douro river) are given in Table 3. Lec¸ a Reg. is a clean matrix, Ave and Douro are medium loaded and Lec¸ a Pte Pedra is a highly polluted river water sample. 2.5 Sampling scheme in Lec¸ a and Douro rivers The assessment of the occurrence of pharmaceuticals in surface waters was carried out in two very distinct rivers in the north of Portugal. Lec¸ a river is a small water stream of 45 km Table 3. Physico-chemical characteristics of the river waters matrices tested in the clean-up and matrix effect studies. Ave river Lec¸ a river Reg. Lec¸ a river Pte Pedra Douro river TOC (mg L ÿ1 ) 1.96 1.41 22.23 2.30 NH4þ(mg L ÿ1 ) 0.512 0.107 43.35 0.032 NO2ÿ(mg L ÿ1 ) 0.042 0.03 0.094 0.235 NO3ÿ(mg L ÿ1 ) 9.50 8.8 17.00 6.4 Conductivity (mS cm ÿ1 ) 146 94.4 371.5 310 pH 6.83 6.16 7.44 7.79 Turbidity (NTU) 1.71 0.81 16.35 4.64 Dissolved Oxygen (mg L ÿ1 ) 7.34 6.84 2.21 7.89 International Journal of Environmental Analytical Chemistry 7 Downloaded by [b-on: Biblioteca do conhecimento online UP] at 04:55 22 January 2013
long and average flow of 3.4 m 3 s ÿ1 that receives in its path different inputs of agricultural and industrial origin as well as domestic effluents from a large urban population in the neighbourhood of Oporto, contributing to its high contamination status. Douro river is an international river with headwaters in Spain and mouth in Oporto with a stream of average 450 m 3 s ÿ1 (ranging from 200 to 710 m 3 s ÿ1 ) running a 927 km course. It is source of drinking water for the population living in the metropolitan area of Oporto. The monitoring scheme in Lec¸ a river included ten sampling stations, one in the headwaters, as control, and so forth until the mouth before and after every contamination source (villages, tributaries, WWTPs). The same rationale was applied in Douro river with a total of 12 sampling stations beginning from 45 km towards the mouth. Samples were collected in 1 L amber glass bottles and transported to the laboratory in a refrigerated environment. Samples were processed immediately or at latest three days after sampling and were kept always at 4C. The samplings took place in the 23 February and 15 June 2009 in Lec¸ a river and 4 March and 16 June 2009 in the case of Douro river. 3. Results and discussion 3.1 LC-MS/MS optimization Instrumental optimization began by setting up the appropriate electrospray ionization source parameters, MS/MS conditions and the chromatographic gradient to analyse a mixture of acidic, basic and neutral pharmaceuticals at the maximum sensitivity and resolution. Just the most critical aspects will be discussed, briefly, which derive from the challenge of analysing the above analytes in the same run. After an initial optimization by direct infusion of the standards, two observations highlighted: some analytes gave an extremely broad chromatographic peak under methanol/water mobile phase while others required better fine tuning for improving the sensitivity. The former group include, noticeably, ciprofloxacin, azithromycin, bisoprolol, paroxetine and fluoxetine. Aqueous mobile phases buffered at the pH of 2.9, 3.2, 3.7 and 4.6 were tested using formic acid and ammonium formate as additives in a concentration of 5 and 10 mM. A mobile phase composed of formic acid 10 mM in water (pH 2.9) revealed essential for obtaining good peak shape and resolution of the abovementioned pharmaceuticals. The tremendous improvement in peak shape is demonstrated in Figure S1. In general, peak widths at half height (W 0.5 ) of the studied pharmaceuticals ranged then from 4.5–11.2 s. Acidification of the mobile phase also contributed to improve the ionization efficiency of several analytes, therefore ESIþwas the preferred ionization mode, namely for amoxicillin, paracetamol, ketoprofen and naproxen, which are normally analysed under ESIÿ. A representative MS/ MS chromatogram of a 500 mg L ÿ1 standard solution obtained under the above conditions is given in Figure 1. A fast and efficient separation was achieved with a total chromatographic time of 13 min and a 22 min cycle time, owing to the use of a 50 mm and 2.4 mm particle size column. MS/MS segments with the least number of analytes (5) were scheduled to keep the required number of data points per peak. To address the improvement of sensitivity for some compounds, further optimization was attempted changing the precursor m/z isolation window and the high mass ejection factor. These conditions were given as footnote to Table 2. It should be noted that a larger isolation window also allowed collecting the isotopic pattern of some halogenated benzodiazepines in the precursor ion and MS/MS spectra, providing better confirmation quality. 8C.M.O. Gonc¸alves et al. Downloaded by [b-on: Biblioteca do conhecimento online UP] at 04:55 22 January 2013
Figure 1. Representative LC-MS/MS chromatogram of a standard solution of the 23 relevant pharmaceuticals at 500 mg L ÿ1 concentration. International Journal of Environmental Analytical Chemistry 9 Downloaded by [b-on: Biblioteca do conhecimento online UP] at 04:55 22 January 2013
3.2 Comprehensive adsorbent comparison Surface modification of styrene-divinylbenzene and synthesis of divinylbenzene-N- vinylpyrrolidone and methacrylate-divinylbenzene co-polymers have been pursued to better retain polar environmental pollutants, such as pharmaceuticals, in SPE cartridges [20]. We have performed a comprehensive evaluation of 17 SPE materials available nowadays, some of them neglected in previous studies [19], for the enrichment of a wide array of pharmaceuticals. Irrespective of their order, the five adsorbents that performed worst in a first round were: Chromabond C18 and HRP, Lichrolut EN, Bond Elut ENV and Isolute ENVþ. The remaining 12 were then systematically compared in four replicates with an average relative standard deviation (RSD) per cartridge of 5.7 to 20% and the ranking of average recoveries/median recoveries at 500 ng L ÿ1 concentration level was as follows: JTBaker H 2 Ophilic (89/92%), JTBaker H 2 Ophobic (86/96%), Abselut Nexus (86/90%), Resprep RDX (80/87%), Oasis HLB (80/85%), JTBaker SDB1 (73/77%), Bond Elut PPL (69/78%), Strata X (67/67%), Oasis MCX (66/81%), Chromabond HRX (62/61%), Chromabond Easy (58/62%) and Evolute ABN (56/64%). Of these, the best five are represented in Figure 2 for the whole group of test compounds. For the majority of pharmaceuticals the five adsorbents are comparable, namely for antidepressives and anxiolytics, nevertheless five cases must be highlighted: paracetamol and simvastatin are extracted much more efficiently by JTBaker H 2 Ophilic and H 2 Ophobic while for ciprofloxacin the H 2 Ophilic is just superpassed by Abselut Nexus with an abnormally high recovery rate. On the contrary, hydrochlorothiazide and ibuprofen are better extracted by Oasis HLB and Resprep RDX. The efficiency obtained for simvastatin, paracetamol and ciprofloxacin has driven the selection of JTBaker H 2 Ophilic as the most appropriate for the set of pharmaceuticals. The full results for the 12 best adsorbents are given in Table S1 (Supplementary Information). According to the literature survey performed by Kostopoulou and Nikolaou until 2008 [6], the most effective SPE adsorbents for the enrichment of pharmaceuticals in aqueous samples were Isolute ENVþ, Oasis HLB, Oasis MAX, Oasis MCX, Strata-X, LiChrolut C18 and LiChrolut EN. The most popular pre-concentration technique is undoubtedly SPE with a few applications of solid-phase microextraction (SPME) using fibres coated with polyacrylate and carbowax/divinylbenzene [6]. Nevertheless, the limits of detection never rival those of SPE [6]. On the other hand, van de Steene concluded that among Oasis HLB, C8, Phenyl, Strata-X and Strata SCX, the former was the best cartridge for the extraction of nine basic pharmaceuticals, mostly antimycotics, but it also led to intense matrix effect (non selective extraction) [8]. The study of Weigel et al. [19] also included some specialized polystyrene-divinylbenzene based adsorbents and co-polymers with noteworthy similarities to our results. The extraction efficiencies of Abselut Nexus, Chromabond Easy and JTBaker SDB1 for the commonly tested pharmaceuticals, although different in the figures, follow the same trend. The recoveries we obtained were generally higher probably because the samples were acidified. That can be important for acidic substances such as bezafibrate, diclofenac and ibuprofen but also for fluoxetine and related analytes as it will be demonstrated later. The above references show that most of the tested cartridges in the present work, namely the most efficient ones JTBaker H 2 Ophilic, H 2 Ophobic and Resprep RDX, are usually disregarded. Hierarchical cluster analysis (HCA) of the recoveries achieved by the twelve adsorbents revealed that JTBaker H 2 Ophilic, H 2 Ophobic and Abselut Nexus have a similar extraction profile, so as Oasis HLB and Resprep RDX. JTBaker SDB1 and Bond Elut PPL cluster 10 C.M.O. Gonc¸alves et al. Downloaded by [b-on: Biblioteca do conhecimento online UP] at 04:55 22 January 2013
together and at a later stage a link is formed with the previous two. The two Chromabond phases HRX and Easy also link together and at a much farther stage they join the remaining group other than the JTBaker H 2 Ophilic’s. The respective dendrogram is given in Figure S2. This information can be useful when deciding from different brands since it shows how the adsorbents compare among them for a representative list of pharmaceuticals. Not to undervalue, of course, the other way round of how the substances behave 0 20 40 60 80 100 120 140 160 Hydrochlorothiazide Paracetamol Ciprofloxacine Azithromycin Bisoprolol Furosemide Indapamide Paroxetine Bromazepam Fluoxetine recovery % JTBaker H2Ophilic JTBaker H2Ophobic Abselut Nexus Resprep RDX Oasis HLB 0 20 40 60 80 100 120 140 Alprazolam Lorazepam Nimesulide Ketoprofen Diazepam Naproxen Bezafibrate Diclofenac Ibuprofen Gemfibrozil Simvastatin % recovery JTBaker H2Ophilic JTBaker H2Ophobic Abselut Nexus Resprep RDX Oasis HLB Figure 2. Recovery rates of the five most efficient SPE adsorbents for the enrichment of pharmaceuticals in Milli Q water at 500 ng L ÿ1 concentration (n¼5). Error bars represent the standard deviation. Omeprazole was not represented since it is not extracted at pH 2 and amoxicillin gives erratic results due to its unstable character. International Journal of Environmental Analytical Chemistry 11 Downloaded by [b-on: Biblioteca do conhecimento online UP] at 04:55 22 January 2013
regarding the different adsorbents. The respective dendrogram is given in Figure S3. Chemical family affinity is apparent for anti-inflammatories (Group 1) and antidepressives (Group 2), however benzodiazepines scatter between two major groups where antiinflammatories and lipid regulators are present (Groups 1 and 3). Bromazepam clusters in a more heterogeneous group (Group 4) composed of hydrochlorothiazide, azithromycin and indapamide. Two aspects should be noted: chemical variability is also present among benzodiazepines expressed in their different chromatographic retention times (from 6.37 to 8.97 min) and; the similar or dissimilar behaviour of bromazepam with the other group members is not clearly seen in Figure 2 because only five adsorbents are displayed whereas HCA encompassed all of them. For Group 1 all five adsorbents present in Figure 2 are equally good with average recoveries above 95%. For Group 3 JTBaker H 2 Ophilic, H 2 Ophobic and Nexus are preferred giving average recoveries between 88 and 92%. For Group 4, Nexus is the most suitable with average recovery of 83%. Group 2 approaches the case of Group 1 except that Oasis MCX is totally unsuitable (Table S1). Of superior evidence, HCA demonstrated that simvastatin, paracetamol and ciprofloxacin showed the most dissimilar pattern thus outweighed the selection of JTBaker H 2 Ophilic as the most suitable adsorbent. 3.3 Effect of sample pH, flow rate and breakthrough volume of the adsorbents Sample pH is a very important parameter when enriching substances containing diverse chemical moieties, such as the pharmaceuticals, in SPE adsorbents by reversed-phase interaction. This assumption was proved by testing the following sample pHs: 2, 3, 4, 7.3 and natural pH 5.4. The results obtained are given in Figure 3. The antibiotics azithromycin and ciprofloxacin require acidification to the lowest pH (pH 2) to improve recovery rates while paroxetine and fluoxetine are efficiently extracted at a moderately acid pH (pH 4). Bisoprolol benefits from acidification in a much smaller scale. On the contrary, omeprazole becomes ionized at low pH and extraction is reduced to complete loss. The extraction of the remaining pharmaceuticals is not significantly affected by sample pH and they well tolerate a pH of 2, therefore this value was considered the most appropriate to undertake the analysis, although omeprazole cannot be analysed. The flow rate of sample percolation was studied between 5 and 20 mL min ÿ1 . Generally, the extraction efficiency was independent of the flow rate except for selected compounds. A continuous loss in retention capacity was observed for azithromycin, ciprofloxacin, paroxetine and fluoxetine, with a more intense drop changing from 5 to 10 mL min ÿ1 for antibiotics (18 and 22%, respectively) than antidepressives (7 and 9%, respectively). Bisoprolol and bromazepam lost retention strength just at 15 and 20 mL min ÿ1 flow rate. Interestingly, the retention efficiency of simvastatin increased three fold at 20 compared to 5 mL min ÿ1 . It has not been proved whether simvastatin degrades during 40 min of sample extraction or if adsorption to tubing is reduced at 20 mL min ÿ1 . A 10 mL min ÿ1 flow rate was finally chosen corresponding to 25 min sample processing time compared to 45 min at 5 mL min ÿ1 . Subsequently, the breakthrough volume was determined under the conditions above. The results in Figure 4 show a good proportionality between the extracted volume and the analytical response at a constant concentration of pharmaceuticals for the majority of analytes. Therefore, the breakthrough volume is above 1000 mL and the retention capacity is independent of the volume of sample. Nevertheless, paracetamol starts leaking from the 12 C.M.O. Gonc¸alves et al. Downloaded by [b-on: Biblioteca do conhecimento online UP] at 04:55 22 January 2013
cartridge at sample volumes around 200 mL, which is related to its hydrophilicity. Omeprazole, on the other hand, is not possible to enrich on the adsorbent under acidic conditions and the recovery is low at 100 mL volume and is further lost. In view of the above results and the necessary sensitivity of pharmaceutical analysis in surface waters, 200 mL of sample are SPE processed and eluted in 2 steps of 3 mL of methanol (1 min-. waiting time in between). The dry residue is redissolved in 200 mL of a mixture of methanol/water 35 : 65 and injected for LC-MS/MS analysis. 3.4 Clean-up procedure to overcome matrix effects Evaluating matrix effects is very important when developing an LC–MS/MS method. The results of van de Steene et al. [8] prove that matrix effects are very difficult to tackle and combined multiple approaches are necessary to reduce their impact in precision and accuracy. Normalization of erroneous analyte response, either signal suppression or enhancement, caused by co-eluting compounds originated from the matrix can be 0 20 40 60 80 100 120 Azithromycin Ciprofloxacin Paroxetine Fluoxetine Bisoprolol Omeprazole Diazepam Alprazolam Lorazepam Bromazepam Simvastatin % Recovery pH 2 pH 3 pH 4 pH 5,4 pH 7,3 0 20 40 60 80 100 120 140 Ketoprofen Paracetamol Nimesulide Naproxen Furosemide Bezafibrate Diclofenac Indapamide Hydrochlorot. Gemfibrozil Ibuprofen % Recovery pH 2 pH 3 pH 4 pH 5,4 pH 7,3 Figure 3. Effect of sample pH on the retention of some critical and representative pharmaceuticals in the JTB H 2 Ophilic adsorbent (n¼4). Error bars represent the standard deviation of the mean. Tests were performed at a concentration level of 500 ng L ÿ1 in Milli Q water. International Journal of Environmental Analytical Chemistry 13 Downloaded by [b-on: Biblioteca do conhecimento online UP] at 04:55 22 January 2013
endeavoured by several methods: use of isotopically labelled internal standards, removal of interfering substances by clean-up, selective extraction of different types or standard addition for calibration and calculation of recoveries. Selective extraction, and clean-up in particular, extend the life-time of the chromatographic column [22]. Sample to sample variability is another issue to keep in mind: compensation by recoveries calculated for each sample or use of isotopically labelled internal standards are the best options to achieve accurate results. The pharmaceuticals most affected by matrix effect (mainly ionization suppression rather than interference in enrichment efficiency) with average (relative) recovery reduction above 30% were: hydrochlorothiazide, azithromycin, furosemide, indapamide, fluoxetine, alprazolam, naproxen and diclofenac, considering the four matrices in Table 4. Signal enhancement was observed for lorazepam (18%). More detailed information on the type of matrix effect (either ionization suppression or effect on recoveries) obtained according to the calculations explained in section 2.4 can be found in Table S2). Indeed, ionization suppression could reach 46% signal reduction in the dirtiest sample. Additionally, azithromycin and bisoprolol suffered peak broadening and increase of retention time whereas paroxetine and fluoxetine were just affected by retention time deviation. Retention time shifting is critical in MS/MS methods scheduled in segments. Fluoxetine D5 was used to monitor retention time variation of antidepressives while paracetamol D4 was used for compensation of matrix effect on the native paracetamol. For the remaining pharmaceuticals correction by recoveries on related matrix was used spiking the sample with pharmaceuticals at 500 ng L ÿ1 concentration. Additionally, clean-up was employed in the dirtiest samples. 0.0E+00 2.0E+07 4.0E+07 6.0E+07 8.0E+07 1.0E+08 1.2E+08 mL Area Azithromycin Paroxetine Bisoprolol Diazepam Alprazolam Lorazepam Ciprofloxacin Fluoxetine Bromazepam Simvastatin Ketoprofen 0.0E+00 2.0E+05 4.0E+05 6.0E+05 8.0E+05 1.0E+06 1.2E+06 1.4E+06 1.6E+06 1.8E+06 2.0E+06 mL Area Paracetamol Nimesulide Naproxen Furosemide Bezafibrate Hydrochlorothiazide 0.0E+00 5.0E+04 1.0E+05 1.5E+05 2.0E+05 0 200 400 600 800 1000 0 200 400 600 800 1000 0 200 400 600 800 1000 mL Area Omeprazole Indapamide Diclofenac Gemfibrozil Ibuprofen Figure 4. Study of breakthrough volume of the JTBaker H 2 Ophilic adsorbent for several pharmaceuticals in a concentration of 500 ng L ÿ1 in Milli Q water. 14 C.M.O. Gonc¸alves et al. Downloaded by [b-on: Biblioteca do conhecimento online UP] at 04:55 22 January 2013
Table 4. Recoveries of pharmaceuticals in Milli Q water and surface water of the Ave, Lec¸ a and Douro rivers at 500 ng L ÿ1 . Douro water was submitted to clean-up with different anion exchange adsorbents to evaluate their suitability. Recoveries (%) without clean-up Recoveries (%) with clean-up Milli Q (n¼4) Ave (n¼3) Lec¸ a Reg. (n¼3) Lec¸ a Pte Pedra (n¼3) Douro (n¼6) Lec¸ a Pte Pedra þ 1, 2amino (n¼3) Douro þ 1, 2amino (n¼6) Douro þNH2 (n¼6) Douro þSAX (n¼6) Hydrochlorothiazide 50 16 19 21 12 33 28 16 11 Paracetamol 68 50 62 29 38 36 44 37 35 Ciprofloxacin 87 73 81 61 70 81 107 100 95 Zolpidem 98 96 97 88 104 92 94 96 103 Azithromycin 68 25 58 0.1 56 51 48 48 47 Bisoprolol 101 97 95 87 94 90 90 94 89 Furosemide 76 54 68 33 20 39 41 33 25 Indapamide 70 33 45 7 12 35 33 19 15 Paroxetine 99 77 77 64 77 86 89 83 36 Bromazepam 75 64 61 61 65 74 58 61 57 Fluoxetine 109 55 57 52 93 67 104 89 47 Alprazolam 89 61 63 55 56 66 62 64 56 Lorazepam 97 118 113 117 113 93 108 112 122 Nimesulide 110 130 133 98 47 89 85 76 66 Ketoprofen 101 88 90 89 81 91 77 73 68 Diazepam 92 90 86 81 85 84 82 80 75 Naproxen 102 71 73 57 53 66 57 41 35 Bezafibrate 96 75 85 58 49 72 59 51 52 Diclofenac 86 71 52 36 60 49 65 58 43 Ibuprofen 81 82 105 62 82 77 102 109 87 Gemfibrozil 86 95 91 75 80 74 81 70 84 Simvastatin 46 28 19 2 50 39 46 36 31 Average Recoveries 86 69 73 57 63 67 71 66 58 International Journal of Environmental Analytical Chemistry 15 Downloaded by [b-on: Biblioteca do conhecimento online UP] at 04:55 22 January 2013
supplying samples of the SPE cartridges. This work was carried out under the project NORTE-01- 0162-FEDER-000023 co-funded by ON.2 O Novo Norte. References [1] A. Kot-Wasik, J. Debska, and J. Namiesnik, Trends in Anal. Chem. 26, 557 (2007). [2] E. Zuccato, S. Castiglioni, R. Fanelli, G. Reitano, R. Bagnati, C. Chiabrando, F. Pomati, C. Rossetti, and D. Calamari, Environ. Sci. Pollut. Res. 13, 15 (2006). [3] W.W. Buchberger, Anal. Chim. Acta 593, 129 (2007). [4] T. Heberer, J. Hydrology 266, 175 (2002). [5] D. Barcelo ´, Trends in Anal. Chem. 22, xiv (2003). [6] M. Kostopoulou and A. Nikolaou, Trends in Anal. Chem. 27, 1023 (2008). [7] T.V. Madureira, J.C. Barreiro, M.J. Rocha, E. Rocha, Q.B. Cass, and M.E. Tiritan, Sci. Total Environ. 408, 5513 (2010). [8] J.C. van de Steene, K.A. Mortier, and W.E. Lambert, J. Chromatogr. A 1123, 71 (2006). [9] J. Debska, A. Kot-Wasik, and J. Namiesnik, Critical Reviews in Anal. Chem. 34, 51 (2004). [10] J.D. Cahill, E.T. Furlong, M.R. Burkhardt, D. Kolpin, and L.G. Anderson, J. Chromatogr. A 1041, 171 (2004). [11] K. Fent, A. Weston, and D. Caminada, Aquatic Toxicol. 76, 122 (2006). [12] D. Fatta, A. Nikolaou, A. Achilleos, and S. Meric¸ , Trends in Anal. Chem. 26, 515 (2007). [13] S.D. Richardson, Anal. Chem. 78, 4021 (2006). [14] C. Prasse, M.P. Schlu ¨sener, R. Schulz, and T. Ternes, Environ. Sci. Technol. 44, 1728 (2010). [15] C. Gonc¸ alves, S. Pe ´rez, V. Osorio, M. Petrovic, M.F. Alpendurada, and D. Barcelo ´, Environ. Sci. Technol. 45, 4307 (2011). [16] M.J. Go ´mez, M.M. Go ´mez-Ramos, O. Malato, M. Mezcua, and A.R. Fe ´rnandez-Alba, J. Chromatogr. A 1217, 7038 (2010). [17] Entec UK limited, Technical Support for the Impact Assessment of the Review of Priority Substances under Directive 2000/60/EC, report addressed to the European Commission DG Environment, August 2010. [18] R. Rodil, J.B. Quintana, P. Lo ´pez-Mahı´a, S. Muniategui-Lorenzo, and D. Prada-Rodrı´guez, J. Chromatogr. A 1216, 2958 (2009). [19] S. Weigel, R. Kallenborn, and H. Hu ¨hnerfuss, J. Chromatogr. A 1023, 183 (2004). [20] N. Fontanals, R.M. Marce ´, and F. Borrull, Trends in Anal. Chem. 24, 394 (2005). [21] M.A. Sousa, C. Gonc¸ alves, E. Cunha, J. Hajs ˇlova ´, and M.F. Alpendurada, Anal. Bioanal. Chem. 399, 807 (2011). [22] N.M. Vieno, T. Tuhkanen, and L. Kronberg, J. Chromatogr. A 1134, 101 (2006). [23] N. Vieno, T. Tuhkanen, and L. Kronberg, Environ. Sci. Technol. 39, 8220 (2005). [24] A.P. Ferreira, Cadernos de Sau´ de Pu´ blica 21, 1884 (2005). [25] J.J. Carvalho, M.G. Weller, U. Panne, and R.J. Schneider, Anal. Bioanal. Chem. 396, 2617 (2010). [26] I.J. Buerge, T. Poiger, M.D. Muller, and H.-R. Buser, Environ. Sci. Technol. 37, 691 (2003). 22 C.M.O. Gonc¸alves et al. Downloaded by [b-on: Biblioteca do conhecimento online UP] at 04:55 22 January 2013
Annexes 255 Annex III – Report of the intercalibration exercise under PHARMAS Project: “Ecological and human health risk assessments of antibiotics and anti-cancer drugs found in the environment”
P PH HA AR RM MA AS S Ecological and human health risk assessments of antibiotics and anticancer drugs found in the environment Contract n° 265346 Operative commencement date of the project: January 1st 2011 Final date of the project: December 31th 2014 Deliverable number: D1.6 “Report of the intercalibration exercise” The deliverable authors are responsible for the content AUTHOR: Benoit Roig AFFILIATION: EHESP ADDRESS: Avenue du prof Leon Bernard – 35043 Rennes – France TEL.: +33 4 7 86 28 45 51 EMAIL: [email protected] FURTHER AUTHORS: Document Information DOCUMENT TYPE Report DOCUMENT NAME: Report of the Intercalibration exercise REVISION: 1 REV.DATE: CLASSIFICATION: STATUS:
Table of content Abstract ...................................................................................................................................... 1 1. Introduction ......................................................................................................................... 2 2. Participants profile .............................................................................................................. 2 3. Material preparation ............................................................................................................ 4 3.1. Standard preparation ....................................................................................................... 4 3.2. Reference sample preparation ......................................................................................... 5 4. Homogeneity of samples ..................................................................................................... 6 5. Stability of samples ............................................................................................................. 7 6. Sample distribution ............................................................................................................. 9 7. Analytical methods ............................................................................................................. 9 8. Evaluation procedure ........................................................................................................ 12 9. Results ............................................................................................................................... 13 9.1. z score ............................................................................................................................ 14 9.2. Results by molecules and matrices ................................................................................ 17 9.3. Coefficients of variation ................................................................................................ 19 9.4. Laboratory performances .............................................................................................. 20 10. Conclusion ..................................................................................................................... 22 11. References ..................................................................................................................... 23 ANNEX 1: ................................................................................................................................ 24 ANNEX 2: ................................................................................................................................ 28 ANNEX 3: ................................................................................................................................ 30 ANNEX 4: ................................................................................................................................ 35 ANNEX 5: ................................................................................................................................ 38 STATUS, CONFIDENTIALITY AND ACCESSIBILITY Status Confidentiality Accessibility S0 Approved/Released R0 General public Work-space x S1 Reviewed R1 Restricted to pharmas partners x Internet S2 Pending for review x R2 Restricted to European. Commission Paper S3 Draft for comments S4 Under preparation
1 Abstract An inter-laboratory exercise was organized under the PHARMAS EU project, by the Advanced School of Public Health (EHESP), in order to evaluate the reproducibility of analytical method for the measurement of antibiotics (ATB) in waters. For the first time, this type of exercise has been organized in Europe, and using different kind of analytical methods and devices. In this exercise thirteen laboratories from five countries (Canada, France, Italy, the Netherland and Portugal) participated, and a total number of 78 samples were distributed. During the exercise, 2 testing samples (3 bottles of each) prepared from tap water and river water respectively, spiked with ATB, were sent to the participants and analyzed along one month. A final number of 77 (98.7 %) testing samples were analyzed. Depending on substances analysis by each participant, 305 values in duplicate were collected, and the results for each sample were expressed as the target concentration. A statistical study was initiated using 611 results. The mean values, standard deviations (σ), variances (σ2) and, upper and lower warning limits (UWL and LWL) between concentrationresult from the participant laboratory at different intervals were obtained. In this exercise, 24 results (3.93 % of accounted values) were outliers according with the Zscore values and the Dixon test. The outlier results were excluded. In order to establish the stability of testing samples along the exercise, differences between variances obtained for every type of sample at different intervals were evaluated. The results showed not significant variations and can be considered that all samples were stable during the exercise. The goals of this inter-laboratory study were to evaluate the repeatability (r) and reproducibility (R) when different laboratories conduct the analysis, the influence of different matrix samples, the variability between different methods, and to determine the rate at which participating laboratories successfully completed the tests initiated.
2 1. Introduction The presence of antibiotics in environmental samples has been reported extensively. Antibiotic are among the most investigated class of pharmaceuticals in the environment. More recently, they have been detected in drinking water (Bruchet et al., 2005; Ye et al., 2007; Focazio et al., 2008; Benotti et al., 2009, Vulliet et al., 2009; Watkinson et al., 2009). However, concentrations are very low and accurate and analytical methods need to be controlled in order to insure a good uncertainty of the data obtained required for a robust assessment of exposure scenarios of human people. In this context, an inter-laboratory comparison exercise (announcement and recommendation have been reported in annex 1) was conducted among thirteen laboratories covering different EU countries. 2 samples were analyzed along one month. The samples correspond to 3 glass bottles of tap water for the first and 3 glass bottles of river water (filtered sample from La Vilaine river in Brittany, France ) for the second. The principal aims of the interlaboratory study were: - To evaluate the variability of results between different laboratories - To evaluate the rate at which participating laboratories successfully completed the exercise - To evaluate the capacity and variability in front of complex real samples 2. Participants profile Interlaboratory exercise is an opportunity for one laboratory to assess their performances by comparison with several other laboratories and to quantify the modification necessary to improve them. 13 laboratories (annexe 2) participed to the exercise. 5 countries are represented (Canada, France, Italy, The Netherland, Portugal). 1 (one) laboratory used two analytical methods and was considered as 2 different ones. Then 14 “labs” have been considered. For reasons of anonymity, they have been numbered randomly from lab1 to lab14. 50% have already participated to EIL 1/3 of laboratories have less than 2 years of experience in the analysis of ATB in water 25% have more than 6 years of experience in the analysis of ATB in water
3 40% have between 3 and 5 years of experience in the analysis of ATB in water 92% perform routine analysis of ATB in their lab (83% of which several times a week and 17% at least once every 6 months). Analytical know-how: Figure 1 show the % of laboratories being able to analysis the ATB proposed. Figure 1 : Analytical Know-how of the participating laboratories Performances of the whole consortium are reported in the Table 1. They take into account the limit of detection of each laboratory. Min and Max correspond to the lower and higher value of LOD among the laboratories. Median and mean have been calculated on the basis on the average of the 14 values of LOD declared by the laboratories. n is the number of laboratories analyzing the corresponding substance. Table 1: Analytical performances of the participating laboratories LOD repartition (ng/L) n Min Max Mean Median Erythromycin (ERY) 10 1 500 80.3 10 Trimethoprim (TMP) 10 1 50 16.6 5.5 0 10 20 30 40 50 60 70 80 90 100 CHL CIP CFT CFU DOX ERY LEV OFL OXY SMX TET TMP % of laboratories
4 Sulfamethoxazole (SMX) 9 1 50 13 10 Ciprofloxacin (CIP) 8 1 100 32.13 20.5 Oxytetracycline (OXY) 7 1 50 26.9 25 Ofloxacin (OFL) 7 1 20 12.2 11 Tetracycline (TET) 6 1 100 39.5 37.5 Chlortetracycline (CHL) 5 1 200 48.4 11 Doxycycline (DOX) 3 5 100 51.7 50 Levofloxacin (LEV) 2 5 50 27.5 27.5 Cefuroxim (CFU) 1 50 Ceftriaxone (CFT) 0 When available, the substances were analyzed by non accredited method except ERY in three laboratories and SMX in two laboratories. Only one laboratory is accredited for 8 substances. Considering the different performances and capabilities of the participating laboratories and the capacity of the host laboratory to produce the reference samples, the following antibiotics were chosen (Table 2) Table 2 : Final substances studied during the exercise Class of antibiotic Molecule Acronym CAS registry number Fluoroquinolones Ofloxacin OFL 83380-47-6 Ciprofloxacin CIP 85721-33-1 Sulfonamide Sulfamethoxazole SMX 723-46-6 Pyrimidin (associated to sulfonamide) Trimethoprim TMP 738-70-5 Macrolide Erythromycin ERY 114-07-8 3. Material preparation 3.1. Standard preparation Acetonitrile (HPLC gradient, Baker), methanol (HPLC gradient, Baker), formic acid (Baker), NH4OH 25% (Carlo Erba), sodium nitrite 99% (ACS Reag. PhEur, Merck) and ultrapure water were used for the preparation of standards solution.
5 Standards are commercialized under powder form with purity between 97 to 99,9%. Trimethoprim standard was purchased from VWR (certified quality,Dr. Ehrenstorfer GmbH, Bgm.-Schlosser-Str. 6 A, 86199 Augsburg, Germany) Standards of other antibiotics were purchased from Sigma Aldrich. Life durations are not certified by the provider but 3 years of preservation are insured by a certificate. Mother solutions are solutions of individual compounds prepared in Methanol. CIP and OFL solutions are prepared as 0.5 g/L; the other ones at 1g/L. CIP solution requires addition of formic acid for a better solubility. Daughter solutions (DS) are prepared in acetonitrile. A first DS (DS1) at 10mg/L followed by a second DS (DS2) at 0.5mg/L are prepared by dilution of the mother solutions. DS are preserved at -20°C. The characterization (stability and homogeneity test) of the reference sample has been conducted by UPLC/MS/MS using internal standard. Internal standard solutions have been produced from the following marked molecule: Ofloxacin-d3, Trimethoprim-13C, Ciprofloxacin-13C-15N, Sulfamethoxazole-13C, Erythromycin-13C purchased from Sigma Aldrich (TMP, ERY and CYP are solutions in 1.2 ml flask) Mother solutions of OFL-d3 and SMX-13C are prepared in Methanol (1g/L) followed by a dilution in acetonitrile up to solution at 10mg/L. ERY-13C, TMP-13C, CIP-13C-15C are prepared directly in acetonitrile (12pp). A solution standard of reference marked compound at 0.5 mg/L is finally prepared in acetonitrile. 3.2. Reference sample preparation Two reference samples (called sample A and sample B) were prepared. Sample A and sample B correspond to treated and filtrated surface water respectively spiked with the 5 antibiotics of interest. The characteristics of the matrices are presented in Table 3.
12 8. Evaluation procedure The statistical evaluation was executed according to ISO 13528 “Statistical methods for use in proficiency testing by interlaboratory comparisons”. For each series the mean value (X), the standard deviation (σ), coefficient of variation (CV), standard error of the mean (M), median (M), the minimum (Min) and maximum (M) values of each series as well as the 95% confidence interval have been calculated or reported. As an acceptance criteria for each result were used the z-score function according with the “Laboratory Accreditation & Audit Protocol: Food Inspection Directorate”3,4, following the AOAC & ISO & IUPAC directives. The Z values where calculated according to the following expression: Where: Xlab= result for a laboratory X = mean values between all laboratories σ = Standard deviation in the correspondent population (accounting the results obtained for the different laboratories in this series in front of this sample) The results whose Z value was over 3 was directly excluded and when the Z-score value was between 2 and 3 was applied the Dixon Test with a 5% of Significance Level. In addition, laboratory biases (D) were estimated for each results (or average of results) reported by each participants. D corresponds to the difference between the value (or average value) of laboratory (x) with the reference value (X) D = x – X According to ISO 13528, the biases were classified into three categories: D≥3.0 indicating an “action signal”, 2.0≤D<3.0 considered as “warming signal” and -2.0 ≤D<2.0indicating “acceptable value”. The outlier results were excluded from the calculation of D.
13 Distribution of data produced by the participants for each substance was displayed in box plot (a.k.a. tuckey and whisker diagram). This standardized representation is based on five numbers: minimum, first quartile, median, third quartile, and maximum. In the simplest box plot the central rectangle spans the first quartile to the third quartile. A segment inside the rectangle shows the median and "whiskers" above and below the box show the locations of the minimum and maximum values (Figure 2: General scheme of simple Box plot). Figure 2: General scheme of simple Box plot Degree of variation of the data obtained between the different participants but also obtained in a same laboratory was estimated by the calculation of the coefficient of variation. It is expressed in % and represents the ratio of the standard deviation to the mean. 9. Results A total number of 13 participants took place of this study, by using the methods detailed in Table 7. One participant using two analytical methods, we have considered 14 series of data. Each participant received 3 bottles of sample A (spiked treated water) and 3 bottles of sample B (spiked surface water). Statistical analysis was performed by considering the spiking value as the reference value of the antibiotic of concern (Table 8). Table 8 : Reference value for each antibiotic in the two matrices Sample A : treated water Sample B : Surface Water ERY OFL SMX TMP ERY CIP OFL SMX TMP Spiking level (ng/L) 55 40 30 50 200 180 200 100 150 The five anibiotics have not been measured by all laboratories. Table 9 shows the repartitions of the antibiotics of interest by laboratories. Maximum value Minimum value Median First quartile Third quartile
14 Table 9 : Substances and laboratories Lab Subst. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 ERY CIP OFL SMX TMP A total number of 611 results were collected (annex 4). The mean values, standard deviations (), coefficient of variation, standard error of mean, median, lower and higher values between results from the participant laboratories at different intervals were calculated. Notice that ciprofloxacin in sample A was determinate only in two laboratories. The number of data is too weak to be considered in the statistical analysis. Consequently, CIP in sample B (surface water) have been only investigated. 9.1. z score z score values were used to calculate the outliers within the results. The calculation gave 24 outliers (3.9% of the total number of results, see Table 10). The outliers were not distributed evenly across all labs. Only 4 labs produced outliers and only on one substance. Figure 3 presents the absolute values of z score after exclusion of the outliers. According to the difference of the analytical protocols used, the number of outliers obtained in this exercise can be considered low. The sample matrix with the higher number of outlier was treated water.
15 Figure 3 : z score absolute values The outlier values were excluded for the final data treatment and the statistical parameters (mean values, standard deviation, variance, coefficient of variation) were calculated. Table 10 shows the corrected statistical values (after outlier exclusion) obtained for each compound in the different samples. 0 0,5 1 1,5 2 2,5 3 1 2 3 4 5 6 7 8 9 10 11 12 13 14 z score Lab N° Sample A Ery OFL SMX A Trimethoprim 0 0,5 1 1,5 2 2,5 3 1 2 3 4 5 6 7 8 9 10 11 12 13 14 z score Lab N° Sample B Ery OFL SMX B Trimethoprim Ciprofloxacin
16 Table 10 : Statistical values corrected after outlier exclusion for each compound in the different types of water: mean (x), standard deviation (), coefficient of variation (CV), standard error of mean (M), median (M), minimum value (Min), maximum value (Max). Substance Matrix Spiking level (ng/L) Nbr of accepted results x (ng/L) (ng/L) CV M (ng/L) M (ng/L) Min (ng/L) Max (ng/L) 95% confidence interval Nbr of outlier from to Cyprofloxacin Surface Water (B) 180 55 137.96 62.44 0.45 8.42 135.03 22.59 298.54 121.12 154.80 0 Erythromycin Treated water (A) 55 60 58.64 34.66 0.59 4.47 45.73 13.66 145.45 49.70 67.59 6 Surface water (B) 200 58 296.87 252.89 0.85 33.21 170.47 54.74 1048 230.46 363.28 6 Ofloxacin Treated water (A) 40 54 45.20 21.64 0.48 2.95 40.43 6.19 93.05 39.31 51.09 0 Surface Water (B) 200 54 211.51 83.12 0.39 11.31 197.78 57.49 433.0 188.89 234.14 0 Sulfamethoxazole Treated water (A) 30 70 20.0 5.67 0.28 0.68 19.83 5.64 30.51 18.64 21.35 6 Surface water (B) 100 76 87.06 26.73 0.31 3.07 89.28 26.86 127.07 80.96 93.16 0 Trimetoprim Treated water (A) 50 80 39.91 12.77 0.32 1.43 39.43 18.86 74 37.05 42.76 4 Surface water (B) 150 80 105.79 32.98 0.31 3.69 110.70 30 178 98.41 113.16 2 Total 587 24
17 9.2. Results by molecules and matrices Results for the different samples expressed in ng/L, mean and reference values are plotted in Figure 4. For all molecules, the tendency is the same in the two matrices concerning the comparison of the mean of all values and the reference value. Two behaviors can be distinguished: ERY and OFL show higher mean than the reference value; SMW, TMP and CIP show the contrary. The deviation between the two values (calculated by the ration between the observed mean concentration and the reference one) is important for ERY in surface water (148%) and SMX in treated water (66%). It corresponds to 106%, 88% 71%, 77% in surface water for OFL, SMX, TMP, CIP respectively and to 106%, 113%, 80% in treated water for ERY, OFL, TMP. This difference can be considered as representative in the case of SMX in treated water and TMP in surface water because the value is higher than the standard deviation (Table 10). 0 200 400 600 800 1000 1200 Lab 1 Lab 2 Lab 6 Lab 7 Lab 8 Lab 9 Lab 11 Lab 12 Lab 13 Lab 14 ng/L Erythromycin, Surface water Mean (296.9) Reference value 0 20 40 60 80 100 120 140 160 Lab 1 Lab 2 Lab 6 Lab 7 Lab 8 Lab 9 Lab 11 Lab 12 Lab 13 Lab 14 ng/L Erythromycin, treated water Mean (58.6) Reference value 0 50 100 150 200 250 300 350 400 450 500 Lab 1 Lab 2 Lab 5 Lab 6 Lab 7 Lab 8 Lab 9 Lab 10 Lab 14 ng/L Ofloxacin , Surface water Mean (211.5) Reference value 0 10 20 30 40 50 60 70 80 90 100 Lab 1 Lab 2 Lab 5 Lab 6 Lab 7 Lab 8 Lab 9 Lab 10 Lab 14 ng/L Ofloxacin , treated water Mean (45.2) Reference value
18 Figure 4: Results obtained for each participant for Erythromycin, Ofloxacin, Sulfamethoxazole, Trimethoprim and Ciprofloxacin expressed in ng/Lin the different samples, and mean values of results. These data can be represented by using Tuckey diagram representing the different characteristics of the values for each molecule in the two matrices (Figure 5) illustrating the distribution of the data. 0 20 40 60 80 100 120 140 Lab 1 Lab 2 Lab 4 Lab 5 Lab 6 Lab 7 Lab 8 Lab 9 Lab 10 Lab 11 Lab 12 Lab 13 Lab 14 ng/L Sulfamethoxazole, Surface Water Mean (87.1) Reference value 0 5 10 15 20 25 30 35 Lab 1 Lab 2 Lab 4 Lab 5 Lab 6 Lab 7 Lab 8 Lab 9 Lab 11 Lab 12 Lab 13 Lab 14 ng/L Sulfamethoxazole, Treated water Mean (20.0) Reference value 0 20 40 60 80 100 120 140 160 180 200 Lab 1 Lab 2 Lab 3 Lab 4 Lab 5 Lab 6 Lab 7 Lab 8 Lab 9 Lab 10 Lab 11 Lab 12 Lab 13 Lab 14 ng/L Trimethoprim, Surface Water Mean (105.8) Reference value 0 10 20 30 40 50 60 70 80 Lab 1 Lab 2 Lab 3 Lab 4 Lab 5 Lab 6 Lab 7 Lab 8 Lab 9 Lab 10 Lab 11 Lab 12 Lab 13 Lab 14 ng/L Trimethoprim, Treated Water Mean (39.9) Reference value 0 50 100 150 200 250 300 350 Lab 1 Lab 3 Lab 5 Lab 6 Lab 7 Lab 9 Lab 10 Lab 13 Lab 14 ng/L Ciprofloxacin, surface water Mean (138) Reference value
19 Figure 5 : Tuckey diagram for the repartition of the value for treated water and surface water 9.3. Coefficients of variation A comparison of the CV values (Figure 6) in both matrices (surface and treated water) for all laboratories resulted in a weak difference. The highest CV was observed for erythromycin in both matrices. It is particularly high in surface water (highest concentration). The lowest CV was observed for sulfamethoxazole and trimethoprim and is similar in the two matrices in spite of a difference in the spiked concentration. According to the heterogeneity of the methods employed we can consider the CV relatively low for the targets except for erythromycin Figure 6 : Coefficient of variation for all laboratories Several laboratories have used the same methodology in particular, same extraction cartridges (Oasis HLB) and chromatography techniques (HPLC/MS/MS). CV between such laboratories has been calculated (Figure 7). Except for ciprofloxacin in surface water, and trimethoprim in 0 20 40 60 80 100 120 140 160 ERY OFL SMX TMP ng/L Sample A (Treated water) 0 200 400 600 800 1000 1200 ERY CIP OFL SMX TMP ng/L Sample B (Surface water) 0 10 20 30 40 50 60 70 80 90 CIP ERY OFL SMX TMP Coefficient of Variation (%) Treated Water (A) Surface Water (B)
20 treated water, CV is lower than considering all laboratories, showing an influence of the analytical procedure in the representativity of the results. Figure 7: CV of laboratories using oasis HLB and HPLC/MS/MS as extraction cartridge and chromatography techniques respectively 9.4. Laboratory performances Intralaboratory repeatability has been evaluated for each substance in the two matrices and for each laboratory by the calculation of the coefficient of variation (Figure 8 and Figure 9). Figure 8: Intralaboratory repetability in sample A (tap water) 0 10 20 30 40 50 60 70 CIP ERY OFL SMX TMP Coefficient of Variation (%) Treated Water (A) Surface Water (B) 0 5 10 15 20 25 30 Lab 1 Lab 2 Lab 3 Lab 4 Lab 5 Lab 6 Lab 7 Lab 8 Lab 9 Lab 10 Lab 11 Lab 12 Lab 13 Lab 14 Coefficient of Variation (%) Treated water (sample A) ERY OFL SMX TMP
21 Figure 9: Intralaboratory repetability in sample B (surface water) Good intralaboratory reproducibility has been observed. CV varies from 0.5 to 25.5 % (median 6.6%) and to 0.8 to 19.8 % (median 5.6%) for sample A and B respectively. This result illustrates the important capability of the different laboratories to produce representative data. On the second hand, the difference between the produced values by each laboratory and the reference value was estimated by the calculation of the bias. By considering all values, the ISO 13528 classification of the laboratory biases resulted in 85% of acceptable values, 11% of value to be survey and 5% of not acceptable values Furthermore, the average performance of the laboratories can be divided up the following (Table 11). The detail of the calculation is reported in annex 5. - 43% of the laboratories have the whole result considered as acceptable - 36% of the laboratories have at minimum one value to be survey - 14% of the laboratories have at minimum one value not acceptable - 7 % of the laboratories have at minimum one value to be survey and one value not acceptable 0 5 10 15 20 25 Lab 1 Lab 2 Lab 3 Lab 4 Lab 5 Lab 6 Lab 7 Lab 8 Lab 9 Lab 10 Lab 11 Lab 12 Lab 13 Lab 14 Coefficient of variation (%) Surface water (sample B) ERY CIP OFL SMX TMP