use of light-supported oxidation processes towards microbiological and chemical contaminants

POUR L'OBTENTION DU GRADE DE DOCTEUR ÈS SCIENCES

acceptée sur proposition du jury:

Prof. C. Ludwig, président du juryProf. C. Pulgarin, Dr L. F. De Alencastro, directeurs de thèse

Prof. D. V. Vione, rapporteurDr P. Fernández-Ibáñez, rapporteuse

Prof. U. von Gunten, rapporteur

Use of light-supported oxidation processes towards microbiological and chemical contaminants elimination in

hospital wastewaters

THÈSE NO 7387 (2016)

ÉCOLE POLYTECHNIQUE FÉDÉRALE DE LAUSANNE

PRÉSENTÉE LE 2 DÉCEMBRE 2016

À LA FACULTÉ DES SCIENCES DE BASEGROUPE PULGARIN

PROGRAMME DOCTORAL EN GÉNIE CIVIL ET ENVIRONNEMENT

Suisse2016

PAR

Stefanos GIANNAKIS

3

“Success consists of going from failure to failure without loss of enthusiasm.”

- Winston Churchill

5

Acknowledgements With these words, I would like to thank everyone and everything that contributed to the successful

delivery of my Doctoral Thesis, and I would like to start with the Swiss National Foundation, for the funding

through the project “Treatment of the hospital wastewaters in Cote d’Ivoire and in Colombia by Advanced

Oxidation Processes”, during the last years.

First and foremost, this journey would have never even started if it was not for Prof. Cesar Pulgarin. I was

a young and foolish dreamer when I started working under his supervision, but with exemplary guidance,

tireless advice and time investment off his own well-being, I can safely say that I am now an old and foolish

dreamer, albeit with a purpose. I had the rare privilege to grow and learn with truly the best, and I hope

to never fail evolving, learning, getting better, as he would always want me to do. Thank you for breaking

all the stereotypes in the student-supervisor relation, and for being a true mentor in every aspect.

Also, I would like to thank my co-supervisor, Dr. Luiz Felippe de Alencastro, as his contribution was more

than crucial in the successful delivery of this Thesis. It takes a lot of patience and good will to work with

me, but he, along with Dominique Grandjean, surely went out of their way to help me complete my work.

My gratitude is also expressed to my Thesis examination Jury, Prof. Christian Ludwig, Dr. Pilar Fernández-

Ibáñez, Prof. Davide Vione and Prof. Urs von Gunten, for their comments, corrections, thoughts and the

very educative experience during the exam day.

All this work, would have never been completed without my trusted partners, my students, who tirelessly

and great zeal worked endless hours to complete their Master Projects: Miquel Pastor Gelabert, Simon

Schindelholz, David Muzard, Barbara Androulaki, Franco Alejandro Gamarra Vives, Margaux Voumard,

Idriss Hendaoui, Samuel Watts, Siting Liu and Christian Pinilla, and rest of the vibrant team members as

Ana Isabel Merino Gamo, Alba Camarasa, Marco Mangayayam. I sincerely wish you the best in your career

and your personal life. May you all find your path and look back at this time at EPFL with the same nostalgia

as I will do.

Special thanks to the rest of my co-authors for their constructive comments and contribution to

successfully publishing this work: Michael Bensimon, Jean-Marie Furbringer, Milica Jovic, Anna Carratala,

Natalia Gasilova, Hubert Girault, Anoys Magnet, Sana Thabet, Pascale Cotton, Maria Inmaculada Polo

Lopez, Jose Antonio Sanchez Perez and Pilar Valero.

A separate mention to the good friends I made through this work, my colleagues who were there seeing

me at my best and worst. Stefanos Papoutsakis and Sami Rtimi, thank you for your friendship and for the

countless times a short or extensive, scientific or not exchange took place among us, even if people around

us had to change a table during lunchtime. Sometimes a word is all people need. Laura Suarez and Paola

6

Villegas, my more-than-officemates, for the moral and scientific support and for being more than my

neighbors, but a balance and a perspective in the difficult times during this period.

A special shout-out to Lambros Alexandrou, Kostas Karalazos and Christopher Zolotas. Sons of Berlusconi,

Macedonian Phalanx, €urogroup, many names of the same face: true friends, 2000 km away. The value

of knowing that you always have someone to turn to: Priceless. Special thanks to my Ballets, in Greece

and abroad: Alexandra (Young and Tall), Olga, Giota, Anastasia, Vanessa and Kiki. Andrew and Panagioti,

many years have gone by, but every time we meet it is High School all over again, thank you guys!

Saving the best for last, I want to thank and apologize to my family, my parents Fotis and Elektra and my

brothers George and Vangelis. Thank them for being there, at the other end of the phone line, the other

side of a Skype call, in any way they could, they supported my every step. They never put their own

happiness before mine, even when I had to expatriate, and for that I feel the need to apologize, for

depriving me from them.

Finally, I would like to thank the EPFL Language Services for the fortunate arrangements during the

German A1 Course placement, and for introducing me to the person who was fated to change me and my

life forever. My inspiration, my muse, my source of strength and happiness, my truly everything. Thank

you for every single time you were there for me, with your words and your care, even throughout the

most difficult of times… Sofia, my anticipation to spend the rest of my life with you cannot be described

by words…

7

Abstract Hospital wastewaters have been long identified as carriers of chemical and microbiological pollutants.

Their amounts and risk levels have initiated numerous works on changing the existing practices of co-

treatment with municipal wastewaters and safe disposal in the environment. In this work, the issue of

hospital wastewater treatment is studied in two different contexts, in Switzerland and in developing

countries (Ivory Coast and Colombia). For this purpose, their treatment with municipal wastewater

effluents is recreated, simulating the developed countries’ context, while cheap and sustainable solutions

are proposed for the developing countries, to form a barrier between hospitals and receiving water

bodies. In both examples, the use of Advanced Oxidation Processes is implemented, focusing on UV-based

and solar-supported ones, in the respective target areas. A list of emerging contaminants and bacteria are

firstly studied to provide operational and engineering details on their removal by AOPs. Fundamental

mechanistic insights are provided as well on the degradation of the effluent wastewater organic matter.

The use of viruses and yeasts as potential model pathogens is also accounted for, treated by the photo-

Fenton process. Emphasis is given on the influence of the wastewater matrix parameters (organic matter,

pH, iron speciation etc.) and the exploration of the internal oxidative events, by the use of genomic and

proteomic analyses, respectively. Finally, two pharmaceutically active compound (PhAC) models of

hospital and/or industrial origin are studied in wastewater and urine, treated by all accounted AOPs, as a

proposed method to effectively control concentrated point-source pollution from industrial and hospital

wastewaters, respectively. Their elimination was modeled and the degradation pathway was elucidated

by the use of state-of-the-art analytical techniques (TOF-MS, Orbitrap). The use of light-supported AOPs

was proven to be effective in degrading the respective target and further insights were provided by each

application, which could facilitate their divulgation and potential application in the field.

Keywords: Advanced Oxidation Process, hospital wastewater, urine treatment, E. coli bacteria,

Saccharomyces cerevisiae yeast, MS2 Coliphage virus, emerging contaminants, UV/H2O2, photo-Fenton,

degradation pathway

9

Résumé Les eaux usées hospitaliers ont longtemps été identifiés comme porteurs des polluants chimiques et

microbiologiques. Leurs montants et les niveaux de risque ont entrepris de nombreux travaux pour

modifier les pratiques existantes de co-traitement avec les eaux usées municipales et l'élimination sûre

dans l'environnement. Dans ce travail, le sujet du traitement des effluents d’eaux usées hospitaliers est

étudié dans deux contextes différents, en Suisse et dans des pays en développement (Côte d'Ivoire et la

Colombie). À cette fin, leur traitement avec les eaux usées municipales est recréé, avec la simulation du

contexte des pays développés, alors que des solutions économiques et durables sont proposées pour les

pays en développement, afin de former une barrière entre les hôpitaux et les eaux réceptrices. Dans les

deux exemples, l’utilisation des Procédés d'Oxydation Avancée (POA) est effectuée, en se concentrant

sur les procédés basés sur la lumière Ultraviolette et ceux solaires, respectivement dans les zones cibles.

Une liste de contaminants émergents et des bactéries sont premièrement étudiés pour fournir des détails

opérationnels et technologiques, concernant leur élimination par les POA. En outre, des aperçus sur les

mécanismes fondamentaux sont fournis sur la dégradation des matières organiques des effluents des

eaux usées. L’utilisation des virus et des levures comme des modèles de pathogènes potentiels ont été

également considérés et traités par le procédé photo-Fenton. L'accent est mis sur l'influence des

paramètres de la matrice des eaux usées (matière organique, pH, spéciation de fer, etc.) et l'exploration

des événements oxydatifs internes, par l'utilisation respective des analyses génomiques et protéomiques.

Enfin, deux modèles de composés pharmaceutiquement actifs d’origine hospitalière et/ou industrielle

sont étudiés dans les eaux usées et dans l’urine, et ils sont traités par tous les représentants POA, comme

une méthodologie proposée pour contrôler la pollution dense et ponctuelle par les eaux usées

respectivement industrielles et hospitalières. Leur élimination a été modélisée et la voie de dégradation

a été élucidée par l'utilisation de techniques d'analyse de pointe (TOF-MS, Orbitrap). L’utilisation de POA

basés sur la lumière a été démontrée efficace pour la dégradation de la cible respective et des conceptions

supplémentaires ont été fournies par chaque application, lesquelles pourraient faciliter leur divulgation

et implémentation potentielle dans le domaine.

Mots-clés : Procédés d'Oxydation Avancée, eaux usées hospitaliers, traitement d’urine, bactéries E. coli,

levure Saccharomyces cerevisiae, virus MS2 Coliphage, contaminants émergents, UV/H2O2, photo-Fenton,

voie de dégradation

11

Contents

Acknowledgements ...................................................................................................................................... 5

Abstract ........................................................................................................................................................ 7

Résumé ......................................................................................................................................................... 9

Contents ..................................................................................................................................................... 11

List of Figures .............................................................................................................................................. 15

List of Tables ............................................................................................................................................... 18

Abbreviations list ........................................................................................................................................ 19

1. Chapter 1 – Introduction & state of the art ....................................................................................... 21

1.1. Basic characteristics of the targets ............................................................................................. 23

1.1.1. Categories/micropollutants classification .......................................................................... 24

1.1.2. Categories/microorganisms classification .......................................................................... 24

1.2. Problems related with microorganisms’ and micropollutants’ presence .................................. 27

1.3. Hospital WW and pollutants ...................................................................................................... 29

1.3.1. HWW and MPs ................................................................................................................... 29

1.3.2. HWW and MOs ................................................................................................................... 31

1.4. Occurrence, treatment and fate of MPs in MWWTPs ............................................................... 32

1.4.1. MP Occurrence in WWTPs ................................................................................................. 32

1.4.2. MP evolution path in WWTPs ............................................................................................ 32

1.4.3. Factors affecting MP removal in WWTPs ........................................................................... 33

1.5. AOPs action in chemical and microbiological pollutants’ degradation ...................................... 34

1.5.1. UV-based processes (UV, UV/H2O2) ................................................................................... 36

1.5.2. Fenton-related reactions (Fenton, photo-Fenton, solar light) ........................................... 41

1.6. Problem identification and contextualization: Micropollutants and microorganisms in developed and developing countries ..................................................................................................... 44

1.7. AOPs vs. Micropollutants and Microorganisms: current status ................................................. 46

1.8. Thesis aims and objectives ......................................................................................................... 48

2. Chapter 2: Effect of advanced oxidation processes on the micropollutants and the effluent organic matter contained in municipal wastewater previously treated by three different secondary methods .. 53

2.1. Introduction ................................................................................................................................ 54

2.2. Materials and Methods .............................................................................................................. 56

2.2.1. Sampling campaign ............................................................................................................. 56

2.2.2. Chemicals and reagents...................................................................................................... 56

2.2.3. Employed reactors .............................................................................................................. 57

12

2.2.4. Advanced Oxidation Processes specifics ............................................................................ 57

2.2.5. Physicochemical parameters .............................................................................................. 58

2.2.6. Analytical methods ............................................................................................................. 58

2.2.7. Secondary treatment systems specifications ..................................................................... 59

2.3. Results ........................................................................................................................................ 59

2.3.1. Initial conditions ................................................................................................................. 59

2.3.2. Efficacy of the various advanced oxidation processes ....................................................... 62

2.3.3. Degradation kinetics evaluation for the 6 different pollutants ......................................... 67

2.3.4. Evolution of the Average Oxidation State during the 5 different treatment processes .... 68

2.4. Discussion ................................................................................................................................... 71

2.4.1. Degradation of micropollutants in wastewater: characteristics, influence and role of the Effluent Organic Matter ..................................................................................................................... 71

2.4.2. Pathways of MP degradation in secondary effluent .......................................................... 72

2.5. Conclusions ................................................................................................................................. 76

3. Chapter 3 - Micropollutant degradation, bacterial inactivation and regrowth risk in wastewater effluents: influence of the secondary (pre)treatment on the efficiency of Advanced Oxidation Processes 79

3.1. Introduction ................................................................................................................................ 80

3.2. Materials and Methods .............................................................................................................. 81

3.2.1. Collection of wastewater samples and treatment plant specifications ............................. 81

3.2.2. Employed chemicals and reagents ..................................................................................... 82

3.2.3. Experimental set-up: reactors and apparatus .................................................................... 82

3.2.4. Application of AOPs: details and specifications ................................................................. 82

3.2.5. Analytical methods, physicochemical and microbiological parameters ............................ 83

3.3. Results ........................................................................................................................................ 84

3.3.1. Micropollutant elimination in the selected wastewater effluents .................................... 84

3.3.2. Microorganism elimination in the different wastewater effluents, per AOP: inactivation and post-treatment regrowth ............................................................................................................ 86

3.4. Discussion ................................................................................................................................... 91

3.4.1. The major threat and treatment focus: micropollutants or microorganisms? .................. 91

3.4.2. Common events and dissimilarities in the treatment of different targets in secondary effluents 93

3.5. Conclusions ................................................................................................................................. 97

4. Chapter 4 - Effect of Fe(II)/Fe(III) species, pH, irradiance and bacterial competition on viral inactivation in wastewater by the photo-Fenton process: Kinetic modeling and mechanistic interpretation. .......................................................................................................................................... 101

4.1. Introduction .............................................................................................................................. 102

4.2. Materials and methods ............................................................................................................ 103

13

4.2.1. Chemicals and reagents.................................................................................................... 103

4.2.2. Sunlight source and reactors ............................................................................................ 105

4.2.3. Microorganisms and quantification methods .................................................................. 105

4.2.4. Inactivation experiments .................................................................................................. 106

4.2.5. Data treatment and analysis ............................................................................................ 106

4.3. Results and Discussion ............................................................................................................. 107

4.3.1. Isolated effect of the photo-Fenton constituents ............................................................ 107

4.3.2. Parametrization of MS2 inactivation by the photo-Fenton process in wastewater ........ 109

4.3.3. Effect of bacterial competition on MS2 inactivation in wastewater ................................ 113

4.3.4. Iron cations solubility in wastewater ............................................................................... 115

4.3.5. MS2 inactivation modeling ............................................................................................... 116

4.3.6. Integrated proposal for the inactivation mechanism of viruses in wastewater .............. 118

4.4. Conclusions ............................................................................................................................... 120

5. Chapter 5 - Castles fall from inside: Evidence for dominant internal photo-catalytic mechanisms during treatment of Saccharomyces cerevisiae by photo-Fenton at near-neutral pH ............................ 121

5.1. Introduction .............................................................................................................................. 122

5.2. Materials and Methods ............................................................................................................ 124

5.2.1. Chemicals .......................................................................................................................... 124

5.2.2. Fe–citrate complex and Goethite preparation ................................................................. 124

5.2.3. Yeast strains and growth media ....................................................................................... 124

5.2.4. Photo-inactivation experiments ....................................................................................... 125

5.2.5. Cultivability assays ............................................................................................................ 125

5.2.6. Analytical methods ........................................................................................................... 126

5.2.7. Biochemical methods ....................................................................................................... 126

5.2.8. Experimental Planning ...................................................................................................... 127


5.3.1. Preliminary assays in simulated wastewater ................................................................... 127

5.3.2. Cultivability assays – Efficiency of treatment ................................................................... 128

5.3.3. Flow cytometry results – Localization of damage ............................................................ 134

5.3.4. Identification of targets – Nuclear DNA, cell wall and cytoplasmic protein damage ....... 141

5.3.5. Holistic proposal for the inactivation mechanism of S. cerevisiae ................................... 144

5.4. Conclusions ............................................................................................................................... 146

6. Chapter 6 - Iohexol degradation in wastewater and urine by UV-based Advanced Oxidation Processes (AOPs): Process modeling and by-products identification. ..................................................... 151

6.1. Introduction .............................................................................................................................. 152

6.2. Materials and Methods ............................................................................................................ 153

6.2.1. Chemicals and Reagents ................................................................................................... 153

14

6.2.2. Reactors and experimental apparatus ............................................................................. 154


6.2.4. Water matrices and treatment conditions ....................................................................... 156

6.2.5. Statistics, modeling and data treatment .......................................................................... 157


6.3.1. Engineering approach – investigation on the operational parameters ........................... 158

6.3.2. Statistical approach – modeling and mathematical optimization of the treatment ........ 163

6.3.3. Analytical approach – Global measurements (COD, TOC, and UV-vis absorbance) combined with specific HPLC and MS analysis ................................................................................. 170

6.4. Conclusions ............................................................................................................................... 175

7. Chapter 7 - Solar photo-Fenton and UV/H2O2 processes against the antidepressant Venlafaxine in urban wastewaters and human urine. Intermediates formation and biodegradability assessment. ..... 177

7.1. Introduction .............................................................................................................................. 178

7.2. Materials and methods ............................................................................................................ 179

7.2.1. Chemicals and reagents.................................................................................................... 179

7.2.2. Water, wastewater and urine matrices............................................................................ 180

7.2.3. Light sources and corresponding reactors-experimental apparatus ............................... 181



7.3.1. UV-based AOPs degradation of Venlafaxine .................................................................... 182

7.3.2. Fenton-related AOPs degradation of Venlafaxine ........................................................... 187

7.3.3. Venlafaxine degradation experiments in wastewater and urine ..................................... 192

7.3.4. Elucidation of the AOP-driven degradation pathway and inherent biodegradability properties of Venlafaxine ................................................................................................................. 197

7.4. Conclusions ............................................................................................................................... 200

8. Chapter 8 - General conclusions, perspectives and future work ..................................................... 201

9. References ........................................................................................................................................ 205

Appendix A: Supplementary material of Chapter 2 ................................................................................. 222

Appendix B: Supplementary material of Chapter 3 ................................................................................. 230

Appendix C: Supplementary material of Chapter 4 ................................................................................. 235

Appendix D: Supplementary material of Chapter 5 ................................................................................. 237

Appendix E: Supplementary material of Chapter 6 .................................................................................. 240

Appendix F: Supplementary material of Chapter 7 .................................................................................. 261

Curriculum Vitae of the Candidate ........................................................................................................... 273

15

List of Figures Figure 1.1 – E. coli structure (adaptation: Source: http://www.bevpease.force9.co.uk/p.Dawn-of-Life_files/image004.gif) ............................................................................................................................ 25 Figure 1.2 – Infection mechanism of the bacteriophage MS2 (Source file: http://faculty.washington.edu/jclara/301/images/ssRNA.jpg). Inset: MS2 structure (Credits: Stephan Spencer, http://www.virology.wisc.edu/virusworld/imgency/ms2MS22.jpeg) .................................... 26 Figure 1.3 – Saccharomyces cerevisiae (yeast model) structure (Thabet et al. 2013)............................. 27 Figure 1.4 – Bioaccumulation of MPs (http://toxics.usgs.gov/regional). ............................................... 29 Figure 1.5 – Contribution of HWW in pollutants integration to the environment (http://www.frontiersin.org/files). .......................................................................................................... 30 Figure 1.6 – Categorization of Advanced Oxidation Process (Poyatos et al. 2010)................................. 36 Figure 1.7 – Light absorption of Fe3+ species at normal solar irradiance (I) on the Earth’s surface ....... 41 Figure 1.8 – Routes of pharmaceutical contamination of the aquatic environment (Ikehata et al. 2006). .................................................................................................................................................................... 45 Figure 1.9 – Thesis graphical representation and organization of aims and objectives. ........................ 48 Figure 2.1 – Simplified overview of the Vidy WWTP and the sampling points used in the study. ......... 56 Figure 2.2 – UV treatment results. a) % degradation vs. time b) % COD & TOC reduction vs. time. ...... 62 Figure 2.3 – UV/H2O2 treatment results. a) % degradation vs. time. b) % COD & TOC reduction vs. time. .................................................................................................................................................................... 63 Figure 2.4 – Solar exposure results. a) % degradation vs. time b) % COD & TOC reduction vs. time. .... 64 Figure 2.5 – Fenton process results. a) % degradation vs. time. b) % COD & TOC reduction vs. time. ... 65 Figure 2.6 – photo-Fenton process results. a) % degradation vs. time. b) % COD & TOC reduction vs. time. ........................................................................................................................................................... 67 Figure 2.7 – Overall mechanistic interpretation for the action of UVC and solar light within the effluent wastewater (adapted from (De la Cruz et al. 2012)). ............................................................................... 76 Figure 3.1 – Schematic representation of the WWTP of Vidy, Lausanne (VD, Switzerland) and the sampling points for this research. ............................................................................................................. 81 Figure 3.2 – Micropollutants’ degradation by AOPs after secondary treatment. a) UV and UV/H2O2 processes. b) Fenton, solar and photo-Fenton process. AS: blue trace, MBBR: red trace, CF: green trace. Continuous lines and colored symbols show the measured evolution of the experiment, while the dashed lines and open symbols indicate the projection of the experiment according to the measured first order degradation rate constant. ...................................................................................................... 85 Figure 3.3 – UV-based disinfection and respective regrowth after 24 h. A) UVC irradiation alone. B) UV/H2O2 process (20 ppm initial H2O2 addition). The shaded part and the dashed lines symbolize the dark storage and regrowth after treatment, for 24 h. ............................................................................. 87 Figure 3.4 – Fenton-related disinfection and respective regrowth after 24 h. A) Fenton process (2:10 ppm initial Fe2+/H2O2 addition). B) bare solar light. C) photo-Fenton process (2:10 10 ppm initial Fe2+/H2O2 addition). The shaded part and the dashed lines symbolize the dark storage and regrowth4 after treatment, for 24 h. .......................................................................................................................... 89 Figure 3.5 – UV-based disinfection and decontamination. A) UVC irradiation alone. B) UVC/H2O2 process. The lines indicate the microorganism inactivation, while the bars the micropollutant degradation (%). The circles indicate the regrowth suppression points with the respective colors indicating the secondary treatment method, while the horizontal lines indicate the minimal micropollutant (brown line) and microorganism removal (orange line). ................................................ 91 Figure 3.6 – Fenton-related disinfection and respective regrowth after 24 h. A) Fenton process. B) Bare solar light. C) photo-Fenton process. The lines indicate the microorganism inactivation, while the bars the micropollutant degradation (%). The circles indicate the regrowth suppression points with the

16

respective colors indicating the secondary treatment method, while the horizontal lines indicate the minimal micropollutant (brown line) and microorganism removal (orange line). ................................. 93 Figure 4.1 – Solar/H2O2 and Solar/Fe control experiments. a) Isolated effect of the operating H2O2 levels of this work. b) Addition of 0.5 or 1 mg/L Fe(II) or Fe(III) salts. DL: detection limit. ................... 108 Figure 4.2 – Effect of solar irradiance on the evolution of the photo-Fenton reaction. A) Fe(II) as starting iron species. B) Fe(III) as starting iron species. A notable difference exists in the kinetic families of Fe(II) or Fe(III). DL: detection limit. ....................................................................................... 109 Figure 4.3 – Effect of the Fe:H2O2 ratio on the evolution of the photo-Fenton process. a) Fe(II) as starting species. b) Fe(III) as starting species. ........................................................................................ 110 Figure 4.4 – Effect of the starting pH on the evolution of the photo-Fenton process. a) Fe(II) as starting species. b) Fe(III) as starting species. ...................................................................................................... 113 Figure 4.5 – Bacterial competition tests: Inactivation of MS2 and E. coli by the photo-Fenton process. a) Bacterial inactivation with increasing Fe:H2O2 ratios, in absence or presence of MS2. b) E. coli inactivation in presence of MS2 and MS2 inactivation in presence or absence of the bacterial host (Fe:H2O2 ratio 1:1). Higher Fenton reagents addition than 1:1 resulted in <2-min inactivation. ......... 114 Figure 4.6 – Iron evolution during (dark) Fenton or photo-Fenton process followed by ICP-MS analysis. A) Fe(II) starting salts. B) Fe(III) starting salts. The dashed lines indicate the dark Fenton experiments, closed trace symbols indicate dissolved iron and open trace symbols the total iron. .......................... 116 Figure 4.7 – Proposed MS2 inactivation pathway by the photo-Fenton process in wastewater at near-neutral pH. The events 1-6 are further analyzed in the text. ................................................................. 119 Figure 5.1 – Overview of the photo-Fenton tests in simulated wastewater. ........................................ 128 Figure 5.2 – Overview of the photocatalytic inactivation tests and their respective controls. a) The plots describe the cultivability evolution over time. b) Comparison between pH 5.5 and 7.5 for the FeSO4-assisted photo-Fenton system. c) Comparison between pH 6.0 and 7.5 for the iron citrate-assisted photo-Fenton system. Standard deviation < 5%. ..................................................................... 129 Figure 5.3 – Control tests and an indicative presentation of the flow cytometry results evolution, during photo-Fenton reaction, at pH = 5.5. ............................................................................................ 135 Figure 5.4 – Flow cytometry results. Control tests: a) Simulated solar light only. b) hv/H2O2 system. FeSO4–assisted photo–Fenton processes: c) pH = 5.5. d) pH = 7.5. Fe-cit–assisted photo–Fenton processes: e) pH =6.0. f) pH = 7.5. Standard deviation < 5%. ................................................................. 140 Figure 5.5 – Nuclear DNA damage in the four different systems. Comparison of the pH effect in FeSO4-assisted photo-Fenton systems. .............................................................................................................. 141 Figure 5.6 – Cell wall (a) and cytoplasmic proteins damage (b) in the four different systems. (i-ii): Comparison of the pH effect in FeSO4-assisted photo-Fenton systems. ............................................... 142 Figure 5.7 – Mechanistic proposition of the pathways towards yeast cell inactivation. a) (direct) Simulated solar light. b) (Indirect) hv/H2O2. c) FeSO4–assisted photo–Fenton process. d) Fe-cit–assisted photo–Fenton process. ............................................................................................................................ 145 Figure 6.1 – UV photolysis and UV/H2O2 experiments in Mili-Q water. Note that the results in the 10-1000 mg/L range are plotted in double-logarithmic scale and axis breaks for clarity purposes only. 159 Figure 6.2 – Effect of pH, dilution and Iohexol, H2O2 and Fe2+ amounts. Dotted lines represent the undiluted matrices, continuous lines indicate the x10 times dilution experiments, and for Figure 3b, the x100 times diluted UR experiments are signified with long dashed lines. Note the mixed axes scales. .................................................................................................................................................................. 161 Figure 6.3 – Real wastewater and urine experiments: UV/H2O2/Fe2+ process. A) Iohexol in untreated or biologically treated WW, and B) diluted/undiluted urine, H2O2 added in 0, 10 or 50 ppm, iron was added in 0, 1, or 5 ppm, and changing of the initial pH value (3, 5 or near-neutral). The two main groups of Figure 3a data are separated by continuous (10 ppm Iohexol) or dashed lines (100 ppm). The

17

respective groups in Figure 3b are designated by color. The vertical bars show the variation in efficiency when pH was changed. Note the mixed axes scales. ............................................................. 162 Figure 6.4 – HPLC peak areas evolution during Iohexol degradation by the UV photolytic and photocatalytic process. A) UV only, B) 10 ppm H2O2, C) 100 ppm H2O2, D) 1000 ppm H2O2. 100 ppm of Iohexol was chosen as initial spiking. ..................................................................................................... 171 Figure 6.5 – Iohexol elimination by the UV-based AOPs. Iohexol degradation was followed by HPLC (blue trace), COD (red trace) and TOC decrease (green trace) during the following treatment methods: UV photolysis (trace: ), UV/H2O2 process (50 ppm H2O2, trace: ), and UV/H2O2/Fe2+ process (5 ppm Fe2+, 50 ppm H2O2, trace: ). H2O2 reduction: brown traces. A system employing 35-W UV-C lamps (instead of the 11-W ones of the previous parts, but otherwise identical) was used here. 100 ppm Iohexol was chosen as initial spiking. ..................................................................................................... 172 Figure 6.6 – Overall mechanistic degradation pathway of Iohexol treated by UV-based AOPs. Products common for all three treatments were marked with P, UV marked with A, for UV/H2O2 with B and for UV/H2O2/Fe2+ with C. Products common for A and C treatment were marked as AC, and accordingly, products common for B and C treatment were marked as BC. .............................................................. 174 Figure 7.1 – Summary of the UV-C photolysis experiments. a) UV-induced degradation of Venlafaxine followed by HPLC, COD removal and TOC reduction during UV photolysis. b) Evolution of the COD/TOC ratio. ......................................................................................................................................................... 183 Figure 7.2 – UV/H2O2 Advanced Oxidation of Venlafaxine: degradation and process optimization. a) Degradation of VFA by UV/H2O2 with addition of 5-50 mg/L H2O2. b) Evolution of COD/TOC ratio (for 50 mg/L initial H2O2 addition). c) Consumption of H2O2 (black axis and traces) and changes in the t90% (blue axis and traces) as a function of initial H2O2 amounts. ................................................................. 185 Figure 7.3 – Treatment of Venlafaxine by the Fenton process in the dark. a) Evolution of the t90% with increasing Fe|H2O2 ratio and pH. b) COD/TOC ratio modification by the Fenton process at various Fe|H2O2 ratios (pH=3) and increasing pH for a 20|50 (mg/L|mg/L) ratio. ........................................... 188 Figure 7.4 – Absorbance spectra during the 24-h Fenton treatment of Venlafaxine, for various Fe|H2O2 ratios (panels a-d) and pH levels (panels i-iii). Fe|H2O2 ratios: a) 5|10, pH=3. b) 5|50, pH=3. c) 12.5|30, pH=3. d) 20|50, pH=3. pH levels: i) 20|50, pH=3. ii) 20|50, pH=5. iii) 20|50, pH=7. ............................. 190 Figure 7.5 – Treatment of Venlafaxine by the photo-Fenton process (irradiance: 900 W/m2). a) Evolution of the t90% with increasing Fe|H2O2 ratio and pH. b) COD/TOC ratio evolution by the solar photo-Fenton process at various Fe|H2O2 ratios (pH=3) and increasing pH for a 20|50 ratio. ............ 191 Figure 7.6 – Absorbance spectra during the 3-h photo-Fenton treatment of Venlafaxine, for various Fe|H2O2 ratios (panels a-d) and pH levels (panels i-iii). Fe|H2O2 ratios: a) 5|10, pH=3. b) 5|50, pH=3. c) 12.5|30, pH=3. d) 20|50, pH=3. pH levels: i) 20|50, pH=3. ii) 20|50, pH=5. iii) 20|50, pH=7. .............. 192 Figure 7.7 – Treatment of Venlafaxine by AOPs in urban WW effluents. The experimental conditions are marked in the corresponding graphs. a) VFA degradation by UV-based AOPs in AS, MBBR and CF effluents. b) VFA degradation by the Fenton-related processes in AS, MBBR and CF effluents. .......... 194 Figure 7.8 – Treatment of Venlafaxine by UV-based methods in human urine. a) VFA degradation by UV-based AOPs (0, 50 or 100 mg/L H2O2 and 0/100% or 10%-90% urine/water ratio. b) COD reduction and DOC (0.45μm filtration) removal in the same conditions. .............................................................. 196 Figure 7.9 – Combined Venlafaxine degradation pathway through the application of the treatment methods analyzed. ................................................................................................................................... 198 Figure 7.10 – Zahn-Wellens inherent biodegradability test of Venlafaxine and treated solutions in MQ. a) ZW test after treatment of 50% of the initial VFA solution. b) ZW test after treatment of 100% of the initial VFA solution. Note that results are normalized towards the initial DOC to enable comparison. .................................................................................................................................................................. 199

18

List of Tables Table 1.1 – Examples of fecal-related viruses contained in the feces and their associated diseases (Carter 2005, Glass et al. 2009, Mayer et al. 2008, Okoh et al. 2010, Rodriguez et al. 2014) ................. 26 Table 1.2 – Comparison between indicative UWW and HWW effluents characteristics (Carraro et al. 2016, El-Ogri et al. 2016, Verlicchi et al. 2010) and references therein). ................................................ 31 Table 1.3 – Classification of commonly studied micropollutants (based on removal efficiency). .......... 33 Table 2.1 – Evolution of indicator pollutants in the Swiss legislation. .................................................... 55 Table 2.2 – Summary of the advanced treatment methods applied. ...................................................... 58 Table 2.3 – Initial physicochemical characteristics of the effluents of the different secondary treatment units. ........................................................................................................................................................... 60 Table 2.4 – Initial micropollutant concentration and limit of quantification (LOQ) for micropollutants in the effluent of the different secondary treatment units. ......................................................................... 61 Table 2.5 – Degradation kinetics of the 6 different pollutants during treatment in the different effluents and treatment methods. ............................................................................................................ 68 Table 2.6 – Evolution of the Average Oxidation State (AOS) during treatment by the various methods in the different effluents. ........................................................................................................................... 69 Table 2.7 – AOS percentile change and correlation with % of degradation. ........................................... 70 Table 3.1 – Basic physicochemical and optical characteristics of the wastewater used in this study (own measurements and (aMargot et al. 2013, bMargot et al. 2011)). ................................................... 83 Table 3.2 – Photochemical characteristics of the various effluents ......................................................... 95 Table 4.1 – Composition of synthetic secondary wastewater (Muthukumaran et al. 2011). ............... 104 Table 4.2 – Effect of photo-Fenton treatment [Fe(II)] on MS2 inactivation. ......................................... 117 Table 4.3 – Effect of photo-Fenton treatment [Fe(III)] on MS2 inactivation. ........................................ 117 Table 5.1 - Timeline of the inflicted damage in the corresponding targets of the different iron-assisted systems ..................................................................................................................................................... 144 Table 6.1 – Synthetic matrices composition. .......................................................................................... 153 Table 6.2 – Physicochemical characteristics of the real wastewater matrices (Giannakis et al. 2015c, Margot et al. 2013, Margot et al. 2011). ................................................................................................ 156 Table 6.3 – Physicochemical characteristics of real urine matrices (own measurements and (Beach 1971)). ...................................................................................................................................................... 156 Table 6.4 – t90% evolution (min) in varied Iohexol (10-1000 ppm) and H2O2 (0-1000) levels ................. 159 Table 6.5 – Wastewater models with S and R2 values. .......................................................................... 165 Table 6.6 – Urine models with S and R2 values ....................................................................................... 167 Table 6.7 – Optimal regions for treatment Iohexol through optimization by the desirability function. .................................................................................................................................................................. 169 Table 7.1 – Venlafaxine characteristics and physicochemical properties (USNLM 2016b). .................. 179 Table 7.2 – Composition of the synthetic matrices used in this study. .................................................. 180 Table 7.3 – Measured pseudo-first order degradation kinetics of Venlafaxine per AOP and matrix. . 185 Table 7.4 – Occurrence and fate of Venlafaxine in urban WW effluents. ............................................. 193

19

Abbreviations list AOPs - Advanced Oxidation Processes

AOS - Average Oxidation State

AS - Activated Sludge

BCF – Bio-concentration Factor

CAT - Catalase

CBS - Carbonate Buffer Solution

CDOM - Chromophoric Dissolved Organic Matter

CF - Coagulation-Flocculation

CFDA - 5-carboxyfluorescein di-acetate

CFU - Colony Forming Units

COD - Chemical Oxygen Demand

CPDs - Cyclobutane Pyrimidine Dimers

D - (Composite) Desirability

DBP - Disinfection By-Product

DL - Detection Limit

DNA - Deoxyribonucleic Acid

DOC - Dissolved Organic Carbon

DOM - Dissolved Organic Matter

EfOM - Effluent Organic Matter

EPS - Extracellular Polymeric Substances

Fe/S - Iron-Sulfur

FOEN - Federal Office for the Environment

HRT - Hydraulic Retention Time

HWW - Hospital Wastewater

hv - Light

ICP-MS - Inductively coupled plasma mass spectrometry

LMCT - Ligand-to-Metal Charge Transfer

MBBR - Moving Bed Bioreactor

MO - Microorganism

MP - Micropollutant

MQ - MiliQ water

MS - Mass Spectrometry

MW - Molecular Weight

MWW - Municipal Wastewater

NOM - Natural Organic Matter

OxOM - Oxidizable Organic Matter

PFU - Plaque Forming Units

PhOM - Photo-sensitizable Organic Matter

pI - Isoelectric point

PI - Propidium Iodide

POM - Particulate Organic Matter

PP - PhotoProduct

ROS - Reactive Oxygen Species

RU - Real Urine

S - Synergy

SDS-PAGE - Polyacrylamide gel electrophoresis

SMP - Soluble Microbial Products

SOD - Superoxide Dismutase

SODIS - Solar Disinfection

SPE - Solid Phase Extraction

SRT - Sludge Retention Time

SS - Suspended Solids

SSNRIs - selective serotonin and norepinephrine reuptake inhibitors

SUR - Synthetic Urine

SUVA - Specific UV Absorbance

SWW - Synthetic Wastewater

T - Transmittance

TDS - Total Dissolved Solids

TKN - Total Kjeldahl Nitrogen

TOC - Total Organic Carbon

20

TSS - Total Suspended Solids

UWW - Urban Wastewater

UV - Ultraviolet

VFA - Venlafaxine

WW - Wastewater

WWTP - Wastewater Treatment Plant

ZW - Zahn-Wellens

21

1. Chapter 1 – Introduction & state of the art

23

Introduction & state of the art

Currently, one of the environmental concerns in global scale is the presence and accumulation of

micropollutants in the natural environment. These substances are comprising an increasing list of

anthropogenic (or not) contaminants, which include among others, pharmaceuticals, personal care

products, steroid hormones, industrial chemicals, pesticides and many other emerging compounds (Luo

et al. 2014). The majority of these substances are designed to be biologically active, and therefore, their

occurrence can affect the receiving environment, even at low concentrations (Santos et al. 2010); this

feature characterized them as “micropollutants”. This exact characteristic, combined with the diversity of

the chemical pollutants, re-instate this topic as high priority and challenging for the treatment facilities.

Wastewater treatment plants have been built, transformed and updated through the years to effectively

prevent solids, organic and inorganic compounds (carbon, nitrogen, phosphorus etc.) and more, that

enter the environment. The challenge posed by the pollutants is a matter that the majority of the WWTPs

are not equipped to handle. The micropollutants are (in a high percentage) invulnerable to biological

treatment; the transfer from source to the environment is therefore facilitated, leading to further

accumulation in the environment. As it will be analyzed later on, their presence has been associated with

minor and major health risks, toxicity and more (Fent et al. 2006, Luo et al. 2014).

The risk of microorganisms’ presence in natural water bodies is more explored compared to

micropollutants. Water scarcity has led to reuse concepts and many countries worldwide have included

legislation concerning the removal thresholds according to the subsequent water reuse (e.g. Italy, (Liberti

et al. 2003)). However, chlorination is still the most widespread technique, with its inherent problems,

such as DBPs (trihalomethanes THM) formation, and where funds are available UV has been applied. In

order to better suggest treatment goals and methods, the present thesis will address the use of Advanced

Oxidation Processes (AOPs) as a greener and sustainable disinfection and decontamination method of

hospital and urban wastewater. The solutions will be studied and provided according to the context of

application (developed or developing country) and the target of reference (microorganisms or

micropollutants). Finally, mechanistic insights will be given to further strengthen the knowledge base and

the know-how on AOPs used in wastewater treatment.

1.1. Basic characteristics of the targets

Before analyzing in detail the problems related with the presence of micropollutants and microorganisms,

the basic categories are hereby mentioned.

24

1.1.1. Categories/micropollutants classification

Concerning the categories of chemical pollutants, among the list of contaminants of emerging concern,

we find:

Pharmaceuticals: drug-related compounds.

Personal care products: Fragrances, disinfectants, UV filters, and insect repellents.

Steroid hormones: Estrogenic or androgenic effects

Surfactants: Detergents, wetting agents, emulsifiers, foaming agents, and dispersants.

Industrial chemicals: Plasticizers, flame retardants.

Biocides: Pesticides - insecticides, herbicides and fungicides.

The present work focuses, but is not limited, on the pharmaceutical pollutants, also known as

Pharmaceutically Active Compounds (PhACs). These are considered special micropollutants (Fatta-

Kassinos et al. 2011), due to the complexity of their structure, the polymorphism and the variation due to

the precedent metabolism in the human body (Cunningham 2004).

The PhACs have been long identified as harmful agents against the occurring flora and fauna in the natural

environment. Reports (Halling-Sørensen et al. 1998, Kümmerer 2001, Kümmerer 2004, Sumpter 1998)

link their presence with a wide array of effects on microorganisms, plants, fish, and possibly to humans.

The classes of PhACs that potentially pose a threat are (Ikehata et al. 2006):

Cytostatic agents, immunosuppressive drug and genotoxic drugs: this class has direct impact

(cytotoxic, carcinogenic, embryotoxic, etc.) on the receptors

Antibiotics: Their accumulation can lead to antibiotic-resistant bacteria

Hormones: Their action is profound in low levels, often leading to endocrine disruption and

feminization of male species.

Iodinated contrast media (ICM): Non-biodegradable, and therefore highly persistent in the

environment

Metals: Drugs that contain heavy metals pose toxicity risks.

Other micropollutants that are investigated in this thesis are pesticides, corrosion inhibitors,

antidepressants and beta-blockers, found in urban effluents.

1.1.2. Categories/microorganisms classification

Concerning the representatives of microbiological risk, the main categories of microorganisms in

wastewater are represented by:

Bacteria: Fecal indicator bacteria, wild type pathogenic strains etc.

Viruses: e.g. Coliphages, adenoviruses, rotaviruses etc.

25

Yeasts: e.g. Candida albicans

Protozoa: e.g. Cryptosporidium parvum (cysts and oocysts)

Parasites: e.g. Helminth eggs

Prions

This thesis will deal only with the first three categories of microorganisms, which represent the largest

fraction of pathogens in wastewater. Therefore, some general and structural details will be provided.

The bacterial presence is associated with either enteric or non-enteric source. E. coli has been for over a

century considered an indicator of enteric origin bacteria, and has been used as such ever since, due to

its relative ease to cultivate and maintain in a laboratory, its adequate survival and proper representation

of various water matrices (Edberg et al. 2000, Lin and Ganesh 2013). Also, the wild types have been

associated with various diseases themselves (Nataro and Kaper 1998) and the use of the lab-strain allows

safe recreation of the desired conditions.

Figure 1.1 – E. coli structure (adaptation: Source: http://www.bevpease.force9.co.uk/p.Dawn-of-

Life_files/image004.gif)

Viruses on the other hand differ significantly in structure complexity and morphology. They are also single-

cell microorganisms, but are composed simply by a capsid and the genome. As an obligate parasite, they

require the presence of a host in order to replicate and infect their target (Figure 1.2). Their presence in

wastewaters has been verified and some indicative ones (and their effect) are presented in Table 1.1. The

systematic investigations necessary led to the use of surrogate strains, such as MS2 coliphage, presenting

similar advantages with the lab E. coli strains.

26

Table 1.1 – Examples of fecal-related viruses contained in the feces and their associated diseases

(Carter 2005, Glass et al. 2009, Mayer et al. 2008, Okoh et al. 2010, Rodriguez et al. 2014)

Type (Family) Associated disease

Rotavirus (Reoviridae) Diarrhea in infants

Enterovirus (Picornavidae) Meningitis, severe conjunctivitis, poliomyelitis

Adenovirus (Adenoviridae) Gastro-enteritis, pneumonia

Norovirus (Calciviridae) Gastro-enteritis

Figure 1.2 – Infection mechanism of the bacteriophage MS2 (Source file:

http://faculty.washington.edu/jclara/301/images/ssRNA.jpg). Inset: MS2 structure (Credits: Stephan

Spencer, http://www.virology.wisc.edu/virusworld/imgency/ms2MS22.jpeg)

27

The fungal kingdom is historically the less intensively studied among the ones in this Thesis, but these

microorganisms present characteristics which vary significantly from bacteria or viruses. Yeasts are

eukaryotic microorganisms, which are created with more complex cell walls, more advanced life functions

and suspected higher resistance to treatment (Temple et al. 2005). Although their expected number in

wastewater is low, these traits state them as an interesting target. Furthermore, their presence has been

linked with serious chronic (such as candidiasis) or opportunistic diseases (Thabet et al. 2014). Some

representatives are the Candida species, the Fusarium and the Aspergillus.

Figure 1.3 – Saccharomyces cerevisiae (yeast model) structure (Thabet et al. 2013)

1.2. Problems related with microorganisms’ and micropollutants’

presence

The problem of microorganisms’ occurrence in water sources has been identified for many decades.

Water related diseases plague developing countries using compromised drinking water sources

(McGuigan et al. 2012), or reusing wastewater for food production (Qadir et al. 2010), but also many

outbreaks have occurred by cross contamination of public access waters in the developed states (Hoebe

28

et al. 2004). Apart from the notorious diseases such as cholera, gastro-enteritis or dengue fever, the

contamination of water sources poses risks to populations which have incorporated their use in their

economical-related activities, such as fishing or other artisanal activities.

The main problem associated with the occurrence of micropollutants (MPs) in the environment is the lack

of information, concerning the side-effects of their presence in the receiving matrix (Deblonde et al. 2011).

For instance, little is known on the long term effects of pollutants, classified as “potentially not harmful”;

there are no studies on the accumulation or the chronic toxicity of such substances neither in plants,

animals, nor in humans. There is a relatively long list of MPs actually found in the environment, which

derived from the various anthropogenic activities (Focazio et al. 2008, Schwarzenbach et al. 2006). To

date, not all substances have been assessed on their potential actions against the environment.

Furthermore, toxicologically speaking, most of the substances that have been assessed for their risks,

have been directed through single-compound investigations, when the reality differs significantly

(Gregorio and Chèvre 2014). The environment contains a mixture of a vast number of compounds, which

could even react with each other, affecting their mode of action (Backhaus et al. 2003). The result has

been demonstrated to be severe, with cases reporting compounds that had no prior effect on species, to

demonstrate harmful properties when found in mixtures (Deneer 2000, Junghans et al. 2006, Rodney et

al. 2013).

Apart from the toxicity problems, other critical PhAC-related issues that have emerged over the years are:

i) The enhancement of antibiotic resistance, by the presence of antibiotics and their

metabolites in the environment (Rizzo et al. 2013),

ii) The problematic identification of the transformed metabolites of drugs (Fatta-Kassinos et al.

2011),

iii) The bioaccumulation of pollutants in living organisms in the environment (Figure 1.4).

29

Figure 1.4 – Bioaccumulation of MPs (http://toxics.usgs.gov/regional).

1.3. Hospital WW and pollutants

1.3.1. HWW and MPs

Hospital wastewater (HWW) is the result of the residue collection from the various water-consuming

activities taking place within its premises. These services include (Verlicchi et al. 2012) (Figure 1.5):

i) Human sewage

ii) Kitchen and laundry

iii) Heating and cooling processes’

iv) Laboratorial discharge (clinics, research centers)

v) Wards and outpatients contribution

The first categories are common also in MWW, which is the reason that led practitioners to suggest co-

treatment in MWWTPs, sometimes with only a pre-treatment (e.g. chlorination) (Emmanuel et al. 2004),

only to limit the microbiological-related risk. The reality suggests that within HWW there are many

substances, such as disinfectants, organic compounds, therapeutic metals, rare microbial agents or

antibiotic-resistant ones (Boillot et al. 2008, Emmanuel et al. 2005, Hawkshead III 2008), often in high

concentrations that modify significantly the composition of HWW compared to MWW. For this reason,

30

many authors have openly objected to the co-treatment practice so far (Altin et al. 2003, Pauwels and

Verstraete 2006, Vieno et al. 2007).

The presence of PhACs in hospital wastewater is a hot topic, with a variety of works dedicated in the

characterization of their nature (Kosma et al. 2010, Kümmerer 2001, Mahnik et al. 2007, Suarez et al.

2009, Verlicchi et al. 2012) and their importance in the overall load (Fatta-Kassinos et al. 2011). The

characteristics of HWW are influenced by a variety of factors, such as the size of the hospital, the range

of services and activities, the season of the year, the time of the day (Verlicchi et al. 2012, Verlicchi et al.

2010), and more. Table 1.2 presents an overview comparing HWW with MWW, in terms of a single unit

(patient/inhabitant).

Figure 1.5 – Contribution of HWW in pollutants integration to the environment

(http://www.frontiersin.org/files).

It is obvious that the composition of either micro or macropollutants in each case is significantly different

(see Table 1.2), indicating one of the reasons for failure of treatment by conventional WWTPs. The

micropollutants arriving in WWTPs are of a range of μg or ng, and also are reported to affect the nature

of the WW in the treatment plant (solubility, volatility, adsorbability, absorbability, biodegradability,

polarity and stability) (Verlicchi et al. 2010).

31

Table 1.2 – Comparison between indicative UWW and HWW effluents characteristics (Carraro et al.

2016, El-Ogri et al. 2016, Verlicchi et al. 2010) and references therein).

UWW HWW

BOD 60 160 mg/L COD 110 280 mg/L SS 80 135 mg/L

Bacteria Total Coliforms 106 7.7*109 MOs/L Viruses Norovirus 1.6*102 2.4*106 MOs/L

Hepatitis A virus 102 104 MOs/L Adenovirus 1.6*102 2.8*106 MOs/L

For the pre-mentioned reasons, HWW should be treated as a separate entity (Verlicchi et al. 2010). The

economic and overall risks should be assessed (Pauwels and Verstraete 2006), on-site treatment should

be implemented as close as possible to the source (H. Jones et al. 2005, Ikehata et al. 2006), the

consideration of no-mix toilets for urine separation must be taken into account (Lienert et al. 2007) and

reducing the quantities can considerably mitigate the effluent amounts if direct discharge in surface water

is expected.

1.3.2. HWW and MOs

Similarly to MPs, HWW are carriers of higher microbiological agent loads, compared to their urban

counterparts. In principle, the UWW are subjected to higher dilution and the intrinsic properties of HWW

imply the presence of higher and possibly more infectious agents. Carraro et al. in their recent review

(Carraro et al. 2016) have presented the differences between UWW and HWW effluents in various

countries, with compelling differences in the distribution of microorganisms’ species and quantities. This

microorganism load is usually led to co-treatment with UWW in MWWTPs and their removal efficiency is

a function of the existing treatment. Among others mentioned before, some authors (El-Ogri et al. 2016)

stressed the importance of treating HWW separately, on-site, to effectively reduce micro- and macro-

pollutants, but also stressing the need for microorganisms’ elimination. Problematic treatment or

inexistent treatment can lead to the problems mentioned before (antibiotic resistance, bioaccumulation

etc) but also, can directly jeopardize the drinking water sources in developing countries (Kilunga et al.

2016).

32

1.4. Occurrence, treatment and fate of MPs in MWWTPs

1.4.1. MP Occurrence in WWTPs

As the removal of microorganisms in this Thesis will focus on the mechanistic interpretation of the

inactivation phenomenon, while micropollutant removal will deal with the application in MWWTPs, in this

part their fate and transport within MWWTPs will be mainly addressed to clarify the series of events based

on the existing literature. There are several works reporting the presence and concentrations of MPs in

WWTPs, recently reviewed by Luo et al. (2014). However, one should take into account that the

concentration of MPs is a dynamically evolving phenomenon; measuring the MP content of a WW sample

is simply an instance, a snapshot that does not represent the factors involved in the process. The

occurrence of MPs in WWTPs is related to:

i) Supply rate (the rate of production)

ii) The drug availability (market restrictions)

iii) Population issues (increasing with higher numbers)

iv) GDP of the country (developed vs. developing countries)

v) Size of the WWTP (inhabitants equivalent)

vi) Special connections (industrial effluent and/or hospital WW)

vii) Water consumption (per capita)

viii) Persistence of compounds (type of pollutants)

ix) Metabolic rate of compounds (excretion rates)

1.4.2. MP evolution path in WWTPs

Municipal WWTPs at their current state are designed to handle bulk substances, such as organic matter,

phosphorus and nitrogen. The treatment of micropollutants is a latest addition and elutes specific

handling in the said installations, harboring the danger of subsequent release in water bodies. Typically,

a WWTP consists of primary treatment, followed by biological treatment/secondary clarification. Tertiary

treatment or disinfection units were additions and improvements during the last decades for effluents

destined for specific purposes. During this sequence, MPs are subjected to a number processes during

their passage through the various stages.

a) Primary treatment: In principle, primary treatment is neither destined nor capable of removing

MPs (Carballa et al. 2004, Kosma et al. 2010, Luo et al. 2014). Their low elimination is due to

sorption in primary sludge or bigger solids (Jones et al. 2006, Kosma et al. 2010, Ternes et al.

2004). In cases of aerated grit removal, increase of phenolic compounds is even expected (Kosma

et al. 2010, Nie et al. 2012).

33

b) Secondary treatment: Biological processes are expected to contribute to the removal of the

majority of the pollutants. The processes involved, such as dispersion, dilution, partition,

biodegradation and abiotic transformation (Luo et al. 2014) and less to volatilization (Verlicchi et

al. 2012). It has also been found (Margot et al. 2013) that from the 70 dissolved organic

micropollutants detected in raw effluent, 50 were removed under a 50% in the activated sludge

process. Adding nitrification improved the process in many instances (Jones et al. 2006, Kosma et

al. 2010, Margot et al. 2013). However, the biodegradability of many substances is limited, hence

leading to low percentage removal rates (Kosma et al. 2010).

c) Tertiary treatment/disinfection units: During this step, and more specifically during

flocculation/sedimentation MPs with logKow>3 can be removed, because they are more likely to

adsorb to flocs. Free chlorine has also been reported to react with some micropollutants (Kosma

et al. 2010). More details on the oxidation processes will be given later.

At this point, it is interesting to note the rather peculiar phenomenon of MP concentration increase during

their presence in WWTPs. Some human metabolites of the parent MP are secreted by the human feces

(Luo et al. 2014). More often, some semi-metabolized substances can be re-transformed back to their

original form (Göbel et al. 2007, Kasprzyk-Hordern et al. 2009). In these cases, the effluent concentration

is higher than the reported influent one.

1.4.3. Factors affecting MP removal in WWTPs

Along with the type of MP itself, the case-specific treatment sequence in WWTPs can influence the

removal rates, aided or hindered by the governing environmental conditions (Kosma et al. 2010). In

general, the physicochemical characteristics of each specific MP defines its removal possibilities. A

summary of commonly encountered micropollutants and their removal rates/classification is given in

Table 1.3.

Table 1.3 – Classification of commonly studied micropollutants (based on removal efficiency).

Low removal (<40%) Atrazine, carbamazepine, diclofenac, metoprolol…

Medium removal (40-70%) Atenolol, ketoprofen, sulfamethoxzole…

Highly removed (>70%) Acetaminophen, bisphenol A, ibuprofen, naproxen…

The two main categories of factors affecting the removal of MPs in WWTPs are the internal (MP-specific)

and the external ones (WWTP-specific). More specifically:

i) Internal factors (Luo et al. 2014): sorption, biodegradability, volatility, hydrophobicity etc.

In general, the chemical structure is defining the removal rates in high percentage. The most

34

easily degraded compounds are the ones with long side chains, the unsaturated aliphatic

compounds and the sulfate, halogen or the ones possessing electron withdrawing groups

(Jones et al. 2006, Tadkaew et al. 2011). Another main mechanism (sorption) relies on the

lipophilicity (logkow) and the acidity of the various functional groups (pKa), for absorption and

adsorption, respectively; pH values above pKa causes phenolic dissociation thus causing

charge repulsion with negatively charged membranes. Therefore, the affinity with the

bacterial enzymes is modified (Siegrist et al. 2005). Finally, other important parameters are

the Henry’s coefficient (H), governing the volatilization of MPs while treatment in the aeration

tank, and the biodegradation rate constant (kbiol); however this value is affected by external

factors and changes in dynamic way.

ii) External factors (Luo et al. 2014): treatment conditions, competition among MPs and WW

physicochemical characteristics. Sludge and hydraulic retention times (SRT and HRT)

determine the fate of the (biodegradable) MPs. Nitrifying conditions through buildup of

bacteria can enhance the removal of various compounds (Fernandez-Fontaina et al. 2012,

Suarez et al. 2009), nitrogen removal through nitrification and denitrification above 10 days

influences favorable removal rates (Clara et al. 2005) while extended HRT prolong the

available time for sorption and biodegradation (Göbel et al. 2007). Finally, redox conditions,

pH and temperature during treatment can affect either the biodiversity of microbial flora

(Göbel et al. 2007), their favorable growth conditions or the solubility of MPs in WW, leading

to low removal rates (Cirja et al. 2008), as well as phosphorus precipitation, which enhances

sorption.

Nevertheless, the aforementioned mechanisms cannot remove the micropollutants from WW, and

complete elimination relies on further, “quaternary” treatment, such as advanced oxidation treatment.

1.5. AOPs action in chemical and microbiological pollutants’

degradation

Due to the recalcitrant nature of many existing MPs, the existing biological and physicochemical treatment

methods have been proven unable to efficiently degrade them in WWTPs (Luo et al. 2014). Their

degradation however has been achieved by the use of ozonation or AOPs. These processes have

successfully mineralized or converted the persistent MPs to less harmful forms (Ikehata et al. 2006). AOPs

can be used as pre-treatment or post-biological treatment processes. Depending on the target, they can

achieve conversion of recalcitrant pollutants to biodegradable ones, or act as a polishing step. In the

respective cases, the residence time in biological treatment is reduced or the residual pollutant content

35

can be eliminated (Gernjak et al. 2006, Mantzavinos and Kalogerakis 2005, Oller et al. 2011, Ribeiro et al.

2015).

Ozonation and AOPs are redox technologies with main characteristic the non-selectivity on the target and

share the production of the highly oxidative hydroxyl radical (HO●) (Luo et al. 2014). After fluorine, the

HO● is the second most powerful oxidant (3.03 eV, compared to 2.80), with reaction rates ranging from

10-6 to 10-9 M-1 s-1 (Michael et al. 2013).

The AOPs typically involve chemical agents (metals, ozone or hydrogen peroxide) and an assistive energy

source, such as UV or visible light, current, ultrasound or γ-irradiation (Oppenländer 2003). Some

examples of AOPs are (figure 1.6):

Ozone-based: O3/H2O2, O3/UV, O3/UV/H2O2

UV-based: UV, UV/H2O2

Fenton-related: (Fe/H2O2), including photo-Fenton, electro-Fenton etc

Heterogeneous photocatalysis, such as (TiO2/hv)

γ-radiolysis

Ultrasound-based: sonolysis, ultrasound-supported Fenton etc.

Although the hydroxyl radical is the main oxidizing agent in these processes, their application often

induces the production and participation of other reactive oxygen species (ROS), such as superoxide

radical anions, hydroperoxyl radicals, singlet and triplet oxygen etc. (Ikehata et al. 2006, Oppenländer

2003). Another main advantage of the AOPs application is the characteristic versatility with which the

method can be achieved. For instance, photolysis acts directly or indirectly, by absorption of energy and

excitation or photosensitizing agents, typically dissolved organic matter (DOM) (Michael et al. 2013).

Finally, the AOPs can not only be used as a final, polishing step in order to degrade the persistent MPs in

WW, but also as a means to cause partial degradation of many compounds. This strategy is known to

cause increase in the biodegradability of organic compounds (Alvares et al. 2001), and is interesting

practice, especially for hospital or industrial effluents. If factors as pH are adjusted before and/or after

treatment (to enhance the AOP action and then prevent residual activity from oxidants) and toxicity is

mitigated, the effluent could be discharged at surface water. In the case of MWW, the organic matter

content combined with the high amounts of influent, is a prohibiting factor in this strategy. Hence, in most

applications AOP units are installed after secondary treatment (Ikehata et al. 2006).

36

Figure 1.6 – Categorization of Advanced Oxidation Process (Poyatos et al. 2010)

The two main groups of AOPs under study in the Thesis will be the UVC-based and Fenton-related ones.

More specifically, the methods drawing the interest will be:

1) UV photolysis

2) UV/H2O2

3) Fenton reaction

4) Photo-Fenton reaction

5) Solar exposure

In this thesis, the UV-based studies concern mostly application possibilities, while the Fenton-related

more fundamental and mechanistic aspects. Hence, special focus will be given here in the respective

literature and parameters of each process.

1.5.1. UV-based processes (UV, UV/H2O2)

UV treatment consists on the direct photo-transformation of organic compounds. In UV direct photolysis,

the micropollutant must absorb the incident radiation and undergo degradation starting from its excited

state. This treatment has been the most known and widely used irradiation method in initiating oxidative

degradation processes. Some organic pollutants effectively absorb UV-C light directly, and absorption of

this high-energy can cause destruction of the chemical bonds and subsequent breakdown of the

contaminant.

37

The main factors which will affect the degradation of MPs in light assisted processes are the UV absorption

and its quantum yield (Legrini et al. 1993). The molar absorption coefficient, i.e. UV absorption is an

indication of the strength with which a molecule absorbs UV, and consequently, cause its degradation

(Kim et al. 2009a, b, Michael et al. 2013). In principle, reaction kinetics of the organic substrate with the

oxidant are described by second order law, as follows (Eq.1.1):

(1.1)

where r(-M) represents the rate of degradation of the MP. At the same time, direct photolysis contributes

in the dual manner mentioned before, provided that other WW constituents and physicochemical

characteristics are present, such as pH conditions (Ikehata et al. 2006). However, it must be noted that

although some MPs are susceptible to both ROS damage and light action, others are not easily affected

by the radicals (such as ICM Iohexol) and others (such as TCEP and TCPP) which are impervious to both

types of attacks (Gerrity et al. 2011). Other important factors include the concentration of the target

compound, the pH of the matrix, the amount of H2O2, the presence/absence of scavenging compounds

(e.g. bicarbonates) and the reaction time.

This treatment is considered an advanced oxidation process because it involves the generation of hydroxyl

radicals (HO•) produced by homolytic cleavage of hydrogen peroxide. Photolysis of H2O2 yields two HO•

radicals per photon absorbed. The hydroxyl radicals are strong oxidants (E°=2.8V) with fast reactivity due

to their non-selectiveness. Hence, the removal rate of micropollutants has at least two contributions:

direct photolysis and hydroxyl radical attack. They might achieve high degrees of elimination (final

products are mainly CO2 and H2O) of several micropollutants. The efficiency of the process will depend

strongly on the HO production velocity. The propagation reactions are as follows (Eqs. 1.2-1.7):

2 ● (1.2)

● ● + (1.3)

● ● (1.4)

2 ● (1.5)

2 ● (1.6) ●+ ● (1.7)

The molar absorption coefficient of H2O2 is only 18.7 M-1.cm-1 at 254 nm (Zapata et al. 2010). Hence, the

efficiency of UV/H2O2 process decreases drastically with the presence of strong photon absorbers or when

the UV absorbance of the target pollutant is high.

38

Low-pressure mercury vapor lamps with 254 nm peak emission are the most common UV source used in

UV/H2O2 systems; however the maximum absorbance of H2O2 occurs at 220 nm. Hence, if low-pressure

lamps are used, a high concentration of H2O2 is needed to generate sufficient hydroxyl radicals because

of this low absorption coefficient. At the same time, at high concentrations, H2O2 scavenges the radicals

making the process less effective. Also, low pH is usually preferred, but in the case of UWW the

acidification/neutralization costs and implications rule out this option. The UV/H2O2 processes have

demonstrated efficiency against all compounds being able to be degraded by the HO radicals, but despite

their low environmental footprint, their operation cost halts the high degrees of commercialization.

Nevertheless, on research level, the use of this process in wastewater treatment is depicted by the

important number of publications produced during the last decade.

In order for UV light to inactivate microorganisms, the UV photons must get in contact with the cell. UV

energy penetrates the outer cell membrane, passes through the cell body and disrupts its DNA, preventing

its replication and therefore the microorganism reproduction. More specifically, UV light disrupts the

dividing of the deoxyribonucleic acid (DNA) (genetic material, chromosomes) and the production of

enzymes (Das 2001). The components within the DNA that absorb the UV light are the nucleotide bases:

adenine, guanine, thymine, and cytosine. Although proteins fulfill many vital functions in cells, their UV

absorption compared with that of DNA is of minor importance. This absorption of UV energy forms new

bonds between adjacent nucleotides, creating dimers. Dimerization of adjacent pyrimidine molecules,

particularly thymine, is the most common form of photochemical damage (Das 2001). The nucleotides

differ in their ability to absorb UV light and undergo a permanent chemical change. The pyrimidines

(thymine and cytosine) are ten times more sensitive to UV light than the purines (adenine and guanine).

Among the pyrimidines, thymine undergoes change the most readily and the dimers are very stable (Das

2001). Ultimately, the effect of numerous dimers forming along the DNA chain inhibits replication of the

organism.

The combined UV/H2O2 process builds on the principles of the UVC disinfection, plus the homolytic

disruption of H2O2. The mechanism of UV/H2O2 inactivation, is both external and internal. Externally, the

massive generation of HO● ensures oxidation of the bacterial membrane and cell lysis, while the H2O2

which penetrates the bacterial membrane is also cleaved and HO●-mediated damage ensues internally.

The natural, enzymatic defense mechanisms of the bacterial cell against ROS, such as catalase, dismutases

and peroxidases are unlikely to protect the microorganism in presence of UVC light. When H2O2 is added

into the bulk, the HO● attacks improve all the processes; till now, there is no microorganism found with

resistance to the oxidation by HO●, contrary to damages by UVC irradiation which can be repaired instantly

(Sinha and Häder 2002).

39

The most important factors that affect the UV and UV/H2O2 disinfection/decontamination capacity are

the following:

1) Water Quality

A number of substances can inhibit the transmission of ultraviolet rays through water, so special emphasis

must be given to the treatment prior to the exposure in UV (Darby et al. 1999). Certain contaminants in

water can reduce the transmission of UV light through the water, which reduces the UV dose that reaches

the bacteria. These UV-absorbing substances include micro-contaminants, but also humic and fulvic acids,

with high absorption coefficients.

2) Suspended Solids

Although in influents the suspended solids are mainly of terrestrial origin, the corresponding ones in

biologically treated effluents are typically composed of bacteria-laden particles of varying number and

size. Some of the suspended solids in wastewater will absorb or reflect the UV light before it can penetrate

the solids to kill any occluded microorganisms. UV light can penetrate into suspended solids with longer

contact times and higher intensities but there is still a limit to pathogen inactivation. Suspended particles

are a problem because microorganisms harbored within particles are shielded from the UV light and pass

through the unit unaffected.

3) Particle Size Distribution

Particle size distribution (PSD) measurements of wastewater effluent are used as an indicator of filter and

clarifier performance (Jolis et al. 2001). Even though the SS concentrations of different effluent samples

can be similar, their particle size distributions are significantly different because of the different treatment

processes before disinfection. Therefore, the influence of particle size may explain why traditional solids

measurements do not provide accurate prediction results of UV disinfection (Brownell and Nelson 2006).

4) Iron

Iron affects UV disinfection by absorbing UV light. If the concentration of dissolved iron is high enough in

the wastewater the UV light will be absorbed by the iron complexes before it can kill any microorganisms

(Das 2001). Another common problem is the iron settling onto the quartz sleeves; iron will precipitate on

the quartz sleeves and absorb the UV light before it reaches wastewater (Das 2001). This can increase

maintenance costs since the sleeves must be cleaned regularly. Finally, adsorption of iron onto suspended

solids, clumps of bacteria and other organic compounds can occur. This adsorbed iron will prevent UV

light from penetrating the suspended solids etc. and inactivating the entrapped microbes (Das 2001). On

the other hand, iron can be removed with proper pre-treatment and eliminate possible downstream

issues.

5) Hardness

40

Calcium and magnesium salts, which are generally present in water as bicarbonates or sulfates, are the

source of water hardness. The main problem with hard water is the formation of mineral deposits. These

products precipitate and coat on any warm or cold surfaces. The optimum temperature of the low-

pressure mercury lamp is between 40 °C and 104 °F. Therefore, a molecular layer of warm water can form,

where calcium and magnesium salts will be precipitated. These precipitates will prevent the UV light from

entering the wastewater (Das 2001).

6) Wastewater Source

This category engulfs special wastewater sources, such as the ones originating from textile industries.

These industries may be periodically discharging low concentrations of dye into the municipal wastewater

collection system. In this case, the effluent will be heavily colored when it reaches the treatment plant.

Dyes can readily absorb UV light thereby preventing UV disinfection (Das 2001). Dyes absorbing UV can

seem contradictory with color not affecting UV light. However many textile dyes and colorless auxiliary

substances are absorbent in the UV range.

7) Temperature

The climatic variation as well as the seasonal changes in the environmental temperature affects the

wastewater and the efficiency of UV disinfection. The kinetics of the UV disinfection process was highly

affected by system operations at extreme temperatures, i.e., at 10 and 45°C (Abu-ghararah 1994). Higher

inactivation constants were noticed in temperatures as high as 45°C and lower when the water

temperature was at 10°C, respectively.

8) Bacterial Strains

As in this thesis, only bacteria will be subjected to UV treatment, special focus will be given in the

difference among bacterial strains. Escherichia coli are commonly employed as an indicator

microorganism, as they are easily propagated and detected in the laboratory. The various strains of E. coli

(wild-type, UV-resistant and antibiotic-resistant strains) frequently encountered when the sanitation

system of hospitals malfunction, can demonstrate differential demand of UV fluence in order to be

eliminated (Quek and Hu 2008). Generally, wild or environmental strains show greater resistance to UV

rays. Also, the required laboratorial disinfection doses are lower than the ones of the real applications

and therefore the constants need correction. The evaluated studies (Hijnen et al. 2006) suggest a two

times increased fluence requirement for bacteria and four times for bacterial spores in drinking water.

For wastewater this is most likely not enough and a factor of seven seems more appropriate (Hijnen et al.

2006).

41

1.5.2. Fenton-related reactions (Fenton, photo-Fenton, solar light)

The Fenton process is an attractive oxidative system for wastewater treatment, due to iron abundance in

nature and low inherent toxicity, as well as the fact that hydrogen peroxide is easy to handle and

environmentally safe, decomposing spontaneously to H2O and O2.

It has been demonstrated that Fenton’s reagent is able to destroy toxic compounds in wastewater

(Andreozzi et al. 1999). Production of HO radicals by Fenton reagent occurs by means of addition of H2O2

to Fe2+ salts trough the following reactions (Neyens and Baeyens 2003, Stasinakis 2008). “R” is used to

describe the reacting organic compound and L is an organic ligand (Eqs. 8-11):

(1.8)

(1.9)

(1.10)

● (1.11)

However, exposure to light enhances the Fenton reaction by the photo-regeneration of Fe (II), when

reducing Fe (III). Hence, there is a double production of hydroxyl radicals (Poyatos et al. 2010) (Eq. 1.12):

(1.12)

Thus, photo-Fenton is a process that is able to use solar radiation as input taking advantage not only of

the UV portion contained in solar radiation but also because of the ability of some compounds such as

ferro-hydroxy and ferro-acid to absorb energy in the visible spectra (Figure 1.7).

Figure 1.7 – Light absorption of Fe3+ species at normal solar irradiance (I) on the Earth’s surface

The use of solar light as source of radiation for activating the hydroxyl radicals is not a new concept:

several researches have proved the efficiency of solar light as an activating agent for the Fenton reaction.

In this process, the Fe(II) is continuously recycled, reducing the amount of iron salts required (and their

42

further disposal) for the Fenton’s reaction. This feature makes the photo-Fenton process more applicable

and attractive for application in sunny regions around the globe.

Recently, Giannakis et al. have reviewed the mechanisms of photo-Fenton inactivation of microorganisms,

as well as the reported applications on water and wastewater disinfection (Giannakis et al. 2016a,

Giannakis et al. 2016b). The necessary steps to inflict inactivation of microorganisms (especially bacteria)

is explained below.

1) Direct action of light: The action of solar light (UVA and UVB light) is the baseline action, although

unimportant for MPs (photolysis), has multi-level contribution against biological targets. UVB

wavelengths lead to the formation of same-strand photo-adducts among nitrogen-containing

bases, or even in double stranded DNA. The photoproducts are cyclobutane pyrimidine dimers

(CPDs), pyrimidine (6-4) pyrimidone dimers, monomeric pyrimidine (cytosine) photoproducts, or

purine base photoproducts. In overall, the direct effects of UVA can be characterized as less

harmful, compared with the rest of the UV light wavelengths, but the direct absorption by DNA,

proteins and other structures is noteworthy.

2) Indirect action of light: Light induces a Fenton process inside the cell, which is the main indirect

pathway. When solar light is provided to the bacterial cells, the chain reaction of events follows a

complex mechanism, initiated by two simultaneous fronts: action of light and action of ROS.

Assuming that a cell is preserving its normal ROS cycle, light addition creates a chain of oxidative

events. UVB can enhance H2O2 accumulation and induce excess production in E. coli cells in

vivo (Gomes et al. 2004, Knowles and Eisenstark 1994). Also, singlet oxygen (1 ), key factor in

cytotoxicity and gene expression (Basu-Modak and Tyrrell 1993, Tyrrell And Pidoux 1989, Tyrrell

et al. 2000) can be generated by UVA irradiation, through excitation of chromophoric substances,

such as porfyrins (Tyrrell et al. 2000).

As it seems, there is an over-accumulation of ROS inside the cell, which is only made worse by the

inactivation of the key enzymes by the action of light; CAT and SOD reduce significantly their

activity when exposed to UVB or UVA light (Imlay 2003, 2008, Santos et al. 2013). It has long been

suggested that near-UV induces mutations in bacteria (in macroscopic level) and the explanation

has been attributed to the excess H2O2 accumulated into the cell and the subsequent reactions

involved with it (Eisenstark 1998). UVA has also been known to affect the respiratory chain of E.

coli, with some of the mechanisms suggested (Bosshard et al. 2010) being verified in this cycle of

events. A malfunctioning electron transport chain would provide electrons, with many reductants

now available to accept them and convert themselves to reactive intermediates.

3) Enhancement of the photo-induced actions by H2O2 and/or Fe: The first instance of synergistic

inactivation by near-UV light and H2O2 was demonstrated by Anathaswamy and Eisenstark

43

(Ananthaswamy and Eisenstark 1976) for phages and Hartman and Eisenstark some years later

(Hartman and Eisenstark 1980) for E. coli K-12. The following years many works have been

developed to assess the H2O2-enhanced photokilling modes and parameters that are involved

(Fisher et al. 2008, Fisher et al. 2012, García-Fernández et al. 2012, Hartman and Eisenstark 1978,

Keenan 2001, Khaengraeng and Reed 2005, Ng et al. 2015, Sciacca et al. 2010, Spuhler et al.

2010)). The majority of the works agree that the involved mechanism is in fact a light-enhanced

internal photo-Fenton reaction. The prevailing mechanism is as follows.

i. The direct damage of the light affects the DNA and the enzymes responsible for its

reparation (direct action).

ii. Light is disrupting the normal ROS-scavenging enzymes into the cells such as catalase,

superoxide dismutase, peroxidases etc. (indirect action)

iii. H2O2 penetrates the cell, causing imbalance of ROS into the cells.

iv. ROS and light release iron into the cytoplasm, with reacts with H2O2 to create ●. Other

ROS are involved into the reduction of iron, or directly attack susceptible moieties (oxidative

stress).

v. Added H2O2 affects bacterial membrane (outer damage), initiating its auto-oxidation.

vi. Light reduces ferric iron to ferrous directly, through ligand-to-metal charge transfer

(LMCT) or indirectly, through the reactive intermediates available by the light-induced

malfunctioning into the cell, initiating a photo-catalytic cycle.

As far as iron addition is concerned, the various steps are presented here:

Step 1: addition of Fe2+ internal action. Iron can diffuse into the bacterial cell quite easily (Braun 2001,

Touati 2000) due to low charge density and difference in osmotic pressure between the cell and the

matrix. From this point and onwards, it is available as a readily oxidizable catalyst.

Step 2: addition of Fe2+ external action. Fe2+ addition, in presence of H2O2 in the matrix, can drive a

homogeneous photo-Fenton process, for a limited period of time. Fe2+ is soluble in water, and by reaction

with H2O2, production of ● is achieved in a big extent, effectively degrading the external cell

membrane and resulting in microorganism degradation

Step 3: Fe3+ formation/addition (in presence of bacteria). Bacteria are known to produce siderophores

such as enterobactin, aerobactin, and ferrichrome, which are able to metabolically chelate Fe3+ present

in the cell (Köster 2001, Upritchard et al. 2007), to cover their needs in Fe3+. These proteins efficiently bind

to Fe3+ and create complexes, therefore facilitating internal photo-assisted LMCT and production of ●.

Step 4: Iron Oxides formation from Fe2+/Fe3+ addition. After conversion of Fe2+ to Fe3+, the Fenton process

is considered as limited, since Fe(OH)2+ has limited solubility at near-neutral pH and therefore, exploitation

44

of its photoactivity is limited (Ruales-Lonfat et al. 2014). Instead, zero-charge complexes are formed, such

as , which are prone to oxidation and formation of solid iron oxides, such as magnetite, goethite,

lepidocrocite, or feroxyhyte (Jolivet et al. 2004).

Step 5: Semiconductor action mode of iron oxides. Iron oxides, either naturally present in water (Cornell

and Schwertmann 2006) or laboratory-prepared (Cornell and Schwertmann 2006) are among the most

reactive components within the matrix. Their chemical activity involves potential photocatalyst activity, if

the hole-electron recombination (electron returning to an empty state) problem is overpassed (Zhang et

al. 1993).

Step 6: Heterogeneous (photo)Fenton reaction. Iron oxides in presence of H2O2 can play the role of an

efficient heterogeneous photo-catalyst, towards, bacterial inactivation (Ruales-Lonfat et al. 2015, Ruales-

Lonfat et al. 2014), in two ways. Firstly, in presence of siderophores, they can contribute to the supply of

dissolved Fe2+ in the bulk (Upritchard et al. 2007). Furthermore, H2O2 can start a series of reactions, at

which iron hydroxide ligands can get reduced, with simultaneous hydroperoxyl radical formation

(Upritchard et al. 2007). Under light, the production of hydroxyl radicals is also favored (Han et al. 2011).

1.6. Problem identification and contextualization: Micropollutants

and microorganisms in developed and developing countries

Since micropollutants have been identified in many cases as high risk compounds, many works have been

initiated to identify their presence in the environment (Kolpin et al. 2002, Luo et al. 2014). Moving in

backward steps, the presence in environmental water matrices is a result of a variety of pathways. One of

the main sources which will be further analyzed later on, are the MWWTPs, due to the collection of urban

and sometimes, (pre-treated) industrial effluents (Kasprzyk-Hordern et al. 2009).

Although the treatment in WWTPs is followed by natural processes, such as sorption, photolysis and

biodegradation, that can reduce the contaminant loads up to 10 times (Gros et al. 2007, Pal et al. 2010),

the MPs’ presence is still unambiguous. In a research conducted among many countries, the most

frequently encountered drugs were the non-steroidal anti-inflammatory drugs (NSAIDs),

Sulfamethoxazole, Carbamazepine and Triclosan (Luo et al. 2014). Generally, the occurrence of MPs was

less frequent in summer (probably due to elevated, temperature-driven biodegradation), and even

though winter rain promoted dilution, sometimes the contribution in natural water was important (Wang

et al. 2011). Finally, the concentrations found in surface waters was well correlated with the population

distribution, linked with the massive utilization of parent chemicals by a bigger number of users (Luo et

al. 2014).

45

Concerning drinking water, the studies are relatively few, because the occurrence is sometimes below the

detection limit (Vulliet et al. 2011, Wang et al. 2011). However, this is often a limitation of the

experimental capabilities of the analytical laboratories. Kummerer has discussed this problem, in the

appearance of “new” compounds, which could have been normally encountered (for pollutants in ng or

μg scale) (Kümmerer 2011), if the technology allowed so. Also, as far as long-term side-effects are

concerned, the presence of certain compounds or their intermediated in drinking water has not been

under study (yet). In overall, in the review published by Luo et al. (Luo et al. 2014) it is mentioned that

most of the countries investigated (France, USA, Spain, Canada) were capable of removing the presence

of some MPs in drinking water. This is a critical step, considering that drinking water treatment is literally

the last line of defense among end-users and micropollutants (Ikehata et al. 2006).

Figure 1.8 – Routes of pharmaceutical contamination of the aquatic environment (Ikehata et al. 2006).

Although in developed countries water is considered a de facto supply in each household, in developing

countries, the reality is sometimes far from this state. Water acquisition can be an everyday struggle for

many families. If in this scenario, one thinks of the potential problems that could appear if MP pollution

is high, the risks are more imminent. The quality of life of the affected population is considerably

endangered, and more specifically not by chronic or potential problems, but from the harsh reality of raw,

untreated wastewater in the water supplies.

46

In many developing countries, the combination of rampant population growth and the lack of financial

means, have led to insufficient (up to inexistent) sanitation facilities. Therefore, the collection and the

treatment of wastewater is problematic. The poorest fractions of the population, who employ themselves

in handcraft, fishing and agricultural activities suffer the most, since the situation in centralized, capital

areas is slightly better. However, these areas have demonstrated unacceptable treatment, especially in

(semi)industrial or hospital effluents, with cases describing direct untreated water being discharged in

rivers and sea.

“Fortunately”, the risk of MPs is relatively less, when compared with developed countries. The availability

of drugs and the capability of purchase restrict the widespread use and the diffuse pollution. The main

areas expected to provide major MP flows are the hospitals and similar facilities. Recent research that has

been performed in the framework of the “Treatment of the Hospital wastewaters in Ivory Coast and in

Colombia by advanced oxidation processes” (unpublished data) indicated that even in this case, the

majority of administered drugs are biodegradable and the MP content is limited in isolated hospitals in

Colombia, but in the University Hospital in Abidjan, the situation is quite problematic.

On the other hand, even when the amounts of MPs is not alarming, the presence of microorganisms in

WW is an important matter, which becomes top priority, since no disinfection process is applied in the

effluents. Therefore, the focus should be directed at least to microbial pollution, when it comes to

discharged WW in developing countries, which poses direct and acute illness risks. Hospitals are an

identified contributor to fecal and overall pathogen microorganisms in surface waters, and the lack of

treatment is directly jeopardizing their use (Kilunga et al. 2016). The current practices in agriculture for

instance, include the use of contaminated water for crops irrigation, and the transfer of pathogens is

highly possible. Therefore, monitoring of total coliform bacteria and aerobic mesophilic bacteria, as

representatives of fecal and non-fecal routes, respectively, should be monitored and their elimination

should precede discharge in natural water bodies.

1.7. AOPs vs. Micropollutants and Microorganisms: current status

Previously, the different classes of micropollutants have been mentioned, in order to classify the various

substances into different categories. Nevertheless, the anthropogenic pollutants surpass the 92 million

registered compounds (CAS registry-not all available on market). Hence, it is relatively impossible for

treatment facilities to take into account all substances in order to draw the strategies for protection.

However, not all of them pose a potential threat, nor cause (acute or chronic) problems, so the focus will

remain around the MP classes described in the previous sections.

47

Concerning the work already carried out, a literature review performed in 2013 (Michael et al. 2013)

indicated 5500 instances in Scopus, concerning treatment of pharmaceuticals by AOPs, and comparable

number of works exist for the rest of the classes of MPs (biocides etc). Also, if the topic is divided by the

efficacy of the different AOPs, for each AOP presented in Figure 1.6, the results are equally overwhelming;

against our targets, there are 3500 returned search results for UV-related processes, 4200 for Fenton-

based ones etc. Therefore, it can be easily understood that it is a very dynamic, hot topic, spanning only

in the last 15-20 years. In the same time, AOPs are a very powerful tool against the control of these

substances; the majority of the publications manage to efficiently degrade their targets, one way or

another. This indicates enormous windows of opportunity in applying AOPs as tertiary treatment

methods.

In general, research focuses on certain pillars, when it comes to treating pollutants. Some of the main

points of interest when working with AOPs are:

efficacy of certain processes against pollutants,

testing of classic AOPs against emerging pollutants

new pathways involved in degradation (due to analytical technological advances)

measurement degradation kinetics (compound-specific or mixture),

bench scale or field applications against new contaminants,

intermediates’ formation during treatment (treatment pathways),

potential toxicity problems during/after treatment,

improvement and optimization of the classic treatment processes,

statistical interpretation/integration of new mathematic tools (artificial neural networks),

development of new/combined treatment methods,

reuse and recycling of materials involved in the process,

replacement of old processes with environmentally friendly treatment methods or non-toxic,

“greener” reagents etc.

Another point worth commenting, are the results of the same search terms concerning AOPs against

microorganisms. The search results returned are orders of magnitude lower, when compared to

micropollutants. The reality is that the targets are significantly less, but of equal importance to MPs. In

any other sense, the focuses remain the same (as presented in the list) while shifting from chemical to

microbial contaminants. Keeping all the previous facts in mind, the critical question is to direct research

in context-specific basis, regarding the particularities of each context (developed/developing countries

and MPs/MOs) and suggesting proper solutions.

48

1.8. Thesis aims and objectives

As described in the previous chapters, the issue of hospital wastewater treatment has multiple contextual,

application and engineering extensions. The necessary strategies need to be specifically addressed

towards developed or developing countries, where HWW is channeled in UWW or is directly discharged,

respectively. The context differs significantly; the developed countries have more or less under control

the problem of MOs and focus on MPs, while developing countries’ priority should be the acute risk

caused by MOs presence. Furthermore, the AOP chosen has to be a function of the technical and

economic status of the place of application, with the UV-based methods being more prominent in

developed countries and the solar based ones more suitable for developing countries. As a result, the

present thesis takes account the aforementioned constraints and focuses on HWW treatment by AOPs in

developed and developing countries, both on the application point of view, as well as the underlying

mechanisms governing micropollutant degradation and microorganism inactivation.

Figure 1.9 – Thesis graphical representation and organization of aims and objectives.

In Switzerland (as an example of developed country), the wastewater effluents are treated and the

implementation of the relative AOPs focuses on the elimination of the chronic risk caused by the presence

of micropollutants in natural waters, and the acute risk of microbial infection due to the pathogens carried

within the flows. The micropollutants chosen in this Thesis derive from the modifications in the Swiss

legislation and the wild, indigenous bacterial population as microbial targets. In the first part, the effect

of different pre-treatment methods against these targets is of key interest.

49

In Ivory Coast and Colombia, as the involved countries in this research project, WWTPs are inexistent.

Therefore, the application of solar photo-Fenton as a feasible AOP is implied only after the construction

of basic treatment before (primary-secondary). The acute risk of microbial infections is prioritized and

studied extensively in the second part, taking a viral and a yeast pathogen model into study, as the photo-

Fenton against bacteria has already been a subject of the Thesis.

In the third and final part, as mass flows of special drugs derive from hospital and production sites, two

pollutants (Iohexol and Venlafaxine) in high amounts have been chosen and their degradation by relevant

AOPs was studied. These drugs can be encountered in the production wastewaters or in urine, due to the

treatment of patients, and the (pre)treatment of concentrated flows at hospital or manufacturing level is

desirable before dilution in the municipal wastes.

These objectives, organized in the respective chapters, can be summarized as follows:

Chapter 1: Introduction.

PART 1: Disinfection and decontamination of municipal WW as carrier of HWW pollutants in developed

countries by all proposed AOPs: implications of AOPs application in WW.

Chapter 2: Effect of advanced oxidation processes on the micropollutants and the effluent organic matter

contained in municipal wastewater previously treated by three different secondary methods.

Chapter 3: Micropollutant degradation, bacterial inactivation and regrowth risk in wastewater effluents:

Influence of the secondary (pre) treatment on the efficiency of Advanced Oxidation Processes.

PART 2: Hospital-derived microorganism inactivation in developing countries by Fenton-related AOPs:

mechanistic interpretation and underlying mechanisms of the photo-Fenton process.

Chapter 4: Effect of Fe(II)/Fe(III) species, pH, irradiance and bacterial presence on viral inactivation in

wastewater by the photo-Fenton process: Kinetic modeling and mechanistic interpretation

Chapter 5: Castles fall from inside: Evidence for dominant internal photo-catalytic mechanisms during

treatment of Saccharomyces cerevisiae by photo-Fenton at near-neutral pH..

PART 3: Degradation of hospital PhACs by AOPs, as a point-source treatment option in HWW and urine:

treatment optimization and degradation pathway elucidation.

Chapter 6: Iohexol degradation in wastewater and urine by UV-based Advanced Oxidation Processes

(AOPs): Process modeling and by-products identification.

50

Chapter 7: Solar photo-Fenton and UV/H2O2 processes against the antidepressant Venlafaxine in urban

wastewaters and human urine. Intermediates formation and biodegradability assessment.

Chapter 8: General Conclusions and Perspectives.

51

PART 1

Disinfection and decontamination of municipal WW as carrier of

HWW pollutants in developed countries by all proposed AOPs:

implications of AOPs application in WW.

53

2. Chapter 2: Effect of advanced oxidation processes on the

micropollutants and the effluent organic matter

contained in municipal wastewater previously treated by

three different secondary methods

Published work:

Stefanos Giannakis, Franco Alejandro Gamarra Vives, Dominique Grandjean, Anoys Magnet, Luiz Felippe

De Alencastro, and César Pulgarin. "Effect of advanced oxidation processes on the micropollutants and

the effluent organic matter contained in municipal wastewater previously treated by three different

secondary methods." Water research 84 (2015): 295-306.

Web link:

http://www.sciencedirect.com/science/article/pii/S0043135415301329

Supplementary material:

Appendix A

Doctoral Candidate’s contribution:

Main investigator and author.

54

2.1. Introduction

Currently, one of the environmental concerns in global scales is the presence and accumulation of

micropollutants (MPs) in the natural environment. These substances are comprising an increasing list of

anthropogenic contaminants, which include among others, pharmaceuticals, personal care products,

steroid hormones, industrial chemicals, pesticides and many other emerging compounds (Luo et al. 2014).

The majority of these substances are designed to be biologically active, and therefore, their occurrence

can affect the receiving environment. Some of the associated problems to MPs’ presence are the

ecotoxicity of their mixture (Gregorio and Chèvre 2014), the enhancement of antibiotic resistance by the

presence of antibiotics and their metabolites in the environment (Fatta-Kassinos et al. 2011), or the

problematic identification of the transformed metabolites of drugs in the nature (Fatta-Kassinos et al.

2011).

Wastewater treatment plants have been built, transformed and updated through the years to effectively

remove -among others- solids, organic and inorganic compounds (carbon, nitrogen, phosphorus etc.),

which were previously disposed in the environment. However, the challenge posed by the hydrophilic

MPs, is an obstacle the majority of the WWTPs are not equipped to handle. Margot et al. (Margot et al.

2013) have reported that from the 70 dissolved organic MPs detected in raw effluents, many of them

were removed at less than 30% in the conventional activated sludge process.

However, micropollutant degradation has been achieved by the use of ozonation or Advanced Oxidation

Processes (AOPs). These processes have successfully degraded or converted the persistent MPs to less

harmful forms (Ikehata et al. 2006). Although the hydroxyl radical (HO●) is the main oxidizing agent in

these processes (oxidative power: 2.80 eV), their application often induces the production and

participation of other reactive oxygen species (ROS), such as superoxide radical anions, hydroperoxyl

radicals, singlet and triplet oxygen etc. (Ikehata et al. 2006) The main advantage of the AOPs application

is the characteristic versatility with which the method can be achieved. For instance, photolysis acts

directly or indirectly, by absorption of energy and excitation or photosensitizing agents, typically dissolved

organic matter (DOM) (De la Cruz et al. 2012).

In Switzerland, actions against the issue of MPs were systematically taken since 2006, when the Micropoll

Strategy was implemented, by the Federal Office for the Environment (FOEN 2014). The outcome of the

project was an adaptation of the Water Protection Ordinance (GSchV) and 80% of the total MPs was

decided to be eliminated, with 5 substances playing the role of indicator (Table 2.1). The adaptation was

finally voted in 2014 and will be effective starting January 2016. WWTPs will have to eliminate at least

80% amount of 6 pollutants (from a list of 12), divided in “very well eliminated” and “well eliminated”, as

55

presented in Table 2.1. Oxidation (for instance by ozone or AOPs) or adsorption (activated carbon) are

suggested to tackle the issue.

Table 2.1 – Evolution of indicator pollutants in the Swiss legislation.

5 Indicators List of 12 indicators Pollutant type

(2009 list) (2016 enforcement)

Very Well Eliminated

Benzotriazole Amisulpride Antidepressant

Carbamazepine Carbamazepine Antiepileptic

Diclofenac Citalopram Antidepressant

Mecoprop Clarithromycin Antibiotic

Sulfamethoxazole Diclofenac Analgesic

Hydrochlorothiazide Diuretic

Metoprolol Beta blocker

Venlafaxine Antidepressant

Well Eliminated

Benzotriazole Anticorrosive

Candesartan Angiotensin II antagonist

Irbesartan Angiotensin II antagonist

Mecoprop Herbicide

In this work, the ability of AOPs to degrade MPs at lab scale, after 3 different secondary treatment

methods is tested. UV/H2O2 and solar photo-Fenton, as well as their three composing sub-processes (UV-

C at 254 nm, solar light and Fenton reaction), are considered as a final treatment step after Activated

Sludge (AS), Moving Bed Bioreactor (MBBR) or Coagulation-Flocculation (CF) treatment. 6 pollutants were

selected from the list of 12 indicators provided by FOEN, also contained in the WW from Vidy WWTP

(Lausanne, Switzerland). The degradation kinetics, the oxidation levels and the effect of the preceding

secondary treatment are given and the dependence of success as a function of the previous treatment

steps. Finally, considerations on the role of effluent organic matter (EfOM) are given.

56

2.2. Materials and Methods

2.2.1. Sampling campaign

In total, 4 sampling campaigns took place. The strategy was as follows: for two consecutive days, 2-L

samples were taken at 9am, 12pm and 15pm. The composite sample (12 L) was then mixed and

analyzed/treated. The physicochemical analysis was immediately performed, while the MPs were

analyzed the day after the creation of the composite (preservation at 4°C). The samples were collected

from the primary effluents (point 1, Figure 1) and at the output of the different secondary treatment

facilities in the WWTP of Vidy: Activated Sludge (AS), Moving Bed Bioreactor (MBBR) and Coagulation-

Flocculation (CF) (points 2, 3 and 4, Figure 2.1).

Figure 2.1 – Simplified overview of the Vidy WWTP and the sampling points used in the study.

2.2.2. Chemicals and reagents

All the chemicals for the experiments were used as received. Hydrogen Peroxide 30%, Iron Sulfate

Heptahydrate and Titanium Oxysulfate (1.9-2.1%) (for H2O2 determination) were acquired from Sigma-

Aldrich (Switzerland) and Sodium Bisulfite from Acros Organics (Switzerland).

Sand removal Screening Primary Clarifier

~20% MPs removal

Activated Sludge

Moving Bed Bioreactor

Coagulationon-n-Flocculation

2

~40% MPs removal

~25% MPs removal

1

3

4

Influent

57

2.2.3. Employed reactors

For the UV-C irradiation and UV/H2O2 treatment methods, two 300 mL double-wall, water-jacketed glass

batch stirred reactors were used in parallel. Each of them contained a 36-W, low-pressure amalgam lamp

(Model: UVI 40 4C P 15/300) from UV-Technik Speziallampen. These lamps have an emission irradiance

of 350 μW/cm² at 254 nm and 9 W output. During the experiments, the temperature of the reactors was

water-controlled at 22 °C with a Neslab RTE-111 recirculating thermostat.

For the Fenton experiments, 100-mL dark Pyrex reactors were used. Wastewater was continuously stirred

(300 rpm) with magnetic bars and the photo-Fenton reaction took place in the same conditions, although

transparent reactors were used. Light was provided by a solar simulator (Hanau, Suntest), with a detailed

description in de la Cruz et al., (De la Cruz et al. 2012). Light intensity was controlled and kept constant at

900 W/m2, by a Global and UV radiometer Kipp & Zonen (Models CM3 and CUV3, respectively).

2.2.4. Advanced Oxidation Processes specifics

2.2.4.1. UV-C and UV/H2O2 irradiation

A volume of 300 mL of the composite wastewater sample was introduced into each UV reactor and they

were exposed to UV-C irradiation during 10 min and 30 min, respectively. The pH was monitored and the

treated WW was stored at 4°C. For the UV/H2O2 treatment, hydrogen peroxide (30%) was added, reaching

an initial concentration of 25 mg H2O2/L in the reactor. The residual H2O2 was neutralized with sodium

bisulphite.

2.2.4.2. Solar irradiation, Fenton and photo-Fenton reaction

Wastewater samples contained in 100-mL brown bottles were treated with 25 mg H2O2/L and 5 mg Fe2+/L.

The iron source was FeSO4 •7H2O. Hydrogen peroxide dose was reached adding sufficient volume of 30%

H2O2 solution. WW samples contained in 100 mL Pyrex glass bottles were exposed to irradiation in the

Suntest solar simulator. For photo-Fenton, the same doses of H2O2 and Fe2+ used in Fenton experiments

were added. During the different treatment methods, pH, hydrogen peroxide and iron concentration

(when applicable) were monitored. Treated WW was stored then at 4°C until analysis.

Table 2.2 summarizes the treatment times and conditions of the AOP experiments. All experiments were

conducted with the MP content of the effluent WW (no spiking took place whatsoever). Assessment of

the MPs’ content before treatment was calculated as an average of 2 campaigns. Also, the results

concerning the MPs degradation (Figures 2.2 to 2.6) are a result of 4 campaigns, during which all

experiments were performed in duplicates, and each matrix (and AOP) was assessed in at least two

campaigns; the presented results are subject to approximately 5% standard error. The percentage results

58

were calculated with weighted arithmetic mean, rather than simple average of degradation %, according

to the following formula:

(2.1)

Where X% is the overall removal %, is the quantity of the pollutant (mol) and (i = 1 to 6) the removal

percentage of each MP. In this way, the quantity of the pollutants is taken into account in the calculations.

Table 2.2 – Summary of the advanced treatment methods applied.

Time

[min]

H2O2

[mg/L]

Fe2+

[mg/L]

UV-C irradiation 10, 30 - -

UV/H2O2 5, 10, 30 25 -

Solar irradiation 30, 60 - -

Fenton 60,120 25 5

photo-Fenton 30, 60 25 5

2.2.5. Physicochemical parameters

A pH meter (Mettler Toledo, GmbH) was used to measure pH values before and after each advanced

treatment method. A Perkin-Elmer UV/Vis Lambda 20 spectrophotometer was used in order to perform

the measurements of H2O2 at 410 nm (modified method: DIN 38 402 H15) and iron (Fe+2/+3) by the

Ferrozine method (De la Cruz et al. 2012). Calibration curves with respective WW effluents were made to

minimize spectral interferences. Total organic carbon (TOC) was measured by a Shimadzu TOC-V

CPH/CPN, while a digestion reactor and a spectrophotometer HACH DR/4000 were used to carry out the

COD measurements.

2.2.6. Analytical methods

For the MPs’ analysis, before and after the application of AOPs in lab scale, solid phase extraction,

followed by UPLC/MS-MS was employed. The analytical procedure applied is presented in detail in

previous works (De la Cruz et al. 2013, De la Cruz et al. 2012). In a summary, 300 mL of acidified samples

to pH=2.0 (32% hydrocloric acid) were filtered through 0.7 μm Whatmann glass fiber and spiked with 200

μL of a standard surrogate, containing the MPs in deuterated form. Solid Phase Extraction (SPE) was

performed automatically (GX-274 ASPEC, Gilson) onto Oasis HLB (200 mg, 6 mL) cartridges (Waters),

59

followed by nitrogen drying. After elution with methanol, the MPs were analyzed by UPLC/MS-MS

(Acquity Xevo-TQ, Waters). Multiple reaction monitoring mode with two transitions was used to detect

MPs and quantification was performed with internal standard calibration.

2.2.7. Secondary treatment systems specifications

The Vidy WWTP receives the municipal effluents from 16 different cities representing 220.000 inhabitants

and 40 million m3/year. It is composed mainly of a pre-treatment phase to separate suspended solids and

fats particles, followed by a primary decantation. Afterwards, the WW is mainly subjected to an aerobic

biological treatment followed by a sedimentation tank. The current secondary treatment consists of

activated sludge with a capacity of 1200 L.s-1. Finally, apart from the main line, the WWTP of Vidy accounts

with two secondary treatment installations: an MBBR and a coagulation-flocculation physicochemical

treatment.

2.2.7.1. Activated Sludge (AS)

It is the current biological treatment used at the WWTP of Vidy. The hydraulic retention time (HRT) is ~4

h and the sludge retention time (SRT) is ~2 days. The process does not include a nitrification step. It has a

reported MPs’ removal efficiency of 23%, as found by Margot et al. (Margot et al. 2011).

2.2.7.2. Moving Bed Bioreactor pilot plant (MBBR)

In this facility, MPs’ removal efficiency is about 44% which is higher than at activated sludge treatment

(Margot et al. 2013). The positive correlation among MP removal and nitrification could be attributed

either to the higher hydraulic residence time, or either to the higher microbial diversity of the nitrifying

bacteria; the latter have the ability to degrade many MPs, attributed to the co-metabolism due to the

ammonium monooxygenase enzyme (Fernandez-Fontaina et al. 2012). The treatment capacity is limited

to 5% compared to the AS unit.

2.2.7.3. Coagulation-Flocculation (CF)

This treatment method is based on chemical coagulation and flocculation processes in WW. The WWTP

of Vidy uses iron (III) chloride (FeCl3 - 40%) as coagulant agent.

2.3. Results

2.3.1. Initial conditions

After the three secondary treatment systems, the effluents recovered were found to differ in their

average physicochemical characteristics. Table 2.3 summarizes the main parameters controlled prior to

the application of advanced treatment. Among the three WW, the effluent of MBBR presents a series of

60

advantages. First of all, the pH is slightly lower than the AS and CF effluents. The organics content is

significantly lower as well, which can work as a precursor of the performance; since the AOPs rely heavily

on the HO● production, which non-selectively attack organics and MPs, less competition is provided for

the radicals produced. Alkalinity plays an important role, as it is an indicator of the (bi)carbonates content

of WW, and MBBR effluents contain 4 times less. Bi-carbonates are known to scavenge the hydroxyl

radicals, thus hampering the degradation efficiency (Carra et al. 2014). The suspended solids are

important for irradiation-based processes, since they cause shielding and pose a physical barrier between

the target and the photons. UV and solar irradiation are expected to perform very well both in AS and

MBBR effluents, since the difference is only 2 mg/L. Finally, the initial iron content in the effluents is an

indicator of the expected Fenton (side)reactions or the action as suspended solids, as far as UV light is

concerned; excess of iron, after a certain point can inflict an inverse effect, and pose a physical barrier in

the transmission of light within a wastewater sample (De la Cruz et al. 2013).

Table 2.3 – Initial physicochemical characteristics of the effluents of the different secondary treatment

units.

Parameter Unit

Wastewater previously treated with

Activated Sludge


Coagulation Flocculation

pH - 7.8 7.4 7.9

TOC mg/L 37 20.2 57.2

COD mg/L 63 35 90

Alkalinity mg CaCO3/L 273 85 231

TSS mg/L 12 14 30

Total iron mg Fe/L 0.9 1.6 1.9

Comparing the three effluents, analysis has been performed to estimate the biodegradability of the

substances during the treatment in the different secondary treatment systems. Figure 2.1 shows the

evolution of the MP concentration during their presence in the WWTP. The treatment efficiency per

method was around: MBBR 40% > AS 25% > CF 20%, near to the values assessed by Margot et al. (Margot

et al. 2013). However, our estimation of removal percentage was approximated with 2 sampling

campaigns before treatment (Sampling point 1) and the average of 4 after the various secondary methods

(sampling points 2, 3 and 4), and was subject to (heavy) temporal variation for some substances (30-40%

difference).

61

Table 2.4 – Initial micropollutant concentration and limit of quantification (LOQ) for micropollutants in

the effluent of the different secondary treatment units.

Micropollutant (ng/L)

LOQ (ng/L)

AS (point 2) MBBR (point 3) CF (point 4)

[C] |Variation| [C] |Variation| [C] |Variation|

Carbamazepine 6 220 31 349 104 238 70

Diclofenac 18 1358 586 1254 328 1579 712

Metoprolol 6 579 134 793 166 855 317

Clarithromycin 6 490 132 363 52 479 153

Benzotriazole 14 4199 3633 7244 1566 7228 2366

Mecoprop 9 235 205 20 13 26 17

TOTAL* 7081 +4721 10023 -2230 10405 +3636

*The values in bold correspond to the values used for the experiments, the signed variation indicates the maximum value measured during the campaigns.

The 4 sampling campaigns results (at points 2, 3 and 4) on the initial MPs concentration are summarized

in Table 2.4. The MPs amount in the effluents of the three different secondary treatment methods slightly

differ from the estimation that derives from their removal efficiency during the secondary treatment

(described in Figure 2.1), because our (post-secondary)treatment experiments were conducted in

different sampling campaigns. Margot et al. (Margot et al. 2013) have mentioned the big variability of the

MPs content in the effluents, which makes the estimations complex. Also, AS presents a lower quantity,

due to the dependence of the average to Benzotriazole. In all cases, Benzotriazole consists of more than

50% of the monitored MP content. Hence, its variation highly influences the presented values. However,

it can be also attributed to high solubility in water (19800 mg/L) and low logkow (1.44); this specific

pollutant is very mobile in water and is not expected to sorb onto the sludge during treatment. Finally,

the total 6 MP content is on the order of 10 μg/L, at an average of 11, 10 and 13 μg/L, for AS, MBBR and

CF wastewater effluents, respectively. For micropollutant degradation in complex matrix such as

wastewater, this difference is not going to influence highly the results; if the initial COD and TOC contents

are considered, all the MPs monitored combined add up to 0.1% of the total organic load.

62

2.3.2. Efficacy of the various advanced oxidation processes

2.3.2.1. UV-C irradiation & UV/H2O2

The UV-C treatment method is based on direct photolysis of the persistent organic compounds by light

emitted at 254 nm. The main mechanism is the electronic excitation of the organic compounds, leading

to electron transfer from the excited state of the target compound to ground-state molecular oxygen or

homolysis to form organic radicals that react with oxygen (Legrini et al. 1993).

Figure 2.2 – UV treatment results. a) % degradation vs. time b) % COD & TOC reduction vs. time.

Figure 2.2a shows the removal efficiency of the 6 selected MPs at 10 min and 30 min of treatment, as well

as the evolution of COD and TOC (Figure 2.2b), for the three secondary treatment methods. After 10 min

of treatment, the removal efficiency was very similar, being 80, 85 and 79% for the AS, MBBR and CF

effluents, respectively. However, applying UV-C irradiation for 30 min, the three different effluents

demonstrated the following order of removal efficiency: Previously treated by MBBR (97%) > AS (93%) >

CF (92%).

Diclofenac and Mecoprop are already degraded 100% by UV-C irradiation alone, when the WW is treated

for a time as short as 10 min, for the three types of WW. This result verifies past evidence that these

compounds absorb well UV irradiation at 254 nm and can be easily photolyzed (De la Cruz et al. 2012).

Despite that Benzotriazole has shown low removal efficiencies against direct photolysis for pH > 7 in a

recent study (Bahnmüller et al. 2014), it showed relatively high efficiency in this case, achieving 100%

removal after 30 min of treatment. Furthermore, Clarithromycin showed a partial degradation after 30

min UV-C irradiation and only after MBBR treatment reached more than 80% removal. On the other hand,

Carbamazepine and Metoprolol remained below 80% removal after 30 min UV-C irradiation for the three

types of WW.

63

Finally, after 30 min of UV-C exposure, TOC was removed by 6.6%, 11.0% and 5.3%, while COD removal

rate was 47%, 71% and 27%, for the three types of WW (AS, MBBR, CF), respectively. The WW effluent

from Coagulation-Flocculation was the less favorable for organic compounds degradation.

Furthermore, UV/H2O2 is one of the most efficient AOPs, combining the immediate UV effect and the HO●

radicals produced from the homolytic disruption of H2O2. The initiation, propagation and termination

reactions involved are the following (Guo et al. 2013, Legrini et al. 1993):

2 ● (1.2)

● ● + (1.3)

● ● (1.4)

2 ● (1.5)

2 ● (1.6) ●+ ● (1.7)

After 5 min treatment, the order of MPs’ removal was the following: AS (99%) > MBBR (97%) > CF (96%).

After 10 min treatment, 100% removals of MPs removal was achieved, except for the WW previously

treated with CF, which showed 99% removal. 100% removal was already achieved in 30 min of exposure,

with the necessary time being calculated to 13 min. Figure 2.3a shows the removal efficiency for the 6

MPs at 5 min, 10min and 30 min of treatment, along with the COD and TOC removal (figure 2.3b) and the

remaining H2O2 content during treatment (Supplementary Figure S1).

Figure 2.3 – UV/H2O2 treatment results. a) % degradation vs. time. b) % COD & TOC reduction vs. time.

More specifically, Diclofenac and Mecoprop were 100% removed after 5 min of treatment. Benzotriazole

was also removed efficiently (over 80%) for the three types of WW. Clarithromycin was the only

compound not removed at 100% after 5 min treatment for the AS WW type. For the WW type

64

corresponding to CF, even after 10min treatment, Carbamazepine and Metoprolol showed persistent

behavior reaching 73% and 95% removal, respectively. After 30 min treatment, all the selected MPs were

eliminated (figure 2.3a). Concerning the consumption of H2O2, a great reactivity is demonstrated.

Hydrogen peroxide is consumed almost at the same rate for the three different types of WW

(Supplementary Figure S1). After 30 min treatment, it was reduced almost completely (3 mg/L).

Concerning the global parameters, after 30 min treatment, TOC removal was 31.5%, %, 25.3% 12.0%,

while COD was removed ~100%, ~100% and 32% for AS, MBBR and CF effluents, respectively. The WW

types that were subjected to a previous secondary biological treatment reached higher TOC and COD

removal efficiency than the WW previously treated with CF. Finally, the two first types of WW achieved

almost 100% of COD removal; therefore, after 30 min UV/H2O2 treatment, extensive mineralization

process has started.

2.3.2.2. Solar irradiation, Fenton reagent and solar photo-Fenton treatment

Solar irradiation is anticipated to have the lowest impact against MPs, since the quantum yield of the

photons and the wavelengths reaching the organic compounds fall within less energetic bands of the light

spectrum. However, the experiments investigating effect of solar light were interesting, providing

information on the photo-transformation taking place in WWTPs and in natural waters, by indicating the

photo-sensible compounds. Figure 2.4a shows the MPs removal efficiency and 2.4b the COD and TOC

removal, respectively.

Figure 2.4 – Solar exposure results. a) % degradation vs. time b) % COD & TOC reduction vs. time.

After 60 min treatment, the order of removal efficiency was: MBBR (17%) > AS (11%) > CF (5%). Diclofenac

and Mecoprop had the most photo-sensible behavior; thus, they achieved more than 25% removal after

60 min treatment for the three types of WW. Carbamazepine, Metoprolol and Clarithromycin presented

65

more than 10% removal, except for the WW coming from the CF, where the Carbamazepine concentration

remained unchanged (0% removal). Benzotriazole was also very resistant to this treatment, and no

significant removal was found. Only the MBBR effluent demonstrated Benzotriazole removal (~15%).

Finally, as it was expected, the TOC and COD removal remained low after 60 min of solar exposure. After

60 min of simulated solar irradiation, TOC was removed of 9.5%, 3.7% and 5.1%, while COD decreased

18%, 21% and 11% for the three types of WW (AS, MBBR and CF), respectively.

The treatment by the Fenton reaction relies on the effect of the hydroxyl radicals produced by H2O2 and

Fe2+, and since this treatment is carried out in dark conditions, very slow regeneration of iron takes place

through the following reactions (Neyens and Baeyens 2003, Stasinakis 2008):

(1.8)

(1.9)

(1.10)

(2.2)

(limiting step) (2.3)

Figure 2.5 summarizes the MPs removal (2.5a), the H2O2 and iron concentration (Supplementary Figure

S2) along with the COD and TOC removal during the Fenton treatment (figure 2.5b). The fast oxidation of

Fe2+ to Fe3+ and the generation of a massive oxidative wave is followed by a steady Fenton cycle leading

to MP degradation. After 60 min of treatment, the order of removal efficiency was: MBBR (19%) > AS

(13%) > CF (3%), while after 120 min treatment, MPs removal of the WW coming from AS approached the

removal efficiency of the WW of MBBR, without significant increase in the CF water.

Figure 2.5 – Fenton process results. a) % degradation vs. time. b) % COD & TOC reduction vs. time.

66

The most susceptible pollutant to treatment was Diclofenac, with 75% elimination rate in 2 h, followed by

Mecoprop (45%) and Carbamazepine (24%). Clarithromycin was almost unchanged (2% reduction), and

Benzotriazole and Metoprolol were mildly affected, demonstrating 12 and 15% reduction, respectively.

Despite that there is faster consumption of H2O2 for the WW effluents of AS and CF than for the one

resulting from MBBR, the MPs removal is higher in this last one (Supplementary Figure S2). This is mainly

attributed to the lowest pH among the others, acting beneficially as far as the Fenton process is

concerned, and the lower alkalinity of the MBBR WW type (85 mg CaCO3/L) compared to the one found

in the AS and CF WW (273 mg CaCO3/L and 231 mg CaCO3/L, respectively), since a lower bicarbonate

concentration was found. Iron concentration decreased accordingly to their initial pH. After 120 min of

treatment by the Fenton reaction, TOC removal was of 27.8%, 41.2%, and 16.0%, while COD decreased

27%, 83% and 19% for AS, MBBR and CF-treated WW, respectively.

Finally, the solar photo-Fenton has the advantageous synergistic action of light on the Fenton reaction,

regenerating Fe3+ back to Fe2+, thus limiting iron precipitation and enhancing further radical production,

through the following reaction (Pignatello et al. 2006):

(2.4)

Figure 2.6a shows the MP removal by the photo-Fenton reaction, 2.6b presents the COD and TOC removal

and the H2O2 and iron monitoring are present in Supplementary Figure S3. Although after 30 min the

removal efficiency is MBBR (21%) > AS (19%) > CF (7%), after 60 min treatment the difference increases,

the order of removal efficiency is: MBBR (31%) > AS (28%) > CF (11%).

Previous research, similarly to ours, has indicated the advantageous treatment against Metoprolol and

Carbamazepine (Klamerth et al. 2010, Prieto-Rodríguez et al. 2013). Diclofenac showed important

removal efficiency for the biologically pre-treated WW, while much lower removal efficiency for the one

coming from the CF treatment. Nevertheless, only Diclofenac achieved more than 80% removal for the

MBBR WW type. The rest of MPs were partially removed after 60 min treatment. Metoprolol,

Clarithromycin and Benzotriazole presented the lowest rate (less than 10% removal) for the CF WW type.

Concerning the organic matter, after 60 min of photo-Fenton treatment, TOC had removal efficiencies of

27.7%, 29.8%, 8.0%, while COD achieved removal efficiencies of 47%, ~100% and 10%, for the three types

of WW, respectively. H2O2 is consumed accordingly with the content of organic compounds (COD, TOC)

and the presence of inhibitors (carbonates and suspended solids) contained in each WW type. Dissolved

iron concentration remains almost constant during photo-Fenton treatment due to the Fe (II)

regeneration (Supplementary Figure S3).

67

The photo-Fenton treatment test was carried out during 60 min, and it could not reach the indicative 80%

MPs removal for any of the WW types. Consequently, it is suggested to extend the experiments residence

time until the removal goal is attained. However, if same degradation rates observed in the experiments

are followed, the estimated residence times would be: 116 min, 92 min and 185 min for AS, MBBR and CF

WW types, respectively. A summary of the degradation percentage per pollutant per process is given in

the Supplementary Material (Supplementary Figure S4).

Figure 2.6 – photo-Fenton process results. a) % degradation vs. time. b) % COD & TOC reduction vs.

time.

2.3.3. Degradation kinetics evaluation for the 6 different pollutants

The MPs chosen in this study were degraded by the different advanced treatment methods were fitted to

first-order degradation kinetics. Due to the limited sampling intervals these kinetics represent a common

ground for comparison only, described by the following equation:

(2.5)

or formulated as:

(2.6)

Table 2.5 summarizes the kinetic constants per AOP and (secondary) pre-treatment method. The

degradation rate constants k indicated by “nc” are the ones that could not be calculated due to the fact

that they were totally degraded before the minimal time span evaluated. For example, Diclofenac was

100% eliminated before 10 min of UV-C irradiation, or Mecoprop, which was also totally degraded before

5 min of UV/H2O2 treatment, for all three types of WW effluents. In addition, k values indicated by zero

represent the compounds that showed an insignificant (or not existent) removal. For example,

carbamazepine remained with the same initial concentration after 120 min of Fenton treatment for the

68

WW coming from CF. The degradation rate constants shown above are estimative values, and in order to

give more accurate kinetics, more and extended time spans (until complete degradation) should be

considered.

Regardless the type of WW analyzed, the order of degradation rates was the following:

Concerning the MP degradation, the degradation trends (the order indicates the increasing degradation

rate) and details on the Achilles’ heel of each pollutant are provided in the Supplementary Material

(Supplementary Table S1 and S2, respectively).

2.3.4. Evolution of the Average Oxidation State during the 5 different treatment processes

One of the most widely used parameters in WW degradability assessment is the Average Oxidation State

(AOS), which use the evolution of Total Organic Carbon (TOC) and Chemical Oxygen Demand (COD):

(2.7)

where COD and TOC values are expressed in mol O2/L and mol C/L, respectively. After each treatment

process, the following changes, summarized in Table 6, are observed.

As a common pattern, at the beginning of the highly oxidative process, the AOS increased rapidly, followed

by a stage when this increase decelerates, suggesting that the chemical nature of the intermediates

formed does not vary significantly. Nevertheless, the change of the AOS may not represent an important

impact on MPs removal, because MPs only represent a small part of the total organic compounds

contained in the WW (De la Cruz et al. 2012). In this study, only ~10 μg/L of the total organic load are the

6 MPs. Therefore, the degree of MP elimination is correlated with the change in AOS in treated WW. Since

the COD and TOC of a WW are relatively easy parameters to monitor, the AOS will predict the elimination

of the MPs.

Table 2.5 – Degradation kinetics of the 6 different pollutants during treatment in the different

effluents and treatment methods.

Degradation constant UV UV/H2O2

k (min-1) AS MBBR CF AS MBBR CF

Carbamazepine 0.027 0.046 0.021 nc 0.644 0.131

Diclofenac nc nc nc nc nc nc

Metoprolol 0.032 0.048 0.034 nc 0.599 0.3

69

Clarithromycin 0.035 0.061 0.024 0.563 0.354 0.277

Benzotriazole 0.23 0.281 0.204 nc 0.701 0.921

Mecoprop nc nc nc nc nc nc

Degradation constant Solar Fenton photo-Fenton

k (min-1) AS MBBR CF AS MBBR CF AS MBBR CF

Carbamazepine 0.004 0.003 0 0.004 0.002 0 0.006 0.01 0.007

Diclofenac 0.005 0.008 0.005 0.005 0.012 0.001 0.021 0.071 0.011

Metoprolol 0.003 0.001 0.002 0.001 0.001 0 0.003 0.01 0.001

Clarithromycin 0.001 0.003 0.002 0.001 0 0 0.002 0.003 0

Benzotriazole 0.001 0.003 0 0.001 0.001 0 0.003 0.003 0.001

Mecoprop 0.005 0.009 0.004 0.004 0.005 0.002 0.009 0.008 0.004

As a first step, the percentage of AOS change during treatment was calculated. The values are summarized

in Table 2.6. Afterwards, the correlation among the percentage of MP degradation per treatment method

(for all effluents) was estimated. The Pearson test was used to detect the correlation between the two

parameters, and the P-value indicating the verification of the hypothesis (that the parameters are

correlated) within the 95% confidence interval. The Pearson correlation values showing mild correlation

are among 0.6 and 0.8, while values among 0.8 and 1 indicate strong correlation among the parameters.

The accepted P-values are lower than 0.05.

Table 2.6 – Evolution of the Average Oxidation State (AOS) during treatment by the various methods

in the different effluents.

UV AS MBBR CF UV/H2O2 AS MBBR CF

0 min 1.45 1.4 1.64 0 min 1.45 1.4 1.64

10 min 1.8 1.95 1.75 10 min 3.43 2.16 2.16

30 min 2.55 3.17 2.18 30 min 4 4 2.18

Solar AS MBBR CF Fenton AS MBBR CF photo-

Fenton AS MBBR CF

0 min 1.45 1.4 1.64 0 min 1.45 1.4 1.64 0 min 1.45 1.4 1.64

30 min 1.48 1.58 1.75 60 min 1.6 1.77 1.66 30 min 1.96 1.6 1.65

60 min 1.51 1.86 1.79 120 min 1.77 2.36 1.71 60 min 2.14 ~4 1.67

70

As Table 2.7 indicates, there is correlation between the MP removal and the AOS, and is within acceptable

rates in the majority of the cases. For UV/H2O2, the fast kinetics pose difficulties in the calculation, since

100% removal is easily achieved; if none or up to 2 values corresponding to 100% efficiency are inserted

in the calculations, then the correlation can increase up to 0.949, while the corresponding P-value drops

to 0.014. Nevertheless, the AOS can offer, under conditions, a predictive value on the fate of MPs during

treatment. The total organic load is correlated with the concentration of the MPs, so the main effort will

be now focused on explaining its degradation pathways, according to the AOP applied.

Table 2.7 – AOS percentile change and correlation with % of degradation.

UV % 0 min % 10 min % 30 min Pearson Correlation

AS 0 59 80 0.819

MBBR 0 62 89 P-Value

CF 0 57 77 0.024

UV/H2O2 % 0 min % 5 min % 10 min Pearson Correlation

AS 0 99 100 0.665-0.949


CF 0 84 95 0.103-0.014

Solar % 0 min % 30 min % 60 min Pearson Correlation

AS 0 8 17 0.766


CF 0 5 11 0.045

Fenton % 0 min % 60 min % 120 min Pearson Correlation

AS 0 16 26 0.822


CF 0 4 7 0.023

photo-Fenton % 0 min % 30 min % 60 min Pearson Correlation

AS 0 20 32 0.819


CF 0 12 19 0.871

71

2.4. Discussion

2.4.1. Degradation of micropollutants in wastewater: characteristics, influence and role of

the Effluent Organic Matter

In order to further elucidate the mechanism of MPs and organics’ degradation by the various processes

involved in this work, the contribution of the organic matter present must be evaluated, since significant

removal of Effluent Organic Matter (EfOM) (and therefore, reduction in COD and TOC) has been found in

the previous section.

The composition of EfOM was recently reviewed by Michael-Kordatou et al., (Michael-Kordatou et al.

2015) and its main substances are organic contaminants, other non-degradable compounds, or products

of the biological treatment (Lee et al. 2013). One part of the EfOM contained is attributed to refractory

compounds (Barker and Stuckey 1999, Jarusutthirak and Amy 2007). Humic-like materials are found, and

their presence is linked with their existence in drinking water sources (Jarusutthirak and Amy 2007) or

their transfer through the non-separating WW collection systems. The EfOM also contains a series of

substances linked with the biological activity of the microorganisms, known as soluble extracellular

polymeric substances (EPS) and soluble microbial products (SMP). These substances are usually a product

of i) microbial metabolism or ii) released during cell degradation (lysis) (Barker and Stuckey 1999,

Jarusutthirak and Amy 2007). Some characteristics related to these substances are their specific UV

absorbance index (SUVA: UV absorbance/DOC) which is an indication of their humic behavior

(Jarusutthirak and Amy 2007) but more importantly, they form the majority of the COD of aerobic

effluents.

However, during physicochemical treatment, i.e. CF in our case, the EPS and SMP fraction is expected to

be less important, since the bigger organic molecules are expected to prevail. Decrease of the negatively

charged colloidal particles is expected in high rates (Shon et al. 2006), but no significant transformation

due to the addition of the coagulants. The removal of the flocs by settling is involved instead.

Nevertheless, the original sources of humic substances (i.e. at least from drinking water) are expected to

exist; this indicates some common reaction pathways involved in water and wastewaters, which will be

further analyzed later.

From the observations above, a differentiation can be made in the EfOM, compared with the organic

matter present in natural waters (NOM). Wastewater contains sensitizers not encountered or found in

lesser amounts in natural waters (Ryan et al. 2011). In EfOM, among others, these substances are

expected to participate in the active fraction during the photochemical reactions and the advanced

oxidation processes applied, while the rest of the organic matter plays the role of the target. Literature

72

does not indicate any participation of the EPS or SMP in the photochemical reactions that resembles the

NOM action mode and therefore we make the previous working hypothesis. Under this condition, the

removal of the MPs, compared to the amount of carbon eliminated by this effect is merely a side reaction,

and will be treated as such; the MP degradation is in fact facilitated by the organic load reduction within

the matrix.

2.4.2. Pathways of MP degradation in secondary effluent

At this point, we separate the UV-based and the solar/Fenton processes and examine the degradation

pathways separately. An accompanying overview of the mechanism is presented in Figure 2.7.

2.4.2.1. MP degradation by UVC-based processes

In principle, the effect of UV-C irradiation at 254 nm can be separated in two major categories, the direct

and the indirect pathway.

1. Direct UV-C photolysis: the direct action of UV-C light on OM in wastewater. If we consider a

sample receiving UV-C irradiation, theoretically the EfOM plays the role of screening filter, but

experimental proof accounted a 10% decrease due to this effect (Ryan et al. 2011). Literature suggests

that the application of solar UV light exerts fragmentation of large conjugated compounds in smaller,

lower molecular weight ones (Parkinson et al. 2001); we expect that this phenomenon is taking place in a

certain extent under UV-C irradiation as well. These targets can be either the MPs (action 1) that present

UV-absorption peaks near the emitted wavelengths (small contribution in the total DOC), as well as the

ESP and SMP contained in WW (action 2). The action of UV against proteins, lipids and sugars involves

protein degradation and lipid peroxidation (Nasibi and Kalantari 2005), protein carboxylation (Krisko and

Radman 2010) etc. Also, the direct effect on humic acids is to be considered (action 3), causing their

polymerization and dissociation of covalent bonds present in the molecule (Sławińska et al. 2002). The

production of free organic radicals (action 4) and electrons (actions 5a&b) can also initiate indirect

mechanisms. As a matter of fact, the direct UV-C effect is never acting alone, but carries with it all the

indirect pathways, further analyzed afterwards; this also explains not only the parent MP elimination, but

also the mineralization taking place during sole UV-C treatment.

2. Indirect UV-C photolysis: the action of excited compounds against the same targets. It has been

reported that UV can electronically excite the photosensitizing molecules (probably in triplet state) and,

in presence of oxygen, cause photolysis through electron transfer to molecular oxygen (action 6) or to

form organic radicals (action 7), which then will react with oxygen (Legrini et al. 1993). Important

parameters that are involved in this pathway are the presence of oxygen (aerobic conditions), the specific

UV absorbance (SUVA) and the ratio of absorbance at 254 to 365 nm (Sharpless et al. 2014).

73

2.i. The energy transfer to oxygen causes the formation of (more commonly found as )

which is responsible for interference to activated double bonds and facilitate electron transfer. By itself,

it reacts with lipids, proteins and acids, as well as other cellular targets. By one-electron transfer, the

creation of superoxide radicals is induced (action 8), and since it is a relatively weak oxidant, its

participation is expected to contribute lightly in redox reactions with metals, such as iron or copper

(Buettner 2013). In turn, these can initiate Fenton-like reactions in presence of H2O2 (which will be further

analyzed later).

2. ii. By two-electron transfer (action 9), and/or the participation of superoxide radicals and water,

H2O2 can be generated (Buettner 2013). In organic-free waters, no H2O2 is accumulated, but in WW, there

was a constant presence of H2O2 in the bulk (Buchanan et al. 2006). The simultaneous presence of H2O2

and dissolved iron in our WW induces the initiation of a Fenton reaction. Also, the dissociation of H2O2

causes further hydroxyl radical production and elimination of the organic content in the bulk, as it has

been suggested before (De la Cruz et al. 2012).

The mineralization extent observed by UV-C alone was limited to 10%. However, if hydrogen peroxide is

added into the bulk from the beginning, the equilibrium and the importance of these pathways is limited,

since the homolytic disruption of the HO-OH bond in H2O2 (action 10) induces the production of the highly

oxidative hydroxyl radical (HO●) (De la Cruz et al. 2013). The hydroxyl radical non-selectively attacks the

organic matter present in the solution, actively eliminating the COD and increasing the TOC removal up

to 30%, from 10% which was before. Finally, if H2O2 is present, as in all of our matrices (from 0.9 to 1.9

ppm), taking into account the iron in complexes present in the effluents, Fenton reaction can work in

parallel, further inducing the radical generation and the subsequent organic matter removal (from 60-

70% to almost 100% COD removal).

2.4.2.2. MP degradation by solar, Fenton and (solar) photo-Fenton processes.

The three processes of solar exposure, Fenton and photo-Fenton are inter-connected, with the action of

solar light alone proposed as a possible basis to facilitate the degradation of EfOM due to the other two

processes as well. Solar light has been long identified to initiate photochemical reactions, in presence of

NOM (Canonica 2007). Since EfOM contains a part of NOM, NOM-like substances or substances with high

SUVA values (as index of humic-containing waters), the action mode of light in natural waters is

extrapolated in our WW effluents.

Compared to UV-C light, the direct effects of solar light (UVA and UVB) on MPs and EfOM are relatively

limited (action i). However, this contribution is in order of some ng/L carbon elimination, does not take

place with all pollutants and therefore could be neglected. This contribution in the degradation of MPs

involves their excitation to singlet-excited state, and through intersystem crossing, to triplet state (Ryan

74

et al. 2011). After this point, the MP can either form products through side reactions or reacts with oxygen,

returning to ground state.

To explain the high mineralization percent in our experiments, we will extrapolate the reaction taking

place in natural water towards our effluents. In natural waters, there is considerable photosensitization

of the organic matter, or more specifically, humic, fulvic acids and other substances (action ii). The

substances involved in this process are either autochthonous or allochthonous, comprising the

chromophoric dissolved organic matter (CDOM), which is mainly responsible of participating in indirect

photoreactions (Canonica 2007). In a lesser extent, the light absorption leads to mineralization of the OM

to inorganic compounds, due to a direct pathway involving its disintegration to smaller constituents (Gao

and Zepp 1998). The main reaction pathway includes the excitation of OM and as an end-product, the

reactive transients participate in energy and electron transfer, as well as free radical reactions (Gao and

Zepp 1998).

The main transient products of the photosensitization are the HO●, ●CO3, 1O2 and triplet states of DOM

(3DOM*) (Vione et al. 2014). The production of 1O2 is associated with the photosensitizing ability of DOM

(Mostafa and Rosario-Ortiz 2013). The action of the singlet oxygen was analytically presented before (now

action iii) and now only the HO● contribution will be now assessed further. The humic and fulvic fraction

of the DOM is producing HO●, as it absorbs light better than the smaller compounds and some of their

triplet states are able to oxidize superoxide or hydroperoxyl to hydroxyl radicals (Vione et al. 2014). All

transient species undergo various pathways afterwards, including the reaction with natural organic

matter, (bi)carbonates, dissolved oxygen and more (Vione et al. 2014).

Extrapolating to the WW in our experiments, since EfOM is a combination of autochthonous (bacterial

metabolism related) and allochthonous organic matter (humic/fulvic acids of drinking water) a similar

behavior is expected (Ryan et al. 2011). The active part EfOM is excited to a triplet state from where it can

react with dissolved compounds by energy/electron or hydrogen transfer (Vione et al. 2014), with the

targets stated before, and return to ground state or react with oxygen and follow a pathway similar to the

one of natural organic matter (Ryan et al. 2011). The low-molecular weight fractions are the ones mostly

responsible for generating the hydroxyl radical as well as the humic fraction, because of their

chromophoric abilities (Lee et al. 2013, Vione et al. 2014). Also, compared to natural waters, WW with

same DOC levels have been suggested to produce hydroxyl radicals more efficiently (Ryan et al. 2011);

higher energy triplet sensitizers may increase the ability of energy/electron transfer than the respective

NOM-induced ones. Consequently, the HO● produced (action iv) are added to double bonds or to aromatic

compounds, abstract H+ or cause electron transfer (Wenk et al. 2011).

75

Other routes that are responsible for generating ROS in WW include H2O2-dependent systems, which

demonstrate Fenton-like behavior (Vione et al. 2014). The contribution of these reactions can be

considered in our work, when H2O2 is added to the system; (organically complexed) iron is already present

and for the Fenton experiments an additional 2.5 mg/L Fe (II) was added to the system. In natural systems,

the presence of iron is known to catalyze the oxidation of organic matter, as well as participate in the

production of reactive transient species (action v) (H2O2, HO●, HO2/O2●─) (Voelker et al. 1997); our reagents

addition ensures the Fenton reaction in the system in the most readily available form. Here, in an analogy

with the natural system, light is catalyzing the ligand-to-metal charge transfer in the Fe-EfOM system,

forming iron (II) and polycarboxylate radicals, which in turn create free radical intermediates (action vi)

(Voelker et al. 1997). At this pH, the contribution of the heterogeneous Fenton system of the solid iron

oxides has to be mentioned, but its action is neglected when compared with the homogeneous system.

Finally, the last possible pathway of radicals’ formation in our system is through the participation of

nitrites (in lesser extent) and nitrates contained in WW (action vii). Their presence is correlated with the

denitrification step after the MBBR, and their contribution to HO● formation should not be overlooked,

and their presence has been suggested as crucial when comparing the radicals’ production between

natural water and WW (Lee et al. 2013, Ryan et al. 2011, Vione et al. 2014). In WW, as in nature, the

concentration of nitrites is less significant than nitrates, but their quantum yield is higher (Vione et al.

2014). The comparison among the AS and the MBBR WW reveals similar levels of nitrites (0.2-0.3 mg N/L),

but the main radical producer, NO3 is 15.3 mg N-NO3/L in MBBR waters compared to 2.3 mg N-NO3/L in

AS effluents (Margot et al. 2011).

The effect and the intensity of the various effects is visible in Figures 2.2-2.6 (b), by the modification of

COD and TOC. The difference among our treated WW types lies in the difference in their physicochemical

features and composition. Firstly, there are physical characteristics that hinder the degradation process,

such as suspended solids, the pH and the alkalinity. Most of the unfavorable conditions are met in CF WW

(higher pH and suspended solids); the extra iron present cannot compensate with the action induced by

the parallel Fenton reaction. Secondly, the organic matter present in CF effluent is 1.5-2 times higher than

the respective aerobic processes and in less favorable forms, when it comes to sensitization, since the

colloids (and other low-molecular weight substances) have been removed. The EfOM present in this WW

forms a significant filter as well as causes self-scavenging of the reactive species produced (Ortega-Gómez

et al. 2014). Nevertheless, in absolute values, the carbon elimination was significant, as the percentage

was lower but the initial value was much higher than the aerobically treated WW; as it seems, the indirect

pathways are enhanced by the presence of EfOM, due to the similar photo-sensible content but the higher

target availability in CF water. The contribution of solar light alone is worth mentioning, as well as the one

induced by UV-C. There is almost 10% mineralization in both cases, strengthening the significance of the

76

multi-level solar light-induced reactions in WW, while adding the Fenton reagents lead to

complementarily increased radicals production in all the systems.

Figure 2.7 – Overall mechanistic interpretation for the action of UVC and solar light within the effluent

wastewater (adapted from (De la Cruz et al. 2012)).

2.5. Conclusions

The unambiguous necessity to adopt advanced treatment methods in municipal WWTPs in order to tackle

the MP issue was addressed by the implementation of 5 treatment techniques. High MP removal was

thereby achieved. In addition, the reduction of EfOM was identified as the main mechanism of organics

removal.

Regardless the type of advanced treatment applied, a previous biological treatment stage (AS or MBBR)

seemed to be more appropriate than a physicochemical one (CF). Moreover, the MBBR effluents had

better physicochemical characteristics (lower alkalinity, TSS, COD and TOC concentration), rendering the

water matrix more suitable to apply a further advanced treatment.

From the five different treatment methods applied: UV-C irradiation, UV/H2O2, solar irradiation, Fenton

reaction, and photo-Fenton, only the UV-based methods removed 80% of the selected MPs, for the times

spans tested. Furthermore, after 30 min treatment, the degree of oxidation was very high in terms of COD

and TOC removal. Thus, high levels of mineralization of the organic compounds could be achieved with

this technique. For the case of solar light, Fenton and photo-Fenton treatment methods, the degradation

EfOM

3PhOM*

PhOM •-

O2

1O2Fe3+-EfOM

EfOM • ox

O2-• /HO2

•

HH2HH2OOOO2

• OH

• OHEfOM

EfOMox

OxOM & MP

OxOM & MP

MP MPmod

(i)

(ii)

(iv)

(vi)

(iii)

(v)+H+

OxOM ox

O2

Fe2+-EfOM

EfOM

UV-C

R•

ROO•

O2OxOM & MP

OxOM ox

UV-C

UV-C

Abbreviations

EfOM: Effluent Organic

Matter

PhOM: Photo-sensitizablefraction of EfOM

OxOM: Oxidizable fraction

of EfOM

MP: Micropollutant

(i)-(vii): solar-induced

pathways

(1)-(10): UVC-induced

pathways

(5a)

(4)

(1)

(5b)

PhOMmod

OxOMmod

PhOM

(2)

(3)

(6)

(6)

(7)

(8)

(9)

(10)(10)

(10)

(ii)

(ii)

OxOM

(vii)

NO2/NO3

OxOM ox

(iii)

77

rates are slower. The photo-Fenton treatment has slower MPs’ degradation rates than the UV-based

treatment methods. However, solar-based processes achieved a (small) degree of overall organics

removal which is not negligible at all. Extending the residence times in future studies is suggested, in order

to verify their efficiency to remove the 80% of MPs. In general, the order of degradation rates was the

following:

This study presented an overview of the impact of different advanced treatment methods, coupled to

three different previous secondary treatment techniques. Promising results were gathered from the

fifteen different combinations assessed for the removal of the selected MPs and the EfOM. Nevertheless,

further studies should be carried out in order to optimize these processes; the main issues are located in

reducing residence time, changing H2O2 dose and feeding method, for the UV/H2O2 methods, and

changing Fenton’s Reagent ratio or extending the treatment time for the photo-Fenton treatment.

79

3. Chapter 3 - Micropollutant degradation, bacterial

inactivation and regrowth risk in wastewater effluents:

influence of the secondary (pre)treatment on the efficiency

of Advanced Oxidation Processes

Published work:

Stefanos Giannakis, Margaux Voumard, Dominique Grandjean, Anoys Magnet, Luiz Felippe De Alencastro,

and César Pulgarin. "Micropollutant degradation, bacterial inactivation and regrowth risk in wastewater

effluents: Influence of the secondary (pre) treatment on the efficiency of Advanced Oxidation

Processes." Water Research 102 (2016): 505-515.

Web link:



Appendix B



80

3.1. Introduction

Throughout the years, urban wastewater treatment plants (WWTPs) have implemented strategies to

eliminate the organic load, chronologically followed by the inorganic one (phosphorus and nitrogen), and

presently, the current average treatment stops at the disinfection level. Chlorination was until some time

ago the most common disinfection method. However, its use has been connected with trihalomethane

(THM) production, a harmful disinfection by-product (DBP) of the reaction with organic matter (Krasner

et al. 2009). Hence, treatment at disinfection and decontamination level was turned towards safer

“greener” techniques (Michael et al. 2012).

Lately, ozone and ultraviolet light have been widely employed to tackle the issue of microorganism

elimination in wastewater (Drinan and Spellman 2012). UVC alone, combination with H2O2 and/or O3 are

some of the most studied and well-understood Advanced Oxidation Processes (AOPs) for this purpose.

The UVC-based processes have a well-established disinfection efficiency when applied in secondary

wastewater effluents (Rodríguez-Chueca et al. 2015), but the main concern of bacterial regrowth is yet to

be resolved. In overall, the AOPs gained supporters during the last two decades, mainly for their non-

selective character against organic matter and microorganisms (Moncayo-Lasso et al. 2012). However,

despite the interest gain, the limited number of full-scale applications engulfs the danger of over- or mal-

dimensioning of such units, since the pre-treatment process differs from plant to plant.

On the other hand, in less wealthy countries of the developing world, ozone and UV-based techniques are

far from applicable. Instead, the use of solar ponds has been widely applied (Von Sperling 2005), as a

simple, and quite efficient method of treating wastewater effluents. As this process has been successfully

applied, enhancing its performance with the photo-Fenton reagents could significantly increase the

removal of microbial and organic loads (Moncayo-Lasso et al. 2012). Iron and H2O2 are abundant and

environmentally safe, respectively, and the inactivation potential of photo-Fenton can improve the

effluent quality while reducing the residence times in such configurations (Von Sperling 2005) or in

Raceway Pond Reactors (RPRs) (Rivas et al. 2015).

In Switzerland, although special effort has been made to effectively remove (Giannakis et al. 2015c,

Margot et al. 2013, Margot et al. 2011), the recommended strategies have not included the

microorganism risk in the design, neither in the legislation. The upgrade of wastewater treatment plants

(WWTPs) affects the >10.000 inhabitant equivalent, thus leaving a large number of WWTPs without

disinfection units. For instance, the WWTP of Vidy (Lausanne, Switzerland) in its current reconstruction

planning, which focuses on the micropollutant removal after the secondary treatment of wastewater

(WW), involves the use of activated carbon, ozonation, followed by UVC light, but mostly for degrading

the by-products of the previous two installations.

81

In this work, we take advantage of the simultaneous presence of 3 parallel secondary treatment systems

of wastewater treatment in the plant of Vidy, in order to study the effect of secondary pretreatment on

the efficiency of AOPs. More specifically, wastewater from Activated Sludge, Moving Bed Bioreactors and

Coagulation-Flocculation units (with primary wastewater effluent as control) has been subjected to

various oxidation methods. The UVC alone, UVC/H2O2, Fenton, solar light and (solar) photo-Fenton

processes (namely UV-based and Fenton-related processes) were tested on the (immediate) bactericidal

removal efficiency, as well as the post-treatment regrowth. Finally, to put things into the real wastewater

context, the evolution of 8 micropollutants were monitored, and insights were given on the comparative

order of removal of micropollutants (MPs) and microorganism (MO) regrowth risks.


3.2.1. Collection of wastewater samples and treatment plant specifications

For the needs of the microbial testing, 6 sampling campaigns were performed. During each visit,

wastewater from the following points was collected: i) before secondary treatment (after primary

decantation) (PT), ii) after secondary treatment by activated sludge and secondary clarification (AS), iii)

after secondary treatment by moving bed bioreactors (MBBR) and iv) after physicochemical treatment by

coagulation-flocculation (CF). The aforementioned points can be found in Figure 3.1. Each time, a 5-L grab

sample was collected and transported immediately to the laboratory for treatment. For the

micropollutants, the strategy has been analyzed in a previous work (Giannakis et al. 2015c).

Figure 3.1 – Schematic representation of the WWTP of Vidy, Lausanne (VD, Switzerland) and the

sampling points for this research.

82

The AS unit has a hydraulic retention time (HRT) of 4h and sludge retention time (SRT) of 2 days,

approximately, without nitrification. The MBBR capacity is only 5% compared to the AS unit, but has a

longer retention time and includes a nitrification step. Finally the CF unit is based on chemical coagulation

and flocculation process by FeCl3 as coagulant. More details on the different units can be found in Margot

et al. (Margot et al. 2013).

3.2.2. Employed chemicals and reagents

For the experiments of bacterial inactivation H2O2 30%, FeSO4.7H2O and Titanium (IV) oxysulfate (for H2O2

determination) was acquired from Sigma-Aldrich (Switzerland) and NaHSO3 for H2O2 elimination from

Sigma-Aldrich and Acros Organics, respectively. Finally, the plate count agar (PCA) was purchased from

Sigma-Aldrich (Switzerland).

3.2.3. Experimental set-up: reactors and apparatus

The experiments are divided in two main groups, namely the UV-based and the Fenton-related ones. For

the UV-based experiments, two double-wall, water-jacketed merry-go-round reactors were used in

parallel, for the UVC and UVC/H2O2 experiments, respectively. The water recirculating in the glass reactors

was controlled at 22°C (for protection of the UVC equipment). UVC light was provided by 35-W low

pressure UVC lamps (Model: UVI 40 4C P 15/300), with an emission of 350 μW/cm2, acquired from UV-

Technik Speziallampen (see Supplementary Figure S1 for the reactor scheme). Among the Fenton-related

experiments, the solar only and solar-assisted photo-Fenton process were performed in 100-mL Pyrex

glass reactors, placed on magnetic stirrers and constantly agitated by magnetic bars (300 rpm). The Fenton

experiment took place in shaded reactors, in the dark, whereas the solar and the photo-Fenton

experiments took place in a solar simulator (Atlas, Suntest CPS+). This artificial solar light source was set

at 900 W/m2 global irradiance (~0.5% UVB, ~5% UVA and ~95% visible light) and has been analytically

presented in previous works, e.g (Giannakis et al. 2015b). The solar UV intensity and global irradiance was

monitored with a coupled CM3 – CUV3 UV radiometer and pyranometer (Kipp & Zonen, Netherlands).

3.2.4. Application of AOPs: details and specifications

A summary of the conditions is given in Supplementary Table S1, regarding the experimental times and

reagents addition. For each group of experiments, an analytical presentation follows.

3.2.4.1. UV-based experiments

For the sole UVC experiments, 300 mL of wastewater were added in the reactors and exposed to the

irradiation. In the case of UVC/H2O2 experiments, H2O2 was also added from a stock solution to reach the

desired initial H2O2 amount. The pH was monitored and the samples were immediately analyzed. Samples

were also kept in order to assess bacterial regrowth after the experiments.

83

3.2.4.2. Fenton-related experiments

For the solar tests, 100 mL of wastewater were inserted in the reactors and placed in the Suntest for

simulation of solar exposure. The temperature of the water never exceeded 35°C. The remaining tests

were performed with the addition of Fe2+ and H2O2, either in the dark (Fenton experiments) or under light

in similar conditions (photo-Fenton experiments). In all cases, samples were also kept for regrowth tests.

3.2.5. Analytical methods, physicochemical and microbiological parameters

3.2.5.1. Analytical methods

The micropollutants presence in wastewater was followed by a UPLC/MS-MS method, used in previous

works (De la Cruz et al. 2013, De la Cruz et al. 2012) and analytically presented in (Giannakis et al. 2015c).

The removal percentage of the 8 micropollutants monitored in this study (Carbamazepine, Diclofenac,

Atenolol, Metoprolol, Venlafaxine, Clarithromycin, Benzotriazole and Mecoprop) was calculated by

weighted arithmetic mean, as follows:

(3.1)

In the graphs, the overall removal X% is given, wi is the micropollutant amount (mol) and xi the removal

percentage of each micropollutant (xi: i=1-8). Finally, we note that the work was effectuated by the

inherent micropollutant content of the collected wastewater; no spiking took place prior to testing.

3.2.5.2. Physicochemical parameters

For the pH monitoring throughout the experiments a Mettler Toledo pH meter was used. A UV-vis Lamda

20 spectrophotometer was used to monitor the evolution of H2O2, followed by the (modified) DIN method

38 402 H15 and the dissolved iron by the Ferrozine method (De la Cruz et al. 2012). The principal

physicochemical characteristics of the wastewater used in this work are summarized in Table 3.1.

Table 3.1 – Basic physicochemical and optical characteristics of the wastewater used in this study

(own measurements and (aMargot et al. 2013, bMargot et al. 2011)).

Parameter Unit Wastewater previously treated with Primary Activated Moving Bed Coagulation

Treatment Sludge Bioreactor Flocculation pH - 7.8-8 7.3-7.8 6.6-7.4 7.3-7.9

TOC mg/L 109.1±25.6 28.08±12.62 14.615±7.9 68.47±15.94 DOC (0.45μm) mg/L 89.32±26.42 20.40±1.41 6.39±1.71 29.98±0.88

COD mg/L 200±19 51±10 20±11 85±5 Alkalinity mg CaCO3/L 282.5±15 230±35 95±10 240±10

TSSa,b mg/L 35±10.4 12.1±2.8 14.2±1.4 28.5±5.7 Total Solidsa,b mg/L 59.5±5.4 54.6±4.1 25.5±4.3 58.2±3.3

Total iron mg Fe/L 2.5±0.55 0.95±0.05 1.75±0.15 5.5±1

84

Dissolved iron mg Fe/L 0.25±0.05 0.04±0.002 0.02±0.001 0.33±0.06 UVAT % 37.78±5.78 87.58±4.95 86.10±5.51 62.22±7.09 UVBT % 22.45±3.32 75.77±2.24 74.07±1.57 44.45±4.09 UVCT % 11.42±0.97 65.59±2.34 67.47±3.23 31.27±6.55

3.2.5.3. Microbiological parameters

The microbial content of wastewater was monitored by cultivation on a non-selective agar (PCA). This

medium allows the enumeration of E. coli, Bacillus subtilis, Listeria monocytogenes, Staphylococcus

aureus and other bacteria contained in water and dairy matrices. The microorganism recovery rate, after

24h incubation at 35°C, is above 70% for a 1.000-100.000 inoculum (initial population). In such manner,

the disinfecting capabilities of the AOPs tested shall not be limited by selectivity on the medium and our

estimation will have the minimum error possible, given the particularities of the cultivation techniques

for inactivation estimation (Michael et al. 2012). The initial CFU/mL in the working matrices was 6.5x105,

6.5x103, 3.5x103 and 4x105, for PT, AS, MBBR and CF, respectively. No microorganisms were spiked in the

wastewater samples; the work was realized by the indigenous population. Samples were drawn at the

respective points (~ 5 mL) and were plated immediately. Appropriate dilutions were done, to achieve 15-

150 colonies per plate. Experiments were done in duplicates, in the different sample campaigns and at

least double plating was performed, enumerating 3 consecutive dilutions. Therefore, the average value is

presented in all the figures, subjected to <5% (max. 10% in few cases) standard deviation; measurements

with higher errors were not considered and error bars are not shown for clarity.

3.3. Results

3.3.1. Micropollutant elimination in the selected wastewater effluents

Figure 3.2 summarizes the degradation of the 8 selected organic micropollutants present in the effluents

of the secondary treatment facilities in the WWTP of Vidy. All the selected AOPs were tested (UVC,

UVC/H2O2, Solar, Fenton, photo-Fenton) against 8 MPs in WW (Carbamazepine, Diclofenac, Atenolol,

Metoprolol, Venlafaxine, Clarithromycin, Benzotriazole, Mecoprop), after treatment by AS, MBBR or CF.

85

Figure 3.2 – Micropollutants’ degradation by AOPs after secondary treatment. a) UV and UV/H2O2

processes. b) Fenton, solar and photo-Fenton process. AS: blue trace, MBBR: red trace, CF: green trace.

Continuous lines and colored symbols show the measured evolution of the experiment, while the

dashed lines and open symbols indicate the projection of the experiment according to the measured

first order degradation rate constant.

In a previous study with 6 micropollutants, we have correlated the degradation of micropollutants with

the pretreatment method and the subsequent AOP (Giannakis et al. 2015c) with an order of increasing

efficiency of CF < AS < MBBR, for the pre-treatment methods and k (UV) < k (UV/H2O2) and k (Fenton) < k (solar

irradiation) < k (photo-Fenton), for the AOPs tested. This consistent behavior is also observed here. Although the

list of MPs has widened, including both photo-sensitive and “resistant” compounds, the effluents of MBBR

again facilitated the highest MP removal. More specifically, the micropollutant removal efficiency

dropped as we move from MBBR to AS and CF, in all types of AOP applied, attributed to the COD and DOC

content of the respective effluents. Furthermore, as far as the comparison among the tested AOPs is

concerned, only a relative order can be established, as the processes differ significantly. UVC/H2O2 and its

massive HO● production has a profound effect in MP elimination (2-10 min), whereas UVC alone requires

a mere 35-40 min for total removal of the parent compounds. The difference among the times in each

process is a function of the pre-treatment. These processes achieved also a very high degree of

mineralization (Giannakis et al. 2015c), therefore their application holds high potential.

On the other hand, solar, Fenton and photo-Fenton processes would require much longer times to fully

degrade the MPs. After 1 or 2h, when our experiments stopped and the apparent first order k constant

was calculated, a maximum of 50% elimination took place, in the photo-Fenton reaction after 1h in MBBR

effluents. The apparent k indicates minimum 2-h exposure to achieve 100% removal. As at the end of our

exposure period more than 75% of the oxidants was available (data not shown, available (Giannakis et al.

2015c), the assumption that the process would continue uniformly for an additional period holds true.

86

Furthermore, an interesting observation takes place if the synergy among the constituents of the photo-

Fenton reaction is concerned. As it appears, the photo-Fenton process is more efficient than simply the

sum of its parts. The Synergy (S) is higher in the processes that were hampered in either way, solar or

organic load (see Supplementary Table S2). The reactivity of the organic pollutants with solar light is

limited and the enhancement of the process with the HO● generation could result in efficient removal in

relatively reasonable retention time.

3.3.2. Microorganism elimination in the different wastewater effluents, per AOP:

inactivation and post-treatment regrowth

3.3.2.1. UVC and UVC/H2O2 processes

Figure 3.3 showcases the bacterial inactivation after the application of UVC, with or without the addition

of H2O2. The effluents of MBBR, AS, CF and PT were exposed to monochromatic UVC irradiation (Figure

3a) or UVC and 20 mg/L H2O2. When UVC was applied alone, two distinct categories of kinetics were

observed, in either MBBR and AS effluents or in CF and PT. In MBBR and AS effluents the time necessary

to completely inactivate bacteria was similar (<1 min actual difference) while the 4-log reduction in CF

and PT effluents delayed significantly (10 min). The interpretation lies in the Achilles’ heel of UVC

applications. The two major drawbacks are the suspended solids and the organic content in the effluents.

The solids, apart from their physical barrier effect, they effectively shield microorganisms, and favor

aggregation. Hence, as the efficiency of the UVC irradiation relies in the transmittance of the medium, the

two effluents with lower SS content and therefore higher transmittance (UVCT in Table 3.1) presented the

lowest treatment times.

87

Figure 3.3 – UV-based disinfection and respective regrowth of the indigenous bacterial population

after 24 h. A) UVC irradiation alone. B) UV/H2O2 process (20 ppm initial H2O2 addition). The shaded

part and the dashed lines symbolize the dark storage and regrowth after treatment, for 24 h.

The combined UVC/H2O2 process builds on the principles of the UVC disinfection, plus the homolytic

disruption of H2O2. The mechanism of UVC/H2O2 inactivation, is both external and internal. When H2O2 is

added into the bulk, the HO● attacks improve disinfection in all effluents; till now, there is no

microorganism found with resistance to the oxidation by HO●, contrary to damages by UVC irradiation

which is repaired instantly (Sinha and Häder 2002). For the aforementioned reasons, we naturally observe

decrease in all necessary treatment times in Figure 3.3b. The major enhancement observed in the CF and

PT effluents lies in the dependence of the disinfection efficiency in the transmission of UVC light. MBBR

and AS effluents contain significantly low organic content, which acts as a HO● scavenger (Ortega-Gómez

et al. 2014), and low SS; the improvement is of some minutes and the HO●-induced oxidation was proven

to affect the overall carbon content of the matrix (Giannakis et al. 2015c). In CF and PT effluents, since

the transmittance levels are lower, the benefit of the oxidative action was higher, with the HO● attacks

compensating the lower inactivation rates demonstrated when UVC alone is applied.

Finally, it should be noted that the wastewater effluents are far more complex to be dissociated only to

solids and organic scavenging, as many actors influence the degradation efficiency. For instance,

phosphorus moderately scavenges the radicals generated (Wu and Linden 2010) and UVC is absorbed by

nitrates, resulting to nitrite. Nitrite reacts with hydroxyl radicals, producing nitrite radicals, which are far

less oxidative, but on the other hand are more long-living (Vione et al. 2014). Margot et al. (Margot et al.

2011) measured high values of ions in MBBR effluents, due to the nitrification step, and their

participation cannot be overlooked. Finally, another dual mechanism is the scavenging of HO● by

carbonates (Carra et al. 2014), or the generation of carbonate radicals (Wu and Linden 2010). The high

alkalinity measured in AS and CF is an indication of high (bi)carbonate content, partially explained by the

lower pH values of the MBBR effluents, and the expected scavenging effect is higher than the beneficial

impact of the oxidation by the mildly oxidative carbonate radicals.

The shaded part of Figure 3.3 presents the regrowth assays performed for the UV-based processes. When

UVC irradiation is applied, the damage is mainly at the genome level, due to the high absorption by the

thymine and cytosine bases. The result is cyclobutane–pyrimidine dimers (CPDs), 6–4 photoproducts (6–

4PPs) plus their Dewar valence isomers (Douki et al. 2003). Within minutes of UVC exposure, the stress

induces responses of chaperones to repair the DNA damages, but soon this response is surpassed. As the

storage took place in the dark, the presented results are exclusively an effect of the base or nucleotide

excision repair, the mutagenic repair and other similar mechanisms (Sinha and Häder 2002). Similarly to

UVC disinfection, the regrowth is influenced by the dose of UVC received by the microorganisms. Hence,

88

for similar treatment times, the different effluents result in a regrowth of increasing rate. After treatment,

the microorganisms that are present in the bulk can be categorized in some states, such as healthy (will

grow in WW), injured (which can regrow), apoptotic (which cannot regrow) and dead ones. The regrowth

presented here is the result of growth or regrowth from the first two categories. What is changing for

instance in MBBR effluents among 2.5 and 5 min, since the number of viable counts is the same, is the

regrowth potential and the shift towards the two last states. Finally, the differentiation among the various

statuses of cell conditions is more evident in CF and PT effluents. These received low UVC doses and the

higher final populations led to significant regrowth.

Concerning the UVC/H2O2 inactivation, the baseline damage is the UVC (common to the previous) adding

the oxidative action of HO●. If the damage was solely attributed to UVC, higher regrowth would be

demonstrated. The CF and PT effluents presented high regrowth rates, since they received low UVC doses

and suffered from high HO● scavenging. For the AS and MBBR effluents, when the inactivation process

was complete, no detectable regrowth occurred. The internal and external oxidative stresses posed by

HO● rendered the microorganisms unable to repair their lesions. Also, compared to the UVC alone, the

proportionality of residual concentration and regrown population was lower, attributed to the different

pathway brought by HO●. Finally, even for the non-favorable conditions met in CF and PT effluents, looking

the experimental process in retrospect, the regrowth was eliminated when the treatment was prolonged

beyond the necessary time for total inactivation. Hence in a field application, increasing the UVC dose

above the threshold for inactivation may ensure limitation of bacterial regrowth.

3.3.2.2. Fenton, solar and photo-Fenton processes

Figure 3.4 presents the results obtained when the Fenton reaction, solar light or the (solar-assisted)

photo-Fenton reaction were the inactivation driving forces. Concerning the Fenton reagents (Figure 3.4a),

H2O2 causes minor oxidation in outer cell wall layers, but also diffuses into the cell. The amounts of H2O2

used in this study are not bactericidal per se, and the modest inactivation achieved was the result of the

external Fenton reaction, by the addition of Fe2+ from our side, resulting to a massive oxidative wave. The

regeneration of ferric iron back to ferrous is very slow and difficult to recover since the Fe3+-dissolved

organic matter (DOM) complexes are very stable (Hakala et al. 2009); hence our low inactivation rates.

When H2O2 was doubled, the efficiency did not increase importantly. The HO● was partially scavenged by

the organic matter present in the solution. A positive effect though lies in the complexation of iron by

DOM after its conversion to Fe3+ and its reduced precipitation (Hakala et al. 2009). All effluents were in

the near-neutral region, in which the Fe3+ aqua-complexes are not soluble (Ruales-Lonfat et al. 2015).

Concerning the precipitated iron in the form of iron oxides, a small contribution of the heterogeneous

Fenton reaction is possible, with the iron oxides acting as iron source (Ruales-Lonfat et al. 2015), since all

effluents contain iron in lesser or higher amounts (Table 3.1).

89

Figure 3.4 – Fenton-related disinfection and respective regrowth after 24 h. A) Fenton process (2:10

ppm initial Fe2+/H2O2 addition). B) bare solar light. C) photo-Fenton process (2:10 10 ppm initial

Fe2+/H2O2 addition). The shaded part and the dashed lines symbolize the dark storage and regrowth4

after treatment, for 24 h.

The exposure of the samples to solar only irradiation (Figure 3.4b) caused the bacterial inactivation in a

dual manner: UVB is documented to cause similar dimerization to the UVC-inflicted one, but in lower rates

(Cadet et al. 2005). UVA on the other hand initiates a chain of reactions in the bacteria, among which,

internal oxidative (Fenton-like) events are the most crucial (Imlay 2003, Spuhler et al. 2010). As it appears,

solar light demonstrated aspects similar to an indirect AOP. The nature of these actions initiated by UVB

and UVA explain the initial lag period (Giannakis et al. 2014b) visible in all graphs of Figure 3.4b and the

accumulation of this damage actually resulted to bacterial inactivation. UVB and UVA wavelengths are

also subjected to physical blocking and scattering by the suspended solids, and therefore their efficiency

was a function of the turbidity of the samples, as shown in Table 3.1.

As far as the photo-Fenton process is concerned (Figure 3.4c), within the time-frame studied, the

inactivation for any effluent was dramatically increased. In a recent review, we described the mechanism

of bacterial inactivation in water and wastewater (Giannakis et al. 2016a, Giannakis et al. 2016b). Briefly,

iron can complex with the organic matter and under light, a ligand-to-metal charge transfer (LMCT) results

to the sacrificial oxidation of the organic ligand (producing organic ligand radicals) and more importantly,

the regeneration of Fe3+ to Fe2+ (Spuhler et al. 2010). Furthermore, the simultaneous presence of iron,

DOM and solar light initiates photochemical cycles which can result in the production of ROS, such as HO●

and H2O2 (Canonica 2007, Ng et al. 2014). This cycle and its potential implications in wastewater effluents

(EfOM) has been previously analyzed (Giannakis et al. 2015c). Other sources of ROS are the nitrates and

carbonates (Vione et al. 2014, Wu and Linden 2010). As a result of the aforementioned actions, significant

microorganism removal took place in MBBR and AS effluents. In 3h the disinfection process was (almost)

complete, while even in CF and PT effluents the removal was important. The combined, multi-level

90

damage inflicted by the photo-Fenton process can also be seen in the synergy calculated for these

experiments (see Supplementary Table S3).

The shaded part of Figure 3.4a summarizes the Fenton-related techniques regrowth. The Fenton process

was not efficient in inactivating the microorganisms present in the effluents. Only 1-2 log units of

inactivation represented the highest amount of bacteria removed from the bulk. Nevertheless, the

residual H2O2 and the iron present in solution (complexed or dissolved) efficiently continued the Fenton

process during the dark storage period. Furthermore, a bacteriostatic behavior is observed in all but the

PT effluents. As discussed before, this is a result of the excess of targets present in the solution. Also, the

final concentration of bacteria after 24 h was correlated with the number of leftover microorganisms at

the end of the “observation” period (since the same action continues for the whole 24 h). This fact is

promising, since this process could be used for suppressing any regrowth risk in treated wastewater,

during storage and before potential reuse.

Solar disinfection and its regrowth was previously studied by our group in synthetic, non-turbid effluents

(Giannakis et al. 2015b), where we correlated the exposure to solar irradiation with the regrowth, and

were able to identify the necessary solar dose for shifting from (post-irradiation) growth to a deterministic

decay phase. In Figure 4b, in all effluents, after 5 h of treatment, regrowth is demonstrated. Only after 6h

of exposure in MBBR effluents a bacteriostatic effect is observed. The main difference in the present work

is the presence of particles in WW which effectively aggregate and shield the microorganisms, thus

minimizing the received solar dose. Furthermore, the action mode of solar irradiation seems to be heavily

influenced by the UVA irradiation. If UVB was the main driving force, enhanced regrowth would be

demonstrated. In MBBR effluents, the bacteriostatic/decay phase observed after the exposure is an

indication of the internal oxidative damage occurring in the cell (Imlay 2003). Finally, the contribution of

the photo-sensitizable organic matter could influence the inactivation, by the production of HO● and H2O2

(Canonica 2007); iron is present in the effluents (especially in CF, Table 3.1), which in combination with

the small H2O2 accumulation could hamper the regrowth in the dark.

The application of solar photo-Fenton combines the benefits presented in the previous two processes.

Firstly, a decay phase is observed for the end-treatment phase in both AS and MBBR effluents, reaching

total elimination during the post-irradiation period. Most probably, the simultaneous exposure to solar

UV and the Fenton reagents limits the potential of recovery for microorganisms; this effect hinders their

growth in the minimal nutritional contents and the oxidative stress present in the WW effluents which

still contain H2O2 and iron (Giannakis et al. 2015c). The low organic content is probably the key point in

managing to effectively stop the growth. The PT effluents on the other hand still possess a great number

of healthy cells which continue to thrive. In conclusion, the short times of photo-Fenton by the exposure

under light and the continuous Fenton action have the best overall effect in inactivation and regrowth

91

inhibition, which makes it a potentially attractive solution for WW treatment; possibly, if higher Fenton

reagent’s concentration was added to reduce the observed residence time, the process could be more

competitive.

3.4. Discussion

3.4.1. The major threat and treatment focus: micropollutants or microorganisms?

The previous chapters have dealt with the isolated problem of establishing proper disinfection or

decontamination in the wastewater effluents. Nevertheless, since the co-existence of MPs and MOs is the

actual situation, in this part we perform a combined approach and try to assess the treatment strategies

for ensuring both adequate MP removal and MO elimination, while suppressing post-treatment bacterial

regrowth. As thresholds, we have chosen the 3-log inactivation of microorganisms as minimum removal

(3 to 4-log for most reuse purposes; for instance (Liberti et al. 2003)) and the Swiss legislation limits for

micropollutants (80% elimination (Giannakis et al. 2015c)).

Figure 3.5 – UV-based disinfection and decontamination. A) UVC irradiation alone. B) UVC/H2O2

process. The lines indicate the microorganism inactivation, while the bars the micropollutant

degradation (%). The circles indicate the regrowth suppression points with the respective colors

indicating the secondary treatment method, while the horizontal lines indicate the minimal

micropollutant (brown line) and microorganism removal (orange line).

Figure 3.5 (a and b) presents in parallel the disinfection and the decontamination processes of the UV-

based processes. Concerning the UVC irradiation alone, we have noted the efficiency in the previous

chapters, and >99.9% MO removal can be achieved after 5 min of treatment. The immediate MP

92

degradation does not satisfy the 80% threshold before 30 min of exposure except for the MBBR effluents.

Therefore, one could suggest that MOs are easier to remove and the design of the process has to be done

taking the MPs as reference. When bacterial regrowth is taken into consideration, different times are

reported to minimize this risk. For MBBR and AS effluents, 5 and 10 min were necessary, respectively. In

our samples, regrowth suppression was not achieved in CF effluents before 10 minutes but extrapolating

from the MBBR and AS effluents, in 30 min it is safe to believe that it would be no longer possible to

contain healthy or repairable microorganisms. Even so, 30 min of exposure is the threshold for MPs

removal and therefore, indeed the design needs to be implemented according to the MPs; their removal

means an already achieved acceptable disinfection level.

For the UV/H2O2 process, the necessary times were decreased. Only 2.5 and 5 min of treatment were

found necessary for MBBR and AS (and CF) effluents, respectively. The regrowth was completely

suppressed after 2.5, 5 and 10 min for MBBR, AS and CF as well. Comparing the necessary time for

minimum 3-log removal and regrowth inhibition, we observe that MPs were removed at least at 85% (up

to 99%) for all effluents. This means that if the design takes place for the minimal MP removal, then

effluents that have been subjected to biological treatment are also microbiologically safe. Only for the

physicochemical process a prolongation would be necessary, and a higher (but not complete) MP removal

was attained. If we take into account the higher oxidative conditions used for MPs removal (25 ppm H2O2

vs. 20 ppm and 5 ppm Fe2+ vs. 2 ppm), the micropollutants should definitely be the primary target.

Therefore, for both the UV-based processes, the reference for micro-contaminant and microbiological

safety is the removal of MPs.

For the Fenton and solar processes presented in Figure 3.6a, and 3.6b neither MPs nor MOs have been

successfully removed; after 6 h, moderate removal was achieved. In the Fenton process, if the k constant

remains quasi-linear, around 6 h would be necessary for efficient MP removal, therefore the exposure

times would be similar. After these exposure/treatment times, regrowth should not be an issue; after 6 h

of solar treatment, bacteria in MBBR effluents demonstrated stationary behavior and when the Fenton

process was used, the presence of the reagents efficiently suppressed regrowth.

93

Figure 3.6 – Fenton-related disinfection and respective regrowth after 24 h. A) Fenton process. B) Bare

solar light. C) photo-Fenton process. The lines indicate the microorganism inactivation, while the bars

the micropollutant degradation (%). The circles indicate the regrowth suppression points with the

respective colors indicating the secondary treatment method, while the horizontal lines indicate the

minimal micropollutant (brown line) and microorganism removal (orange line).

On the other hand, the photo-Fenton process shows (Figure 3.6c) promising potential, as the order of

magnitude of bacterial decay is reduced to the hour range. More specifically, after 2 h of treatment the

MBBR and AS effluents are almost bacteriologically safe (99 and 97% removal, respectively) and above

the 80% threshold for MPs. Furthermore, after 3 h, apart from the attained disinfection, regrowth is

suppressed for biologically treated effluents and almost for all effluents decontamination is achieved

(~79% for CF). These results demonstrate that for the Fenton-related processes, a shift in the importance

takes place and bacteriological targets are gaining importance over the organic contaminant ones. We

believe that if similar oxidative conditions were to be used (higher were used in MPs experiments), after

2 h, complete disinfection could have been achieved and the only open question should be regrowth.

Nevertheless, the results indicate with a high level of certainty that MOs inactivation and regrowth could

be taken as reference, since >80% MP removal is estimated to precede the disinfection events.

3.4.2. Common events and dissimilarities in the treatment of different targets in

secondary effluents

Although we cannot possibly compare the two families of AOPs applied, i.e. UV-based and Fenton-related

ones, a separation can be made according to the effluents. MBBR effluents always allowed degradation

of the targets faster than AS, CF, and PT, in order of decreasing efficiency. However, an intriguing

controversy among the investigated effluents is the shift in the order of importance, when the family of

AOPs used is modified. When the UV-based technologies were applied, the MPs were the reference

target, while, in the Fenton-related processes the MOs disinfection and their regrowth risk are more

important. Even though this difference is marginal, it contradicts the clear importance of the MPs in the

UV-based processes (and MOs for the Fenton-related) and at this point we will try to assess the main

parameters behind this issue.

1) The nature of the target plays a key role.

The micropollutants are complex, high molecular weight (HMW) structures with the ones studied here

almost all exceeding the 200 g/mol. Sensitive LC-MS analyses have identified that the various attacks

against the molecules can be either photonic (UVC induced damage) or ROS-related. These mechanisms

can result to ipso, para, meta attacks, -OH addition, halogen scissions etc. for UVC light, while ketone

formation, decarboxylation, side chain breakage, H+ abstraction (and other) have been found for the HO●-

94

induced attacks (Zhao et al. 2014). Even a moderate structural modification, such as short side-chain

breakage induced by e.g. a radical attack, modifies the micropollutant significantly, and the by-products,

even those structurally similar to the parent compound, are no more considered for the calculation of the

degradation percentage. However, these by-products continue to get degraded in the bulk, but compared

to the carbon content of the wastewater, their amount is negligible (Giannakis et al. 2015c).

Microorganisms on the other hand, seem to require significantly more attacks in order to get inactivated.

The influence of the light or the radicals in key structures (e.g. lipids in cell membrane) has a mitigated

effect, leaving the microorganism viable. As this damage progresses, the result could be bacterial death.

It is an intriguing question however, the effect of each process on bacterial viability, i.e. quantify the UVC

photons required to inactivate a microorganism, hinder its regrowth, or the ROS-induced attacks

necessary for membrane rupture etc. For the HO● effect alone and the necessary amount to inactivate a

single E. coli it has been found to be in the order of 109 (Marugán et al. 2008).

2) The heterogeneity of targets (number and size) acts in a dual manner.

The amount of each micropollutant in WW effluents is found most frequently in the ng/L and low μg/L

range. If a theoretical sum of all the contaminants of emerging concern is assumed at μg level, then in our

experiments, a g/mol is estimated and therefore at 300 mL of sample we get around of 1015

“pollutant targets”. Similarly, the maximum amount of bacteria encountered was of 105 CFU/mL, hence

~108 CFU are expected. It becomes clear that the micropollutants are in order of millions more than the

bacteria. However, the size of a molecule is Å to nm scale, while bacteria range around the μm range;

hence, the difference is in the order of thousands. All things considered, although there are far more MP

“molecules” to degrade, since the HO● radicals are limited by diffusion, it is probably easier to inflict

attacks on a big and rare-to-find microorganism, rather than reach a small and abundant pollutant.

Furthermore, the assumptions made in this study have to be re-considered when experimenting with

urban WW. For instance, bacteria are not the sole microorganism present (especially the monitored ones),

although our medium has optimal recovery for the spread plate method. Viruses, protozoa and others

compete for the oxidants generated, however in lesser extent than the EfOM. Also for MPs, we chose a

list of contaminants relevant to the research trends, but among the list, there are contaminants with

limited light reactivity (e.g. Atenolol) or high reaction rate constants with HO● (e.g. Diclofenac). As a result,

the percentage of removal is a relative measurement, and in future studies, the wider the list of

contaminants, the better approximation will be achieved.

3) Technical aspects related with the application of AOPs are implicated.

This study is a bench-scale simulation of a field application, using real WW effluents, but employing a

certain reactor geometry. The experiments were performed under a smaller optical path than the

95

dimensions of, for instance, a UVC-based plant. This fact explains mainly the Figure 2.2 and the MP

degradation results, where although a linear change is observed, the difference among the pre-treated

effluents and the effect on degradation is lower than the respective one in microorganisms. Another

example is the efficiency of the UVC process after primary treatment only. Indeed, for a short exposure

time, with a short optical path, the difference is mitigated, but in a real application, the scale-up problems,

the effect of suspended solids, the cleaning of the quartz sleeves due to particle settling and

polymerization would make this application impossible. As UVC exposure requires low SS and high UVCT

waters (min. 35%), the PT effluents are rejected and the CF ones would be marginally accepted (see Table

1) to be treated by such a system. Therefore, although in some cases the results do not justify it

completely, a biological pre-treatment is necessary before AOPs application for disinfection.

4) Effluent characterization complexity hides the mode of action of the different AOPs.

Another controversial point comes at the attribution of the effect of AOPs to each effluents, as a direct

effect from the physicochemical characteristics presented in Table 3.1. As found for both microorganisms

and micropollutants, the order of degradation in all cases is MBBR > AS> CF > PT, for all AOPs used. One

could dissociate the two groups, of MBBR with AS (biologically treated effluents) and the CF with PT

(physicochemical processes). In the first group, the effluents present similar light transmittance, and

therefore, their optical-based effects are quite similar. The AS has higher alkalinity and organics content,

so the disinfection is slightly hindered. In the second group, the effluents present similar SS content, with

the CF effluents having been through a physicochemical removal of smaller (colloidal) particles.

Table 3.2 – Photochemical characteristics of the various effluents

Index PT AS MBBR CF

E2 (254 nm) 0.430±0.16 0.147±0.04 0.109±0.04 0.391±0.22 E3 (365 nm) 0.084±0.05 0.017±0.01 0.013±0.01 0.074±0.06

E2:E3 5.432±1.03 14.545±14.18 24.386±29.15 5.935±1.68

E4 (465 nm) 0.040±0.03 0.006±0.01 0.005±0.01 0.008±0.01 E6 (665 nm) 0.021±0.01 0.004±0.01 0.004±0.01 0.004±0.01

E4:E6 1.909±0.13 0.625±0.88 0.563±0.8 0.813±1.15

ε280 0.327±0.10 0.121±0.03 0.091±0.03 0.171±0.01 SUVA (DOC/E2) 273.15±15.88 168.59±3.21 70.22±4.66 178.32±5.52

Slope 275-295 nm 0.00250 0.00130 0.00055 0.00275 Slope 350-450 nm 0.00055 0.00015 0.00012 0.00051

SR (slope ratio) 4.697±0.78 10.472±5.62 8.750±8.84 5.413±0.11

96

As a result, parameters with profound differences seem to merely affect the process (e.g. COD), as

ultimately, we assess the net degradation force at each effluent. To better investigate this effect related

with the organic matter, a look into the characterization of the organics content, which can act either as

(photo-) sensitizable or oxidizable matter, is useful (Giannakis et al. 2015c); to assess this possibility, a set

of photo-activity indices and measurements were applied to these effluents (see Table 3.2). The detailed

absorption spectra for the four effluents, are given in the supplementary material (Supplementary Figure

S2).

The specific wavelengths are used for photochemical activity evaluation purposes. The E2:E3 and the E4:E6

ratios, are measures of aromaticity, which indirectly imply photo-activity. The MBBR effluent presented

the highest E2:E3 ratio. The E4:E6 indicates higher colored dissolved organic matter (CDOM) presence; in

fact, higher E4:E6 ratio means low aromaticity (Fu et al. 2016). Since sometimes the absorbance at 665 nm

is zero, the ε280 (Helms et al. 2008) or SUVA (specific UV absorbance) were suggested as aromaticity

indexes. In our case, it corroborates with the rest of the findings. Also, the dimensionless slope ratio (SR)

(Helms et al. 2008) increased as we move from PT to MBBR (lowest to highest ratio) and suggested lower

molecular weight fractionated organic matter present in the effluent (Fu et al. 2016). Ultimately, since

LMW organic matter is proposed to act as photo-sensitizable organic matter (PhOM) rather than

oxidizable (OxOM) (Giannakis et al. 2015c, Vione et al. 2014), the possibility to participate in generating

ROS is higher. The generated ROS then participate in the controversial points 1 and 2, as analyzed before.

With this overview, it is safe to suggest that the physicochemical parameters followed in the effluents of

this work give only a view on some key factors influencing the process. This linear behavior is merely the

geometrical sum of the different actors present in WW water. In each effluent, the presence of

“secondary” parameters, such as the carbonates, or the organic matter, are vectors which can act either

synergistically or in an antagonistic manner.

5) Differences in the response between the two families of AOPs, for the different targets.

The UVC and UV/H2O2 processes were found to inactivate microorganisms and degrade pollutants rather

efficiently in all cases, whereas the Fenton and solar light (and less the photo-Fenton) require special

conditions (e.g. prolonged exposure) to ensure the success of the application. Apart from the obvious

differences in kinetics due to massive HO● generation in the UV/H2O2 process and higher energy in the

UVC photons (compared to solar UVB and UVA ones), there are other biological issues implicated in the

observed change in the reference and the importance attributed to each target.

When targeting micropollutants, the type and intensity of light applied, the addition of reagents etc. affect

the apparent 1st order degradation kinetics. However, in microorganisms, when light based processes are

involved, more issues should be considered. For instance, when UVC light is applied, the absorbance by

97

the thymine and cytosine bases is much higher than the UVB supplied by (simulated) solar light (Cadet et

al. 2005). As a result, the time necessary to inactivate microorganisms is in order of minutes, compared

to hours, for UVC and solar light, respectively. On the contrary, if many hours are required for completing

a solar disinfection experiment, given that bacterial (E. coli) cell division takes place every ~30 mins, many

generations of microorganisms could be formed, thus constantly repopulating the sample. In conclusion,

the forces of inactivation and salvation of microorganisms are far from collinear. This delicate balance,

which depends on light irradiance, solar dose, temperature and more, may contribute significantly in the

change of order of importance among the chemical and microbiological contaminants observed in the

Fenton related processes.

3.5. Conclusions

The use of five AOPs against microorganisms and contaminants of emerging concern was assessed in this

work, focusing in the differences among the two broad categories tested here, the UV-based and the

Fenton-related ones. In combination, three effluents were treated, from which two involved biological

treatment (AS, MBBR) and a physicochemical one (CF), while presenting the primarily treated as control

(PT).

The quality of the effluents influenced the outcome of the experiment, for micropollutants and

microorganisms, both for immediate removal or regrowth suppression. In general, the order of effluent

quality was MBBR>AS>CF. The measured physicochemical characteristics of these three effluents

influenced in the same manner the application of the AOPs. Among the two AOP groups, the order of

efficiency for the UV-based processes was UVC<UVC/H2O2 for both targets, while for the Fenton-related

ones was solar<Fenton<photo-Fenton for micropollutants and Fenton<solar<photo-Fenton for

microorganisms. Even so, high levels of synergy where observed for the constituents of the photo-Fenton

process and its utilization in neutral pH was encouraged. The post-treatment events of bacterial regrowth

were monitored, and prolongation of the treatment beyond the required time for inactivation was the

solution for the UV-based processes, while the Fenton-related ones, as they employ different inactivation

mechanisms that can hinder regrowth, need optimization of the Fenton reagent’s addition.

Taking into account the levels proposed by the Swiss legislation for micropollutant removal and the

threshold for water reuse of treated wastewater, if a quaternary treatment unit based on UVC irradiation

(with or without H2O2) was to be considered, the removal of micropollutants is a better indicator of water

quality; the use of H2O2 reduces significantly the residence times but increases the operational cost and

design implications. In other contexts than Switzerland, i.e. sunny or developing countries, if the Fenton-

related processes are to be considered, the bacteria were (marginally) considered to be a better reference

98

for depollution of wastewater, as for these higher residence times (compared to UV) the micropollutants

demonstrated faster degradation kinetics than the respective micropollutant ones. The variation

appearing among the two families of AOPs in the secondary effluents are a result i) of the structural

differences among the targets, ii) the degradation/inactivation pathway, iii) the reactor design

specifications, iv) the photochemical characteristics of the effluents and v) the biological implications of

the targets, such as bacterial growth and regrowth during treatment. As a result, apart from the

physicochemical data which can be easily monitored, photo-chemical indicators, design implications and

a battery of tests (microbiological, analytical) should be considered, in order to firstly well-characterize

the effluents and then refine the treatment strategies applied. Finally, although adequate MP removal

can be achieved, the degradation by-products and the potential toxicity problems related with the studied

processes are yet to be determined.

99

PART 2

Hospital-derived microorganism inactivation in developing countries

by Fenton-related AOPs: mechanistic interpretation and underlying

mechanisms of the photo-Fenton process.

101

4. Chapter 4 - Effect of Fe(II)/Fe(III) species, pH, irradiance and

bacterial competition on viral inactivation in wastewater by

the photo-Fenton process: Kinetic modeling and

mechanistic interpretation.

Work accepted for publication in Applied Catalysis B: Environmental

Stefanos Giannakis, Siting Liu, Anna Carratala, Rtimi Sami, Michaël Bensimon, César Pulgarin (2017).

Effect of Fe(II)/Fe(III) species, pH, irradiance and bacterial presence on viral inactivation in wastewater by

the photo-Fenton process: Kinetic modeling and mechanistic interpretation.

Supplementary Material:

Appendix D



102

4.1. Introduction

Wastewater disinfection is of major importance to prevent the microbial contamination of downstream

water resources. Treatment strategies such as filtration, chlorination or UV-radiation for microbial

inactivation have been developed over the last decades and their efficiency against bacteria (McGuigan

et al. 2012), viruses (Carratalà et al. 2016) and parasites (Carter 2005, Weir et al. 2002) has been assessed

in a number of studies. Nevertheless, treatment strategies are turning towards greener and more

sustainable techniques, such as the Advanced Oxidation Processes (AOPs).

Among these techniques, the photo-Fenton process has emerged as a prominent solution to treat

chemical contaminants (Malato et al. 2009), but the number of studies focusing on microorganisms is

significantly inferior. This process has been found to inactivate structurally simple (Kim et al. 2010),

complex (Giannakis et al. 2016c) or resistant microorganisms (Karaolia et al. 2014), and is promoted

because of its simplicity, low cost and limited environmental footprint (Pulgarin 2015). In the Fenton

reaction, hydrogen peroxide reacts with iron generating hydroxyl radicals, which are the predominant

reactive oxidizing species (ROS) responsible for microorganism inactivation in AOPs, and effectively

oxidize microbial components, such as amino acids and nucleotides (Giannakis et al. 2016a, Giannakis et

al. 2016b). In this process, iron acts as a catalyst, is repeatedly oxidized and reduced. In wastewater, the

process becomes significantly more complicated, due to the chemical and biological complexity of the

matrix and the imminent iron precipitation due to the near-neutral pH (6-8). Also, the presence of effluent

organic matter (EfOM) in wastewater (e.g. humic acid, fulvic acid), can scavenge a significant part of the

generated HO●, leading to a weakened inactivation (Rincón and Pulgarin 2004). However, other reports

have pointed out that light radiation on EfOM components (i.e. the dissolved fraction of the organic

matter, DOM) can create intermediate radical species, which react with water to generate HO● (Kohn and

Nelson 2007), and partially compensate for the loss of HO● by DOM.

Furthermore, Fe ions form complexes with organic matter and these Fe-organo species not only absorb

light, but also stabilize at near neutral pH. This extends the application of the homogenous photo-Fenton

reaction with less pH dependency (Spuhler et al. 2010). Under solar light exposure, the organic ligands

that compose the DOM form complexes with Fe(III) and participate in a ligand-to-metal charge transfer

(LMCT) type reaction (Ruales-Lonfat et al. 2015). However, during the photo-Fenton process, both

organics degradation and microorganism inactivation would compete. DOM may scavenge ROS and

provide targets of degradation with protection sites (Pignatello et al. 2006).

Most of the previous works on the efficacy of photo-Fenton processes against microorganisms at near

neutral-pH were done targeting bacteria (e.g. (Giannakis et al. 2015d, Ndounla et al. 2014, Ndounla et al.

2013, Ortega-Gómez et al. 2014, Ortega-Gomez et al. 2012), either in pure water or simulated wastewater

103

with NOM-like substances. To date, however, there is much less experimental information on the

efficiency and parameters governing the photo-Fenton inactivation of viruses (Kim et al. 2010, Kohn and

Nelson 2007, Nieto-Juarez and Kohn 2013, Nieto-Juarez et al. 2010, Ortega-Gómez et al. 2015). MS2

bacteriophage is a single-stranded RNA virus, which infects Escherichia coli thorough the pili and

resembles certain human enteric viruses in size (27.5 nm) and structural complexity. Also, it can be rapidly

cultivated in laboratory conditions up to the high concentrations needed for most inactivation studies, is

easily purified, not pathogenic to humans (Kuzmanovic et al. 2006, Templeton et al. 2006), which explains

its frequent selection as a model virus in microorganism inactivation studies (Bradley et al. 2011).

The main purpose of this work is to assess the overall efficiency of photo-Fenton inactivation against

viruses in wastewater, using MS2 bacteriophage as a model. Furthermore, the effect of major parameters

implicated in the process (namely Fe species and concentration, sunlight irradiance, initial pH and

bacterial competition) is also assessed, emphasizing the alterations inflicted by the presence of organic

matter in the matrix. The obtained datasets were used to propose a mathematical and a mechanistic

model to describe the pathways exerted during the photo-Fenton inactivation of MS2 bacteriophage in

wastewater.

4.2. Materials and methods


In the experiments, all the chemicals were reagent grade or above, and all the solutions were prepared in

water purified at analytical grade using a Millipore Elix 3 system combined with a Progard filter (Millipore

AG, Zug, Switzerland). The pH measurement was handled by a digital pH-meter (S220 SevenCompactTM

pH/Ion, Mettler Toledo).

4.2.1.1. Fenton reagents

Iron salts of FeSO4 7H2O (≥ 99.0%, Sigma-Aldrich) or Fe2(SO4)3 xH2O, (97%, Sigma-Aldrich) were used

according to the required starting iron species. The H2O2 stock solution in water was prepared with

PerdrogenTM (H2O2, 30% w/w, refrigerated, Sigma-Aldrich) and the quenching agent for H2O2 residuals was

aqueous solution prepared with a mixture of NaHSO3 and Na2S2O5, (99% Acros Organics).

4.2.1.2. Synthetic secondary wastewater

The synthetic wastewater was chosen rather than the secondary effluent from WWTP, for its ability to

provide an identical water matrix for every single experiment. The composition of the synthetic secondary

wastewater is shown in Table 4.1 (Muthukumaran et al. 2011). A concentrated stock solution (×10 times)

104

was prepared and dilution to reach the levels of Table 4.1 was done. The dilution was regulated to neutral

pH (pH 6, 7 or 8 in different series of experiment with 0.1 M NaOH or HNO3) before use.

4.2.1.3. Fe(II)/Fe(III) determination

The Fe(II)/Fe(III) concentrations in experimental samples were measured by a modified Ferrozine method

(Viollier et al. 2000), in which Fe ions chelated with the organic ligand Ferrozine and the formed magenta

iron complex was determined by absorbance measurement at 562 nm. This modified method could be

applied to detect μM level of Fe ions in water samples containing DOM. ICP-MS was also used to monitor

trace Fe amounts during experiments. The FinniganTM ICP-MS 7-238- NU1700 used was equipped with a

double focusing reverse geometry mass spectrometer presenting low background signal and high ion-

transmission coefficient. The spectral signal resolution was 1.2×10^5 cps/ppb. Fe from the solutions was

digested in 69% nitric acid (1:1 ratio HNO3:H2O), ensuring organics removal in solution and ions adhesion

to the vial wall. MS quantification of Fe concluded the analyses.

Table 4.1 – Composition of synthetic secondary wastewater (Muthukumaran et al. 2011).

Substances Composition [mg/L]

Meat extract (Sigma-Aldrich) 1.8 Peptone from meat, peptic digest (Sigma-Aldrich) 2.7

Humic acid (Carl Roth) 4.25 Tannic acid (AppliChem) 4.18

Lignosulfonic acid, sugared sodium salt (Sigma-Aldrich) 2.4 Sodium dodecylsulfate (NaC12H25SO4, 98%, Abcr) 0.9

Arabic gum powder (Acros Organics) 4.7 Ammonium sulfate ((NH4)2SO4, ≥ 99.0%, Carlo Erba Reagents) 7.1

Potassium phosphate dibasic (K2HPO4, ≥ 99.0%, Sigma-Aldrich) 7 Ammonium bicarbonate (NH4HCO3, ≥ 99.0%, Sigma-Aldrich) 19.8

Magnesium sulfate heptahydrate (MgSO4 7H2O, Sigma-Aldrich) 0.71

4.2.1.4. H2O2 determination

During the photo-Fenton experiments, the H2O2 concentration in the system was monitored. By reacting

with 10 μL of titanium(IV) oxysulfate solution (TiOSO4, 1.9-2.1%, Sigma-Aldrich), the H2O2 concentration

in a 1 mL experimental sample was quantified by measuring the colorimetric absorbance of the produced

yellow-colored pertitanic acid (H2TiO4) at 410 nm (Eisenberg 1943).

105

4.2.2. Sunlight source and reactors

4.2.2.1. Suntest solar simulator

The experiments were conducted using a solar simulator (CPS Suntest System Heraeus Noblelight, Hanau,

Germany) with infrared and UVC cut-off filters in order to simulate solar global radiation from outdoor

daylight and to prevent from the influence of UVC radiation and thermal heating (the emitted spectrum

can be found in the Supplementary material, Figure S1). The solar radiation intensities used in the assays

were 300, 600 and 900 W/m2 (global irradiance) which were monitored by the combination of a UV

radiometer and a pyranometer connected to a data-logger (CUV3 and CM6b respectively, Kipp & Zonen,

Delft, Holland). These values represent typical intensities achieved by solar light for different seasons,

latitude and/or time points within a day. According to previous experiments and own current

measurements performed under the same conditions, the temperature in the simulator never exceeded

38 °C (Spuhler et al. 2010).

An irradiance of 300 W/m2 was selected in the experiments concerning pH variations and different ratios

of Fenton reagents. At this relatively low light intensity, the MS2 inactivation was slower and the

inactivation kinetics could be better distinguished. However, the inactivation experiments in systems of

MS2 and host E. coli coexistence were carried out under the irradiance of 600 W/m2 in order to enhance

the E. coli inactivation rates.

4.2.2.2. Glass reactors

All the solar radiation experiments were performed in Pyrex, UVB-transparent glass vials of 100 mL, while

brown ones were used in the dark Fenton controls. Inside the solar simulator, a rectangular stirrer (MIX

15 eco, 2Mag Magnetic Motion, München, Germany) was used to place the reactors and the reaction

solutions were continuously stirred at 350 rpm. In order to avoid iron cross-contamination, after every set

of experiments, all glass reactors were soaked in 10% HNO3 overnight and then rinsed with deionized

water before heat-sterilization.

4.2.3. Microorganisms and quantification methods

4.2.3.1. Microorganisms

MS2 phage (DSMZ 13767) and the antibiotic-resistant (2 mg/L streptomycin) strain of bacterial host

Escherichia coli (DSMZ 5695) were obtained from Deutsche Sammlung von Mikroorganismen und

Zellkulturen (DSMZ, German Collection for Microorganisms and Cell Cultures, Braunschweig, Germany).

The propagation and purification of the coliphage was following the procedure described by Ortega-

Gómez et al. (Ortega-Gómez et al. 2015). The preparation of the bacteria for the co-culture experiments

followed a protocol similar to the one for E. coli K-12 (Giannakis et al. 2015b, Giannakis et al. 2014c).

106

4.2.3.2. Phage and bacteria quantification

Infective MS2 Coliphage was measured by the double agar layer technique (DAL, EPA Method 1602, 2001).

Plates were incubated at 37 °C for 18-24 h in a CO2-controlling incubator (B 5060 EK-CO2, Heraeus

Instruments, Hanau, Germany) and then the plaque forming units (PFU) were counted manually. The

detection limit of these experimental methods was found to be 10 PFU/mL. E. coli were handled with the

spread plate technique and colonies were quantified by the standard plate count method and results were

collected similarly.

4.2.4. Inactivation experiments

At the beginning of each experiment, 50 mL of the spiked wastewater matrix in the glass reactor was

stirred in the dark for 10 min to get evenly mixed. By adding 50 μL of 109 PFU/mL MS2 stock in carbonate

buffer solution (CBS, 8.401 mg NaHCO3, Sigma-Aldrich, and 876.6 mg NaCl, Sigma-Aldrich, dissolved in

1000 mL of water, adjusted pH to 8), the initial concentration of infective MS2 was around 106 PFU mL-1.

The reactors were then spiked with Fe(II)/(III) from freshly prepared stock solutions (500 mg/L) to reach

the final concentration at 0.25, 0.5 or 1 mg/L. Lastly, H2O2 was added from a fresh stock solution (1000

mg/L) and the final concentration was 0.5 or 1 mg/L. Next, the Xenon lamp was turned on to photo-

inactivate viruses under continuous irradiation at constant intensity. Corresponding control experiments

containing MS2 and Fe ion or H2O2 alone, under solar light or in the dark, were also conducted. To monitor

MS2 inactivation in the presence of its host, experiments were conducted at the light intensity of 600

W/m2, using 1 mg/L of Fe and 1 mg/L of H2O2.

During the experimental process, samples of 0.5 mL were taken at certain time intervals and immediately

mixed with 10 μL NaHSO3 aqueous solution (100 mg/L) to scavenge exceeding H2O2. Inactivation of MS2

and E. coli caused by NaHSO3 and dilution solutions was negligible over the experimental period (Kim et

al. 2010). Finally, before use, all the materials and solutions were autoclaved at 15 psi, 121 °C for 15 min

(Fedegari FVG1, Vitaris AG, Baar, Switzerland) to achieve complete sterilization.

4.2.5. Data treatment and analysis

Experimental data were expressed as the measured plaque forming units over time (PFU/mL vs. time),

where the instant MS2 concentrations were presented as arithmetic means ± standard deviations

calculated from the last three serial sample dilutions. For experiments involving host E. coli, the same

operation was done in interpreting its colony forming units (CFU/mL).

From the slope of a linear regression of ln([virus]0/[virus]) vs. ln(time), the observed k was determined as

the first-order inactivation rate in each individual experiment. In the systems involving Fe(II), where the

inactivation was better approximated by a two-phase exponential approach (Kim et al. 2010), for each

107

phase a first-order kobs was determined. The curve fitting for each series of experiment was conducted

with Maple 18 (Waterloo Maple Inc.) or Excel 2010 (Microsoft Corp.).

4.3. Results and Discussion

4.3.1. Isolated effect of the photo-Fenton constituents

First and foremost, all the experiments conducted in this work took place in a (simulated) secondary

wastewater matrix, therefore the inherent challenges encountered when stepwise constructing the

photo-Fenton process merit a separate mention.

4.3.1.1. Effect of the Fenton reagents in absence of light and the effect of solar irradiance

MS2 infectivity was well preserved in absence of light and Fenton reagents’ addition (see Supplementary

Figure S2). These results were similar to the viral survival curve obtained in CBS (Ortega-Gómez et al.

2015), although here at pH 7 and a complex dilution matrix. Adding 0.25 mg/L of Fe(II), Fe(III) or 0.5 mg/L

of H2O2 had a negligible effect within 60 min (<0.5-log inactivation). When both Fe ion and H2O2 were

applied simultaneously, the (dark) Fenton system with Fe(II) showed a 1.2-log inactivation while almost

no virucidal effect was observed for the Fe(III) system; this differentiation is also parallel (but lower) to

Kim et al. (Kim et al. 2010), where CBS served as the MS2 suspension medium. Although 1.2-log

inactivation is low, the higher efficiency compared to the other (milder or Fe(III)-driven) processes reveals

the potential of viral infectivity decrease in the complex WW matrix.

When samples were exposed to (global) irradiance of 300, 600 or 900 W/m2 for 60 min, the infectivity of

MS2 did not demonstrate a significant decrease, which indicated the relatively high resistance of MS2

towards UV/visible light, in accordance to previous findings (Nieto-Juarez et al. 2010, Ortega-Gómez et al.

2015). In addition, the wastewater matrix itself did not show a notable impact on virus inactivation under

solar radiation. The operating pH (initial pH: 7) and the presence of particles did not significantly affect

the infectivity of the viruses present in the matrix.

108

4.3.1.2. Effect of sole H2O2 or Fe addition on MS2 sunlight inactivation

Figure 4.1 – Solar/H2O2 and Solar/Fe control experiments. a) Isolated effect of the operating H2O2

levels of this work. b) Addition of 0.5 or 1 mg/L Fe(II) or Fe(III) salts. DL: detection limit.

In Figure 4.1a, under 600 W/m2 of solar irradiance, H2O2 doses of 0.5 and 1 mg/L contribute at about 0.5

log MS2 inactivation further than the effect of solar radiation alone. This result contradicts what was

reported by Ortega-Gómez et al. (Ortega-Gómez et al. 2015) in buffered water, which showed a total

inactivation after 50 min by 1 mg/L of H2O2. The virucidal effect was explained as a consequence of solar

irradiation that made MS2 more sensitive to oxidants, rather than the impact of HO itself (Ortega-Gómez

et al. 2015). The apparent contradiction of low H2O2 efficiency here, however, could be interpreted by the

presence of NOM-like substances in the specific water matrix, which can act as efficient scavengers of

HO (Westerhoff et al. 1999) and non-selectively react with H2O2 in competition with MS2 particles.

As shown in Figure 4.1b, dissolved Fe ions in wastewater induced a more effective MS2 inactivation than

hydrogen peroxide under solar irradiation. The addition of 0.5 mg/L of Fe(II) resulted in a decay of 2.9-log,

while in the presence of 1 mg/L MS2 inactivation reached 3.5 log units. For Fe(III), the amount of 0.5 mg/L

had a decrease by 1.8 logs, and 1 mg/L led a 2.5-log inactivation. Nieto-Juarez et al. (Nieto-Juarez et al.

2010) observed that MS2 infectivity dropped by half over 30 min in their CBS matrix containing 0.056 mg/L

of Fe(III), and considered this degradation as either virus inactivation by Fe ion alone or virus aggregation.

The possibility of spontaneous aggregation of MS2 is unlikely here, as Fe-induced MS2 removal by

excessive aggregation was found negligible by filtering samples prior to plating as well (0.2 μm filtering)

and comparing the results with un-filtered ones (data not shown). Thus, it is implied that Fe ions serve as

the main actors of inactivation in the Fe/light system.

109

According to previous works concerning virus-metal interaction, the LMCT process would occur in the

iron- and MS2-spiked water matrix, where the iron-MS2 “complex” acted as a light sensitizer (Ortega-

Gómez et al. 2015). These iron-MS2 complexes resulted from the interaction of Fe cations and negatively-

charged MS2 virions (pI = 3.9) in the wastewater. However, in our experiments, the observation that Fe(II)

contributed to a higher MS2 inactivation than Fe(III) in wastewater contradicts the aforementioned

results, and implies an efficient electron transfer from iron first, most probably a direct reduction of viral

capsid elements with oxidation of Fe(II) to Fe(III). Afterwards, the events in the two iron systems follow

the same catalytic cycle. Nevertheless, the enhanced inactivation events observed, despite the

scavenging effects of DOM, suggest complexation possibilities of iron in this matrix, and higher

participation in MS2 inactivation than in the buffered waters (Kim et al. 2010, Ortega-Gómez et al. 2015),

and called for further experimentation.

4.3.2. Parametrization of MS2 inactivation by the photo-Fenton process in wastewater

The following section summarizes the experimental assays combining the main actors in wastewater

disinfection by the photo-Fenton reaction, namely irradiance, Fe:H2O2 ratio, Fe starting species and initial

pH.

4.3.2.1. Effect of solar irradiance and Fe species

Figure 4.2 – Effect of solar irradiance on the evolution of the photo-Fenton reaction. A) Fe(II) as

starting iron species. B) Fe(III) as starting iron species. A notable difference exists in the kinetic

families of Fe(II) or Fe(III). DL: detection limit.

Figure 4.2 shows that solar light significantly enhanced the virucidal effect of Fenton reaction. For Fe(II)

(Figure 2a), at initial concentrations as low as 0.5 mg/L Fe(II) and 1 mg/L H2O2, an increasing sunlight

110

irradiance caused a dramatic decay of MS2 titers (3-log) as compared to the inactivation observed in the

dark, reaching the detection limit (DL) after a 6-log decay in 10 min. In the Fe(III)-involving system (Figure

2b), only an irradiance of 900 W/m2 could achieve the same extent of inactivation after an exposure during

60 min, while exposure at 300 W/m2 and 600 W/m2 still improved MS2 inactivation by 2.4 and 3 log,

compared to the Fenton process alone, respectively. The dependence of inactivation efficiency on light

intensity in presence of Fe(III) implies a photonic limitation of the system and dependence of the solar-

driven actions in order to achieve viral inactivation. This effect suggests that while the Fe(II) photo-assisted

process is not light intensity- dependent, it has a direct HO●-related inactivation, but for Fe(III), the Fe(III)

complexes are indeed photo-active and the light-assisted LMCT process is the driving force of the

inactivation process (Eqs. 1-3) (Giannakis et al. 2016a, Giannakis et al. 2016b):

● (4.1)

● (4.2)

(4.3)

4.3.2.2. Effect of Fe:H2O2 Ratio and Fe starting species

Figure 4.3 – Effect of the Fe:H2O2 ratio on the evolution of the photo-Fenton process. a) Fe(II) as

starting species. b) Fe(III) as starting species.

Increasing the initial concentration of Fe ions in the experimental water matrix led to an improved MS2

inactivation effect in both Fe(II) and Fe(III) starting form (Figures 4.3a and 4.3b, respectively). Using Fe(II)

at low concentrations (0.25:0.5) required 30 min for total inactivation and after doubling, the necessary

time reached the 2’ experimental limitation (minimum time from sampling to plating). On the contrary,

only the high ratios, employing 1 mg/L of Fe(III) and 1 mg/L of H2O2 under 300 W/m2 of sunlight, 6-log

111

inactivation was achieved, in 50 min. In the meantime, lower concentrations of Fe(III) (0.25 mg/L and 0.5

mg/L) ended in only 1- and 2.2-log inactivation. The inactivation rate showed a dependency on the initial

Fe concentration, as well as verifying the dependence of the starting iron species. Ortega-Gómez et al.

(Ortega-Gómez et al. 2015) interpreted that the formation of Fe-MS2 complex was favored by the

increasing amount of Fe ion, resulting in the generation of oxidants close to virus particles, crucial for their

efficient inactivation (Nieto-Juarez et al. 2010).

The variation of inactivation effects by different initial concentrations of H2O2 was also displayed in Figures

4.3a and 4.3b. In sunlit Fe(III) (0.25 mg/L) systems, 1 mg/L of H2O2 increased the proportion of inactivated

MS2 by 0.5 log, when compared with the employment of 0.5 mg/L. When reacting with 0.25 mg/L of Fe(II),

by applying more H2O2, time used to reach complete inactivation was shortened from 30 min to 15 min.

Clearly, the difference due to varied H2O2 concentrations exhibited in a second phase (after 2 min) of

inactivation. As mentioned before, since the second period was driven by the regeneration of Fe(II) from

Fe(III), a higher concentration of H2O2 helps accelerate the overall inactivation process by the produced

ROS (Giannakis et al. 2016a, Giannakis et al. 2016b).

● ( ) (1.8) ● ( ) (2.2) ● ( ) (4.4)

● ( ) (4.5)

( ) (4.5)

( ) (4.6) ● ( ) (4.7)

● ( ) (4.8) ● ( ) (4.9) ● ( ) (4.10)

Because of the interaction between Fe(II)/Fe(III) and negatively-charged DOM (e.g. the pI of humic acid is

around 2) (Kohn and Nelson 2007), H2O2 was more efficiently used in the inactivation by photo-Fenton

process with Fe-DOM complexes than having its ROS quenched by it (Giannakis et al. 2016a, Giannakis et

al. 2016b).

Facilitator: (4.11)

Antagonist: ● (4.12)

112

This explanation might be confirmed by the colorimetric H2O2 measurements during the experiments,

where the concentrations of H2O2 in the system can be considered constant (data not shown). Our findings

agree with (Timchak and Gitis 2012) who assumed that H2O2 regenerated constantly in HO2 and HO

radical reactions, through the following reaction (Giannakis et al. 2016a, Giannakis et al. 2016b):

● (4.13)

The MS2 survival curves in photo-Fenton systems containing Fe(II) or Fe(III) have different shapes. The

Fe(II)-induced photocatalysis performed a sudden decrease and all viruses were inactivated in 5 min;

comparing to the curve of 0.25 mg/L of Fe(II) and 0.5 mg/L of H2O2, which showed a distinction of two

inactivation phases, the proposal was put that the Fe(II) performance includes two distinct phases,

although due to its relatively high ability of disinfection, the second phase was only observed as a tailing

pattern under low concentrations (<0.5 mg/L). Tailing in UV-inactivation studies has been attributed to

recombination of viruses (Olsthoorn and Van Duin 1996) which requires that multiple (inactivated) MS2

virions to infect the same host cell (Mattle and Kohn 2012). However, the onset of a tailing in MS2 decay

was not observed for Fe(III). In this case, the overall inactivation rate was lower, and the decrease of MS2

infectivity followed a logarithmic curve without phase distinction. At a ferric dose of 0.5 mg/L, it was not

able to inactivate MS2 beyond 2.5 logs in 60 min. The second phase in the Fe(II) system at low

concentrations is correlated with the dependence to the Fe(III) presence, whose inactivation kinetics are

slower.

4.3.2.3. Effect of the starting pH

113

Figure 4.4 – Effect of the starting pH on the evolution of the photo-Fenton process. a) Fe(II) as starting

species. b) Fe(III) as starting species.

The influence of pH on the photo-Fenton system is shown in Figure 4.4. The pH of applied water matrices

was measured before and after each experiment, with negligible drop noticed after 60 min of

experimental time. Although all the experiments were handled at near neutral pH, the variation of the

starting pH from 6 to 8 created a remarkable difference between inactivation rates. In photo-Fenton

experiments driven by Fe(II), when pH was 6 or 7, a complete inactivation was achieved in less than 5 min;

at pH 8, in 60 min the inactivation of MS2 was no more than 3.5 logs. When the starting Fe ion was

trivalent, the results of pH 6 and 7 were even more different. At pH 6, viruses were all inactivated in 10

min, at pH 7, the inactivation was completed in 50 min at a ferric dose of 1 mg/L and a pH of 8 did not

favor the photo-Fenton reaction, with inactivation of 0.8 log being attained.

4.3.3. Effect of bacterial competition on MS2 inactivation in wastewater

The experiments mentioned until now were all performed in sterile synthetic wastewater spiked only with

MS2. Normally in wastewater, a number of different microorganisms (bacteriophages, their bacterial

hosts and other microorganisms) co-exist and may play an important role in each other’s life cycle; certain

bacteria can grow and reproduce in the environment, while virus are capable of infecting them. To verify

if the proposed conditions of the photo-Fenton reaction could also apply to a more realistic situation in

which different microorganisms coexist, experiments were conducted in reactors simultaneously spiked

with MS2 bacteriophage and its bacterial host E. coli (Figure 4.5). Since no significant increase of MS2

titers was observed during the course of these experiments the effect of phage infection to the host could

be ignored.

When E. coli alone were exposed to the treatment conditions (figure 4.5a) at which we achieved total

MS2 inactivation (1:1 ratios and 600 W/m2 solar intensity), the viability of E. coli slightly dropped in 60

min, comparing to a 0.5-log decrease when MS2 was present (figure 4.5b). Naturally, viruses consume

oxidants in competition with bacteria. However, since a virion is one or two orders of magnitude smaller

than a bacterium and has a simpler structure, in the photocatalytic system an identical amount of

generated ROS could kill more individual viruses, while E. coli inactivation requires multiple ROS hits to

damage a single cell, and is subjected to endogenous inactivation events. Even among virus strains, the

ones with thinner capsids are more susceptible to inactivation by ROS (Carratalà et al. 2016), therefore

the analogy with bacteria is compelling. In both Figures 4.5a and 4.5b, Fe(II) was again proven to

contribute significantly to bacterial inactivation when MS2 were absent. A noteworthy change was

observed when MS2 were present in the solution, lowering the already small bacterial inactivation with

114

the same reagent concentration. This difference is mitigated when higher reactants’ addition was assayed

(0.5:10 and 1:10).

Figure 4.5 – Bacterial competition tests: Inactivation of MS2 and E. coli by the photo-Fenton process.

a) Bacterial inactivation with increasing Fe:H2O2 ratios, in absence or presence of MS2. b) E. coli

inactivation in presence of MS2 and MS2 inactivation in presence or absence of the bacterial host

(Fe:H2O2 ratio 1:1). Higher Fenton reagents addition than 1:1 resulted in <2-min inactivation.

Finally, Figure 4.5b demonstrates the effect the bacterial presence has on MS2 inactivation. The time

necessary for total inactivation was prolonged from 5 to 10 min for MS2 with the presence of E. coli, since

viruses and bacteria are in competition by the generated ROS in the system. Under the condition of steady

stirring, the microbes were homogenously distributed in the water matrix and had random contacts with

systemic ROS. Nevertheless, the viruses could still be totally inactivated in a short period. Furthermore, in

this experiment, 1 mg/L of Fe(II) and 1 mg/L of H2O2 were introduced into the tested system and the

simulated irradiance was set at 600 W/m2. Recent disinfection approaches (Ortega-Gómez et al. 2014,

Ruales-Lonfat et al. 2015) of Fenton reagents application in the micro- to milli-molar range have been

reported so far, while (Spuhler et al. 2010) and (Giannakis et al. 2016d) observed the efficiency of bacterial

inactivation at neutral pH by 0.6 mg/L Fe(II) or Fe(III) and 10 mg/L of H2O2 under 550 W/m2 of solar

radiation. Based on the results above, it can be assumed that a general dose of the Fenton reagents and

115

sunlight exposure for bacterial disinfection is also sufficient for a fast viral inactivation in the wastewater

matrix.

4.3.4. Iron cations solubility in wastewater

Normally, effluents that have been subjected to a biological (or other secondary) treatment, are good

candidates for applying AOPs, such as the photo-Fenton process. Nevertheless, there are some

characteristics that often hinder the process, such as the presence of organic matter, the alkalinity, the

suspended solids and the quasi-neutral pH, among others. However, the organic matter is often

considered acting in a dual manner, with capabilities of complexing metals not only antagonizing the

process, but also able to facilitate the photo-Fenton process (Giannakis et al. 2016a, Giannakis et al.

2016b).

Similarly, the simulated wastewater used in this study was adjusted to neutral pH, at which normally iron

precipitates. We studied the evolution of the photo-Fenton process, in presence of MS2 and Fe(II) or

Fe(III), comparing Milli-Q (MQ) water with the studied wastewater. Over the course of two hours, there

is a slight decrease in the UV-vis absorbance of the bulk in each of the iron additions in MQ (see

Supplementary Figure S3). The reference absorbance belongs to MQ water and 1, 2 and 5 mg/L of Fe(III)

were added for further analyses. Additionally, WW forms more photo-absorbing, and therefore, photo-

active complexes. During a 2-h test, there was practically no decrease in the absorbance, indicating the

stability of the Fe-DOM complexes.

In order to verify the observations, a similar test was performed for 1 hour in WW, with 1 mg/L of either

Fe(II) or Fe(III). The goal of the test was to dissociate the dissolved and total iron (by 0.2 μm filtering)

during an experiment at pH 7 in presence or absence of light. The main two general trends can be

summarized in Figure 6 under the higher values of dissolved iron of Fe(II) (Figure 4.6a) than Fe(III) (Figure

4.6b) as starting iron species, and the slightly lower precipitation rates in the dark tests. On one hand,

Fe(II) readily reacts with the H2O2 in the matrix and then passes to Fe(III), which precipitates faster in near-

neutral values, hence the fast losses in total iron presented with Fe(III) as starting iron species. The

presence of organic matter helps complex the iron, efficiently perform an LMCT, and then possibly re-

complex the iron. This could possibly explain the high iron availability during the tests. Finally, TOC

measurements (data not shown) verify the degradation of the organic matter, however in negligible rates

when low Fenton reagents’ concentration was used. In conclusion, iron, even at small amounts is available

to react and plays an important role in the MS2 inactivation in wastewater.

116

Figure 4.6 – Iron evolution during (dark) Fenton or photo-Fenton process followed by ICP-MS analysis.

A) Fe(II) starting salts. B) Fe(III) starting salts. The dashed lines indicate the dark Fenton experiments,

closed trace symbols indicate dissolved iron and open trace symbols the total iron.

4.3.5. MS2 inactivation modeling

Various kinetic models have been proposed in the literature for microbial inactivation (Marugán et al.

2008). According to Hiatt (Hiatt 1964), the virus survival data can be plotted as ln c/c0 vs. ln t, obtaining

straight lines which are curvilinear for ln c/c0 vs. t. Here c/c0 is the reduction in the infective MS2

concentration and t refers to the treatment time.

Thus, the relationship between c/c0 and t is described as the following function:

(4.14)

where kobs behaves as the observed “relative velocity constant”, corresponding to a pseudo-first-order

kinetic. This model was applied to fit experimental data of the photo-Fenton inactivation at pH 7 under

different conditions (Table 4.2 and 4.3).

However, as seen in before, the curves of MS2 photo-inactivation with Fe(II) and H2O2 proceeded a sudden

decrease at the very beginning of the whole process, and then changed to a tailing pattern if MS2 had not

been totally inactivated. These “biphasic” survival curves can be drawn as two linear components (Hiatt

1964, Kamolsiripichaiporn et al. 2007):

The first stage: (4.14)

117

The second stage: (4.15)

where m is the intercept of the second function on the y-axis.

Table 4.2 – Effect of photo-Fenton treatment [Fe(II)] on MS2 inactivation.

LLight [[W/m22]

Fe(II) [mg/L]

H2O2 [mg/L]

T999.99% [min]

k1,obs [min-1]

r12

t1 [min]

k2,obs [min-1]

m r22

t2 [min]

0 0.5 1 - 2.4152 0.5770 0-5 0.3806 3.6564 0.9903 5-40 300 0.25 0.5 10 7.1218 1 0-2 1.3106 5.9716 0.9146 2-20 300 0.25 1 7.5 7.1218 1 0-2 2.3580 5.0404 0.9686 2-10 300 0.5 1 4 7.1218 1 0-2 300 1 1 3 7.7527 1 0-2 600 0.5 1 2 8.4208 0.9969 0-2 600 1 1 2 20.2658 0.9986 0-0.5 2.818 7.0741 1 0.5-1 900 0.5 1 1.5 9.8486 1 0-2

For different iron starting species of Fe(II) and Fe(III), klight can be determined from the slopes of kobs vs.

[intensity].

Table 4.3 – Effect of photo-Fenton treatment [Fe(III)] on MS2 inactivation.

Light [W/m2]

Fe(III) [mg/L]

H2O2 [mg/L]

T99.99% [min]

kobs [min-1] r2 t

[min] 0 0.5 1 - 0.1606 0.9561 0-60

300 0.25 0.5 - 0.3746 0.9142 0-30 300 0.25 1 - 0.7197 0.9094 0-60 300 0.5 1 - 1.4366 0.9653 0-40 300 1 1 33 2.4794 0.9772 0-30 600 0.5 1 32 1.8140 0.9698 0-60 900 0.5 1 30 2.2894 0.9258 0-30

Ultimately, the three investigated parameters light intensity, Fe(III) and H2O2 concentrations can be

linked together as a single function, within the boundaries of the experimental space, i.e. Intensity [I] =

300-900 W/m2, [Fe(III)]ini = 0.25-1 mg/L, [H2O2]ini = 0.5-1 mg/L), solving the Eqs. 4.14-4.15 and by non-

linear regression fitting we get:

(4.16)

Here klight,Fe(III) corresponds to 0.0016 m2 W-1min-1, kFe(III) to 2.310 L mg-1min-1 and to 0.7996 L mg-1

min-1.

As Eq. 4.16 shows, klight and kFe can be directly applied in the multi-parameter function in Fe(III) systems.

However, in Fe(II)-induced systems cannot be calculated due to the fast inactivation and relatively

118

long sampling time from which the obtained kobs were the same. The correlation between experimental

data and the function is acceptable in the majority of the sets (Tables 4.2 & 4.3). However it must be noted

that during the first two minutes of the experiments, the inactivation was so fast that it was possible to

have already inactivated all the viruses or entered the second phase of disinfection.

4.3.6. Integrated proposal for the inactivation mechanism of viruses in wastewater

Based on our experimental findings, the inactivation presented in one of our previous works (Ortega-

Gómez et al. 2015) and the relevant literature, a modified framework for wastewater is suggested in

Figure 4.7, displaying the possible pathways occurred in virus inactivation, driven by the photo-Fenton

process in wastewater. This framework postulates a modified inactivation pattern in close vicinity to the

virus, also suggested as a caged mechanism by previous works (Nieto-Juarez et al. 2010). Concerning the

MS2 inactivation (Figure 4.7, events 1-6):

1. Sunlight directly affects the viral genome, decreasing its infectivity. Other direct actions involve

the damages to the coat protein, as well as considerable A protein decay (Wigginton et al. 2012).

2. Oxidative stress exerted by H2O2 on the virus is not significant, due to its low oxidation potential,

while under light, due to both its poor photolysis and the presence of DOM that scavenges reactive

hydroxyl radicals the expected contribution is limited. Only by combination with UVA important damages

have been reported (Nelson et al. 2008, Romero et al. 2011).

3. Irradiation of the DOM present in WW generates a small amount of H2O2, O2-, 1O2 and other ROS.

The superoxide radical anion is always generated in lower steady-state concentrations than other ROS.

Also, the presence of these trace metals is reported to enhance its production (Voelker and Sedlak 1995).

As it was reported, O2- has minor contribution in direct MS2 infectivity decrease. On the other hand,

singlet oxygen significantly affects viral infectivity (Kohn and Nelson 2007).

4. In a dark Fenton reaction, Fe(IV) species participates predominantly in the inactivation of viruses;

however, an implementation of solar light greatly enhances the production of HO , which is more

effective in virus inactivation (Nieto-Juarez et al. 2010). When high organic matter concentrations were

involved in previous works, the steady state concentrations and the use of specific HO● quenchers proved

that the HO● contribution of the bulk is low (Kohn et al. 2007). Later works based on caged mechanisms

of inactivation (Chevion 1988, Kocha et al. 1997), suggested that HO● plays a role only when generated in

the vicinity of the virion (Nieto-Juarez et al. 2010).

5. In the wastewater matrix, Fe ions form aquo-complexes by hydrolysis and organo-complexes with

the DOM present. As indicated by Rose and Waite (Rose and Waite 2002), the reaction rate of Fe

complexes towards ROS is the same as for the free ions. The complex formation allows the photo-Fenton

reaction to proceed at neutral pH where Fe(III)-organo complexes are generally stable in particular. They

show higher absorption in the visible range of solar light than aquo-complexes, favoring LMCT reaction

119

that generates ROS under sunlight (Spuhler et al. 2010). The resulting ROS (HO , O2-) have been analyzed

before and the oxidized ligand could possibly proceed with further ROS production (Cieśla et al. 2004,

Šima and Makáňová 1997).

6. Finally, Fe(II) and Fe(III) may directly interact with any of the amino acids in MS2 capsid, forming

organo-complexes. In this case, the Fenton reactions can proceed on the MS2 surface (Kim et al. 2010),

facilitating an LMCT reaction with the viral capsid as sacrificial ligand.

Figure 4.7 – Proposed MS2 inactivation pathway by the photo-Fenton process in wastewater at near-

neutral pH. The events 1-6 are further analyzed in the text.

Fe(II) Fe(III)

Fenton,Photo-Fenton LMCT, Fenton,

Photo-Fenton

ssRNA

Fe(II)Aqua-complex

Organo-complex

Fe(III)Aqua-complex

Organo-complex

,

Lox- & other ROS

Produces

Importantif in

proximity

Dissolved Organic Matter

6 6 1 3 5

4

Fe(IV)

Sunlight (280 – 800 nm)

2

ssRNA

A

A

A

A

C

C

CU

U

U

G

G

CU

CA

ACCCCCCCCCCCCCCCC

CU

U

G

UUUC

A

A

A

A

C

C

CU

U

U

G

G

CU

C

A

AAAAAAA

CU

GC

A

AAAAAAAAAA

A

C

CU

U

U

GGG

A

A

A

A

C

C

CU

U

U

G

G

CU

C

A

AA

CCCCCCCCCCCCCCU

GC

UUUC

A

AAAAAAAAA

A

A

C

CCCCCCCCCCCCCCCC

CU

U

U

G

GGG

CCUUU

C

120

4.4. Conclusions

The use of μm concentrations of Fe and H2O2 has been proven, under certain conditions, enough to

enhance the viral model inactivation in simulated wastewater. The photo-Fenton process in near-neutral

conditions under the presence of scavenging DOM has proceeded to effectively reduce MS2 infectivity,

and the effects of the key parameters of the process were further analyzed.

The photo-Fenton reagents’ concentration was kept at <1 mg/L level, in order to facilitate easier kinetics

studies. Even so, it proves the feasibility of removing MS2 from wastewaters. Fe(II) proven was more

efficient than Fe(III), as its interaction with key components of the capsid enables the ROS production in

the vicinity of the virus. Lowering the ratio among Fe:H2O2 (increasing iron concentration) improved the

disinfection process indicating the importance of iron presence for efficient disinfection. Similar trend was

observed for the pH, which improves iron solubility and therefore its participation to the photo-Fenton

process. Nevertheless, the most important contribution derives from the DOM complexation of iron,

which, apart from the additional ROS production, also mitigates precipitation and facilitates the

maintenance of the initially added amounts. Furthermore, a simple model describing the MS2

inactivation, including the implicated parameters, was effectively constructed.

In conclusion, we identified iron and organic matter as the key factors playing a role in the process when

a complex, multi-target matrix was involved, proposing an inactivation scheme, as well as presenting the

indicative difference of the needs for addition of oxidants. As in actual applications the levels of Fe and

H2O2 addition are higher, the MS2 removal is normally expected sooner than the corresponding bacterial

inactivation. Finally, further work concerning human pathogenic viral strains and other public health

related microorganisms is required before establishing a relative order of removal (compared to other

pathogens) and asserting the efficacy of photo-Fenton.

121

5. Chapter 5 - Castles fall from inside: Evidence for dominant

internal photo-catalytic mechanisms during treatment of

Saccharomyces cerevisiae by photo-Fenton at near-neutral

pH

Published work:

Stefanos Giannakis, Cristina Ruales-Lonfat, Sami Rtimi, Sana Thabet, Pascale Cotton, and César Pulgarin.

"Castles fall from inside: Evidence for dominant internal photo-catalytic mechanisms during treatment of

Saccharomyces cerevisiae by photo-Fenton at near-neutral pH." Applied Catalysis B: Environmental 185

(2016): 150-162.

Web link:

http://www.sciencedirect.com/science/article/pii/S092633731530299X


Appendix C



122

5.1. Introduction

Historically, one of the biggest side-effects of urbanization, was the high concentration of inhabitants in

relatively small areas, which in turn created massive flows of wastewater. It was not until after the

Industrial Revolution that the these wastes started to become a threat against the environment, since up

to that point, the self-cleaning capacity of natural water bodies was able to regulate the incoming nutrient,

organic and inorganic pollution. The last decades, apart from the chemical constituents, a systematic

effort to upgrade the wastewater treatment plants has been undertaken, in order to eliminate

microbiological contaminants contained in wastewater. As first approaches, chlorination, peracetic acid

and other chemical methods were employed (White 2010), gradually replaced by more eco-friendly

methods (such as UV or ozonation) due to the lower harmful by-products connected with their

application.

Moreover, hospital wastewater has been categorized as a main contributor of chemical substances (drugs,

wastes) (Verlicchi et al. 2012) and microbiological agents normally not encountered in municipal

wastewater, such as high concentration of antibiotic resistant bacteria, viruses, yeasts or fungi (Li et al.

2007, Prado et al. 2011, Schwartz et al. 2003). The similar origin of wastewater in health facilities led the

practitioners to apply one sterilization step before leading the treated wastewater in the municipal

wastewater treatment plants (Emmanuel et al. 2004). During the last years in Europe, there is an

increasing demand of treatment plants specifically designed for hospitals, taking into account the special

form of pollution contained (Lienert et al. 2011). Nevertheless, in less favorable regions of the world, these

wastewater treatment plant facilities are either a luxury, non-existing or have never operated properly.

As a result, the wastewater is directly discharged into nature and the natural water bodies are

transformed into disease carriers. Bacterial induced illnesses, such as diarrhea, viral or fungal-related

infections have been encountered, problems craving a simple, cheap and sustainable solution (Gibson

2014, McGuigan et al. 2012).

The advanced oxidation processes (AOPs) have emerged during the last years as a practical and in some

cases, easily applicable solution, as a barrier to stop pollution in contaminated drinking water sources. As

a common denominator, the extremely oxidative hydroxyl radical is produced, and secondarily, other

reactive oxygen species (Ruales-Lonfat et al. 2014). Ozonation (at basic pH), heterogeneous photocatalysis

by TiO2, UV-induced production of HO● by the homolytic disruption of H2O2 and the iron-catalyzed radical

induction are some of the most popular AOPs in decontaminating water or wastewater (Comninellis et al.

2008, Poyatos et al. 2010). Especially the Fenton reaction, and its photo-enhanced version, gained much

merit during the last decade, after the proof of its considerable effectivity at near-neutral pH (Rincón and

Pulgarin 2006); in the past, its acidic pH-bound character acted as a limiting step towards its application

123

against microorganism removal. Its low cost and easy application has proven itself a powerful ally in sunny

regions around the globe (Ndounla et al. 2014, Ndounla et al. 2013).

A significant number of studies discuss the efficiency of the Fenton reaction as an antibacterial agent,

using most commonly E. coli, but also Enterococci, and Salmonella (Ndounla et al. 2013, Ortega-Gómez et

al. 2013, Ruales-Lonfat et al. 2015, Spuhler et al. 2010). The efficiency of the Fenton reaction was lately

tested for the inactivation of viruses, such as MS2 Coliphage and Echovirus (Ortega-Gómez et al. 2014).

Although prokaryotic unicellular microorganisms have been more studied, due to their omnipresence in

the environment and their impact on health, expanding works on other groups of microorganisms is

necessary. The impact of eukaryotic microorganisms on environment and human health is not negligible.

The effects of photocatalysis on fungal cell survival has been showed for environmental species, like

Penicillium, Fusarium, and Aspergillus that are frequently recovered from water and soil (Brinkman et al.

2003, Park et al. 2015, Pigeot-Rémy et al. 2012) and for yeast species like Candida closely associated to

human opportunistic infections (Thabet et al. 2014).

In eukaryotic cells, there are significant structural modifications compared to prokaryotes which may

facilitate difference in interesting traits, disinfection-wise, such as resistance, higher stress responses and

repair mechanisms (Temple et al. 2005). A series of photocatalysis studies using TiO2 have focused on F.

solani as a model of multicellular microbial structure (Sichel et al. 2009, Sichel et al. 2007b), and at the

unicellular level on Saccharomyces cerevisiae as a model for oxidative response (Thabet et al. 2014,

Thabet et al. 2013). It was proven that there is limited penetration of TiO2 nanoparticles into the yeast

cell, and that photocatalysis induces the establishment of an intracellular oxidative environment (Thabet

et al. 2014). In bacteria, the effect of an internal Fenton mechanism has been brought to surface (Ruales-

Lonfat et al. 2014), complementing the external pathways but yet, it is still unclear how it would act, when

yeast cells are the target of the Fenton reaction.

In this work, we have used S. cerevisiae, as a model of eukaryotic microorganisms, and photo-Fenton as

the antimicrobial AOP. The mechanisms of yeast inactivation by photo-Fenton process at near-neutral pH

were investigated. The reactions were fueled by different iron sources, namely iron sulfate and iron

citrate, in presence of H2O2, under simulated solar light. The different inactivation pathways were

interpreted by using flow cytometry, and assessment of the damage at both DNA and protein level was

also performed. DNA and cell wall protein damages were depicted by electrophoresis, to elucidating the

photo-Fenton mode of action.

124


5.2.1. Chemicals

Ferrous sulfate heptahydrate (FeSO4˙7H2O) (Riedel-de Haën 99-103.4%); Ferric chloride (FeCl3) (98% carlo

erba), Trisodium citrate dihydrate (Na3C6H5O7.2H2O) (99% Merck); Sodium hydrogen carbonate (NaHCO3)

(analytical grade, Merck); Hydrogen peroxide (H2O2) 30% w/v (Riedel de Haën); Titanium (IV) oxysulfate

(TiOSO4) (Fluka); Sodium hydroxide (NaOH, 98%) and hydrochloric acid (HCl, 36.5%), were purchased from

Sigma-Aldrich, Switzerland. The spin-trap, 5,5-dimethyl-1-pyrroline-N-oxide (DMPO), was purchased from

Enzo Life Sciences (ELS) AG (Switzerland). All solutions were prepared immediately prior to irradiation

with the use of Milli-Q water (18.2 MΩ-cm).

5.2.2. Fe–citrate complex and Goethite preparation

Fe–citrate complex was prepared according at modified patent European from Bayer (Antonini and Vidic

1994). Ferric chloride (4.1 g) and sodium hydrogen carbonate (3.0 g) were dispersed in 50 mL of distilled

water and dissolved therein by stirring. This solution was degassed under vacuum for 2 hours followed by

constant stirring and Trisodium citrate dihydrate (4.0 g) was added. The color of the solution turned to a

pale brown. The solution was stored in the dark for 24 h. Then, 40 mL of methanol was added to the

brown solution under constant stirring at 25 °C and a brown precipitate was formed. The resulting solution

was centrifuged (5 min at 5000 rpm) to remove the precipitate, and the clear supernatant suspension was

separated by filtration. The precipitate was washed with methanol at least three times and dried under

vacuum at room temperature. Goethite preparation and characterization was analytically presented in

previous works of our group (Ruales-Lonfat et al. 2015).

5.2.3. Yeast strains and growth media

The laboratory strain S. cerevisiae (BY4742) was used for all yeast inactivation experiments. The strain was

maintained on YPD medium (1% yeast extract, 1% peptone, 2% glucose, 2% agar for plates). Yeast cells

were grown in liquid YPD overnight under aerobic conditions with constant shaking at 28 °C. The yeast

culture growth was checked by measuring optical density at 600 nm using a spectrophotometer. For all

experiments, cell samples were collected at the beginning of the exponential growth phase (OD600 = 1,)

washed twice and suspended in the photoreactor in 20 mL of sterile ultra-pure (UP) water (Simplicity™,

Millipore), resulting into a concentration of 107 cells/mL.

125

5.2.4. Photo-inactivation experiments

All yeast inactivation experiments were performed in Pyrex reactors (4 cm x 9 cm, 100 mL). The Pyrex

reactors containing the yeast suspension in Milli-Q water (approximately 107 CFU/mL) were placed in the

dark at 25 °C under magnetic stirring for at least 30 min to let the yeast adapt to the new matrix and to

allow the die-off and equilibration of the most stress-sensitive species.

The following systems were analyzed for the inactivation effect on yeast. (i) photo-Fenton process

mediated by Fe–citrate (0.6 mg Fe/L) at pH: 6.0 or 7.5; (ii) photo-Fenton process mediated by FeSO4 (0.6

mg Fe/L) at pH: 5.5 and 7.5; and (iii) control experiments: H2O2/dark; light alone and H2O2/light. Goethite

addition (0.6 mg Fe/L) was assessed complementarily, to assess the function of the iron after its

precipitation in the near-neutral environments. In the experiments iron was added to a yeast suspension.

Then, the pH was adjusted depending on the experiment. Neighboring near-neutral values were assayed

to better investigate the behavior of a system which partially permits soluble Fe and compare with realistic

pH. Finally, H2O2 (10 mg/L) was added to the reactor as the last component.

Experiments were carried out using a solar simulator CPS Suntest System (Heraeus Noblelight, Hanau,

Germany). This solar simulator was equipped with a basic uncoated quartz glass light tube, a filter E and

an IR screen (neither UVC nor IR is reaching the sample, the intermediate wavelengths are a simulation of

the solar radiation); more information can be found at (Giannakis et al. 2015a, b). The irradiance was

measured by a spectro-radiometer, Model ILT-900-R (International Light Technologies) and corresponded

to 820 W/m2 of light global irradiance (from which ~0.5% UVB, ~5% UVA is emitted). Temperature was

monitored and always remained <38°C.

5.2.5. Cultivability assays

Samples were collected at regular intervals during yeast inactivation reaction. Serial dilutions were

immediately made in liquid YPD medium and spread onto YPD agar plates. After 2 days of incubation at

28 °C, the colony forming units (CFU) detected on appropriate dilution plates were counted, in order to

determine the concentration of surviving cells. Triplicate plating was performed for each dilution of the

samples. All experiments were performed in triplicates and the results presented in the graphs are the

average value (<5% statistical error).

126


5.2.6.1. Optical epifluorescence microscopy

Microscopy observations were performed using Axioscop 2 plus Zeiss optical microscope equipped with

AxioCam MRm camera. Data were collected using AXioVision software.

Cell viability was investigated using PI (Propidium iodide, Invitrogen, ex/em 490/635 nm) and CFDA-AM

dye (Carboxyfluorescein diacetate- acetoxymethyl, ROCH, ex/em 492/517 nm). PI enters only cells with

damaged cytoplasmic membranes, whereas CFDA-AM enters all cells and is non-fluorescent until it is cut-

off by active cytoplasmic esterases. CFDA-AM reflects cell metabolic activity. Treated cell samples (100 μL)

were diluted in PBS to 106 cells/mL. After addition of dyes (1 μg/mL PI and 5 μg/mL CFDA-AM), the mix

was incubated for 20 min at 37 °C.

5.2.6.2. Flow cytometry

Flow cytometry was carried out using FACS CantoII instrument (BD Biosciences) fitted with three lasers:

blue (488 nm, aircooled, 20 mW solid state), red (633 nm, 17 mW HeNe) and violet (405 nm, 30 mW solid

state). Diffracted light (related to cell surface: Forward scatter FSC) and reflected light (related to

granularity: Side scatter SSC) of blue laser, as well as CFDA-AM and PI, were collected. Data from

10,000 cells were collected using FACSDIVA software (6.1.2 version, BD Biosciences).

5.2.7. Biochemical methods

5.2.7.1. DNA extraction and analysis

Yeast chromosomal DNA was extracted from 2 mL of stationary phase YPG culture at 28° C and 150 rpm.

Cells were collected by centrifugation 30 s at 16000 g, washed in 0.5 mL of sterile water and suspended

in 0.2 mL of lysis buffer (2% Triton X-100, 1% SDS, 100 mM NaCl, 10 mM Tris-HCl pH 8,0, 10 mM EDTA pH

8.0). Glass beads (0.5 mm diameter, 0.4 g) and phenol chloroform isoamylalcool (0.2 mL) were then

added. Samples were vortexed for 5 min and 0.2 mL TE pH 8.0 (10 mM Tris-HCl pH 8.0, 1 mM EDTA, pH

8.0) was added. After centrifugation (10 min, 16000 g) at room temperature, the supernatant was

transferred into a new tube and 0.4 mL chloroform was added. After centrifugation for 2 min at 16000 g,

the aqueous phase was transferred and 4 μL of 10mg/mL RNAse were added. Samples were incubated at

37°C for 15 min and DNA was precipitated with 1 mL 100% ethanol and incubated for 10 min at -20°C.

After centrifugation (10 min, 16000 g), the pellet was solubilized in 0.4 mL TE pH 8.0. The samples’ DNA

was quantified using a nanodrop spectrometer at 260 nm. DNA was loaded on a 0.8% agarose gel and

separated by electrophoresis according to (Sambrook et al. 1989) and visualized under UV irradiation after

staining the gels with ethidium bromide (1μg/mL).

127

5.2.7.2. Protein extraction and analysis of protein profiles

To analyze the proteins, the protocol described by Thabet et al. (Thabet et al. 2014) was followed. SDS-

PAGE, was performed with 10% (wt/vol) polyacrylamide gels as decribed by Laemmli (Laemmli 1970). 100

μg of proteins were loaded in each well. LC/MS (HPLC Ultimate 3000; Dionex coupled with LTQ Velos;

Thermo Scientific) was used to identify the proteins, followed by discoloration of the bands of interest (by

trypsin digestion). A second MS was undertaken for the 10 most significant peaks, and analysis through

Proteome Discoverer software (Thermo Electron). The Mascot software (v2.3) was finally used to perform

a UniP_Sacchar_cerev search, with the following criteria applied: MS/MS ion search, electrospray

ionization (ESI-TRAP), trypsin (digestion enyme), carbamidomethyl and oxidation (modifications), max 2

missed cleavages, peptide and fragment mass tolerance ±1.5 and ±0.6 Da, respectively, ion scores > 37,

P<0.01 (statistical identification significance).

5.2.8. Experimental Planning

The strategy for unveiling the complex inactivation mechanism was as follows. In principle, the photo-

Fenton process was stepwise constructed: first light only, then addition of H2O2 and finally addition of the

iron source. This construction was evaluated in three levels: i) inactivation efficiency measured by

cultivability, ii) localization of the damage (internal, external) and iii) identification of the targets deriving

from the various processes.


5.3.1. Preliminary assays in simulated wastewater

Initially, the efficacy of the photo-Fenton process in S. cerevisiae inactivation in simulated wastewater was

assayed (recipe and conditions identical to Chapter 4), and the results are presented in Figure 5.1.

Although solar light exposure in a 3-h period failed to decrease the viability significantly, the addition of

H2O2 inflicted a 2 to 3-log decrease as the addition increased from 10 to 25 mg/L. The addition of both Fe

(di- and tri-valent starting iron species) and H2O2 under illumination inflicted a 4 to 6-log reduction in the

population, indicating that the photo-Fenton process is a prominent technique towards the elimination

of this specific pathogen, even in the presence of organic matter and ROS scavengers (see chapter 4 for

composition and simulated WW characteristics). However, the focus of this work is to elucidate the

underlying mechanisms of S. cerevisiae inactivation, hence for simplification and interference avoidance

purposes, the following detailed study was conducted in MQ water.

128

Figure 5.1 – Overview of the photo-Fenton tests in simulated wastewater.

5.3.2. Cultivability assays – Efficiency of treatment in MQ water

Figure 5.2 summarizes the various photocatalytic inactivation tests carried out in the framework of this

work in MQ water. Figure 5.2a contains the heterogeneous and homogeneous photocatalytic systems

with the respective blank tests, while Figs. 5.2b and 5.2c depict the effect of pH on the efficiency of the

(homogeneous) photo-Fenton action.

More specifically, in Figure 5.2a the boundary conditions, concerning the oxidative stress applied to S.

cerevisiae are shown first ( trace). Hydrogen peroxide at high concentrations has been reported to have

detrimental effect on the survival of S. cerevisiae (Oyane et al. 2009). Normally, H2O2 acts on its cell wall

and plasma membrane, causing carbonylation and thiolation of surface proteins (Cabiscol et al. 2000,

Costa et al. 2002, Grant et al. 1999). The fungal wall is suggested to protect from diffusion of H2O2 into

the cell of F. solani (Sichel et al. 2009), with increasing thickness and efficacy as the cell ages (Sousa-Lopes

et al. 2004). Here, the control test was performed at 10 mg/L initial H2O2 concentration. As a result,

negligible inactivation was observed during the monitoring period (2 h).

129

Figure 5.2 – Overview of the photocatalytic inactivation tests and their respective controls. a) The

plots describe the cultivability evolution over time. b) Comparison between pH 5.5 and 7.5 for the

FeSO4-assisted photo-Fenton system. c) Comparison between pH 6.0 and 7.5 for the iron citrate-

assisted photo-Fenton system. Standard deviation < 5%.

When light was introduced to the system, significant inactivation of yeast cells was initiated ( trace).

The fungal kingdom is known to be affected by solar (Sichel et al. 2007a) or pulsed light (Takeshita et al.

2003), which essentially has similar mode of action, but higher intensity. After 30 min of minor

inactivation, resembling the shoulder period of inactivation found for E. coli (Giannakis et al. 2015a, b), a

linear (at log-scale) pattern was observed, without significant tailing. The necessary dose to initiate the

linear inactivation was 72 kJ/m2 and 288 kJ/m2 were required for total inactivation, respectively;

compared to other (multicellular) fungi, this is 40% less (Sichel et al. 2007b). The emitted light from the

solar simulator includes wavelengths at both UVB and UVA regions. Although UVB is notably less than the

UVA (0.5% compared to 5%) its biological effects are significant (Giannakis et al. 2014a, b). UV light affects

both external sites (cytoplasmic membrane) and internal ones, such as the enzymatic activity and the

genomic structure (Schenk et al. 2011). Pfeifer et al. (Pfeifer et al. 2005) have shown that artificially

irradiated cells present a higher percentage of UVB-induced lesions, namely CPDs (cyclobutane pyrimidine

dimers), rather than (6-4) photoproducts from type I or II reactions, that UVA is held usually responsible.

However, UVA can i) produce H2O2 (under certain conditions) by activating oxidases (Hockberger et al.

1999) ii) or, release iron from ferritins (Pourzand et al. 1999) and Fe/S clusters (De Freitas et al. 2003); this

ends up in initiating an internal (photo)Fenton process. Although DNA has low absorbance in the UVA

a)

b)

c)

130

region (maxima around 260) (Lindberg and Horneck 1991, Schenk-Meuser et al. 1992); strand breaks have

been reported, supporting this hypothesis, and guanine to cytosine transversions (Pfeifer et al. 2005).

Finally, apoptotic responses have been reported, and are initiated as a result of disrupted or altered yeast

life cycle or UV sensitivity (Chen et al. 2014, Del Carratore et al. 2002).

The last control test of the system assessed the combined action of simulated solar light and hydrogen

peroxide ( trace). Compared to the simulated solar light test, the required time for 6-log inactivation

has decreased almost in half (75 min instead of 120) and the shoulder length has also been decreased to

10 min from 30. In this experiment, notable consumption of H2O2 was observed compared to the dark test

(see Supplementary Figure S1). Since it is known that the cleavage of H2O2 with the subsequent formation

of hydroxyl radicals is unlikely to be achieved within the solar spectrum (Zapata et al. 2010), the possible

inactivation mechanism is as follows: the action of H2O2 and UV on the external proteins affects the

composition and stability of the membrane, and subsequent regional changes in permeability occur. Since

the H2O2 can now enter the cell, the internal Haber-Weiss reactions are enhanced (Spuhler et al. 2010).

The first step of the catalytic cycle involves reduction of ferric ion to ferrous (Eq. 5.1):

● (5.1)

The second step is the Fenton reaction (Eq. 1.8):

(1.8)

Net reaction (Eq. 5.2):

● ● (5.2)

This hypothesis is reinforced by data acquired by F. solani, where notable decrease of H2O2 occurred,

during UVA irradiation (Sichel et al. 2009), and entrance of H2O2 in the spore was highly probable. The

second action mode suggests a damage in the oxidative stress regulation mechanisms in the cell by the

action of UV light, such as SOD or CAT, followed by accumulation of ROS into the cell. If the SOD, the

responsible for regulation of superoxide radical ● inside the cell is no longer functional, it can lead to

the accumulation of ● . Superoxide radicals attack the DNA (Keyer and Imlay 1996), lead to the

accumulation of H2O2 and lately, ● has been suggested that in aqueous environments can lead to the

formation of hydroxyl radicals ( ●) (Ruales-Lonfat et al. 2015, Xu et al. 2013). CAT is also strongly

affected by light (Imlay 2008) and suspending its functions can cause internal stress by over-accumulation

of H2O2. Once the internal regulation mechanisms have been dropped, the cells cultivability is decreasing

dramatically, succumbing to the internal and external stresses initiated by UV and H2O2.

The remaining inactivation curves describe the photo-assisted Fenton processes undertaken in this study,

marked with the square traces. Iron in the form of FeSO4 ( trace), iron citrate ( trace) and iron oxides,

as goethite ( trace), were added to the solution 30 min prior to illumination, and H2O2 addition (and

131

light supply) indicated the initiation of the experimental assay time. According to the different starting

forms of iron in the solution, different responses were expected and subsequently monitored in the

system. In general, contrary to solar only or solar/H2O2 systems, an hour or less was required to achieve

total inactivation. In addition, apart from the goethite-powered photo-Fenton process, no significant

delay (shoulder) was measured; almost constant inactivation kinetics were observed. An estimated order

of efficiency according to the initial iron source was as follows: Goethite < FeSO4 < Citrate.

Goethite: Goethite is one of the most abundant forms of iron in nature (Ruales-Lonfat et al. 2015) and

can play a significant role in S. cerevisiae inactivation as heterogeneous catalyst in presence of H2O2

(Ruales-Lonfat et al. 2014). At near-neutral pH, when iron salts (such as the ferrous sulfate used here) are

added in presence of H2O2, zero-charge ferrous complexes [Fe(OH)2] are formed, which are very sensitive

to oxidation and rapidly form solid ferric (hydr)oxide compounds, such as goethite and lepidocrocite

(Ruales-Lonfat et al. 2014). Therefore, according to the process described in our previous work, goethite

was prepared (Ruales-Lonfat et al. 2015) and used in this study. It is necessary to differentiate the actions

attributed to iron oxides, at least for the cultivability assays, as their formation in near-neutral

environments is ubiquitous. As explained below, goethite contribution is divided to action as a photo-

excited semiconductor with the cells, or a heterogeneous photo-catalyst.

Goethite, in absence of H2O2 has demonstrated semiconductor properties, towards bacterial inactivation

(band gap: 2.1 eV). Even more, photo-activity has also been reported (Leland and Bard 1987). Hence, HO●

radicals can be formed only by presence of light, and the particles that could adsorb onto the yeast surface

due to negative charging (Dunlap et al. 2005, Polo-López et al. 2010) may contribute to surface holes and

aid in inactivation. Furthermore, apart from the oxidative damage from the holes, the iron oxides can lead

to damage on the cell surface through another semiconductor action mode. The goethite particles bound

in the surface of the yeast cell are illuminated and electron excitation is following. Oxygen plays the role

of electron acceptor, leading to superoxide radicals ( ● ) and subsequent reaction with water locally

produces hydroxyl radicals (Ruales-Lonfat et al. 2015). In our experiments, a stronger electron acceptor

(H2O2) is present and the heterogeneous (Fenton with Fe-oxide as iron source) action mode is more likely

to affect the inactivation process. Nevertheless, we cannot neglect the possibility of semiconductor

pathways’ contribution to cell damage.

In presence of H2O2, the following action modes are more prominent. Firstly, radicals’ production is

expected, through the excitation of electrons on the surface of the oxide, and transformation of H2O2

through the following reaction:

→ ● (5.3)

132

The other pathway includes a heterogeneous photo-Fenton mechanism for hydroxyl radical production,

where H2O2 is responsible for initiating a series of reactions, leading firstly to HO2● when in contact with

Fe3+ from iron oxides’ surface, reducing it to Fe2+. Fe2+ participates in the Fenton reaction and a cycle is

initiated, described by the following reactions (Ruales-Lonfat et al. 2015) (>Fe corresponds to the iron in

the surface of the goethite particle).

● (5.4)

● (5.5)

● ● (5.6)

● (5.7)

Finally, Fe3+ could form complexes with organic compounds in the cell wall (Polo-López et al. 2014, Polo-

López et al. 2013, Spuhler et al. 2010). Through ligand-to-metal charge (LMCT) transfer, Fe3+ is reduced,

leading to Fe2+ and initiating the Fenton reaction anew, in presence of light.

● (5.8)

Iron sulfate: FeSO4 was used as a starting source of iron, where Fe2+ is available to react with H2O2,

according to the following reactions:

(1.8) ● (5.9)

● (5.10) ● (5.11)

The treatment here takes place under (simulated) solar light. Hence, regeneration of the iron catalyst is

taking place according to the last two equations (Eqs. 5.10-5.11). Equation 5.11 takes into account the

possible complexes with organic by-products (R) deriving from the initial yeast cell wall attack, which can

maintain the iron available in solution for the reaction, as follows:

Fe2+ feeds the photo-Fenton cycle outside the cell, damaging the external membranes by the ROS

produced, in the same way the HO● radicals do due to heterogeneous photocatalysis with iron oxides.

133

However, after addition of FeSO4 at neutral pH, Fe2+ has a half-life of minutes before its oxidation in Fe3+,

and the subsequent precipitation as iron oxide (Barona et al. 2015). It also has been reported to diffuse

through cell walls (Polo-López et al. 2014, Spuhler et al. 2010), increasing the inactivation efficiency

through radicals’ production. Contrary to bacteria, S. cerevisiae do not produce any siderophores to

facilitate this transport, but there are i) surface binding mechanisms, ii) opportunistic use of siderophores

from other microorganisms (in real water samples) as well as reductive and non-reductive iron transport

mechanisms reported in literature (Gaensly et al. 2014, Lesuisse et al. 2001, Stearman et al. 1996, Yun et

al. 2000).

After diffusion, although iron does not affect the carrier proteins per se, it blocks their synthesis and the

transport energy resources (Khansuwan and Kotyk 2000); as a result, halving of the ATPase was observed.

Fe2+ is also proposed to bind to internal membranes to participate in the internal (photo-Fenton) process

and further promotes the superoxide radical production, ultimately increasing HO● radical production

(Sigler et al. 1998). The free iron inside saccharomyces is in Fe3+ form (Srinivasan et al. 2000), so reduction

to Fe2+ (and further HO● production) by LMCT or by the superoxide radical is also expected (Temple et al.

2005).

Finally, since the iron is not likely to remain in solution for long (Barona et al. 2015), the heterogeneous

action mode due to the formation of iron oxides is highly likely to be the driving force, after a certain

point. For simplicity reasons, we do not repeat here the heterogeneous and the semiconductor action of

the iron oxides as explained before; nevertheless their participation cannot be neglected.

Iron citrate: The experiments with complexed iron (by citrate ligand) simulate the mild chelating

properties of naturally available iron in natural conditions. These tests presented the fastest inactivation

rates for the operated pH region (pH = 6.0). In principle, the optimal efficiency for the photo-Fenton

reaction is found at pH = 2.8. If the experiments with FeSO4 were performed at that pH the efficiency

would have significantly improved. However, the acidification costs for treatment and the subsequent

neutralization necessary for use of the water after treatment demands viable solutions at the neutral pH.

Iron citrate has been successfully used in high pH (up to 8) for pollutants’ degradation (Katsumata et al.

2006, Trovó and Nogueira 2011) and bacterial inactivation (Ruales-Lonfat et al. 2016), therefore its

success is promising also for drinking water disinfection, since its related toxicity is very low (Silva et al.

2007). Here, the citrate complexes lead to higher solubility and stabilization of iron cations in the solution.

In general, the action mode of citrate can be categorized as a homogeneous Fenton promoter. The

generation of HO● radicals is induced by photoactive [FeOH-cit]– complex, which is the main species

formed at neutral pH (Chen et al. 2011); first a LMCT transition and then a Fenton reaction take place. The

(photoactive) [FeOH-cit]– complex will generate a ligand radical (HGA2●─) and Fe2+, which will in turn result

134

to superoxide radical anion and HO● production, respectively. This Fe2+ can also participate in the photo-

Fenton cycle mentioned before for Fe2+ from FeSO4.

pH dependence: In order to test the efficiency of the two processes initiated by FeSO4 and Fe-cit and

verify the extent of homogeneous, heterogeneous action mode and side-effects of iron addition, assays

in higher pH (7.5) were initiated. In Figs. 2b and 2c, the results for FeSO4 and Fe-cit are summarized. As a

rule of thumb, increase of the pH leads to faster precipitation of Fe3+, and lower inactivation rates are

expected. Nevertheless, for FeSO4 the reaction duration was not affected and 45 min were necessary to

completely inactivate the Saccharomyces. In pH 7.5, the oxide forms of iron are favored, and since the

overall inactivation rate was not affected, it indicates the significance of the heterogeneous process in

yeast inactivation.

The iron citrate inactivation was more affected by the pH increase. Since the iron and the citrate complex

were synthesized by a 1:1 ratio, after the ligand to metal charge transfer and citrate’s sacrifice, iron is

more likely to precipitate at neutral pH, rather than re-complexing with another citrate ligand. Therefore,

the homogeneous action is not affected. The difference in the inactivation rates could also probably a

consequence of the side-reactions influenced by the citrate ligand, which acts as an extra target for the

non-selective oxidative species generated. Our overall suggestion is that the reactions initiated by extra

iron intake are limited by the pH increase, and the subsequent oxide formation.

5.3.3. Flow cytometry results – Localization of damage

Very often, the viability assay through cultivation, works on the assumption that viable cells are the one

able to reproduce. Hence, cultivability is the required measurement. However, Davey in her recent review

emphasizes on the problematic dependence on these measurements, since microbial cells are not

classified only as “live/dead”, but cryptobiotic, dormant, moribund and latent states have been suggested

throughout the years (Davey 2011). Therefore, flow cytometry combined with CFDA/Propidium Iodide

staining has also been used in this study, for an assessment of the type and the extent of damage at yeast

cells. Indicative results of one process, are presented in Figure 5.3.

The staining protocol used allows the identification of living cells (appearing green in P2 quadrant) dead

cells (red in P4 quadrant), as well as the unlabeled and intermediate states in quadrants P1 and P5

respectively (Figure 5.3i). The green staining is an indirect indication of a living cell, since an acetoxymethyl

ester of the 5-carboxyfluorescein diacetate (CFDA) added is entering the cell through the membrane, and

once inside, gets hydrolyzed into acid and alcohol by non-specific enzymes (esterases) resulting in a

fluorescent green stain. On the contrary, the Propidium Iodide (PI) test indicates the non-viable cells. The

propidium ion is excluded from permeating the membrane and the loss of this ability suggests a loss of

135

viability. Even so, literature suggests cases that this is not definitive (Davey and Hexley 2011) and live cells

can get the red staining while maintaining their viability; up to 7% of these cells can still perform repair.

Figure 5.3 – Control tests and an indicative presentation of the flow cytometry results evolution,

during photo-Fenton reaction, at pH = 5.5.

Figure 5.3ii demonstrates the evolution of cell state in 30 min of exposure to light and the addition of the

photo-Fenton reagents, at pH = 5.5. This case is presented as an indicative test, and the results of the

other tests will be summarized instead (Figure 5.4). At time 0, 99% of the cells fall within the live state,

and 1% to the other states; we remind that 30 min have preceded all experiments before illumination, to

allow die-off of the most sensitive cells and acclimatization time for the rest of the cells. The intermediate

cell state appearing in quadrant P5 in Figs. 5.3ii-b and c, indicate vital cells but with compromised

membrane. As the exposure time passed and cells were subjected to the actions induced by the photo-

Fenton reaction, viability, as it is defined in this test, diminished within 30 min. For the rest of the tests,

136

the conventional graphs will be presented instead. By processing the data (number of cells) on CFDA and

PI staining for the first 30 or 45 min of the treatment, a correlation was found among the classic

cultivability assay and the flow cytometry results (more information can be found in the Supplementary

Material, Figure S2 and Table S1).

In our work, for each different test (hv, hv/H2O2, photo-Fenton with FeSO4 and Fe-cit as iron sources), flow

cytometry has been performed and the results were always juxtaposed with the cultivability assays

presented before, for comparison. An overview of the CFDA decrease and the PI increase during the

various experiments: hv, hv/H2O2, photo-Fenton (FeSO4 - pH = 5.5 & 7.5, Fe-cit – pH = 6.0 & 7.5) can be

found in the Supplementary Figure S3, while the analytical data will be presented in Figure 5.4. Following,

an analytical explanation of the results of CFDA, PI and the intermediate cell state is presented, with

suggestions on the inactivation mechanisms for each case. As this process is constructed in a step-wise

manner, the explanations and the mechanisms suggested in the first steps (solar degradation and H2O2

oxidation), will not be repeated in the more complex systems.

Simulated solar light: During exposure of S. cerevisiae to simulated solar light, the changes in esterase

activity and membrane permeability were recorded and summarized in Figure 5.3a. Firstly, we compared

the time necessary for 50% and 90% reduction of the initial microbial load. As it appears, the cultivability

assays indicate a required time somewhat higher than 30 min. However, CFDA and the PI-stained cells are

almost in agreement (6% intermediate state cells) at 50 minutes. For 1-log reduction there is a small

convergence between the two methods, with 10 min difference instead. Nevertheless, the remaining

population is consistently underestimated by the cultivability assays, and viable cells that are not able to

form colonies are left out of the estimation.

A critical point in this study is the timeframe of 45 min (Figure 5.4a), where the number of viable cells, but

with compromised membranes is the highest (11.5%). Therefore, the principal mechanism of light-

induced inactivation is an internal process. More specifically, the proposed mechanism is as follows:

1) If we consider either CFDA or PI staining as the accurate viability assay, and not cultivability, the

inactivation presents a shoulder, a latency (in linear scale). Therefore, there should be either an

accumulation of damage before inactivation or the specific type of damage can be repaired.

2) The low accumulation of purple stains during flow cytometry (11.5% versus 26% PI-stained or

generally 37.5% esterase inactive), suggests that lipid peroxidation is limited, and external

proteins are rather intact. Also, it could signify that the cells are inactivated through a failure in

their internal functions.

3) The above indicate that this result is probably, the effect of the actions of UVB and UVA light on

(nucleic/mitochondrial DNA and internal enzymes, respectively.

137

hv/H2O2: The flow cytometry results for the hv/H2O2 treatment of S. cerevisiae are summarized in Figure

5.4b. In principle, the shapes of the curves are very similar to the ones already shown during the solar-

only exposure. The cultivability decreases rapidly in time, and the 4-log reduction time dropped to 30 min

(instead of 60). However, the CFDA/PI staining levels presented the same lag in the beginning, with

different time to reach >99.5% (30 instead of 45 min). Also, the intermediate cells were at the same

percentage with the solar-only process. The main difference that modifies the inactivation mechanism

lies in the higher levels of non-viable cells:

1) The latent character of inactivation indicates the necessity to accumulate damage prior to

inactivate the cells. The shoulder length is again around 30 min, indicating similar accumulation

of photo-induced damage.

2) The low number of intermediates could mean a low peroxidation-related killed cells. Hence, there

are not many vital cells with compromised membrane.

3) In 45 min, there were significantly more inactivated cells compared with the solar-only process,

with the same number of intermediate cell-states. Therefore, there is a synergy in the action of

light and H2O2.

4) Since the suggested mechanism is the CPD-formation in DNA and enzymatic failure (CAT, SOD),

the mutations and enzymatic activity loss could increase the permeability of the membrane, thus

enhancing the H2O2 diffusion into the cell.

5) The entry of H2O2 in the cell enhances the oxidative damage, now unable to be controlled by SOD,

or even more, enhancement of the internal Fenton process taking place into the cell.

Photo-Fenton with FeSO4, at pH = 5.5 and 7.5: Figs. 5.4c and d summarizes the evolution of cultivability,

plus CFDA and PI staining through time. The photo-catalytic process is profoundly more efficient than

solar, or hv/H2O2. The time mark of 15 min offer a solid ground for comparisons. At pH 5.5, the level of

viable cells is 86%, compared with the 94% of the higher pH. Also, 4-log of cultivability loss occurs at 20

min (here recorded at 30 min for pH 7.5 due to sampling interval settings) and total inactivation takes

place at 30 min.

For the samples treated at pH = 5.5, there is a latent period marked with CFDA staining but probably the

levels of viability are similar (Figure 5.4c); the cultivability is almost equal but the marked difference is at

the intermediate cells, were 5% and 40% yeasts with compromised membranes appear, for pH 5.5 and

7.5, respectively. At 20 min, the peak of intermediate cells appear for pH 5.5, reaching 30%, and 67% dead

cells (PI stain) but the striking difference at pH 7.5 is that even 5 min earlier, the peak of intermediate cells

is reached, almost double in quantity (55%) and only 6% dead cells (Figure 5.3d). Also at 30% where the

inactivation is almost complete, the number of vital but membrane-compromised cells is higher. Hence,

the proposed inactivation mechanism is as follows:

138

1) Although a shoulder appears in the beginning, and a build-up in damage is required, the PI stain

indicates similar death rates. Therefore, at the early stage the inactivation is similar, and probably

attributed to the same mechanism, of the homogeneous photo-Fenton action mode.

2) The markedly high difference after 15 min of treatment (40-50% live intact cells) but the notable

difference of compromised membranes shows that the action mode is probably the external

photo-Fenton, and the production of reactive oxygen species is responsible for membrane

peroxidation.

3) At pH = 5.5, total inactivation is achieved faster than at pH 7.5 (>99% of PI stained cells). This

result is logical if the main mechanism is the external photo-Fenton, which is more profound at

lower pH due to favorable form of iron and higher dissolution levels. However, the maintenance

of iron as Fe2+ and its soluble form explain the difference in intermediate cells. The diffusion of

iron in the cell is higher, and the internal photo-Fenton process is enhanced. For this reason,

inactivation is not heavily dependent on the external ROS action, marked by the higher red

staining levels and lower membrane-compromised, compared to the process at pH 7.5.

4) At pH = 7.5, iron is either participating in heterogeneous photo-Fenton reaction in form of oxides,

whose efficacy is lower, or by attachment to the cell surface. The notable difference in cell

integrity, due to the inflicted damage can be explained by reactions taking place at the cell

surface, such as LMCT between iron forms and yeast cell wall, due to local contact-related

promotion of photo-Fenton and subsequent damage at the proximity of a surface.

Iron citrate-driven photo-Fenton reaction at pH = 6.0 and 7.5: Figs. 5.4e and f summarize the results of

the iron-citrate assisted photo-Fenton processes. As seen also in the previous section, the inactivation

measured by cultivability, among the Fe-citrate-fueled experiments did not differ significantly. In Figure

5.3e, at around 20-25 min of treatment, up to 4-log reduction has been achieved and at 30 min, almost

complete loss of cultivability. The viability assay differs significantly from the cultivability once more.

Initially, there is a lag period, 15 and 30 min for the experiments at pH 6.0 and 7.5 respectively, as

measured by the loss of esterase activity, due to the similarities in the absorption spectrum among yeast

and the citrate complex (Robertson et al. 2013, Ruales-Lonfat et al. 2016, Ułaszewski et al. 1979). Even by

monitoring the PI staining, an initial 10-min vs. 15-min period of latency was followed by rapid increase,

to lead in 30 vs 45 min period for >99% viability loss. The peak of the intermediate states was noted at 15

minutes for pH = 6.0, whereas at 30’ for the 7.5 pH experiments. Finally, there is a notable difference at

these last time points, where for pH = 6.0, at 15 min, the composition is 25% viable, 15% dead and 60%

viable, but with compromised membranes. If this point characterizes the maximal damage point, for pH

7.5 at 30 min (Figure 5.4f), the corresponding composition was 32%, 22% and 45%, respectively. This

indicates that the damage made was rather internal, and less related with cell membrane lesions.

However, this 15-min delay has to be taken into account in the mechanism suggestion that follows:

139

1) The lag is explained by the competition of yeast cells and the citrate complex, as they absorb in

the same wavelengths. Hence, a delay in the ROS production is expected, and lower peroxidation

of membranes.

2) At pH = 6.0 (compared to 7.5), the solubility of iron (Fe2+) is higher and therefore, a higher fraction

is expected to participate at the homogeneous Fenton reaction. Consequently, higher ROS

production and higher damage is recorded.

3) Also, the fast inactivation and the concomitant increase of dead and injured cells indicates a

notable participation of the internal mechanisms of inactivation (through Fe2+ penetration).

4) At pH = 7.5, the delay expressed alters the proportions of the dead cells compared to the live

ones; at the point where the intermediates reaches its maximum, more PI-stained cells appear. A

heterogeneous action mode is probable as well.

5) Generally, there is less iron intake compared with the processes initiated by FeSO4. This is

probably attributed to the initial form of iron (complexed Fe3+, compared to free Fe2+). The

importance of damage as depicted by the dead fractions and the time achieved is appearing to

be more dependent on the internal, and less in the external photo-Fenton reactions.

140

Figure 5.4 – Flow cytometry results. Control tests: a) Simulated solar light only. b) hv/H2O2 system.

FeSO4–assisted photo–Fenton processes: c) pH = 5.5. d) pH = 7.5. Fe-cit–assisted photo–Fenton

processes: e) pH =6.0. f) pH = 7.5. Standard deviation < 5%.

e) f)

c) d)

a) b)

141

5.3.4. Identification of targets – Nuclear DNA, cell wall and cytoplasmic protein damage

In order to further shed light in the suggested mechanisms of S. cerevisiae inactivation, an assessment of

the damage on nuclear DNA and the proteins of cell wall and cytoplasm, was performed. The analysis of

DNA fragmentation was monitored by gel electrophoresis, and the results are summarized in Figure 5.5,

while Figure 5.6 a) and b) show the results of the cell wall and cytoplasmic protein damage, respectively.

Protein damage is analyzed by SDS-PAGE and Coomasie blue staining. Simulated solar light only or

combined with H2O2, and the two main photo-Fenton systems were compared.

Figure 5.5 – Nuclear DNA damage in the four different systems. Comparison of the pH effect in FeSO4-

assisted photo-Fenton systems.

In the case of simulated solar light, as it is clearly depicted (Figure 5.5), DNA damage progresses over time,

showing a dose-dependence with exposure. After 120 min of treatment, the fragmentation levels were

too high (trace disappears). The change among the state at 60 and 120 min corroborates with the acute

loss of viability recorded via the CFDA and PI staining. The cell wall proteins pool profile appears intact

(Figure 5.6a), and negligible degradation is was detected; the intermediate cell state was also very low, as

measured by flow cytometry in Figure 5.3a. Nevertheless, after 120 min there is a notable reduction of

cytoplasmic proteins, which was not high until 60 min of exposure (Figure 5.6b). Therefore, a double

action is the probable pathway where i) DNA is damaged severely, no repair/defense mechanisms are

able to be deployed against the (mild) internal photo-Fenton action (whose damage becomes later

significant), or ii) concomitantly with the DNA destruction, the internal photo-Fenton action is becoming

profound.

0’ 30’ 60’ 120’ 0’ 60’ 120’30’ 45’ 0’ 60’ 120’30’ 45’ 0’ 60’ 120’30’ 45’

hv only hv/H2O2 hv/FeSO4/H2O2at pH = 5.5

hv/Fe-cit/H2O2at pH = 6.0

Nuclear DNA damage Comparison:pH 5.5 & 7.5

hv/FeSO4/H2O2at pH = 5.5

Nuclear DNA damage

hvh /F/FeSOSOvv 4/H/H2OO2 hv/FeSO4/H2O2at pH = 7.5

0’ 60’ 120’30’ 45’ 0’ 60’ 120’30’ 45’

142

Figure 5.6 – Cell wall (a) and cytoplasmic proteins damage (b) in the four different systems. (i-ii):

Comparison of the pH effect in FeSO4-assisted photo-Fenton systems.

Cytoplasmic Protein Damage

170130

95

72

55

43

34

26

1710

0’ 60’ 120’

30’ 45’’’’170

1309572

55

43

34

26

1710

0’ 60’ 120’

30’ 45’

170130

95

72

55

43

34

26

1710

0’ 60’ 120’

30’170130

95

72

55

43

34

261710

0’ 60’ 120’

30’ 45’

hv only hv/H2O2





Comparison: pH 5.5 & 7.5

Cell wall proteins

170130

9572

55

43

34

26

1710

170130

9572

55

43

34

26

1710

170130

95

72

55

43

34

26

1710

170130

9572

55

43

34

26

1710

0’ 60’ 120’30’ 0’ 60’ 120’30’

0’ 60’ 120’30’ 0’ 60’ 120’30’

hv only hv/H2O2





Comparison:pH 5.5 & 7.5

a)

b)

ii)

ii)

i)

i)

143

The addition of H2O2 in the system resulted in similar DNA damages (Figure 5.5). The presence of

intermediates cells cannot really be linked to DNA and protein damages visualized by electrophoresis. The

results of the cell wall proteins’ damage presents the first notable, but limited damage (Figure 5.6a), which

probably does not lead to loss of viability, but is presented only as 3% increase of intermediate cell states.

However, the results of the cytoplasmic protein damage indicate the participation of H2O2 to the internal

mechanisms, as the staining is less intense and more bands disappeared completely (Figure 5.6b). Hence,

the increased inactivation kinetics are a result of DNA damage, combined with limited cell wall

peroxidation and increased internal photo-Fenton action, after the regulatory mechanisms fail to cope

with the high accumulation of H2O2 and the rest of the ROS in the cell.

Adding iron to the system has a detrimental effect in all levels. First of all, DNA damages were extensive,

with disappearing bands at 45 and 60 min, for FeSO4 and Fe-cit respectively (Figure 5.5). These time frames

are significantly lower than the ones recorded for the solar and hv/H2O2 systems. Probably there is an

increased loss of membrane integrity (as seen in Figs. 5.4c-f) and oxidative damage inside the yeast cell.

As it appears on the damage of the cell wall proteins, even at time 0 the staining has a different, lighter

pattern, compared with the previously presented systems. This disappearing pattern is deteriorating

slowly (until 60 minutes) and accelerates afterwards (Figure 5.6a). However, since the cytoplasmic

proteins are actively getting degraded after 60 min for FeSO4 and Fe-cit, the strong dependence on the

internal contribution is verified (Figure 5.6b).

Cultivability is lost even before the damage (at any level) is highly accumulated. Furthermore, flow

cytometry suggested that the loss of viability was accelerated after 15-20 min for the iron assisted systems

(Figs. 5.4c-f), but the protein damage is highly notable after 30 minutes. For the iron-supported systems,

the order of events in the form of a suggested timeline can be found in Table 5.1. The loss of cultivability

is related with the first and direct oxidative stress conditions that the cell is faced with. The simultaneous

damage of DNA and cytoplasmic proteins, verify our initial hypothesis that the driving force in S. cerevisiae

inactivation is the internal photo-Fenton process. For the two iron salts, the difference in the appearance

of this effect is related with the ROS production in the system; FeSO4 at pH 5.5 presented significantly

higher activity. Then, as these electrophoresis processes are quantitative, the cell wall proteins appear as

the last to degrade because they comprise a very big part of the total mass of the cell. This is also explained

by the fact that in all figures depicting the protein damage, the bigger ones (highest kDa values) are

targeted first. Finally, in order to notice oxidative damage inside the cell, only a small part of the

membrane must be breached, enabling the introduction of iron into the cell (extra amount added to the

diffusing one). The fact that the flow cytometry data indicate more than 50% viable but compromised

membrane-cells, is the signature of the breach and the subsequent internal action.

144

Table 5.1 - Timeline of the inflicted damage in the corresponding targets of the different iron-assisted

systems

Event FeSO4 Fe-cit Loss of cultivability 10-20 min 15-20 min

Loss of viability (by CFDA or PI) 10-30 min 10-30 min

DNA damage 30-45 min 45-60 min

Cytoplasmic proteins damage 30-45 min 45-60 min

Cell wall proteins damage 60-120 min 60-120 min

At a higher pH for the FeSO4 system, where solid Fe-oxides (goethite) is present, weaker, less active

systems are involved. The DNA damages were delayed, compared to the corresponding system at 5.5 pH.

However, a possible contribution of the iron oxides is noted, when the internal and external protein

damage is compared. When the pH of the solution promoted the oxide formation, increased cell wall

degradation was observed. Accordingly, lower internal protein damage has been recorded. The

combination of these two events indicate a contribution of the iron oxides at the degradation of the cell

membrane, either by promotion of local oxidative species at the surface of the (attached to the cell) iron

oxide, by the oxide itself acting as a semiconductor and causing local damage to the structure, or by

heterogeneous photo-catalysis in presence of H2O2.

5.3.5. Holistic proposal for the inactivation mechanism of S. cerevisiae

Having analyzed the cultivability and flow cytometry results, as well as the DNA and protein damages,

together with an extensive literature review, a mechanism proposed for the inactivation of S. cerevisiae

is given in Figure 5.7. The four sub-figures represent distinct actions, as before: a) simulated solar light

alone (here, direct action), b) hv/H2O2 (+indirect light actions, important at this stage), c) FeSO4-assisted

(including oxides) and d) Fe-cit-assisted photo-Fenton systems. Once again, the step-wise construction of

the complex photo-Fenton mechanism is not repeated; the actions of the solar-only system and the

hv/H2O2 are present in the photo-Fenton actions, but not presented again, for simplicity. Also, the

references that support our observations and the subsequent suggestions are not repeated again in this

section, as they were discussed before. A brief explanation on the different actions per sub-figure follows:

a) (Direct) Solar light: (1) Simulated solar light induces mutations at DNA level (e.g. CPDs, (6-4)

photo-products, guanine to cytosine transversions) at both nucleic and mitochondrial genome (2). Also,

mainly UVB and in a lesser extent UVA induce damage at the external cell surface (3). The cell structures

suffer damage at external and internal level (4): internal groups and structural functions (enzymes,

clusters, proteins etc) are physically affected by the direct illumination. Cell death can be the final

outcome of the loss of viability.

145

b) (Indirect) hv/H2O2: Light affects internal enzymes, such as superoxide dismutase (SOD) and

peroxidases, such as catalase, oxidase etc. (1) and (2). As a result, the internal respiratory chain is affected,

with subsequent H2O2 and ● accumulation. Since Fe/S clusters are affected, release of iron is expected,

and an internal Fenton mechanism is initiated with the hydroxyl radicals damaging internally the cell. H2O2

plays both the role of oxidant for Fe2+ and the reductive for Fe3+, and since light is present, internal photo-

Fenton is taking place. The superoxide radical also participates in iron reduction and the internal oxidation

actions. As far as the external actions are concerned, mild peroxidation of the external cell wall proteins

at small extent is expected (3) and possible penetration of H2O2 into the cell (4). These actions further

enhance the internal photo-Fenton actions.

Figure 5.7 – Mechanistic proposition of the pathways towards yeast cell inactivation. a) (direct)

Simulated solar light. b) (Indirect) hv/H2O2. c) FeSO4–assisted photo–Fenton process. d) Fe-cit–assisted

photo–Fenton process.

c) FeSO4-assisted photo-Fenton: The addition of FeSO4 in the solution provides (for a limited period

of time) Fe2+, which produces radicals externally (1). Fe3+ complexes are reduced by light, further

producing hydroxyl radicals and Fe2+. The rupture of the cell wall can allow the penetration of Fe2+ and

Fe3+ into the cell, enhancing the Fenton reactions taking place inside the cell (2). Also, the iron transport

mechanisms carry Fe2+ and Fe3+ into the cell, to maintain homeostasis, with the same effect as before (3).

However, the presence of dissolved oxygen in the sample and the near-neutral pH cause the oxidation of

iron to goethite and lepidocrocite; the heterogeneous Fenton action is initiated (4). The attachment of

146

the (positively charged) iron to the (negatively charged) cell wall induces local damage to the cell through

either (5) a photo-assisted reduction of Fe3+ (LMCT) on cell wall and the initiation of a photo-Fenton action

(since H2O2 and light are present and Fe2+ is produced) or (6) semiconductor action mode by the iron oxide,

which includes oxidative damage from the hole (h+) and the excitation of electrons (e-). These electrons,

in presence of oxygen or H2O2, acting as electron acceptors, generate ROS that cause extra oxidative

damage in the exterior of the cell.

d) Fe-cit-assisted photo-Fenton: Iron citrate is stable in water, in absence of light (equilibrium 1).

Light on the Fe-citrate, on the other hand, induces another pathway, involving LMCT and a [Fe2+─cit2─●]

product. Its dissociation gives Fe2+ and cit2─●, which can participate in the Fenton reaction (2) and

formation of superoxide radical anions (3), respectively. Alternatively, the cit2─● can react with H2O2

(4).The end-product is a hydroxyl radical, which inflicts external damage to the cell.

5.4. Conclusions

A eukaryotic, unicellular microorganism (S. cerevisiae) was subjected to a multi-level, systematic

investigation on its inactivation mechanisms. The contribution of photo-Fenton and its constituents were

put under study, and light was shed on the separate or synergistic pathways participating in yeast

inactivation.

The cultivability results indicated the best conditions and starting iron forms, to achieve the best

inactivation rates. Furthermore, flow cytometry data coupled with the electrophoresis data on DNA and

protein suggested the pathways towards inactivation. A significant contribution of the internal photo-

Fenton process was measured, in addition to the external oxidative stress by the ROS produced.

In principle, the action of light was monitored to affect mainly the DNA, and secondarily, the internal

proteins. The sequence of events suggests a photocatalytic-like induction of damage. When H2O2 was

added to the system, the non-viable cells were a result of increased internal photocatalytic reactions,

when compared with bare light. The addition of iron greatly enhanced the process, reducing the

inactivation time significantly. The generation of ROS inside and outside the cell reduced the viability by

destroying DNA and internal proteins, and when the process was prolonged, total destruction of the cell

was monitored.

According to the pH levels, iron oxides participate in heterogeneous pathways. Efficient photo-Fenton

inactivation was observed at pH 7.5, and in cellular level, a mixed mode between the diffusion of iron into

the cell and the damage caused from iron particles attached to the surface of the cell. Finally, iron citrate,

a relatively cheap organic complex was investigated, to increase the applicability of the process. The

significant inactivation measured indicates promising application potentials.

147

In overall, a wide view into the pathway of S. cerevisiae inactivation was given, helping understand the

inactivation of more complex microorganisms. Nevertheless, its complexity did not offer great resistance

against the photo-Fenton reaction, but can offer great insight on the function of eukaryotic cells when

present in similar oxidative stresses.

149

PART 3

Degradation of hospital PhACs by AOPs, as a point-source treatment

option in HWW and urine: treatment optimization and degradation

pathway elucidation

151

6. Chapter 6 - Iohexol degradation in wastewater and urine by

UV-based Advanced Oxidation Processes (AOPs): Process

modeling and by-products identification.

Published work:

Stefanos Giannakis, Milica Jovic, Natalia Gasilova, Miquel Pastor Gelabert, Simon Schindelholz, Jean-

Marie Furbringer, Hubert Girault, and César Pulgarin. "Iohexol degradation in wastewater and urine by

UV-based Advanced Oxidation Processes (AOPs): Process modeling and by-products

identification." Journal of Environmental Management (2016), pp. XXX - XXX

Web link:



Appendix E



152

6.1. Introduction

During the last decades, research in wastewater (WW) treatment has focused on the elimination of

emerging contaminants. Having efficiently tackled the classical WW issues of macro-pollution (organics,

phosphorus, nitrogen etc), combined with the leaps in analytical chemical capabilities, micropollutants

are the hottest topic in WW treatment for the last 15 years. These substances are comprising an increasing

list of anthropogenic contaminants, which include among others, pharmaceuticals, personal care

products, steroid hormones, industrial chemicals, pesticides and many other emerging compounds (Luo

et al. 2014). The polymorphism and the diversity of the chemical pollutants, state this topic as high priority

for the treatment facilities.

A distinctive category of micropollutants are the Pharmaceutically Active Compounds (PhACs), and

especially the ones that are exclusively administered from hospitals. Verlicchi et al. (Verlicchi et al. 2012)

analyzed 73 PhACs from 12 different therapeutic classes, and they concluded that most compounds are

found in consistently higher concentrations in HWW than in urban WW. The antibiotics, analgesics and

lipid regulators were the most concentrated, and 9 compounds posed a high risk at the concentrations

detected in hospital effluent and 5 in urban WWTPs influent and effluent. Ort et al. (Ort et al. 2010)

analyzed 59 PhACs in HWW, from which 2 had a contribution higher than 15% of the total of WWTP

influent. Within the “Pills-Project” (Pills-Project 2012) 16 key substances out of eight substance groups in

different hospitals were studied, concluding that 5 of these groups are exclusive contributors in 10 to 60%

of the load found in urban WWTPs, and are attributed to hospital use.

Among these categories, the Iodinated Contrast Media (ICM) are identified as a major threat. In a recent

research in a University hospital in Lausanne, Switzerland, contrast media contributed in 59% of the total

PhAC load. Globally, the consumption of ICM is 35.000 tons/year (Sprehe and Geissen 2000), from which

95% are excreted unchanged from the human body. Their use in Germany has been estimated at 500 tons

of ICM per annum (Haiß and Kümmerer 2006), while in Switzerland, the total consumption per year of

ICM is estimated at 35 tons (McArdell et al. 2011). In the case of a standard Swiss University hospital, the

consumption is of 1149 g/day and 725 g/day only for Iohexol (Weissbrodt et al. 2009); half of the

consumed amount is rejected in HWW. This amount is reached because the individual dosage in imaging

treatment is up to 300 g (Haiß and Kümmerer 2006). ICM are non-biodegradable and only partially

removed in WWTP, so their concentration in surface and drinking water is increasing, with the

concentration of Iohexol in Lake Leman, Switzerland being 0.03 μg/L (Chèvre 2014), or Iopamidol, which

was found in groundwater up to 2.4 μg/l (Ternes and Hirsch 2000).

Since the proven incapability of the existing WWTPs to handle ICM is established, their discharge to

natural water bodies will contain the aforementioned amounts (Pérez and Barceló 2007, Putschew et al.

153

2000, Ternes and Hirsch 2000). Advanced Oxidation Processes (AOPs) have long been suggested as an

alternative in treating non-biodegradable compounds (Herrera et al. 1998, Pulgarin and Kiwi 1996) and

many works have demonstrated the efficiency of AOPs against ICM. For example, for various ICM, such as

Iopamidol, Iomeprol, Iohexol and others, very good removal was attained at basic pH (Seitz et al. 2008),

or exposure to gamma irradiation (Jeong et al. 2010), while UV/TiO2 although it requires higher treatment

times, is a treatment that can achieve high removal rates (Borowska et al. 2014, Doll and Frimmel 2004,

Sugihara et al. 2013). Nevertheless, the works that have addressed the removal in matrices as hospital

wastewater or urine are scarce.

In this work, Iohexol has been chosen as a model non-ionic ICM, and has been subjected to extensive

investigation concerning the various UV-based processes (UV, UV/H2O2, UV/H2O2/Fe2+). In order to assess

the feasibility of the treatment by UV-based AOPs, an engineering approach has firstly been made, where

the investigation focuses on the matrices that Iohexol can be potentially found and measure the effect of

the operational parameters in achieving 90% degradation of the initial amount in water, wastewater and

urine. Also, in the view of treatment optimization by these AOPs, a modeling approach has been made for

the reactants addition and pH. Finally, in order get insights on the structures affected by the different

components of the process, the degradation pathway is studied for the three AOPs applied.


6.2.1. Chemicals and Reagents

Table 6.1 presents the composition of the synthetic urine and synthetic wastewater employed in most

experiments; the chemicals were used as received. Iohexol (Histodenz), hydrogen peroxide (30%) and iron

sulfate heptahydrate, used for the degradation experiments, as well as KCl, Peptone, CaCl2·2H2O and

MgSO4·7H2O were purchased from Sigma Aldrich (Switzerland), NaCl, Na2SO4, Meat Extract and NH4Cl

were acquired from Fluka (Switzerland), KH2PO4 and K2HPO4 from Merck (Switzerland), while urea and

creatinine from ABCR (France). Finally, titanium oxysulfate for the colorimetric determination of H2O2 and

Ferrozine for iron detection were purchased from Fluka.

Table 6.1 – Synthetic matrices composition.

Synthetic Urine Synthetic Wastewater Name Chemical

formula SUR

composition [g/L]

Name Chemical formula

SWW composition

[mg/l]

154

Urea CH4N2O 25 Peptone - 160 Sodium Chloride

NaCl 2.925 Meat Extract - 110

Sodium Sulfate Na2SO4 2.25 Urea CH4N2O 30

Potassium chloride

KCl 1.6 Dipotassium Phosphate

HK2PO4 28

Potassium phosphate monobasic

KH2PO4 1.4 Sodium Chloride NaCl 7

Calcium Chloride

dihydrate

CaCl2·2H2O

1.103 Calcium Chloride dihydrate

CaCl2·2H2O

4

Creatinine C4H7N3O 1.1 Magnesium Sulfate

Heptahydrate

MgSO4·7H2O

2

Ammonium chloride

NH4Cl 1

6.2.2. Reactors and experimental apparatus

Three “merry-go-round” reactors were used for the Iohexol degradation experiments, presented in

Supplementary Figure S1. These double coated glass vessels recirculate water at 22°C (Neslab RTE-111

thermostat), for the protection of the UV-C equipment. UV-C light at 254 nm (5x10-3 mW/cm2) was

supplied to the system by low pressure (LP) mercury discharge lamps (Philips TUV 11W/G11 T5 UV). The

lamps were placed in the interior of quartz glass and then submerged in the solution inside the reactor.

Mixing is ensured by a magnetic bar at the bottom of the reactor and the placement of the apparatus on

a magnetic stirrer.


6.2.3.1. Iohexol determination

The determination of Iohexol concentration was achieved through HPLC analysis (HP Agilent 1100 Series),

including a G1315A diode array detector, set at 254 nm. The HPLC method was as follows: The mobile

phase was held at an isocratic mode during all the analysis and consisted in the mixture of 95 % ultrapure

(Mili-Q) water with 0.1 % of formic acid (phase A) and 5 % of methanol (phase B). The flow was 1 ml/min

and the temperature of the column is 40 °C and the injection volume was 50 μL. This configuration led to

a retention time of 10.15 min, with a C18 reverse phase column (Merck Lichrospher 100 RP-18, 5 μm, 250

· 4 mm).

155

6.2.3.2. H2O2, Fe, COD and TOC measurements

A Shimadzu UV 1800 spectrophotometer was used for the colorimetric determination of H2O2 and iron.

H2O2 was quantified by adding 10 μL of Ti(IV) oxysulfate in 1 mL of sample and subsequent measurement

at 410 nm (DIN 38402H15 method). In some experiments, due to color interferences, Merck Milipore

peroxide detection strips were used for semi-quantitative measurement of H2O2 (measurement ranges:

<1. 1-3 and 3-10 or <1, 1-5 and 5-10 ppm). They were employed to detect residual (<10 ppm) of H2O2 in

real WW samples and H2O2 in high concentrations of Iohexol (color interference). Dissolved iron was

followed with the Ferrozine method, as described elsewhere (Viollier et al. 2000). Briefly, after filtration

of a 5 mL sample (0.22 μm), 0.2 mL of hydroxylamine hydrochloride, 0.2 mL of acetate buffer at pH 4.65

and 0.1 mL of 10 mM Ferrozine solution were added in the bulk. The iron determination took place by

spectrophotometric measurement of the magenta color formation at 562 nm.

Pre-acquired COD (HR/LR vials, HACH Lange) were used to determine the chemical oxygen demand, and

the corresponding colorimetric methods were used, measured by a HACH DR3900 Spectrophotometer.

Total organic carbon and inorganic carbon of the samples during treatment were followed by a Shimadzu

TOC-VCSN analyzer, with an ASI-V automatic sampling module. Finally, pH was measured with a Seven

Easy pH meter (Mettler-Toledo).

6.2.3.3. Orbitrap MS analysis for determination of the degradation pathway

The products of Iohexol degradation were analyzed by HPLC-HR-MS. Prior to the analysis the samples

were desalted using C18 SPE spin columns (Pierce Biotechnology, Rockford, IL, USA) following the

manufacturer protocol. Desalted samples were separated using Dionex UltiMate 3000 UPLC system with

Nucleodur C18 Gravity-SB precolumn (4 x 2 mm, 1.8 m) and Nucleodur C18 Gravity-SB separation column

(4 x 2 mm, 1.8 m, Macherey-Nagel, Düren, Germany). The solvent A was composed of water with 0.1 %

FA, while solvent B contained acetonitrile with 0.1 % FA. The flow rate of the mobile phase was set to 250

l/min. The gradient consisted in the linear increase of solvent B percentage from 3 to 60 % within 11

min. The sample injection volume was set to 35 l. The order of events were i) column activation:

MeOH/TFA 50%/0.01%, ii) column equilibration: 5% ACN 0.5% TFA, iii) sample binding: 150 μL sample, iv)

washing: 5% ACN 0.5% TFA and v) elution: 80% ACN 0.1% TFA.

The MS was performed using a Q Exactive-HF-Orbitrap MS instrument (Thermo Scientific, Bremen,

Germany) in positive ion mode within 200-1000 m/z. For all LC-MS runs survey scans were acquired with

15000 resolution (at 400 m/z), automatic gain control (AGC) value of 3E6 and maximum injection time of

100 ms. Dynamic exclusion duration for the precursor ions was set to 30 s. Top 5 data-dependent MS/MS

scans were recorded also with 15000 resolution (at 400 m/z), AGC value set at 1E5 and maximum injection

time of 50 ms. The isolation width for the precursor ion was set to 1.8 m/z. Higher-energy collision induced

dissociation (HCD) was used for the fragmentation of isolated precursor ion with normalized collision

156

energy (NCE) of 26 % and minimum signal threshold for MS/MS triggering was fixed to 20000 counts. The

obtained data were processed using Xcalibur software (3.0.63 version, Thermo Scientific, San Jose, CA,

USA).

6.2.4. Water matrices and treatment conditions

All experiments were carried out in triplicates (3 reactors) and in different categories of matrices, i.e.

water, wastewater and urine. The systematic studies took place in ultrapure water (MQ), synthetic

wastewater (WW) and synthetic urine (UR). The graph data represent the average, with <5% standard

deviation in the majority of cases (error bars not shown). The operational parameters tested were the

following: i) Specific matrices investigated: Mili-Q water (MQ), synthetic WW, diluted synthetic WW,

untreated (biologically) WW, secondary WW, synthetic urine, real urine and diluted real urine, ii) initial

Iohexol concentration 10 – 1000 ppm, iii) initial H2O2 concentration 0 – 1000 ppm, iv) initial Fe2+ addition:

0 – 50 ppm, v) Dilution factor: undiluted, x10 diluted, x100 diluted and vi) starting pH: 3 – 11.

Table 6.2 – Physicochemical characteristics of the real wastewater matrices (Giannakis et al. 2015c,

Margot et al. 2013, Margot et al. 2011).

(mg/L) WWTP Influent Activated Sludge Effluent Total Suspended Solids (TSS) 30 13.9

Dissolved Organic Carbon (DOC) 91 7.9

Chemical Oxygen Demand (COD) 256 30.2

Total Nitrogen (in NH4, NO3, NO2) 30 19.1

Total Phosphorus 5.8 0.55

Alkalinity (CaCO3) n.m. 273

pH 8 7.5 n.m: not measured,

The composition of the synthetic WW and UR matrices was presented above. The real WW experiments

involved i) the sampling from the influent and ii) the sampling after biological treatment and secondary

clarification, from the WWTP of Vidy, Lausanne, Switzerland, whereas the real UR experiments succeeded

the collection of urine from healthy individuals. Their main characteristics are presented in Tables 6.2 and

6.3, respectively. Finally, spiking with Iohexol to the desired level took place before each experiment.

Table 6.3 – Physicochemical characteristics of real urine matrices (own measurements and (Beach

1971)).

Parameter low high unit TDS 24.8 37.1 g/L pH 6.2 8.3

157

COD 6.1 10.6 g/L TKN 4.8 7.9 g/L TOC 3.6 6.7 g/L

Average

Inorganic salts 14.2 g/L Urea 13.4 g/L Organic compounds 5.37 g/L Organic ammonium salts 4.1 g/L Total solutes 37.1 g/L

6.2.5. Statistics, modeling and data treatment

For the statistical and modeling part of the investigation, the collected data were organized under

separate designs of experiment. Their treatment was achieved through MINITAB software for Windows

including the ANOVA and the proposed models for the first order degradation constant k. The evaluation

of the models is performed through the standard error (S) and the coefficient of determination (R2).

i) Quadratic model

The model is formulated as follows:

(6.1)

where k is the (first order) reaction constant, xi are the model parameters and aij the respective weights.

For each experiment the first-order k constant was determined (dependent variable) and was described

as a function of initial Iohexol concentration [I], H2O2 addition [H2O2], starting iron addition [Fe2+] and the

pH (independent variables).

ii) Multiplicative model

The second model presented here is formulated as follows:

(6.2)

k refers to is the reaction kinetics, xi are the model parameters, γi the corresponding weight, while c is a

numerical constant.

Given that all the parameters can be expressed in logarithmic terms (as concentrations), the model can

be written as above. The independent variables assume 0.01 instead of 0 in the Iohexol, H2O2 and Fe2+

concentrations, and the pH of the solution is expressed as –log[H+]. Therefore, the equation 2 is expressed

as follows:

158

(6.3)

With the aforementioned transformation, the model takes a rather simple form of a product between the

independent variables, directly connected with the reaction constant k.


6.3.1. Engineering approach – investigation on the operational parameters

6.3.1.1. Iohexol, H2O2 and water matrix effect on t90%

In the first part of this study, the operational parameters involved in the UV/H2O2 AOP process were

investigated, namely the concentration of Iohexol, the addition of H2O2 and the matrix containing the

contaminant. Figure 6.1 depicts the influence of concentration of the drug (I) and the oxidant (H2O2), in

Mili-Q (MQ) water. As it has been previously reported (Pereira et al. 2007), Iohexol is rather susceptible

to UV treatment alone. The high molar absorption coefficient of Iohexol at 240 nm (Borowska et al. 2014)

makes the treatment by the high energetic UV-C photons effective. According to the level of drug addition,

the time necessary to degrade the pollutant starts from a matter of minutes (10 ppm) and exceeds 10 h

(1000 ppm) for this relatively low light intensity used in our experiments. The good linear profile of the

fitted lines in log Y scale (R2>99%) indicate that the degradation follows pseudo first order kinetics. The

threshold of 90% degradation is set to ensure elimination of the compound and offer a ground of

comparison among the various processes that will follow, which is calculated as t90%=-(ln0.1)/k.

In the former experiment, H2O2 was also added in a log-stepwise manner. The homolytic disruption of the

HO-OH bond results in the release of the second most powerful oxidant, the hydroxyl radical (HO●) (Guo

et al., 2013). In the past, adding small quantities of H2O2, when LP UV was used, has not sufficiently

improved the degradation of Iohexol in water matrices (De la Cruz et al. 2013, De la Cruz et al. 2012,

Pereira et al. 2007), but the concentrations never exceeded the “economic” range.

159

Figure 6.1 – UV photolysis and UV/H2O2 experiments in Mili-Q water. Note that the results in the 10-

1000 mg/L range are plotted in double-logarithmic scale and axis breaks for clarity purposes only.

Here, we reached up to 1000 ppm H2O2 initial addition, and the results are presented in the different color

traces of Figure 6.1. Iohexol is mostly affected by the hydroxyl radical addition to the aromatic ring, as

hydrogen abstraction or electron transfer are either slower or less common pathways (Zhao et al. 2014).

Nevertheless, it appears that the process is mildly affected in its totality, since the t90% is moderately

improved. Also, since the first order models also fit well this process, we can corroborate with previous

findings (Pereira et al. 2007, Sharpless and Linden 2003) that most probably, the efficiency of direct

photolysis is very high.

Table 6.4 – t90% evolution (min) in varied Iohexol (10-1000 ppm) and H2O2 (0-1000) levels

Iohexol/H2O2

1000/0 1000/10 1000/100 1000/1000

Mat

rix MQ 324 320 311 268

WW 371 360 329 291

UR 535 523 501 461

100/0 100/10 100/100 100/1000

Mat

rix MQ 33 28 26 19

WW 35 29 27 22

160

UR 204 184 141 122

10/0 10/10 10/100 10/1000 M

atrix

MQ 7 5 4 3

WW 10 7 6 3

UR 168 154 139 122

Similar investigation also took place in (synthetic) wastewater (WW) and (synthetic) urine (UR) matrices.

These matrices represent the main conditions in which Iohexol is encountered, and the effect of the

matrix is here investigated. Table 6.4 summarizes the t90% calculated in the different experiments, varying

the initial Iohexol and H2O2 amount, as a first step. A marginal improvement is observed as the ratio of

Iohexol/H2O2 is modified, as found in the MQ matrix before. However, although t90% times for WW

remained close to the observed ones for MQ, synthetic urine values did not drop significantly for low (10

ppm) Iohexol content, but were rather similar to the 100 ppm ones. The explanation lies in the

competition for the HO● radicals generated by the process. In wastewater, the oxidizable organic and

inorganic components are significantly less than in urine, which makes the degradation of mg/L quantities

easier. Normally, for compounds that are highly photo-oxidized, the effect of the organic matter is usually

not very profound (Canonica et al. 2008). Hence, a parameter which could greatly affect the application

of the process and would need further investigation is the possible dilution of the concentrated WW and

UR, and then treatment with AOPs.

6.3.1.2. Effect of matrix dilution and Fenton-initiated enhancement of the process on t90%

Figure 6.2 showcases the wastewater and urine experiments plus the comparison between undiluted and

diluted (x10 times) matrix, while featuring the addition of Fe2+ in the system and the drop of the pH up to

3. For WW (Figure 6.2a), dilution had an effect on the degradation efficiency, decreasing as the initial

Iohexol content is decreasing. The diluted matrix now presents less competition for the HO , effectively

targeting the contaminant. However, for the same reason, milder effect on the addition of iron now

occurs, as, along with the antagonistic nature, the synergistic metal complexing effects of the organic

matter, leading to a higher Fe3+ solubility and consequently a more efficient Fenton and photo-Fenton

cycle, are now mitigated. Nevertheless, the dilution plays an important role and ~20% improvement in

degradation is measured. This suggests confidently that reducing the organic matter of the water, for

instance with an activated sludge process could enhance the efficiency of the downstream UV/H2O2/Fe2+

AOP applied, the participation of the matrix, but also the recalcitrance of the compound while treated,

since the improvement is not impressive.

161

Figure 6.2 – Effect of pH, dilution and Iohexol, H2O2 and Fe2+ amounts. Dotted lines represent the

undiluted matrices, continuous lines indicate the x10 times dilution experiments, and for Figure 3b, the

x100 times diluted UR experiments are signified with long dashed lines. Note the mixed axes scales.

Furthermore, in Figure 6.2b we observe a similar improvement in UR treatment when x10 dilution is

applied in the system. In average, the improvement is higher than the respective one in WW, with almost

~40% minimum increase in the efficiency. The new matrix composition absorbs less UV inefficiently (nitro-

, amino- and phosphoric compounds from the synthetic recipe), thus efficiently targeting Iohexol, and also

the generated radicals target the contaminant without being overly wasted. Technically, using a

secondary water from other sources in a hospital could be feasible; for example water from deionization

units, or even greywater (having considerably lower organic and inorganic content than urine) could be

used for the dilution. Of course, hydraulic optimization of the reactor and construction costs will play an

important role, which is beyond the scope of this research.

6.3.1.3. Experiments in real wastewater and urine matrices

i) Real wastewater

In municipal wastewater, we tested the lower two concentrations assayed in the synthetic matrix, before

and after secondary treatment of the WW inflow, varying the Iohexol spiking, the H2O2/Fe2+ addition levels

and the pH in the beginning of the experiment.

162

Figure 6.3 – Real wastewater and urine experiments: UV/H2O2/Fe2+ process. A) Iohexol in untreated or

biologically treated WW, and B) diluted/undiluted urine, H2O2 added in 0, 10 or 50 ppm, iron was

added in 0, 1, or 5 ppm, and changing of the initial pH value (3, 5 or near-neutral). The two main

groups of Figure 3a data are separated by continuous (10 ppm Iohexol) or dashed lines (100 ppm). The

respective groups in Figure 3b are designated by color. The vertical bars show the variation in

efficiency when pH was changed. Note the mixed axes scales.

Figure 6.3a suggests that, as a rule of thumb, acidification increased efficiency and moving towards the

natural pH, the removal is hampered. The initial physicochemical conditions differ significantly among the

two matrices with suspended solids, which block light transmission, being double before treatment.

Secondly, the organic content removed in the activated sludge unit greatly benefited the process

demonstrating 50% reduction in the removal of Iohexol (10 ppm) and almost 20% in the 100 ppm

experiments. The reduction of the organics content permitted the redirection of the HO● radicals against

the contaminant, which is less effective when high amounts of Iohexol were added.

However, the degradation in WW is a complex system, with various forces which aid or act antagonistically

either to the photolysis or the production of hydroxyl radicals. Other parameters that influence the

efficiency are the nitrogen, phosphorus and carbonate-related compounds. Nitrate exposed to UVC at

254 nm has been shown to undergo photo-transformations. With the participation of peroxynitrite and

peroxinitrous acid, either nitric acid or nitrite is produced (Mack and Bolton 1999). The nitrite, reacts with

the hydroxyl radicals, producing nitrite radicals (Vione et al. 2014). Phosphorus on the other hand

moderately consumes hydroxyl radicals, but precipitates by the iron salts, which actively reduces the

available iron for the complementary photo-Fenton action. Finally, (bi)carbonates are known not only to

scavenge the hydroxyl radicals, but also to form the carbonate radicals, which have mild oxidative action

(Wu and Linden 2010).

163

ii) Real urine

For the experiments involving real urine, the initial COD values ranged between 2.8-9.5 g/L and DOC 2.5-

7 g/L, corroborating with the literature suggesting similar values, previously presented in Table 6.3.

Treatment Iohexol in urine (Figure 6.3b) revealed only two different families of graphs, i.e. the diluted and

undiluted urine experiments. Firstly, pH modification and the addition of the Fenton reagents did not

enhance significantly the degradation rate. The vertical bars indicating the improvement are smaller than

the respective wastewater ones. Secondly, when treating Iohexol in undiluted urine, almost regardless of

the method or reactant addition, both 100 and 1000 ppm experiments required a significantly elevated

time to complete. Apart from the suspended solids, the reactants involved in wastewater, concerning

nitro-, phosphoro- and carbonate-compounds here are in higher amounts, compared to wastewater, and

the implications are expected to make the degradation scheme more complex.

Also, urine contains large amounts of proteins, such as urobilin, serum albumins, transferrins etc, some

of which contain groups which absorb in the UV region. On the other hand, transferrins can bind the iron

(Davis et al. 1962), which limit the participation in the Fenton reaction (Papoutsakis et al. 2015a). This

hypothesis is further strengthened when examining the diluted urine matrix. If x10 dilution is applied, the

need to acidify and add the Fenton reactants is lower, as they marginally contribute to the degradation.

Although a bigger reactor theoretically would be necessary, the economical and operational costs are

considerably lower than the ones necessary for improving the process in undiluted urine.

6.3.2. Statistical approach – modeling and mathematical optimization of the treatment

A general finding of the previous part was that there were no significant differences between the results

found in synthetic and real WW and UR matrices, therefore the laboratory tests can be extrapolated in

the real context. Also, it means that in two different matrices, we could predict the time necessary for

degradation (t90%), based only on broad indications about the target matrix and the order of magnitude

of Iohexol concentration. For the aforementioned reasons, we attempt to model the degradation process

(first order degradation constant k), based on the laboratory experiments already presented in the

previous chapter.

The models, to which the data will be fitted, have been presented in the Materials and Methods section.

For identifying the coefficients of these models, we have completed the experiments presented in the

previous section with additional experimental points, to constitute a Central Composite Design (CCD),

which summary is as follows:

Iohexol was kept constant, at 100 mg/L,

H2O2 and Fe2+ values were kept in a low amounts, in all matrices

164

pH was tested among 3 and 5 (to better take advantage of the conditions that favor the Fenton

reaction).

Table S1 of the Supporting Information presents the data points and the levels of the parameters used in

the CCD. The experiments added from the previous section bring different Iohexol and Fenton reagent

amounts, as well as alternative pH values (near neutral and basic, only for WW).

As a general strategy for both matrices, a step-wise construction of the model took place, as follows: After

removal of the outliers and through regression, different models were fitted. Their expressions are

presented in Table 6.5 for WW and Table 6.6 for UR. For the multiplicative model, the logarithmic values

of the various levels and the response variable k were used; we remind that for pH, its definition of the “–

log[H+]” was used.

Two evaluation criteria have been used, i.e. the standard error (S) and the coefficient of determination

(R2, %). ANOVA was also performed (the detailed ANOVA tables can be found in the Supplementary

material) and put in evidence the order of importance among the factors.

Iohexol > H2O2 or pH > Fe2+.

This order is only qualitative since often the Fenton reagents and the pH failed to pass the P-test of 95%

confidence interval (e.g. Tables S1.1.1 or S1.1.5. in Supplementary Material). Also, the order among H2O2

and pH depends on the matrix, with WW regarding the H2O2 addition as most important and UR matrices

(especially the undiluted) depending more on the pH. This effect can be attributed to the efficient

photolysis or the homolysis of H2O2 in the more transparent matrices, compared to UR, whereas

acidification favors the photo-Fenton participation.

As it is clearly shown in the two previous Tables (6.5 and 6.6), the process cannot be described fairly by a

linear model. The interactions and/or the square terms need to be added. In all cases, Iohexol amount is

the most important factor in determining the order of magnitude of the k constant. The use of the

multiplicative model provides a far simpler expression, without conceding much in terms of accuracy in

most of the cases. As the scale is different, the importance of the parameters is also changed, with pH

becoming more important for wastewater and Fe2+ for urine.

165

Table 6.5 – Wastewater models with S and R2 values.

WASTEWATER Linear Linear w/squares Quadratic WASTEWATER Multiplicative

Linear

Linear

Constant 1.46E-01 3.56E-01 3.36E-01

Constant 9.33E-01

[I] -1.43E-04 -2.86E-03 2.85E-03

log[I] -9.75E-01

[H2O2] 2.30E-05 2.93E-04 3.33E-04

log[H2O2] 2.79E-02

[Fe2+] -8.00E-05 2.89E-03 5.08E-03

log[Fe2+] 1.07E-01

pH -1.69E-03 -3.14E-03 -3.00E-04

log[H+] 2.32E-02

Squares

[I] x [I]

3.00E-06 3.00E-06

[H2O2] x [H2O2]

-1.00E-06 1.00E-06

[Fe2+] x [Fe2+]

-6.10E-05 -1.20E-04

pH x pH

-7.30E-05 -1.97E-04

Interactions

[I] x [H2O2]

-1.00E-06

[I] x [Fe2+]

-2.00E-06

[I] x pH

-5.00E-06

[H2O2] x [Fe2+]

3.00E-06

[H2O2] x pH

1.20E-05

[Fe2+] x pH

-1.01E-04

S 0.07 0.02 0.02

S 0.10

R2 38.44 92.38 93.90

R2 96.62

DILUTED WASTEWATER Linear Linear

w/squares Quadratic DILUTED WASTEWATER Multiplicative

Linear

Linear

Constant 2.48E-01 6.68E-01 6.29E-01

Constant 1.37E+00

[I] -2.39E-04 -3.33E-03 -3.79E-03

log[I] -9.18E-01

[H2O2] 9.80E-05 8.70E-05 2.69E-04

log[H2O2] 5.60E-03

[Fe2+] -1.74E-03 1.50E-03 1.11E-02

log[Fe2+] 6.56E-02

pH 3.00E-04 -8.45E-02 -7.70E-02

log[H+] 8.60E-02

Squares

[I] x [I]

3.00E-06 3.00E-06

166

[H2O2] x [H2O2]

-1.00E-06 2.00E-06

[Fe2+] x [Fe2+]

4.10E-05 -1.08E-04

pH x pH

7.69E-03 7.00E-03

Interactions

[I] x [H2O2]

-1.00E-06

[I] x [Fe2+]

-2.00E-06

[I] x pH

1.70E-05

[H2O2] x [Fe2+]

-3.00E-05

[H2O2] x pH

2.00E-05

[Fe2+] x pH

-6.30E-04

S 0.08 0.04 0.04

S 0.17

R2 61.38 92.50 95.76

R2 92.24

167

Table 6.6 – Urine models with S and R2 values

URINE Linear Linear w/squares Quadratic URINE Multiplicative

Linear

Linear

Constant 9.28E-02 1.86E-01 1.02E-01

Constant -3.39E-01

[I] -2.50E-05 8.90E-05 -2.30E-04

log[I] -3.37E-01

[H2O2] -9.00E-05 -1.22E-04 -2.47E-04

log[H2O2] 7.90E-03

[Fe2+] 2.38E-04 2.47E-03 9.33E-03

log[Fe2+] 2.20E-01

pH -1.02E-02 -6.29E-02 -2.06E-02

log[H+] 1.33E-01

Squares

[I] x [I]

-1.00E-06 1.00E-06

[H2O2] x [H2O2]

1.00E-06 -1.00E-06

[Fe2+] x [Fe2+]

-5.00E-05 -1.41E-04

pH x pH

5.56E-03 1.12E-03

Interactions

[I] x [H2O2]

-1.00E-06

[I] x [Fe2+]

2.00E-06

[I] x pH

2.90E-05

[H2O2] x [Fe2+]

5.00E-06

[H2O2] x pH

4.20E-05

[Fe2+] x pH

-1.25E-03

S 0.02 0.01 0.01

S 0.17

R2 59.52 79.59 95.81

R2 83.42

DILUTED URINE Linear Linear w/squares Quadratic

DILUTED URINE

Multiplicative

Linear

Linear

Constant 1.19E-01 2.41E-01 2.30E-01

Constant 2.08E-01

[I] -1.60E-04 -1.31E-03 -1.05E-03

log[I] -6.35E-01

[H2O2] 6.14E-04 9.10E-05 -1.63E-04

log[H2O2] 5.05E-02

[Fe2+] -1.33E-03 -8.40E-04 -5.40E-04

log[Fe2+] 1.43E-01

pH 2.33E-03 -1.10E-03 -5.50E-03

log[H+] 3.39E-02

Squares

[I] x [I]

1.00E-06 1.00E-06

168

[H2O2] x [H2O2]

3.00E-06 2.00E-05

[Fe2+] x [Fe2+]

4.70E-05 6.20E-05

pH x pH

-7.70E-04 4.00E-04

Interactions

[I] x [H2O2]

-1.00E-06

[I] x [Fe2+]

-

[I] x pH

-

[H2O2] x [Fe2+]

-4.00E-06

[H2O2] x pH

1.10E-04

[Fe2+] x pH

-1.54E-04

S 0.03 0.01 0.01

S 0.08

R2 76.14 95.08 95.39

R2 95.39

The prediction of degradation time holds high importance for the technical applications, therefore

optimization of the models has been assayed. The problem is broken down to maximizing an objective

function k = f (Iohexol, H2O2, Fe2+, pH). The optimization took place through the desirability function

(Derringer and Suich 1980, Papoutsakis et al. 2015b). This method allows the simultaneous optimization

of several equations. The desirability function for each equation is given by the following expression (6.3):

(6.3)

Since normalization of the parameters has taken place, the R different desirability functions d can be

combined to the (overall) desirability function, for the k constant, as follows (6.4):

(6.4)

R is the number of functions,

d the desirability of each function and

D the desirability of the system.

169

The optimization results are presented in Table 6.7, showing the desirability function values and the

operating regions. As it appears, the optimal region for maximizing the k constant is when Iohexol is

minimal, as the addition of higher amounts of Iohexol changes dramatically the k values. As expected, for

the linear models we found the optimum at the border of the experimental domain.

Table 6.7 – Optimal regions for treatment Iohexol through optimization by the desirability function.

Wastewater Iohexol H2O2 Fe2+ pH D Urine Iohexol H2O2 Fe2+ pH D

Linear 10 1000 50 lowest 0.3947 Linear 10 1000 50 lowest 0.7815

Quadratic 10 1000 32.3 7 1 Quadratic 10 1000 22.7 lowest 1

Diluted WW Iohexol H2O2 Fe2+ pH D Diluted UR Iohexol H2O2 Fe2+ pH D

Linear 10 1000 50 lowest 0.7401 Linear 10 1000 50 7 0.7829

Quadratic 10 1000 1 7 1 Quadratic 10 1000 1 7 0.9986

For the quadratic model in the undiluted matrices, the gains from the increase of the Fe2+ amounts starts

to get mitigated and values around 20-30 ppm are suggested. This is probably caused by the physical

blocking of UV light by the iron particles. Between WW and UR, the difference is found in the Fe2+ amount

added, as more iron is suggested in the case of WW. Since in UR acidification of the matrix was

recommended, less iron was suggested for the optimal performance in this case. Also, Fe2+ is an efficient

HO● scavenger, hence higher amounts will not necessary mean more advantageous performance. As far

as the desirability of the proposed operating regions is concerned, the linear models did not produce

solutions very close to D=1. On the other hand, the quadratic models often found the optimal regions and

the desirability of the system is ~1.

In conclusion, our statistical approach resulted in models with satisfactory performance (based on S and

R2 values). However, the optimization with a single response variable is relatively blindsided by other

factors, such as the cost of the process. In that case, the optimal region would be a compromise among

the efficiency and the cost. It is recommended that the quadratic model has to be preferred but in further

work, more response variables should be taken into consideration, such as involving the cost of reagents,

iron reclamation, residual H2O2 elimination, pH neutralization and others as in other works (e.g.

(Papoutsakis et al. 2015b).

170

6.3.3. Analytical approach – Global measurements (COD, TOC, and UV-vis absorbance)

combined with specific HPLC and MS analysis

The combination of the findings of the previous two parts do not fully support the idea of using H2O2 to

enhance the degradation efficiency of Iohexol in aqueous matrices. Therefore in this part, an effort has

been made to decode the reasons behind this effect and propose the proper operational conditions under

a new prism.

In Figure 6.4 (a-d) we present the profiles of the transformation products (TPs) by Peaks (area) vs. time

graphs. The operational conditions were set to Iohexol at 1000 ppm and H2O2 logarithmically increasing

from 0 (UV photolysis) to 1000 ppm (highly oxidative conditions). In all figures, the treatment was stopped

after 11 hours and the colored lines of the graphs represent the TPs by their elution time, corresponding

to Iohexol and the generated by-products under the different studied conditions. We have excluded the

peaks below 3.5 min elution time as they represent highly-polar aliphatic acids, with low absorbance and

non-linear correspondence to the UV detector set at 254 nm. Also, a representative chromatogram can

be found at the supplementary material (Figure S2).

As it can be observed, the degradation of the parent compound is proceeding in all cases as time

progresses (peak at 10.12 min) towards complete elimination, but achieved only when 100 and 1000 ppm

H2O2 were used. During the degradation of Iohexol, intermediate peaks appear in distinctive times,

representing the more polar by-products resulting from the elimination of Iodine atoms, the -OH addition

and the side chains breakage from the central aromatic ring. Also, it can be observed that increasing the

H2O2 concentration also leads to faster peak maxima in mid-range intermediates. This effect is profound

at 1000 ppm addition, where the fastest oxidation of the lower range intermediates is observed. In terms

of peak areas, Figure 6.4d shows the lowest areas, thus demonstrating the effect of H2O2. The HPLC

method used for the analysis of Iohexol was set at 254 nm, which is in-between the maximum absorbance

of the central aromatic ring (Weast 1985), and therefore, the detected intermediates have their central

ring intact. Since H2O2 addition has caused lower detection in overall, it means that it actively contributes

to the mineralization of the parent compound and the generated by-products, due to the non-selectivity

of the HO● radicals.

171

Figure 6.4 – HPLC peak areas evolution during Iohexol degradation by the UV photolytic and

photocatalytic process. A) UV only, B) 10 ppm H2O2, C) 100 ppm H2O2, D) 1000 ppm H2O2. 100 ppm of

Iohexol was chosen as initial spiking.

In order to assess the extent of mineralization, further investigation was initiated, focusing in the

degradation of the compound and the reduction of organic load in the bulk. The three tests assayed and

presented in Figure 6.5 are UV, UV/H2O2 and UV/H2O2/Fe2+ process. Enhancing the UV degradation process

with H2O2 and then with Fe2+ and H2O2 inflicted a t90% reduction of 40 and 50% respectively for the two

aforementioned additions. Nevertheless, the addition of iron enhanced only the early stages of the

treatment, as the time necessary for complete elimination was similar with the UV/H2O2 process alone.

Furthermore, the COD evolution shows a very similar behavior for UV photolysis and UV/H2O2 oxidation.

Up to 30 min of treatment, the two processes are quasi-identical which means that there is transformation

to more readily oxidized forms of carbon rather than actual degradation of the total carbon content. On

172

the other hand, the presence of iron in the solution that enhances the hydroxyl radical generation

demonstrates an immediate and constant rhythm of reduction. Most probably, the degradation pathway

of the combined UV/H2O2/Fe2+ process has different steps than the oxidation process, such as enhanced

side chains breakage instead of simple substitutions.

Figure 6.5 – Iohexol elimination by the UV-based AOPs. Iohexol degradation was followed by HPLC

(blue trace), COD (red trace) and TOC decrease (green trace) during the following treatment methods:

UV photolysis (trace: ), UV/H2O2 process (50 ppm H2O2, trace: ), and UV/H2O2/Fe2+ process (5 ppm

Fe2+, 50 ppm H2O2, trace: ). H2O2 reduction: brown traces. A system employing 35-W UV-C lamps

(instead of the 11-W ones of the previous parts, but otherwise identical) was used here. 100 ppm

Iohexol was chosen as initial spiking.

Finally, as far as the TOC removal is concerned, the compound confirms its highly recalcitrant behavior,

with the three processes demonstrating similar and limited reduction for the first 15-20 mins. After the

total (parent) Iohexol removal, the TOC is further reduced, up to 30 and 55% for UV/H2O2 and

UV/H2O2/Fe2+, respectively. This TOC removal justifies experimentally for the first time so far in this

investigation that extended treatment in presence of H2O2 and Fe2+ will eliminate the majority of the

organic carbon present in the solution.

If we consider the Average Oxidation State as a normalized measure to assess the overall oxidation state

of the solution (Bandara et al. 1997), we get:

173

(2.6)

where COD and TOC values are expressed in mol O2/L and mol C/L, respectively, and ranges from -4 (fully

oxidizable, e.g. CH4) to +4 for CO2 (completely oxidized)

For each case, for time 0 to time 120 min:

A) UV AOS: from 0.1 to 0.25

B) UV/H2O2 AOS: from 0.1 to 1.03

C) UV/ H2O2/Fe2+ AOS: from 0.1 to 3.24.

As it appears, for the same conditions, only with a moderate iron addition, the overall system did not

leave residual H2O2, which strengthens the economic design of the system and secondly, the overall state

of the system demonstrates an almost complete carbon elimination. Also, as H2O2 in high amounts is toxic

to microorganisms, complete elimination during treatment will reduce the post-treatment removal costs.

Therefore, apart from the Iohexol removal, which is moderately enhanced, the intermediates can be

effectively removed, which is often a common question when treating (recalcitrant) pharmaceuticals

(Malato et al. 2009, Sarria et al. 2003).

What was made evident, is that the degradation process differs among the three processes. For this

reason, the degradation products from the different treatment methods (after 5 min) were identified by

HPLC-HR-MS analysis according to the corresponding spectral characteristics: mass spectra, accurate mass

and characteristic fragmentation. Supplementary Tables S2.1 and S2.2 show the molecular formulas,

double bond equivalent (DBE), theoretical and experimental masses along with mass accuracy (∆m)

expressed in ppm.

In order to depict the differences, a synthetic representation of the intermediate products is shown in

Figure 6.6. For all the studied treatment methods, the degradation of Iohexol starts by scission of iodine

from the aromatic ring and subsequent addition of -OH at iodo-sites, which results in single or double de-

iodination, forming the phenolic products P1 and P6. Additionally, two more products were produced via

direct attack of HO on the side chain of Iohexol molecule (C1 and P3), with ketone formation and side-

chain breakage, respectively. Identified products are in the agreement with some products identified by

other authors (Jeong et al. 2010).

After P1, the degradation continues with HO attack on side chains via oxidation, -OH addition and

decarboxylation reactions, as seen in relevant works (Jeong et al. 2010, Tian et al. 2014, Zhao et al. 2014),

with minor differences between the treatments (different color arrows). Indeed, the influence of the UV

irradiation is the main actor in the degradation process. However, degradation at P1, which also includes

174

HO attacks by treatment B and C (UV/H2O2 and UV/H2O2/Fe2+) continues with the loss of a second iodine

atom, resulting to product P6. At this point, the degradation with UV (treatment A) is finished. The biggest

differences in the products appeared after P6 degradation, with four products identified for treatment B

and three for treatment C. The degradation of the P6 product is a pathway that is based on UV exposure,

as all processes, but proceeds further only in presence of H2O2 and/or Fe2+. The structure of these by-

products suggests that the degradation continues via the removal of the third iodine atom, plus further

oxidation and decarboxylation of side chains. However, the appearance of unique and different products

for treatments B and C imply that formation rate of HO plays crucial role in all stages of the degradation

process. Also, it is also noteworthy that in several pathways the formation of a non-ring hydroxylated

derivative of Iohexol is derived from ring-hydroxylated TP of Iohexol, which is rare, but has occurred again

in relevant literature (Csay et al. 2012, Jović et al. 2013).

Figure 6.6 – Overall mechanistic degradation pathway of Iohexol treated by UV-based AOPs. Products

common for all three treatments were marked with P, UV marked with A, for UV/H2O2 with B and for

UV/H2O2/Fe2+ with C. Products common for A and C treatment were marked as AC, and accordingly,

products common for B and C treatment were marked as BC.

Additionally, the use of Fe2+ and its affinity to the side chain structures, results to their higher substitution

or breaking. From the products’ structure, it could be concluded that the addition of iron is increasing the

efficiency of the treatment, giving products with shorter side chains (products C4 and C5), and even

nitrogen removal (product C4). The presence of hydroxyl radicals is responsible for their further

degradation, as well as for the oxidation of the side chains removed. The MS analysis at 5 min corroborates

175

with the HPLC results, which detection took place at 254 nm, indicating that the aromatic ring remains

intact. Nevertheless, at 15 min, there is identified presence of products without iodine atoms, with the

aromatic ring, but with degraded side chains. Finally, the absorbance spectra (see supplementary Figure

S3) shifts significantly after 15 min of treatment, there are no exclusive UV pathways and mineralization

initiates after this point, therefore we can conclude that the processes B and C involving HO attacks and

especially C, are the most efficient. The addition of iron is strongly recommended for efficient parent

compound and by-product degradation.

6.4. Conclusions

Iodinated contrast media, such as the investigated Iohexol, can burden the environment with their

presence for a relatively long time, due to their refractory nature. Since AOPs gained more attention over

the last decades, the abilities of the synthetic UV/H2O2/Fe2+ process were assayed. To achieve efficient

degradation and deep insight on the inactivation pathway, in the present work we assessed three

approaches in degrading this drug, namely the operational parameter testing, the statistical optimization

and the analytical chemical investigation.

As it appears, the dominant driving force in Iohexol degradation is the UV-C irradiation. The undertaken

assays however, showed that the t90%, as a measure of comparison among the various experiments, can

be moderately reduced if H2O2 and/or Fe2+ are added in the bulk. Also, depending on the matrix used,

dilution was proven very effective in reducing organic matter and solids concentration, thus enhancing

the removal of Iohexol.

The process was very well described by a quadratic model, which provided the best prediction of the

kinetics constant for both wastewater and urine, and diluted or undiluted matrix. Also, a multiplicative

model was produced, which sustained adequate accuracy while offering a simple formula. Additionally,

the target of indicating the optimal operation regions for Iohexol degradation was achieved.

Finally, evidence for the main degradation actor and the evolution of the process, as well as the

intermediates formed during the degradation of the parent compound were obtained. H2O2 and Fe, while

macroscopically had a modest effect, their contribution in the mineralization is noteworthy. This is of high

importance, as the ICM are notorious for their recalcitrance and their subsequent presence in the

environment. Hence, although initially H2O2 and Fe2+ presence seem as an economic side-effect, our

investigation suggests the optimization of their quantities and their addition to the UV process, which can

effectively reduce the organic pollution in the subsequent matrices.

177

7. Chapter 7 - Solar photo-Fenton and UV/H2O2 processes

against the antidepressant Venlafaxine in urban

wastewaters and human urine. Intermediates formation

and biodegradability assessment.

Published work:

Stefanos Giannakis, Idriss Hendaoui, Milica Jovic, Dominique Grandjean, Luiz Felippe De Alencastro,

Hubert Girault, and Cesar Pulgarin (2017). Solar photo-Fenton and UV/H2O2 processes against the

antidepressant Venlafaxine in urban wastewaters and human urine. Intermediates formation and

biodegradability assessment. Chemical Engineering Journal 308, 492-504.

Web link:



Appendix F



178

7.1. Introduction

During the last years, there is a lot of word on the new challenges and problems that have emerged in

Environmental Engineering and more specifically, wastewater (WW) treatment. Water scarcity and reuse

(Jimenez and Asano 2008), emerging pollutants (Richardson and Ternes 2005) and the rise of antibiotic

resistance (Rossolini et al. 2014) are only a few that are closely related to the absence of potent measures

to limit or mitigate their effect. With that being said, wastewater treatment plants (WWTPs) hold their

fair share of blame for inefficiently stopping some of the new threats.

One of the most crucial problems the WWTPs face, and is connected with the improvement of life

standards globally, is the perception of illness and the ways drugs are administered nowadays. Apart from

the physical illnesses, injuries etc. which require antibiotics, sanitizers and similar products, the taboos of

psychological conditions have crumbled and as their recognition as illnesses is established, treatment is

now considered as nothing out of the ordinary. The intense ways of living have been proven to increase

the stress levels of people, leading to anxiety and chronic symptoms, as serious as sleep deprivation or

depression (Melchior et al. 2007). The administration of proper medication has led to the second problem

WWTPs have to face: emerging contaminants.

Venlafaxine (VFA, brand name: Effexor XR, Lanvexin or Trevilor), is an anti-depressant, which belongs in

the general family of selective serotonin and norepinephrine reuptake inhibitors (SSNRIs) (Andrews et al.

1996, Harvey et al. 2000). Practically, it treats depression, anxiety and panic disorders by increasing the

concentrations these natural substances in the body and brain of the patient (USNLM 2016a, Weller et al.

2000). Chemically, it belongs to the class of benzene and substituted derivatives and is a tertiary amino

compound that is N,N-dimethylethanamine substituted at position 1 by a 1-hydroxycyclohexyl and 4-

methoxyphenyl group (Table 1). Its excretion from the body follows the renal route, ends up in WWTPs

and therefore is found in natural waters, as VFA and its metabolites are escaping in almost 50% rate the

wastewater treatment process (Lajeunesse et al. 2012, Metcalfe et al. 2010, Rúa-Gómez and Püttmann

2012, Stadler et al. 2015). Measured concentrations range from 18 to 122 ng L−1 for VFA (Rúa-Gómez and

Püttmann 2013), or 102 and 690 ng L−1 (Rúa-Gómez and Püttmann 2012, Schultz et al. 2010) which could

affect the natural biota.

Since this compound lacks functional groups that hydrolyze under environmental conditions (pH 5 to 9),

hydrolysis is not expected to be an important environmental process, and therefore the risk of

bioaccumulation is possible. With an estimated bio-concentration factor (BCF) of 60, the potential for bio-

concentration in aquatic organisms is classified as moderate (USNLM 2016b). Furthermore, the exposure

experiments of Bisesi et al. (Bisesi et al. 2014) have indicated that Venlafaxine exposures of bass increased

time to capture their prey (minnows), and the analysis of brain tissues revealed that VFA caused decrease

179

in brain serotonin concentrations, thus explaining the behavior changes. Fong and Molnar (Fong and

Molnar 2013) investigated the biological effects of antidepressants comprising VFA on the mollusks and

crustaceans, i.e. foot detachment, a potentially sub-lethal effect that could result in transport to

unfavorable habitats for the target organisms; VFA exposure caused this effect even by exposure to

concentrations significantly lower than the ones found in WWTP effluents (Fong and Ford 2014). The

question is due to the food chain and cycle of water, how this is going to demonstrate on human beings.

As there risks similar to VFA in WW are increasing, WWTPs need to adapt to modern era threats more

effectively. Under this spirit, the new Swiss regulations for wastewater treatment include a list of 12

priority contaminants for elimination from WWTPs, and VFA has become one of the recent additions to

that list (Giannakis et al. 2015c). The said regulation involves the upgrade of WWTPs to employ activated

carbon, ozone or another advanced oxidation process (AOP) and ensure 80% removal of the chosen

micropollutants. As it appears, AOPs can play an important role acting as a barrier for contaminants of

emerging concern before reaching natural waters (Comninellis et al. 2008).

Recently, works have been initiated on the degradation of VFA by UV/H2O2 and TiO2 photocatalysis

(García-Galán et al. 2016, Lambropoulou et al. 2016) dealing with the elucidation of the degradation

pathway by these methods and also focusing on the toxicological safety of the degradation by-products.

To contribute to this end, in our work we employ 5 Advanced Oxidation Processes (UV, UV/H2O2, solar

light, Fenton, solar photo-Fenton) to degrade VFA in the matrices mostly expected to be encountered.

After a systematic investigation of the opportunities and pitfalls of treatment in water, urban wastewater

effluents and human urine containing VFA are employed and the degradation efficiency is assessed.

Finally, we investigate the degradation pathway of VFA inflicted by the various AOPs and explore the use

of AOPs as a pre-treatment step to increase the biodegradability of this contaminant with the Zahn-

Wellens tests.

7.2. Materials and methods


The chemicals for the experiments were used as received. Venlafaxine HCl (see Table 7.1) was acquired

from TCI (Germany), the HPLC solvents (acetonitrile, acetic acid and ammonium acetate) and the Fenton

reagents (hydrogen peroxide 30% and iron sulfate heptahydrate) were acquired from Sigma-Aldrich

(Switzerland).

Table 7.1 – Venlafaxine characteristics and physicochemical properties (USNLM 2016b).

180

Compound Chemical structure

Molecular

Weight

(g/mol)

Water solubilit

y

(mg/L)

log kow pKa

Henry’s coefficient

(H)

(atm.m3/mol)

HO• reaction rate constant

(M-1.s-1)a,b,c,d,e

Venlafaxine

C17H27NO2

277.402

267 3.20

10.01 2.0 x 10-11

(8.15±0.4)x109

(8.46±0.5)x109

(8.8±1.5)x109

1010

a: (Wols et al. 2013), b: (Santoke et al. 2012), c: (García-Galán et al. 2016), d: (Abdelmelek et al. 2011), e:

(Lee et al. 2014)

7.2.2. Water, wastewater and urine matrices

The preparation of the synthetic matrices involved dissolution of 100 mg/L VFA in either Mili-Q (MQ)

water (18.2 MΩ cm-1), synthetic wastewater (SWW) or synthetic urine (SUR). The composition of the

synthetic matrices is presented in Table 7.2. Real wastewater samples were collected from the local

wastewater treatment plant of Vidy, Lausanne (Switzerland), after an activated sludge process, a moving

bed bio-reactor or a coagulation-flocculation unit. The corresponding (initial) concentrations before

treatment were determined by HPLC/MS and were around 300 ng/L. For the real urine experiments, 10

μg/L VFA was added prior to experimentation.

Table 7.2 – Composition of the synthetic matrices used in this study.

Synthetic Wastewater

Name Chemical formula SWW composition [mg/l]

Peptone - 160

Meat Extract - 110

Urea CH4N2O 30

Dipotassium Phosphate HK2PO4 28

Sodium Chloride NaCl 7

Calcium Chloride dihydrate CaCl2·2H2O 4

Magnesium Sulfate Heptahydrate MgSO4·7H2O 2

Synthetic Urine

Name Chemical formula SUR composition [g/L]

Urea CH4N2O 25

181

Sodium Chloride NaCl 2.925

Sodium Sulfate Na2SO4 2.25

Potassium chloride KCl 1.6

Potassium phosphate monobasic KH2PO4 1.4

Calcium Chloride dihydrate CaCl2·2H2O 1.103

Creatinine C4H7N3O 1.1

Ammonium chloride NH4Cl 1

7.2.3. Light sources and corresponding reactors-experimental apparatus

For the UV and UV/H2O2 experiments, two double-wall, water-jacketed glass batch reactors were used in

parallel. The sample was placed within and the lamps, covered by a quartz sleeve were then submerged

in it. The monochromatic, Hg discharge UV-C lamps were 11-W Philips TUV mini (11W/G11 T5 UV) with

I11W = 26 μW/cm2. For the real WW and UR tests, a 35-W lamp was used instead. For protection of the UV

equipment and standardized conditions, water at 22 °C was recirculated with a Neslab RTE-111

thermostat.

For solar and photo-Fenton experiments, a Hanau Suntest CPS solar simulator was used, employing a

1500-W Xenon lamp, equipped with UVC and IR cut-off filters. The intensity was set at 900 W/m2 and

temperature was kept below 38 °C at all times by air-cooling. The Pyrex-glass reactors were kept in

constant agitation (350 rpm) by magnetic bars. The (dark) Fenton tests were performed in identical

reactors and conditions without providing light to the system.


7.2.4.1. Venlafaxine determination routine by HPLC

The determination of Venlafaxine concentration was performed through HPLC. An HP 1100 Agilent series

HPLC was used. Briefly, the mobile phase consisted of 0.14 M ammonium acetate buffer, (1.079 g/L

acidified with glacial acetic acid (pH = 4). This was then mixed with 10% methanol/acetonitrile solution

and sonicated for 15 min. Finally, filtration from 0.45 μL membrane was done. The HPLC conditions

consisted of 40 °C temperature, 20 μL injection volume, RP-C18 column (4.6mm x 250 mm) and detection

of the peaks at 254.4 nm.

7.2.4.2. Venlafaxine quantification by UPLC/MS in real WW and RU

An online SPE-UPLC®/MS-MS (Acquity Xevo-TQ, Waters) was used. Samples were acidified to pH = 2.0

(32% hydrichloric acid), spiked with a standard mixture of surrogate, containing the MPs in deutered form

and filtered with glass fiber filter (Simplepure PP+GF, 0.22 μm, 25 mm, BGB). Standard solutions have

182

followed the same preparation. Five ml of each sample were loaded on an SPE column (Oasis HLB 25 μm,

2.1 x 20 mm, Waters) with ultrapure water, acidified at 1% of formic acid, as eluent. The online transfer

of VFA to the analytical column (Acquity HSS T3, 1.8 μm, 2,1 x 100 mm, Waters) was made with a gradient

of ultrapure water and acetonitrile acidified at 0.1% of formic acid. Multiple reaction monitoring mode

with two transitions was used to detect MPs and quantification was performed with internal standard

calibration.

7.2.4.3. Intermediates identification by TOF-MS analysis

HR-MS analyses were conducted on a Xevo G2-S QTOF mass spectrometer coupled to the Acquity UPLC

Class Binary Solvent Manager and BTN Sample Manager (Waters, Corporation, Milford, MA). Mass

spectrometer detection was operated in positive ionization using the ZSpray™ dual-orthogonal multimode

ESI/APCI/ESCi® source. Samples were diluted in H2O and directly infused into the mass spectrometer at a

flow rate of 100 μL/min. The TOF mass spectra were acquired in the sensitive mode over the range of m/z

50-1200 at an acquisition rate of 1 sec/spectra. A mass accuracy better than 5 ppm was achieved using a

leucine-encephalin solution as lock-mass (200 pg/ L in ACN/H2O (50:50)) infused continuously using the

LockSpray source (5 sec reference Scan frequency). Source settings were as follows: cone, 25V; capillary,

3 kV, source temperature, 120° C; desolvation temperature, 500° C, cone gas, 100 L/h, desolvation gas,

500 L/h. Data were processed using MassLynx™ 4.1 software.

7.2.4.4. Global chemical analyses (TOC, COD, H2O2 and UV/Vis Absorbance)

The COD of the solution was monitored with HR/LR dichromate vials (HACH Lange, Switzerland) and TOC

was followed by a Shimadzu TOC-VCSN analyzer, with an ASI-V automatic sampling module. H2O2 was

determined spectrophotometrically, after the addition of 10 μL of titanium oxysulfate in 1 mL of sample

and measurement at 410 nm (DIN 38402H15 method). Finally, the absorbance spectra was recorded at

each sampling point (Shimadzu 1800 UV spectrophotometer) and the pH was followed by a Mettler-

Toledo Seven Easy pH meter.


7.3.1. UV-based AOPs degradation of Venlafaxine

7.3.1.1. Monochromatic UV-C photolysis

Figure 7.1 presents the photolysis of Venlafaxine (VFA), under the exposure to monochromatic UV-C (peak

at 253.7 nm) irradiation. The photolysis rates, quantum yields and the proven, but limited efficiency to

degrade VFA by UV light in a collimated beam apparatus have been recently documented (García-Galán

et al. 2016, Wols et al. 2013). The kphot has been in the order of 1.5x10-4 cm2/J in the corresponding work.

183

0 10 20 30 40 50 60 70 80

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

TOC COD VFA

ln(C

/C0) VF

A

Dose (mJ/cm2)

0 10 20 30 40 50 60 70 80

0.0

0.1

2.2

2.3

2.4

COD/TOC ratio Asymptotic fit

CO

D/T

OC

ratio

Dose (mJ/cm2)

Figure 7.1 – Summary of the UV-C photolysis experiments. a) UV-induced degradation of Venlafaxine

followed by HPLC, COD removal and TOC reduction during UV photolysis. b) Evolution of the COD/TOC

ratio.

In our work, a merry-go-round reactor with submerged lamp was employed and the corresponding

kinetics are depicted in Figure 7.1. In principle, the degradation follows a linear trend in (natural)

logarithmic scale which allows the determination of first-order reaction rate constant, being 0.0104

cm2/mJ, calculated as follows:

(7.1)

or as:

(7.2)

By re-arranging the equation 1, we get:

(7.3)

Due to the reactor design and configuration, the actual irradiance received by the system and determined

by iodide/iodate actinometry (Rahn 1997) is 0.005 mW/cm2, which is much higher than the collimated

beam apparatus. Since the degradation follows first-order kinetics, the time necessary to degrade 90% of

the initial concentration (t90%) has been determined and will be used for comparison among the various

processes. From (7.3), by substitution we get:

(7.4)

or:

184

(7.5)

The required time for 90% was 163 min and until the t90% was reached, the UV-C system removed 7% and

4% of COD and TOC. Even by extending the treatment for 4 h, UV-C alone is not sufficient to degrade more

than 20% of COD or remove more than 15% of TOC. As it appears, the complex VFA structure is affected

by UV in the double bonds present but light alone cannot efficiently mineralize the carbon content of the

solution attributed to the degradation by-products and intermediates. Figure 7.1b corroborates with the

findings, and the depicted COD/TOC ratio presents asymptotic, plateau-like tendencies. This reveals the

formation and accumulation of stable by-products, which after an initial fast oxidation, do not further

undergo significant modification or mineralization. Furthermore, Figure S1 of the supplementary material

also confirms this tendency, where the absorbance spectra indicate that although VFA is removed (1st

peak) a plateau is reached after 2 h of exposure, where a formation of intermediates is stable and fails to

proceed further. From the wavelength of the peak (260-280 nm) we postulate that the cyclohexanol cycle

is not affected by UV and the intermediate formation stops there.

7.3.1.2. UV/H2O2 Advanced Oxidation Process

Following the UV photolysis, the same set-up was used to induce UV/H2O2 advanced oxidation of VFA. As

such, in a H2O2 dosing optimization process 5 different H2O2 addition levels were tested, and the results

are summarized in Figure 7.2 (7.2a and 7.2b).

The homolytic disruption of the HO-OH bond results in the release of hydroxyl radicals (HO●) (Guo et al.

2013). The addition of H2O2 and the production of HO● has a direct effect on the degradation of

Venlafaxine. Hydroxyl radicals can act on molecules via oxidation, -OH substitution, as well as water

abstraction and decarboxylation, as seen in related works (García-Galán et al. 2016, Lambropoulou et al.

2016). These attacks drastically modify the properties of VFA and proceed to more efficient degradation

than UV alone.

0 5 10 15 20 25 30

0.0

0.2

0.4

0.6

0.8

1.0

5 ppm 10 ppm 20 ppm 50 ppm

ln(C

/C0) VF

A

Dose (mJ/cm2)0 10 20 30 40 50 60 70 80 90 100 110

0.0

0.2

0.4

0.6

0.8

1.0

2.0

2.2

2.4

COD/TOC

TOC COD H2O2

Ct/C

0

Dose (mJ/cm2)

185

0 5 10 15 20 25 30 35 40 45 50

0

2

4

6

8

10

12 t90%

H2O2 consumed

Initial H2O2

H2O

2 con

sum

ptio

n

0

25

50

75

100

125

150

175

t 90%

Figure 7.2 – UV/H2O2 Advanced Oxidation of Venlafaxine: degradation and process optimization. a)

Degradation of VFA by UV/H2O2 with addition of 5-50 mg/L H2O2. b) Evolution of COD/TOC ratio (for

50 mg/L initial H2O2 addition). c) Consumption of H2O2 (black axis and traces) and changes in the t90%

(blue axis and traces) as a function of initial H2O2 amounts.

The addition of even 5 mg/L H2O2 improved the VFA degradation 20% respectively, compared to the sole

UV experiments. VFA has –OCH3 and –OH groups that react fast when faced to hydroxyl radicals (order of

kHO ≈ 109 M-1.s-1 and 108 M-1.s-1, respectively), but also a –CN group that has a much slower rate constant

(kHO ≈ 107 M-1.s-1) (USNLM 2016b), hence the overall degradation rate will be determined by these groups

and the direct photolysis rate (kphot) when treated by the UV/H2O2 process.

The stepwise increase in H2O2 concentration reveals the changes in degradation kinetics and the

limitations of the employed experimental system (Figure 7.2a & 7.2b). After 50 mg/L the improvement in

reaction kinetics is marginal (data for 100 mg/L not shown). In addition, in (García-Galán et al. 2016) and

(Wols et al. 2013), the photolysis rate was considerably lower than the corresponding rate for oxidation

due to the hydroxyl radicals, and was considered negligible. Here, the oxidation kinetics are estimated as

follows:

(7.6)

Where, i: the H2O2 addition (mg/L).

Table 7.3 – Measured pseudo-first order degradation kinetics of Venlafaxine per AOP and matrix.

Pseudo first order reaction kinetics k (min-1)

186

UV alone MQ

0.0141

UV/H2O2 MQ

0.0392 1.78

0.0757 4.37

0.1412 9.01

0.2948 19.91

0.3726 25.43

Solar light MQ Synergy12.5|58.3 =

0.0002 89.46

Fenton (pH = 3) MQ

0.0026

photo-Fenton (pH = 3) MQ

0.0161

0.1443

0.1814

0.2505

0.2363

As an example, the kHO /kphot ratio was calculated 5.5 for 10 mg/L (more details are given in Table 7.3).

Hence, although these measurements reveal the high contribution of the photolysis in the process, they

can be partially attributed to the specific reactor geometry that has relatively short optical path, and light

attenuation is small. Therefore, this design influences the economical parameters of the degradation

process.

Nevertheless, the consumption of H2O2 increases with increasing addition. We further notice that the

overall oxidation of the system proceeds towards mineralization of the existing carbon content for each

case. A 4-h exposure to UV/H2O2 system with 50 mg/L H2O2 (i.e. for as long as there was H2O2 present) the

COD and TOC removal is improved compared to the plain UV system, and an additional 20% was removed

187

for both parameters. However, according to the absorbance spectra recorded, after the removal of VFA,

the remaining intermediates and by-products are not removed equally fast, and a lower second order kHO

must be in effect (detailed graphs can be found in the supplementary Figure S2).

7.3.2. Fenton-related AOPs degradation of Venlafaxine

In order to fully attribute the effects of the synthetic photo-Fenton process against the degradation of

VFA, a stepwise construction of the process took place. Hence, solar exposure, Fenton treatment in the

dark and the combined process took place and the results are presented in the respective groups. To our

knowledge, this is the first instance where VFA is systematically treated by the photo-Fenton process and

therefore, the different parts will be analyzed separately.

7.3.2.1. Solar photolysis of Venlafaxine

The experiments of simulated solar exposure of VFA were performed in order to establish solar photolysis

rates and check the potential photo-transformation of the drug. Santoke et al. (Santoke et al. 2012)

proved that Venlafaxine is undergoing limited photolysis. Here, after 24 h of irradiation at relatively high

solar irradiance (900 W/cm2) a mere 12% of the initial VFA amount has been removed. As such, a ksol =

0.0002 min-1 was measured (Table 7.3). As far as the COD and TOC of the solution are concerned, limited

removal was observed. COD was removed at 13% and 4% TOC was eliminated during the course of 24h

(for more details, see Supplementary Figure S2). However, the direct action of solar light includes 1) the

excitation of the organic compound at singlet-excited state (Ryan et al. 2011), 2) its intersystem crossing

to triplet state and 3) its reaction with oxygen to form singlet oxygen (Vione et al. 2014). Afterwards, the

micropollutant returns to ground state, but the singlet oxygen created by the reaction with water

participates in i) the superoxide radical anion formation from oxygen and consequently ii) to the formation

of H2O2 from water (Vione et al. 2014) or iii) the direct attack to double bonds present in the molecule.

Although of lesser importance, these results will play the role of reference when the photo-Fenton process

will be described from its parts.

7.3.2.2. Fenton-driven degradation of Venlafaxine in the dark

The degradation of VFA in the dark due to the Fenton reaction were assessed in a range of parameters,

such as the initial pH and the starting Fenton reagents concentration. Literature suggests that the

reactivity of Venlafaxine with hydroxyl radicals is ranging among 8x109 to 1010 M-1 s-1 (see Table 7.1). Figure

3 presents 4 of the Fe|H2O2 ratios tested (indicatively chosen), in the 3 different pH levels of operation,

i.e. 3, 5 and 7. The results of the optimization are summarized in Figure 7.3a (analytical data in

supplementary Figure S3a-S3c), for 24 h of treatment for each batch process.

188

5|10 5|50 12.5|30 20|501

10

100

1000

t 90% (h

)

Fe|H2O2 ratio

pH= 3 pH= 5 pH= 7

0 4 8 12 16 20 24

0.0

0.1

1.8

1.9

2.0

2.1

2.2

2.3

2.4

2.5

CO

D/T

OC

ratio

Time (h)

5|10 5|50 12.5|30 20|50-3 20|50-5 20|50-7

Figure 7.3 – Treatment of Venlafaxine by the Fenton process in the dark. a) Evolution of the t90% with

increasing Fe|H2O2 ratio and pH. b) COD/TOC ratio modification by the Fenton process at various

Fe|H2O2 ratios (pH=3) and increasing pH for a 20|50 (mg/L|mg/L) ratio.

As expected, Figure 7.3 shows that at pH=3 the results were optimal (see also Table 7.3), as the Fenton

process greatly benefits of the soluble iron at that pH (Pignatello et al. 2006). Although limited by the

regeneration of Fe3+ back to Fe2+, 3 out of the 4 processes were able to degrade more than 50% of the

initial content. These processes contained H2O2 in the highest amount, which played the role of the

reductant of ferric iron, as indicated in the following reactions:

● (1.8)

● (7.7)

● ● (7.8)

● ● (7.9)

● (7.10)

● (7.11)

● (7.12)

● ● (7.13)

189

The generation of the Reactive Oxygen Species (ROS) by the Fenton process is dependent on the catalyst

regeneration, i.e. the Fe3+ back to Fe2+. However, the reaction kinetics of equation 1.8 compared with the

others is orders of magnitude higher, and for this reason, the regeneration of the catalyst, especially by

Eq. 7.8, is considered a limiting step. In water, Fe3+ forms different aqua-complexes, according to its pH,

with precipitation tendencies as the pH increases. The iron speciation in water changes according to the

Pourbaix diagram and Salgado et al. (Salgado et al. 2013) have updated the original contribution, with the

most recent knowledge. At acidic pH, the most abundant form of iron is [Fe2+] which is the strongest

reductive form. The optimal value to perform the Fenton reaction is near 3; the rate increases with the

pH (Millero and Sotolongo 1989). After pH 4, the majority of the iron species are insoluble (Fe2O3, or other

colloidal forms) and therefore, the reactivity drops.

However, moving up to pH=5 the degradation potential are diminished and only the 20|50 addition

Fe|H2O2 ratio was sufficient to inflict a 70% degradation after 24 h. At pH = 7 even this ratio was limited

to a 35% degradation of VFA. Nevertheless, as 24 and 22 mg/L H2O2 were still measured after 24 h, the

process could continue, albeit in lower rates. At pH 5, solid Fe(OH)2 species dominate, since they are

more readily oxidized (compared to Fe2+ and FeOH+) (Morgan and Lahav 2007). Nevertheless, until pH 5.8

the contribution of the homogeneous photo-Fenton is still considerable (Barona et al. 2015). Finally

among pH 5 and 7 (up to 8), solid iron species dominate and no further increase in its concentration

appear (Morgan and Lahav 2007). However, at this pH HO2●─ is formed in higher quantities and indirectly

helps the Fe2+ formation through the Equation 7.12 (Papoutsakis et al. 2015b). The soluble iron species

have very different rate constants reported, which practically means that their participation in redox

reactions depends on the distribution of these species (Fe2+, FeOH+, and solid Fe(OH)2) (Morgan and Lahav

2007).

As pseudo-first order kinetics were established for the degradation process, the t90% was established for

each ratio and pH level. It can be observed that the theoretical t90% can reach 1000 hours in the near-

neutral pH and low Fe|H2O2 ratios, although increasing the amounts can reduce it to merely a day, which

is great improvement. The acidic pH on the other hand ensures proper removal and never exceeds 100 h

of treatment, while 11 h are necessary with high reactants concentration.

Finally, during the 24-h treatment by the Fenton process the mineralization rate of the organic matter

remains low for high pH and low concentrations of Fe and H2O2. The biggest removal noted was at pH=3

and 20|50 ratio. Nevertheless, the COD/TOC ratio indicates a fast initial degradation step and a

decelerated process afterwards. As no process was H2O2-limited for any pH or ratio, this indicates that the

VFA structure contains some easily removed groups, which readily react with the HO● radicals (see

supplementary Figure S4 and Table S1, for analytical COD & TOC measurements, and H2O2, respectively).

The corresponding absorbance spectra in Figure 7.4 indicate the formation of different intermediates and

190

complexes with iron, depending on the pH and the Fe|H2O2 ratio. Nevertheless, the VFA removal is

confirmed to be low for most cases, and the participation of the iron is demonstrated by the absorbance

in higher wavelength UV and visible light; Fe can bind to acidic groups or side-chains and create stable

organo-complexes. In Figure 7.4, axis x shows the wavelength (nm), y the absorbance (a.u.) and z either

the Fe|H2O2 ratio (left group) or the pH level (right group). Also, the stability of the solution after 6 h,

indicates the low levels of reaction with the organics present. However, these absorbance spectra

suggests that these complexes are photo-active, as they absorb UV and visible light in higher rates than

the original solution with iron (t=0) and therefore we anticipate their possible involvement in the photo-

Fenton process.

Figure 7.4 – Absorbance spectra during the 24-h Fenton treatment of Venlafaxine, for various Fe|H2O2

ratios (panels a-d) and pH levels (panels i-iii). Fe|H2O2 ratios: a) 5|10, pH=3. b) 5|50, pH=3. c) 12.5|30,

pH=3. d) 20|50, pH=3. pH levels: i) 20|50, pH=3. ii) 20|50, pH=5. iii) 20|50, pH=7.

7.3.2.3. The photo-Fenton driven degradation of Venlafaxine

The final step in this section considers the combined photo-Fenton process. In Figure 7.5, we summarize

the experiments performed, in a similar manner to the Fenton process. As pseudo-first order kinetics were

established, Figure 7.5a shows the evolution of t90% as the Fenton reactants and the pH levels increase

(detailed data on the photo-Fenton action can be found in the Supplementary Figure S5 and the H2O2

consumption at Table S1). A very similar trend with the Fenton process is observed, but the t90%, even the

theoretical one is now measured in minutes rather than hours. The synergy of the Fenton with light is very

high, yielding t90% as low as 10 min for 20|50 at pH=3.

191

5|10 5|50 12.5|30 20|501

10

100

1000t 90

% (m

in)

Fe|H2O2 ratio

pH= 3 pH= 5 pH= 7

0 30 60 90 120 150 180

0.0

0.1

1.8

1.9

2.0

2.1

2.2

2.3

2.4

2.5

CO

D/T

OC

ratio

Time (min)

5|10 5|50 12.5|30 20|50-3 20|50-5 20|50-7

Figure 7.5 – Treatment of Venlafaxine by the photo-Fenton process (irradiance: 900 W/m2). a)

Evolution of the t90% with increasing Fe|H2O2 ratio and pH. b) COD/TOC ratio evolution by the solar

photo-Fenton process at various Fe|H2O2 ratios (pH=3) and increasing pH for a 20|50 ratio.

As the mode of action of the Fenton process has been previously explained, here we will assess only the

changes and improvements inflicted by light. The most notable difference among these results and the

Fenton process in the dark, is the change in the time scale, from hour to minute range. Here, all

experiments were handled within 3 hours. Although some processes (5|10 ratio) did not conclude, the

great number of completed experiments within the experimental time demonstrates the efficiency of the

photo-Fenton process. In presence of light, the iron recycling is facilitated by the absorption of light by

the photosensitive aqua-hydroxy- and organo-complexes. At pH=3, the prevalent form is [Fe(H2O)5(OH)]2+;

in general, the photo-reduction can be summarized with the following two reactions (15-16), also leading

to an extra hydroxyl radical production:

(7.14)

● (7.16)

At higher pH, the contribution in radicals’ formation is reduced, along with the photo-active compounds

concentration. The production of the hydroxyl radical, as described before, induces hydroxylation of two

possible sites simultaneously or independently, or even reaction with the nitrogen group (Santoke et al.

2012), resulting to efficient degradation (Klamerth et al. 2012, Mackuľak et al. 2015).

As the mineralization rate indicates (figure 7.5b), he overall removal of organic matter in the solution stays

limited, within the duration of the experiment. According to the H2O2 consumption rates, and the residual

H2O2 at the end of 3h, which is lower than the respective after 24 in the Fenton process, the process is, or

will be H2O2 limited for some combinations. Hence, the COD and TOC removal is halting and re-dosing

192

would be necessary to continue the degradation (detailed COD and TOC measurements can be found in

the supplementary material).

Figure 7.6 – Absorbance spectra during the 3-h photo-Fenton treatment of Venlafaxine, for various

Fe|H2O2 ratios (panels a-d) and pH levels (panels i-iii). Fe|H2O2 ratios: a) 5|10, pH=3. b) 5|50, pH=3. c)

12.5|30, pH=3. d) 20|50, pH=3. pH levels: i) 20|50, pH=3. ii) 20|50, pH=5. iii) 20|50, pH=7.

Finally, the combined photo-Fenton process effects are also demonstrated in the changes in absorbance

spectra, in Figure 7.6. The photoactivity of the complexes remains high, which means an active ligand-to-

metal charge transfer could be facilitated:

● (7.17)

The sacrificial ligand is offered from either VFA or some intermediate and the regenerated Fe2+ will re-

participate in the Fenton reaction. The changes found can be grouped under either i) the higher complex

formation (higher absorbance) compared to the Fenton system, ii) the significantly faster plateau

achievement or iii) the bigger differences in the neutral and near-neutral process (differences between

pH=3 and 7).

7.3.3. Venlafaxine degradation experiments in wastewater and urine

7.3.3.1. Experiments in real secondary wastewater effluents and human urine

The occurrence of VFA and its metabolites in surface waters (Rúa-Gómez and Püttmann 2012) indicates

the improper elimination in WWTPs and their extreme adverse effects underlines the need for their

193

degradation prior to their discharge. Table 7.4 summarizes literature values and own measurements of

VFA in WWTPs, at various stages, from influents to effluents.

Table 7.4 – Occurrence and fate of Venlafaxine in urban WW effluents.

Venlafaxine occurrence in

WWTPs

Treatment stage

Quantity (ng/L) Source

Pre-

trea

tmen

t Influent

500-1100 (Metcalfe

et al. 2010)

800

(Rúa-Gómez

and Püttmann

2012)

623±18 Own

measurements

235±21 (Margot

et al. 2013)

Primary treatment 374±111

Own measure

ments

Seco

ndar

y Tr

eatm

ent

Biological treatment (high/low

HRT)

900-1000 (Metcalfe

et al. 2010)

Degradation of Venlafaxine (this work)

140-150 (Petrie et al. 2015) Measured pseudo-first order kinetics k (min-1)

95-188 (Margot

et al. 2013)

UV-C UV/H2O2 Solar Fenton photo-Fenton

AS 281±83 Own

measurements

0.073 nc 0.0001 0.006 0.033

MBBR 224+67 Own

measurements

0.08 nc 0.0001 0.007 0.035

CF 299+9 Own

measurements

0.042 nc 0.0001 0.001 0.032

194

Effluents 95-188 (Margot

et al. 2013)

In our work, we assessed the degradation of VFA with the same AOPs analyzed in the previous parts, in

the amounts found in three different secondary effluents, namely activated sludge (AS), moving bed bio-

reactors (MBBR), and coagulation-flocculation effluent. For more information on the nature and the

composition of the effluents, as well as COD and TOC removal, interested readers should refer to

(Giannakis et al. 2015c), (Cunningham 2004) and (Edberg et al. 2000) and the supplementary material

(Table S2). Figure 7.7 depicts the degradation measurements in the two families of AOPs per process and

per effluent. As expected, the UV-based AOPs presented the fastest degradation kinetics, as summarized

also in Table 7.4.

Figure 7.7 – Treatment of Venlafaxine by AOPs in urban WW effluents. The experimental conditions

are marked in the corresponding graphs. a) VFA degradation by UV-based AOPs in AS, MBBR and CF

effluents. b) VFA degradation by the Fenton-related processes in AS, MBBR and CF effluents.

Recent works have demonstrated their efficiency in VFA degradation by the Fenton and Fenton-like

processes (Mackuľak et al. 2016, Mackuľak et al. 2015), confirming the feasibility of its application in real

WW samples. Venlafaxine degradation is a function of contradicting factors in real effluents. On the

antagonists of the process, we can mention the i) suspended solids, blocking UV and solar light

transmission, ii) the Effluent Organic Matter (EfOM), consisting in still particulate organic matter (POM),

biodegradable organic matter, refractory organic matter and other MPs, which all compete for the

oxidants generated by AOPs (Giannakis et al. 2015c) and screen the light (Ryan et al. 2011), iii) the ROS

scavengers, such as (bi)carbonates, nitrate and nitrite (Vione et al. 2014), and iv) the microorganisms. On

195

the other hand, the very presence of some substances has been proven to enhance the self-purification

capabilities of the effluents, such as a) the presence of photo-sensitizable organic matter (PhOM), which

further produces ROS, and b) the nitrates and the carbonates, which contribute in producing nitrate

radicals, carbonate radicals and ROS, all with mild oxidative potential (Giannakis et al. 2015c). In the end,

what we perceive as “degradation” is the net force of all these factors, which leans on the negative side

overall, compared to simulated WW or water.

As far as the initial content is concerned, the similar values in AS and MBBR have been recently verified

(Margot et al. 2013) and the close values in CF effluents are a result of the low solubility and the

hydrophobicity, as expressed by the logkow; VFA tends to adsorb to the generated flocs and is “removed”

by settling. More specifically, the degradation kinetics follow a similar trend to the content of suspended

matter and organic content in the effluents, as well as the alkalinity of the matrix, an indicator of

(bi)carbonates content and therefore a precursor of HO scavenging. As shown in (Margot et al. 2013) and

Chapter 6, the physicochemical characteristics of the effluents measured, verify a trend, as MBBR has the

best characteristics, followed by AS and then CF.

Concerning the experiments in human urine, based on manufacturer, medical and pharmaco-kinetical

data, we have found that the normal dose of Venlafaxine (as Effexor) for patients is 75 mg/day and can

reach up to 150 mg/day in severe cases. Out of the administered amount, 92% is recovered in urine, as

the renal excretion pathway is prevailing. However, VFA in urine appears in only 5% (1-10%, (Metcalfe et

al. 2010), unconjugated O-desmethyl Venlafaxine (29%), conjugated O-desmethyl Venlafaxine (26%), 1%

N-desmethyl Venlafaxine, and the rest as the other intermediates. Hence, a normal person excretes ~2 L

urine per day, it is normal to expect μg/L concentrations in patients. As such, 10 μg/L spiking was

performed in urine collected by healthy individuals. The COD of the solution varied significantly from 2.5

to 8.5 g/L. Therefore, collection and homogenization over the course of 6h was performed to mitigate the

differences in chemical and optical properties. The average urine characteristics can be found in the

supplementary material (Table S2). Besides, 10% diluted urine experiments took place, to assess the

possibility of yielding higher UV transmittance in exchange of higher treatment volumes. The results of

the study are summarized in Figure 7.8.

196

Figure 7.8 – Treatment of Venlafaxine by UV-based methods in human urine. a) VFA degradation by

UV-based AOPs (0, 50 or 100 mg/L H2O2 and 0/100% or 10%-90% urine/water ratio. b) COD reduction

and DOC (0.45μm filtration) removal in the same conditions.

During the experiments in undiluted urine, the efficacy of UV alone in degrading VFA was low. As urine

contains light absorbing compounds (organic matter, nitro-, phosphoro- and other groups), light

attenuation was a limiting step in the degradation process (Iohexol, Chapter 6). The step-wise addition of

H2O2 (50 and 100 mg/L) was beneficial, reaching up to ~30% degradation of VFA. Other researchers have

also demonstrated the increase in PhACs’ degradation by the addition of H2O2 in the bulk (Zhang et al.

2015). The positive aspects of H2O2 addition are found also in the COD removal, where up to 15% of COD

and 13% of DOC were removed. In real urine, this amount corresponds to 750 mg/L, from a 5000-average

COD. Diluting the urine x10 times has modified the matrix significantly, allowing up to 80% degradation

of VFA (spiking was done after dilution) and the H2O2 addition further improved degradation, up to 100%.

An interesting correlation can be derived from the COD removal and the VFA degradation, where a 20%

removal of COD (or 15% DOC) was found to indicate the end of the VFA treatment period. Hence, this

process could be monitored by following global parameters instead of sophisticated LC/MS methods.

Furthermore, the dilution of urine has been studied in another contaminant (Iohexol, Chapter 6), and the

optimization experiments revealed that a near-optimal performance could be obtained by 10% dilution

of the matrix. The effect of dilution improves the transmittance of urine, thus improving light absorption

by the organic matter in solution, but also improve the homolytic disruption of H2O2, the subsequent

197

increase in HO● production and the decrease in scavenging by the organic matter. On the other hand, the

x10 times increase of the treated volume holds technical and engineering implications, as well as

questions on the water used for dilution. Here, as a proof of concept we have shown that the effects are

multiple and since the human urine production is small, even this dilution is not a limiting agent to a

potential application.

Finally, no photo-Fenton experiments were performed as the k in synthetic urine were too slow and the

treatment of urine in open vessels seems impractical. Nevertheless, medium (MP) UV lamps could be

suggested as a potential substitute for monochromatic UV or solar light, as they emit an array of peaks

and photo-Fenton reaction could be sustained.

7.3.4. Elucidation of the AOP-driven degradation pathway and inherent biodegradability

properties of Venlafaxine

7.3.4.1. Degradation pathway of Venlafaxine by AOPs through TOF-MS

The identification of degradation products is essential for providing the risk assessment information of

drug residues in the environment, as well as for the improvement of water treatment technologies. In our

study, Venlafaxine degradation products were identified after each degradation procedure (UV, UV/H2O2,

solar treatment, Fenton and solar photo-Fenton). All identified products were shown in supplementary

Table S3, including their molecular formula, theoretical and experimental m/z value, double bond

equivalent (DBE) and mass accuracy in ppm (accepted structures with error less than 5 ppm). Ten

degradation products were identified in overall: one for solar treatment, six for UV treatment, five for

UV/H2O2 treatment, seven for Fenton degradation and seven for photo-Fenton treatment. Based on the

identified structures, a simple mechanistic scheme was proposed (Figure 7.9).

Transformation of Venlafaxine can occur via four dominant reactions: 1) sequential hydroxylation of the

aromatic ring, 2) transformation of the methoxy- group, 3) hydroxylation and shortening of the

cyclohexanol ring and 4) attack on the nitrogen group.

Firstly, the degradation pathway of VFA was sequential hydroxylation: once hydroxylated VFA yielding

m/z 294.2072 (P8) was identified in all treatments except solar treatment; followed by di-hydroxylated

product with m/z 310.202 (P9) identified in UV, UV/H2O2 and photo-Fenton treatment; and tri-

hydroxylated product with m/z 326.1956 (P10) identified only in the UV treatment (Treatment B). The

sequence of the identified hydroxylated degradation products implies that UV light was the main driving

force in the multiple hydroxylation of the aromatic ring. Here, the identified products were also reported

in the literature (García-Galán et al. 2016, Lambropoulou et al. 2016, Lester et al. 2014, Santoke et al.

2012). Santoke et al. identified two products formed by the attack of HO● radical on the aromatic ring and

198

the N-chain, but they were not identified in this study (marked with red dashed arrow in the degradation

pathway) (Santoke et al. 2012).

Figure 7.9 – Combined Venlafaxine degradation pathway through the application of the treatment

methods analyzed.

A second degradation pathway started with dehydrated Venlafaxine’s product m/z 260.2012 (P6),

identified in all degradation procedures. Combinations of dehydration and hydroxylation reactions

present a transformation pathway also dominant for VFA UV/TiO2 treatments (Lambropoulou et al. 2016).

HO● attack on the tertiary C-atom and cyclohexanol structure led to the formation of products with m/z

215.1431, m/z 229.1429 and m/z 292.1914; P4, P5 and P7, respectively.

Demethylation presents a well-known VFA transformation route in biological and chemical reactions (Boix

et al. 2016, Li et al. 2015, Santoke et al. 2012). Transformation of the methoxy group was identified within

the products m/z 121.0654 (P1), m/z 178.1231 (P2) and m/z 194.1182 (P3), which were at the same time

the final degradation products. Aliphatic nitrogen products identified by Garcia-Galan et al. (García-Galán

et al. 2016), were not identified in this study, however they were also included in the degradation scheme

(marked with blue dashed arrow in the degradation pathway) to complement the overall VFA degradation

routes. It should be noted that the structures of identified products for different UV/H2O2 treatments

depend not only on the reaction time, but also on the H2O2 concentration used in the experiments. Finally,

the appearance of apparent biodegradable compounds calls for assessment of the biodegradability

assessment of VFA and the AOP-treated effluents containing it.

199

7.3.4.2. Zahn-Wellens (ZW) biodegradability test of AOP-treated Venlafaxine solutions

As literature suggests low removal of VFA in WWTPs, we assessed the biodegradability of VFA, by

subjecting it first through an AOP. This strategy has been successfully used in various effluents (Malato et

al. 2009, Sarria et al. 2003). Therefore, treatment of VFA solutions until 50% and 100% initial

concentration degradation, along with a reference compound (diethylene glycol) and untreated VFA were

subjected to a 28-day Zahn-Wellens inherent biodegradability tests (Lin and Ganesh 2013, Nataro and

Kaper 1998). In parallel, DOC was followed at the corresponding blanks (test suspension and inoculum

blank). In order to avoid self-inhibition problems, the initial VFA amount was reduced to 10 mg/L. The

results of the study are summarized in Figure 7.10.

Figure 7.10 – Zahn-Wellens inherent biodegradability test of Venlafaxine and treated solutions in MQ.

a) ZW test after treatment of 50% of the initial VFA solution. b) ZW test after treatment of 100% of the

initial VFA solution. Note that results are normalized towards the initial DOC to enable comparison.

First of all, the test is considered valid as 70% of the initial DOC of the reference compound has been

degraded (>72%) within 14 days. This indicates the suitability of the activated sludge inoculum. Secondly,

VFA alone was removed at 35%, which corresponds to similar degradation rates of VFA in biological

treatment facilities (see Table 7.4). As far as the applied AOPs are concerned, 50% pre-treatment of VFA

resulted in 20-25% biodegradability improvement. The most efficient process was the photo-Fenton

reaction, only by marginal difference. According to TOF- MS analysis for VFA in this work, and Orbitrap-

MS for Iohexol (Chapter 6), using AOPs where iron is involved always leads to enhanced modifications on

the target contaminant. If VFA was eliminated 100% a further 10-15% was achieved, depending on the

process. This indicates that over the course of 28 days, almost 70% of the initial DOC was eliminated,

reaching the threshold for considering the solution biodegradable. Of course, further treatment of the

parent solutions would achieve the threshold with greater ease, and correlation with the initial DOC

200

removal should be made instead. Hence, by extrapolation, it could be possible to propose a pre-treatment

step in industries or hospitals, where mass flows of similar contaminants are released, if the said facilities

do not employ their own WWTPs, as it would seriously ease the burden off the municipal WWTPs.

7.4. Conclusions

The ubiquitous presence of drugs in surface waters demands strict control frameworks and efficient

removal methods at the level of WWTPs. Under this scope, the degradation of the antidepressant

Venlafaxine was systematically investigated, through the application of 5 AOPs. UV-based technologies

(UV-C light alone and UV/H2O2) and Fenton-related techniques (solar photolysis, Fenton and photo-

Fenton oxidation) were assessed as control measures and their efficiency was estimated.

The investigation on the degradation kinetics has shown that Venlafaxine demonstrates moderate

photolysis under UV, and the addition of H2O2 with the simultaneous HO● generation enhances the

degradation potential of the chosen treatment. On the other hand, solar photolysis was found limited,

but in combination with the action of the Fenton process (in the dark), the photo-Fenton process was

efficient in degrading the contaminant, with decreasing, but not diminishing performance tendencies as

we approached the neutral pH.

The tests in wastewater and urine revealed a drop in efficiency, due to the presence of antagonists in the

matrix. Urban wastewater and human urine tests indicated that the actual conditions expected in the field

demand intensive treatment; in wastewater the degradation of Venlafaxine is subjected to similar

problems as most ng/L contaminants present, but the efficient removal is possible, and the human urine

experiments indicate an innovative treatment proposal, by the use of UV to collect and treat on-site the

emerging contaminants, and addition of H2O2, if high simultaneous DOC removal is desired, before

dispersion in the wastewater matrices.

The mechanistic interpretation (degradation pathway) based on our own TOF-MS experiments and recent

advances in the field revealed the opportunity of converting Venlafaxine to its bio-degradable

intermediates. The Zahn-Wellens tests performed showed that pre-treatment of Venlafaxine solutions

increases biodegradability, and under certain conditions, conversion of the mixture of intermediates into

biodegradable is possible.

In the light of the above findings, we conclude that the non-selective and highly oxidative character of the

Advanced Oxidation Processes is capable in controlling substances before their discharge in natural

aquifers and their upcoming environmental consequences. Environmental protection has a well-

201

established ally in traditional contaminant categories (priority pollutants, organic matter) and the

application of AOPs can play a role of outmost importance towards this direction in the near future.

8. Chapter 8 - General conclusions, perspectives and future

work

In this thesis, the use of AOPs was selected towards pollutant decontamination and disinfection of

effluents. UV, UV/H2O2, solar light (shown to work as an AOP), Fenton and solar photo-Fenton are

established as powerful allies in the ongoing task of wastewater purification. From the various works

analyzed, the key conclusions are the following:

1) UV-based AOPs are efficient for MP removal and MO inactivation. Although changing dynamically,

the Swiss reality on hospital wastewater treatment dictates their discharge in the municipal

collection network, and therefore imply their co-treatment with municipal wastes. The UV-based

AOPs (UV and UV/H2O2) were found to be effective micropollutant removal strategies in ng/L level

and bacterial inactivating processes, after biological secondary pre-treatment, as found in

municipal wastewaters. When used in simulated hospital wastewaters and urine treatment, as

alternative micropollutant elimination strategies, their efficiency was measured and established

against a list of contaminants, with parallel elimination of the contained organic matter. The

degradation was fast although the engineering parameters and costs were not part of this study,

the reactants addition and necessary light doses were moderate.

2) The solar photo-Fenton process and its constituents can be very effective in the proper context.

Despite the lower apparent efficiency of this process when compared with its UV-based

counterparts, photo-Fenton was found to effectively and non-selectively remove micropollutants

and effluent organic matter. Furthermore, their application resulted in high bacterial removal,

regrowth suppression, yeasts and viruses inactivation from water and wastewater effluents. Most

importantly, through systematic studies the mechanism and the key points of the process against

the aforementioned targets were characterized. Special emphasis was given to the organic matter

present in WW, as it is found to hinder the inactivation process but other benefits, such as iron

complexation, also occur.

3) The selected model hospital/industrial contaminants (Iohexol, Venlafaxine) helped elucidate the

pitfalls and opportunities in HWW treatment by AOPs. The AOPs were found to work particularly

well against the concentrated, (simulated) industrial wastewater, hospital flows and urine.

Therefore, their application in hospitals and related industrial activities is promising. Also, the

structural deformation of the selected pollutants provided helpful insights on the operational and

chemical constraints on applying the various AOPs; for instance the use of iron (when H2O2 is

202

present) is strongly recommended for faster and more intense degradation of the contaminants

in HWW. Finally, apart from the degradation point of view, the AOPs studied increased the

biodegradability of the selected compounds treated solutions, which could allow their use as pre-

treatment methods in HWWTPs.

In conclusion, more work is necessary to establish these methods as suitable for application in hospital

environments. However, the initial results strongly support their further development, and future work

stemming from the present research is encouraged to be sought.

To begin with, the combination of three secondary pre-treatment methods with an AOP as a post-

treatment showed great potential in micropollutant elimination and enhanced bacterial disinfection, with

regrowth repression. Although certain combinations (e.g. MBBR+UV/H2O2) were very efficient, the overall

findings indicate great potential for the UV-based methods in developed countries. Some topics that need

to be addresses involve:

- The treatment combinations in pilot- or full-scale.

- Assessment of AOPs in lab- or pilot scale in conjunction with the recent developments in

secondary treatment (SBR, Annamox etc.).

- The optimization of the full-scale treatment for micropollutant removal and microorganism

elimination; Optimization of operational parameters, such as H2O2, UV light lamp types etc.

- Engineering proper reactors for WW treatment by AOPs.

- Techno-economical assessment of the proposed treatment processes.

Nevertheless, in developing countries the thought of WW treatment is at infant stage at most. Therefore,

the focus should be turned towards the improvement of the existing strategies. Almost intuitively, the

populations have developed the lagoons as a means of WW retention and (accidental) treatment, the

photo-Fenton process could offer a barrier towards surface water decontamination. The sunlight and the

inherent iron content of waters provide a good starting point for further exploration of these techniques.

Also, developing simple operating and performance monitoring aspects could be explored.

The study of the photo-Fenton process as a microorganism disinfection technique revealed that under the

eye of hydroxyl radicals, all microbes are the same. From the simple MS2 coliphage, to the “fortified”

yeast model tested, microbial disinfection was eventually achieved. However, the proposed mechanistic

model for both viruses and yeasts, contains many variables that escaped the focus of this work and

demand further investigation. Some propositions are:

- The quantification of ROS production in WW and their contribution towards inactivation.

- The characterization of the kinetic constants throughout the complex mechanisms proposed.

- The contribution of endogenous inactivation, for the different microbial species.

203

- Extension towards human pathogens, in real WW (urban and hospital), and achievement of

proper inactivation, through the establishment of an order of removal among the pathogenic

species.

As far as the hospital WW treatment is concerned, setting up a local HWWTP with biological treatment

and AOPs as polishing steps would be ideal. Nevertheless, the primary evidence collected indicated great

potential for the in-situ urine treatment in health facilities. As it appears, dispersion of the chemical

contaminants in HWW and then in UWW collection systems will require higher residence times of

advanced treatment solution for their elimination. On the other hand, focusing on small, concentrated

volumes at hospital level could remove a burden from the downstream treatment. The use of medium-

pressure UV lamps and the photo-Fenton treatment is a potentially powerful process, as shown by the

corresponding low-pressure UV/H2O2/Fe2+ process. A wider list of MPs, and exploration of a portable

reactor for urine treatment could facilitate this treatment process. Also, although urine leaves the kidneys

microorganism-free, microbiological assessment should be conducted, in order to ensure lack of cross-

contaminations, safe recovery and explore potential reuse (e.g. for phosphorus recovery).

205

9. References

Abdelmelek, S.B., Greaves, J., Ishida, K.P., Cooper, W.J. and Song, W. (2011) Removal of pharmaceutical and personal care products from reverse osmosis retentate using advanced oxidation processes. Environmental Science & Technology 45(8), 3665-3671. Abu-ghararah, Z.H. (1994) Effect of temperature on the kinetics of wastewater disinfection using ultraviolet radiation. Journal of Environmental Science & Health Part A 29(3), 585-603. Altin, A., Altin, S. and Degirmenci, M. (2003) Characteristics and treatability of hospital(medical) wastewaters. Fresenius Environmental Bulletin 12(9), 1098-1108. Alvares, A.B.C., Diaper, C. and Parsons, S.A. (2001) Partial Oxidation by Ozone to Remove Recalcitrance from Wastewaters - a Review. Environmental Technology 22(4), 409-427. Ananthaswamy, H. and Eisenstark, A. (1976) NEAR-UV-INDUCED BREAKS IN PHAGE DNA: SENSITIZATION BY HYDROGEN PEROXIDE (A TRYPTOPHAN PHOTOPRODUCT). Photochem Photobiol 24(5), 439-442. Andreozzi, R., Caprio, V., Insola, A. and Marotta, R. (1999) Advanced oxidation processes (AOP) for water purification and recovery. Catalysis Today 53(1), 51-59. Andrews, J.M., Ninan, P.T. and Nemeroff, C.B. (1996) Venlafaxine: a novel antidepressant that has a dual mechanism of action. Depression 4(2), 48-56. Antonini, E. and Vidic, H. (1994) Complejo de hierro-citrato, procedimiento para su fabricación y su aplicación farmacéutica. marca, Oedpy (Ed.), Spanish. Backhaus, T., Altenburger, R., Arrhenius, Å., Blanck, H., Faust, M., Finizio, A., Gramatica, P., Grote, M., Junghans, M., Meyer, W., Pavan, M., Porsbring, T., Scholze, M., Todeschini, R., Vighi, M., Walter, H. and Horst Grimme, L. (2003) The BEAM-project: prediction and assessment of mixture toxicities in the aquatic environment. Continental Shelf Research 23(17–19), 1757-1769. Bahnmüller, S., von Gunten, U., Loi, C.H., Linge, K.L. and Canonica, S. (2014) Degradation rates of benzotriazoles and benzothiazoles under UV-C irradiation and the advanced oxidation process UV/H 2 O 2. Water Res. Bandara, J., Pulgarin, C., Peringer, P. and Kiwi, J. (1997) Chemical (photo-activated) coupled biological homogeneous degradation of p-nitro-o-toluene-sulfonic acid in a flow reactor. Journal of Photochemistry and Photobiology A: Chemistry 111(1), 253-263. Barker, D.J. and Stuckey, D.C. (1999) A review of soluble microbial products (SMP) in wastewater treatment systems. Water Res 33(14), 3063-3082. Barona, J.F., Morales, D.F., González-Bahamón, L.F., Pulgarín, C. and Benítez, L.N. (2015) Shift from heterogeneous to homogeneous catalysis during resorcinol degradation using the solar photo-Fenton process initiated at circumneutral pH. Applied Catalysis B: Environmental 165(0), 620-627. Basu-Modak, S. and Tyrrell, R.M. (1993) Singlet oxygen: a primary effector in the ultraviolet A/near-visible light induction of the human heme oxygenase gene. Cancer research 53(19), 4505-4510. Beach, H. (1971) COMPOSITION AND CONCENTRATIVE PROPERTIES OF HUMAN URINE. Bisesi, J.H., Bridges, W. and Klaine, S.J. (2014) Effects of the antidepressant venlafaxine on fish brain serotonin and predation behavior. Aquatic Toxicology 148, 130-138. Boillot, C., Bazin, C., Tissot-Guerraz, F., Droguet, J., Perraud, M., Cetre, J.C., Trepo, D. and Perrodin, Y. (2008) Daily physicochemical, microbiological and ecotoxicological fluctuations of a hospital effluent according to technical and care activities. Science of The Total Environment 403(1–3), 113-129. Boix, C., Ibáñez, M., Sancho, J.V., Parsons, J.R., de Voogt, P. and Hernández, F. (2016) Biotransformation of pharmaceuticals in surface water and during waste water treatment: Identification and occurrence of transformation products. J Hazard Mater 302, 175-187. Borowska, E., Felis, E. and Żabczyński, S. (2014) Degradation of Iodinated Contrast Media in Aquatic Environment by Means of UV, UV/TiO2 Process, and by Activated Sludge. Water, Air, & Soil Pollution 226(5), 1-12.

206

Bosshard, F., Bucheli, M., Meur, Y. and Egli, T. (2010) The respiratory chain is the cell's Achilles' heel during UVA inactivation in Escherichia coli. Microbiology 156(7), 2006-2015. Bradley, I., Straub, A., Maraccini, P., Markazi, S. and Nguyen, T.H. (2011) Iron oxide amended biosand filters for virus removal. Water Res 45(15), 4501-4510. Braun, V. (2001) Iron uptake mechanisms and their regulation in pathogenic bacteria. International journal of medical microbiology 291(2), 67-79. Brinkman, N.E., Haugland, R.A., Wymer, L.J., Byappanahalli, M., Whitman, R.L. and Vesper, S.J. (2003) Evaluation of a Rapid, Quantitative Real-Time PCR Method for Enumeration of Pathogenic Candida Cells in Water. Applied and environmental microbiology 69(3), 1775-1782. Brownell, S.A. and Nelson, K.L. (2006) Inactivation of single-celled Ascaris suum eggs by low-pressure UV radiation. Applied and environmental microbiology 72(3), 2178-2184. Buchanan, W., Roddick, F. and Porter, N. (2006) Formation of hazardous by-products resulting from the irradiation of natural organic matter: Comparison between UV and VUV irradiation. Chemosphere 63(7), 1130-1141. Buettner, G. (2013) Molecular Targets Of Photosensitization. Cabiscol, E., Piulats, E., Echave, P., Herrero, E. and Ros, J. (2000) Oxidative Stress Promotes Specific Protein Damage inSaccharomyces cerevisiae. Journal of Biological Chemistry 275(35), 27393-27398. Cadet, J., Sage, E. and Douki, T. (2005) Ultraviolet radiation-mediated damage to cellular DNA. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis 571(1), 3-17. Canonica, S. (2007) Oxidation of aquatic organic contaminants induced by excited triplet states. CHIMIA International Journal for Chemistry 61(10), 641-644. Canonica, S., Meunier, L. and Von Gunten, U. (2008) Phototransformation of selected pharmaceuticals during UV treatment of drinking water. Water Res 42(1), 121-128. Carballa, M., Omil, F., Lema, J.M., Llompart, M.a., Garcıa-Jares, C., Rodrıguez, I., Gomez, M. and Ternes, T. (2004) Behavior of pharmaceuticals, cosmetics and hormones in a sewage treatment plant. Water Res 38(12), 2918-2926. Carra, I., Sánchez Pérez, J.A., Malato, S., Autin, O., Jefferson, B. and Jarvis, P. (2014) Performance of different advanced oxidation processes for tertiary wastewater treatment to remove the pesticide acetamiprid. Journal of Chemical Technology & Biotechnology, n/a-n/a. Carraro, E., Bonetta, S., Bertino, C., Lorenzi, E., Bonetta, S. and Gilli, G. (2016) Hospital effluents management: Chemical, physical, microbiological risks and legislation in different countries. Journal of Environmental Management 168, 185-199. Carratalà, A., Calado, A.D., Mattle, M.J., Meierhofer, R., Luzi, S. and Kohn, T. (2016) Solar Disinfection of Viruses in Polyethylene Terephthalate Bottles. Applied and environmental microbiology 82(1), 279-288. Carter, M. (2005) Enterically infecting viruses: pathogenicity, transmission and significance for food and waterborne infection. J Appl Microbiol 98(6), 1354-1380. Chen, K., Liang, N., Yang, J. and Zhao, H. (2014) UV-B Irradiation Regulates Apoptosis in Yeast, pp. 1869-1879, Springer. Chen, Y., Liu, Z., Wang, Z., Xue, M., Zhu, X. and Tao, T. (2011) Photodegradation of propranolol by Fe (III)–citrate complexes: kinetics, mechanism and effect of environmental media. J Hazard Mater 194, 202-208. Chevion, M. (1988) A site-specific mechanism for free radical induced biological damage: the essential role of redox-active transition metals. Free Radical Biology and Medicine 5(1), 27-37. Chèvre, N. (2014) Pharmaceuticals in surface waters: sources, behavior, ecological risk, and possible solutions. Case study of Lake Geneva, Switzerland. Wiley Interdisciplinary Reviews: Water 1(1), 69-86. Cieśla, P., Kocot, P., Mytych, P. and Stasicka, Z. (2004) Homogeneous photocatalysis by transition metal complexes in the environment. Journal of Molecular Catalysis A: Chemical 224(1), 17-33. Cirja, M., Ivashechkin, P., Schäffer, A. and Corvini, P.F. (2008) Factors affecting the removal of organic micropollutants from wastewater in conventional treatment plants (CTP) and membrane bioreactors (MBR). Reviews in Environmental Science and Bio/Technology 7(1), 61-78. Clara, M., Strenn, B., Gans, O., Martinez, E., Kreuzinger, N. and Kroiss, H. (2005) Removal of selected pharmaceuticals, fragrances and endocrine disrupting compounds in a membrane bioreactor and conventional wastewater treatment plants. Water Res 39(19), 4797-4807.

207

Comninellis, C., Kapalka, A., Malato, S., Parsons, S.A., Poulios, I. and Mantzavinos, D. (2008) Advanced oxidation processes for water treatment: advances and trends for R&D. Journal of chemical technology and biotechnology 83(6), 769-776. Cornell, R.M. and Schwertmann, U. (2006) The iron oxides: structure, properties, reactions, occurrences and uses, John Wiley & Sons. Costa, V.M.V., Amorim, M.A., Quintanilha, A. and Moradas-Ferreira, P. (2002) Hydrogen peroxide-induced carbonylation of key metabolic enzymes in Saccharomyces cerevisiae: the involvement of the oxidative stress response regulators Yap1 and Skn7. Free Radical Biology and Medicine 33(11), 1507-1515. Csay, T., Rácz, G., Takács, E. and Wojnárovits, L. (2012) Radiation induced degradation of pharmaceutical residues in water: Chloramphenicol. Radiation Physics and Chemistry 81(9), 1489-1494. Cunningham, V. (2004) Pharmaceuticals in the Environment, pp. 13-24, Springer. Darby, J., Emerick, R. and Loge, F. (1999) The effect of upstream treatment processes on UV disinfection performance, Water Environment Research Foundation. Das, T.K. (2001) Ultraviolet disinfection application to a wastewater treatment plant. Clean Products and Processes 3(2), 69-80. Davey, H.M. (2011) Life, Death, and In-Between: Meanings and Methods in Microbiology. Applied and environmental microbiology 77(16), 5571-5576. Davey, H.M. and Hexley, P. (2011) Red but not dead? Membranes of stressed Saccharomyces cerevisiae are permeable to propidium iodide. Environ Microbiol 13(1), 163-171. Davis, B., Saltman, P. and Benson, S. (1962) The stability constants of the iron-transferrin complex. Biochemical and biophysical research communications 8(1–2), 56-60. De Freitas, J., Wintz, H., Kim, J.H., Poynton, H., Fox, T. and Vulpe, C. (2003) Yeast, a model organism for iron and copper metabolism studies. Biometals 16(1), 185-197. De la Cruz, N., Esquius, L., Grandjean, D., Magnet, A., Tungler, A., de Alencastro, L.F. and Pulgarín, C. (2013) Degradation of emergent contaminants by UV, UV/H2O2 and neutral photo-Fenton at pilot scale in a domestic wastewater treatment plant. Water Res 47(15), 5836-5845. De la Cruz, N., Giménez, J., Esplugas, S., Grandjean, D., de Alencastro, L.F. and Pulgarín, C. (2012) Degradation of 32 emergent contaminants by UV and neutral photo-fenton in domestic wastewater effluent previously treated by activated sludge. Water Res 46(6), 1947-1957. Deblonde, T., Cossu-Leguille, C. and Hartemann, P. (2011) Emerging pollutants in wastewater: a review of the literature. Int J Hyg Environ Health 214(6), 442-448. Del Carratore, R., Della Croce, C., Simili, M., Taccini, E., Scavuzzo, M. and Sbrana, S. (2002) Cell cycle and morphological alterations as indicative of apoptosis promoted by UV irradiation in S. cerevisiae. Mutation Research/Genetic Toxicology and Environmental Mutagenesis 513(1), 183-191. Deneer, J.W. (2000) Toxicity of mixtures of pesticides in aquatic systems. Pest Management Science 56(6), 516-520. Derringer, G. and Suich, R. (1980) Simultaneous optimization of several response variables. Journal of quality technology 12(4), 214-219. Doll, T.E. and Frimmel, F.H. (2004) Kinetic study of photocatalytic degradation of carbamazepine, clofibric acid, iomeprol and iopromide assisted by different TiO 2 materials—determination of intermediates and reaction pathways. Water Res 38(4), 955-964. Douki, T., Reynaud-Angelin, A., Cadet, J. and Sage, E. (2003) Bipyrimidine photoproducts rather than oxidative lesions are the main type of DNA damage involved in the genotoxic effect of solar UVA radiation. Biochemistry 42(30), 9221-9226. Dunlap, C.A., Biresaw, G. and Jackson, M.A. (2005) Hydrophobic and electrostatic cell surface properties of blastospores of the entomopathogenic fungus Paecilomyces fumosoroseus. Colloids and Surfaces B: Biointerfaces 46(4), 261-266. Edberg, S., Rice, E., Karlin, R. and Allen, M. (2000) Escherichia coli: the best biological drinking water indicator for public health protection. J Appl Microbiol 88(S1). Eisenberg, G. (1943) Colorimetric determination of hydrogen peroxide. Industrial & Engineering Chemistry Analytical Edition 15(5), 327-328. Eisenstark, A. (1998) Bacterial gene products in response to near-ultraviolet radiation. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis 422(1), 85-95.

208

El-Ogri, F., Ouazzani, N., Boraâm, F. and Mandi, L. (2016) A survey of wastewaters generated by a hospital in Marrakech city and their characterization. Desalination and Water Treatment 57(36), 17061-17074. Emmanuel, E., Keck, G., Blanchard, J.-M., Vermande, P. and Perrodin, Y. (2004) Toxicological effects of disinfections using sodium hypochlorite on aquatic organisms and its contribution to AOX formation in hospital wastewater. Environment International 30(7), 891-900. Emmanuel, E., Perrodin, Y., Keck, G., Blanchard, J.M. and Vermande, P. (2005) Ecotoxicological risk assessment of hospital wastewater: a proposed framework for raw effluents discharging into urban sewer network. J Hazard Mater 117(1), 1-11. Fatta-Kassinos, D., Meric, S. and Nikolaou, A. (2011) Pharmaceutical residues in environmental waters and wastewater: current state of knowledge and future research. Anal Bioanal Chem 399(1), 251-275. Fent, K., Weston, A.A. and Caminada, D. (2006) Ecotoxicology of human pharmaceuticals. Aquatic Toxicology 76(2), 122-159. Fernandez-Fontaina, E., Omil, F., Lema, J.M. and Carballa, M. (2012) Influence of nitrifying conditions on the biodegradation and sorption of emerging micropollutants. Water Res 46(16), 5434-5444. Fisher, M., Keenan, C., Nelson, K. and Voelker, B. (2008) Speeding up solar disinfection (SODIS): effects of hydrogen peroxide, temperature, pH, and copper plus ascorbate on the photoinactivation of E. coli. Journal of water and health 6(1), 35-51. Fisher, M.B., Iriarte, M. and Nelson, K.L. (2012) Solar water disinfection (SODIS) of Escherichia coli, Enterococcus spp., and MS2 coliphage: effects of additives and alternative container materials. Water Res 46(6), 1745-1754. Focazio, M.J., Kolpin, D.W., Barnes, K.K., Furlong, E.T., Meyer, M.T., Zaugg, S.D., Barber, L.B. and Thurman, M.E. (2008) A national reconnaissance for pharmaceuticals and other organic wastewater contaminants in the United States — II) Untreated drinking water sources. Science of The Total Environment 402(2–3), 201-216. FOEN (2014) Swiss Federal Office for the Environment. Fong, P.P. and Ford, A.T. (2014) The biological effects of antidepressants on the molluscs and crustaceans: a review. Aquatic Toxicology 151, 4-13. Fong, P.P. and Molnar, N. (2013) Antidepressants cause foot detachment from substrate in five species of marine snail. Marine environmental research 84, 24-30. Fu, H., Liu, H., Mao, J., Chu, W., Li, Q., Alvarez, P.J.J., Qu, X. and Zhu, D. (2016) Photochemistry of Dissolved Black Carbon Released from Biochar: Reactive Oxygen Species Generation and Phototransformation. Environmental Science & Technology 50(3), 1218-1226. Gaensly, F., Picheth, G., Brand, D. and Bonfim, T. (2014) The uptake of different iron salts by the yeast Saccharomyces cerevisiae. Brazilian Journal of Microbiology 45(2), 491-494. Gao, H. and Zepp, R.G. (1998) Factors Influencing Photoreactions of Dissolved Organic Matter in a Coastal River of the Southeastern United States. Environmental Science & Technology 32(19), 2940-2946. García-Fernández, I., Polo-López, M., Oller, I. and Fernández-Ibáñez, P. (2012) Bacteria and fungi inactivation using Fe 3+/sunlight, H 2 O 2/sunlight and near neutral photo-Fenton: A comparative study. Applied Catalysis B: Environmental 121, 20-29. García-Galán, M.J., Anfruns, A., Gonzalez-Olmos, R., Rodríguez-Mozaz, S. and Comas, J. (2016) UV/H 2 O 2 degradation of the antidepressants venlafaxine and O-desmethylvenlafaxine: Elucidation of their transformation pathway and environmental fate. J Hazard Mater 311, 70-80. Gernjak, W., Fuerhacker, M., Fernández-Ibañez, P., Blanco, J. and Malato, S. (2006) Solar photo-Fenton treatment—process parameters and process control. Applied Catalysis B: Environmental 64(1), 121-130. Gerrity, D., Gamage, S., Holady, J.C., Mawhinney, D.B., Quiñones, O., Trenholm, R.A. and Snyder, S.A. (2011) Pilot-scale evaluation of ozone and biological activated carbon for trace organic contaminant mitigation and disinfection. Water Res 45(5), 2155-2165. Giannakis, S., Darakas, E., Escalas-Cañellas, A. and Pulgarin, C. (2014a) The antagonistic and synergistic effects of temperature during solar disinfection of synthetic secondary effluent. Journal of Photochemistry and Photobiology A: Chemistry 280(0), 14-26. Giannakis, S., Darakas, E., Escalas-Cañellas, A. and Pulgarin, C. (2014b) Elucidating bacterial regrowth: Effect of disinfection conditions in dark storage of solar treated secondary effluent. Journal of Photochemistry and Photobiology A: Chemistry 290(0), 43-53.

209

Giannakis, S., Darakas, E., Escalas-Cañellas, A. and Pulgarin, C. (2015a) Environmental considerations on solar disinfection of wastewater and the subsequent bacterial (re) growth. Photochemical & Photobiological Sciences. Giannakis, S., Darakas, E., Escalas-Cañellas, A. and Pulgarin, C. (2015b) Solar disinfection modeling and post-irradiation response of Escherichia coli in wastewater. Chemical Engineering Journal 281, 588-598. Giannakis, S., Gamarra Vives, F.A., Grandjean, D., Magnet, A., De Alencastro, L.F. and Pulgarin, C. (2015c) Effect of advanced oxidation processes on the micropollutants and the effluent organic matter contained in municipal wastewater previously treated by three different secondary methods. Water Res 84, 295-306. Giannakis, S., López, M.I.P., Spuhler, D., Pérez, J.A.S., Ibáñez, P.F. and Pulgarin, C. (2016a) Solar disinfection is an augmentable, in situ-generated photo-Fenton reaction—Part 2: A review of the applications for drinking water and wastewater disinfection. Applied Catalysis B: Environmental 198, 431-446. Giannakis, S., Merino Gamo, A.I., Darakas, E., Escalas-Cañellas, A. and Pulgarin, C. (2014c) Monitoring the post-irradiation E. coli survival patterns in environmental water matrices: Implications in handling solar disinfected wastewater. Chemical Engineering Journal 253(0), 366-376. Giannakis, S., Papoutsakis, S., Darakas, E., Escalas-Cañellas, A., Pétrier, C. and Pulgarin, C. (2015d) Ultrasound enhancement of near-neutral photo-Fenton for effective E. coli inactivation in wastewater. Ultrason Sonochem 22, 515-526. Giannakis, S., Polo López, M.I., Spuhler, D., Sánchez Pérez, J.A., Fernández Ibáñez, P. and Pulgarin, C. (2016b) Solar disinfection is an augmentable, in situ-generated photo-Fenton reaction—Part 1: A review of the mechanisms and the fundamental aspects of the process. Applied Catalysis B: Environmental 199, 199-223. Giannakis, S., Ruales-Lonfat, C., Rtimi, S., Thabet, S., Cotton, P. and Pulgarin, C. (2016c) Castles fall from inside: Evidence for dominant internal photo-catalytic mechanisms during treatment of Saccharomyces cerevisiae by photo-Fenton at near-neutral pH. Applied Catalysis B: Environmental 185, 150-162. Giannakis, S., Voumard, M., Grandjean, D., Magnet, A., De Alencastro, L.F. and Pulgarin, C. (2016d) Micropollutant degradation, bacterial inactivation and regrowth risk in wastewater effluents: Influence of the secondary (pre)treatment on the efficiency of Advanced Oxidation Processes. Water Res 102, 505-515. Gibson, K.E. (2014) Viral pathogens in water: occurrence, public health impact, and available control strategies. Current opinion in virology 4, 50-57. Glass, R.I., Parashar, U.D. and Estes, M.K. (2009) Norovirus gastroenteritis. New England Journal of Medicine 361(18), 1776-1785. Göbel, A., McArdell, C.S., Joss, A., Siegrist, H. and Giger, W. (2007) Fate of sulfonamides, macrolides, and trimethoprim in different wastewater treatment technologies. Science of The Total Environment 372(2), 361-371. Gomes, A., Asad, L., Felzenszwalb, I., Leitão, A., Silva, A., Guillobel, H. and Asad, N. (2004) Does UVB radiation induce SoxS gene expression in Escherichia coli cells? Radiation and environmental biophysics 43(3), 219-222. Grant, C.M., Quinn, K.A. and Dawes, I.W. (1999) Differential protein S-thiolation of glyceraldehyde-3-phosphate dehydrogenase isoenzymes influences sensitivity to oxidative stress. Molecular and cellular biology 19(4), 2650-2656. Gregorio, V. and Chèvre, N. (2014) Assessing the risks posed by mixtures of chemicals in freshwater environments: case study of Lake Geneva, Switzerland. Wiley Interdisciplinary Reviews: Water 1(3), 229-247. Gros, M., Petrović, M. and Barceló, D. (2007) Wastewater treatment plants as a pathway for aquatic contamination by pharmaceuticals in the Ebro river basin (Northeast Spain). Environmental Toxicology and Chemistry 26(8), 1553-1562. Guo, H.-G., Gao, N.-Y., Chu, W.-H., Li, L., Zhang, Y.-J., Gu, J.-S. and Gu, Y.-L. (2013) Photochemical degradation of ciprofloxacin in UV and UV/H2O2 process: kinetics, parameters, and products. Environmental Science and Pollution Research 20(5), 3202-3213. H. Jones, O., Voulvoulis, N. and Lester, J. (2005) Human pharmaceuticals in wastewater treatment processes. Critical Reviews in Environmental Science and Technology 35(4), 401-427.

210

Haiß, A. and Kümmerer, K. (2006) Biodegradability of the X-ray contrast compound diatrizoic acid, identification of aerobic degradation products and effects against sewage sludge micro-organisms. Chemosphere 62(2), 294-302. Hakala, J.A., Fimmen, R.L., Chin, Y.-P., Agrawal, S.G. and Ward, C.P. (2009) Assessment of the geochemical reactivity of Fe-DOM complexes in wetland sediment pore waters using a nitroaromatic probe compound. Geochimica et Cosmochimica Acta 73(5), 1382-1393. Halling-Sørensen, B., Nors Nielsen, S., Lanzky, P.F., Ingerslev, F., Holten Lützhøft, H.C. and Jørgensen, S.E. (1998) Occurrence, fate and effects of pharmaceutical substances in the environment- A review. Chemosphere 36(2), 357-393. Han, S.K., Hwang, T.-M., Yoon, Y. and Kang, J.-W. (2011) Evidence of singlet oxygen and hydroxyl radical formation in aqueous goethite suspension using spin-trapping electron paramagnetic resonance (EPR). Chemosphere 84(8), 1095-1101. Hartman, P. and Eisenstark, A. (1978) Synergistic killing of Escherichia coli by near-UV radiation and hydrogen peroxide: distinction between recA-repairable and recA-nonrepairable damage. Journal of bacteriology 133(2), 769-774. Hartman, P. and Eisenstark, A. (1980) Killing of Escherichia coli K-12 by near-ultraviolet radiation in the presence of hydrogen peroxide: role of double-strand DNA breaks in absence of recombinational repair. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis 72(1), 31-42. Harvey, A.T., Rudolph, R.L. and Preskorn, S.H. (2000) Evidence of the dual mechanisms of action of venlafaxine. Archives of General Psychiatry 57(5), 503-509. Hawkshead III, J.J. (2008) Hospital wastewater containing pharmaceutically active compounds and drug-resistant organisms: a source of environmental toxicity and increased antibiotic resistance. Journal of Residuals Science & Technology 5(2), 51-60. Helms, J.R., Stubbins, A., Ritchie, J.D., Minor, E.C., Kieber, D.J. and Mopper, K. (2008) Absorption spectral slopes and slope ratios as indicators of molecular weight, source, and photobleaching of chromophoric dissolved organic matter. Limnology and Oceanography 53(3), 955-969. Herrera, F., Pulgarin, C., Nadtochenko, V. and Kiwi, J. (1998) Accelerated photo-oxidation of concentrated p-coumaric acid in homogeneous solution. Mechanistic studies, intermediates and precursors formed in the dark. Applied Catalysis B: Environmental 17(1–2), 141-156. Hiatt, C. (1964) Kinetics of the inactivation of viruses. Bacteriological reviews 28(2), 150. Hijnen, W., Beerendonk, E. and Medema, G.J. (2006) Inactivation credit of UV radiation for viruses, bacteria and protozoan (oo) cysts in water: a review. Water Res 40(1), 3-22. Hockberger, P.E., Skimina, T.A., Centonze, V.E., Lavin, C., Chu, S., Dadras, S., Reddy, J.K. and White, J.G. (1999) Activation of flavin-containing oxidases underlies light-induced production of H2O2 in mammalian cells. Proceedings of the National Academy of Sciences 96(11), 6255-6260. Hoebe, C.J., Vennema, H., de Roda Husman, A.M. and van Duynhoven, Y.T. (2004) Norovirus outbreak among primary schoolchildren who had played in a recreational water fountain. Journal of Infectious Diseases 189(4), 699-705. Ikehata, K., Jodeiri Naghashkar, N. and Gamal El-Din, M. (2006) Degradation of Aqueous Pharmaceuticals by Ozonation and Advanced Oxidation Processes: A Review. Ozone: Science & Engineering 28(6), 353-414. Imlay, J.A. (2003) Pathways of oxidative damage. Annual Reviews in Microbiology 57(1), 395-418. Imlay, J.A. (2008) Cellular defenses against superoxide and hydrogen peroxide. Annual review of biochemistry 77, 755. Jarusutthirak, C. and Amy, G. (2007) Understanding soluble microbial products (SMP) as a component of effluent organic matter (EfOM). Water Res 41(12), 2787-2793. Jeong, J., Jung, J., Cooper, W.J. and Song, W. (2010) Degradation mechanisms and kinetic studies for the treatment of X-ray contrast media compounds by advanced oxidation/reduction processes. Water Res 44(15), 4391-4398. Jimenez, B. and Asano, T. (2008) Water reclamation and reuse around the world. Water reuse: an international survey of current practice, issues and needs, 3-26. Jolis, D., Lam, C. and Pitt, P. (2001) Particle effects on ultraviolet disinfection of coliform bacteria in recycled water. Water environment research 73(2), 233-236.

211

Jolivet, J.P., Chanéac, C. and Tronc, E. (2004) Iron oxide chemistry. From molecular clusters to extended solid networks. Chemical Communications 10(5), 481-487. Jones, O., Voulvoulis, N. and Lester, J. (2006) Partitioning behavior of five pharmaceutical compounds to activated sludge and river sediment. Archives of environmental contamination and toxicology 50(3), 297-305. Jović, M., Manojlović, D., Stanković, D., Dojčinović, B., Obradović, B., Gašić, U. and Roglić, G. (2013) Degradation of triketone herbicides, mesotrione and sulcotrione, using advanced oxidation processes. J Hazard Mater 260, 1092-1099. Junghans, M., Backhaus, T., Faust, M., Scholze, M. and Grimme, L.H. (2006) Application and validation of approaches for the predictive hazard assessment of realistic pesticide mixtures. Aquatic Toxicology 76(2), 93-110. Kamolsiripichaiporn, S., Subharat, S., Udon, R., Thongtha, P. and Nuanualsuwan, S. (2007) Thermal inactivation of foot-and-mouth disease viruses in suspension. Applied and environmental microbiology 73(22), 7177-7184. Karaolia, P., Michael, I., García-Fernández, I., Agüera, A., Malato, S., Fernández-Ibáñez, P. and Fatta-Kassinos, D. (2014) Reduction of clarithromycin and sulfamethoxazole-resistant Enterococcus by pilot-scale solar-driven Fenton oxidation. Science of The Total Environment 468, 19-27. Kasprzyk-Hordern, B., Dinsdale, R.M. and Guwy, A.J. (2009) The removal of pharmaceuticals, personal care products, endocrine disruptors and illicit drugs during wastewater treatment and its impact on the quality of receiving waters. Water Res 43(2), 363-380. Katsumata, H., Kaneco, S., Suzuki, T., Ohta, K. and Yobiko, Y. (2006) Photo-Fenton degradation of alachlor in the presence of citrate solution. Journal of Photochemistry and Photobiology A: Chemistry 180(1–2), 38-45. Keenan, C. (2001) The effect of additional hydrogen peroxide on solar water disinfection. Massachusetts Institute of Technology, Cambridge MA, USA. Keyer, K. and Imlay, J.A. (1996) Superoxide accelerates DNA damage by elevating free-iron levels. Proceedings of the National Academy of Sciences 93(24), 13635-13640. Khaengraeng, R. and Reed, R. (2005) Oxygen and photoinactivation of Escherichia coli in UVA and sunlight. J Appl Microbiol 99(1), 39-50. Khansuwan, U. and Kotyk, A. (2000) Effects of the Fenton reagent on transport in yeast. Folia microbiologica 45(6), 515-520. Kilunga, P.I., Kayembe, J.M., Laffite, A., Thevenon, F., Devarajan, N., Mulaji, C.K., Mubedi, J.I., Yav, Z.G., Otamonga, J.-P. and Mpiana, P.T. (2016) The impact of hospital and urban wastewaters on the bacteriological contamination of the water resources in Kinshasa, Democratic Republic of Congo. Journal of Environmental Science and Health, Part A, 1-9. Kim, I., Yamashita, N. and Tanaka, H. (2009a) Performance of UV and UV/H2O2 processes for the removal of pharmaceuticals detected in secondary effluent of a sewage treatment plant in Japan. J Hazard Mater 166(2–3), 1134-1140. Kim, I., Yamashita, N. and Tanaka, H. (2009b) Photodegradation of pharmaceuticals and personal care products during UV and UV/H2O2 treatments. Chemosphere 77(4), 518-525. Kim, J.Y., Lee, C., Sedlak, D.L., Yoon, J. and Nelson, K.L. (2010) Inactivation of MS2 coliphage by Fenton's reagent. Water Res 44(8), 2647-2653. Klamerth, N., Malato, S., Agüera, A., Fernández-Alba, A. and Mailhot, G. (2012) Treatment of Municipal Wastewater Treatment Plant Effluents with Modified Photo-Fenton As a Tertiary Treatment for the Degradation of Micro Pollutants and Disinfection. Environmental Science & Technology 46(5), 2885-2892. Klamerth, N., Rizzo, L., Malato, S., Maldonado, M.I., Agüera, A. and Fernández-Alba, A.R. (2010) Degradation of fifteen emerging contaminants at μg L−1 initial concentrations by mild solar photo-Fenton in MWTP effluents. Water Res 44(2), 545-554. Knowles, R.L. and Eisenstark, A. (1994) Near-ultraviolet mutagenesis in superoxide dismutase-deficient strains of Escherichia coli. Environmental Health Perspectives 102(1), 88. Kocha, T., Yamaguchi, M., Ohtaki, H., Fukuda, T. and Aoyagi, T. (1997) Hydrogen peroxide-mediated degradation of protein: different oxidation modes of copper-and iron-dependent hydroxyl radicals on the

212

degradation of albumin. Biochimica et Biophysica Acta (BBA)-Protein Structure and Molecular Enzymology 1337(2), 319-326. Kohn, T., Grandbois, M., McNeill, K. and Nelson, K.L. (2007) Association with natural organic matter enhances the sunlight-mediated inactivation of MS2 coliphage by singlet oxygen. Environmental Science & Technology 41(13), 4626-4632. Kohn, T. and Nelson, K.L. (2007) Sunlight-mediated inactivation of MS2 coliphage via exogenous singlet oxygen produced by sensitizers in natural waters. Environmental Science & Technology 41(1), 192-197. Kolpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M., Zaugg, S.D., Barber, L.B. and Buxton, H.T. (2002) Pharmaceuticals, Hormones, and Other Organic Wastewater Contaminants in U.S. Streams, 1999−2000: A National Reconnaissance. Environmental Science & Technology 36(6), 1202-1211. Kosma, C.I., Lambropoulou, D.A. and Albanis, T.A. (2010) Occurrence and removal of PPCPs in municipal and hospital wastewaters in Greece. J Hazard Mater 179(1-3), 804-817. Köster, W. (2001) ABC transporter-mediated uptake of iron, siderophores, heme and vitamin B 12. Research in microbiology 152(3), 291-301. Krasner, S.W., Westerhoff, P., Chen, B., Rittmann, B.E., Nam, S.-N. and Amy, G. (2009) Impact of wastewater treatment processes on organic carbon, organic nitrogen, and DBP precursors in effluent organic matter. Environmental Science & Technology 43(8), 2911-2918. Krisko, A. and Radman, M. (2010) Protein damage and death by radiation in Escherichia coli and Deinococcus radiodurans. Proceedings of the National Academy of Sciences 107(32), 14373-14377. Kümmerer, K. (2001) Drugs in the environment: emission of drugs, diagnostic aids and disinfectants into wastewater by hospitals in relation to other sources – a review. Chemosphere 45(6–7), 957-969. Kümmerer, K. (2004) Resistance in the environment. Journal of Antimicrobial Chemotherapy 54(2), 311-320. Kümmerer, K. (2011) Emerging Contaminants versus Micro-pollutants. CLEAN – Soil, Air, Water 39(10), 889-890. Kuzmanovic, D.A., Elashvili, I., Wick, C., O'Connell, C. and Krueger, S. (2006) The MS2 coat protein shell is likely assembled under tension: a novel role for the MS2 bacteriophage A protein as revealed by small-angle neutron scattering. Journal of molecular biology 355(5), 1095-1111. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-685. Lajeunesse, A., Smyth, S.A., Barclay, K., Sauvé, S. and Gagnon, C. (2012) Distribution of antidepressant residues in wastewater and biosolids following different treatment processes by municipal wastewater treatment plants in Canada. Water Res 46(17), 5600-5612. Lambropoulou, D., Evgenidou, E., Saliverou, V., Kosma, C. and Konstantinou, I. (2016) Degradation of venlafaxine using TiO 2/UV process: Kinetic studies, RSM optimization, identification of transformation products and toxicity evaluation. J Hazard Mater. Lee, E., Glover, C.M. and Rosario-Ortiz, F.L. (2013) Photochemical Formation of Hydroxyl Radical from Effluent Organic Matter: Role of Composition. Environmental Science & Technology 47(21), 12073-12080. Lee, Y., Kovalova, L., McArdell, C.S. and Von Gunten, U. (2014) Prediction of micropollutant elimination during ozonation of a hospital wastewater effluent. Water Res 64, 134-148. Legrini, O., Oliveros, E. and Braun, A. (1993) Photochemical processes for water treatment. Chemical Reviews 93(2), 671-698. Leland, J.K. and Bard, A.J. (1987) Photochemistry of colloidal semiconducting iron oxide polymorphs. Journal of Physical Chemistry 91(19), 5076-5083. Lester, Y., Ferrer, I., Thurman, E.M. and Linden, K.G. (2014) Demonstrating sucralose as a monitor of full-scale UV/AOP treatment of trace organic compounds. J Hazard Mater 280, 104-110. Lesuisse, E., Blaiseau, P.-L., Dancis, A. and Camadro, J.-M. (2001) Siderophore uptake and use by the yeast Saccharomyces cerevisiae. Microbiology 147(2), 289-298. Li, C.S., Chia, W.C. and Chen, P.S. (2007) Fluorochrome and flow cytometry to monitor microorganisms in treated hospital wastewater. Journal of Environmental Science and Health, Part A 42(2), 195-203. Li, X., Wang, Y., Zhao, J., Wang, H., Wang, B., Huang, J., Deng, S. and Yu, G. (2015) Electro-peroxone treatment of the antidepressant venlafaxine: Operational parameters and mechanism. J Hazard Mater 300, 298-306.

213

Liberti, L., Notarnicola, M. and Petruzzelli, D. (2003) Advanced treatment for municipal wastewater reuse in agriculture. UV disinfection: parasite removal and by-product formation. Desalination 152(1–3), 315-324. Lienert, J., Burki, T. and Escher, B. (2007) Reducing micropollutants with source control: substance flow analysis of 212 pharmaceuticals in faeces and urine. Water Science & Technology 56(5), 87-96. Lienert, J., Koller, M., Konrad, J., McArdell, C.S. and Schuwirth, N. (2011) Multiple-criteria decision analysis reveals high stakeholder preference to remove pharmaceuticals from hospital wastewater. Environmental Science & Technology 45(9), 3848-3857. Lin, J. and Ganesh, A. (2013) Water quality indicators: bacteria, coliphages, enteric viruses. International journal of environmental health research 23(6), 484-506. Lindberg, C. and Horneck, G. (1991) Action spectra for survival and spore photoproduct formation of Bacillus subtilis irradiated with short-wavelength (200–300 nm) UV at atmospheric pressure and in vacuo. Journal of Photochemistry and Photobiology B: Biology 11(1), 69-80. Luo, Y., Guo, W., Ngo, H.H., Nghiem, L.D., Hai, F.I., Zhang, J., Liang, S. and Wang, X.C. (2014) A review on the occurrence of micropollutants in the aquatic environment and their fate and removal during wastewater treatment. Sci Total Environ 473-474, 619-641. Mack, J. and Bolton, J.R. (1999) Photochemistry of nitrite and nitrate in aqueous solution: a review. Journal of Photochemistry and Photobiology A: Chemistry 128(1–3), 1-13. Mackuľak, T., Birošová, L., Bodík, I., Grabic, R., Takáčová, A., Smolinská, M., Hanusová, A., Híveš, J. and Gál, M. (2016) Zerovalent iron and iron (VI): Effective means for the removal of psychoactive pharmaceuticals and illicit drugs from wastewaters. Science of The Total Environment 539, 420-426. Mackuľak, T., Mosný, M., Grabic, R., Golovko, O., Koba, O. and Birošová, L. (2015) Fenton-like reaction: A possible way to efficiently remove illicit drugs and pharmaceuticals from wastewater. Environmental Toxicology and Pharmacology 39(2), 483-488. Mahnik, S., Lenz, K., Weissenbacher, N., Mader, R. and Fuerhacker, M. (2007) Fate of 5-fluorouracil, doxorubicin, epirubicin, and daunorubicin in hospital wastewater and their elimination by activated sludge and treatment in a membrane-bio-reactor system. Chemosphere 66(1), 30-37. Malato, S., Fernández-Ibáñez, P., Maldonado, M.I., Blanco, J. and Gernjak, W. (2009) Decontamination and disinfection of water by solar photocatalysis: Recent overview and trends. Catalysis Today 147(1), 1-59. Mantzavinos, D. and Kalogerakis, N. (2005) Treatment of olive mill effluents: Part I. Organic matter degradation by chemical and biological processes—an overview. Environment International 31(2), 289-295. Margot, J., Kienle, C., Magnet, A., Weil, M., Rossi, L., De Alencastro, L.F., Abegglen, C., Thonney, D., Chèvre, N. and Schärer, M. (2013) Treatment of micropollutants in municipal wastewater: Ozone or powdered activated carbon? Science of The Total Environment 461, 480-498. Margot, J., Magnet, A., Thonney, D., Chèvre, N., Alencastro, D., Felippe, L. and Rossi, L. (2011) Traitement des micropolluants dans les eaux usées-Rapport final sur les essais pilotes à la STEP de Vidy (Lausanne), Ville de Lausanne. Marugán, J., van Grieken, R., Sordo, C. and Cruz, C. (2008) Kinetics of the photocatalytic disinfection of Escherichia coli suspensions. Applied Catalysis B: Environmental 82(1–2), 27-36. Mattle, M.J. and Kohn, T. (2012) Inactivation and tailing during UV254 disinfection of viruses: contributions of viral aggregation, light shielding within viral aggregates, and recombination. Environmental Science & Technology 46(18), 10022-10030. Mayer, B.K., Ryu, H. and Abbaszadegan, M. (2008) Treatability of US Environmental Protection Agency contaminant candidate list viruses: removal of coxsackievirus and echovirus using enhanced coagulation. Environmental Science & Technology 42(18), 6890-6896. McArdell, C.S., Kovalova, L., Siegrist, H., Kienle, C., Moser, R. and Schwartz, T. (2011) Input and elimination of pharmaceuticals and disinfectants from hospital wastewater. Eawag, Duebendorf Switzerland, 95. McGuigan, K.G., Conroy, R.M., Mosler, H.J., du Preez, M., Ubomba-Jaswa, E. and Fernandez-Ibanez, P. (2012) Solar water disinfection (SODIS): a review from bench-top to roof-top. J Hazard Mater 235-236, 29-46.

214

Melchior, M., Caspi, A., Milne, B.J., Danese, A., Poulton, R. and Moffitt, T.E. (2007) Work stress precipitates depression and anxiety in young, working women and men. Psychological medicine 37(08), 1119-1129. Metcalfe, C.D., Chu, S., Judt, C., Li, H., Oakes, K.D., Servos, M.R. and Andrews, D.M. (2010) Antidepressants and their metabolites in municipal wastewater, and downstream exposure in an urban watershed. Environmental Toxicology and Chemistry 29(1), 79-89. Michael-Kordatou, I., Michael, C., Duan, X., He, X., Dionysiou, D., Mills, M. and Fatta-Kassinos, D. (2015) Dissolved effluent organic matter: Characteristics and potential implications in wastewater treatment and reuse applications. Water Res 77, 213-248. Michael, I., Frontistis, Z. and Fatta-Kassinos, D. (2013) Removal of pharmaceuticals from environmentally relevant matrices by advanced oxidation processes (AOPs). Comprehensive Analytical Chemistry 62, 345-407. Michael, I., Hapeshi, E., Michael, C., Varela, A.R., Kyriakou, S., Manaia, C.M. and Fatta-Kassinos, D. (2012) Solar photo-Fenton process on the abatement of antibiotics at a pilot scale: Degradation kinetics, ecotoxicity and phytotoxicity assessment and removal of antibiotic resistant enterococci. Water Res 46(17), 5621-5634. Millero, F.J. and Sotolongo, S. (1989) The oxidation of Fe (II) with H 2 O 2 in seawater. Geochimica et Cosmochimica Acta 53(8), 1867-1873. Moncayo-Lasso, A., Rincon, A.G., Pulgarin, C. and Benitez, N. (2012) Significant decrease of THMs generated during chlorination of river water by previous photo-Fenton treatment at near neutral pH. Journal of Photochemistry and Photobiology A: Chemistry 229(1), 46-52. Morgan, B. and Lahav, O. (2007) The effect of pH on the kinetics of spontaneous Fe (II) oxidation by O 2 in aqueous solution–basic principles and a simple heuristic description. Chemosphere 68(11), 2080-2084. Mostafa, S. and Rosario-Ortiz, F.L. (2013) Singlet Oxygen Formation from Wastewater Organic Matter. Environmental Science & Technology 47(15), 8179-8186. Muthukumaran, S., Nguyen, D.A. and Baskaran, K. (2011) Performance evaluation of different ultrafiltration membranes for the reclamation and reuse of secondary effluent. Desalination 279(1), 383-389. Nasibi, F. and Kalantari, K. (2005) The effects of UV-A, UV-B and UV-C on protein and ascorbate content, lipid peroxidation and biosynthesis of screening compounds in Brassica napus. Iranian J Sci Technol 29, 39-48. Nataro, J.P. and Kaper, J.B. (1998) Diarrheagenic escherichia coli. Clinical microbiology reviews 11(1), 142-201. Ndounla, J., Kenfack, S., Wéthé, J. and Pulgarin, C. (2014) Relevant impact of irradiance (vs. dose) and evolution of pH and mineral nitrogen compounds during natural water disinfection by photo-Fenton in a solar CPC reactor. Applied Catalysis B: Environmental 148–149(0), 144-153. Ndounla, J., Spuhler, D., Kenfack, S., Wéthé, J. and Pulgarin, C. (2013) Inactivation by solar photo-Fenton in pet bottles of wild enteric bacteria of natural well water: Absence of re-growth after one week of subsequent storage. Applied Catalysis B: Environmental 129, 309-317. Nelson, C.D., Minkkinen, E., Bergkvist, M., Hoelzer, K., Fisher, M., Bothner, B. and Parrish, C.R. (2008) Detecting small changes and additional peptides in the canine parvovirus capsid structure. Journal of virology 82(21), 10397-10407. Neyens, E. and Baeyens, J. (2003) A review of classic Fenton’s peroxidation as an advanced oxidation technique. J Hazard Mater 98(1), 33-50. Ng, T.-W., Chow, A.T. and Wong, P.-K. (2014) Dual roles of dissolved organic matter in photo-irradiated Fe(III)-contained waters. Journal of Photochemistry and Photobiology A: Chemistry 290(0), 116-124. Ng, T.W., An, T., Li, G., Ho, W.K., Yip, H.Y., Zhao, H. and Wong, P.K. (2015) The role and synergistic effect of the light irradiation and H 2 O 2 in photocatalytic inactivation of Escherichia coli. Journal of Photochemistry and Photobiology B: Biology 149, 164-171. Nie, Y., Qiang, Z., Zhang, H. and Ben, W. (2012) Fate and seasonal variation of endocrine-disrupting chemicals in a sewage treatment plant with A/A/O process. Separation and Purification Technology 84, 9-15. Nieto-Juarez, J.I. and Kohn, T. (2013) Virus removal and inactivation by iron (hydr) oxide-mediated Fenton-like processes under sunlight and in the dark. Photochemical & Photobiological Sciences 12(9), 1596-1605.

215

Nieto-Juarez, J.I., Pierzchła, K., Sienkiewicz, A. and Kohn, T. (2010) Inactivation of MS2 coliphage in Fenton and Fenton-like systems: role of transition metals, hydrogen peroxide and sunlight. Environmental Science & Technology 44(9), 3351-3356. Okoh, A.I., Sibanda, T. and Gusha, S.S. (2010) Inadequately treated wastewater as a source of human enteric viruses in the environment. International journal of environmental research and public health 7(6), 2620-2637. Oller, I., Malato, S. and Sanchez-Perez, J.A. (2011) Combination of Advanced Oxidation Processes and biological treatments for wastewater decontamination--a review. Sci Total Environ 409(20), 4141-4166. Olsthoorn, R. and Van Duin, J. (1996) Evolutionary reconstruction of a hairpin deleted from the genome of an RNA virus. Proceedings of the National Academy of Sciences 93(22), 12256-12261. Oppenländer, T. (2003) Photochemical Purification of Water and Air: Advanced Oxidation Processes (AOPs)-Principles, Reaction Mechanisms, Reactor Concepts, John Wiley & Sons. Ort, C., Lawrence, M.G., Reungoat, J., Eaglesham, G., Carter, S. and Keller, J. (2010) Determining the fraction of pharmaceutical residues in wastewater originating from a hospital. Water Res 44(2), 605-615. Ortega-Gómez, E., Ballesteros Martín, M.M., Carratalà, A., Fernández Ibañez, P., Sánchez Pérez, J.A. and Pulgarín, C. (2015) Principal parameters affecting virus inactivation by the solar photo-Fenton process at neutral pH and μM concentrations of H2O2 and Fe2+/3+. Applied Catalysis B: Environmental 174-175, 395-402. Ortega-Gómez, E., Ballesteros Martín, M.M., Esteban García, B., Sánchez Pérez, J.A. and Fernández Ibáñez, P. (2014) Solar photo-Fenton for water disinfection: An investigation of the competitive role of model organic matter for oxidative species. Applied Catalysis B: Environmental 148–149(0), 484-489. Ortega-Gómez, E., Esteban García, B., Ballesteros Martín, M.M., Fernández Ibáñez, P. and Sánchez Pérez, J.A. (2013) Inactivation of Enterococcus faecalis in simulated wastewater treatment plant effluent by solar photo-Fenton at initial neutral pH. Catalysis Today 209, 195-200. Ortega-Gomez, E., Fernandez-Ibanez, P., Ballesteros Martin, M.M., Polo-Lopez, M.I., Esteban Garcia, B. and Sanchez Perez, J.A. (2012) Water disinfection using photo-Fenton: Effect of temperature on Enterococcus faecalis survival. Water Res 46(18), 6154-6162. Oyane, I., Takeda, T., Oda, Y., Sakata, T., Furuta, M., Okitsu, K., Maeda, Y. and Nishimura, R. (2009) Comparison between the effects of ultrasound and γ-rays on the inactivation of Saccharomyces cerevisiae: Analyses of cell membrane permeability and DNA or RNA synthesis by flow cytometry. Ultrason Sonochem 16(4), 532-536. Pal, A., Gin, K.Y.-H., Lin, A.Y.-C. and Reinhard, M. (2010) Impacts of emerging organic contaminants on freshwater resources: Review of recent occurrences, sources, fate and effects. Science of The Total Environment 408(24), 6062-6069. Papoutsakis, S., Afshari, Z., Malato, S. and Pulgarin, C. (2015a) Elimination of the iodinated contrast agent iohexol in water, wastewater and urine matrices by application of photo-Fenton and ultrasound advanced oxidation processes. Journal of Environmental Chemical Engineering 3(3), 2002-2009. Papoutsakis, S., Miralles-Cuevas, S., Oller, I., Garcia Sanchez, J.L., Pulgarin, C. and Malato, S. (2015b) Microcontaminant degradation in municipal wastewater treatment plant secondary effluent by EDDS assisted photo-Fenton at near-neutral pH: An experimental design approach. Catalysis Today 252, 61-69. Park, M., Eom, J., Fong, J. and Lim, Y. (2015) New record and enzyme activity of four species in Penicillium section Citrina from marine environments in Korea. Journal of Microbiology 53(4), 219-225. Parkinson, A., Barry, M.J., Roddick, F.A. and Hobday, M.D. (2001) Preliminary toxicity assessment of water after treatment with uv-irradiation and UVC/H2O2. Water Res 35(15), 3656-3664. Pauwels, B. and Verstraete, W. (2006) The treatment of hospital wastewater: an appraisal. J Water Health 4, 405-416. Pereira, V.J., Linden, K.G. and Weinberg, H.S. (2007) Evaluation of UV irradiation for photolytic and oxidative degradation of pharmaceutical compounds in water. Water Res 41(19), 4413-4423. Pérez, S. and Barceló, D. (2007) Fate and occurrence of X-ray contrast media in the environment. Anal Bioanal Chem 387(4), 1235-1246. Petrie, B., Barden, R. and Kasprzyk-Hordern, B. (2015) A review on emerging contaminants in wastewaters and the environment: Current knowledge, understudied areas and recommendations for future monitoring. Water Res 72, 3-27.

216

Pfeifer, G.P., You, Y.-H. and Besaratinia, A. (2005) Mutations induced by ultraviolet light. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis 571(1–2), 19-31. Pigeot-Rémy, S., Simonet, F., Atlan, D., Lazzaroni, J. and Guillard, C. (2012) Bactericidal efficiency and mode of action: A comparative study of photochemistry and photocatalysis. Water Res 46(10), 3208-3218. Pignatello, J.J., Oliveros, E. and MacKay, A. (2006) Advanced Oxidation Processes for Organic Contaminant Destruction Based on the Fenton Reaction and Related Chemistry. Critical Reviews in Environmental Science and Technology 36(1), 1-84. Pills-Project (2012). Polo-López, M.I., Castro-Alférez, M., Oller, I. and Fernández-Ibáñez, P. (2014) Assessment of solar photo-Fenton, photocatalysis, and H2O2 for removal of phytopathogen fungi spores in synthetic and real effluents of urban wastewater. Chemical Engineering Journal 257(0), 122-130. Polo-López, M.I., Oller, I. and Fernández-Ibáñez, P. (2013) Benefits of photo-Fenton at low concentrations for solar disinfection of distilled water. A case study: Phytophthora capsici. Catalysis Today 209(0), 181-187. Polo-López, M., Fernández-Ibáñez, P., García-Fernández, I., Oller, I., Salgado-Tránsito, I. and Sichel, C. (2010) Resistance of Fusarium sp spores to solar TiO2 photocatalysis: influence of spore type and water (scaling-up results). Journal of chemical technology and biotechnology 85(8), 1038-1048. Pourzand, C., Watkin, R.D., Brown, J.E. and Tyrrell, R.M. (1999) Ultraviolet A radiation induces immediate release of iron in human primary skin fibroblasts: the role of ferritin. Proceedings of the National Academy of Sciences 96(12), 6751-6756. Poyatos, J., Muñio, M., Almecija, M., Torres, J., Hontoria, E. and Osorio, F. (2010) Advanced oxidation processes for wastewater treatment: state of the art. Water, Air, and Soil Pollution 205(1-4), 187-204. Prado, T., Silva, D.M., Guilayn, W.C., Rose, T.L., Gaspar, A.M.C. and Miagostovich, M.P. (2011) Quantification and molecular characterization of enteric viruses detected in effluents from two hospital wastewater treatment plants. Water Res 45(3), 1287-1297. Prieto-Rodríguez, L., Oller, I., Klamerth, N., Agüera, A., Rodríguez, E.M. and Malato, S. (2013) Application of solar AOPs and ozonation for elimination of micropollutants in municipal wastewater treatment plant effluents. Water Res 47(4), 1521-1528. Pulgarin, C. (2015) Fe vs. TiO<sub>2</sub> Photo-assisted Processes for Enhancing the Solar Inactivation of Bacteria in Water. CHIMIA International Journal for Chemistry 69(1), 7-9. Pulgarin, C. and Kiwi, J. (1996) Overview on photocatalytic and electrocatalytic pretreatment of industrial non-biodegradable pollutants and pesticides. CHIMIA International Journal for Chemistry 50(3), 50-55. Putschew, A., Wischnack, S. and Jekel, M. (2000) Occurrence of triiodinated X-ray contrast agents in the aquatic environment. Science of The Total Environment 255(1), 129-134. Qadir, M., Wichelns, D., Raschid-Sally, L., McCornick, P.G., Drechsel, P., Bahri, A. and Minhas, P. (2010) The challenges of wastewater irrigation in developing countries. Agricultural Water Management 97(4), 561-568. Quek, P.H. and Hu, J. (2008) Indicators for photoreactivation and dark repair studies following ultraviolet disinfection. Journal of industrial microbiology & biotechnology 35(6), 533-541. Rahn, R.O. (1997) Potassium iodide as a chemical actinometer for 254 nm radiation: use of lodate as an electron scavenger. Photochem Photobiol 66(4), 450-455. Ribeiro, A.R., Nunes, O.C., Pereira, M.F. and Silva, A.M. (2015) An overview on the advanced oxidation processes applied for the treatment of water pollutants defined in the recently launched Directive 2013/39/EU. Environment International 75, 33-51. Richardson, S.D. and Ternes, T.A. (2005) Water Analysis: Emerging Contaminants and Current Issues. Analytical Chemistry 77(12), 3807-3838. Rincón, A.-G. and Pulgarin, C. (2004) Bactericidal action of illuminated TiO2 on pure Escherichia coli and natural bacterial consortia: post-irradiation events in the dark and assessment of the effective disinfection time. Applied Catalysis B: Environmental 49(2), 99-112. Rincón, A.-G. and Pulgarin, C. (2006) Comparative evaluation of Fe3+ and TiO2 photoassisted processes in solar photocatalytic disinfection of water. Applied Catalysis B: Environmental 63(3-4), 222-231.

217

Rivas, G., Carra, I., García Sánchez, J.L., Casas López, J.L., Malato, S. and Sánchez Pérez, J.A. (2015) Modelling of the operation of raceway pond reactors for micropollutant removal by solar photo-Fenton as a function of photon absorption. Applied Catalysis B: Environmental 178, 210-217. Rizzo, L., Manaia, C., Merlin, C., Schwartz, T., Dagot, C., Ploy, M.C., Michael, I. and Fatta-Kassinos, D. (2013) Urban wastewater treatment plants as hotspots for antibiotic resistant bacteria and genes spread into the environment: A review. Science of The Total Environment 447(0), 345-360. Robertson, J.B., Davis, C.R. and Johnson, C.H. (2013) Visible light alters yeast metabolic rhythms by inhibiting respiration. Proceedings of the National Academy of Sciences 110(52), 21130-21135. Rodney, S.I., Teed, R.S. and Moore, D.R.J. (2013) Estimating the Toxicity of Pesticide Mixtures to Aquatic Organisms: A Review. Human and Ecological Risk Assessment: An International Journal 19(6), 1557-1575. Rodríguez-Chueca, J., Ormad, M.P., Mosteo, R., Sarasa, J. and Ovelleiro, J.L. (2015) Conventional and Advanced Oxidation Processes Used in Disinfection of Treated Urban Wastewater. Water environment research 87(3), 281-288. Rodriguez, R.A., Bounty, S., Beck, S., Chan, C., McGuire, C. and Linden, K.G. (2014) Photoreactivation of bacteriophages after UV disinfection: role of genome structure and impacts of UV source. Water Res 55, 143-149. Romero, O.C., Straub, A.P., Kohn, T. and Nguyen, T.H. (2011) Role of temperature and Suwannee River natural organic matter on inactivation kinetics of rotavirus and bacteriophage MS2 by solar irradiation. Environmental Science & Technology 45(24), 10385-10393. Rose, A.L. and Waite, T.D. (2002) Kinetic model for Fe (II) oxidation in seawater in the absence and presence of natural organic matter. Environmental Science & Technology 36(3), 433-444. Rossolini, G.M., Arena, F., Pecile, P. and Pollini, S. (2014) Update on the antibiotic resistance crisis. Current Opinion in Pharmacology 18, 56-60. Rúa-Gómez, P.C. and Püttmann, W. (2012) Occurrence and removal of lidocaine, tramadol, venlafaxine, and their metabolites in German wastewater treatment plants. Environmental Science and Pollution Research 19(3), 689-699. Rúa-Gómez, P.C. and Püttmann, W. (2013) Degradation of lidocaine, tramadol, venlafaxine and the metabolites O-desmethyltramadol and O-desmethylvenlafaxine in surface waters. Chemosphere 90(6), 1952-1959. Ruales-Lonfat, C., Barona, J.F., Sienkiewicz, A., Bensimon, M., Vélez-Colmenares, J., Benítez, N. and Pulgarín, C. (2015) Iron oxides semiconductors are efficients for solar water disinfection: A comparison with photo-Fenton processes at neutral pH. Applied Catalysis B: Environmental 166–167(0), 497-508. Ruales-Lonfat, C., Barona, J.F., Sienkiewicz, A., Vélez, J., Benítez, L.N. and Pulgarín, C. (2016) Bacterial inactivation with iron citrate complex: A new source of dissolved iron in solar photo-Fenton process at near-neutral and alkaline pH. Applied Catalysis B: Environmental 180, 379-390. Ruales-Lonfat, C., Benítez, N., Sienkiewicz, A. and Pulgarín, C. (2014) Deleterious effect of homogeneous and heterogeneous near-neutral photo-Fenton system on Escherichia coli. Comparison with photo-catalytic action of TiO 2 during cell envelope disruption. Applied Catalysis B: Environmental 160, 286-297. Ryan, C.C., Tan, D.T. and Arnold, W.A. (2011) Direct and indirect photolysis of sulfamethoxazole and trimethoprim in wastewater treatment plant effluent. Water Res 45(3), 1280-1286. Salgado, P., Melin, V., Contreras, D., Moreno, Y. and Mansilla, H. (2013) Fenton reaction driven by iron ligands. Journal of the Chilean Chemical Society 58, 2096-2101. Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Molecular cloning, Cold spring harbor laboratory press New York. Santoke, H., Song, W., Cooper, W.J. and Peake, B.M. (2012) Advanced oxidation treatment and photochemical fate of selected antidepressant pharmaceuticals in solutions of Suwannee River humic acid. J Hazard Mater 217, 382-390. Santos, A.L., Oliveira, V., Baptista, I., Henriques, I., Gomes, N.C., Almeida, A., Correia, A. and Cunha, Â. (2013) Wavelength dependence of biological damage induced by UV radiation on bacteria. Archives of microbiology 195(1), 63-74. Santos, L.H.M.L.M., Araújo, A.N., Fachini, A., Pena, A., Delerue-Matos, C. and Montenegro, M.C.B.S.M. (2010) Ecotoxicological aspects related to the presence of pharmaceuticals in the aquatic environment. J Hazard Mater 175(1–3), 45-95.

218

Sarria, V., Kenfack, S., Guillod, O. and Pulgarin, C. (2003) An innovative coupled solar-biological system at field pilot scale for the treatment of biorecalcitrant pollutants. Journal of Photochemistry and Photobiology A: Chemistry 159(1), 89-99. Schenk-Meuser, K., Pawlowsky, K. and Kiefer, J. (1992) Inactivation and mutation induction in Saccharomyces cerevisiae exposed to simulated sunlight: evaluation of action spectra. Journal of Photochemistry and Photobiology B: Biology 14(3), 231-245. Schenk, M., Raffellini, S., Guerrero, S., Blanco, G.A. and Alzamora, S.M. (2011) Inactivation of Escherichia coli, Listeria innocua and Saccharomyces cerevisiae by UV-C light: Study of cell injury by flow cytometry. LWT - Food Science and Technology 44(1), 191-198. Schultz, M.M., Furlong, E.T., Kolpin, D.W., Werner, S.L., Schoenfuss, H.L., Barber, L.B., Blazer, V.S., Norris, D.O. and Vajda, A.M. (2010) Antidepressant pharmaceuticals in two US effluent-impacted streams: occurrence and fate in water and sediment, and selective uptake in fish neural tissue. Environmental Science & Technology 44(6), 1918-1925. Schwartz, T., Kohnen, W., Jansen, B. and Obst, U. (2003) Detection of antibiotic-resistant bacteria and their resistance genes in wastewater, surface water, and drinking water biofilms. FEMS Microbiology Ecology 43(3), 325-335. Schwarzenbach, R.P., Escher, B.I., Fenner, K., Hofstetter, T.B., Johnson, C.A., von Gunten, U. and Wehrli, B. (2006) The Challenge of Micropollutants in Aquatic Systems. Science 313(5790), 1072-1077. Sciacca, F., Rengifo-Herrera, J.A., Wéthé, J. and Pulgarin, C. (2010) Dramatic enhancement of solar disinfection (SODIS) of wild Salmonella sp. in PET bottles by H 2 O 2 addition on natural water of Burkina Faso containing dissolved iron. Chemosphere 78(9), 1186-1191. Seitz, W., Jiang, J.-Q., Schulz, W., Weber, W.H., Maier, D. and Maier, M. (2008) Formation of oxidation by-products of the iodinated X-ray contrast medium iomeprol during ozonation. Chemosphere 70(7), 1238-1246. Sharpless, C.M., Aeschbacher, M., Page, S.E., Wenk, J., Sander, M. and McNeill, K. (2014) Photooxidation-induced changes in optical, electrochemical, and photochemical properties of humic substances. Environmental Science & Technology 48(5), 2688-2696. Sharpless, C.M. and Linden, K.G. (2003) Experimental and Model Comparisons of Low- and Medium-Pressure Hg Lamps for the Direct and H2O2 Assisted UV Photodegradation of N-Nitrosodimethylamine in Simulated Drinking Water. Environmental Science & Technology 37(9), 1933-1940. Shon, H.K., Vigneswaran, S. and Snyder, S.A. (2006) Effluent Organic Matter (EfOM) in Wastewater: Constituents, Effects, and Treatment. Critical Reviews in Environmental Science and Technology 36(4), 327-374. Sichel, C., De Cara, M., Tello, J., Blanco, J. and Fernández-Ibáñez, P. (2007a) Solar photocatalytic disinfection of agricultural pathogenic fungi: Fusarium species. Applied Catalysis B: Environmental 74(1), 152-160. Sichel, C., Fernández-Ibáñez, P., de Cara, M. and Tello, J. (2009) Lethal synergy of solar UV-radiation and H2O2 on wild Fusarium solani spores in distilled and natural well water. Water Res 43(7), 1841-1850. Sichel, C., Tello, J., de Cara, M. and Fernández-Ibáñez, P. (2007b) Effect of UV solar intensity and dose on the photocatalytic disinfection of bacteria and fungi. Catalysis Today 129(1–2), 152-160. Siegrist, H., Joss, A., Ternes, T. and Oehlmann, J. (2005) Fate of EDCs in wastewater treatment and EU perspective on EDC regulation. Proceedings of the Water Environment Federation 2005(13), 3142-3165. Sigler, K., Gille, G., Vacata, V., Stadler, N. and Höfer, M. (1998) Inactivation of the plasma membrane ATPase ofSchizosaccharomyces pombe by hydrogen peroxide and by the fenton reagent (Fe2+/H2O2): Nonradicalvs. Radical-induced oxidation. Folia microbiologica 43(4), 361-367. Silva, M., Trovó, A. and Nogueira, R. (2007) Degradation of the herbicide tebuthiuron using solar photo-Fenton process and ferric citrate complex at circumneutral pH. Journal of Photochemistry and Photobiology A: Chemistry 191(2), 187-192. Šima, J. and Makáňová, J. (1997) Photochemistry of iron (III) complexes. Coordination Chemistry Reviews 160, 161-189. Sinha, R.P. and Häder, D.-P. (2002) UV-induced DNA damage and repair: a review. Photochemical & Photobiological Sciences 1(4), 225-236.

219

Sławińska, D., Polewski, K., Rolewski, P., Pluciński, P. and Sławiński, J. (2002) Spectroscopic studies on UVC-induced photodegradation of humic acids. Development 5(2). Sousa-Lopes, A., Antunes, F., Cyrne, L. and Marinho, H. (2004) Decreased cellular permeability to H 2 O 2 protects Saccharomyces cerevisiae cells in stationary phase against oxidative stress. FEBS letters 578(1), 152-156. Sprehe, M. and Geissen, S. (2000) Verfahrensauswahl zur AOX-eliminierung im krankenhausabwasserbereich. Halogenorganische verbindungen, Herausgeber ATV-DVWK, ATV-DVWK Schriftenreihe 18, 257-268. Spuhler, D., Andrés Rengifo-Herrera, J. and Pulgarin, C. (2010) The effect of Fe2+, Fe3+, H2O2 and the photo-Fenton reagent at near neutral pH on the solar disinfection (SODIS) at low temperatures of water containing Escherichia coli K12. Applied Catalysis B: Environmental 96(1-2), 126-141. Srinivasan, C., Liba, A., Imlay, J.A., Valentine, J.S. and Gralla, E.B. (2000) Yeast Lacking Superoxide Dismutase(s) Show Elevated Levels of “Free Iron” as Measured by Whole Cell Electron Paramagnetic Resonance. Journal of Biological Chemistry 275(38), 29187-29192. Stadler, L.B., Su, L., Moline, C.J., Ernstoff, A.S., Aga, D.S. and Love, N.G. (2015) Effect of redox conditions on pharmaceutical loss during biological wastewater treatment using sequencing batch reactors. J Hazard Mater 282, 106-115. Stasinakis, A. (2008) Use of selected advanced oxidation processes (AOPs) for wastewater treatment—a mini review. Global NEST Journal 10(3), 376-385. Stearman, R., Yuan, D.S., Yamaguchi-Iwai, Y., Klausner, R.D. and Dancis, A. (1996) A permease-oxidase complex involved in high-affinity iron uptake in yeast. Science 271(5255), 1552-1557. Suarez, S., Lema, J.M. and Omil, F. (2009) Pre-treatment of hospital wastewater by coagulation–flocculation and flotation. Bioresour Technol 100(7), 2138-2146. Sugihara, M.N., Moeller, D., Paul, T. and Strathmann, T.J. (2013) TiO 2-photocatalyzed transformation of the recalcitrant X-ray contrast agent diatrizoate. Applied Catalysis B: Environmental 129, 114-122. Sumpter, J.P. (1998) Xenoendocrine disrupters — environmental impacts. Toxicology Letters 102–103(0), 337-342. Tadkaew, N., Hai, F.I., McDonald, J.A., Khan, S.J. and Nghiem, L.D. (2011) Removal of trace organics by MBR treatment: the role of molecular properties. Water Res 45(8), 2439-2451. Takeshita, K., Shibato, J., Sameshima, T., Fukunaga, S., Isobe, S., Arihara, K. and Itoh, M. (2003) Damage of yeast cells induced by pulsed light irradiation. International Journal of Food Microbiology 85(1), 151-158. Temple, M.D., Perrone, G.G. and Dawes, I.W. (2005) Complex cellular responses to reactive oxygen species. Trends in Cell Biology 15(6), 319-326. Templeton, M., Andrews, R. and Hofmann, R. (2006) Impact of iron particles in groundwater on the UV inactivation of bacteriophages MS2 and T4. J Appl Microbiol 101(3), 732-741. Ternes, T.A. and Hirsch, R. (2000) Occurrence and behavior of X-ray contrast media in sewage facilities and the aquatic environment. Environmental Science & Technology 34(13), 2741-2748. Ternes, T.A., Joss, A. and Siegrist, H. (2004) Peer reviewed: scrutinizing pharmaceuticals and personal care products in wastewater treatment. Environmental Science & Technology 38(20), 392A-399A. Thabet, S., Simonet, F., Lemaire, M., Guillard, C. and Cotton, P. (2014) Impact of photocatalysis on fungal cells: depiction of cellular and molecular effects on Saccharomyces cerevisiae. Applied and environmental microbiology 80(24), 7527-7535. Thabet, S., Weiss-Gayet, M., Dappozze, F., Cotton, P. and Guillard, C. (2013) Photocatalysis on yeast cells: Toward targets and mechanisms. Applied Catalysis B: Environmental 140–141(0), 169-178. Tian, F.-X., Xu, B., Lin, Y.-L., Hu, C.-Y., Zhang, T.-Y. and Gao, N.-Y. (2014) Photodegradation kinetics of iopamidol by UV irradiation and enhanced formation of iodinated disinfection by-products in sequential oxidation processes. Water Res 58, 198-208. Timchak, E. and Gitis, V. (2012) A combined degradation of dyes and inactivation of viruses by UV and UV/H 2 O 2. Chemical Engineering Journal 192, 164-170. Touati, D. (2000) Iron and oxidative stress in bacteria. Arch Biochem Biophys 373(1), 1-6. Trovó, A.G. and Nogueira, R.F. (2011) Diclofenac abatement using modified solar photo-Fenton process with ammonium iron (III) citrate. Journal of the Brazilian Chemical Society 22(6), 1033-1039.

220

TYRRELL, R.M. and PIDOUX, M. (1989) SINGLET OXYGEN INVOLVEMENT IN THE INACTIVATION OF CULTURED HUMAN FIBROBLASTS BY UVA (334 nm, 365 nm) AND NEAR-VISIBLE (405 nm) RADIATIONS. Photochem Photobiol 49(4), 407-412. Tyrrell, R.M., Pourzand, C.A., Brown, J., Hejmadi, V., Kvam, V., Ryter, S. and Watkin, R. (2000) Cellular studies with UVA radiation: a role for iron. Radiation protection dosimetry 91(1-3), 37-39. Ułaszewski, S., Mamouneas, T., Shen, W.-K., Rosenthal, P.J., Woodward, J.R., Cirillo, V.P. and Edmunds, L.N. (1979) Light effects in yeast: Evidence for participation of cytochromes in photoinhibition of growth and transport in Saccharomyces cerevisiae cultured at low temperatures. Journal of bacteriology 138(2), 523-529. Upritchard, H.G., Yang, J., Bremer, P.J., Lamont, I.L. and McQuillan, A.J. (2007) Adsorption to metal oxides of the Pseudomonas aeruginosa siderophore pyoverdine and implications for bacterial biofilm formation on metals. Langmuir 23(13), 7189-7195. USNLM (2014) U.S. National Library of Medicine. USNLM (2016a) Venlafaxine, U.S. National Library of Medicine. USNLM (2016b) Venlafaxine Hydrochloride. Verlicchi, P., Al Aukidy, M., Galletti, A., Petrovic, M. and Barcelo, D. (2012) Hospital effluent: investigation of the concentrations and distribution of pharmaceuticals and environmental risk assessment. Sci Total Environ 430, 109-118. Verlicchi, P., Galletti, A., Petrovic, M. and Barceló, D. (2010) Hospital effluents as a source of emerging pollutants: An overview of micropollutants and sustainable treatment options. Journal of Hydrology 389(3-4), 416-428. Vieno, N., Tuhkanen, T. and Kronberg, L. (2007) Elimination of pharmaceuticals in sewage treatment plants in Finland. Water Res 41(5), 1001-1012. Viollier, E., Inglett, P.W., Hunter, K., Roychoudhury, A.N. and Van Cappellen, P. (2000) The ferrozine method revisited: Fe(II)/Fe(III) determination in natural waters. Applied Geochemistry 15(6), 785-790. Vione, D., Minella, M., Maurino, V. and Minero, C. (2014) Indirect Photochemistry in Sunlit Surface Waters: Photoinduced Production of Reactive Transient Species. Chemistry – A European Journal 20(34), 10590-10606. Voelker, B.M., Morel, F.M.M. and Sulzberger, B. (1997) Iron Redox Cycling in Surface Waters: Effects of Humic Substances and Light. Environmental Science & Technology 31(4), 1004-1011. Voelker, B.M. and Sedlak, D.L. (1995) Iron reduction by photoproduced superoxide in seawater. Marine Chemistry 50(1), 93-102. Vulliet, E., Cren-Olivé, C. and Grenier-Loustalot, M.-F. (2011) Occurrence of pharmaceuticals and hormones in drinking water treated from surface waters. Environmental Chemistry Letters 9(1), 103-114. Wang, C., Shi, H., Adams, C.D., Gamagedara, S., Stayton, I., Timmons, T. and Ma, Y. (2011) Investigation of pharmaceuticals in Missouri natural and drinking water using high performance liquid chromatography-tandem mass spectrometry. Water Res 45(4), 1818-1828. Weast, R.C. (1985) Handbook of data on organic compounds. Weir, S.C., Pokorny, N.J., Carreno, R.A., Trevors, J.T. and Lee, H. (2002) Efficacy of Common Laboratory Disinfectants on the Infectivity of Cryptosporidium parvum Oocysts in Cell Culture. Applied and environmental microbiology 68(5), 2576-2579. Weissbrodt, D., Kovalova, L., Ort, C., Pazhepurackel, V., Moser, R., Hollender, J., Siegrist, H. and McArdell, C.S. (2009) Mass Flows of X-ray Contrast Media and Cytostatics in Hospital Wastewater. Environmental Science & Technology 43(13), 4810-4817. Weller, E.B., Weller, R.A. and Davis, G.P. (2000) Use of venlafaxine in children and adolescents: a review of current literature. Depression and anxiety 12(S1), 85-89. Wenk, J., von Gunten, U. and Canonica, S. (2011) Effect of Dissolved Organic Matter on the Transformation of Contaminants Induced by Excited Triplet States and the Hydroxyl Radical. Environmental Science & Technology 45(4), 1334-1340. Westerhoff, P., Aiken, G., Amy, G. and Debroux, J. (1999) Relationships between the structure of natural organic matter and its reactivity towards molecular ozone and hydroxyl radicals. Water Res 33(10), 2265-2276. White, G.C. (2010) White's handbook of chlorination and alternative disinfectants, Wiley.

221

Wigginton, K.R., Pecson, B.M., Sigstam, T.r., Bosshard, F. and Kohn, T. (2012) Virus inactivation mechanisms: impact of disinfectants on virus function and structural integrity. Environmental Science & Technology 46(21), 12069-12078. Wols, B., Hofman-Caris, C., Harmsen, D. and Beerendonk, E. (2013) Degradation of 40 selected pharmaceuticals by UV/H 2 O 2. Water Res 47(15), 5876-5888. Wu, C. and Linden, K.G. (2010) Phototransformation of selected organophosphorus pesticides: Roles of hydroxyl and carbonate radicals. Water Res 44(12), 3585-3594. Xu, J., Sahai, N., Eggleston, C.M. and Schoonen, M.A. (2013) Reactive oxygen species at the oxide/water interface: Formation mechanisms and implications for prebiotic chemistry and the origin of life. Earth and Planetary Science Letters 363, 156-167. Yun, C.-W., Ferea, T., Rashford, J., Ardon, O., Brown, P.O., Botstein, D., Kaplan, J. and Philpott, C.C. (2000) Desferrioxamine-mediated iron uptake in Saccharomyces cerevisiae Evidence for two pathways of iron uptake. Journal of Biological Chemistry 275(14), 10709-10715. Zapata, A., Oller, I., Rizzo, L., Hilgert, S., Maldonado, M., Sánchez-Pérez, J. and Malato, S. (2010) Evaluation of operating parameters involved in solar photo-Fenton treatment of wastewater: Interdependence of initial pollutant concentration, temperature and iron concentration. Applied Catalysis B: Environmental 97(1), 292-298. Zhang, R., Sun, P., Boyer, T.H., Zhao, L. and Huang, C.-H. (2015) Degradation of Pharmaceuticals and Metabolite in Synthetic Human Urine by UV, UV/H2O2, and UV/PDS. Environmental Science & Technology 49(5), 3056-3066. Zhang, Z., Boxall, C. and Kelsall, G. (1993) Photoelectrophoresis of colloidal iron oxides 1. Hematite (α-Fe 2 O 3). Colloids and Surfaces A: Physicochemical and Engineering Aspects 73, 145-163. Zhao, C., Arroyo-Mora, L.E., DeCaprio, A.P., Sharma, V.K., Dionysiou, D.D. and O'Shea, K.E. (2014) Reductive and oxidative degradation of iopamidol, iodinated X-ray contrast media, by Fe (III)-oxalate under UV and visible light treatment. Water Res 67, 144-153.

222

Appendix A: Supplementary material of Chapter 2

1. MP degradation details

Observing the degradation rates of each MP, in combination with the specific pollutant information

of Supplementary Table 2, a detailed explanation is given below.

Diclofenac

This compound has a relative high k value when treated with UV-C, which could not be estimated

because it was eliminated before the sampling interval. It also presented a high degradation constant

when exposed to solar irradiation, indicating its photo-sensitive behavior. It also reacted quickly with

OH• radicals, which is normal because diclofenac has reported the highest apparent second-order rate

constant among the eight selected MPs (1.0 x 1011 M-1.s-1). It showed high degradation rate for Fenton,

and photo-Fenton, except for the WW produced by CF which was very slow, and the delay was related

with the physicochemical characteristics of the WW instead of the pollutant (Klamerth et al. 2010).

The UV/H2O2 process degraded Diclofenac very fast in all three types of WW.

Mecoprop

Together with diclofenac, this compound showed the highest degradation rate when treated with UV-

C irradiation. This is attributed to its chromophores that absorb UV at wavelengths between 280 and

290 nm. During solar irradiation, the degradation rate was high in the AS WW type, while very low for

the MBBR and CF WW types. The higher suspended solid concentrations of these two WW types may

be a reason of solar irradiation low degradation rates. Mecoprop again (De la Cruz et al. 2013) showed

moderate degradation rates which comply with previous apparent second-order rate constants

observed (kOH = 1.0 x 1010 M-1.s-1). The presence of –OH and –OCH3 radicals in its chemical structure

facilitates its degradation (kOH = 97 x 107 M-1.s-1 and 100 x 107 M-1.s-1, respectively).

Metoprolol

This compound showed (after diclofenac) a remarkable degradation rate when treated with UV-C

radiation. Indeed, Metoprolol UV-absorption peak is at 223 nm, very close to 254 nm emission of the

UV-C lamps. It was also well degraded when exposed to solar irradiation, pointing out a general photo-

sensible nature. Again in this case, the degradation rate constant was higher in MBBR than CF WW

type; which might be partly explained by the inexistent potential desorption effect of Metoprolol in

CF (high solubility = 16900 mg/L and low kow = 1.88). On the contrary, this compound degraded

223

moderately fast when exposed to hydroxyl radicals in Fenton and photo-Fenton treatment (Prieto-

Rodríguez et al. 2013) which agrees with the HO rate constant (kOH = 9.6 x 1010 M-1.s-1). UV/H2O2

treatment entailed high (measured) degradation rates of this compound.

Benzotriazole

This compound demonstrated the highest degradation rate constants, compared to the rest of MPs,

when treated with UV-C irradiation. However, it is suggested that this behavior is attributed to the

fact that the initial Benzotriazole concentration was ~50 times higher than the other MPs for all types

of effluents. Degradation rate constants were also high when it was exposed to solar irradiation, which

is expected, given that the UV-absorption peak of this compound is at 465 nm. Only Benzotriazole

contained in CF WW type showed resistance to solar irradiation (k = 0 min-1), probably because of the

high concentration of suspended solids that may have blocked sunlight transmittance. Benzotriazole

showed relatively low degradation rates when exposed to hydroxyl radicals during Fenton and photo-

Fenton treatment, which is consistent with previous HO apparent second-order rate constant

observed (6.0 x 1010 M-1.s-1). The small rate reaction of Benzotriazole with hydroxyl radicals (De la Cruz

et al. 2013) is explained by the presence of –CN compound in its chemical structure (kOH = 2.2 x 107 M-

1.s-1).

Carbamazepine

This compound withholds moderate degradation constants for UV-C and solar irradiation because it

contains chromophores that absorb at wavelengths greater than 290 nm and therefore may be

susceptible to direct photolysis by UV-C and even by sunlight. Only the WW coming from CF showed

insignificant removal after solar irradiation (ksolar = 0 min-1). Moreover, Carbamazepine treated with

Fenton, photo-Fenton and UV/H2O2 also showed moderate degradation rate constants (Klamerth et

al. 2010), which is coherent with the apparent second-rate constant reported in literature (k = 4.9 x

1010 M-1.s-1). Carbamazepine also has a -NH2 compound, which reacts fast with hydroxyl radicals (kOH

= 420 x 107 M-1.s-1). However, further degradation is expected to decelerate due to the formation of a

carboxylic radical (kOH= 1.6 x 107 M-1.s-1). Only carbamazepine contained in the CF WW type showed

very low degradation constants (e.g. kFenton = 0 min-1) when subjected to these treatment methods.

Clarithromycin

This compound showed the lowest degradation rates for UV-C, UV/H2O2, photo-Fenton, solar

irradiation and Fenton reaction. Furthermore, it was the only compound not degraded after 5 minutes

of UV/H2O2 treatment for the AS WW type. This persistent behavior might be attributed to its higher

224

molecular weight (748 g/mol), compared to the other MPs and mainly to its complex chemical

structure which includes several deactivating groups. The degradation pathway of this pollutant

involves a significant number of intermediates prior to mineralization. Previous works reported low k

for Clarithromycin as well (Kim et al. 2009b).

2. Supplementary Figures

Figure S1 – H2O2 consumption during UV/ H2O2 treatment

225

Figure S2 – Dissolved iron content and H2O2 consumption during Fenton treatment

226

Figure S3 – Dissolved iron content and H2O2 consumption during photo-Fenton treatment

227

Figure S4 – Summary of the degradation results, per secondary (pre)treatment method and

advanced (post)treatment process.

0102030405060708090

100

UV - 10' UV - 30' UV/H2O2 - 5' UV/H2O2 - 10' Solar - 30' Solar - 60' Fenton - 60' Fenton - 120' Photo-Fenton -30'

Photo-Fenton -60'

% D

egra

datio

nActivated Sludge

Carbamazepine

Diclofenac

Metoprolol

Clarithromycin

Benzotriazole

Mecoprop

0102030405060708090

100


Photo-Fenton -60'

% D

egra

datio

n


Carbamazepine

Diclofenac

Metoprolol

Clarithromycin

Benzotriazole

Mecoprop

0102030405060708090

100


Photo-Fenton -60'

% D

egra

datio

n

Coagulation-Flocculation

Carbamazepine

Diclofenac

Metoprolol

Clarithromycin

Benzotriazole

Mecoprop

228

3. Supplementary Tables

Table S1: Micropollutant degradation order, per process and matrix involved.

UV-based all matrices Carbamazepine Clarithromycin Metoprolol Benzotriazole Mecoprop Diclofenac Fenton-related all matrices Clarithromycin Benzotriazole Metoprolol Carbamazepine Mecoprop Diclofenac

UV-based

AS Benzotriazole Carbamazepine Metoprolol Clarithromycin Mecoprop Diclofenac

MBBR Carbamazepine Clarithromycin Metoprolol Benzotriazole Mecoprop Diclofenac

CF Carbamazepine Clarithromycin Metoprolol Benzotriazole Mecoprop Diclofenac

Fenton-related

AS Clarithromycin Benzotriazole Metoprolol Mecoprop Carbamazepine Diclofenac

MBBR Clarithromycin Benzotriazole Metoprolol Carbamazepine Mecoprop Diclofenac

CF Clarithromycin Benzotriazole Metoprolol Carbamazepine Mecoprop Diclofenac

Table S2: Main physical and chemical properties of the eight selected micropollutants (USNLM 2014).

Compound Chemical structure

Molecular

weight

(g/mol)

Water

solubility

(mg/L)

log

kow pka

Henry’s

coefficient

(H)

(atm.m3/mol)

OH•

reaction

rate

constant

(M-1.s-1)

Carbamazepine

C15H12N2O

236.269 18 2.45 13.9 1.1 x 10-10

4.9 x

1010

Diclofenac

C14H11Cl2NO2

296.149 2.4 4.51 4.15 4.7 x 10-12

1.0 x

1011

229

Clarithromycin

C38H69NO13

747.953 1.7 3.16 8.99 - 5 x 109

Metoprolol

C15H25NO3

267.364 16900 1.88 9.68 2.1 x 10-11

9.6 x

1010

Benzotriazole

C6H5N3

119.124 19800 1.44 8.37 3.17 x 10-7 6.0 x 108

Mecoprop

C10H11ClO3

214.646 880 3.13 3.78 3.8 x 10-9

1.0 x

1010

230

Appendix B: Supplementary material of Chapter 3

Supplementary Figure S1 – UV reactor configuration.

231

Supplementary Figure S2 – Absorbance and transmittance spectra of the four effluents. For the

transmittance, samples were analyzed as received, and for the absorbance spectra, filtering through

0.45 μm filter took place beforehand.

232

Table S1 – AOPs used, sampling times and reagents addition for all effluents.

Microorganisms Initial H2O2 Initial Fe2+

[mg/L] [mg/L] UV-C irradiation - -

UV/H2O2 20 -

Solar irradiation - - Fenton 10, 20 2

photo-Fenton 10, 20 2

Micropollutants Initial H2O2 Initial Fe2+

[mg/L] [mg/L] UV-C irradiation - -

UV/H2O2 25 -

Solar irradiation - - Fenton 25 5

photo-Fenton 25 5

233

Table S2 – Synergy (S) among the photo-Fenton constituents for micropollutant degradation in the

treated effluents.

Degradation % per minute Pretreatment Process Activated Sludge Moving Bed BioReactor Coagulation-Flocculation

Fenton process 0.2417 0.3884 0.1550 Solar light 0.2092 0.3076 0.1279

Photo-Fenton 0.6894 0.8516 0.4289

Solar + Fenton 0.4509 0.6960 0.2829

1.5289 1.2236 1.5161

234

Table S3 – Synergy (S) among the solar photo-Fenton constituents in microorganism elimination: first

order reaction kinetics constant.

Kinetics constant (k) Pretreatment

process Primary

Treatment Activated

Sludge Moving Bed BioReactors

Coagulation-Flocculation

Fenton process 0.32 0.45 0.51 0.45

Solar light 0.61 0.76 0.81 0.62 Photo-Fenton 0.95 2.14 2.21 1.09

Solar + Fenton 0.93 1.21 1.32 1.07

1.02 1.77 1.68 1.02

235

Appendix C: Supplementary material of Chapter 4

Figure S1 – Suntest solar simulator light wavelength emission spectrum (Manufacturer: Suntest Xenon

Test-Instruments Brochure)

Figure S2 – (dark) Fenton and solar disinfection experiments. a) Individual effect of the Fenton

reagents (Fe(II), Fe(III) and H2O2). b) Solar wastewater exposure under 300, 600 and 900 W/m2 global

irradiance.

236

Figure S3 - Absorbance spectra of Fe(III) in Mili-Q water or wastewater and iron solubility experiments, 1, 2 or 5 mg/L iron addition in MQ or wastewater (reference: pure MQ).

237

Appendix D: Supplementary material of Chapter 5

Figure S1 – H2O2 evolution during the assays. The different tests are: H2O2 only, light/ H2O2 (hv/H2O2)

and photo-Fenton (pF) with FeSO4 as starting iron salt (FeSO4 pF), with iron citrate (Fe-cit) and

Goethite (Goethite pF). The numbers following indicate the starting pH of the corresponding test.

238

Statistical connection among the cultivability and viability assays

Table S1 summarizes the findings of the Pearson correlation test and the P-Values of the statistical

significance test. As it appears, the three parameters are connected with each other. The CFDA

diminishing is well correlated with the cultivability (0.71 Pearson value) and P<0.05 for the 95% confidence

interval, while PI increase is not well correlated (<0.6) with the cultivability.

Table S1 – Correlation among CFDA or PI staining techniques results and cultivability

CFDA IP

IP -0.855 - Pearson

value

0 - P-value

Cultivability 0.71 -0.512 Pearson

value

0.001 0.03 P-value

This difference is attributed to the intermediate cell states existing, which count as non-cultivable, but do

not repel the propidium ion (yet). Our values are lower than the ones obtained by other authors with TiO2

[27], probably because of the difference in the mode of action, among TiO2 and photo-Fenton. The

contour plot in FigureS2 depicts this correlation among the three assays.

Figure S2 – Contour plot of the correlation among the staining techniques results and the cultivability.

Cultivability (%) vs CFDA (%) and PI (%)Cultivability (%)

PI

CFDA

239

In order to decrease 1-log the cultivability, the readings of CFDA are among 40 and 70%, for all possible PI

values. Consequently, 2-log reduction for CFDA < 40%, 3-log for CFDA < 10% and 50-70% IP and total

inactivation (>4-log) at simultaneous 95% PI and 95% decrease of CFDA.

Figure S3 – Overview of the results of the different staining techniques. a) CFDA staining, indicating

the live cells. b) PI staining, indicating the dead cells.

In Figure S3, there is a consistent trend, where the decrease of esterase activity is slower in simulated

solar light, followed by the hv/H2O2 system. The photo-Fenton systems executed at pH = 7.5 are more

efficient, and finally the pH 6.0 and 5.5 tests are the fastest to disrupt the esterase activity. Conversely, PI

staining appears in the opposite order. The yeast inactivation efficiency is as follows: photo-Fenton (pH<7)

> photo-Fenton (pH=7.5) > hv/H2O2 > hv.

240

Appendix E: Supplementary material of Chapter 6

Supplementary figures

Figure S1 – Experimental installation. UV reactors and related apparatus.

Figure S2 – Indicative chromatogram and peaks of Iohexol TPs during UV/H2O2 treatment.

241

Figure S3 – Absorbance spectra for UV, UV/H2O2 and UV/H2O2/Fe2+ treatment. The UV process is

marked with A, UV/H2O2 with B and for UV/H2O2/Fe2+ with C, in MQ water.

Supplementary Table

Table S1: Detailed information on the CCD for each of the following matrices: WW, diluted WW, UR,

diluted UR.

Factors 3 Run H2O2 Fe2+ pH Levels 2 1 -1 -1 -1

2 1 -1 -1 a 1.68179 3 -1 1 -1

Base runs 20 4 1 1 -1 Replicates 2 5 -1 -1 1 Total Runs 40 6 1 -1 1

7 -1 1 1 Cube points 8 8 1 1 1

Center points 6 9 -1.68179 0 0 Axial points 6 10 1.681793 0 0

11 0 -1.68179 0 Levels -1 1 12 0 1.681793 0 H2O2 20 50 ppm 13 0 0 -1.68179 Fe2+ 10 20 ppm 14 0 0 1.681793 pH 3 5 15 0 0 0

16 0 0 0 Response k min-1 17 0 0 0

18 0 0 0 19 0 0 0 20 0 0 0

242

Statistical analyses nomenclature and abbreviations

DF: Degrees of Freedom

Seq SS: Sequential Sum of Squares

Adj SS: Adjusted Sum of Squares

Adj MS: Adjusted Mean of Squares

F-Value: Ratio of Mean Squares

P-Value: Hypothesis test value

S: Standard Error

R2: Coefficient of Determination

R2 (adj): Adjusted Coefficient of Determination

PRESS: Prediction Sum of Squares

R2 (pred): Predicted Coefficient of Determination

243

1. Statistical Analysis: Models and ANOVA tables for each model

1.1. ANOVA and Model Summary for (Undiluted) WW models

Linear:

Table S1.1.1. Analysis of Variance

Source DF Seq SS Contribution Adj SS Adj MS F-Value P-Value

Model 4 0.087581 38.44% 0.087581 0.021895 5.15 0.002

Linear 4 0.087581 38.44% 0.087581 0.021895 5.15 0.002

[I] 1 0.086400 37.92% 0.057602 0.057602 13.55 0.001

[H2O2] 1 0.000773 0.34% 0.000602 0.000602 0.14 0.709

[Fe] 1 0.000034 0.01% 0.000022 0.000022 0.01 0.943

pH 1 0.000374 0.16% 0.000374 0.000374 0.09 0.769

Error 33 0.140245 61.56% 0.140245 0.004250

Total 37 0.227826 100.00%

Table S1.1.2. - Model Summary

S R2 R2 (adj) PRESS R2 (pred)

0.0651909 38.44% 30.98% 0.163796 28.10%

Linear with Squares:

Table S1.1.3. - Analysis of Variance


Model 8 0.210459 92.38% 0.210459 0.026307 43.93 <0.001

Linear 4 0.087581 38.44% 0.200111 0.050028 83.54 <0.001

[I] 1 0.086400 37.92% 0.166283 0.166283 277.66 <0.001

[H2O2] 1 0.000773 0.34% 0.000468 0.000468 0.78 0.384

[Fe] 1 0.000034 0.01% 0.000043 0.000043 0.07 0.790

pH 1 0.000374 0.16% 0.001882 0.001882 3.14 0.087

244

Square 4 0.122878 53.94% 0.122878 0.030720 51.30 <0.001

[I] x [I] 1 0.115482 50.69% 0.109366 0.109366 182.62 <0.001

[H2O2] x [H2O2] 1 0.004193 1.84% 0.003042 0.003042 5.08 0.032

[Fe] x [Fe] 1 0.003198 1.40% 0.003157 0.003157 5.27 0.029

pH x pH 1 0.000006 0.00% 0.000006 0.000006 0.01 0.922

Error 29 0.017367 7.62% 0.017367 0.000599

Total 37 0.227826 100.00%



0.0244717 92.38% 90.27% 0.0494703 78.29%

Quadratic Model:



Model 14 0.213923 93.90% 0.213923 0.015280 25.28 <0.001

Linear 4 0.087581 38.44% 0.092411 0.023103 38.22 <0.001

[I] 1 0.086400 37.92% 0.008808 0.008808 14.57 0.001

[H2O2] 1 0.000773 0.34% 0.002171 0.002171 3.59 0.071

[Fe] 1 0.000034 0.01% 0.000093 0.000093 0.15 0.699

pH 1 0.000374 0.16% 0.000111 0.000111 0.18 0.673

Square 4 0.122878 53.94% 0.123821 0.030955 51.21 <0.001

[I] x [I] 1 0.115482 50.69% 0.104702 0.104702 173.21 <0.001

[H2O2] x [H2O2] 1 0.004193 1.84% 0.000068 0.000068 0.11 0.741

[Fe] x [Fe] 1 0.003198 1.40% 0.000738 0.000738 1.22 0.281

pH x pH 1 0.000006 0.00% 0.000025 0.000025 0.04 0.842

2-Way Interaction

6 0.003464 1.52% 0.003464 0.000577 0.95 0.477

[I] x [H2O2] 1 0.002745 1.20% 0.001254 0.001254 2.07 0.163

[I] x [Fe] 1 0.000231 0.10% 0.000248 0.000248 0.41 0.529

245

[I] x pH 1 0.000034 0.01% 0.000004 0.000004 0.01 0.933

[H2O2] x [Fe] 1 0.000398 0.17% 0.000412 0.000412 0.68 0.417

[H2O2] x pH 1 0.000038 0.02% 0.000034 0.000034 0.06 0.814

[Fe] x pH 1 0.000018 0.01% 0.000018 0.000018 0.03 0.865

Error 23 0.013903 6.10% 0.013903 0.000604

Total 37 0.227826 100.00%



0.0245865 93.90% 90.18% * *

Multiplicative model:



Model 4 9.03575 96.62% 9.03575 2.25894 242.94 <0.001

Linear 4 9.03575 96.62% 9.03575 2.25894 242.94 <0.001

log[I] 1 8.55057 91.43% 8.43966 8.43966 907.65 <0.001

log[H2O2] 1 0.07906 0.85% 0.02190 0.02190 2.35 0.134

log[Fe] 1 0.34392 3.68% 0.10067 0.10067 10.83 0.002

log[H+] 1 0.06220 0.67% 0.06220 0.06220 6.69 0.014

Error 34 0.31614 3.38% 0.31614 0.00930

Total 38 9.35189 100.00%



0.0964280 96.62% 96.22% 0.454375 95.14%

246

1.2. ANOVA and Model Summary for Diluted WW models

Linear:



Model 4 0.192009 61.38% 0.192009 0.048002 7.15 0.001

Linear 4 0.192009 61.38% 0.192009 0.048002 7.15 0.001

[I] 1 0.186773 59.71% 0.136364 0.136364 20.32 <0.001

[H2O2] 1 0.000720 0.23% 0.003609 0.003609 0.54 0.473

[Fe] 1 0.004511 1.44% 0.002598 0.002598 0.39 0.542

pH 1 0.000004 0.00% 0.000004 0.000004 0.00 0.980

Error 18 0.120790 38.62% 0.120790 0.006711

Total 22 0.312798 100.00%



0.0819179 61.38% 52.80% 0.181728 41.90%




Model 8 0.289349 92.50% 0.289349 0.036169 21.59 <0.001

Linear 4 0.192009 61.38% 0.262954 0.065738 39.25 <0.001

[I] 1 0.186773 59.71% 0.223651 0.223651 133.53 <0.001

[H2O2] 1 0.000720 0.23% 0.000342 0.000342 0.20 0.658

[Fe] 1 0.004511 1.44% 0.000503 0.000503 0.30 0.592

pH 1 0.000004 0.00% 0.004705 0.004705 2.81 0.116

Square 4 0.097341 31.12% 0.097341 0.024335 14.53 <0.001

[I] x [I] 1 0.093373 29.85% 0.068851 0.068851 41.11 <0.001

247

[H2O2] x [H2O2] 1 0.000512 0.16% 0.000418 0.000418 0.25 0.625

[Fe] x [Fe] 1 0.000584 0.19% 0.000049 0.000049 0.03 0.867

pH x pH 1 0.002872 0.92% 0.002872 0.002872 1.71 0.211

Error 14 0.023449 7.50% 0.023449 0.001675

Total



0.0409260 92.50% 88.22% 0.0775774 75.20%

Quadratic Model:



Model 14 0.299548 95.76% 0.299548 0.021396 12.92 0.001

Linear 4 0.192009 61.38% 0.089026 0.022256 13.44 0.001

[I] 1 0.186773 59.71% 0.009570 0.009570 5.78 0.043

[H2O2] 1 0.000720 0.23% 0.000914 0.000914 0.55 0.479

[Fe] 1 0.004511 1.44% 0.000406 0.000406 0.24 0.634

pH 1 0.000004 0.00% 0.000408 0.000408 0.25 0.633

Square 4 0.097341 31.12% 0.078487 0.019622 11.85 0.002

[I] x [I] 1 0.093373 29.85% 0.076154 0.076154 45.98 <0.001

[H2O2] x [H2O2] 1 0.000512 0.16% 0.000561 0.000561 0.34 0.577

[Fe] x [Fe] 1 0.000584 0.19% 0.000115 0.000115 0.07 0.799

pH x pH 1 0.002872 0.92% 0.000758 0.000758 0.46 0.518

2-Way Interaction

6 0.010199 3.26% 0.010199 0.001700 1.03 0.472

[I] x [H2O2] 1 0.003070 0.98% 0.001021 0.001021 0.62 0.455

[I] x [Fe] 1 0.005907 1.89% 0.000070 0.000070 0.04 0.842

[I] x pH 1 0.000438 0.14% 0.000395 0.000395 0.24 0.639

[H2O2] x [Fe] 1 0.000697 0.22% 0.000436 0.000436 0.26 0.622

248

[H2O2] x pH 1 0.000002 0.00% 0.000076 0.000076 0.05 0.836

[Fe] x pH 1 0.000086 0.03% 0.000086 0.000086 0.05 0.826

Error 8 0.013250 4.24% 0.013250 0.001656

Total 22 0.312798 100.00%



0.0406972 95.76% 88.35% 107.505 0.00%




Model 4 6.56526 92.24% 6.56526 1.64132 56.47 <0.001

Linear 4 6.56526 92.24% 6.56526 1.64132 56.47 <0.001

log[I] 1 5.89394 82.81% 4.87605 4.87605 167.77 <0.001

log[H2O2] 1 0.06083 0.85% 0.00026 0.00026 0.01 0.926

log[Fe] 1 0.34575 4.86% 0.01315 0.01315 0.45 0.509

log[H+] 1 0.26473 3.72% 0.26473 0.26473 9.11 0.007

Error 19 0.55222 7.76% 0.55222 0.02906

Total 23 7.11748 100.00%



0.170482 92.24% 90.61% 0.985649 86.15%

249

1.3. Correlation among diluted and undiluted WW

Linear:

logK_dil = 0.1730 + 0.9879 logK_undil

S = 0.0823160 R2 = 97.7% R2 (adj) = 97.6%


Source DF SS MS F P

Regression 1 5.57677 5.57677 823.03 0.000

Error 19 0.12874 0.00678

Total 20 5.70551

250

1.4. ANOVA and Model Summary for (undiluted) UR models

Linear:



Model 4 0.021003 63.78% 0.021003 0.005251 14.97 <0.001

Linear 4 0.021003 63.78% 0.021003 0.005251 14.97 <0.001

[I] 1 0.006919 21.01% 0.002533 0.002533 7.22 0.011

[H2O2] 1 0.000356 1.08% 0.000280 0.000280 0.80 0.378

[Fe] 1 0.006179 18.76% 0.000229 0.000229 0.65 0.425

pH 1 0.007550 22.93% 0.007550 0.007550 21.52 0.000

Error 34 0.011928 36.22% 0.011928 0.000351

Total 38 0.032931 100.00%



0.0187304 63.78% 59.52% 0.0164007 50.20%




Model 8 0.026208 79.59% 0.026208 0.003276 14.62 <0.001

Linear 4 0.021003 63.78% 0.008111 0.002028 9.05 <0.001

[I] 1 0.006919 21.01% 0.001088 0.001088 4.85 0.035

[H2O2] 1 0.000356 1.08% 0.000005 0.000005 0.02 0.878

[Fe] 1 0.006179 18.76% 0.000003 0.000003 0.01 0.906

pH 1 0.007550 22.93% 0.006649 0.006649 29.67 <0.001

Square 4 0.005206 15.81% 0.005206 0.001301 5.81 0.001

[I] x [I] 1 0.000045 0.14% 0.000232 0.000232 1.03 0.317

251

[H2O2] x [H2O2] 1 0.000166 0.51% 0.000551 0.000551 2.46 0.127

[Fe] x [Fe] 1 0.002209 6.71% 0.003093 0.003093 13.80 0.001

pH x pH 1 0.002785 8.46% 0.002785 0.002785 12.43 0.001

Error 30 0.006723 20.41% 0.006723 0.000224

Total 38 0.032931 100.00%



0.0149695 79.59% 74.14% 0.0113284 65.60%

Quadratic model:



Model 14 0.031552 95.81% 0.031552 0.002254 39.25 <0.001

Linear 4 0.021003 63.78% 0.001325 0.000331 5.77 0.002

[I] 1 0.006919 21.01% 0.000328 0.000328 5.71 0.025

[H2O2] 1 0.000356 1.08% 0.000031 0.000031 0.54 0.472

[Fe] 1 0.006179 18.76% 0.000011 0.000011 0.18 0.672

pH 1 0.007550 22.93% 0.000937 0.000937 16.31 <0.001

Square 4 0.005206 15.81% 0.000972 0.000243 4.23 0.010

[I] x [I] 1 0.000045 0.14% 0.000004 0.000004 0.07 0.790

[H2O2] x [H2O2] 1 0.000166 0.51% 0.000043 0.000043 0.75 0.394

[Fe] x [Fe] 1 0.002209 6.71% 0.000914 0.000914 15.92 0.001

pH x pH 1 0.002785 8.46% 0.000052 0.000052 0.90 0.353

2-Way Interaction

6 0.005344 16.23% 0.005344 0.000891 15.51 <0.001

[I] x [H2O2] 1 0.000098 0.30% 0.000013 0.000013 0.22 0.640

[I] x [Fe] 1 0.001262 3.83% 0.000262 0.000262 4.57 0.043

[I] x pH 1 0.001757 5.34% 0.002120 0.002120 36.91 <0.001

[H2O2] x [Fe] 1 0.000458 1.39% 0.000801 0.000801 13.95 0.001

252

[H2O2] x pH 1 0.000014 0.04% 0.001577 0.001577 27.46 <0.001

[Fe] x pH 1 0.001754 5.33% 0.001754 0.001754 30.55 <0.001

Error 24 0.001378 4.19% 0.001378 0.000057

Total 38 0.032931 100.00%



0.0075780 95.81% 93.37% 0.0047446 85.59%




Model 4 5.04941 83.42% 5.04941 1.26235 41.50 <0.001

Linear 4 5.04941 83.42% 5.04941 1.26235 41.50 <0.001

log[I] 1 1.16137 19.19% 1.40367 1.40367 46.15 <0.001

log[H2O2] 1 0.07470 1.23% 0.00149 0.00149 0.05 0.826

log[Fe] 1 2.57934 42.61% 0.42109 0.42109 13.84 0.001

log[H+] 1 1.23400 20.39% 1.23400 1.23400 40.57 <0.001

Error 33 1.00373 16.58% 1.00373 0.03042

Total 37 6.05314 100.00%



0.174402 83.42% 81.41% 1.39151 77.01%

253

1.5. ANOVA and Model Summary for diluted UR models

Linear:



Model 4 0.033269 76.14% 0.033269 0.008317 11.97 <0.001

Linear 4 0.033269 76.14% 0.033269 0.008317 11.97 <0.001

[I] 1 0.024963 57.13% 0.029512 0.029512 42.46 <0.001

[H2O2] 1 0.005612 12.84% 0.005586 0.005586 8.04 0.013

[Fe] 1 0.002572 5.89% 0.000996 0.000996 1.43 0.250

pH 1 0.000122 0.28% 0.000122 0.000122 0.18 0.681

Error 15 0.010425 23.86% 0.010425 0.000695

Total 19 0.043694 100.00%



0.0263631 76.14% 69.78% 0.0422853 3.22%




Model 8 0.041546 95.08% 0.041546 0.005193 26.60 <0.001

Linear 4 0.033269 76.14% 0.040179 0.010045 51.44 <0.001

[I] 1 0.024963 57.13% 0.037442 0.037442 191.75 <0.001

[H2O2] 1 0.005612 12.84% 0.001893 0.001893 9.69 0.010

[Fe] 1 0.002572 5.89% 0.000015 0.000015 0.08 0.786

pH 1 0.000122 0.28% 0.000585 0.000585 2.99 0.111

Square 4 0.008277 18.94% 0.008277 0.002069 10.60 0.001

[I] x [I] 1 0.008184 18.73% 0.004800 0.004800 24.58 <0.001

254

[H2O2] x [H2O2] 1 0.000059 0.13% 0.000035 0.000035 0.18 0.679

[Fe] x [Fe] 1 0.000021 0.05% 0.000034 0.000034 0.17 0.686

pH x pH 1 0.000014 0.03% 0.000014 0.000014 0.07 0.794

Error 11 0.002148 4.92% 0.002148 0.000195

Total 19 0.043694 100.00%



0.0139739 95.08% 91.51% 0.0263597 39.67%

Quadratic Model:



Model 12 0.043335 99.18% 0.043335 0.003611 70.46 <0.001

Linear 4 0.033269 76.14% 0.037618 0.009405 183.49 <0.001

[I] 1 0.024963 57.13% 0.037442 0.037442 730.54 <0.001

[H2O2] 1 0.005612 12.84% 0.000019 0.000019 0.36 0.566

[Fe] 1 0.002572 5.89% 0.000000 0.000000 0.00 0.978

pH 1 0.000122 0.28% 0.000131 0.000131 2.55 0.154

Square 4 0.008277 18.94% 0.003233 0.000808 15.77 0.001

[I] x [I] 1 0.008184 18.73% 0.002854 0.002854 55.68 <0.001

[H2O2] x [H2O2] 1 0.000059 0.13% 0.000006 0.000006 0.12 0.743

[Fe] x [Fe] 1 0.000021 0.05% 0.000040 0.000040 0.78 0.406

pH x pH 1 0.000014 0.03% 0.000003 0.000003 0.05 0.829

2-Way Interaction

4 0.001789 4.09% 0.001789 0.000447 8.73 0.007

[I] x [H2O2] 1 0.001651 3.78% 0.001764 0.001764 34.43 0.001

[H2O2] x [Fe] 1 0.000097 0.22% 0.000002 0.000002 0.03 0.867

[H2O2] x pH 1 0.000036 0.08% 0.000039 0.000039 0.76 0.412

[Fe] x pH 1 0.000005 0.01% 0.000005 0.000005 0.11 0.754

255

Error 7 0.000359 0.82% 0.000359 0.000051

Total 19 0.043694 100.00%



0.0071591 99.18% 97.77% 0.592478 0.00%




Model 4 2.01639 95.39% 2.01639 0.50410 82.74 <0.001

Linear 4 2.01639 95.39% 2.01639 0.50410 82.74 <0.001

log[I] 1 1.61151 76.23% 1.61151 1.61151 264.50 <0.001

log[H2O2] 1 0.11925 5.64% 0.01715 0.01715 2.82 0.113

log[Fe] 1 0.26743 12.65% 0.03630 0.03630 5.96 0.027

log[H+] 1 0.01820 0.86% 0.01820 0.01820 2.99 0.103

Error 16 0.09748 4.61% 0.09748 0.00609

Total 20 2.11388 100.00%



0.0780558 95.39% 94.24% 0.305412 85.55%

256

1.6. Correlation among diluted and undiluted UR

Linear:

logK_dil = - 0.2781 + 0.5143 logK_undil

S = 0.249791 R2 = 43.9% R2 (adj) = 41.0%


Source DF SS MS F P

Regression 1 0.92836 0.928357 14.88 0.001

Error 19 1.18552 0.062396

Total 20 2.11388

Quadratic:

logK_dil = - 3.092 - 3.217 logK_undil - 1.149 (logK_undil)2

S = 0.148008 R2 = 81.3% R2 (adj)= 79.3%


Source DF SS MS F P

Regression 2 1.71956 0.859781 39.25 0.000

Error 18 0.39431 0.021906

Total 20 2.11388

Cubic:

logK_dil = 4.436 + 11.52 logK_undil + 8.025 (logK_undil)2 + 1.821 (logK_undil)3

S = 0.108540 R2 = 90.5% R2 (adj) = 88.9%


Source DF SS MS F P

257

Regression 3 1.91360 0.637867 54.14 0.000

Error 17 0.20028 0.011781

Total 20 2.11388

258

2. MS supplementary material: List of identified products

Table S2.1. Mass spectrum characteristics of Iohexol products identified in degradation studies after 5 min

No Product Treatment Formula m/z experimental m/z theoretical ppma DBEb

1 P1 UV C19H27I2N3O10 711.9888 711.9859 4.07 6.5

2 P2 UV C19H25I2N3O11 709.9695 709.9702 0.99 7.5

3 P3 UV C14H18I3N3O6 705.8411 705.8402 1.28 5.5

4 P4 UV C18H21I2N3O9 677.9464 677.9440 3.54 8.5

5 P5 UV C18H21I2N3O10 675.9652 675.9647 0.74 8.5

6 AC1 UV C16H21I2N3O7 621.9553 621.9542 1.77 6.5

7 P6 UV C19H28IN3O11 602.0859 602.0841 2.99 6.5

8 P7 UV C14H19I2N3O6 579.9445 579.9436 1.55 5.5

1 P1 UV/H2O2 C19H27I2N3O10 711.9888 711.9859 4.07 6.5

2 P2 UV/H2O2 C19H25I2N3O11 709.9695 709.9702 0.99 7.5

3 P3 UV/H2O2 C14H18I3N3O6 705.8411 705.8402 1.28 5.5

4 BC1 UV/H2O2 C19H25I2N3O9 693.9762 693.9753 0.90 7.5

5 BC2 UV/H2O2 C18H23I2N3O9 679.9585 679.9596 0.90 7.5

6 P4 UV/H2O2 C18H21I2N3O9 677.9464 677.9440 3.54 8.5

7 P5 UV/H2O2 C18H21I2N3O10 675.9652 675.9647 0.74 8.5

8 P6 UV/H2O2 C19H28IN3O11 602.0859 602.0841 2.99 6.5

9 B1 UV/H2O2 C19H26IN3O11 600.0715 600.0684 5.17 13

10 P7 UV/H2O2 C14H19I2N3O6 579.9445 579.9436 1.55 5.5

11 B2 UV/H2O2 C18H24IN3O9 554.0636 554.063 1.08 7.5

12 B3 UV/H2O2 C19H27N3O11 474.1726 474.1718 1.69 7.5

13 B4 UV/H2O2 C18H25N3O10 444.1617 444.1613 0.90 7.5

1 C1 UV/Fenton C16H20I3N3O7 747.8510 747.8508 0.27 6.5

2 C2 UV/Fenton C19H26I2N3O11 725.9645 725.9651 0.83 7.5

3 P1 UV/Fenton C19H27I2N3O10 711.9888 711.9859 4.07 6.5

4 P2 UV/Fenton C19H25I2N3O11 709.9695 709.9702 0.99 7.5

5 P3 UV/Fenton C14H18I3N3O6 705.8411 705.8402 1.28 5.5

6 BC1 UV/Fenton C19H25I2N3O9 693.9762 693.9753 0.90 7.5

7 BC2 UV/Fenton C18H23I2N3O9 679.9585 679.9596 0.90 7.5

8 P4 UV/Fenton C18H21I2N3O9 677.9464 677.9440 3.54 8.5

9 P5 UV/Fenton C18H21I2N3O10 675.9652 675.9647 0.74 8.5

10 C3 UV/Fenton C16H21I2N3O8 637.9501 637.9491 1.57 6.5

11 AC1 UV/Fenton C16H21I2N3O7 621.9553 621.9542 1.77 6.5

12 P6 UV/Fenton C19H28IN3O11 602.0859 602.0841 2.99 6.5

259

13 P7 UV/Fenton C14H19I2N3O6 579.9445 579.9436 1.55 5.5

14 C4 UV/Fenton C14H21IN2O7 456.0416 456.0388 6.14 5

15 C5 UV/Fenton C11H14IN3O8 442.9820 442.982 0.90 6

16 C6 UV/Fenton C17H23N3O9 414.1519 414.1507 0.90 7.5

Table S2.2. Mass spectrum characteristics of Iohexol products identified in degradation studies after 15 min


1 1 UV C19H27I2N3O10 711.9888 711.9859 4.07 6.5

2 2 UV C19H25I2N3O11 709.9695 709.9702 0.99 7.5

3 3 UV C19H23I2N3O12 707.9576 707.9546 4.24 8.5

4 4 UV C14H18I3N3O6 705.8411 705.8402 1.28 5.5

5 5 UV C19H25I2N3O9 693.9762 693.9753 1.30 7.5

6 6 UV C18H21I2N3O9 677.9464 677.9440 3.54 8.5

7 7 UV C17H25I2N3O9 669.9750 669.9753 0.45 5.5

8 8 UV C17H23I2N3O8 651.9648 651.9647 0.15 6.5

9 9 UV C16H21I2N3O7 621.9553 621.9542 1.77 6.5

10 10 UV C19H28IN3O11 602.0859 602.0841 2.99 6.5

11 11 UV C14H19I2N3O6 579.9445 579.9436 1.55 5.5

12 12 UV C18H24IN3O9 554.0636 554.063 1.08 7.5

13 13 UV C19H27N3O11 474.1726 474.1718 1.69 7.5

14 14 UV C19H25N3O10 456.1614 456.1613 0.22 8.5

15 15 UV C14H20IN3O6 454.0473 454.047 0.66 5.5

16 16 UV C18H25N3O10 444.1617 444.1613 0.90 7.5

1 17 UV/H2O2 C19H16IN3O15 652.9575 652.9621 7.04 13

2 18 UV/H2O2 C18H16IN3O16 640.9571 640.9621 7.80 12

3 20 UV/H2O2 C13H14IN3O11 514.9654 514.9667 2.52 8

1 21 UV/Fenton C17H24I3N3O8 779.8792 779.8770 0.90 5.5

2 22 UV/Fenton C14H18I3N3O7 721.8367 721.8352 0.90 5.5

3 4 UV/Fenton C14H18I3N3O6 705.8411 705.8402 0.90 5.5

4 23 UV/Fenton C19H25I2N3O9 693.9762 693.9753 0.90 7.5

5 9 UV/Fenton C16H21I2N3O7 621.9553 621.9542 0.90 6.5

6 10 UV/Fenton C19H28IN3O11 602.0859 602.0841 0.90 6.5

7 24 UV/Fenton C14H19I2N3O7 595.9375 595.9385 0.90 5.5

8 11 UV/Fenton C14H19I2N3O6 579.9445 579.9436 0.90 5.5

9 25 UV/Fenton C11H14IN3O8 442.9820 442.982 0.90 6

260

10 26 UV/Fenton C17H23N3O9 414.1519 414.1507 0.90 7.5

11 27 UV/Fenton C13H16N2O11 377.0846 377.0827 0.90 6.5

appm - Mass Accuracy in ppm bDBE - Double Bond Equivalent A – Exclusive products from UV treatment B – Exclusive products from UV/H2O2 treatment C – Exclusive products from UV/Fenton treatment P - Products common for all three treatment methods AC - Products common for UV (A) and UV/Fenton (C) treatment BC - Products common for UV/H2O2 (B) and UV/Fenton (C) treatment

261

Appendix F: Supplementary material of Chapter 7

Figure S1 – Summary of the UV-C photolysis experiments for the two different energy output systems.

Absorbance spectra changes during the UV exposure.

225 250 275 300 325 350 375 4000.0

0.5

1.0

1.5

2.0

2.5

3.0

Abs

orba

nce

(a.u

.)

Wavelength (nm)

0' 15' 30' 60' 120' 180' 240'

Absorbance spectra during UV treatment

262

Figure S2 – UV/H2O2 Advanced Oxidation of Venlafaxine: global analyses. a) COD removal, TOC

reduction (red traces and axis) and H2O2 consumption (blue trace and axis) with addition of 50 mg/L

H2O2. b) Absorbance changes during UV/H2O2 treatment (50 mg/L H2O2).

0 30 60 90 120 150 180 210 240

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

COD TOC H2O2

Time (min)

CO

D/C

OD

0 & T

OC

/TO

C0

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

(H2O

2)/(H

2O2) 0

11 W

225 250 275 300 325 350 375 4000.0

0.5

1.0

1.5

2.0

2.5

3.0

50 ppm H2O2

Abs

orba

nce

(a.u

.)

Wavelength (nm)

0' 11' 30' 60' 90' 120' 150' 180'

Absorbance spectraduring UV/H2O2 treatment

11 W

264

Figure S3 – Treatment of Venlafaxine by the Fenton process in the dark. a) VFA degradation at pH=3 and varied Fe|H2O2 ratio. b) VFA degradation at pH=5 and varied Fe|H2O2 ratio. c) VFA degradation

at pH=7 and varied Fe|H2O2 ratio

265

Figure S4 – Treatment of Venlafaxine by the Fenton process in the dark. a) COD reduction by the Fenton process in by the various Fe|H2O2 ratios at pH=3 and increasing pH for 20|50 ratio. b) TOC

removal by the Fenton process in by the various Fe|H2O2 ratios at pH=3 and increasing pH for 20|50 ratio.

267

Figure S5 - Treatment of Venlafaxine by the photo-Fenton process (irradiance: 900 W/m2). a) VFA

degradation at pH=3 and varied Fe|H2O2 ratio. b) VFA degradation at pH=5 and varied Fe|H2O2 ratio.

c) VFA degradation at pH=7 and varied Fe|H2O2 ratio.

268

Figure S6 - Treatment of Venlafaxine by the photo-Fenton process (irradiance: 900 W/m2). a) COD

reduction by the photo-Fenton process by the various Fe|H2O2 ratios at pH=3 and increasing pH for

20|50 ratio. b) TOC removal by the photo-Fenton process by the various Fe|H2O2 ratios at pH=3 and

increasing pH for 20|50 ratio.

269

Supplementary Tables

Table S1 – H2O2 evolution during the Fenton and photo-Fenton experiments in MQ water

Fenton photo-Fenton pH= 3 pH= 3

Sample ID

initial [H2O2]

H2O2 consumption in 24 hours [ppm]

Sample ID

initial [H2O2]

H2O2 consumption during the experiment [ppm]

5|10 10 4.83 5|10 10 5.32 5|50 50 40.74 5|50 50 48.15

20|50 50 28.34 20|50 50 37.27 12.5|5

8.3 58.3 48.3 12.5|58.3 58.3 54.13

12.5|30 30 23.62 12.5|3

0 30 23.24

pH= 5 pH= 5

Sample ID [H2O2]


Sample ID [H2O2]


5|10 10 1.28 5|10 10 6.5 5|50 50 15.11 5|50 50 45.25

20|50 50 21.27 20|50 50 39.97 12.5|5

8.3 58.3 26.24 12.5|58.3 58.3 49.26

12.5|30 30 6.44 12.5|3

0 30 20.79

pH= 7 pH= 7

Sample ID [H2O2]


Sample ID [H2O2]


5|10 10 0.39408867 5|10 10 6.3 5|50 50 11.5 5|50 50 33.9

20|50 50 14.02 20|50 50 26.3 12.5|5

8.3 58.3 17.94 12.5|58.3 58.3 35.55

12.5|30 30 4.14 12.5|3

0 30 20.9

270

Table S2 – Basic physicochemical characteristics of the used effluents in the study.

Parameter Unit Wastewater previously treated with Activated Moving Bed Coagulation

Sludge Bioreactor Flocculation pH - 7.3-7.8 6.6-7.4 7.3-7.9

TOC mg/L 28.08±12.62 14.615±7.9 68.47±15.94 COD mg/L 51±10 20±11 85±5

Alkalinity mg CaCO3/L 230±35 95±10 240±10 TSS mg/L 12.1±2.8 14.2±1.4 28.5±5.7

Total iron mg Fe/L 0.95±0.05 1.75±0.15 5.5±1

271

Table S3 – Basic average physicochemical characteristics of the used urine in the study.

Parameter Value unit

TDS 60 g/L

pH 6.5

COD 8.3 g/L

TOC 5 g/L

272

Table S4 - Mass spectrum characteristics of Venlafaxine's products identified in degradation studies


1 P6 Solar C17H25NO 260.2016 260.2009 2.69 5.5

1 P3 UV C11H15NO2 194.1179 194.1175 2.06 4.5

2 P5 UV C12H20O4 229.1429 229.1434 2.18 2.5

3 P6 UV C17H25NO 260.2012 260.2009 1.15 5.5

4 P8 UV C17H27NO3 294.2057 294.2064 2.38 4.5

5 P9 UV C17H27NO4 310.2021 310.2013 2.58 4.5

6 P10 UV C17H27NO5 326.1956 326.1962 1.84 4.5

1 P1 UV/H2O2 C8H8O 121.0652 121.0648 3.30 4.5

2 P4 UV/H2O2 C15H18O 215.1431 215.1430 0.46 6.5

3 P6 UV/H2O2 C17H25NO 260.2017 260.2009 3.07 5.5

4 P8 UV/H2O2 C17H27NO3 294.2072 294.2064 2.72 4.5

5 P9 UV/H2O2 C17H27NO4 310.202 310.2013 2.26 4.5

1 P1 Fenton C8H8O 121.0653 121.0648 4.13 4.5

2 P2 Fenton C11H15NO 178.1231 178.1226 2.81 4.5

3 P3 Fenton C11H15NO2 194.1181 194.1175 3.09 4.5

4 P4 Fenton C15H18O 215.1433 215.1430 1.39 6.5

5 P6 Fenton C17H25NO 260.2017 260.2009 3.07 5.5

6 P7 Fenton C17H25NO3 292.1914 292.1907 2.40 5.5

7 P8 Fenton C17H27NO3 294.2070 294.2064 2.04 4.5

1 P1 Photo-Fenton C8H8O 121.0654 121.0648 4.96 4.5

2 P2 Photo-Fenton C11H15NO 178.1231 178.1226 2.81 4.5

3 P3 Photo-Fenton C11H15NO2 194.1182 194.1175 3.61 4.5

4 P4 Photo-Fenton C15H18O 215.1434 215.1430 1.86 6.5

5 P6 Photo-Fenton C17H25NO 260.2018 260.2009 3.46 5.5



appm - mass accuracy in ppm bDBE - Double Bond Equivalent

273

Curriculum Vitae of the Candidate

Stefanos GIANNAKIS

Address: Av. Ed. Dapples 12 CH-1006, Lausanne Switzerland Tel: +41787690465 Email: [email protected] Web: https://www.researchgate.net/profile/Stefanos_Giannakis

EDUCATION

2009-2010 M.Sc. Environmental Protection and Sustainable development Aristotle University of Thessaloniki, Greece School of Engineering, Department of Civil Engineering Grade: 9.15, Excellent (top 3% for academic year 2010-2011) Thesis Grade 9.95/10

2003-2009 Diploma (M.Sc. Equivalent) Aristotle University of Thessaloniki, Greece School of Engineering, Department of Civil Engineering Double major: Civil Engineering, Hydraulics and Environmental Engineering Grade: 7.26, Very good Thesis Grade 10/10

AWARDS AND SCHOLARSHIPS

2012-2013 “Swiss Government Excellence Scholarships” Type: Research Fellowship Place: École Polytechnique Fédérale de Lausanne (EPFL), Switzerland Department: Institute of Chemical Sciences and Engineering Research Domain: Environmental Engineering

2011-2012 “Mediterranean Office for Youth – Mobility Grant” Type: Research Fellowship Place: Universitat Politècnica de Catalunya (UPC), Catalonia, Spain Department: Institut d'Investigació Tèxtil i Cooperació Industrial de Terrassa (INTEXTER) Research Domain: Environmental Engineering

RESEARCH AND TEACHING EXPERIENCE

Teaching

274

2015-2016 Chemical Technology and Biology of the Environment, EPFL Teaching Assistant Professors: Cesar Pulgarin

2014-2015 Chemical Technology and Biology of the Environment, EPFL Teaching Assistant Professors: Christos Comninellis, Cesar Pulgarin

2014-2015 Chemical engineering lab & project, EPFL Laboratory Assistant Professors: Liubov Kiwi, Kevin Sivula, Jeremy Luterbacher

2013-2014 Chemical engineering lab & project, EPFL Laboratory Assistant Professors: Liubov Kiwi

2010-2011 Environmental Chemistry, AUTh Laboratory assistant Professors: Darakas Efthymios, Aikaterini Tasoula

2010-2011 Environmental Engineering, AUTh Laboratory assistant Professors: Darakas Efthymios, Aikaterini Tasoula

Student Projects and Theses

Student Year Level Origin/University

Ana Isabel Merino Gamo 2013 MSc Erasmus (University of Barcelona, Spain)

Simon Schindelholz 2014 MSc EPFL, Lausanne, Switzerland

Miquel Pastor Gelabert 2014 MSc Erasmus (University of Barcelona, Spain)

Angelica Varon Lopez 2014 PhD/Trainee University of Cali, Colombia

David Muzard 2014 MSc GPAO, EPFL, Lausanne, Swizerland

Lula Dind 2014 Trainee GPAO, EPFL, Lausanne, Swizerland

Idriss Hendaoui 2014 Trainee GPAO, EPFL, Lausanne, Swizerland

Idriss Hendaoui 2014 MSc GPAO, EPFL, Lausanne, Swizerland

Franco Alejandro Gamarra Vives 2014 MSc EPFL, Lausanne, Switzerland

Barbara Androulaki 2014 Trainee University of Patras, Greece

Margaux Voumard 2014 Project EPFL, Lausanne, Switzerland

Siting Liu 2015 MSc GPAO, EPFL, Lausanne, Swizerland

Siting Liu 2015 Trainee GPAO, EPFL, Lausanne, Swizerland

Samuel Watts 2015 MSc EPFL, Lausanne, Switzerland

Samuel Watts 2015 Trainee GPAO, EPFL, Lausanne, Swizerland

Cristian Pinilla 2015 Trainee GPAO, EPFL, Lausanne, Swizerland

275

Paola Villegas Guzman 2015 PhD/Trainee University of Antioquia, Colombia

Margaux Voumard 2015 MSc EPFL, Lausanne, Switzerland

Margaux Voumard 2016 Trainee GPAO, EPFL, Lausanne, Swizerland

Pilar Valero 2016 PhD/Trainee University of Zaragoza, Spain

Casto Ramos 2016 Trainee EPFL, Lausanne, Switzerland

Cristian Pinilla 2016 MSc Co-supervision (University of Geneva, Switzerland)

Projects

2014-2016 Contributor Member Swiss National Foundation Project (No. 146919) “Treatment of the hospital wastewaters in Côte d'Ivoire and in Colombia by advanced oxidation processes”

2014-2015 Contributor Member Swiss-Hungarian Cooperation Program, No. SH7/2/14 “Towards a sustainable fine chemical and pharmaceutical industry: screening and re-utilisation of carbon-rich liquid wastes”

2013-2014 Contributor Member FP7 LIMPID project (Project No.: 310177) “Nanocomposite materials for photocatalytic degradation of pollutants”

TRAINING COURSES – LIFELONG EDUCATION

2016 COSEC Responsible for Laboratory Safety and Security, GPAO, EPFL

2016 Participant CM1305 ECOSTBio Summer School, Groningen, Netherlands, 8-16 July, 2016

2015 Participant 1st AOPs PhD Summer School, Salerno (Fisciano), Italy, June 15-19, 2015

2015 Participant Swiss Program for Research on Global Issues and Development, Filzbach, Switzerland, 18-20 March, 2015

2015 Participant Teaching Toolkit Workshop, EPFL, 19 May 2015, Lausanne, Switzerland

2014-2015 Participant Introduction to University Teaching, EPFL, completed: 29 June 2015, Lausanne, Switzerland

2013 Participant Mediterranean Office for Youth-Campus France, “Mediterranean Programs Forum”, 20 June 2013, Marseille, France,

2012 Participant 18th European Students Symposium on the Environment “Conservation is not isolation”, 19-26 May 2012, Bratislava, Slovakia

2011 Participant

276

17th European Students Symposium on the Environment “Biodiversity and Urbanism”, 20-28 May 2011, Zagreb, Croatia

2011 Participant 4th Summer School "Climate change in cities and city regions – Time to adapt?", 26–30 September 2011, Hamburg, Germany

2011 Participant Full scholarship for participation in the 4th Summer School "Environmental Decision Support Systems (EDSS): A Tool for Wastewater Management in the XXI Century", 3-9 July, Girona, Spain

2011 Participant 2nd Training Course “Water Supply in a Changing Environment”, Aristotle University of Thessaloniki, 11-15/4/2011, Thessaloniki, Greece

PUBLICATIONS

Peer-reviewed Journals

1. Mangayayam, M., Kiwi, J., Giannakis, S., Pulgarin, C., Zivkovic, I., Magrez, A. and Rtimi, S. (2017) FeOx magnetization enhancing E. coli inactivation by orders of magnitude on Ag-TiO2 nanotubes under sunlight. Applied Catalysis B: Environmental 202, 438-445.

2. Giannakis, S., Hendaoui, I., Jovic, M., Grandjean, D., De Alencastro, L.F., Girault, H. and Pulgarin, C. (2017) Solar photo-Fenton and UV/H2O2 processes against the antidepressant Venlafaxine in urban wastewaters and human urine. Intermediates formation and biodegradability assessment. Chemical Engineering Journal 308, 492-504.

3. Giannakis, S., Jovic, M., Gasilova, N., Pastor Gelabert, M., Schindelholz, S., Furbringer, J.M., Girault, H., Pulgarin, C., Iohexol degradation in wastewater and urine by UV-based Advanced Oxidation Processes (AOPs): Process modeling and by-products identification." Journal of Environmental Management (2016), In press, Accepted Manuscript

4. Giannakis, S., Polo López, M.I., Spuhler, D., Sánchez Pérez, J.A., Fernández Ibáñez, P., Pulgarin, C. Solar disinfection is an augmentable, in situ-generated photo-Fenton reaction—Part 1: A review of the mechanisms and the fundamental aspects of the process (2016) Applied Catalysis B: Environmental, 199, pp. 199-223.

5. Giannakis, S., López, M.I.P., Spuhler, D., Pérez, J.A.S., Ibáñez, P.F., Pulgarin, C. Solar disinfection is an augmentable, in situ-generated photo-Fenton reaction-Part 2: A review of the applications for drinking water and wastewater disinfection (2016) Applied Catalysis B: Environmental, 198, pp. 431-446.

6. Giannakis, S., Voumard, M., Grandjean, D., Magnet, A., De Alencastro, L.F., Pulgarin, C. Micropollutant degradation, bacterial inactivation and regrowth risk in wastewater effluents: Influence of the secondary (pre)treatment on the efficiency of Advanced Oxidation Processes (2016) Water Research, 102, pp. 505-515.

277

7. Rtimi, S., Giannakis, S., Bensimon, M., Pulgarin, C., Sanjines, R., Kiwi, J. Supported TiO2 films deposited at different energies: Implications of the surface compactness on the catalytic kinetics (2016) Applied Catalysis B: Environmental, 191, pp. 42-52.

8. Giannakis, S., Ruales-Lonfat, C., Rtimi, S., Thabet, S., Cotton, P., Pulgarin, C. Castles fall from inside: Evidence for dominant internal photo-catalytic mechanisms during treatment of Saccharomyces cerevisiae by photo-Fenton at near-neutral pH (2016) Applied Catalysis B: Environmental, 185, pp. 150-162.

9. Rtimi, S., Giannakis, S., Sanjines, R., Pulgarin, C., Bensimon, M., Kiwi, J. Insight on the photocatalytic bacterial inactivation by co-sputtered TiO2-Cu in aerobic and anaerobic conditions (2016) Applied Catalysis B: Environmental, 182, pp. 277-285.

10. Giannakis, S., Darakas, E., Escalas-Cañellas, A., Pulgarin, C. Solar disinfection modeling and post-irradiation response of Escherichia coli in wastewater (2015) Chemical Engineering Journal, 281, pp. 588-598.

11. Giannakis, S., Gamarra Vives, F.A., Grandjean, D., Magnet, A., De Alencastro, L.F., Pulgarin, C. Effect of advanced oxidation processes on the micropollutants and the effluent organic matter contained in municipal wastewater previously treated by three different secondary methods (2015) Water Research, 84, pp. 295-306.

12. Giannakis, S., Darakas, E., Escalas-Cañellas, A., Pulgarin, C. Temperature-dependent change of light dose effects on E. coli inactivation during simulated solar treatment of secondary effluent (2015) Chemical Engineering Science, 126, pp. 483-487.

13. Giannakis, S., Darakas, E., Escalas-Cañellas, A., Pulgarin, C. Environmental considerations on solar disinfection of wastewater and the subsequent bacterial (re)growth (2015) Photochemical and Photobiological Sciences, 14 (3), pp. 618-625.

14. Giannakis, S., Rtimi, S., Darakas, E., Escalas-Cañellas, A., Pulgarin, C. Light wavelength-dependent E. coli survival changes after simulated solar disinfection of secondary effluent (2015) Photochemical and Photobiological Sciences, 14 (12), pp. 2238-2250.

15. Giannakis, S., Papoutsakis, S., Darakas, E., Escalas-Cañellas, A., Pétrier, C., Pulgarin, C. Ultrasound enhancement of near-neutral photo-Fenton for effective E. coli inactivation in wastewater (2015) Ultrasonics Sonochemistry, 22, pp. 515-526.

16. Giannakis, S., Merino Gamo, A.I., Darakas, E., Escalas-Cañellas, A., Pulgarin, C. Monitoring the post-irradiation E. coli survival patterns in environmental water matrices: Implications in handling solar disinfected wastewater (2014) Chemical Engineering Journal, 253, pp. 366-376.

17. Giannakis, S., Darakas, E., Escalas-Cañellas, A., Pulgarin, C. Elucidating bacterial regrowth: Effect of disinfection conditions in dark storage of solar treated secondary effluent (2014) Journal of Photochemistry and Photobiology A: Chemistry, 290 (1), pp. 43-53.

278

18. Giannakis, S., Darakas, E., Escalas-Cañellas, A., Pulgarin, C. The antagonistic and synergistic effects of temperature during solar disinfection of synthetic secondary effluent (2014) Journal of Photochemistry and Photobiology A: Chemistry, 280, pp. 14-26.

19. Giannakis, S., Merino Gamo, A.I., Darakas, E., Escalas-Cañellas, A., Pulgarin, C. Impact of different light intermittence regimes on bacteria during simulated solar treatment of secondary effluent: Implications of the inserted dark periods (2013) Solar Energy, 98 (PC), pp. 572-581.

Conferences

1. Near-Neutral Photo-Fenton And The Differentiated Response Of Microbiological And Chemical Contaminants In Wastewater, César Pulgarin, Stefanos Giannakis, SPEA9, 13-17/6, Strasbourg, France (Oral Presentation)

2. Operational And Economical Optimization Of Venlafaxine Treatment By The Photo-Fenton Reaction Through Response Surface Methodology And Desirability Functions, Stefanos Giannakis, Idriss Hendaoui, Jean-Marie Furbringer, César Pulgarin, SPEA9, 13-17/6, Strasbourg, France (Poster Presentation)

3. New insight on the bacterial inactivation by Co-sputtered TiO2-Cu in aerobic and anaerobic media under low intensity actinic light, S. Rtimi, S. Giannakis, C. Pulgarin, J. Kiwi, COST MP1106 Symposium “Smart and Green Interfaces: Fundamentals and Diagnostics (SGI-FunD 2015)”, Sofia, Bulgaria, 29 – 31 October 2015

4. Comparative Evolution Between Oxidation Processes Used For Bacterial Inactivation After Three Different Secondary Treatment Methods, C. Pulgarin, M. Voumard, S. Giannakis – EAAOP4, Athens, October 2015 (Oral presentation)

5. Optimization And Modeling Of Iohexol Treatment By Advanced Oxidation Processes In Environmentally Relevant Matrices, S. Giannakis, M. Pastor Gelabert, S. Schindelholz, J.M. Furbringer, C. Pulgarin – EAAOP4, Athens, October 2015 (Poster Presentation)

6. New Pathways In Heterogeneous And Homogeneous Near-Neutral Photo-Fenton Bacterial Inactivation By Iron Oxides And Iron Citrate Complexes, S. Giannakis, C. Ruales-Lonfat, C. Pulgarin – EAAOP4, Athens, October 2015 (Poster Presentation)

7. Degradación por UV254/H2O2 de contaminantes emergentes en efluentes de PTAR domesticas: desde el laboratorio hasta el piloto final. C. Pulgarin, S. Giannakis, Primer congreso Colombiano de Procesos avanzados de Oxidación, Manizales, Colombia, 21-24 September 2015

8. Emergent chemical and microbiological pollutants related with hospital wastewater in Colombia, Ivory Coast and Switzerland, S. Giannakis, C. Pulgarin. Invited talk. Simposio ACIS 2015: ¿Hasta cuándo tendremos agua en Colombia? September 11th–12th, 2015, Geneva, Switzerland (Oral Presentation)

279

9. Internal photo-Fenton leads the solar inactivation of Saccharomyces cerevisiae in hv/H2O2/Fe systems at neutral pH, S. Giannakis, C. Ruales-Lonfat, C. Pulgarin, Swiss Chemical Society Photochemistry Section Annual Meeting 2015, September 8, 2015 - ETH Zurich, Switzerland (Oral Presentation)

10. Environmental Considerations On Solar Disinfection Of Wastewater And The Subsequent Bacterial (Re)growth, S. Giannakis, E. Darakas, A. Escalas-Cañellas, C. Pulgarin, SPEA8, Thessaloniki, Greece, 25-28 June 2014 (Poster presentation)

11. Usage of solar radiation and artificially induced UV irradiation for water and wastewater disinfection, S. Giannakis and E. Darakas, 2012., International Conference on Protection and Restoration of the Environment, PRE XI, Thessaloniki, Greece, July 3-6.

12. Investigation of the environmental impact of Thessaloniki’s main streams on the coastal pollution of Thermaikos Gulf, S. Giannakis and E. Darakas, 2011, 4th Environmental Conference of Macedonia, Thessaloniki, Greece, March 18-20.

13. Statistical analysis of fecal indicator bacteria kinetics based on the initial concentration of the population and the dilution rate, S. Giannakis, E. Darakas and M. Vafeiadis, 2010, 2nd COST 929 Symposium: Future Challenges in food and environmental virology, Istanbul, Turkey, October 7-9.

14. Effects of extreme weather conditions in natural wastewater systems, S. Giannakis and E. Darakas, 2010, 3rd Conference: Small and Decentralized Water and Wastewater Treatment Plants, Skiathos, Greece, May 14-16.

LANGUAGE SKILLS

Language Level (Common European Framework)

Reading Writing Speaking Greek C2 C2 C2

English C2 C2 C2 French Β2 Β2 Β2 Spanish Β2 Β1 Β1 Catalan Α2 Α2 Α2 German Α2 Α1 Α1

SCIENTIFIC COMMITTEE & PROFESSIONAL/ACADEMIC MEMBERSHIPS

AOPs PhD School: Student Representative

Journal Referee: Applied Catalysis B: Environmental, Chemical Engineering Journal, Water Research, Environmental Science and Pollution Research, Water Air and Soil Pollution, Journal of Environmental Management, Process Safety and Environmental Protection, Photochemical and Photobiological Sciences, Journal of Photochemistry and Photobiology A: Chemistry, Desalination and water treatment, Water Science & Technology, RSC Advances

use of light-supported oxidation processes towards microbiological and chemical contaminants

Documents