Towards a renewable, reliable and robust electrochemical sensing principle for arsenic(III) detection in environmental freshwater systems

(1)

Thesis

Reference

Towards a renewable, reliable and robust electrochemical sensing principle for arsenic(III) detection in environmental freshwater

systems

TOUILLOUX, Romain Yannick Claude

Abstract

The first aim of this work has been to develop an iridium-based microelectrode plated with a renewable gold nanoparticles coating. This system showed good reproducibility and reliability at the nanomolar concentration level for repeated gold film renewal and for the long term As(III) monitoring at a single film (7 d). In the second hand, the microelectrode covered by an inert antifouling membrane made of 1.5% LGL agarose was characterized for the analysis of As(III). The membrane demonstrated an ability to eliminate fouling by fulvic acid and inorganic colloid particles by size exclusion. Subsequently, this system was applied to the measurement of non-polluted freshwater samples from Lake Greifen (ZH, Switzerland). Finally, the concept of ion exchange nanospheres (IENS) as heterogeneous pH buffer incorporated in the antifouling membrane was put forward. Microelectrodes covered with an IENS-doped antifouling membrane demonstrated the ability to eliminate the pH interference for As(III) in the natural range of pH and are a promising direction for future in situ applications in environmental systems.

TOUILLOUX, Romain Yannick Claude. Towards a renewable, reliable and robust electrochemical sensing principle for arsenic(III) detection in environmental freshwater systems. Thèse de doctorat : Univ. Genève, 2015, no. Sc. 4897

URN : urn:nbn:ch:unige-815712

DOI : 10.13097/archive-ouverte/unige:81571

Available at:

http://archive-ouverte.unige.ch/unige:81571

Disclaimer: layout of this document may differ from the published version.

(2)

UNIVERSITÉ DE GENÈVE Section de chimie et biochimie Département de chimie minérale et analytique

FACULTÉ DES SCIENCES Prof. Eric Bakker

Towards a Renewable, Reliable and Robust Electrochemical Sensing Principle for Arsenic(III) Detection in Environmental

Freshwater Systems

THÈSE

Présentée à la Faculté des sciences de l’Université de Genève Pour obtenir le grade de Docteur ès sciences, mention chimie

par

Romain TOUILLOUX de

Fribourg (Fribourg)

Thèse N° 4897 GENÈVE

(3)

(4)

Acknowledgements

The completion of my dissertation and subsequent Ph.D. has been a long journey. It’s true that “Life is what happens” when you are completing your dissertation. Life doesn’t stand still, nor wait until you are finished and have time to manage it. Much as happened and changed at the time I’ve been involved in this project.

First of all, I would like to express my deepest gratitude to my advisor, Prof. Dr. Eric Bakker.

I have been amazingly fortunate to have an advisor who gave me the freedom to explore on my own and at the same time the guidance to recover when my steps faltered. Eric taught me how to question thoughts and express ideas. His patience and support helped me overcome many difficult situations. I hope that one day I would become an advisor as good as Eric has been to me.

I am grateful to Mary-Lou Tercier-Waeber for the discussions that helped me sort out the technical and scientific details of my work. I am also thankful to her for encouraging the use of consistent notation in my writings.

I am very grateful to the members of my dissertation committee, Prof. Dr. Thomas Bürgi and Prof Dr. Jörg Schäfer. Their academic support is greatly appreciated. Thank you.

I am indebted to my colleagues for supporting me all these years. Many thanks for all these good moments to Dr. Bastien Néel, Dr. Matthieu Masson, Dr. Sandrine Mongin, Zdenka Jarolimova, Dr. Xiaojiang Xie, Jingying Zhai, Miquel Coll Crespi, Nadja Pankratova, Thomas Cherubini, Agustin Gutierrez, Dr. Denis Dorokhin, Anne-Marie Loup, Stephane Jeanneret, Serge Rodak, Dr. Plinio Maroni and Olivier Vassalli.

Of course no acknowledgments would be complete without giving thanks to my mother. She has instilled many admirable qualities in me and given me a good foundation with which to meet life. I would like also to acknowledge my younger brother, Brice, who eventually has become a doctor one year before me.

Last, but certainly not least, I must acknowledge with tremendous and deep thanks my partner, Dr. Tanja Jürgens. Through her love, patience, support and unwavering belief in me, I’ve been able to complete this long dissertation journey. She is my biggest fan and supporter.

(5)

Abstract

Water is an important natural resource as it is used every day in domestic and industrial supplies and is of growing demand for agricultural production. With increasing exploitation of water resources, a wide range of problems related to its availability and quality has emerged during the past three decades. The presence of inorganic contaminants from natural and anthropogenic sources threatens the suitability of water for human consumption and for maintaining a sustainable development. In natural water, inorganic arsenic is present as oxyanions and exists in two valence states, As(III), the most toxic form, and As(V). Under moderately reducing conditions, As(III) may be the dominant form, but As(V) is generally the stable oxidation state in oxic water. Elevated As(III) concentrations in drinking water have been observed in many regions of the world.

Monitoring of As(III) at appropriate spatial and temporal states is of special importance due to its carcinogenic properties. For this reason, extensive research has been carried out to develop suitable monitoring methods. Voltammetry is one of a few analytical methods suitable for the on-line monitoring of trace metals and is in principle capable of detection concentrations in the low subnanomolar range. Nevertheless, in the case of As(III), only a limited number of systems can be used at natural pH and their robustness for environmental applications is not well demonstrated. Moreover, most research is focused on macroelectrodes, which is less applicable to field deployment compared to devices incorporating microelectrodes.

The first aim of this work has been to develop an iridium-based microelectrode plated with a renewable gold nanoparticles coating. This system showed good reproducibility and reliability at the nanomolar concentration level for repeated gold film renewal and for the long term As(III) monitoring at a single film (7 d). In the second hand, the microelectrode covered by an inert antifouling membrane made of 1.5% LGL agarose was characterized for the analysis of As(III).

The membrane demonstrated an ability to eliminate fouling by fulvic acid and inorganic colloid particles by size exclusion. Subsequently, this system was applied to the measurement of non- polluted freshwater samples from Lake Greifen (ZH, Switzerland). Finally, the concept of ion exchange nanospheres (IENS) as heterogeneous pH buffer incorporated in the antifouling membrane was put forward. Microelectrodes covered with an IENS-doped antifouling membrane demonstrated the ability to eliminate the pH interference for As(III) in the natural range of pH and are a promising direction for future in situ applications in environmental systems.

The introduction of an innovative antifouling membrane integrated with a renewable gold- nanoparticle plated iridium-based microelectrode made the use of voltammetry at microelectrodes suitable for monitoring As(III) in freshwater without the need for chemical modification of the samples.

(6)

Résumé

L’eau constitue une ressource naturelle importante, car elle est utilisée tous les jours comme source de l’approvisionnement domestique et industriel ainsi que de la demande croissante pour la production agricole. Avec l’augmentation de l’exploitation des ressources aquifères, un large éventail de problèmes liés à la disponibilité et la qualité a émergé au cours des trois dernières décennies. La présence de contaminants inorganiques provenant de sources naturelles et anthropiques menace l’usage de l’eau pour la consommation et pour le maintien d’un développement durable. Dans les eaux naturelles, l’arsenic inorganique se trouve sous la forme d’oxyanions et existe dans deux états de valence, l’As(III), la forme la plus toxique, et l’As(V).

Dans des conditions réductrices modérées, l’As(III) peut être la forme dominante, mais l’As(V) est généralement l’état d’oxydation stable dans les eaux oxiques. Des concentrations élevées d’As(III) dans l’eau potable ont été observées dans de nombreuses régions du monde.

La surveillance de l’As(III) est d’une importance particulière en raison de ses propriétés cancérogènes. Pour cette raison, des recherches approfondies ont été menées pour développer des méthodes de surveillance appropriées. La voltampérométrie est l’une des quelques méthodes analytiques appropriées pour le contrôle en continu des métaux traces et est en principe capable de détecter des concentrations de l’ordre du subnanomolaire. Néanmoins, dans le cas de l’As(III), seul un nombre limité de systèmes peut être utilisé à pH naturel et leurs robustesses pour les applications environnementales ne sont pas bien démontrées. En outre, la plupart des recherches se concentre sur des macroélectrodes, qui sont moins applicables à un déploiement sur le terrain par rapport aux dispositifs incorporant des microélectrodes.

Le premier objectif de ce travail a été de développer une microélectrode d’iridium plaquée avec une couche renouvelable de nanoparticules d’or. Ce système a montré une bonne reproductibilité et une fiabilité à des concentrations de l’ordre du nanomolaire entre différents films renouvelés, ainsi qu’à long terme pour le contrôle de l’As(III) avec un seul film (7 jours).

Dans un deuxième temps, l’efficacité d’une membrane antifouling faite de 1.5% d’agarose recouvrant la microélectrode a été caractérisée pour l’analyse de l’arsenic. La membrane a montré une capacité à éliminer par exclusion de taille le fouling venant des acides fulviques et des colloïdes inorganiques particulaires. Par la suite, ce système a été appliqué pour des mesures d’échantillons d’eau non polluée du lac de Greifensee (ZH, Suisse). Enfin, le concept de nanosphères échangeuses d’ions (NEI) comme tampon de pH hétérogène incorporé dans la membrane antifouling a été mis en avant. Une microélectrode recouverte d’une membrane antifouling dopée par les NEI a démontré sa capacité à éliminer l’interférence du pH pour l’As(III) dans la plage de pH naturel. Cela constitue une voie prometteuse pour de futures applications in situ dans les systèmes environnementaux.

L’introduction d’une membrane antifouling novatrice intégrée à une microélectrode d’iridium plaqué avec une couche renouvelable de nanoparticules d’or a rendu l’utilisation de la voltamétrie sur des microélectrodes appropriée pour le contrôle de l’As(III) dans les eaux douces, sans la nécessité d’une modification chimiques des échantillons.

(7)

List of figures

Figure 1-1 A simplified, comprehensive cycle transfer of arsenic ... 16 Figure 1-2 Effect of arsenic exposure to humans due to As contaminated groundwater.^{19, 20} . 18 Figure 1-3 Known areas with natural arsenic contamination around the world.²⁵ (Reproduced

with permission of Royal Society of Chemistry via Copyright Clearance Center). ... 19 Figure 1-4 Areas with elevated concentrations of arsenic in Switzerland²⁶. (Reprinted with

permission from IWA Publishing) ... 20 Figure 1-5 Number of papers on environmental and chemical arsenic studies during each

year from 1980 to 2014. ... 25 Figure 1-6 Arsenic species in water ... 29 Figure 1-7 Mole fraction of total dissolved arsenic as a function of pH in water for (a) As(III)

and (b) As(V).¹⁰⁰ (Reproduced with permission of Springer via Copyright Clearance Center). ... 30 Figure 1-8 Eh-pH diagram of aqueous arsenic species for the As at 25°C.¹⁰⁰ (Reproduced

with permission of Springer via Copyright Clearance Center). ... 31 Figure 1-9 Arsenicosis patient with hyperkeratosis on the palm due to drinking groundwater

containing high concentration of inorganic arsenic, primarily As(III)¹²⁷ (Reproduced with permission from Elsevier via Copyright Clearance Center) ... 33 Figure 1-10 Electrochemical cell with the working (WE), counter (CE) and reference (RE)

electrodes. ... 34 Figure 1-11 Ag/AgCl reference electrode ... 36 Figure 1-12 Double junction reference electrode ... 37 Figure 1-13 Chronoamperometric experiment: (a) Potential applied to the cell versus time

where E1 is the initial potential and E2 the final potential; (b) Current response versus time. ... 38 Figure 1-14 Principle of anodic stripping voltammetry (ASV) – schematic depiction. ... 39 Figure 1-15 (a) Schematic representation of square wave pulse technique. Esw, the amplitude;

Es, the step potential and current measurement times, t1 and t2. (b) Current measured over time during SW. (c) Sampled current over potential during the forward and the reverse pulse along with the difference from the two values.(Reproduced from Bakker E.¹⁵²). ... 40 Figure 1-16 Schematic representation of the electrical double layer. ... 41 Figure 1-17 Schematic diagram illustrating the movement of a substance down its

concentration gradient. It is also illustrated the resulting change on the concentration when Flux in and Flux out are equal or different. ... 42 Figure 1-18 Number of papers on microelectrodes/ultramicroelectrodes in the field of

electrochemistry during each year from 1980 to 2014. ... 45 Figure 1-19 Types of microelectrodes: (a) ring-disk, (b) disk, (c) ring, (d) spherical (e)

hemispherical, (f) finite conical, (g) cylindrical and (h) microband. ... 46 Figure 2-1 Cyclic voltammetry in HNO3 (pH 2), 0.01 M NaNO3 on an IrM at 50 mV/s, with

initial and reversal potentials of -500 mV and +800 mV vs. Ag/AgCl, respectively. Edep

= -300 mV. ... 103

(11)

Figure 2-2 Consecutive chronoamperograms (1-4) of Au deposition on an IrM obtained before (a) and after (b) the application of ten consecutive pulses (+800 mV, 50 ms) to oxidize/desorb impurities at the Iridium microdisk substrate. Edep = -300 mV; solution:

1 mM Au(III), 0.1 M NaNO3, HNO3 pH 2. ... 104 Figure 2-3 Consecutive chronoamperograms (1-5) on an IrM with a gold deposition layer

obtained before (a) and after (b) the application of ten consecutive pulses (+800 mV, 50 ms) to oxidize/desorb impurities at the Iridium microdisk substrate. Edep = -300 mV;

Solution: 0.1 M NaNO3, HNO3 pH 2. ... 104 Figure 2-4 Scanning Electron Microscopy (SEM) image of an Au layer electroplated on a

IrM-arrays using the following conditions: Edep = -300 mV; solution: 1 mM Au(III), 0.1 M NaNO3, HNO3 pH2. ... 105 Figure 2-5 As(III) peak current at 50 nM As(III) obtained as a function of the indicated gold

layer thickness. Sample and SWASV conditions as in Figure 2-7. ... 106 Figure 2-6 Linear sweep voltammograms on Au-IrM resulting in complete reoxidation of

the Au. Parameters: Ei = +300 mV, Ef = +800 mV, Eincrement = 2.5 mV, scan rate E

= 5 mV.s-1. Solution: 5 mM Hg(CH3COO)2 in 1 M KSCN. ... 107 Figure 2-7 Calibration plot (n = 5) of stripping peak currents as a function of increasing As(

III ) concentration (0, 10, 20, 30, 40 and 50 nM) in 0.01 M NaNO3 buffered at pH 8 (10 mM phosphate buffer). Inset: corresponding calibration curves. SWASV conditions:

Eprecleaning = +500 mV (30 s); Eprec = -1000 mV; tprec = 3 min; E i = -1000 mV; Ef = +300 mV; f = 200 Hz; ESW = 25 mV; Es = 8 mV. ... 108 Figure 2-8 Calibration plot (n = 3) of stripping peak currents as a function of increasing

As(III) concentration (0, 1, 3, 5 and 10 nM), sample otherwise as in Figure 2-7. Inset:

corresponding calibration curves. SWASV conditions as in Figure 2-7, except tprec = 36 min. ... 108 Figure 2-9 Reproducibility of the As(III) stripping peak current for three different gold layers

(20 measurements for each layer). Sample: 50 nM As(III), otherwise as in Figure 2-7.

SWASV conditions as in Figure 2-7. ... 110 Figure 2-10 (a) Reproducibility of the As( III ) stripping peak currents of 3063 consecutive

measurements performed over a period of 7 days; (b) Superposition of ﬁrst (solid line) and 7 days later (dashed line) voltammograms. Sample and SWASV conditions as in Figure 2-9. ... 110 Figure 2-11 Lifetime comparison for consecutives replicates between a solid gold

microelectrode (circle) and the Au-IrM (square) for a concentration of 25 nM and 50 nM As (III) respectively. Sample and SWASV conditions as in Figure 2-7, except tprecleaning = 2 min for the solid gold microelectrode. ... 112 Figure 2-12 Two calibrations plot in presence of increasing concentration of As(III) (5, 10,

15, 20, 25, 30 nM) performed on a cleaned solid gold microelectrode in two different days. Sample and SWASV conditions as in Figure 2-7, except tprecleaning = 2 min. ... 112 Figure 2-13 SWASV peak currents obtained for 5 nM As(III) in the presence of increasing

concentrations of Cu(II) (0, 10, 30, 60 and 100 nM), average of 3 replicate measurements. Inset: corresponding curves. Sample and SWASV conditions as in Figure 2-8. ... 113

(12)

Figure 2-14 Calibration plots of stripping current as a function of increasing Cu(II) concentration (0, 10, 30, 60, 100 nM) in presence of 5 nM As(III) corresponding to Figure 2-13. Sample and SWASV conditions as in Figure 2-8. ... 114 Figure 2-15 SWASV peak currents obtained for 5 nM As(III) in the presence of 30 nM Cu(

II ) with (dashed line) and without (solid line) 0.6 M Cl^- on the Au-IrM (a) and a solid gold microelectrode (b). Sample and SWASV conditions as in Figure 2-8. ... 115 Figure 2-16 SWASV peak current obtained with the addition of 5 nM As(III) (tprec = 36 min)

(a) without (b) or with the presence of a 1.5% LGL agarose layer of 375 (±25) µm in unfiltered/filtered Arve river. The pH was fixed at 8 by a gas mix made of N2 and CO2

for the Arve river samples. SWASV conditions as in Figure 2-8. ... 116 Figure 3-1 Scheme of the gel integrated Gold nanoparticles-plated Iridium-based

microelectrode (Au-GIME). ... 125 Figure 3-2 Results of arsenic speciation analysis of spiked freshwater samples by proposed

separation method using Lewatit MonoPlus M500 and analyzed by ICP-MS (n=3) at different ratios of As(III):As(V) (0:100, 10:90, 90:10, 100:0 and 50:50 %). Comparison is made between expected (left column) and experimental values (right column). ... 126 Figure 3-3 (a) Experimental (symbols) and theoretical (in grey) diffusion profiles of As(III)

toward (open) or from (solid) the microelectrode surface, in 375 ± 25 µm (●), 500 ± 25 µm (■) and 625 ± 25 µm (▲) thick 1.5% LGL agarose membrane (b) Gel equilibration time is plotted in function of (gel thickness)2 from Table 3-2. Sample: 50 nM As(III) in 10 mM phosphate, 0.01 M NaNO3, pH 8.0. SWASV conditions: Eprecleaning = +500 mV (10s); Eprec = -1.0 V; tprec = 110 sec; Ei = -1.0 V; Ef = +0.3 V; f = 200 Hz; ESW = 25 mV; Es = 8 mV. ... 128 Figure 3-4 As(III) calibration curves (n = 3) obtained (○) from the Au-GIME and (□) from

the Au-IrM (without gel) SWASV measurements in 10 mM phosphate at pH 8 spiked with increasing concentrations of As(III) in a range of 1 to 7 nM. The current are presented (a) before and (b) after the normalization by Dgel/Dfree sol = 0.69. SWASV conditions: Eprecleaning = +0.5 V (30 s); Eprec = -1.0 V; tprec = 36 min; Ei = -1.0 V; Ef = +0.3 V; f = 200 Hz; ESW = 25 mV; Es = 8 mV. ... 130 Figure 3-5 As(III) stripping peak currents measured in presence of FA on the Au-GIME (▲)

and on the Au-IrM (□, ○). Currents obtained by Au-GIME are normalized by Dgel/Dfree sol = 0.69. SWASV conditions: Eprecleaning = +0.5 V (30 s); Eprec = -1.0 V; tprec = 36 min;

Ei = -1.0 V; Ef = +0.3 V; f = 200 Hz; ESW = 25 mV; Es = 8 mV. ... 131 Figure 3-6 As(III) calibration curves (n = 3) obtained from Au-GIME SWASV

measurements in (□) unfiltered Arve river water and (○) 10 mM phosphate at pH 8 spiked with increasing concentrations of As(III) in a range of 1 to 7 nM. SWASV conditions as in Figure 3-5. ... 132 Figure 3-7 As(III) calibration curves (n = 3) obtained from Au-IrM SWASV measurements

in (□) unfiltered Arve river water , (○) 10 mM phosphate at pH 8 and in (Δ) filtered Arve river water spiked with increasing concentrations of As(III) in a range of typically 1 to 10 nM . SWASV conditions as in Figure 3-5. ... 132 Figure 3-8 Corresponding calibration curve of Figure 3-5 ... 133 Figure 3-9 Corresponding calibration curves of Figure 3-6 ... 133

(13)

Figure 3-11 Influence of the temperature on the As(III) stripping peak current intensities. (a) Temperature ramp varies from 25°C to 5°C (solid) and 5°C to 25°C (open), repeated three times.(b) Corresponding voltammograms for the first temperature ramp. Sample as in Figure 3-3. SWASV conditions as in Figure 3-5, except tprec = 3 min. ... 135 Figure 3-12 (a) As(III) concentration profiles in Lake Greifen (19th August 2015)

determined from Au-GIME SWASV measurements and ICP-MS after separation on LM500 resin. Also shown are profiles of total dissolved As measured in pH 1 acidified samples by ICP-MS, (b) dissolved oxygen and chlorophyll a measured in situ using an Idronaut OS316-multiparameter probe. ... 136 Figure 3-13 (a) As(III) concentration profiles in Lake Greifen (22th August 2015)

determined from Au-GIME SWASV measurements and ICP-MS after separation on LM500 resin. Also shown are profiles of total dissolved As measured in pH 1 acidified samples by ICP-MS, (b) dissolved oxygen and chlorophyll a measured in situ using an Idronaut OS316-multiparameter probe. ... 137 Figure 4-1 Schematic of buffering nanospheres principle incorporated in an Agarose gel

deposited on a Au-IrM ... 144 Figure 4-2 Confocal image of 7.4 mM IENS doped with fluorophore trapped in 1.5% LGL

agarose ... 146 Figure 4-3 Absorbance of IENS suspension for different samples at the indicated pH. ... 147 Figure 4-4 Titration of 10 mL of Arve river with (solid) and without (dashed) suspended

IENS (22 µmol ion-exchanger capacity) with 10 mM hydrochloric acid. ... 148 Figure 4-5 (a) Curves for the titration of 10 mL of Arve river in the presence (solid) and

absence of IENS (dashed) with a 100 mM HCl (black) and the back titration with 100 mM NaOH (red). (b) Zoom of (a). ... 148 Figure 4-6 Influence of pH on the As(III) peak potential (a) and peak current (b) by SWASV.

Measurements were performed in the presence (○, solid line) or absence (□, dashed line) of IENS. At pH 8.5, the data were recorded before and after a 4h incubation time in HNO3 at pH 2 (▲) to chemically destroy the buffering particles ... 149 Figure 4-7 (a) Light intensity emitted at 645 ± 38 nm during the SWASV protocol of 2 µM

As(III) (solid) and in the absence of As(III) (dashed) in the Arve river at pH 8.0. (b) A, D, G and J correspond to the precleaning steps (reoxidation). B, E and H represent the preconcentration steps (reduction). C, F and I denote the signal acquisition times. K is an open circuit period. (c) Microscope images acquired during steps B and D. (d) As(III) peak voltammogram obtained during step E. ... 150 Figure 4-8 SWASV voltammograms in performed on gold rotating macrodisk electrode in

the presence of 2 µM As(III). ... 151 Figure 4-9 (a) Oxidation scans of cyclic voltammograms at different scan rates (20, 50 and

100 mV/s) and (b) the current intensity related to the peak at -0.45V. Sample: Arve River in 0.1 M NaNO3 adjusted at pH 8 with N2/CO2 mixture. ... 151

(14)

List of tables

Table 1-1 Total arsenic concentrations in some freshwater systems (rivers and lakes) around

the world. ... 22

Table 1-2 Analytical Methods currently approved for the analysis of arsenic in drinking water. ... 26

Table 1-3 Other analytical methods for arsenic (not currently approved for drinking water analysis). ... 27

Table 1-4 Estimates of the timescales where (a) Cottrell behavior and (b) a steady state response can be expected for various sized microsphere electrodes.¹⁸⁰ ... 48

Table 1-5 Papers involving a Nafion membrane ... 54

Table 1-6 Papers involving a cellulose acetate membrane ... 56

Table 1-7 Papers involving an agarose gel membrane ... 57

Table 1-8 A comparison of selected methods for determining As(III) with macroelectrode (NB: 1 ppb = 13.3 nM). ... 67

Table 1-9 A comparison of selected methods for determining As(III) with microelectrode (NB: 1 ppb = 13.3 nM). ... 71

Table 2-1 Reproducibility of Au layer deposition (Figure 2-2b) for a ﬁxed time of 86 s... 105

Table 2-2 Calibration slopes obtained for SWASV measurements of As(III) in a range of 10 nM to 50 nM using renewed Au-IrM. Preconcentration time = 3 min. Electrolyte: 0.01 M NaNO3 buffered at pH 8 with 10 mM phosphate ... 109

Table 2-3 As(III) current recovery for different natural samples (NF = non-ﬁltered, F = ﬁltered) ... 115

Table 3-1 Reproducibility of Au layer deposition on an Au-GIME for a fixed time of 86 s using the following conditions: Edep = -300 mV; Solution: 1 mM Au(III), 0.1 M NaNO3, HNO3 pH 2. ... 127

Table 3-2 Experimental diffusion coefficients and gel equilibration time of As(III) in 1.5% LGL agarose determined by SWASV ... 128

(15)

List of publications

This thesis is partially based on papers published in peer-reviewed journals:

 Chapter 2: R. Touilloux, M. L. Tercier-Waeber and E. Bakker, Direct arsenic(III) sensing by a renewable gold plated Ir-based microelectrode, Analyst, 2015, 140, 3526-3534.

 Chapter 3: R. Touilloux, M. L. Tercier-Waeber and E. Bakker, Antifouling membrane integrated renewable gold microelectrode for in situ detection of As(III), Anal. Methods, 2015, DOI: 10.1039/C5AY01941A.

 Chapter 4: R. Touilloux, X. Xie, M. L. Tercier-Waeber and E. Bakker, Hydrogel entrapped pH buffer based on ion-exchange nanospheres for electrochemical As(III) detection in environmental media, 2016, manuscript in preparation.

Other contributions:

 E. Bakker, M. L. Tercier-Waeber, T. Cherubini, M. C. Crespi, G. A. Crespo, M. Cuartero, M. G. Afshar, Z. Jarolimova, S. Jeanneret, S. Mongin, B. Néel, N. Pankratova, R. Touilloux, X. J. Xie and J. Y. Zhai, Environmental Sensing of Aquatic Systems at the University of Geneva, Chimia, 2014, 68, 772-777.

(16)

1 Introduction

1.1 Arsenic in water: solution chemistry and environmental behavior

1.1.1 Arsenic over the history

Arsenic was first isolated and documented by Albertus Magnus in 1250. Arsenic has a long history for its toxic and medical properties. Arsenical compounds were used in medicine 2000- 3000 years ago in the Orient and therapeutic applications have continued through the ages.

Arsenic also gained reputation as a dangerous poison in the Middle Ages, and was a preferred poison until the nineteenth century.^1-3 In 1815, the first accidental death from arsine (AsH3) poisoning was reported, and in 1900-1903 accidental poisonings from consumption of arsenic- contaminated beer were widely reported.² In the nineteenth century, the therapeutic use of arsenicals against trypanosomiasis (“sleeping sickness”) , syphilis or the “great pox” became classical examples of chemotherapy.⁴ Nowadays, research on arsenic trioxide (As2O3) is still a focus for cancer therapy, especially in China.^{5, 6} This demonstrates the widely differing toxicity and effectiveness of arsenic in its different oxidation states and in combination with other chemical species. Arsenic and its compounds have also been widely used in pigments, as insecticides and herbicides, in metal alloys, and were developed as chemical warfare agents.

Since the Second World War, however, synthetic organic compounds have substantially supplanted arsenicals for these uses.⁷

1.1.2 General chemistry of Arsenic

Arsenic is a relatively common element that occurs in air, water, soil and all living tissues.⁸ It ranks 20^th in abundance in the earth’s crust, 14^th in seawater, and 12^th in the human body.

Normal background concentration are 0.2-15 mg.kg^-1 in the lithosphere, less than 15 mg.kg^-1 in soils, 0.02-2.8 ng.m^-3 in the atmosphere, and less than 1 µg.L^-1 in the aquatic environment.⁸ Since arsenic is an element, it is indestructible. It can only change from one form to another, and be transported from one medium to another.⁹

Arsenic, described with the symbol As, is a transition element with the atomic number 33 and the atomic weight of 74.9 g.mol^-1. Arsenic occurs in group 15 (called also Va) in the periodic table, the same group as nitrogen and phosphorous. Consequently the chemistry of arsenic is in many respects similar to these elements. Contrary to popular belief, arsenic is not a metal, but a metalloid. Its chemical properties are between those of metals and non-metals: its electronical and thermic conductivity set arsenic closer to metals whereas its behaviour in solution is akin to non-metals (anion formation). Arsenic is a highly toxic compound that is brittle, crystalline, odourless and tasteless in its elemental form. It exhibits several known allotropic forms where the most stable allotrope is the steel grey form, similar to the rhombohedral form of phosphorous.¹⁰

(17)

1.1.3 Sources of Arsenic

Arsenic occurs naturally in soil and ore minerals and it enters in the environment by both anthropogenic and natural sources.

The cycle of arsenic through the environment is dependent on the various forms of arsenic occurrence in the different media. A simplified cycle that is useful in depicting the typical pathways of arsenic in the environment is presented in Figure 1-1.¹¹ This figure shows multiple potentials for human exposure from different environmental pathways. According to Figure 1-1, humans can be exposed through direct contact with water, soil, pesticides, and the atmosphere.

Each of these pathways present serious dangers based on the toxicity of arsenic.¹¹

Figure 1-1 A simplified, comprehensive cycle transfer of arsenic

1.1.3.1 Natural sources

Weathering of arsenic-containing minerals^{12, 13} and volcanic activities are the main natural sources of arsenic.¹⁴ Arsenic is found as a major constituent in more than 200 minerals. The majority are ore minerals or alteration products of ore minerals. Due to the similarity in volume densityof arsenic and sulfur (S), arsenic is often found in sulfide minerals where it substitutes for S in the mineral structure. The most abundant sulfide mineral is pyrite (FeS2), which can be formed under reducing conditions in the sedimentary environment. Pyrite is unstable under aerobic conditions and arsenic will be released upon oxidation of the mineral. The chemical behavior of arsenic is also similar to phosphorus (P) and arsenic is therefore found in many phosphate minerals as hydroxyapatite (Ca5(PO4)3(OH)).¹⁵

Non-agriculture:

Fossil fuels, Industrial wastes Municipal wastes Atmosphere:

Volatiles

Pesticides, Fertilizers

Biota:

Animals, Humans, Plants, Microbes

Soils, Rocks, Sediments Mining,

Smelting, Volcanoes

Waters:

Oceans, Freshwaters, Groundwaters

(18)

As mentioned above, arsenic is released to the aquatic environment by natural erosion processes and most of the arsenic contamination of aquifers results from the mobilization of arsenic under natural conditions. High arsenic concentrations are common in areas with geothermal activity. However, arsenic contamination is not exclusive to these areas, as arsenic can also be released to the aquatic environment through dissolution of minerals containing arsenic.¹⁵

1.1.3.2 Anthropogenic sources of arsenic

Humanity has also contributed to the contamination of arsenic through industrial discharges, mining activity, use of arsenic herbicides, wood preservation and crop desiccants, combustion of fossil fuels and by using arsenic as an additive in livestock feed.^{13, 15} Arsenic is also used in the manufacturing of glass, semiconductors and paper.¹⁶ Perhaps paradoxically, arsenic trioxide (As2O3) is used therapeutically in medicine as a treatment of a certain type of leukemia.^{5, 6, 17} The world production of arsenic in 2012 was estimated to be 46’700 tons as As2O3 (mainly in China, Chile and Morocco), of which around 1’000 tons were estimated to be produced in the European Union (mainly Belgium).¹⁸

1.1.4 Arsenic contamination of natural waters

1.1.4.1 Worldwide

It has been observed that drinking arsenic tainted groundwater is the primary pathway for exposure of humans.^{19, 20} The effect on human health through exposure of naturally occurring arsenic in groundwater is by exposure through three possible pathways: i) ingestion of arsenic by drinking groundwater; ii) ingestion of arsenic through cooking processes; and iii) intake of crops irrigated with arsenic contaminated groundwater (Figure 1-2).

(19)

Figure 1-2 Effect of arsenic exposure to humans due to As contaminated groundwater.^{19, 20}

Figure 1-3 lists the countries most concerned by arsenic contamination around the world.

The most important case in Bangladesh where ~50 millions of people are exposed to drinking water with high arsenic concentrations up to 1000 µg.L^-1.²¹ This is the worst global intoxication that humanity has never known.²² There is no known treatment available for arsenic related diseases. The only way is to drink As-safe water to avoid poisoning, otherwise arsenic poisoning could become the principal cause of mortality in the country.^{20, 23, 24} For this purpose the World Health Organization (WHO) has, in 1993, set a 10 µg.L^-1 arsenic threshold value, decided upon according to the practical quantification limits and practical difficulties in removing arsenic compounds from contaminated water. Any amount of arsenic above 10 µg.L^-1 in water is considered unsafe for consumption.²⁴

Groundwater as source of drinking and irrigation water with elevated concentrations

of arsenic

Possible options not adopted by people due to technical, social,

economical aspect, etc…

Malfunction of possible mitigation-options

Arsenic in water for cooking

Insufficient knowledge concerning possible and awareness level in

rural and disadvantaged communities

Arsenic in soil and crop Exposure through

food chain Arsenicosis

Health effects Social and economical impacts

Exposure through drinking water

Arsenic in drinking water

Exposure through food

Exposure of high levels of As to population

(20)

Figure 1-3 Known areas with natural arsenic contamination around the world.²⁵ (Reproduced with permission of Royal Society of Chemistry via Copyright Clearance Center).

1.1.4.2 Switzerland

Switzerland has three main areas with elevated natural arsenic concentrations: i) The northern part, where a number of arsenic containing thermal and mineral springs are located; ii) the Jura mountain, with its iron containing iron-rich limestones and clay: iii) the Alps, where arsenic bearing ore deposits and crystalline rock formations can be found. In addition, there are other isolated thermal and mineral springs (Figure 1-4).^{26, 27}

(21)

Figure 1-4 Areas with elevated concentrations of arsenic in Switzerland²⁶. (Reprinted with permission from IWA Publishing)

The discovery of rich arsenic soils (200 mg.kg^-1) and waters (5-12 µg.L^-1) started in 1989, when a Canadian mining company proposed to start an exploration campaign on the site of a former gold mine in Ticino. The first case of arsenic pollution was accidentally found after a microbiological contamination in the village of Astano (Ticino) in 1996. Concentrations up to 80 µg.L^-1 were measured in one of the two drinking water reservoirs. This fact lead the authorities to perform a systematic survey of the drinking water of the whole Ticino. Also because of the crisis in Bangladesh, all Swiss drinking and mineral waters have been checked from 1997 to 2002 for elevated arsenic concentrations. This complete survey of public drinking water supplies revealed that about 20’000 people were exposed to arsenic concentrations above 10 µg.L^-1.²⁷ 80% of the waters are well oxygenated and contain the arsenic in its pentavalent oxyanion form.

The local presence of rocks and soils with elevated arsenic concentrations and often pH values above 7.5 seem to be at the origin of the observed values. Cold and warm mineral springs contain between 20 and 50% trivalent arsenic. In water flooded forest soils and especially in wetlands, arsenic concentrations may reach 1000 µg.L^-1 and the percentage of As(III) is up to 90% at Eh- values close to 0 mV.²⁸ These facts, along with WHO guidelines, led the authorities to decrease the authorized level of arsenic to 10 µg.L^-1 in drinking and mineral waters in December 2013.

Owing to the required cost investments to implement this new limit, it was decided to agree for a delay of 5 years, until January 2019.²⁹

(22)

1.1.5 Abundance and distribution of Arsenic

Arsenic concentrations in freshwaters may vary by several orders of magnitude depending on the source, availability and chemistry of the environment.¹⁵ Typically total As concentrations in freshwaters are less than 10 µg.L^-1 but range between 0.02 and 160’000 µg.L^-1 in rivers and between < 0.2 and 21’000 µg.L^-1 in lakes can be found (Table 1-1). Data presented in Table 1-1 are of various contaminated and non-contaminated sites, and thus will give an idea of the known ranges and their variations in the freshwater environment.

(23)

Table 1-1 Total arsenic concentrations in some freshwater systems (rivers and lakes) around the world.

Water systems and Location As concentrations

/ µg.L^-1 Year of

analysis Reference River water

Norway 0.3 (<0.02-1.1) 1978,2002 15, 30

Walker River, Sierra Nevada, USA 0.20-264 1983 31

Schelde catchment, Belgium 0.75-3.8 (up to 30) 1989 32

Lake Pavin, Dordogne, France 0.7 1984 33

Lao River, Northern Chile 19-21000 1992 34

Po River, Italy 1.3 1992,1994,1997 35-37

Waikato, New Zeland 32 (28-36) 1995 38, 39

Willow River, B.C., Canada 0.6 (0.3-0.8) 1995 40

Madison and Missouri rivers, USA 44 (19-67), 10-370 1995,1998 39, 41

Rio Tercero, Cordoba, Argentina 7-114 1996 42

Ron Phibun, Thailand 218 (4.8-583) 1996 43

Anllóns–Laxe, Spain 0.2-40 2006 44

Lot-Garonne River system 0.57-6.90 1999-2005 45

Bijiang river, China 50-250 2010 46

Himalayan regions 0.6-195 2010-2012 47

Elbe river, Czech Republic 0.5-10.5 2011 48

Guguletu river, South Africa 1.10 (0.62-2.03) 2011-2012 49 Langa river, South Africa 0.44 (0.08-0.68) 2011-2012 49

Harlech Dome, UK 0.2 (0.1-5.0) 2013 50

Lake water

Lake Echols, Tampa 3.58 1973 51

Lake Magdalene, Tampa 1.75 1973 51

Western USA 0.38-1000 1983 31

Lake Greifen, Switzerland 0.15-0.35 1989-1991 52

Sweden 0.06-1.2 1992 53

Mono Lake, California, USA 10000-21000 1992 54

Moira Lake, Ontario, Canada 20.4 (22.0-47.0) 1995 55 Jack of Clubs Lake, B.C., Canada 0.3 (0.2-0.4) 1995 40

Bowron Lake, B.C., Canada <0.2 1995 40

Lowhee Creek, B.C., Canada 1.5 (0.2-2.0) 1995 40

Northwest territories, Canada 270 (64-530) 1996 56

Lake Nakaumi, Japan 1-12 2003-2005 57

Lake Mohawk, New Jersey USA 23-25 2005 58

Lake Biwa, Japan 2.2 (0.6-1.7) 2010 59

Lake Kahuku 1.1 (0.4-1.7) 2010 59

Lake Kiba, Japan 0.5 (0.2-0.7) 2010 59

Lake Yangzonghai, China 50-75 2013 60

Lake Greifen, Switzerland 0.5-0.6 2014 This work⁶¹

(24)

1.1.5.1 River water

The baseline concentrations of As in various river waters range between 0.1 and 4.8 µg.L^-1 with an average of 1.3 µg.L^-1 (Table 1-1). These variations might be related to the surface recharge, baseflow, and the bedrock lithology. Low average As concentrations of about 0.25 µg.L^-1 (range < 0.02-1.1 µg.L^-1) in rivers draining basement rocks in Norway,³⁰ and 0.15-0.45 µg.L^-1 in river waters of the south-eastern USA have been reported.⁶² On the other hand, high concentrations of naturally-occurring As have been reported in New Zealand (Waikato River, 32 µg.L^-1),^{38, 39} USA (Madison and Missouri Rivers, 10-370 µg.L^-1),⁴¹ Sierra Nevada, USA (264 µg.L^-1),³¹ and California, USA (Owens River, 85-153 µg.L^-1).⁶³ High concentration of As in these rivers occurred as a result of inputs from geothermal sources.¹⁵ Extremely high concentrations of naturally-occurring As have been reported in waters from the Lao River of northern Chile (up to 21’000 µg.L^-1),³⁴ which might be due to geothermal input (volcanic area), evaporation, and groundwater input. The River Zenne in Belgium also contains high concentrations of As (up to 30 µg.L^-1) which have been affected by inputs from urban and industrial sources, especially sewage.³² Mining activity may also result in the occurrence of high As in river waters. Stream waters adjacent to tailing deposits in the Clubs Lake (Canada) contained up to 556 µg.L^-1 As.⁶⁴ Waters of River Ron Phibum (Thailand),⁴³ the Ashanti (Ghana),⁶⁵ River Bijiang (China)⁴⁶ and Caudal (Spain),⁶⁶ have been reported to have 50-300 µg.L^-1 and up to 160 mg.L^-1 As, due to the effects of Sn, Au, Zn and Hg mining activities, respectively. In Australia, mining and processing of arsenopyrite ore at the Mole River mine of New South Wales during the 1920–1930s has resulted in As contamination of the Mole River.⁶⁷ This phenomenon is not isolated as observed in Italy at the Pecora and Bruna River.⁶⁸ Even the Himalayan Rivers are not spared by arsenic pollution due to anthropogenic activities as demonstrated recently by Zhang et al.⁴⁷

Arsenic concentrations in river waters show distinct seasonal variations (e.g. Madison River, the Himalayans rivers). As concentrations have been reported to be highest during the low-flow season which might be due to the greater contribution of As input from geothermal water and spring runoff.^{41, 47, 49} It can be also explained by the snow melting period during the early summer which increase the volume of the river and decreased the As concentration “by dilution”. On the other hand, maximum As concentration in Waikato river (New Zealand)³⁸ and in the Lot- Garonne River system (France),⁴⁵ during summer was supposed to be due to the influence of temperature controlled microbial activity (reduction of As(V) to As(III) and the consequent mobility of As(III) in the water column). Buzek et al.⁴⁸ showed that the seasonal variations in Elbe river is directly related to the Dissolved Organic Matter (DOM) concentration and the As adsorption on Fe colloids. The DOM stabilizes As in solution and reduces its re-adsorption on Fe colloids and consequently As concentration in the stream increases. The effects of iron on the As concentration on the seasonal changes were also supported by other studies.^{44, 68, 69}

1.1.5.2 Lake water

Data presented in Table 1-1 reveal that As concentrations in lake waters are similar to or lower than those in river waters. Arsenic concentrations in a number of lakes in British Columbia (Canada) have been investigated.^{40, 55, 64} Results showed that the lowest concentration of As was

(25)

of As in these lakes was thought to be the abandoned Cariboo Gold mine tailings,⁴⁰ from where As was transported to the lakes. Most of the As was accumulated in bottom sediments of these lakes and its concentration in the lake waters was observed to be low (between < 0.2 and 2.08 µg.L^-1). Elevated concentrations of As (between 22.0 and 47.0 µg.L^-1) in the output of Moira Lake (Canada) have also been reported (Table 1-1). The persistent input of soluble As into Moira Lake indicates the existence of a continuous source of As.⁵⁵ In common with river waters, As concentrations in lake waters are also influenced by geothermal inputs and mining activities.¹⁵ Although mining effluents have increased As concentrations in river waters,43, 64, 65, 67 some lake waters affected by such effluents have been found to have decreased As concentrations.⁴⁰ The lower concentration of As in mining-affected lake waters might be due to its adsorption on Fe- oxides under neutral or acidic conditions¹⁵ and accumulation in the bottom sediments.⁴⁰ Similar As concentrations in the surface and bottom waters in mining-affected Canadian lakes reveal its adsorption and accumulation in the bottom sediments and its limited movement from the sediment into the overlying water.⁴⁰ Phosphate has been shown to compete with arsenate for binding sites on solids. In contrast to the low As concentrations in some lake waters, the occurrence of high As concentrations, due to extreme evaporation and/or geothermal inputs, has been reported in alkaline lakes (pH 9.5-10). Maest et al.⁵⁴ reported extremely high dissolved As concentrations (between 10’000 and 20’000 µg.L^-1) in Mono Lake (USA) (Table 1-1) resulting from geothermal water input, evaporation and high pH. Variations in As concentrations with thermal stratification in lake waters have also been reported by several researchers.55, 59, 70, 71 The dissolved As concentrations in surface water of Moira Lake (Canada) have been reported to be highest during summer, with an average concentration of 47.0 µg.L^-1 , compared to that of 22.0 µg.L^-1 in winter.⁵⁵ Hasegawa et al.⁵⁹ investigated the seasonal changes of As speciation in 18 lakes around Ishikawa, Nagano, Fukui and Shiga prefectures in Japan and found that the total As concentrations in surface waters of these lakes were higher in summer than in winter. Similar results in the occurrence and distribution of As in lake waters have also been reported by other researchers.^72-74 Seasonal variations in the occurrence of As in lake waters were due to the release of inorganic As from the sediments into the water of the bottom layer under anaerobic conditions in summer, adsorption of As(V) onto Fe/Mn oxides, and accumulation into the sediments in winter.⁵⁹ Depletion of O2 levels in the bottom layer due to increased biological activities during summer have also been considered to cause higher As concentrations in lake waters.¹⁵ In addition to the biological activities of aquatic organisms, organic matter also plays an important role in the distribution of As species in freshwaters. Due to these factors previously cited, Liu et al.⁶⁰ expect that the Lake Yangzonghai, previously treated against an As pollution, will know a secondary pollution. Their results showed that the As that had already been precipitated into the sediment by FeCl3 would be released again into the water body due to the increasing activity of anaerobic microorganisms. Sohrin et al.⁷¹ showed that the speciation of As in lake waters was affected by biological processes such as decomposition of organic matter by bacteria, and by the primary production of phytoplankton. In aquatic systems As concentrations are usually much higher in sediment (mg.kg^-1 level) than in the overlying water (μg.L^-1 level) since As is easily non-specifically bound or adsorbed to suspended and settling particles such as Mn/Fe oxides, organic matter, sulfides, and carbonates.⁷⁵ The concentration of As in lake sediment often

(26)

correlates well with the amount of Fe and Mn hydrous oxides, which supports the idea that As co-precipitates in water and settles into sediment with Fe/Mn oxyhydroxides.^{58, 76-81}

1.1.6 Analytical techniques for Arsenic measurement in natural waters Historical uses of arsenic in products, pharmaceuticals, and industry have led to important insights on arsenic toxicity. In the modern era, most interest in arsenic toxicology comes from naturally occurring background exposure to food, water, and soil as presented above.

Understanding the environmental levels that can cause a public health concern is a key area of research. Figure 1-5 shows the increasing interest of the study of arsenic chemically and environmentally from 1980 to 2014. The interest started to rise after the WHO decided in 1993 to establish an arsenic threshold at 10 µg.L^-1 in water.

Figure 1-5 Number of papers on environmental and chemical arsenic studies during each year from 1980 to 2014.

1.1.6.1 Approved EPA methods

Six methods are currently approved by the Environmental Protection Agency (EPA) for the analysis of arsenic in drinking water. Table 1-2 lists the approved methods and associated detection limits. The table also shows two methods that were approved in 1999 but removed in 2006 after the new arsenic maximum contaminant level was determined by the EPA as 10 µg.L^-

1 instead of 50 µg.L^-1in water.

0 100 200 300 400 500 600 700 800 900

P a pe rs

Year

(27)

Table 1-2 Analytical Methods currently approved for the analysis of arsenic in drinking water.

Method Technique Detection

Limit (µg.L^-1) Multi-Element

Methods

EPA 200.8⁸² ICP-MS 1.4

EPA 200.7⁸³ ICP-AES 8 Removed

in 2006

SM 3120B⁸⁴ ICP-AES 50 Removed

in 2006

Single Analyte Methods

(SAM)

EPA 200.9⁸⁵ GF-AAS 0.5

SM 3113B⁸⁶ GF-AAS 1

ASTM 2972-

15C⁸⁷ GF-AAS 5

SM 3114B⁸⁸ HG-AAS 0.5

ASTM 2972-

93B⁸⁹ HG-AAS 1

ICP-MS:Inductively Coupled Plasma-Mass Spectrometry

ICP-AES: Inductively Coupled Plasma-Atomic Emission Spectrometry GF-AAS: Graphite Furnace Atomic Absorption Spectrometry HG-AAS: Hydride Generation Atomic Absorption Spectrometry EPA: Environmental Protection Agency

SM: Standard Method

ASTM: American Society for Testing and Materials

One of the currently approved methods is multi-element, which means that other analytes besides arsenic can be measured during the analysis. The method is Inductively Coupled Plasma- Mass Spectrometry (ICP-MS) and is listed as EPA 200.8.⁸²

The remaining five analytical methods approved by EPA for the measurement of arsenic in drinking water are all single-analyte methods (SAM), as they can only measure arsenic. Three of them (EPA 200.9,⁸⁵ SM 3113B⁸⁶ and ASTM 2972-15C⁸⁷) are Graphite Furnace Atomic Absorption Spectrometry (GF-AAS) that have traditionally been used to quantify arsenic in drinking water. The methods EPA 200.9 and SM 3113B employ the use of Stabilized Temperature Platform Graphite Furnace Atomic Absorption (STP-GFAAS) technology that significantly reduces interferences and improves analytical sensitivity. ASTM 2972-15C employs a regular hollow graphite tube.

Two of the five SAM, SM 3114B⁸⁸ and ASTM 2972-93B,⁸⁹ use hydride generation atomic absorption spectrometry (HG-AAS). These methods employ zinc in hydrochloric acid or sodium borohydride to convert arsenic to its volatile hydride. In ASTM 2972-93B, the arsenic hydride is removed from the sample by a flow of nitrogen into argon or nitrogen entrained hydrogen flame where it is determined by atomic absorption at 193.7 nm. In SM 3114B, the volatile hydrides may also be swept into an entrained hydrogen flame, or alternatively, into a quartz atomization cell positioned in the optical path of an atomic spectrophotometer. Quartz atomization cells provide the most sensitive arsenic hydride determinations and minimize background noise associated with hydrogen flames. In both methods, the absorption of the light source is proportional to the concentration of arsenic. Both hydride methods provide method- specific sample digestion procedures that are required prior to analysis.

(28)

It has to be noted that the hydride generation system (HG-AAS) has an important advantage compared to ICP-MS, as it can perform speciation analysis. Indeed, As(V) can be first reduced to As(III) followed by transformation to AsH3, but As(III) can be also directly reduced to AsH3. The difference in reaction kinetics is more obvious at high pH. Thus, the reduction rate is slower at high pH than at low pH. By controlling the pH of the reaction solution, inorganic arsenic speciation can be achieved. Hydride generation being a device which can be connected before a ionization source, it can be adapted the ICP-MS to become HG-ICP-MS. The analysis performed with this kind of coupling give better results in term of sensitivity than by HG-AAS.

1.1.6.2 Non-approved EPA methods

There are other methods cited by the EPA that can detect arsenic at 10 µg.L^-1 but that are currently not approved. Table 1-3 lists these techniques with their detection limits. Two multi- element methods are cited, SW-846 6020⁹⁰ and EPA 200.15.⁹¹ The first one is an ICP-MS technique that is very similar to the approved EPA 200.8. The second one is an ICP-AES technique that employs an ultrasonic nebulization to introduce the aqueous sample into the plasma torch.

There are four other SAM methods for the analysis of arsenic, SW-846 7060A,⁹² SW-846 7062,⁹³ EPA 1632⁹⁴ and SW-846 7063.⁹⁵ The first one is a classical GF-AAS method whereas the others are based on HG-AAS.

Finally, SW-946 7063 provides an alternative analysis procedure that uses an anodic stripping voltammetry (ASV) technique to quantify free, dissolved arsenic in aqueous samples.

The chief advantage of this technique is that it does not require expensive instrumentation and has potential as a field technique. This method also has the advantage to perform arsenic speciation analysis by adapting the pH of the solution.⁹⁵

Table 1-3 Other analytical methods for arsenic (not currently approved for drinking water analysis).

Method Technique Detection

Limit (µg.L^-1) Multi-element

methods

SW-846 6020 ICP-MS 0.4

EPA 200.15 ICP-AES 3

Single analyte methods

SW-846 7060A

(EPA) GF-AAS 1

SW-846 7062 (EPA) HG-AAS 1

EPA 1632 HG-AAS 0.002

SW-846 7063 (EPA) ASV 0.1

EPA: Environmental Protection Agency ASV: Anodic Stripping Voltammetry

(29)

To conclude, it has to be noted that most of the listed methods require the acidification of the sample prior to analysis. Methods based on these techniques require expensive instrumentation, complicated procedures and special sample pre-treatment. Besides, most of these methods are essentially sensitive to total arsenic.

1.1.6.3 Techniques used in Switzerland

No standard methods are required legally in Switzerland and EPA methods are not commonly applied in Switzerland because they serve no purpose (i. e. do not warrant better results). Laboratories which fulfill the norm ISO 17025⁹⁶ are recognized to perform arsenic measurements in water. For example, EAWAG (Swiss Federal Institute of Aquatic Science and Technology, ZH) and most of the Cantonal Laboratories in Switzerland use ICP-MS. It allows them to perform arsenic measurements below 1 µg.L^-1 in all kinds of freshwater (environmental, drinking water etc.) but not in sweater due to the strong interference of chloride on As. The law being based only on the total arsenic, no speciation is performed.

The methods previously discussed are developed to perform control analysis every month and are not suited for on-line monitoring to rapidly alert the inhabitants in case of a sudden contamination of drinking water. It was shown by Masson et al.⁹⁷ the importance of high- resolution temporal monitoring (i.e. hourly time scale or shorter) for both environmental studies and water quality monitoring strategies for arsenic. Moreover, it is always the total arsenic which is measured and the speciation is not taken in account. As explained below, the toxicity of the different species of arsenic is not equal.

1.1.7 Speciation of Arsenic

Arsenic forms a number of inorganic and organic compounds (Figure 1-6). Naturally occurring inorganic arsenic is stable in the oxidation state –III, as in arsine gas (AsH3), 0 as in crystalline arsenic, +III as in arsenite, and +V as in arsenate. Organic species include monomethylarsonic acid (MMA), and dimethylarsinic acid (DMA).^{9, 98} They may be produced by biological activity, mostly in surface waters, but are rarely important quantitatively. Organic forms may, however, occur where waters are significantly impacted by industrial pollution.⁹⁹ In relatively pristine natural ground water environments, As(III) and As(V) are typically the dominant forms of arsenic.

(30)

Figure 1-6 Arsenic species in water

1.1.7.1 Acid-base chemistry

Depending on pH, the different forms of As(III) are H₄AsO₃⁺, H₃AsO₃, H₂AsO₃^-, HAsO₃^2-, and AsO₃^3-, whereas As(V) occurs as H₃AsO₄, H₂AsO₄^-, HAsO₄^2-, and AsO₄^3-.¹⁰⁰ The amount of protonation of both As(III) and As(V) is an important factor governing the mobility of these chemicals species. For example, within the range of groundwater (~7 to 9), the uncharged species of As(III) is predominated while As(V) is negatively charged (Figure 1-7). As a result, As(III) is more mobile than As(V). Transport of As(V) is retarded by electrostatic attraction to positively charges particles, such as iron hydroxides.¹⁰¹ This information is also useful for designing effective removal strategies and for determining the speciation of arsenic by ion separation techniques.¹⁰²

Dimethylarsinic acid (DMA)

Monomethylarsonic acid (MMA)

Arsenious acid As(III)

Arsenic acid As(V)