DISS. ETH NO.

19937

Treatment of arsenic-contaminated water with

polyaluminum nanoclusters

A dissertation submitted to
ETH ZURICH

for the degree of

Doctor of Sciences

presented by
Jasmin M. Mertens
Dipl.-Geol. RWTH Aachen University

Born 1st November, 1983

Citizen of Germany

Accepted on the recommendation of

Prof. Dr. Bernhard Wehrli, examiner

Prof. Dr. Gerhard Furrer, co-examiner
Prof. Dr. Calin Baciu, co-examiner
Dr. Jérôme Rose, co-examiner

2011
- ii -
 Contents

Contents

Summary.......................................................................................................... vii

Zusammenfassung ............................................................................................ ix

1. Introduction ................................................................................................... 1
1.1. Occurrence of arsenic in the environment ..............................................................2

1.2. Arsenic speciation and toxicity ................................................................................3

1.3. Global distribution of geogenic arsenic in groundwater.........................................4

1.4. Study site: The East Pannonian Basin.....................................................................6

1.5. Methods for arsenic removal from groundwater....................................................7

1.6. Nanomaterials ........................................................................................................10

1.7. Objectives ...............................................................................................................11

1.8. Thesis Outline.........................................................................................................13

2. Al nanoclusters in coagulants and granulates: application in arsenic

removal from water.....................................................................................15
2.1. Introduction............................................................................................................16

2.2. Working principles.................................................................................................18

2.3. Applications............................................................................................................20

2.4. Training of young researchers ...............................................................................22

3. Geochemistry and arsenic behaviour in groundwater resources of the

Pannonian Basin (Hungary and Romania) ................................................24
3.1. Introduction............................................................................................................26
3.1.1. Occurrence of arsenic in groundwater..............................................................................26
3.1.2. General palaeogeography and geological development of the Pannonian Basin .......29
3.1.3. Geology and mineralogy......................................................................................................30
3.1.4. Ground water system of the field area ..............................................................................31
3.1.5. Aims of this study .................................................................................................................33

3.2. Methodology ...........................................................................................................34

- iii -
Contents

3.2.1. Sample location and collection ...........................................................................................34

3.2.2. Aqueous phase analytical methods ....................................................................................35
3.2.3. Cluster analysis.....................................................................................................................37

3.3. Results and discussion ........................................................................................... 38

3.3.1. Group 4 – thermal and saline influences ..........................................................................39
3.3.2. Group 3 – dug wells .............................................................................................................46
3.3.3. Groups 1 and 2 – general groundwaters...........................................................................47
3.3.4. Arsenic, source and mobilisation mechanisms ................................................................52

3.4. Conclusions ............................................................................................................ 56

3.5. Acknowledgements ................................................................................................ 58

4. Polyaluminum chloride with high Al30 content as removal agent for

arsenic-contaminated well water ............................................................... 59
4.1. Introduction ........................................................................................................... 60

4.2. Materials and methods .......................................................................................... 62

4.2.1. Chemicals...............................................................................................................................62
4.2.2. Analytical methods...............................................................................................................63
4.2.3. Ferron method ......................................................................................................................63
4.2.4. 27Al MAS NMR .....................................................................................................................64
4.2.5. Titration experiments ..........................................................................................................65
4.2.6. Coprecipitation experiments ..............................................................................................66
4.2.7. Field tests ...............................................................................................................................66

4.3. Results and discussion ........................................................................................... 67

4.3.1. Characterization of PAClAl30 ..............................................................................................67
4.3.2. Arsenic removal with PAClAl30 ...........................................................................................70
4.3.3. Field validation .....................................................................................................................76

4.4. Conclusion.............................................................................................................. 77

4.5. Supplemental material .......................................................................................... 79

4.5.1. Supplemental figures ...........................................................................................................79
4.5.2. Supplemental table...............................................................................................................80

5. Adsorption of arsenic on polyaluminum granulate ................................... 81

5.1. Introduction ........................................................................................................... 82

5.2. Materials and methods .......................................................................................... 84

- iv -
 Contents

5.2.1. Chemicals...............................................................................................................................84
5.2.2. Preparation of Al granulate ................................................................................................84
5.2.3. Characterization of PAG.....................................................................................................85
5.2.4. Adsorption experiments ......................................................................................................85
5.2.5. Chemical analysis .................................................................................................................86
5.2.6. Desorption experiments.......................................................................................................87
5.2.7. As distribution on individual grains..................................................................................87

5.3. Results.....................................................................................................................88
5.3.1. Physicochemical characteristics of polyaluminum granulate .......................................88
5.3.2. As(III) and As(V) adsorption..............................................................................................89
5.3.3. Adsorption kinetics ..............................................................................................................91
5.3.4. Arsenic desorption................................................................................................................94
5.3.5. Arsenic distribution in PAG ...............................................................................................95

5.4. Discussion ...............................................................................................................99

5.4.1. Arsenic adsorption ...............................................................................................................99
5.4.2. Sorption reversibility .........................................................................................................100
5.4.3. Adsorption kinetics and intra-particle distribution......................................................100

5.5. Conclusion ............................................................................................................101

5.6. Supplemental material .........................................................................................102

5.6.1. Supplemental figures .........................................................................................................102

6. Arsenate adsorption on Al nanoparticles and Al-based sorbents during

water treatment – an EXAFS study .........................................................105
6.1. Introduction..........................................................................................................106

6.2. Materials and methods.........................................................................................108

6.2.1. Chemicals.............................................................................................................................108
6.2.2. Aluminum sorbents ............................................................................................................109
6.2.3. Adsorption isotherm experiments....................................................................................109
6.2.4. X-ray absorption spectroscopy.........................................................................................110

6.3. Results...................................................................................................................111
6.3.1. Surface area.........................................................................................................................111
6.3.2. As(V) adsorption isotherms ..............................................................................................112
6.3.3. Surface coverage for EXAFS samples.............................................................................114
6.3.4. X-ray absorption spectroscopy.........................................................................................115

6.4. Conclusions...........................................................................................................120

-v-
Contents

7. Discussion and conclusions........................................................................ 121

7.1. Introduction ......................................................................................................... 122

7.2. Arsenic contamination and water chemistry in the East Pannonian Basin ....... 122

7.3. Efficiency of Al nanoclusters for the removal of arsenic.................................... 123

7.3.1. Arsenic removal using PAClAl30 .......................................................................................123
7.3.2. Arsenic removal using polyaluminum granulate (PAG)..............................................124
7.3.3. Comparison of PAClAl30 and PAG ...................................................................................125
7.3.4. Molecular structure of arsenic during water treatment with Al nanoclusters.........126

7.4. Recommendations for further research.............................................................. 126

Appendix ........................................................................................................ 128

Appendix 1: Column Study - Adsorption of arsenate on polyaluminum granulate.. 129

Appendix 2: Case study - Arsenic removal from groundwater using polyaluminum

granulate in a pilot plant, Sepreus, Romania........................................................ 132

Appendix 3: Additional data ....................................................................................... 143

References ...................................................................................................... 144

Acknowledgements ........................................................................................ 157

Curriculum Vitae........................................................................................... 161

- vi -
 Summary

Summary
The World Health Organization identified arsenic-contaminated drinking water as a dominant
source of arsenic intake. Long- and short-term exposure to arsenic (As) causes cancer, vascular
disease, skin lesions and neurological disorders. Therefore, a maximum concentration level of
10
g/L was implemented in European and US drinking-water directives. Four chemical and
physical water treatment processes are generally used to remove arsenic from water:
adsorption, ion exchange, coagulation-coprecipitation, and membrane filtration. The
performance of all treatment methods highly depends on pH, raw water chemistry, arsenic
concentration and speciation, and the material used. In the last decade, nanoparticles have been
tested in water treatment applications due to properties that could enhance treament efficiency,
e.g. high reactivity and high surface area to volume ratio. Aluminum nanoclusters with 1-2 nm
in size have been evaluated in a few studies to be good removal agents for humic acids and
extremely effective coagulants due to strong flocculation.
This thesis focused on the evaluation of aluminum nanoclusters as removal agents for the two
inorganic arsenic species As(V) and As(III) from water. Aluminum nanoclusters were applied
in two ways: (i) in a coagulation process as main constituents of the commercial coagulant
polyaluminum chloride (PACl), and (ii) in an adsorption process as adsorbent material
produced from Al nanocluster aggregates, named polyaluminum granulate (PAG). As a main
approach, laboratory-scale batch studies were performed with deionized and synthetic water.
The presence of Al nanoclusters over the duration of experiments was confirmed in
polyaluminum chloride solution by the Ferron method and in derived precipitates by 27Al
NMR. The interaction of arsenic with aluminum surfaces during the two removal processes
was investigated at molecular scale by extended X-ray absorption fine structure (EXAFS),
micro X-ray fluorescence and scanning electron microscopy.
The key factors for the efficiency of an arsenic treatment method, arsenic concentration and
speciation, and chemical water composition, were analysed for various groundwater types in
the Pannonian Basin (East Hungary and West Romania), the most severely arsenic-affected
area in Europe. Measured arsenic concentrations ranged from < 0.5 to 240
g/L, with As(III)
being the predominant arsenic species. Consistently high arsenic concentrations were found in
methanogenic waters. Arsenic mobility was controlled by reductive dissolution of iron
minerals and by retention mechanisms, which were governed by sulfate reduction and the
organic carbon concentration.

- vii -
Summary

Arsenic removal during the coagulation process with PACl was investigated regarding pH
dependence, Al and As concentrations. Best results were achieved with the application of 0.4 –
6 mmol Al/L at pH 7-8. Laboratory findings were verified with removal tests on As-rich raw
groundwater from two different field sites in Europe: oxidized groundwater in irrigation tanks
in Northern Greece and reduced artesian water in Western Romania.
In order to implement Al nanoclusters in a flow-through adsorption system, polyaluminum
granulate (PAG) of 1-3 mm was produced as part of this work and its physical properties were
characterized. The adsorption of arsenic was investigated with isotherms, and showed that the
Langmuir model with a limited amount of sorption sites described best As(V) sorption,
whereas linear As(III) adsorption followed the Freundlich model. Adsorption kinetics were
described by a reaction of pseudo-second order. Sorption reversibility showed that one half of
adsorbed As(III) could be removed from PAG with synthetic water; As(V) was desorbed by a
maximum of 5 %. In general, As(III) removal was less efficient than As(V) removal with both
methods.
To predict the efficiency and the fate of arsenic in water purification systems it is essential to
understand how arsenic interacts with the adsorbent surfaces. The application of micro X-ray
fluorescence and scanning electron microscopy helped to reveal new insights into arsenic
adsorption processes to aluminum granulate. It was shown that As(III) was distributed
uniformly within the entire grain, while As(V) diffused only into the grain’s rim due to its
strong binding affinity to Al hydroxide surfaces. The ability to identify arsenic distribution is
important to increase efficiencies of adsorbents. The arsenate sorption mechanism was
determined with EXAFS and revealed the formation of inner-sphere bidentate binuclear
complexes on the surfaces of Al nanoclusters, both in polyaluminum chloride and
polyaluminum granulate. Direct chemical bonds of As(V) on Al surfaces implied a robust
sorption mechanism.
The results of this thesis showed that Al nanoclusters used in a coagulation-coprecipitation
system have the ability to reduce As(V) concentrations between 20 and 2300
g/L in
groundwater to below the drinking-water limit, and that the employment of this method is most
efficient in a common pH range for groundwater. The adsorbent polyaluminum granulate has
the potential to strongly reduce As(V) in higher concentration ranges. Based on this work,
removal efficiency of Al sorbents could be enhanced by adjusting the grain size to kinetic
conditions.

- viii -
 Zusammenfassung

Zusammenfassung
Arsen-kontaminiertes Trinkwasser wurde von der Weltgesundheitsorganisation als eine der
Hauptquellen für die Aufnahme von Arsen identifiziert. Der Kontakt mit Arsen führt zu lang-
und kurzfristigen Beschwerden, zu denen Krebs, Gefässerkrankungen, Hautläsionen, und
neurologische Störungen gehören. Aufgrund dessen wurde in Europäischen und US
Trinkwasserverordnungen der maximale Arsengehalt auf 10
g/L festgelegt. Generell werden
vier chemische und physikalische Verfahren zur Entfernung von Arsen angewandt:
Adsorption, Ionenaustausch, Koagulation-Kopräzipitation, und Membranfiltration. Die
Leistung von allen Behandlungsmethoden ist stark abhängig vom pH, von der Wasserchemie,
der Arsenkonzentration und –speziierung, sowie vom Behandlungsmaterial. Innerhalb des
letzten Jahrzehnts wurden Nanopartikel zunehmend in der Wasseraufbereitung getestet, da sie
sich durch besondere Eigenschaften auszeichnen, z.Bsp. hohe Reaktivität und ein grosses
Oberflächen Volumen Verhältnis. Aluminium-Nanocluster haben eine Grösse von 1-2 nm und
wurden in mehreren Studien als gutes Entfernungsmittel für Huminsäuren und als effektive
Koagulanten wegen ihrer starken Ausflockung beschrieben.
Der Schwerpunkt dieser Dissertation liegt auf der Evaluation von Aluminium-Nanoclustern als
Entfernungsmittel für die zwei anorganischen Arsenspezies As(V) und As(III). Aluminium-
Nanocluster wurden in zwei Methoden angewandt: (i) in einem Koagulierungsprozess als
Hauptbestandteil des kommerziellen Fällungsmittels Polyaluminium Chlorid (PACl), (ii) in
einem Adsorptionsprozess als Adsorptionsmittel, benannt als Polyaluminium-Granulat (PAG).
Der Hauptansatz lag in der Durchführung von Laborstudien mit deionisiertem und
synthetischem Wasser. Die Präsenz der Aluminium-Nanocluster wurde in der
Polyaluminiumchlorid-Lösung sowie in den erhaltenen Fällungen über den
Experimentzeitraum durch die Ferron-Methode und 27Al NMR bestätigt. Die Wechselwirkung
von Arsen mit Aluminiumoberflächen wurde auf molekularer Ebene mit erweiterter
Röntgenabsorptionsspektroskopie (EXAFS), Mikro-Röntgenfluoreszenz (
XRF) und
Elektronenmikroskopie (EM) untersucht.
Die Schlüsselfaktoren für die Effizienz einer Wasseraufbereitungsmethode, die
Arsenkonzentration und –speziierung, sowie die chemische Wasserzusammensetzung, wurden
für verschiedene Grundwässerarten aus dem Pannonischen Becken (Ostungarn und
Westrumänien), das am meisten Arsen-belastete Gebiet Europas, analysiert. Gemessene
Arsenkonzentrationen liegen zwischen < 0.5 und 240
g/L, mit As(III) als Hauptspezies.
Durchgehend hohe Arsengehalte wurden in methanogenen Wässern gefunden. Arsenmobilität

- ix -
Zusammenfassung

wurde durch die reduzierende Auflösung von Eisenmineralen, sowie Retentionsmechanismen

kontrolliert, letzteres gesteuert durch Sulfatreduktion und die Konzentration von organischem
Kohlenstoff.
Die Arsenentfernung beim Koagulierungsprozess mit PACl wurde auf die Abhängigkeit von
pH, Aluminium und Arsen Konzentration untersucht. Beste Resultate wurden mit 0.4-6
mmol/L Al/L bei pH 7-8 beobachtet. Die Laborresultate wurden durch Feldtests mit
Grundwasser von zwei unteschiedlichen Gebieten in Europa verifiziert: oxidiertes
Grundwasser in Bewässerungsbehältern von Nord-Griechenland und reduziertes artesisches
Wasser von West-Rumänien.
Um Al-Nanocluster auch in Durchfluss-Adsorptionssystemen einsetzen zu können, wurden im
Rahmen dieser Arbeit Polyaluminium-Granulat (PAG) von Al-Nanocluster Aggregaten in 1-3
mm Grösse hergestellt, und seine physikalischen Eigenschaften charakterisiert. Die Adsorption
von Arsen wurde durch Isotherme quantifiziert. Das Langmuir-Modell mit einer limitierten
Anzahl an Sorptionsplätzen beschreibt am besten die As(V)-Sorption, wogegen As(III)-
Adsorption linear ist und besser dem Freundlich-Modell folgt. Die Adsorptionskinetik wurde
durch eine Reaktion pseudo-zweiter Ordnung modelliert. Die Reversibilität der Sorption
zeigte, dass die Hälfte des adsorbierten As(III) mit synthetischem Wasser von PAG entfernt
werden konnte, As(V) hingegen kaum (5 %). Generell war die Entfernung von As(III) weniger
effizient als As(V) mit beiden Methoden.
Um die Effizienz und den Verbleib von Arsen während der Wasseraufbereitung vorhersagen
und verbessern zu können ist es essentiell, die Wechselwirkung mit der Adsorberoberfläche zu
verstehen. Die Anwendung von
XRF und EM haben zu neuen Einsichten in die
Adsorptionsprozesse von Arsen an Aluminiumgranulat geführt. Es konnte gezeigt werden, dass
As(III) einheitlich im ganzen Korn verteilt ist, während As(V) durch seine starke Affinität nur
in den Rand des Korns diffundiert. Der Sorptionsmechanismus von As(V) an Al Oberflächen
bei dem Koagulationsprozess sowie bei Adsorption an PAG wurde durch EXAFS als
innersphärische bidentat-binukleare Komplexbildung bestimmt, und durch die direkte
chemische Bindung von As(V) an Al-Oberflächen sehr robust ist.
Die Resultate dieser Dissertation haben gezeigt, dass Al-Nanocluster in einem Koagulations-
Kopräzipitationssystem fähig sind, Arsenatkonzentrationen von 2300
g/L bis unter den
Trinkwassergrenzwert zu reduzieren, und dass die Höchstleistung im pH-Bereich der meisten
Grundwässer liegt. Das Adsorbermaterial Polyaluminum-Granulat hat das Potential, höhere
Arsenatkonzentrationen stark zu reduzieren. Diese Arbeit ermöglicht es die Effizienz von Al
Sorptionsmitteln durch Anpassung der Korngrösse an kinetische Bedingungen zu erhöhen.

-x-
Chapter 1

Introduction

-1-
Introduction

1.1. Occurrence of arsenic in the environment

Arsenic is a metalloid, whose discovery is attributed to the German alchemist Albertus Magnus
in the 13th century (Emsley, 2001). Its name is derived from the Greek word „arsenikon“,
which was adapted from Persian and has the meaning „yellow orpiment“ (Bentley and
Chasteen, 2002).
Arsenic (As) is a major constituent in over 320 minerals in the Earth’s crust, most of them
associated with anoxic hydrothermal ore deposits or metamorphic or intrusive igneous rocks
(Henke and Hutchison, 2009). The most common As minerals are arsenopyrite (FeAsS) and
the As sulphides orpiment (As2S3) and realgar (As4S4), which was already described by
Aristotle around 400 B.C. (Bentley and Chasteen, 2002). Arsenic is also present in common
rock-forming minerals, either as part of the crystalline structure or as sorbed species. Highest
concentrations are found in sulphur minerals (e.g. pyrite), where it acts as a substitute for
sulphur due to the similar chemistry, and in iron oxides. Adsorption to the surfaces of hydrous
aluminum and manganese oxides, and to a lesser extent to clay and calcite are other
considerable sources for arsenic (Smedley and Kinniburgh, 2002).
The mobilization of As from sediments and soils is associated with various driving
mechanisms (Mukherjee et al., 2009). Under aerobic conditions, oxygen is consumed to
decompose organic matter bound to As (Mukhopadhyay and Sanyal, 2004), or to oxidize As-
bearing iron sulfides (Kim et al., 2000). In reducing environments, arsenic could be desorbed
from oxide surfaces due to the reduction of sorbed As(V) to the more mobile As(III), or by
reduction of Fe(III) and manganese oxides (Van Geen et al., 2004; Zobrist et al., 2000; Bose
and Sharma, 2002). Reductive dissolution of iron minerals via microbial activity (Nickson et
al., 1998; Lloyd and Oremland, 2006, Islam et al., 2004) result in the release of arsenic to the
aqueous phase. Under strongly reducing conditions sulfate is reduced to insoluble sulfide,
which precipitates. Arsenic is removed from solution by sorption to the surfaces of the formed
sulfide mineral (e.g. pyrite).
The presence of natural organic matter (NOM) in groundwater and in the sediment contributes
to retention, competition and complexation processes and therefore is of major importance for
arsenic mobility. NOM has high affinity for forming water-soluble complexes with arsenic
(Mukhopadhyay and Sanyal, 2004; Buschmann et al., 2006) and concurrently inhibits the
sorption of arsenic to solid phases. Additionally, it competes with arsenic for sorption sites on
iron (oxy)(hydr)oxides, leaving As in solution. Degraded organic matter promotes the

-2-
 Chapter 1

reductive dissolution of As-bearing iron (oxy)(hydr)oxides and impedes iron oxidation and
precipitation (Nickson et al., 1998; 2000; McArthur et al., 2004).

1.2. Arsenic speciation and toxicity

In natural waters, arsenic is present mainly in its inorganic form in the oxidation states +III and
+V, depending on the pH and the redox potential (Eh). The oxidized species arsenate, As(V), is
present as H2AsO4- below pH 6.9, and as HAsO42- at higher pH. Under reducing conditions,
As(III) is the stable arsenic form and dominant as the uncharged species H3AsO30 at pH
9.2
(Smedley and Kinniburgh, 2002). The pH-Eh diagram (Figure 1.1) gives an overview of
arsenic species in the aqueous phase. Both arsenic species exhibit a strong, pH-dependent
sorption affinity for oxyhydroxide minerals.
Arsenic is known as a toxic substance from ancient times. One of the earliest written records is
the arsenic poisoning of Britanicus by Nero 55 AD (Bentley and Chasteen, 2002). Apart from
the very toxic gaseous arsine (AsH3), the toxicity of inorganic arsenic species is generally
higher than that of organic arsenic forms such as arsenobetaine or arsenocholine (Tretner,

As(V)

As(III)

Figure 1.1: Eh-pH diagram of arsenic species in the aqueous phase at 25 °C and 1 bar total pressure
(Smedley and Kinniburgh, 2002). Molecule structures are modified from O’Day (2006).

-3-
Introduction

2003). Ingested inorganic arsenic is absorbed rapidly via the gastrointestinal tract into the
bloodstream, and transported to various organs. It is transformed by methylation to
monomethylarsenic acid (MMA) and dimethylarsinic acid (DMA), which is then excreted in
the urine (Chen et al., 1999). The toxicological effects on human health caused by chronic or
acute arsenic exposure include several types of cancer, diabetes, vascular disease,
hypertension, neurological disorder, reproductive problems and skin damage. In particular,
arsenic is considered by the WHO as a skin, bladder, lung, and kidney carcinogen (Hopenhayn,
2006). The arsenious acid (H3AsO3) impedes the function of enzymes by forming stable
compounds with their sulfohydryl groups (Tretner, 2003).
First wide-spread poisoning of bigger populations by geogenic arsenic were reported from
southwest Taiwan in 1920’s, where people suffer from vascular diseases that can lead to
gangrene of extremities, known as „Blackfoot Disease“ (Hopenhayn, 2006, Chen et al., 2003).
The World Health Organization (WHO) identified drinking water as the dominant source of As
intake, and the guideline value for maximum As concentration in drinking water was
recommended to be 10
g/L (WHO, 2003).

1.3. Global distribution of geogenic arsenic in groundwater

Millions of people worldwide are affected by elevated As concentrations in groundwater -

many of them in developing countries with no other water resource as an alternative. Aquifer
systems with natural As release have been reported in different environments from all
continents. High arsenic concentrations (> 50
g/L) in groundwater of arid and oxidizing
environments are found in Latin America (Mexico, Chile, Argentina), Nevada (USA), and
Mongolia (Smedley et al., 2003, Mukherjee et al., 2009). The development of anaerobic
conditions was associated with the mobilization of As from sediments and soils in many
countries in South and Southeast Asia: Bangladesh (McArthur et al., 2001; van Geen et al.,
2004), India (Chakraborti et al., 2003), Vietnam (Berg et al., 2008; Winkel et al., 2011),
Cambodia (Polya et al., 2005; Buschmann et al., 2007), China (Yu et al., 2007). In West
Bengal and Bangladesh, water supply was changed from surface water to water from shallow
tube wells to avoid microbiological contamination, but led to an „arsenic epidemic“ with
millions of people affected in the densely populated areas (Nickson et al., 1998; Chen et al.,
1999), and also crops are contaminated due to irrigation practices (Roberts et al., 2007; Garnier
et al., 2010).

-4-
 Chapter 1

Geothermal areas are known for As-enriched water, such as the Yellowstone National Park in
the West of the United States, the Donargarh rift belt of Central India (Mukherjee, 2009),
Greece (Kouras et al., 2007; Varnavas and Cronan, 1988) and Central Italy (Angelone, 2009).
Most of the contaminated aquifers are related to sediments of Quaternary age (< 1.75 million
years old) and aquifer depths are reported to vary between a few meters (Bengal Basin, Nepal,
Myanmar, Ghana) and 660 m (Xinjiang, China) (Mukherjee et al., 2009). The comparison
between the river basins of the Ganges-Brahmaputra, Mekong, and Red River showed
similarities in their geological characteristics, namely the rapid burial of organic-bearing and
young sediments, river drainage from weathered Himalayan rocks, and low basin gradients
(Charlet and Polya, 2006). Winkel et al. (2008) showed that Holocene deltaic and organic-rich
surface sediments are key indicators for the prediction of arsenic contaminated aquifers in
South-east Asia. Recently, geogenic arsenic distribution was modeled on global scale based on
geological, geochemical and geomorphological data (Amini et al., 2008). The obtained
probability map corresponded well with known contaminated areas and additionally depicts
regions with possibly high As levels in groundwater (Figure 1.2).

Figure 1.2: Modeled global probability of geogenic arsenic contamination for reducing and for high-
pH/oxidizing aquifer conditions (Amini et al., 2008).

-5-
Introduction

1.4. Study site: The East Pannonian Basin

The border region of Romania and Hungary is located in the East Pannonian Basin and covers
the Hungarian counties Hajdu-Bihar, Bekes and Csongrad, and the Romanian counties Bihor,
Arad, and Timis. It is part of the Tisza River basin, and the main rivers influencing the
hydrogeology of the area are the Körös and the Mures, both confluencing with the Tisza River
in Hungary.

1.4.1. Geology and Hydrogeology

The Pannonian Basin is an extensional rift basin in East-central Europe, which developed
during Early to Middle Miocene, about 19 million years ago (Horváth and Cloetingh, 1996),
and is bound by the Southern and Eastern Alps in the West, the Carpathian mountains to the
North and East, and the Southern Carpathians or Dinaric Alps to the South. It stretches over
Hungary, Croatia, Serbia, Slovakia and Western Romania. This area has a complex geological
history controlled by major tectonic activity, which resulted in the formation of many different
subbasins and varying subsidence rates (Trunko, 1996). From Middle Miocene, a brackish to
fresh water lake, the lake Pannon, formed in the depression, with fluviatile-dominated deltaic
depositions (Horváth and Cloetingh, 1996). From Late Miocene, the extensional movement
was terminated, accompanied by the uplifting of large areas and the development of subsiding
plains between the highs. Sedimentation rate was significantly higher than subsidence rates and
lacustrine-terrestrial environment prevailed over the basin in Pliocene (Horváth and Cloetingh,
1996). During Quaternary, reactivation of tectonic zones occured and fluvial sedimentation
started approximately 2.4 million years ago (Varsányi et al., 1997). Throughout the area, large
alluvial fans and gravel sheets developed.
Changes in subsidence rates in space and time throughout the Quaternary developed a complex
drainage pattern (Nadòr et al., 2007). The field area is located in a continuously subsiding area
between different alluvial fans from paleo rivers draining the North-east Carpathians and the
Apuseni mountains. Three types of groundwater flow were distinguished in East Hungary:
regional flow, mostly in the deeper horizon down to the middle of the Upper Miozene, local
flow in the shallowest layers, and intermediate flow, between the local and the regional flow,
mostly in Quaternary sediments (Varsányi and Kovács, 2006). The deepest aquifer has a depth
of about 2500 m (Varsányi et al., 1999). Within the Pleistocene sediments groundwater consist
of two main types: a calcium and magnesium-bicarbonate water type in coarser sediments, and
a sodium-bicarbonate water held within the finer grained sediments (Varsányi et al, 1997).

-6-
 Chapter 1

1.4.2. Arsenic contamination

Elevated arsenic concentrations have been reported from groundwater of the Pannonian Basin
since the 1940s in the lowland areas of the Danube and the River Tisza. A survey undertaken
by the Hungarian National Institute of Public Health (OKI) in 1980-1981 found water in 148
settlements in 6 Hungarian counties exceeded 50
g/L, predominantly in the South-east of the
country with about 400'000 people that might be affected (Csalagovitis, 1999). High As
concentrations have been found from aquifers at 300 - 400 m depth and are related to
Quaternary sediments of deltaic and fluviatil background (Varsányi and Kovács, 2006). Most
of the published arsenic-related groundwater studies are geographically restricted to Hungary.
Only a few surveys are available for the Romanian side of the Pannonian Basin, which covers
the counties Bihor, Arad and Timis with an area of 23’995 km2. Arsenic concentration
measured in water of 23 artesian wells in this area range between 13
g/L and more than 200

g/L (Baciu et al., 2006). In some rural areas, water from artesian village wells presented a big
part of the daily water supply for cooking, drinking, and washing due to the lack of a water
pipe system. Gurzau and Gurzau (2001) estimated that about 36’000 people are thought to be
exposed to arsenic in drinking water higher than 11
g/L in West Romania. A strong
correlation of the intake of arsenic-contaminated water and arsenic concentrations in urine
were found in Hungary, Romania and Slovakia (Lindberg et al., 2006), and increased cases of
skin, bladder and kidney cancers have been found in As-affected areas of these countries
(Fletcher et al., 2008).

1.5. Methods for arsenic removal from groundwater

Natural arsenic contamination of subsurface water bodies due to the mobilization from
minerals and rocks can affect shallow and deep aquifers, and often spreads over a wide area.
The source therefore cannot be eliminated, and groundwater has to be treated at the point of
use. Four principal physico-chemical processes are generally used for the on-site removal of
contaminants from groundwater: coagulation-precipitation, adsorption, ion exchange, and
membrane filtration (Gray, 2005). Table 1.1 summarizes advantages and constraints of arsenic
removal with the presented methods using different materials. Regarding the arsenic removal
efficiency, adsorption on activated alumina and hydrous ferric iron (HFO), as well as
coagulation with alum belong to the best treatment methods.

-7-
Introduction

1.5.1. Coagulation and coprecipitation

This removal process is started by the addition of a coagulant to the contaminated water. The
coagulant hydrolizes and forms flocs in the water, and arsenic co-precipitates with or adsorbs
at the surface of the coagulant, and the aggregates settle by gravity or get removed by
filtration. In general, best removal is achieved at the pH at which the precipitated species is
least soluble. In addition, the chemistry of the contaminated water plays an important role, as
other elements can compete for the sorption sites on the flocs. There are two main coagulants
used for arsenic removal from groundwater: alum (KAl(SO4)·12H2O) and ferric chloride
(FeCl2). For both of them, arsenic in the pentavalent oxidation state could be removed
successfully with efficiencies > 95%, and As(III) is removed to a lesser degree under the same
conditions. It was demonstrated that the best pH range for alum coagulation is between 5 and
7, and for ferric coagulation it is 5 to 8 (Hering et al., 1996; Edwards, 1994). Sulfate
significantly reduced As(III) removal with ferric chloride below pH 7, whereas calcium
increased As(V) removal at higher pH (> 7) (Hering et al., 1996). The polymeric coagulation
agents of ferric sulfate or chloride as well as aluminum sulfate or chloride were more efficient
for arsenic removal than the non-polymeric coagulants of the same material at pH 5.5 – 8.0
(Fan et al., 2003).

1.5.2. Adsorption

Removal by adsorption takes place by the exchange of arsenic with the surface hydroxides of
the adsorbent. When the surface sites are saturated, the material needs to be replaced or
regenerated. Most adsorption materials for As are based on aluminum and iron due to their
natural sorption affinity and widespread availability. The two most common are activated
alumina (AA), an amorphous aluminum oxide, and hydrous ferric oxide (HFO), a natural
adsorber material, occurring as coatings on the surface of sediment, mineral, and soil particles.
The removal with activated aluminium is very sensitive to pH, with an optimum pH range for
arsenic removal of 5.5 – 6.0. Activated alumina material may foul during the treatment, with
mica and silica being identified as particular foulants (EPA, 2000). For HFO, best pH for
adsorption of As(III) and As(V) as been observed to be 7 and 4, respectively (Wilkie and
Hering, 1996). Zero-valent iron (ZVI) has the advantage that it effectively removes As(III) and
As(V) by adsorption and/or precipitation with Fe(II) and Fe(III) oxides and hydroxides formed
by the oxidation of ZVI on its surface (Nikolaidis et al., 2003), with greater removal achieved
at high temperatures (40°C) (Tyrovola et al., 2006). The presence of silicate, phosphate,

-8-
 Chapter 1

chromate and molybdate inhibited the removal of arsenic, with phosphate being the strongest
inhibitor (Su and Puls, 2001).

1.5.3. Ion exchange

Many geogenic contaminants can effectively be removed from solution by electrostatic

attachment. Synthetic resins are chosen depending on the contaminant and the chemistry of the
water. The contaminated water flows steadily through a bed of ion exchange resin beads until
all the ions on the surface of the resin are exchanged with the contaminating ions. The resins
are regenerated by rinsing with a solution of high resin-ion concentration. Arsenate is
predominantly removed by the ion exchange process, whereas it is not efficient for the
uncharged arsenite species. Strongly basic resins are usually not pH sensitive in the range of
pH 6.5 – 9 (EPA, 2001), which is commonly the pH range of water with natural As
contamination. A range of contaminants and metals can be treated by most strong-base anion-
exchange resins, and therefore the effectiveness for As treatment highly depends on the other
ions in the solution. The presence of nitrate, phosphate and bicarbonate can inhibit the sorption
of arsenic. Especially high concentrations of sulfate and total dissolved solids can cause
problems.

1.5.4. Membrane filtration

Membranes act as filter barriers, allowing some of the constituents of water to pass through,
while keeping others out. The driving force behind this process can be pressure, concentration,
electrical potential and temperature. Pressure-driven membranes are frequently used for the
removal of geogenic elements, and can be classified in high-pressure and low-pressure
membranes, based on the pore size of the membrane and the resulting pressure. Microfiltration
(MF) (pore size between 0.015 μm and 1.5 μm) and ultrafiltration (UF) (pore size between
0.0015 μm and 0.13 μm) are low-pressure processes with 0.3 to 7 bar, removing contaminants
mainly by physical sieving. Nanofiltration (NF) (pore size between 0.7 nm and 10 nm) and
reverse osmosis (RO) (pore size between 0.1 nm and 10 nm) with pores in the range of a metal
ion, create high pressures of 3.5 – 10 bar (Gray, 2005; EPA, 2000). They remove contaminants
mainly by chemical diffusion. Due to the small pore size high-pressure membranes can remove
a range of elements from the water, but the high energy consumption at high pressures
prevents it from installation at large scale. Microfiltration (MF) and ultrafiltration (UF) can be

-9-
Introduction

used as a pretreatment for high turbidity water to remove suspended solids and particulate
organic matter.
In general, the water temperature and the quality of the source water are of high importance for
a successful removal with all membranes. Low water temperatures decrease the water flux. In
order to maintain the same recovery with lower temperatures, more surface area must be given
or the pressure must be increased. The chemistry of the inflowing water has to be known in
order to decide for a membrane process. Calcium can be rejected on low-pressure membranes,
causing scale and precipitation on the membrane surface. Metals and organic compounds can
cause fouling of the membrane. Furthermore the removal of a membrane depends on the
characteristics of the membrane itself: the material chemistry, and here mainly surface charge
and hydrophobicity, as well as the membrane configuration such as the pore size, have a great
effect on the rejection of a contaminant. Reverse osmosis (RO) and nanofiltration are suitable
membrane processes for the removal of both arsenic species (Jekel and Amy, 2006). The
removal of As(V) (> 95%) is better than for As(III) (5 – 75%) at circum-neutral pH with RO.
Turbidity, iron, silica and manganese have a negative influence on As removal by RO, while
high DOC content only had a slight effect. Therefore, pretreatment of natural groundwater is
necessary prior to application of RO (EPA, 2000).

1.6. Nanomaterials

Within the last decade, nanomaterials have been produced and adapted for water treatment.
The small size of about 1-100 nm increases the surface-to-volume ratio, and gives hope for
higher arsenic removal capacities than with conventional materials (Theron et al., 2008). In
addition, nanoparticles possess high surface reactivity and unique catalytic activity (Auffan et
al., 2009). An overview of the performance of different nanomaterials is given in Table 1.2.
Nanomaterials have the potential to reduce the production of waste, to use smaller amounts of
resources, and to provide clean drinking water, with the great advantage to have surfaces that
have the ability to be tailored to increase adsorption capacities or to recognize specific
contaminants (Bottero et al., 2006). The process of arsenic removal with all investigated
nanomaterials is based on the adsorption in suspensions or fixed on a support. Particularly two
iron-based nanomaterials stand out due to their high uptake capacity. For As(III) removal,
nano-maghemite (NMag) has been shown to be most effective (Auffan et al., 2008). HFO fixed
on fibrous ion exchanger (FIBAN®) could remove more than 1200
mol/g of both As species
and could be implemented easily in households or, on larger scale, in water-treatment plants.

- 10 -
 Chapter 1

The efficiency of nanocrystalline titanium dioxide was strongly reduced by silicate, phosphate
and bicarbonate. However, the high reactivity and small size of nanoparticles also comes with
higher potential risks due to a better uptake and interaction with biota (Auffan et al., 2009).
While Al nanomaterials have not been found to expose higher toxicity than non-nano Al (Zhu
et al., 2008), TiO2 nanoparticles were found to be potentially genotoxic (Ghosh et al., 2010).

1.7. Objectives

Within the frame of the European Marie-Curie research training network AquaTRAIN, this
dissertation covers two main areas: the groundwater chemistry and the fate of arsenic in the
Eastern Pannonian Basin, and the evaluation of Al nanoparticles as removal agent for arsenic
in groundwater. The main goals of this work were to:
• Identify the speciation and mobilization mechanisms of arsenic in different
groundwater types in the East Pannonian Basin.
• Explore the removal efficiency of aluminum nanoclusters for the two inorganic As
species As(III) and As(V)
• Examine the molecular structure of As-Al complexes

Table 1.1.: Arsenic removal techniques (after Mertens, 2010) (n.d.: no data found)
Treatment Material Removal Best pH Constraints References
process window
Coagulation/ Alum As(V): > 95% 5-7 pH window Hering et al. (1996),
Precipitation Sorg & Lorgsdon
(1978)
Ferric As(V): > 95% 5-8 Inhibited by sulfide Hering et al. (1996),
chloride Sorg & Lorgsdon
(1978)
Adsorption AA As(V): 88% 5.5 - 6 Fouling, inhibited by US EPA (2000),
As(III): 60% DOC, sulfide and chloride AWWA (2002)
ZVI As(V): 60% 4 Inhibited by phosphate, Bang et al. (2005),
As(III): 95% silicate Nikolaidis and
Tyrovola (2006), Su
and Puls (2001)
HFO As(V): 100% 4-7 Inhibited by sulfate Wilkie and Hering
As(III): 80% 7-9 (1996)
Ion exchange Strong n.d. 6.5 - 9 Inhibited by nitrate, EPA (2001)
base resins phosphate, carbon
bicarbonate, TDS
Membrane NF As(V): 85 90% n.d. Fouling, high energy AWWA Research
filtration As(III): 12% consumption Foundation (1998)
RO As(V): > 95% n.d. Fouling, high energy AWWA Research
As(III): 5–75 % consumption, inhibition of Foundation (1998)
Turbidity, silicate, Fe, Mn,
low recoveries

- 11 -
 Introduction

Table 1.2: Summary of listed nanoparticles (modified from Piazzoli, 2011), n.d.: no data found.
Nanoparticles Size and pH Uptake Regeneration Advantages Disadvantages Reference
characteristics dependence capacity properties

mol As/g
Iron nZVI 10-100 nm; 4-10 46.7 n.d. Possible in situ Strongly impeded Kanel et al. (2005)
pore size 20 nm remediation application by H4SiO40,
As(III) removal H2PO42- and HCO3-
Nmag 6 nm n.d. 2300 n.d. As(III) removal No information on Auffan et al. (2008)
pH and anion
dependence
HFO FIBAN® 10-20 nm As(V): 5.64 As(V): 1400- n.d. Can be implemented at Strongly impeded Vatutsina et al. (2007)
fixed on fibrous As(III): 9 1650 household level by phosphate
ion-exchanger As(III): 1200-
1500
ArsenXnp 2-5 nm wide pH range 109 87.9% using 2% No backwashing Silica and Sylvester et al. (2007)
impregnated into NaOH/ needed phosphate

- 12 -
300-1200
m 3% NaCl interference
polymeric beads solution
Titanium Powder 6 nm < 8.5 500 n.d. As(III) oxidized to Strong competition Pena et al. (2005)
As(V) in air sunlight with silicate,
system phosphate and
bicarbonate
Fixed on TiO2-SiO2-PAN As(V): 6 As(V): 470 < 11% using 1 M Simply production Competition with Nilchi et al. (2010)
support As(III): 8 As(III): 390 H2SO4 solution Improved permeability PO43-, Fe(III) and
Stable in water Co(II)
Aluminum Al-loaded 0.35 nm pore size 3-10 As(V): 75 71% using 0.04 Easy availability of Strong competition Xu et al. (2002)
Shirasu- M NaOH, adsorbent with phosphate
zeolite restored to 94% Slight pH dependence
with aluminum Possible to recharge
sulfate solution
Zirconium 20-30 nm n.d. n.d. n.d. >98% As removal Low As removal Hristovski et al. (2007)
Competition with
carbon dioxide and
carbonate
Nickel 10-20 nm >8.5 n.d. n.d. >98% As removal Dissolves at pH Hristovski et al. (2007)
<8.5
 Chapter 1

1.8. Thesis Outline

A brief overview of each manuscript chapter is given below. Four of the chapters have been
published or submitted for publication, the remaining chapter is in the progress to be
submitted.

Chapter 2: Al nanocluster in coagulants and granulates; application in arsenic removal

from water
This short chapter was published as an invited article under the rubric heading “science and
career” in Environmental Science and Biotechnology in 2011 and elucidates some of the basic
principles and applications of the presented work. It brings the study in context with the
European Marie-Curie research training network AquaTRAIN, which funded a large part of
the actions and mobility during this project.

Chapter 3: Geochemistry and arsenic behaviour in groundwater resources of the

Pannonian Basin (Hungary and Romania)
The natural mobilization mechanism of arsenic in the Eastern part of the largest sediment basin
in Europe was investigated by a groundwater study, which was published in Applied
Geochemistry in 2010. The groundwater from shallow and deep wells of the complex and
interconnected aquifer system along the Hungarian-Romanian border was sampled in three
field campaigns and could be categorized in three major water types. High arsenic
concentrations were associated with generally reducing groundwaters, and the release of As
was coupled with the reductive dissolution of As-bearing iron oxides. However, retention
mechanisms by forming As-sulfides or the sorption of As onto Fe-sulfides in the presence of
sulfide were found to play a major role for As mobility.

Chapter 4: Polyaluminum chloride with high Al30 content as removal agent for arsenic-
contaminated well water
This chapter was published in Water Research and describes the removal of inorganic arsenic
species As(III) and As(V) from water by the coagulation-coprecipitation process using
polyaluminum chloride with high content of Al nanoclusters as coagulant. The Al species in
suspensions and precipitates of polyaluminum chloride have been identified and Al
nanoparticles corresponded to > 90 % and 50 % of the total Al share, respectively. Titrations of
As-enriched Al solutions revealed that As(III) is not as efficiently eliminated as As(V), but
best removal for both species is achieved at pH 7-8. Field tests in natural groundwater from
Romania and Greece were in agreement with laboratory studies.

- 13 -
Introduction

Chapter 5: Adsorption of arsenic on polyaluminum granulate

The coagulation removal process acquires time for the settlement of particles and is
inconvenient for filter systems with constant flow at small scale like artesian wells. For the
packing in column filters, a solid-state polyaluminum granulate (PAG) was produced from
polyaluminum chloride precipitates, and characterized. The adsorption of As(III) and As(V) on
PAG was described by adsorption isotherms. Kinetic investigations revealed that As(III)
reached equilibrium in 40 and most As(V) was removed after 20 hours. The use of
X-ray
fluorescence and electron microscopy techniques allowed to determine the distribution of
arsenic in the grain. The strong affinity of As(V) for the polyaluminum granulate resulted in
quantitatively high As accumulation on a small rim close to grain’s surface, while uncharged
H3AsO3 was distributed over the entire PAG area at lower concentrations. This manuscript was
submitted to Environmental Science and Technology.

Chapter 6: Arsenate adsorption on Al nanoparticles and Al-based sorbents during water

treatment – an EXAFS study
Al nanoclusters exhibit high sorption capacity for arsenate. This chapter compares the reactive
surface sites and the As(V) surface coverage of aluminium nanoclusters and gibbsite. EXAFS
studies of the molecular structure of arsenic adsorbed to gibbsite and various Al nanoparticle-
based removal agents showed no major differences in the atomic environment of arsenic. For
all adsorbents, two Al atoms surround the As atom, forming a bidentate-binuclear complex.
Theoretical As-to-Al ratios were calculated for Al nanoclusters and gibbsite, showing that Al
nanoclusters exhibit 100 As sorption sites per mol Al more than gibbsite.

- 14 -
Chapter 2

Al nanoclusters in coagulants and

granulates: application in arsenic removal
from water

Jasmin Mertens

Reviews of Environmental Science and Biotechnology (2011),

10(2): 111-117

- 15 -
Al nanoclusters in coagulants and granulate

Abstract

The contamination of drinking and irrigation water by arsenic is a severe health risk to millions
of people, particularly in developing countries. Arsenic treatment methods therefore need to
advance to more durable and cost-effective solutions. In recent years, the unique properties of
nanomaterials have received much attention in water treatment research, and their properties
(e.g., high number of reactive surface binding sites) may make them suitable for arsenic
removal. The aluminum nanoclusters Al13 (AlO4Al12(OH)24H2O127+) and Al30
(Al2O8Al28(OH)56(H2O)2618+) have high specific surface charge, deprotonate over a wide pH
range and exhibit a high reactivity due to a great number of OH- and H2O groups. This
contribution evaluates these chemical properties of aluminum nanoclusters and their efficiency
for water treatment, particularly for arsenic removal. It assesses the advantages and constraints
when applied in an industrially produced aluminum coagulant or in Al granulate during water
treatment.

2.1. Introduction

The abundance of the carcinogenic arsenic in drinking and irrigation water is a severe and
widespread health threat to more than 100 million people worldwide, especially in South and
Southeast Asia (Charlet and Polya 2006, Winkel et al. 2008). Arsenic enters groundwater
either naturally by dissolution of arsenic-bearing minerals or by anthropogenic sources such as
mining industry or metal processing plants. Arsenic in natural waters is mostly present in
inorganic form as trivalent arsenide (As(III)) under reducing conditions or as pentavalent
arsenate (As(V)) under oxidizing conditions (Smedley and Kinniburgh 2002).
For arsenic treatment technologies to become a realistic option for more people, new
developments need to address the challenges of reducing both treatment costs and waste
generation while simultaneously enhancing durability and efficiency. Recently, inorganic
nanoparticles have received increasing interest as removal agents for arsenic from water due to
their high surface area-to-mass ratio, chemical properties and high surface reactivity (Auffan et
al. 2008). Nanoparticles are added to water as powder or are incorporated in conventional
remediation methods to increase their efficiency, e.g. in membranes or polymers (Nilchi et al.
2011, Theron et al. 2008).
Polyaluminum chloride (PACl) is a coagulant widely used in water remediation. It generally
consists of Al species such as [Al2(OH)2]4+, [Al3(OH)4]5+, [Al8(OH)20]4+ and

- 16 -
 Chapter 2

[AlO4Al12(OH)24H2O12]7+ (Fan et al, 2003). Commercial polyaluminum coagulant forms large

Al nanoclusters when produced with high total Al concentrations at high temperatures for a
long time (Chen et al. 2006). A high content of nanosized aluminum clusters may increase the
efficiency and reduce the costs of arsenic treatment.
Two main robust aluminum nanoclusters exist in aqueous PACl solutions: Al13
(AlO4Al12(OH)24H2O127+) and Al30 (Al2O8Al28(OH)56(H2O)2618+) are the largest aqueous
aluminum hydroxide complexes with 1 nm diameter and 1-2 nm length. Al13 is a cluster of four
Al trimers placed around one tetrahedral coordinated aluminum atom. Al30 was characterized
by 27Al nuclear magnetic resonance (NMR) to be composed of two Al13 clusters connected to
each other by four Al(O)6 octahedra (Allouche et al. 2000) (Figure 2.1). These Al nano
complexes stand out due to their high specific surface charge (+7 for Al13 and +18 for Al30),
deprotonation over a wide pH range (Furrer et al. 1992; Casey et al. 2005), and high specific
surface area depending on pH and [OH]/[Al] ratio (Bottero and Bersillon 1988). In addition,
Al13 and Al30 exhibit a high reactivity due to a great number of hydroxide ions and water
groups. The presence of Al30 in polyaluminum coagulants enhances turbidity due to strong floc
formation (Chen et al. 2006). These characteristics make Al nanoclusters promising removal
agents, but the extent to which Al13 and Al30 bind arsenic is poorly known. This Marie-Curie
project aims to evaluate how the aqueous chemistry of Al nanoclusters affects the removal of
the two inorganic arsenic species As(III) and As(V) from water when used as main constituents
in polyaluminum chloride (PAClAl30) and PAClAl30 – based granulates.

Figure 2.1: Structure of Al13 (AlO4 Al12(OH)24 H2O127+) and Al30 (Al2 O8Al28(OH)56(H2O)2618+): Al13 is a
tretrahedral coordinated aluminum (Al(O)4, shown in green) surrounded by four trimers of octahedral
Al. Al30 is composed by two Al13 Keggins, linked by a group of four octahedral coordinated Al (adapted
after Casey et al. 2005).

- 17 -
Al nanoclusters in coagulants and granulate

2.2. Working principles

After dissolution in water, industrially manufactured PACl with the formula AlCl0.5(OH)2.5
produces an aluminum solution with a content of > 70 % Al nanoclusters. Depending on the
pH of the solution, the Al clusters undergo coagulation, form aggregates, and precipitate
resulting in the production of an amorphous solid. These processes are driven by the
protonation state of dissolved Al nanoclusters - the more the Al clusters deprotonate, the more
their surface charge decreases and the electrostatic repulsion between single Al clusters
decreases. Lower repulsion enhances the aggregation of Al nanoclusters. The structure of Al13
and Al30 contains a high number of OH- and H2 O groups, and deprotonation proceeds over
almost two pH units from pH 5 to 6.8 (Furrer et al. 1992; Casey et al. 2005). The state of
protonation of the PAClAl30 solution is obtained from base titration data by taking into account
the initial PACl solution volume, the total Al30 concentration, and the concentration and
volume of the added base. Zero charge of the Al30 nanocluster is reached when all 18 positive
charges are lost, and the state of protonation equals 0. The pH where the electrical charge
density on the Al surface is zero (pH point of zero charge; pHPZC) is determined by titration to
be at 6.7 (Figure 2.2).
By use of a laser with
= 633 nm the formation of Al clusters > 633 nm by aggregation in a
PAClAl30 solution with 15 mM Al(tot) was observed after the pH reached a value of 5.5.
Already at pH 6.5, where Al nanoclusters keep a charge of +2, a small amount of aggregates
settles. However, the bulk of aluminum particles settle in big white flocs above a pHPZC of 6.7,
where most of the Al can be separated by sedimentation or centrifugation. A schematic view
illustrates the Al removal processes in dependence of pH for PACl with high Al30 content
(Figure 2.2).
Arsenate adsorbs to Al30 and Al13 clusters during coagulation and aggregation by ligand-
exchange reactions at the hydroxy and oxygen groups of the aluminum, and together they form
soluble complexes. Because the PAClAl30 solution deprotonates from pH 5 to 7, As(V)
adsorption can take place over a wide pH range. Removal of As(V) occurs with the
precipitation of Al-As(V) complexes at a pH > 6.5. The highly charged Al nanoclusters are
strong adsorbents for arsenate ions. However, the neutral arsenite H3 AsO3 has a weak affinity
for Al. Adsorption most likely occurs to low or uncharged Al clusters due to electrostatic

- 18 -
 Chapter 2

Figure 2.2: Deprotonation of a PAClAl30 solution with [Altot] = 15 mM and Al removal processes with
pH. Optimum Al removal takes place above pHPZC at 6.7.

interaction, Van-der-Waals forces and dipole-dipole interaction. As(III) coprecipitates with Al

flocs above pH 7 where solid and liquid phase can be separated effectively.
In order to minimize the risk of aluminum contamination of treated water and to avoid a
filtration step for small Al particles, granulates of 1-2 mm in size were generated from highly
concentrated PAClAl30 solutions after repetitive aggregation, precipitation, washing and
consecutive drying. Dried material was grinded and compressed using a pill press operation.
The adsorption of arsenate occurs mainly on the surface of PAClAl30 granulate as As(V) has a
strong sorption affinity to the aluminum (Figure 2.3a). Contrary, due to weak chemical
interaction with the Al solid, As(III) diffuses into the grain pores, where it is adsorbed, and
therefore it is evenly distributed over the entire grain (Figure 2.3b).

- 19 -
Al nanoclusters in coagulants and granulate

a b

Figure 2.3: a) As(V) and b) As(III) adsorption processes on Al granulate (not to scale).

2.3. Applications

Within the framework of the European project AquaTRAIN, polyaluminum chloride and Al
granulate with a high content of Al nanoparticles were applied to groundwater from highly
contaminated aquifers, but with different groundwater chemistry and usage at two field sites.
The Pannonian Basin, a sedimentary basin stretching over the countries of Serbia, Hungary,
Romania, Slovakia, and Croatia, is the largest area in Europe where groundwater is affected by
geogenic arsenic contamination. Within this region, an estimated one million people are or
have been exposed to water containing 5 to 30 times more arsenic than the World Health
Organisation (WHO) and EU drinking water guide value of 10 ppb, mainly present as As(III)
(Rowland et al. 2010; Gurzau and Gurzau 2001, Varsanyi and Kovacs 2006). This is a
particular risk because many people in this area still rely on unfiltered water. For example,
local artesian wells are the main source for drinking and household water in some villages in
Western Romania.
In Northern Greece, groundwater in geothermal fields contains up to 3000 ppb As. It is mostly
used for irrigation in agriculture during dry summer months, and for this purpose stored in
open water tanks, where As(III) undergoes oxidation in contact with air (Casentini et al. 2009).
Effective application of arsenic treatment systems in natural water needs to take into account
the specifics of arsenic chemistry (e.g., As speciation) and the chemical properties of the bulk
water that might influence the removal efficiency, such as total dissolved solids, organic
substances, pH and other elements that might compete with As for sorption sites. Al
nanoclusters as main constituents of an aluminum coagulant and granulate show advantages
and limitations for As removal (Table 2.1). This work showed that PAClAl30 has the potential
to remove As(V) to below the EU drinking water directive. As(III) removal, however, is less

- 20 -
 Chapter 2

efficient, because the neutral As(III) species has a low affinity for Al. Therefore, oxidation of
As(III) before removal is essential. The optimum pH for As removal with PAClAl30 is at or
above pH 6.7 due to the deprotonation of Al30. Conventional PACl achieved the highest arsenic
removal at pH 5.5 (Fan et al. 2003). The typical groundwater pH ranges from 6.5 to 8.5, and
therefore no pH adjustment is needed for groundwater treatment with Al30. Although field
applications showed that Al precipitation at this pH is effective with no remaining Al measured
in the treated solution, some laboratory tests resulted in aluminum concentrations slightly
above the WHO drinking water guide value of 200 ppb (WHO 1998).
The use of solid granulates allows water to be treated in a flow-through system without the
need for solid-liquid separation. Upward water flow through a column packed with adsorbent
material increases the contact time with the granulates (Figure 2.4). Polyaluminum granulate

Table 2.1: Comparison between PAClAl30 and PAClAl30-based granulate in water treatment applications.
As removal As Removal Advantages Constraints Application
mechanism
PAClAl30 Coagulation/ As(V): 90- • Easy applicable Al Standing water
coprecitation 100% • No pH adjustment remaining bodies/
As(III): < 40% necessary in treated Specialized
water treatment
facilities
PAClAl30 Adsorption As(V): 70-99% • High As uptake Inhibited Flow-through
Granulate As(III): 33-57 capacity by iron, systems
% • No Al in treated phosphate
water

Figure 2.4: Single-column

set-up for the application of
PAClAl30-based granulate in
a flow-through system for
As-contaminated
groundwater. Blue arrows
indicate water flow.

- 21 -
Al nanoclusters in coagulants and granulate

with a high content of Al nanoclusters removed 70 % to 99 % As(V) in a concentrations range

of 20 ppb - 200 ppm. As(III) is removed from 57 % at low concentrations (23 ppb) to 33 % at
high concentrations (200 ppm). Field tests showed furthermore that iron and phosphate
compete for sorption sites vs. As(III) on polyaluminum granulate.
High contents of Al nanoclusters improved the coagulation and precipitation process while
successfully reducing arsenate concentrations below the required limit. In water treatment
facilities operating with coagulation/precipitation, the application of PAClAl30 might help
reducing treatment steps -and therefore treatment costs- by removing the pH adjustment.
Application of PAClAl30 in standing water bodies like open irrigation tanks in Greece is easy
and convenient. For running water systems like water taps or artesian water wells in Romania,
Al30-rich granulate is a better alternative.

2.4. Training of young researchers

The Marie-Curie research training network (MCRTN) AquaTRAIN (www.aquatrain.eu)

focused on the research of geogenic contamination and remediation of groundwater and soils
within Europe and provided direct knowledge transfer to involved early-stage researchers
(ESRs) and experienced researchers (ERs) by training and cross-disciplinary workshops
(Polya, 2010). It incorporated the expertise from 15 leading research institutes across 10
European countries structured in four work packages (Table 2.2) and was funded for four years
under the 6th European Framework Programme.
Pursuing my PhD within AquaTRAIN gave me the opportunity for international scientific
exchange and involvement in various research projects. In addition, participation in project
management, interdisciplinary teamwork and the contact with political stakeholders was very
valuable for personal skills development.
The encouragement of mobility made it possible to gain working experience in different
research groups throughout Europe. A research visit of six months at the European Centre for
research and education in Geosciences and Environment (CEREGE, France) gave me insights
into X-Ray spectroscopic studies of As-Al interactions and the characterization of Al material
by 27Al NMR. During another six months period at the Faculty of Environmental Sciences of
the University Babes-Bolyai (Romania) the application and installation of a water treatment
system at a village well was realized. These trainings widened my experience in the field of
water treatment processes from the nano- to macroscale.

- 22 -
 Chapter 2

Table 2.2: AquaTRAIN coordinator and partner institutions and their specialization in the project.
Institute Specialization City, Country Website
AquaTRAIN Co-ordinator
University of Manchester Microbiology and As Manchester, UK http://www.manchester.ac.
(UNIMAN) mobility uk/
Host Institution
Dept. Environmental Science, Al nanocluster Zurich, http://www.env.ethz.ch/
Swiss Federal Institute of chemistry Switzerland
Technology (ETH)
Collaborators
Babes-Bolyai University (UBB) Geology and field Cluj-Napoca, http://www.ubbcluj.ro/
studies Romania
Centre National de la Recherche Nanoparticle research, Aix-en- http://www.cerege.fr/
Scientifique Scientifique aluminum Provence/Paris,
(CEREGE/ CNRS) characterization France
Swiss Federal Institute of Aquatic Water sampling and Dübendorf, http://www.eawag.ch/
Science and Technology analysis Switzerland
(EAWAG)
Other AQUATRAIN Partners
University Joseph Fourier-Grenoble As speciation and Grenoble, France http://www.ujf-
1 (UJF) cycling grenoble.fr/
University Utrecht (UU) Reactive transport Utrecht, The http://www.geo.uu.nl/Rese
modeling Netherlands arch/Geochemistry/
University of Girona (UdG) Selenium cycling Girona, Spain http://www.udg.edu/
Central European University (CEU) Science-Policy Budapest, http://www.ceu.hu/
interactions Hungary
JRC-Institute for Environment and Mapping Ispra, Italy http://eusoils.jrc.ec.europa
Sustainability (EC-DG, JRC) .eu/
Technical University of Crete Inorganic water Chania, Greece http://www.herslab.tuc.gr/
(TUC) treatment
Bureau de Rechereches Isotope analysis Orleans, France http://www.brgm.fr/
Geologiques et Minieres (BRGM)
De Montfort University (DMU) As biochemistry and Leicester, UK http://www.dmu.ac.uk/
uptake in humans
The University Court of the Speciation techniques Aberdeen, UK http://www.abdn.ac.uk/
University of Aberdeen and analytics
(UNIABDN)
Vlaamase Instelling voor Biological remediation Mol, Belgium http://www.vito.be/
Technologisch Onderzoek (VITO) technologies

- 23 -
Chapter 3

Geochemistry and arsenic behaviour in

groundwater resources of the Pannonian Basin
(Hungary and Romania)

Helen A.L. Rowland, Enoma Omoregie, Romain Millot, Cristina

Jimenez, Jasmin Mertens, Calin Baciu, Stephan J. Hug, Michael Berg

Applied Geochemistry (2011), 26: 1-17

This chapter was a published team effort of AquaTRAIN. The author

contributed to the planning and conductance of water sampling
campaigns, analysis, and data evaluation.

- 24 -
 Chapter 3

Abstract

Groundwater resources in the Pannonian Basin (Hungary, Romania, Croatia and Serbia) are
known to contain elevated naturally occurring arsenic (As). Published estimates suggest
nearly 500,000 people are exposed to levels greater than the EU maximum admissible
concentration of 10
g/L in their drinking water, making it the largest area so affected in
Europe. In this study, a variety of groundwaters were collected from Romania and Hungary to
elucidate the general geochemistry and identify processes controlling As behaviour.
Concentrations ranged from <0.5 to 240
g/L As(tot), with As predominantly in the reduced
As(III) form. Using cluster analysis, 4 main groups of water were identified. Two groups (1
and 2) showed characteristics of water originating from reducing aquifers of the area with
both groups having similar ranges of Fe concentrations, indicating that Fe-reduction occurs in
both groups. However, As levels and other redox characteristics were very different. Group 1,
indicative of waters dominated by methanogenesis contained high As levels (23 to 208
g/L,
mean 123
g/L), with group 2 indicative of waters dominated by SO42--reduction containing
low As levels (<0.5 to 58
g/L, mean 11.5
g/L). The remaining two groups were influenced
either by (i) geothermal and saline or (ii) surface contamination and rain water inputs. Near
absence of As in these groups, combined with positive correlations between 7Li (an indicator
of geothermal inputs) and As(tot) in geothermal/saline influenced waters indicate that
elevated As is not from an external input, but is released due to an in-aquifer process.
Geochemical reasoning, therefore, implies As mobilisation is controlled by redox-processes,
most likely microbially mediated reductive dissolution of As bearing Fe-oxides, known to
occur in sediments from the area. More important is an overlying retention mechanism
determined by the presence or absence of SO42-. Ongoing SO42--reduction will release S2-,
removing As from solution either by the formation of As-sulfides, or from sorption onto Fe-
sulfide phases. In methanogenic waters, As released by reductive dissolution is not removed
from solution and can rise to the high levels observed. Levels of organic C are thought to be
the ultimate control on the redox conditions in these 2 groups. High levels of organic C (as
found in group 1) would quickly exhaust any SO42- present in the waters, driving the system
to methanogenesis and subsequent high levels of As. Group 2 has much lower concentrations
and so SO42- is not exhausted. Therefore, As levels in waters of the Pannonian Basin are
controlled not by release but by retention mechanisms, ultimately controlled by levels of TOC
and SO42- in the waters. D and 18O analysis showed that groundwaters containing elevated
As dated mostly from the last ice-age, and are sourced from Late Pliocene to Quaternary

- 25 -
Geochemistry and arsenic behaviour in groundwater

aquifers. The importance of TOC and retention capabilities of SO42--reduction have only
previously been suggested for recent (Holocene) sediments and groundwater, most notably
those in SE Asia as these are the most likely to contain the right combination of factors to
drive the system to the correct redox situation. In contrast, it is shown here that a much older
system containing As bearing Fe-oxides, also has the potential to produce elevated levels of
As if the TOC is suitable for the microbial population to drive the system to the correct redox
situation and SO42- is either absent or wholly consumed.

3.1. Introduction

3.1.1. Occurrence of arsenic in groundwater

The presence of naturally occurring As in shallow reducing aquifers used for irrigation and
drinking water is a major health issue for millions of people worldwide. Research on
mobilisation processes has commonly focused on SE Asia, in aquifers within the Ganges-
Brahmaputra-Meghna, Mekong and Red River deltas (Smith et al., 2000; BGS and DPHE,
2001; Berg et al., 2001; Chakraborti et al., 2003; Charlet and Polya, 2006; Buschmann et al.,
2007; Winkel et al., 2008), and the US (Welch et al., 2000; Warner, 2001; Saunders et al.,

Figure 3.1: Topographical map of Europe (A) showing location of field area and important
topological features (B).

- 26 -
 Chapter 3

2005). Within Europe, many countries have reducing aquifers with levels of As above the EU
drinking water limit of 10 μg/L, requiring some form of remediation before waters reach
consumers. Countries include Greece (Katsoyiannis and Katsoyiannis, 2006), Belgium
(Coetsiers and Walraevens, 2006), Netherlands (Frapporti et al., 1996), Spain (Garcia-
Sanchez et al., 2005; Gomez et al., 2006) and the UK (Smedley and Edmunds, 2002). For the
most part As is geographically restricted affecting only small numbers of people. However,
within Eastern Europe, the Pannonian Basin, spanning Hungary, Romania, Serbia, Slovakia
and Croatia (Fig. 3.1), naturally occurring As in aquifers utilised for drinking waters is known
to affect far larger populations. Estimates suggest that nearly 1 million people are exposed to
naturally occurring As in drinking waters at levels greater than the 10 g/L WHO and EU
standards (Csalagovitis, 1999; Gurzau and Gurzau, 2001; avar et al., 2005; Varsányi and
Kovács, 2006; Vidovic et al., 2006; Lindberg et al., 2006; Habuda-Stani et al., 2007;
Stauder, 2007; Djuric and Jevtic, 2008; Jimenez et al., 2009; Rowland et al., 2009b; Ujevi et
al., 2010), making it the most severely affected region in Europe (Table 3.1). The geology in
the region is complex, with interactions between thermal, glacial palaeo-, fossil marine,
surface and rain waters giving rise to a complex aquifer system, with a wide range of
hydrogeological conditions.

Table 3.1: Values of As in drinking water, and estimates of people affected in countries of the
Pannonian Basin. For locations see Figure 3.1
Country Arsenic Population References
(g L-1) affected
Hungary
- Great Hungarian Plain 0 - 220 400,000 to Csalagovitis, 1999; Varsányi
500,000 and Kovács, 2006
Romania
- Bihor and Arad counties (Western 0 - 176 50,000 Gurzau and Gurzau, 2001
Romanian Plain)
Croatia
- Osijek-Baranja and Vukovar- <1 – 610 200,000 and/or Habuda-Stanic et al., 2007,
Syrmia counties 3% of population avar et al., 2005, Ujevi et
(Eastern Croatia) al., 2010
Serbia
- Northern Banat (Vojvodina 11 - 222 200,000 Vidovic et al., 2006; Stauder,
province) 2007; Djuric and Jevtic, 2008
Slovakia
- Banska Bystrica and Nitra counties 37 - 39 Not reported Lindberg et al., 2006

- 27 -
Geochemistry and arsenic behaviour in groundwater

It is commonly thought that As release is controlled by microbial processes via the dissolution
of As bearing Fe-oxides due to the onset of reductive conditions during sediment burial and
diagenesis (Nickson et al., 1998; McArthur et al., 2001; Harvey et al., 2002; Islam et al.,
2004) with the presence and type of organic matter (Rowland et al., 2009a) and the presence
of SO42- also playing an important role (Kirk et al., 2004; Quicksall et al., 2008; Buschmann
and Berg, 2009). For Hungary, based on groundwater geochemical investigations, similar
process were suggested to be the primary cause of As release (Csalagovitis, 1999).
Mineralogical and aqueous geochemical investigations by Varsányi and Kovács (2006)
showed that As was correlated with both extractable organic matter and Fe-oxides within
sediments of SE Hungary. However, they suggested that two processes were responsible for
As release, (i) dissolution of As bearing Fe-minerals (as previously suggested) occurring in
regions of low pH (~7.6) in areas of groundwater recharge, and (ii) high concentrations of
organic ligands promoting mobilisation in areas of groundwater discharge with longer
residence times (Varsányi and Kovács, 2006). A recent study conducted in Eastern Croatia
showed that the spatial distribution of As in the groundwater is linked with geological,
geomorphological and hydrogeological development of the alluvial basin, and that reductive
dissolution of Fe oxides, desorption of As from Fe oxides and/or clay minerals as well as
competition for the sorption sites with organic matter and PO43- could be the principal
mechanisms that control As mobilization (Ujevi et al. 2010). Arsenic release mechanisms
within other parts of the Pannonian basin have not been discussed in great detail within the
literature, but elevated As levels in conjunction with higher Fe and organic matter
concentrations in waters from the Vojvodina Province in northern Serbia (Vidovic et al.,
2006; Stauder, 2007; Djuric and Jevtic, 2008), imply that processes suggested by Varsányi
and Kovács (2006) and Csalgovitis (1999) could be occurring over a more widespread area.
The region is also known for its geothermal waters (Korim, 1972; Cohut and Bendea, 1999;
Antics, 2000; Antics and Rosca, 2003). Natural waters associated with these systems
commonly contain high As concentrations ranging from 0.1 to nearly 50 mg/L (Henley and
Ellis, 1983; Ballantyne and Moore, 1988). In Europe, high levels of As within groundwaters
due to hydrothermal activity have been found within Italy (Aiuppa et al., 2003, 2006) and
Greece (Tyrovola et al., 2006) and within Hungary, high levels of As have been reported from
geothermal waters (Csalagovits, 1999). Arsenic in such geothermal waters can be sourced
either from the thermal reservoir by fluid-rock interactions of As bearing minerals such as
pyrite, or by scavenging of the aquifer rocks as hot fluids move through the subsurface
(Ballantyne and Moore, 1988). Therefore, in a region known to contain thermal waters,

- 28 -
 Chapter 3

elevated concentrations of As could be due to an ‘external input’ from geothermal waters,

with variations in concentrations due to dilution with non-As bearing waters.
The health impact of As in drinking water on the population in the Pannonian Basin has
shown predominantly negative results. Dermatological studies of populations in SE Hungary
found cases of arsenical hyperkeratosis and hyperpigmentation, as well as elevated levels of
As in hair occurring in people who drank water with levels above 50 μg/L, in comparison to a
control population from the same region (Borzsinyi et al, 1992). Work by Varsányi et al.
(1991) showed mixed results with no increase in mortality due to the consumption of waters
above 50 μg/L when all causes of death were taken into account, but potential increases in
mortality observed when the data-set was separated into males and females, and with certain
diseases examined. In Croatia, positive correlations between As in drinking water and hair
have been shown (avar et al., 2005; Habuda-Stanic et al., 2007). Research as part of the EU
project ASHRAM (Arsenic Health Risk Assessment and Molecular Epidemiology), looking
at the impact of As in waters on the population of Hungary, Romania and Slovakia has shown
associations between As and increased cases of cancers in the skin, bladder and kidney
(Fletcher et al., 2008). The ASHRAM project also found, for the same countries, that there
were correlations between the level of As in drinking water and urine (Lindberg et al., 2006).

3.1.2. General palaeogeography and geological development of the Pannonian

Basin
The sediments of the Pannonian Basin are heterogeneous, thick and complex. The average
thickness of Neogene and Quaternary sediments in the basin is 2 – 3 km, but in the deep
troughs the thickness of sediments can reach 7 - 8 km (Lenkey et al., 2002). To help
understand the groundwater geochemistry, knowledge of the palaeogeographic setting of the
former Lake Pannon that formed the sedimentary deposits now containing the aquifers of
interest is required. The tectonic setting for the basin was formed as the Tethys sea, being
closed by the northward drift of Africa into Europe, was split into 2 parts, the Mediterranean
Tethys in the south, and the Paratethys (‘besides Tethys’) during the middle Cenozoic (Rögl,
1999). The Paratethys consisted of a series of elongated, epicontinental basins stretching in its
entirety from the Western Alps to the Transcaspian Basin (east of the Caspian sea), of which
the Pannonian is situated in the central part (Rögl , 1999; Geary et al., 2002). However, a full
description of the sedimentation history of the entire basin is outside the scope of this study,
and readers are directed to the following for more detail (Rögl, 1999; Magyar et al., 1999;
Juhász et al., 1999, 2004, 2007; Gábris and Nádor 2007) However, the sedimentation history
from Late Miocene, Pliocene to Quaternary can be highly simplified (oldest to youngest) as

- 29 -
Geochemistry and arsenic behaviour in groundwater

W E

field area
Zone of
Quaternary ~1.8 Ma
- fluvial sediments
- Ca/Mg-HCO3 (coarse grained) and Na-HCO3 (fine grained) waters Sediment types
Fluvial
NOT TO Delta/lacustrine
SCALE
Marine
Late Miocene – Pliocene (Lower/Upper Pannonian)
- Brackish (~11 Ma) – lacustrine - deltaic environment
- NaCl – NaHCO3 waters
Middle Miocene ~13-14 Ma (Samartian)
- thermal waters present in Lower Pannonian - marine environment
- Na-Cl (east) to Na-HCO 3 (central) waters
- thermal waters

Figure 3.2: Simplified cross section of the Pannonian Basin (not to scale) with dominant sediment and
groundwater types (Magyar et al., 1999; Varsányi et al., 1997; Juhász et al., 1999; Juhász et al., 2004,
2007; Gábris and Nádor, 2007).

the following (i) shallow marine (Sarmatian), (ii) brackish – fresh water lacustrine (Lower
Pannonian/Pannonian s.s.), (iii) lacustrine/deltaic/fluvial (Upper Pannonian/Pontian), and (iv)
fluvial channel and flood plain (Quaternary). Terrestrial deposits from the Upper Miocene to
the Pliocene sequence (as discussed above) are commonly referred to as the Pannonian sensu
lato (Pannonian s.l.) and subdivided into the Pannonian sensu stricto (s.s.) and Pontian, with
both stages also informally referred to as ‘Lower Pannonian’ and ‘Upper Pannonian’
respectively (Sacchi and Horvath, 2002). A cross section of the Pannonian basin showing
these dominant sediment types is shown in Figure 3.2.

3.1.3. Geology and mineralogy

The facies development of the Pannonian Basin is extremely complex and variable
throughout the region. Therefore, a more detailed overview of the region of study is given. In
the field area, the boundary between the Quaternary and Pliocene (Upper Pannonian) deposits
are closely related and so difficult to define (Viczián, 2002). In the area, variegated clays of
the Vészt and Nagyalföld Formations make up the Quaternary and Pliocene (Upper
Pannonian) sediments, respectively, with both units having similar facies development,
dominated by siltstone and clay with interbedded sandy river deposits. However, the Pliocene
(Upper Pannonian) is more marshy and lacustrine in nature, reflected by the greater
abundance of clay in these deposits (Viczián, 2002), with the uplift of the Pannonian basin, at
approximately 2.4 Ma, initiating more fluvial sedimentation in the Quaternary (Viczián, 2002;
Gábris and Nádor, 2007). In addition, evidence of swamp and wetland deposits are found

- 30 -
 Chapter 3

throughout the Pliocene and Quaternary deposits is some areas (Koros basin) (Viczián, 2002;
Juhász et al., 2004).
The Quaternary sediments, reaching up to 600m in thickness, consist of gravel, fine grained
sand with silt and clay (20 to 50 m bed thickness) sourced from weathering of the Apuseni
Mountains to the east and the redistribution of tertiary sediments from the same region
(Juhász et al., 2004; Viczián, 2002; Nádor et al., 2003). Patterns and type of fluvial sediments
(including grain size) deposited during the Quaternary were controlled by (i) large scale
Milankovitch orbital cycles of 40 and 100 ka cyclicity, with sediment load into the basin
controlled by the alternating glacials (decreased sediment supply, finer grained sediments)
and interglacials (greater transport capacity of rivers leading to higher sediment flux and
coarser grain size) (Varsányi and Kovács, 1997; Juhász et al., 1999; Nádor et al., 2003) and
(ii) subsidence of the basin, which impacted on the sediment accumulation rates in the area
(Varsányi and Kovács, 1997; Juhász et al., 1999; Nádor et al., 2007; Gábris and Nádor, 2007).
Sediments in the Upper and Middle Pleistocene contain thicker and more abundant sand
layers than the Lower Pleistocene (Varsányi and Kovács, 1997), with mineralogical
homogeneity throughout the Quaternary (and Pliocene) in the area implying that the sediment
source remained the same, and it is thought to have been so since the Late Micoene (Viczián,
2002).
Studies of Hungarian sediments (reviewed by Viczián, 2002, typically detected by XRD)
show a variety of Fe minerals, including goethite, limonite, pyrite and siderite. Iron-
oxyhydroxides have also been identified by chemical extraction of sediments from the
Quaternary (Varsányi and Kovács, 2006), with magnetite also identified within sediments of
the same age (Nádor et al., 2003). Generally, upper horizons (Quaternary and Upper
Pliocene), still retain poorly crystallised/amorphous Fe(III) oxides (goethite and limonite)
(Viczián, 2002). In the Lower Pliocene of the Maros fan, these are replaced by Fe carbonates,
though they can be present in small quantities in the Quaternary also (Viczián, 2002). In
addition, pyrite can be found in more organic-rich deposits, indicative of marshy
environments (Viczián, 2002).

3.1.4. Ground water system of the field area

Groundwater type and geochemical composition is strongly controlled by the sedimentary and
palaeogeographical history of the region (Figure 3.2). Within the Lower Pannonian
(Pannonian s.s.) sediments, groundwaters are typically stagnant with Na-Cl to Na-HCO3
geochemistry (Varsányi and Kovács, 1997). Sodium-HCO3 waters are then present within the

- 31 -
Geochemistry and arsenic behaviour in groundwater

Upper Pannonian (Pontian) sediments. Younger Quaternary sediments, dependant on the

dominant size fraction, contain Na-HCO3 dominated waters within finer sediments and
Ca/Mg-HCO3 waters in the coarser layers (Varsányi and Kovács 1997). There are 3 defined
groundwater flow systems in the basin (i) regional flow around the deeper Lower Pannonian
(Pannonian s.s.) and Upper Pannonian (Pontian) boundary, (ii) a local flow within the upper
layers of the basin and (iii) an intermediate flow that connects the regional and local flow
(Varsányi et al., 1999). The
D and
18O isotope signatures of waters from the Upper
Pannonian (Pontian) to Quaternary sediments in the area show a strong palaeometeoric
signature with a depletion in
D and
18O due to the lower temperatures of the last ice-age
(Rozanski, 1985; Deak et al., 1987; Stute and Deak, 1989; Varsányi et al., 1997, 1999;
Varsányi and Kovács, 2009). With infiltration of these waters thought to have occurred due to
tectonic and paleoclimatic events occurring at the end of the Pleistocene (Varsányi et al.,
1997, 1999). Deeper waters, from the Lower Pannonian (Pannonian s.s.) show
D and
18O
signatures enriched in heavy isotopes with respect to the Local Meteoric Water Line
(LMWL), thought to be due to the influence of deeper oil field waters being squeezed from
the underlying fine-grained sediments (Varsányi et al., 1997, 1999; Varsányi and Kovács,
2009). Connectivity between the different bodies of groundwater can also be seen from
geochemical and petrographic studies of fracture and pore filling minerals showing meteoric
water has had a ‘longstanding (late Miocene to recent)’ hydraulic connection throughout the
sedimentary column from the basement to recent sediments (Juhász et al., 2002).
Artesian waters are common in the region, with overpressures of 15% found below the Lower
Pannonian (Pannonian s.s.) sediments (Varsányi and Kovács, 1997). Dewatering, uneven
compaction and rapid subsidence of thick units of shaly Pannonian sediments (deposited at
rates of up to 200-1000 m/Ma) are thought cause these high levels of overpressure (Juhász et
al., 2002). This can also lead to a complex distribution of pressure, with evidence of
alternation of overpressured pools with hydrostatic pools in a single vertical section (Juhász et
al., 2002).
The region is also known for its geothermal waters. The Pannonian geothermal aquifer in
Romania is multilayered and confined at the base of the Upper Pannonian in an area of ~2500
km2, following the Western border of Romania from Satu Mare (north) to Timisoara (south)
(Cohut and Bendea, 1999; Antics and Rosca, 2003). The main geothermal systems in
Romania are found in porous permeable formations such as sandstones and siltstones
(Western Plain) or in fractured carbonate formations (near Oradea) (Cohut and Bendea,
1999). Geothermal waters are found at a depth of 0.8 to ~ 2 km depth, with surface

- 32 -
 Chapter 3

temperatures of 50 – 90 oC, are Na-HCO3/Cl type with dissolved gases (especially CH4) and
total dissolved solids of 4 - 6 g/L, with little variation in composition with time, implying
hydrologic unity over the entire region (Antics, 2000; Antics and Rosca, 2003). Geothermal
waters are also found in shallow groundwater in the Vojvodina Province in Northern Serbia,
in aquifers from Quaternary, Upper and Lower Pannonian sediments (Mrazovac and Basic,
2009). Across the border in South Eastern Hungary, geothermal waters are also sourced from
Upper Pannonian sediments (sandstones, clays and clayey marls), found at a depth of 0.8 to
2.4 km and contain CH4 (Korim, 1972). Waters in this area are less dominated by Cl-, being
predominantly Na-HCO3 type (Korim, 1972). The presence of geothermal waters in the
region, are thought to be due to high heat flow occurring after Middle Miocene extension
which caused thinning of the lithosphere, as shown by the thin crust, thin lithosphere and
normal faults in the basement of Neogene sediments (Lenkey et al., 2002).

3.1.5. Aims of this study

EU countries within the Pannonian Basin (Hungary, Croatia and Romania) have to comply
with the Drinking Water Directive (98/83/EC), in which levels of As in waters intended for
human consumption must be below 10 μg/L. Therefore, groundwater quality assessments and
an understanding of the modes of As release and subsequent predictions of its distribution
within the natural aquifer systems are urgently needed. Such knowledge is important as it can
aid local government and water management to locate new water sources with As levels
below the EU limits. A complete and detailed knowledge of geochemical processes within
groundwater requires extensive data sets, based on water and sediment composition, as well
as analysis of microbial communities present. Such data is not always readily available. For
example the collection of sediment for geochemical and microbial analysis from the levels of
interest can be both very expensive and time consuming. In contrast, ground water samples
from existing wells can be collected within reasonable time and at much less cost. Since the
mobilization (and retention) of As is largely controlled by the redox conditions of the
groundwater, but only to a lesser degree related to the chemical composition of the sediments,
groundwater can provide much of the required information about As in different types of
groundwater and the underlying geochemical processes.
Therefore, the aim of this work was to elucidate the occurrence of As and the dominant
geochemical processes based on the chemistry and isotopic characteristics of groundwater
samples collected from a variety of existing wells in the area spanning the Western Romanian
Plain, and Eastern Hungary (Figure. 3.1), a region known to contain elevated As in its

- 33 -
Geochemistry and arsenic behaviour in groundwater

groundwaters (Csalagovitis, 1999; Varsányi and Kovács, 2006; Gurzau and Gurzau, 2001).
Based on this evidence and work by others in the region and elsewhere, possible modes of As
release will then be examined.

3.2. Methodology

3.2.1. Sample location and collection

The study area is located between 20.1–22.6° N and 45.8–47.5° E (Figure 3.3). A total set of
73 groundwater samples was collected in December 2007 and May/June 2008 from a region
covering ~7,500 km2 of the Western Romanian Plain and ~8,000 km2 of Eastern Hungary,
bounded to the east by the Tisza River and the foothills of the Apuseni mountains to the west.
This area is known to contain elevated levels of As in its groundwaters (Csalagovitis et al.,
1999; Gurzau and Gurzau, 2001; Varsányi and Kovács, 2006; Jimenez et al., 2009; Rowland
et al., 2009b). Although preferable to have an even distribution of wells throughout this
region, sample distribution relied on the availability and access to public wells and so is
spatially uneven (Figure 3.3).

Figure 3.3: Location map of studied wells in the Pannonian Basin. The altitude of the study area is
100±25 m above sea level.

- 34 -
 Chapter 3

To sample the broadest range of aquifers, a variety of well types tapping different depths were
sampled including artesian (depth 50 to 800 m), hand pump (depth 70 to 200 m), open dug
(depth to water table ~ 2 to 4 m), and thermal wells which were typically artesian (depth 400
to 3000 m). Well age information was also collected and ranged from <1 to 200 a for artesian
wells, 1 to 50 a for hand pump wells and 25 to 79 a for thermal wells. However, depth and
age information was not always available, and when obtained, independent verification was
not always possible, so this information can only be used as an indication. To ensure that the
water sampled was as representative to that at depth, hand-pump wells were pumped for at
least 10 min to remove standing water from within the borehole with samples taken only after
redox and pH readings stabilised. Artesian and thermal artesian wells were sampled directly
from the well head as it was presumed that the water emerging was a direct comparison to
that at depth. Water from open dug wells was lifted from depth using the water receptacle
used by local users, and deep thermal wells when not artesian were sampled only after they
had been pumped for a considerable time period (4 h or more). All water samples and water
characterisation were undertaken from an open PVC container which was rinsed 3 times with
sample water prior to use, and sporadically emptied and refilled as samples and data was
collected.
On site water data was obtained using a handheld unit (WTW Multi 340i) with daily
calibrated probes for the following parameters; pH and temperature (WTW SenTix 41-3),
conductivity (WTW TetraCon) and Eh (WTW SenTix ORP). Probes were calibrated daily
using standard reagents prior to use. Samples for cation and NH4 analysis were filtered using
a 0.45 μm nylon filters (Cronus, Sigma-Aldrich) into opaque acid washed poly-propylene
bottles and were preserved by the addition of HNO3 (1 mL, 1 M HNO3 suprapure, Merck, into
60 mL of sample). Samples for anion, TOC and
7Li analysis were also filtered with 0.45 μm
nylon filters (Cronus, Sigma-Aldrich) into transparent acid washed polypropylene bottles.
Waters for
D and
18O analysis were collected without filtering and stored in brown
borosilicate bottles with no head space. Samples for CH4,
13CCH4, ethane and propane
measurements were preserved with NaOH (6 mL of sample, 3 mL of 20% NaOH) in glass
vials, sealed with rubber stoppers, crimped and stored in the dark. All samples were kept cool
on the day of sampling by storage in a freezer box (typically <10 oC), and then stored at 4 oC
until analysis, except during transit from Romania to Switzerland (typically 1 day) to
minimise microbial activity and sample quality degradation.

3.2.2. Aqueous phase analytical methods

- 35 -
Geochemistry and arsenic behaviour in groundwater

Water samples for analysis of total cation and anion analysis were measured within 4 weeks
of collection, with the majority of samples analysed well within this time frame. Each
parameter was analysed in triplicate. Aqueous major and trace elements concentrations were
measured after dilution of acidified groundwater samples with 1% HNO3 supapure (Merck).
Arsenic, Na, Mg, Si, K, Ca, Mn, Fe, B, V, Mo, Li, Sr and U were determined by inductively
coupled plasma-mass spectrometry (ICP-MS; Agilent 7500cx, Agilent US) with total S
(S(tot)) being quantified with inductively coupled plasma optical emission spectroscopy (ICP-
OES; Spectro Ciros CCD, Kleve, Germany). Standards for both ICP-MS and ICP-OES were
prepared from dilution of single element standards (Merck) with a detection limit for the ICP-
MS of 0.5
g/L and ICP-OES of 0.1 mg/L.
Arsenic(III) from samples collected in December 2007 were determined by hydride
generation atomic fluorescence spectrometer (AFS) (PS Analytical Ltd, Kent, UK.,
Millenium Merlin/Millenium Excalibur System), with those collected in May 2008
determined by infield separation using an As speciation cartridge (for details see Roberts et
al., 2007). Total As (As(tot)) in samples from December were determined after the addition of
a reducing solution (2.5 g KI and 0.5 g ascorbic acid in 5 mL of H2O) prior to analysis by
AFS, with samples from May 2008 measured by ICP-MS (as described previously). During
AFS analysis (both As(III) and As(tot) after reduction) a pH 4.8 disodium citrate buffer (0.5
M) was used as only As(III) is converted to AsH3 under these pH conditions (Yamamoto et
al., 1981). The detection limit for the AFS was 0.7
g/L. Analytical accuracy for ICP-MS,
AFS and ICP-OES was ensured by the use of certified reference materials (Nist 1643e,
TM28.3 Lake Ontario, and Merck X CertiPUR Lot no. HC626403 and reference standards
from interlaboratory quality evaluations ARS 21-28, Berg and Stengel (2008)) and by cross
analysis using different techniques.
Hydrogen sulfide (detection limit 85 μg/L), NH4 (detection limit 0.05 mgN/L) and alkalinity
(detection limit 1 mM) were measured using photometric methods as described by Gilboa-
Barber (1971), Berg et al. (2008) and Sarazin et al. (1999), respectively. Analysis for H2S was
undertaken within 8 h of sampling, with analysis for alkalinity undertaken within 3 days of
sampling. A select number of samples for alkalinity measurements were double checked
using filtered chilled samples on return to the laboratory by Gran Titration, with result
variations not greater than 5%. Total organic carbon (TOC) was measured with a TOC 5000
A analyzer (Shimadzu, Switzerland), with a detection limit of 0.5 mgC/L. Chloride (detection
limit 0.5 mg/L), SO42- (detection limit 5 mg/L) and NO3- (detection limit 0.25 mg/L) were
measured by ion-chromatography (Metrohm 761 Compact IC, Switzerland).

- 36 -
 Chapter 3

Methane was determined by gas-chromatography (GC). Concentrations were determined by

injecting 50 to 100 μL of headspace into an Agilent 7890A GC (Agilent Technologies UK),
equipped with a HP-PLOT-Q column and an Agilent 5975C mass spectrometer detector.
Analytical accuracy was ensured by the use of certified standards (Sigma-Aldrich, UK) and
detection limits were 50 μg/L. 13CCH4 in a selected number of samples was also determined
in a method similar to Sansome et al. (1997). Measurements were done using an IsoPrime
mass spectrometer coupled to a TraceGas preconcentrator (GV Instruments, UK). The amount
of injected gas depended on the CH4 concentration in the sample, ranging from a few
L to
several mL with all samples measured twice. Results are noted in the standard -notation
relative to Vienna PeeDee Belemnite (VPDB).
Ethane and propane were measured using an Agilent GC with a Carboxen 1010 Plot column
(Supelco) with flame ionization detection (FID). The temperature was kept constant at 100˚C
for 4 min, raised to 230˚C at 10˚C/min and then held constant for 7 min. The GC had a 500
μL sample loop. Scotty Transportable gas standards were used for calibration (Scott Specialty
Gases, US).
In waters D and 18O were determined by laser spectroscopy with a liquid-water isotope
analyzer (Los Gatos Research DT-100). The resulting D and 18O values were normalized
using internal standards which were calibrated against VSMOW. D and 18O in waters
collected in December 2007 were also calibrated against VSMOW and analysed by the
method outlined by Berg et al., (2008). D and 18O mean monthly rainfall estimates and
subsequent Local Meteoric Water Line (LMWL) were taken from Bowen (2009) for
Bekescaba (latitude 46.6845, longitude 21.0870, altitude 84 m). The LMWL as calculated by
Deak et al. (1987) of D = 818O + 6.4, and values from geothermal waters from Pliocene
aquifers from Hungary D = 5.718O – 16, and D = 5.618O – 30 were also used.
Lithium isotopic compositions were measured using a Neptune Multi-Collector ICP-MS
(Millot et al., 2004). 7Li/6Li ratios were normalized to the L-SVEC standard solution (NIST
SRM 8545, Flesch et al., 1973) following the standard-sample bracketing method. Typical in-
run precision on the determination of 7Li is about 0.1-0.2‰ (2sm).

3.2.3. Cluster analysis

Samples were collected from a range of well types (dug, artesian, drilled, thermal), with
hierarchical cluster analysis used to divide the data set into groups. From the dendogram
produced, groups can be identified which are created based on their ‘similarity’, which in turn
can identify key parameters that can help explain the geochemical variation (Güler et al.,

- 37 -
Geochemistry and arsenic behaviour in groundwater

2002). Hierarchical cluster analysis was done with R (open source statistical program), using
Euclidean distance as a similarity measurement combined with Wards method for linkage,
after the data was scaled and centred to minimise variations due to differing units. The
combination of Ward and Euclidean distance typically produces the most distinctive groups,
and is commonly used for analysis of geochemical data sets (see for example Güler et al.,
2002; Rao and Srinivas, 2006). Hierarchical cluster analysis does not however provide a
statistical test of the different groupings, nor gives reasons for the groupings (Guler et al.,
2002), so the cophenetic correlation coefficient (a value between 0 and 1) was used as a
validity measurement of the data set as a whole (Rao and Srinivas, 2006).
A total of 23 geochemical variables was used during the hierarchical cluster analysis; pH,
temperature, conductivity, total organic C (TOC), Cl-, NO3-, PO4, NH4, Alkalinity, As(tot),
Na, Mg, Si, K, Ca, U, Mn, Fe, B, Mo, Li, Sr and S(tot). These particular parameters were
selected as variances between them should define major influences on groundwater
geochemistry such as mineral dissolution (i.e. Mg, Ca, Cl-, Na, Alkalinity, Si), surface
pollution (i.e. NO3-), redox indicators (i.e. Fe, Mn, U, NH4, NO3-), inputs from thermal
activity (i.e. B, Cl-, Li) and marine influences (i.e. NH4, Na, Cl-, Sr, B). Other key parameters
such as isotopic analysis (18O, D, 13CCH4 and 7Li) and CH4 were not included as these
were not measured in all samples. Although hierarchical cluster analysis is a useful tool to
identify groups of samples with similar characteristics, it cannot inform on why these groups
are similar. Therefore care was taken not to split the data set into too many small groups. As
the main focus of this work was to identify modes of As release, the different groups were
taken at the point within the cluster dendogram that a distinct group of As-rich samples were
observed (Figure 3.4). The cophenetic correlation coefficient measured in this study was
0.703.

3.3. Results and discussion

Cluster analysis revealed 4 major groups. Distinctive geochemical characteristics are shown
in Table 3.2, with all data shown in Table 3.3. Geographical distribution of the wells is shown
in Figure 3.3. Arsenic within the waters sampled ranged from below detection limits (<0.5

g/L) to 240
g/L, and was predominantly As(III) with geochemical parameters indicative of
reducing conditions (Table 3.3). Arsenic (tot) was found at levels above the EU limit of 10

g/L in all groups, apart from the dug wells of group 3 (Figure 3.4). Despite the initial range

- 38 -
 Chapter 3

250

200

Total As (μg/L)
150

100

50

Grp 1 Grp 2 Grp 3 Grp 4

General groundwaters Dug wells Thermal

Methan- Sulphate and saline
ogenic reducing

Figure 3.4: Arsenic box-plot of different groups. Although levels of As above the EU limit of 10
g/L
are found in all groups apart from group 3, group 1 is the only one with consistently high
concentrations.

in geochemical parameters, the groups fall into 3 main categories, (i) group 4, which
contained all waters with unusual characteristics, including all thermal wells sampled (ii)
group 3, which contained most (7 out of 10) dug wells sampled, and (iii) groups 1 and 2,
representing general groundwaters with reducing conditions in all but 3 wells, which were
dug wells and consequently showed oxidising conditions (Tables 3.2 and 3.3). However,
waters in group 1 had consistently higher As levels (mean 123
g/L As(tot), range 23 to 210

g/L). Therefore, identification of the geochemical characteristics that dominate these groups,
especially those that separate groups 1 and 2 is important, to determine why certain waters
contain elevated As. Geochemical reasoning for these groupings is discussed, followed by a
more detailed look at the potential mechanisms of As release within these waters.

3.3.1. Group 4 – thermal and saline influences (<0.5–240
g/L As, mean 33

g/L)

Group 4 contained all thermal wells sampled with several distinct and highly variable
geochemical characteristics (Table 3.2). Thermal waters in this study are classified as waters
with surface temperatures above 35 oC (Korim, 1972). Waters within this group included Na-
Cl, Na-HCO3 and Mg/Ca-HCO3 types (Figure 3.5), with high levels of a variety of
constituents including NH4, Li, Sr, Fe, B, Na, Cl-, alkalinity and conductivity (Table 3.3) and
low 7Li isotope values (Figure 3.6). The group is widely geographically separated
throughout the area (Figure 3.3). Lithium, Si, NH4 and B are all common constituents of
geothermal waters being easily leached from the surrounding rock by circulating warm waters

- 39 -
Geochemistry and arsenic behaviour in groundwater

(Fouillac and Michard, 1981; Henley and Ellis, 1983) and are all found in higher
concentrations within the thermal waters of group 4 as would be expected. Bulk geochemistry
of the majority of group 4 matches that of other geothermal waters sampled in the same

Table 3.2: Distinguishing characteristics and As concentrations of the main groups of groundwaters
Group name n Predominant characteristics As(tot) μg/L
Mean
average and
range
Group 1 15
Consistent elevated As(tot) in all samples
– General
Elevated CH4, TOC 123
groundwaters
Low Fe and S(tot)
– Methanogenic 23.4 to 208
Redox characteristics and Eh indicate reducing
conditions

Na -HCO3 groundwater types

Lighter
7 Li isotopic characteristics

D and
18 O values similar to palaeowaters

Group 2 35
Range of As(tot) values

– General
Moderate S(tot) 11.5
groundwaters
Low Fe
– Sulfate- <0.5 to 58.0
reducing Redox characteristics and Eh indicate predominantly
reducing conditions

Mg/Ca-HCO3 to Na -HCO3 groundwater types

Lighter
7 Li isotopic characteristics
Range of
D and
18O values from rainwater to
palaeowaters

Group 3 7
Redox characteristics and Eh indicate oxidising

– Dug wells conditions 1.0

Elevated SO42-, NO3-, U, Cl-, K and
conductivity <0.5 to 2.1

As(tot) below detection limit

Mg/Ca-HCO3 groundwater type

D and
18 O values similar to rain water

Group 4 16
Redox characteristics and Eh indicate reducing

– Thermal and conditions 33.1
saline waters
Highly variable geochemistry

Elevated levels of NH4, Li, Cl-, Na, Sr, Fe, B, <0.5 to 240
alkalinity

Group contained all thermal wells (+35 o C)

Na-Cl, Na-HCO3 and Mg/Ca-HCO3 groundwater

types
Heavier
7 Li isotopic characteristics
Variable
D and
18O values

- 40 -
 Chapter 3

region (as discussed in section 3.2.3.), with waters (Na-HCO3 type) associated with
geothermal waters from Upper Pannonian sediments (Korim, 1972, Antics, 2000, Antics and
Rosca, 2003) (Fig. 3.2). Well depths are variable, but are typically deeper than those of
groups 1 and 2. However, lack of independent verification of this information makes it
difficult to accurately place waters in specific sedimentary. regions. Not all samples in these
groups are ‘thermal’ (i.e. above 35 oC at the surface), and in addition a small number have
Ca/Mg-HCO3 type, which is in contrast to the dominant geothermal characteristics in the area.
Despite this, even these samples show elevated Li, Si, NH4 and B.
This can be more clearly seen when comparing Li contents with temperature (Figure 3.7)
which shows good positive correlation between high temperature and high Li waters.
However, some samples from group 4 have high Li contents but low temperatures common to
non-thermal waters. It could then be argued that the presence of these geothermal indicators
are due to other processes such as the dissolution of mineral phases and rocks rich in these
constituents (Arnórsson and Andrésdóttir, 1995), or from the breakdown of organic matter
causing elevated NH4 (Appelo and Postma, 2005). However, further evidence of geothermal
influence can be investigated from the Li isotopic compositions (
7Li) measured in a select
number of samples. Recent studies have shown that the extent of fractionation of Li isotopes

Group 1 – Methanogenic Group 3 – Dug wells

Group 2 – Sulphate-reducing Group 4 – Thermal/saline

2- + - 2+ 2+
SO4 Cl Ca + Mg SO4
2- +
Cl
-
Ca
2+
+ Mg
2+

Saline
waters

2+ - 2+ -
Mg SO4 Mg SO4
Freshest
waters

2+ 2+ + - - 2+ 2+ + - -
Ca Na + K HCO3 Cl Ca Na +K HCO3 Cl
Increased interaction with
aquifer minerals

Figure 3.5: Piper plot of groundwaters (Winston, 2000). Samples display a wide variation in dominant group
types including saline waters (Na-Cl), typical groundwater gradients representing cation exchange from freshest
(Mg/Ca-HCO3) and to older (Na-HCO3) groundwaters.

- 41 -
Geochemistry and arsenic behaviour in groundwater

Increased level of
16 water/rock
interaction

12
Group 1 – Methanogenic

7Li (‰)

Group 2 – Sulphate-reducing
8
Group 4 – Thermal/saline

0
1 10 100 100 1000
Li (μg/L)
Figure 3.6: Comparison between Li concentrations and
7Li values (‰). Group 4 display the lowest
7L
signatures, with groups 1 and 2 the highest, indicating that waters with the highest Li contents also show the
highest levels of water/rock interaction.

in waters is controlled by temperature and the intensity of water/rock interaction. Hence, Li

isotope fractionation during water-rock interaction is a function of temperature with more
extensive fractionation at lower temperatures (Millot and Négrel, 2007; Millot et al., 2007,
2010). Therefore geothermal waters, or waters that have mixed with geothermal waters will
typically have lower
7Li values than those which have had no interaction with such systems.
Analysis of samples within group 4 shows lower
7Li values in comparison with the other
groups, which is characteristic of greater water/rock interaction as would be seen in
geothermal waters (Figure 3.6). High levels of CH4 in most samples from these groups are
also expected as this is the dominant gas present in geothermal waters from the area (Korim,
1972; Antics, 2000; Antics and Rosca, 2003; Veto et al., 2004). Within the subsurface, CH4
can be produced from a variety of sources including biogenic (generated from microbial
activity), thermocatalytic (generated from the cracking of hydrocarbons), abiogenic
(generated by minerals acting as reductants) or mantle (produced from mantle gases) sources
(Clark and Fritz, 1997), with both thermogenic and biogenic CH4 known to occur in the
region (Veto et al., 2004).
C13CH4 for a selected number of samples analysed are all less than
-40 ‰, with only trace levels of propane and ethane (Table 3.3) indicative of biogenic CH4
(Clark and Fritz, 1997). A small number of samples have Na-Cl type waters which are a mix
of thermal and cooler temperatures. Geothermal waters of Na-Cl type are found in Romania
(Antics, 2000; Antics and Rosca, 2004). These waters show brackish/marine influences from
these sediments as seen from the piper plot (Figure 3.5), and have Na/Cl ratios of 0.55,
similar to seawater (data not shown) (Nordstrom et al., 1989). Other indicators in this small
group include high B, Li and NH4 concentrations that can occur in waters that circulate in
marine sediments (Duchi et al., 1992), which occur within this region (Figure 3.2).

- 42 -
 Table 3.3: Geochemical parameters for Pannonian Basin groundwaters

- 43 -
Chapter 3
 Table 3.3 - Continued

- 44 -
Geochemistry and arsenic behaviour in groundwater

a
Denotes group number (see text for details)
b
Denotes date of field work and location of sample; D = December 2007, M = May 2008, H = Hungary, R = Romania
c
Denotes well type; A = artesian, P = pump, Th = thermal, Dg = dug
d
Analysis by ICP-OES
e
Analysis by ICP-MS or AFS, dependant on date of collection, (see Methodology for details)
f
Analysis by ICP-MS
- Denotes analysis not undertaken
 Chapter 3

High Li,
10000 low temperature Thermal waters (+ 35 C)

1000
Group 1 – Methanogenic
Li (μg/L)

Group 2 – Sulphate-reducing
100
Group 3 – Dug wells
Group 4 – Thermal/saline
10

1
0 10 20 30 40 50 60 70
o
Temperature ( C)

Figure 3.7: Comparison between Li concentrations and temperature. Thermal waters of group 4 (+ 35 oC)
typically have higher Li contents, with concentrations increasing with temperature. Presence of high Li but low
temperature (<35 oC) waters (circled area) indicates water with strong geothermal influences despite cool surface
temperatures.

D and
18O values for these groups are variable (Figure 3.8). Previous studies have shown
that Na-Cl waters in SE Hungary are sourced from trapped marine/brackish water (Varsányi
and Kovács, 2009), and thermal waters sourced from the Pliocene (Deak et al., 1987) with
D
and
18O isotope values controlled by evaporative processes which cause them to plot to the
right of the LMWL (as indicated in Figure 3.8). Such a trend is not observed within the
geothermal, geothermally influenced and saline waters of group 4 within the present sample
close to the LMWL. However, samples from these groups cover a wide geographical area
(Figure 3.3) and in the absence of detailed hydrogeological information especially due to the
lack of accurate depth data explaining the connectivity within the subsurface, the relatively
small data set makes it difficult to make any definitive conclusions.
Group 1 Group 2 Group 3 Group 4
Methanogenic Sulphate-reducing Dug wells Thermal/saline
-30

-40

D ‰ (VSMOW)

Dug
-50 wells
Dug
-60 wells
Infiltration
-70
Palaeowaters Ca/Mg-HCO
3

-80

-90 Na-HCO3 Na-HCO3

-100
-14 -12 -10 -8 -6 -14 -12 -10 -8 -6 -14 -12 -10 -8 -6 -14 -12 -10 -8 -6

18

O ‰ (VSMOW)
LMWL – Bekescaba (Bowen 2009)
Yearly range of rainfall (Bowen 2009)
LMWL – Deak et al., (1987)
Thermal waters – Deak et al., (1987)
Figure 3.8: Comparison of
D and
18O values of waters with LMWL, thermal and rainfall values from the area
(Rozanski et al., 1986, Deak et al., 1987, Bowen 2009). All samples plot on or close to the LMWL for the area,
with dug wells and those with Ca/Mg-HCO3 plotting in the range for infiltration waters. Sodium-HCO3 waters
plot within the range of palaeowaters from the previous ice-age.

- 45 -
Geochemistry and arsenic behaviour in groundwater

3.3.2. Group 3 – dug wells (<0.5–2.1
g/L As, mean 1.0
g/L)

Wells in group 3 are all dug-wells. These wells are open pits, varying from 40 to 100 cm in
diameter which tap the surface aquifer, and all show certain geochemical traits (Table 3.2).
Waters are Ca/Mg-HCO3 type (Figure 3.5), have positive Eh values and are the only group
containing significant NO3- and U (Table 3.3). Ignoring group 4 (thermal/saline), the dug wells
also contain elevated Cl-, S(tot), K and conductivity in comparison to groups 1 and 2 (general
groundwaters). In the area of study, dug-wells are typically found in gardens and courtyards of
individual houses, with the land around them commonly being used for raising livestock and
growing food. Although most dug-wells sampled had some form of protective covering to
prevent physical contamination, they were not sealed and sometimes used as waste receptacles.
High Ca/Mg contents imply waters dominated by calcite dissolution and are indicative of
young fresh groundwaters (Appelo and Postma, 2005). D and 18O isotopes values of these
waters are typically enriched and plot within the range of rainwater values calculated for the
area (Figure 3.8) (Bowen, 2009). The presence of U, which under reducing conditions is found
in the highly insoluble U(IV) form but, under oxygenated conditions is found in the soluble
U(VI) form (Appelo and Postma, 2005), coupled with positive Eh values both indicate
oxygenated conditions. These characteristics all would be expected from wells taping the
surface aquifers with strong rain water influence.
Despite this, Cl- contents are very high in comparison to general groundwaters (groups 1 and
2) (Table 3.3). This can indicate contamination from a variety of sources including agricultural
chemicals, animal and human waste, sea water, basin brines, road de-icers and municipal
landfill leachate (Panno et al., 2006). The dug wells are sampled over a wide geographical area
(Figure 3.3) so contamination from a single point source such as landfill leachate is unlikely.
Basin brines and sea water could also be the cause, as seen in the wells of groups 4
(thermal/saline), but the dug wells do not have the other characteristics such as elevated Li,
NH4 and B which would be indicative of such contamination. Although the use of salt as a
road de-icer cannot be discounted as a possible contaminant, especially as the dug wells were
sampled during the winter (Table 3.3), the proximity of the dug wells to small holdings means
that the most likely source of the elevated Cl- is anthropogenic activities from agriculture and
waste disposal. Studies have shown elevated K, NH4, Cl-, Na, PO43- and SO42- are all indicators
of contamination by animal waste (Krapac, et al., 2002; Panno et al., 2006), K, NO3- and Cl-
can indicate contamination by fertilizers and from agricultural run off (Marie and Vengosh,
2001; Panno et al., 2006), and high SO42- and NO3- can indicate domestic waste (Marie and

- 46 -
 Chapter 3

Vengosh, 2001). Therefore, the geochemistry of group 3 is indicative of fresh groundwater,

contaminated by surface human and animal wastes.

3.3.3. Groups 1 and 2 – general groundwaters (<0.5–208
g/L As)

3.3.3.1. General characteristics

Waters in groups 1 and 2 form the bulk of waters sampled, have much more neutral
characteristics and have no obvious geochemical overprints from processes such as surface
pollution (group 3, dug wells) and thermal activity (group 4) (see Table 3.2). Because of this,
they can be used to represent the background geochemistry of waters in the region, and
therefore are termed ‘general groundwaters’. Waters in group 2 (<0.5–58
g/L As, mean 11.5

g/L) vary from Mg/Ca-HCO3 to Na-HCO3 dominated, with group 1 (23–208
g/L As, mean
123
g/L) being dominated by Na-HCO3 type water (Figure 3.5). This range in geochemistry
is observed in waters sourced from Late Pliocene to Quaternary sediments in the area, with
Mg/Ca-HCO3 dominated waters shown to occur within coarser parts of the aquifer, and Na-
HCO3 dominated waters associated with finer grained sediments (Varsányi et al., 1997).
Quaternary sediments in the area exceed 500 m in depth (Varsányi and Kovács, 1997; Nádor et
al., 2003; Juhász et al., 2004), and although depth data collected could not be independently
verified, most wells fit up to and below this depth (Table 3.2).
Within group 2, the fresher Ca/Mg dominated waters have more enriched D and 18O values
that plot close to both the dug well (group 3) and rainwater values (Fig. 8). This is to be
expected as waters with higher Ca and Mg are indicative of groundwater with shorter residence
times (Appelo and Postma, 2005). Group 2 also contains a small number of dug wells (3)
which show geochemistry that is a mix between those of group 3 (contaminated dug wells) and
group 2. Most clear is the lower NO3-, Cl-, K and Na contents in comparison to those of group
3, despite similar redox and D and 18O values (Figure 3.8). This suggests that all dug wells
are strongly influenced by infiltration of modern rainwater and exposure to the atmosphere, but
those contained within group 2 are less contaminated by anthropogenic activities.
Waters in group 2 with more depleted D and 18O values in comparison to modern rainwaters
have Na as their dominant cation, with group 1 waters (all Na-HCO3 type) also having depleted
values (Figure 3.8). Throughout Europe during the last ice-age, lower temperatures led to a
depletion in D and 18O being observed between Pleistocene and Holocene natural
groundwaters (Rozanski, 1985). Within the Upper Pannonian and Quaternary sediments of the
Pannonian Basin, palaeowaters dating from the last ice-age are well known (Rozanski, 1985;

- 47 -
Geochemistry and arsenic behaviour in groundwater

Stute and Deak, 1989), with infiltration thought to occur between 70 and 12 ka BP at the end
of the Pleistocene (Varsányi et al., 1997, 1999). Over time as water moves through the
subsurface, cationic exchange and the dissolution of minerals such as mica and feldspars
(known to occur within these aquifer sediments, eg. Viczián, 2002; Juhász et al., 2002) causes
groundwater to become more dominated by Na and K (Appelo and Postma, 2005). These Na-
HCO3 waters are also thought to be characteristic of waters held within finer grained sediments
(Varsányi et al., 1997), so it makes sense that these waters have been present for a long time
within the aquifer with time for the cationic composition to be dominated by Na. It would,
therefore, seem that waters within group 2 plot on a mixing line, with infiltration waters
containing
18O and
D signatures typical of modern meteoric waters occurring in dug wells
and waters of a palaeometeoric origin (depleted
18O and
D and with Na-HCO3 geochemistry)
forming the end members, with a small number of Ca/Mg-HCO3 waters having a slightly
depleted
D and
18O values that plot close to the dug wells (Figure 3.8). Although such a
theory would require further evidence from hydrogeological investigations, evidence of the
long term infiltration of meteoric waters (late Miocene to present) and mixing between
palaeometeoric and present day meteoric waters have been observed within Hungary (Varsányi
et al., 1997, 1999; Juhász et al., 2002; Veto et al., 2004).

3.3.3.2. Redox characteristics

Although the geochemistry of groups 1 and 2 are similar, there are some striking differences,
most notably in more redox sensitive parameters (Tables 3.2 and 3.3). Although SO42- was
measured in all samples, values were often below the detection limit (5 mg/L, data not shown).
Comparison of measured SO42- and S(tot) found that the majority of S(tot) was SO42-, with H2S
forming only a minor component (Table 3.3), therefore S(tot) is taken as a proxy for SO42-.
Reliance on groundwater geochemistry to determine redox conditions can be problematic, and
it is better to couple such information with solid phase sediment analysis and identification of
the microbial community present. However, as long as such limitations are understood, and in
the absence of additional information, groundwater geochemistry can provide good evidence
pertaining to the dominant processes occurring at depth.
Group 1 contains consistently higher concentrations of CH4 and low to negligible S(tot)
concentrations in comparison to group 2 which contains lower CH4 and higher S(tot)
concentrations (Table 3.3). Trace levels of H2S were found in both group 1 (2 out of 13
samples) and 2 (3 out of 27 samples) waters, but levels were close to the detection limit.

13CCH4 measured in a selection of samples from both groups showed values < -40 ‰ with

- 48 -
 Chapter 3

those in group 1 being more depleted than group 2 (Fig. 8) and only trace levels of ethane and
propane (Table 3.3). All samples show a similar range of Fe concentrations (Table 3.3) which
shows no correlation with either CH4 or S(tot).
The
13C signature of CH4 can be used to identify the source, with samples depleted in
13C (<
-40 ‰) indicative of a microbial source. Although CH4 is known to migrate within the
subsurface (Christensen et al., 2000; Appelo and Postma, 2005), the presence of such high
concentrations in group 1, coupled with the more depleted signatures (in comparison to group
2) indicate that methanogenesis is a dominant process in these waters. An active SO42--
reducing community within sediments can be detected either by a reduction in SO42- content or
the presence of S2- (Christensen et al., 2000). As mentioned, levels of H2S detected in both
groups are low or absent which would indicate that SO42- reduction is not a dominant process.
However measurement of S2- (as H2S) can be problematic due to its volatile nature, and in the
presence of Fe it has a very low solubility, so values are liable to be an underestimation and its
absence is not surprising (Christensen et al., 2000). The presence of SO42- as a single
measurement is not an indicator of an active SO42--reduction, but it has been suggested that the
minimum concentrations of SO42- required for SO42--reduction is 2.8 mg/L (Lovley and Klug,

0

13CCH4 ‰ (VPDB)

Group 1 – Methanogenic
Group 2 – Sulphate-reducing
-20

-40

-60

-80

-100
0 10 20 30 40
25
S(Tot) (mg/L)

20

15

10

0
0 10 20 30 40
CH4 (mg/L)

Figure 3.9: Comparison of CH4 with
13CCH4 and S(tot) of waters from groups 1 and 2 (general
groundwaters). Waters of group 1 show a more depleted
13CCH4 signature than group 2. Levels of
S(tot) are higher in group 2, with CH4 higher in group 1, with a negative covariance observed between
CH4 and S(tot).

- 49 -
Geochemistry and arsenic behaviour in groundwater

1986), which is the case in most samples from group 2 (Table 3.3). Sulfate-reduction and
methanogenesis tend to be mutually exclusive as active SO42--reducers and methanogens often
compete for the same substrates. For example, active SO42--reducing populations have been
shown to maintain acetate concentrations at levels inhibiting the activity of methanogens, with
consumption of acetate by methanogens then outcompeting SO42--reducers in environments
low in SO42- (Lovley and Klug, 1986). This could explain the negative covariance observed
between S(tot) and CH4 (Figure 3.9). Iron(III) reduction can occur in mixed metabolic zones
with SO42--reduction or methanogenesis, when Fe(III) concentrations are low (Lovley and
Goodwin, 1988), as seen here (Fe < 1 mg/L), and explains why Fe levels are similar in both
groups with no relationship observed with CH4 and SO42- (Figure 3.10). Evidence then
suggests that (i) waters in group 1 are dominated by methanogenesis, (ii) those in group 2 more
likely to be undergoing active SO42--reduction and (iii) Fe reduction occurs in both sets of
waters.
Differences in the dominant microbial behaviour between the two groups could be due either
to the availability of SO42-, or from the availability and suitability of the TOC present. Sources
of SO42- within the subsurface can be from the dissolution of gypsum, oxidation of pyrite and
from mixing with seawater (Appelo and Postma, 2005). The presence of fossil seawater is
known to occur within the Pannonian Basin (see for example Deak et al., 1987), but samples
with Na-Cl geochemistry observed in this study (see section 3.3.3.1) have S(tot) levels below
detection limits, implying that SO42- has already been depleted, preventing it from being a
more wide-scale source of SO42- within the system. Absence of SO42- in fossil seawater has

1600 1600

1200 1200
Fe (μg/L)

800 800

400 400

0 0
0 10 20 30 40 0 5 10 15 20 25 30 35

CH4 (mg/L) S(tot) mg/L

Figure 3.10: Comparison of Fe with CH4 and S(tot) of waters from groups 1 and 2 (general
groundwaters). No correlation is observed between Fe, S(tot) or CH4 in waters from either group.

- 50 -
 Chapter 3

mineralogy, such as pyrite, known to occur in the region (Viczián, 2002), could be a reason
also been observed in other parts of the Pannonian Basin (Deak et al., 1987). Variation in
for the difference in the availability of SO42-, but previous work has suggested mineralogical
homogeneity in the Quaternary and Pliocene sediments of the region (Viczián, 2002), though
fine-scale mineralogical variations cannot be discounted.
If variations in the availability and flux of SO42- within the subsurface cannot satisfactorily
explain the differences between the two groups, then the availability of organic C to drive the
process could be to blame (Appelo and Postma, 2005). Comparison of the TOC between both
groups shows much greater concentrations in group 1 than group 2, with a positive correlation
between TOC and CH4 and a negative covariance between TOC and S(tot) (Figure 3.11).
Higher levels of TOC within group 1 (methanogenic) would quickly exhaust any SO42- that
had been present, driving the system to methanogenesis. Within group 2 (SO42--reducing)
waters, the lower concentrations of TOC does not allow the flux of SO42- within the system to
be exhausted as quickly, and so SO42--reduction can dominate. Therefore, the difference
between the behaviour of these two groups of waters is likely to be due to the very high TOC
within group 1 driving the system to methanogenesis after SO42- supplies have been exhausted,
with the lower TOC in group 2 allowing active SO42--reduction to be maintained.

40 40

30 30
TOC (mgC/L)

20 20

10 10

0 0
0 10 20 30 40 0 5 10 15 20 25 30 35

CH4 (mg/L) S(Tot) mg/L

Figure 3.11: Comparison of TOC with CH4 and S(tot) of waters from groups 1 and 2 (general
groundwaters). Positive correlation between TOC and CH4 is observed within waters of group 1. Group
1 waters contain much higher concentrations of TOC with little S(tot), with waters in group 2
containing low concentrations of TOC but with higher concentrations of S(tot).

- 51 -
Geochemistry and arsenic behaviour in groundwater

3.3.4. Arsenic, source and mobilisation mechanisms

3.3.4.1. Geothermal and anthropogenic sources?

Correlation between Cl- and As has been used to show the relationship between the mixing of
As-rich hydrothermal fluids and As-depleted meteoric waters (Aiuppa et al., 2006) but this is
not observed in waters sampled here (data not shown). However, this presumes that both Cl-
and As are behaving conservatively within the system. Arsenic does not behave conservatively
in groundwater systems, and As/Cl relationships are not always seen due to the scavenging of
As into mineralogical phases, redox processes (Ballantyne and Moore, 1988). Indeed
comparison of the total abundance of As(tot) shows group 4 (thermal/saline) to have
predominantly low As(tot) concentrations (Figure 3.4), though some individual samples do
contain appreciable concentrations (Table 3.3). As discussed in section 3.3.3.1 (Figure 3.6) Li
isotopes (
7Li) have been used as a tracer for geothermal input with waters containing the
lowest
7Li values (group 4, thermal/marine) being those strongly influenced by geothermal
waters. Comparison of the
7Li with As(tot) values show a positive correlation, with heaviest

7Li values being found in waters with the highest As(tot) content (Figure 3.12), implying that
waters with higher As(tot) contents are those with little geothermal influence. The presence of
individual samples in group 4 with high As is, therefore, contradictory, but these samples have
heavier
7Li signatures more typical of general groundwaters (groups 1 and 2). The presence of
notable As in group 4 implies that these samples are more likely a combination of geothermal
waters and general groundwaters, with As present as an overprint from the latter. Dug wells
(group 3) have negligible As(tot) values (Figure 3.4), as would be expected with As strongly

16

12

7Li (‰)

8 Group 1 – Methanogenic
Increased contribution Group 2 – Sulphate-reducing
of geothermal input
4 Group 4 –Thermal/saline

0
0 1 10 100 1000
As(tot) (μg/L)

Figure 3.12: Comparison between As(tot) and Li isotopes. Positive correlation implies waters with
lowest As contents are those that have undergone the highest levels of water/rock interaction as
associated with geothermal waters.

- 52 -
 Chapter 3

sorbing to minerals under neutral, oxidising conditions, but also suggest that As is not entering
the system via surface run-off. Therefore, elevated As within the Pannonian Basin is not due to
an external input (geothermal or anthropogenic), but from an in-aquifer process.

3.3.4.2. Redox processes

The reducing nature of the waters in groups 1 and 2 (general groundwaters) and the
predominance of As in the reduced form (Table 3.3), coupled with the knowledge that in the
area Fe-oxides are (i) known to occur (Viczián, 2002), and (ii) found associated with As in
aquifer sediments (Varsányi and Kovács, 2006), suggests that reductive dissolution of As-
bearing oxides also plays an important role in the present samples. The lack of correlation
between As(tot) and Fe (Figure 3.13), has been seen in other similar situations as Fe can be
highly reactive in the subsurface, removed from solution due to the formation of Fe-phases
such as siderite, previously suggested within the Pannonian Basin (Varsányi and Kovács,
2006), or from the sorption of Fe onto Fe-oxides not already reduced (Appelo et al., 2002), as
observed in groundwaters of Bangladesh (Horneman et al., 2004). However, this does not
explain why waters in group 1 have much higher As(tot) levels than in group 2, especially as
both sets have very similar ranges in aqueous Fe concentrations (Table 3.3).

3.3.4.2.1. Sulfate-reducing waters (Group 2)

Despite similarities with regard to general characteristics (groundwater type, Figure 3.5;
D
and
18O, Figure 3.8), there are major differences dependant on the key microbial processes
occurring as indicated by certain redox indicators (see section 3.3.3.3). Group 1 is dominated
by methanogenesis (higher CH4, lower S(tot), lighter
13CCH4 in comparison to group 2), group
2 seem to be undergoing SO42---reduction which would limit methanogenesis (indicated by
lower CH4 and higher S(tot) in comparison to group 1), with Fe-reduction occurring in both
(Table 3.3). The reduction of SO42- by microbial communities leads to the production of S2-,
which can in turn react with other components such as Fe and trace metals to form sulfide
precipitates, removing them from the aqueous phase (Aggett and O’Brian, 1985; Moore et al.,
1988; Huerta-Diaz et al., 1998). This process has been shown to ameliorate As levels within
waters in both laboratory (Jong and Parry, 2003; O’Day et al., 2004) and field scale studies
(Aggett and O’Brien, 1985; Moore et al, 1988; Huerta-Diaz et al., 1998; Kirk et al., 2004;
Quicksall et al., 2008; Buschmann and Berg, 2009). Arsenic is removed from solution under
these conditions either by (i) precipitation to form As-sulfides such as realgar (AsS) and

- 53 -
Geochemistry and arsenic behaviour in groundwater

240 240 Group 1 – Methanogenic

Group 2 – Sulphate-reducing
200 200

160 160

120 120

80 80

40 40
As(tot) μg/L

0 0
0 10 20 30 40 0 5 10 15 20 25 30 35

CH4 (mg/L) S(tot) (mg/L)

240 240

200 200

160 160

120 120

80 80

40 40
0 0
0 10 20 30 40 0 0.4 0.8 1.2 1.6
TOC (mgC/L) Fe (mg/L)
Figure 13: Comparison of As(tot) with CH4, S(tot), TOC and Fe of waters from groups 1 and 2
(general groundwaters). Arsenic, CH4 and TOC concentrations are much higher in group 1, with group
2 waters having much higher abundances of S(tot). Range in Fe concentrations is similar in both
groups.

orpiment (As2S3) or (ii) adsorption by and/or coprecipitation with other sulfide phases, most
notably Fe-sulfide (Jong and Parry, 2003; O’Day et al., 2004; Kirk et al., 2004). Removal of
As released into solution by dissolution of Fe-oxides in the presence of SO42--reduction and
subsequent S2- production could occur in the waters of group 2, and explains the negative
correlation observed between As and S(tot) (Figure 3.13).
Work by O’Day et al., (2004) has shown that Fe-rich sediments favour the adsorption of As
onto surfaces of Fe-sulfides as a retention mechanism (as opposed to precipitation of As-
sulfides), as any free S2- present is likely to form Fe-sulfide phases such as pyrite. Although
sedimentological data was not collected as part of this study, the relatively low concentrations
of Fe found within these waters (<1 mg/L) implies an Fe-limited environment, suggesting that
the formation of As-sulfide, typically orpiment, is the dominant retention mechanism in these
waters (O’Day et al., 2004). Even so, the precipitation of any mineral is dependant on the
aqueous concentrations of each component exceeding its solubility product. Batch experiments

- 54 -
 Chapter 3

have shown the abiotic precipitation of orpiment requires concentrations of As and S2- in
excess of 750 μg/L and 341 ug/L, respectively (Newman et al., 1997) which are not observed
in the waters of group 1 and 2. However, these results are based on bulk groundwater
characterisations and it is well known that microbes can create supersaturated
microenvironments favourable for mineralization (Beveridge, 1988). Microbes have also been
shown to directly precipitate As2S3, by the reduction of SO42- coupled to the reduction of
As(V) (Newman et al., 1997).

3.3.4.2.2. Methanogenic waters (Group 1)

Under the dominantly methanogenic conditions prevailing within group 1, SO42--reduction is
absent or occurring only at minimal levels. Arsenic released from the reduction and dissolution
of Fe-oxides is then not subjected to removal by precipitation with S2-, allowing levels to build
up over time to the high concentrations as observed within the groundwater. Indeed a positive
correlation between As and CH4 is observed (Figure 3.13). Therefore, the presence of As in
the groundwaters of the Pannonian Basin is not controlled by mobilisation processes, but by
subsequent retention mechanisms determined by the presence or absence of SO42- (Kirk et al.,
2004; Quicksall et al., 2008; Buschmann and Berg, 2009).

3.3.4.2.3. The role of organic carbon

There is no clear correlation observed between As and TOC (Figure 3.13), but the presence of
such high levels of TOC in group 1 waters (up to 33.6 mgC/L, mean average 7.73 mgC/L)
could also enhance As levels in the water by competitive sorption for surface sites between
organic matter and As (Grafe et al., 2001). This process was suggested by Varsányi and
Kovács (2006) as an additional As release mechanism observed in the discharge area of their
study. Within Hungary, TOC of waters are dominated by humic acids of a terrestrial source
(Varsányi et al., 1997, 2002). Studies by Buschmann et al. (2006) have shown that terrestrially
sourced humics have a higher affinity for binding with As than those of an aquatic source, with
potential for up to 10% of As(V) to be attracted. However, the dominant form of As in the
waters studied here is As(III) which binds to humic acids only at very high humic acid
concentrations (Buschmann et al., 2006). In waters of group 1 with much higher TOC
concentrations, binding of As(III) could be possible, and would have implications for the
retention in the aqueous phase, again enhancing aqueous As concentrations.

- 55 -
Geochemistry and arsenic behaviour in groundwater

3.4. Conclusions

The studied groundwaters of the Pannonian Basin showed a wide range of chemical
composition. The major types of groundwater are classified as follows:
(i) Generally reducing groundwaters (group 1 and 2), with group 1 containing the
highest levels of As (mean 123 μg/L, range 23.4 to 208 μg/L) in contrast to group 2
(mean 11.5 μg/L, range <0.5 to 58 μg/L),
(ii) Dug wells with very low As (group 3, mean 1 μg/L, range <0.5 to 2.1 μg/L), and,
(iii) Wells characterized by strong geothermal and saline influences with low levels of
As (group 4, mean 33.1 μg/L, range <0.5 to 240 μg/L).
High levels of Li, Si, NH4 and B occur in geothermal and saline waters, which have low
7Li
values. General water characteristics (Na-HCO3) and the presence of CH4 of biogenic origin
suggest that these waters are likely to be sourced from Upper Pannonian sediments. A small
number of samples from this group were Na-Cl type, with elevated B, NH4 and Li, indicative
of marine and/or brackish influences. These waters are likely to be sourced from Lower
Pannonian sediments in the region.
The nature of reducing waters, coupled with knowledge that Fe-oxides are known to occur in
these aquifers and are found in association with As, suggest that reductive dissolution of As
bearing Fe-oxides is responsible for releasing As into groundwaters within the Pannonian
Basin. More important however is the presence or absence of SO42-. In waters of group 2,
ongoing SO42--reduction is supposed to release S2- which can then remove As from solution,
either by the formation of As-sulfides, or from sorption of As onto Fe-sulfide phases formed
from the concurrent release of Fe. This leads to the low As(tot) concentrations found within
this group (mean 11.5 μg/L).
By contrast, in waters that are dominated by methanogenesis (group 1), As released by the
reduction of Fe-oxides would not be exposed to any retention mechanisms, and so can build up
to the high levels as observed in the groundwater (mean 123 μg/L). High TOC levels could
also further exacerbate As levels by competitive sorption for surface sites and also by binding
onto TOC. Therefore, the magnitude of As levels within waters of the Pannonian basin as
studied here are controlled largely by retention processes. Levels of SO42- within groundwaters
appear to be in part governed by the amount of organic C present, leading to low As
concentrations in the presence of SO42-, but high As levels for waters in which SO42- levels
have been exhausted.

- 56 -
 Chapter 3

D and
18O analysis showed that the reducing groundwaters containing elevated As dated
mostly from the last ice-age (palaeowaters), and well depths when obtained suggest all are
sourced from Late Pliocene to Quaternary aquifers. Holocene deltaic sediments, i.e. deposited
in the last 10–15 ka BP, have been shown in recent studies to be key in predicting the presence
of elevated As in groundwaters of South and SE Asia due to reductive dissolution of As
bearing Fe-oxides (Charlet and Polya, 2006; Winkel et al., 2008). Waters within these aquifers
typically have low flushing rates (half lives of tens of thousands of years or more, Charlet and
Polya, 2006), but in essence are relatively young, and influenced strongly by surface inputs
(Harvey et al., 2002). The importance of the young sediment and groundwater age in aquifers
of SE Asia is assigned to the fact that these sediments contain abundant organic matter at
levels capable of driving the system to reducing conditions required for mobilisation. As a
broad simplification, As-rich aquifers then tend to be limited to younger Holocene sediments at
shallower depths typically up to 150 m depth, with deeper wells taping older Pleistocene
sediments then contain much lower As levels (BGS and DPHE, 2001).
However, it is shown here that much older aquifer systems with sediments from the Late
Pliocene/Quaternary and containing palaeowaters dating from the last ice-age (70 and 12 ka
BP) are also capable of mobilising As under similar conditions. Arsenic-rich waters are found
up to ~ 600 m at depth, which is much greater than those observed in SE Asia. Although the
age and facies development of these two regions is different (taking the West Bengal delta as a
proxy for other SE Asian aquifers) both could be said as a broad generalisation to be similar
consisting of a complex terrestrial system of alternating sand, mud and silt layers, interbedded
with organic rich deposits (Umitsu, 1993; McArthur et al., 2004; Juhász et al., 2004). Evidence
from the geochemistry of the groundwater studied in this work suggests that in aquifers
containing As bearing Fe-oxides, As mobilisation is controlled only by redox conditions,
regardless of age of the groundwater or aquifer sediments. The implications of these findings
suggest that although the combination of young groundwaters and young aquifer sediments are
more likely to provide the correct combination of factors, i.e. high enough levels of TOC to
drive a system to the correct redox state, it is not a prerequisite. What is of greater importance
is the type and abundance of TOC and its relative suitability as an energy source to the
microbes present within the aquifer in question. This suggests that any reducing aquifer
containing As-bearing Fe-oxides, regardless of age, has the potential to produce elevated levels
of As if TOC levels are high enough and SO42- is totally consumed, or absent to begin with.
Therefore concentrations of SO42- and TOC should be taken into account, especially with

- 57 -
Geochemistry and arsenic behaviour in groundwater

regards to older aquifer systems, in future developments of groundwater prediction models

determining As enrichment due to reductive dissolution.

3.5. Acknowledgements

The authors would like to thank Torsten Diem and Oliver Scheidegger (Eawag
Kastanienbaum) for
18O,
D, CH4 and
13CCH4 analysis and Colorado Plateau Stable Isotope
Laboratory at Northern Arizona University for
18O and
D analysis. Thomas Ruttimann and
Caroline Stengel are gratefully acknowledged for laboratory support, and Julia Leventon,
Cristian Pop, Bogdan Rentea and Zsofi Szi-Ferenc for invaluable translation and field
assistance. This is a contribution of the AquaTRAIN Marie Curie Research Training Network
(Contract No. MRTN-CT-2006-035420) funded under the European Commission Sixth
Framework Programme (2002–2006) Marie Curie Actions, Human Resources & Mobility
Activity Area – Research Training Networks. It reflects the views of the authors but not
necessarily those of the European Community, which is not liable for any use that may be
made of the information contained therein.

- 58 -
Chapter 4

Polyaluminum chloride with high Al30 content as

removal agent for arsenic-contaminated well water

Jasmin Mertens, Barbara Casentini, Armand Masion, Rosemarie Pöthig, Bernhard

Wehrli, and Gerhard Furrer

Water Research (2012), 46: 53-62

- 59 -
PAClAl30 as removal agent for arsenic

Abstract

Polyaluminum chloride (PACl) is a well-established coagulant in water treatment with high

removal efficiency for arsenic. A high content of Al30 nanoclusters in PACl improves the
removal efficiency over broader dosage and pH range. In this study we tested PACl with 75%
Al30 nanoclusters (PAClAl30) for the treatment of arsenic-contaminated well water by
laboratory batch experiments and field application in the geothermal area of Chalkidiki,
Greece, and in the Pannonian Basin, Romania. The treatment efficiency was studied as a
function of dosage and the nanoclusters’ protonation degree. Acid-base titration revealed
increasing deprotonation of PAClAl30 from pH 4.7 to the point of zero charge at pH 6.7. The
most efficient removal of As(III) and As(V) coincided with optimal aggregation of the Al
nanoclusters at pH 7-8, a common pH range for groundwater. The application of PAClAl30 with
an Altot concentration of 1 to 5 mM in laboratory batch experiments successfully lowered
dissolved As(V) concentrations from 20 – 230
g/L to less than 5
g/L. Field tests confirmed
laboratory results, and showed that the WHO threshold value of 10
g/L was only slightly
exceeded (10.8
g/L) at initial concentrations as high as 2300
g/L As(V). However, As(III)
removal was less efficient (< 40%), therefore oxidation will be crucial before coagulation with
PAClAl30. The presence of silica in the well water improved As(III) removal by typically 10%.
This study revealed that the Al30 nanoclusters are most efficient for the removal of As(V) from
water resources at near-neutral pH.

4.1. Introduction

Arsenic-contaminated groundwater is a major health threat to millions of people worldwide.

Specifically in SE Asia arsenic poisoning via drinking water was identified as cause for skin
lesions and cancer (Chen et al., 1999, Berg et al, 2001). In Europe, severe geogenic arsenic
contamination of groundwater occurs in sedimentary basins due to reductive dissolution of iron
(hydr)oxides, e.g. in the Pannonian Basin spanning the countries Hungary, Romania, Slovakia,
and Serbia (Rowland et al., 2011), in the Duero Basin in Spain (Gómez et al., 2006), in
geothermal areas in Ischia and Central Italy (Daniele, 2004; Aiuppa, 2003; Angelone, 2009)
and in Northern Greece (Kouras et al., 2007). In the Pannonian Basin approximately one
million people are exposed to arsenic-contaminated water by artesian village wells that serve
as source for drinking, household and irrigation water (Gurzau & Gurzau, 2001, Varsanyi and
Kovacs, 2006, Lindberg et al., 2006). In the Chalkidiki geothermal area (Northern Greece),

- 60 -
 Chapter 4

arsenic-contaminated groundwater (up to 3000 g/l) is largely used in agriculture as the sole
water resource for irrigation (Casentini et al., 2011, Kouras et al., 2007).
The European Directive 98/83/EC lowered the limit of arsenic in drinking water to 10 g/L
following WHO and EU guidelines (WHO, 1998, EC, 1998), and efforts are needed to
improve the efficiency of arsenic treatment methods. Coagulation and coprecipitation has been
identified as one of the most effective and cheapest treatment technologies (EPA, 2002,
Mondal et al., 2006).
Polyaluminum chloride (PACl) is widely used as coagulant in water and wastewater treatment
due to its high efficiency at low dosage, low cost and convenient application (Duan and
Gregory, 2003; Bratby, 2006). The removal of soluble arsenic from solution by coagulation
and coprecipitation with PACl includes four main steps: (1) coagulation of aluminum clusters
as a result of deprotonation, (2) adsorption of arsenic to the coagulant by outer-sphere or inner-
sphere complexation (3) flocculation, i.e. the formation of flocs from coagulates, and (4)
precipitation of the flocs resulting in the formation of amorphous solids, including As removal
by coprecipitation. The process of precipitation as separation of the aqueous and solid phase is
time dependent and can be enhanced by centrifugation and filtration. Compared with
conventional aluminum coagulants such as AlCl3 and alum, polyaluminum chlorides have the
advantage of coagulating more effectively over a wider pH range with smaller pH change of
the solution (Duan and Gregory, 2003). A comparative study showed that removal efficiencies
for As(V) with PACl are higher than with polyaluminum sulfate, aluminum chloride or
aluminum sulfate, and reached 99% at pH 5.5 at a dosage of 70 M Altot (Fan et al., 2003). The
application of PACl to groundwater in the Pannonian Basin (Serbia) with high organic matter
content reduced initial arsenic concentrations of 50 g/L by 50%, and in combination with
FeCl3 by 84% (Tubi et al., 2010).
PACl generally consists out of Al monomers, Al oligomers and Al13
(AlO4Al12(OH)24(H2O)127+). The latter is composed of four AlO6 trimers that surround a central
AlO4 tetrahedron (Jolivet et al., 2003). Recently, industrial coagulants have been produced
containing also the Al nanocluster Al30 (Al2O8Al28(OH)56(H2O)2618+) immediately after
dissolution in water. The larger Al30 was identified to be composed of two Al13 clusters
connected to each other by four monomeric aluminum, being 1 nm in diameter and 2 nm
length (Allouche et al., 2000; Rowsell and Nazar, 2000).
The titration of dissolved Al(III) salts with bases is well-known to produce Al13 (Equation 1).
13 Al 3+ + 32 OH -  Al13 7+ (4.1)
The formation of Al13 also occurs naturally in streams affected by acidic soils and acidic
effluents of mines (Furrer et al., 2002). Long-term aging (Casey et al., 2001) or heating (127

- 61 -
PAClAl30 as removal agent for arsenic

°C for 5 h, or 95 °C for 48 h) of concentrated Al13 solutions leads to the conversion of Al13 to

Al30 via two pathways, which can contribute in parallel to the overall reaction (Equations 2 and
3):
234 Al13 7+  91 Al30 18+ + 312 Al(OH)3 (s) (4.2)
7+ + 18+
30 Al13 + 24 H  13 Al30 (4.3)
The dissolution of Locron (aluminochlorohydrate with the general formula Al2(OH)5Cl) in
water results in a solution containing a mixture of the nanoclusters Al30 and Al13, as well as
non-characterized oligomers. Aging of such a solution shifts the composition of the mixture
towards a higher contribution of Al30.
The high surface area of Al nanoclusters has the potential to enhance adsorption efficiency.
The specific surface area of Al13 was estimated by Bottero and Bersillon (1988) to be between
540 and 1100 m2/g, depending on the pH and the [OH]/[Al] ratio. The presence of Al
nanoclusters influences the coagulation and flocculation processes in the polyaluminum
solution due to differences in protonation behavior and surface charge. In comparison to Al13
and AlCl3, Al30 developed the strongest floc formation, and showed the highest turbidity
removal over wider dosage and pH range than Al13 (Chen et al., 2006). Deprotonation of Al
nanoclusters allows for adsorption of As by exchange with H+ at H2O and OH- groups, which
are abundant in the structure of Al30 and Al13.
In the present study PACl with high Al30 content (PAClAl30) is applied to remove As(III) and
As(V) from natural well water. The distribution and transformation of Al species in a
polyaluminum solution and precipitated material, and their stability over time is addressed. We
show how acid-base properties and concentration of polynuclear Al affect the pH and the
aggregation of aluminum, and consequently the removal of arsenic. Finally, the removal
efficiencies from artificial well water are compared to field tests in concentration ranges of 70
to 2300
g/L in the Pannonian Basin and in the Chalkidiki geothermal area.

4.2. Materials and methods

4.2.1. Chemicals

High purity (>18 M
) water (Millipore water purification system) was used for the preparation
of all stock solutions and synthetic groundwater. All chemicals were reagent grade from Fluka,
Merck or Sigma-Aldrich and were used as received. Polyaluminum stock solutions were
prepared by dissolving Locron-S® powder Al(OH)2.5Cl0.5 (24.4% - 25.4% m/m Al, Clariant) in
1 L water. Synthetic water with a similar chemical composition than Pannonian Basin
groundwater (Table 1) was prepared according to Roberts et al. (2004) in a 5 L volumetric

- 62 -
 Chapter 4

flask using 1 mM CaCO3, 8 mM NaHCO3 and 0.6 mM MgCl2. The pH was adjusted to 7.6 ±
0.4 by addition of CO2 after Roberts et al. (2004). Silicate was added under rapid mixing from
an alkaline stock solution prepared daily by dissolving 483 mg Na2SiO3 • 9 H2O in 10 mL
water, yielding to 0.34
M Si in 5 L.

4.2.2. Analytical methods

Water samples were stored at 4 °C and were measured for anions and cations within four
weeks of collection. As(tot), Altot, Si, Fe, K, Na, Ca, Mg, Mn, Sr and U were determined on
inductively coupled plasma-mass spectrometry (ICP-MS, Agilent 7500cx). Standards were
prepared by dilution of single element standards (Merck), and the detection limit was 0.5
g/L.
As(III) was determined by hydride generation atomic fluorescence spectrometer (HG AFS, PS
Analytical Ltd, Kent) with a detection limit of 0.7
g/L. Total organic carbon (TOC) was
measured with a TOC 5000A analyzer (Shimadzu, detection limit 0.5 mg C/L). Chloride was
measured with a detection limit of 0.5 mg/L by ion-chromatography (Metrohm 761 Compact
IC) and alkalinity has been determined by Gran titration.

Table 4.1: Composition of synthetic and natural groundwater

Synthetic groundwater Hungarian/Romanian groundwater*
pH 8.0 ± 0.4 8.2 ± 0.2
Alkalinity (mmol/L) 8.0 9.2 ± 5.27
mg/L mmol/L Average (mg/L) Range (mg/L)
Chloride 21.3 0.6 21.0 2 - 159
Sodium 184.0 8.0 198.0 92 - 299
Calcium 40.0 1.0 16.0 4.2 – 33.7
Magnesium 15.0 0.6 4.6 0.7 – 13.8
Silica as Na2SiO3 9.5 0.3 9.5 7.3 – 14.7

g/L
mol/L
g/L
mol/L
As(tot) 123 23 - 210
20 – 260 a) 0.3 – 3.5 a)
As(III) 74.8 20 - 202
* Data from Rowland et al. (2011), group 1 general groundwater
a)
Arsenic was added either as NaAsO2 or as Na2HAsO4

4.2.3. Ferron method

The Ferron method has been established by Smith and Hem (1972) to differentiate between
mononuclear and polynuclear Al species. It is based on the photometrical determination of the
time-dependent complex formation of Al species in acetate-buffered solution with Ferron (8-

- 63 -
PAClAl30 as removal agent for arsenic

Hydroxy-7-iodchinoline-5-sulfonic acid) and has been modified and improved by several

research groups. The here employed method is based on the studies by Schönherr et al. (1983,
1987) and Bertram et al. (1994). The buffer was prepared by dissolving 24.2 g NaCl and 26 g
concentrated acetic acid in 750 mL distilled water, and adding 100 mL 2M NaOH. The
solution was titrated to pH 5 ± 0.05 with NaOH before it was filled up to 1 L. The Ferron
solution was obtained by dissolving 0.5 g Ferron (Sigma-Aldrich) in 250 mL distilled water at
50 °C under rapid stirring. Calibration solutions with 10-80
g Al/mL were prepared by the
addition of 0.5 – 4 mg Al stock solution (1 g/L) to 50 mL water. Sample solutions with a
concentration of 15 mmol/L Altot were obtained by dissolving 43.8 mg Locron-S® in 15 ml
distilled water. After 1, 3, 24, 48, and 120 hours aging, samples were diluted to the
concentration range of the calibration solutions and measured immediately. A volume of 1 mL
of the Al samples was added to 9 mL buffer solution and 3 ml Ferron solution, resulting in a
Ferron/Al ratio >15. The measurements were carried out using a UV-Vis-scanning
spectrophotometer (UV-2401PC) from Shimadzu. The decomposition of Al polymers by the
Ferron reagent was measured in 1 cm cuvettes at = 368 nm within a time span of 40-3600 s.

The kinetic curves were analyzed according to the method of Schönherr et al. (1983, 1987) by
computer-aided calculation. The absorbance data of the time curves were converted into the
logarithm of non-reacted Al. The concentrations of monomeric Al and polymeric Al were
determined from the intersections of the tangents at t=0 of the initial and the final period of the
converted time curves and described in this text as Ala and Alpoly, respectively. The
concentration of Alb was calculated from the difference between Alpoly and monomeric Ala. A
comparison with 27Al NMR data showed that Alb measured by the Ferron method is equitable
to Al13 (Schönherr et al., 1983; Parker and Bertsch, 1994). Alpoly described by Schönherr et al.
(1983, 1987) and Bertram et al. (1994) correspond to the Al30 polycations first described by
Allouche et al. (2000) and Rowsell and Nazar (2000). The fractions of Al species in a solution
determined via Ferron kinetic or 27Al NMR has been assessed to be equivalent (Bertram et al,
1996; Parker u. Bertsch, 1992; Changui et al, 1990).

27
4.2.4. Al MAS NMR

A PAClAl30 solution with 15 mM Altot and 0.1 M KCl was prepared by dissolving 0.73 g
Locron-S® and 1.8 g KCl in 250 mL water at room temperature. Samples were obtained by
removing 50 mL from PAClAl30 solution after 1, 3, 6, and 24 h. Subsequently, sample pH was
increased to 7 with 0.01 M KOH, and the aggregated flocs were separated from solution by
centrifugation at 27.7 103 G and freeze-dried for three days. Precipitated Al flocs were

- 64 -
 Chapter 4

analyzed by solid-state 27Al multi-quantum magic-angle spinning nuclear magnetic resonance

(27Al MQMAS NMR). All 27Al NMR spectra were obtained on a Bruker Avance 400 WB
spectrometer operated at 104.3 MHz with a spinning frequency of 12 kHz. For every sample a
pulse of
/3 and a recycle delay of 1 s was used. The free-induction-decays (FIDs) were
Fourier transformed with a line broadening of 30 Hz. Line fitting was performed with Gaussian
peaks using the Igor Pro software package. The species at around 7 ppm was assigned to
octahedral aluminum (Aloct). The Al concentrations for 61 ppm and at 66 - 67.6 ppm signals
were multiplied by 13 and 15, respectively, to obtain the concentrations for Al13 and Al30. The
Al species represented by the peak at 35 ppm is referred to as Ali in this text.

4.2.5. Titration experiments

Polyaluminum solutions of 100 mL with total aluminum concentrations of 15 mM and an ionic

strength of 0.1 M KCl were titrated with 0.1 M KOH in a 150 mL water jacket vessel at 25 °C
under continuous magnetic stirring and Argon atmosphere. All acid-base titration experiments
were carried out by means of a computer-controlled titration device (736 GP Titrino,
Metrohm). Titrations were started after 30 minutes aging and 10 minutes pH stability control.
Aggregation of Al nanoparticles in the titrated PAClAl30 solution with [Altot] of 15 mM was
observed using a laser beam (632.8 nm). Surface-specific charge (
) and protonation state (Z)
of the PACl solution were calculated according to Casey et al. (2005) from titration data under
the assumption that most Al is present as Al30 with positive charges of +18 at fully protonated
state:

=
(Z +
) max
(4.4)
Al surf

vB K
[ ]
+ H +
w+
with Z =
( v 0 + v) H [ ] (4.5)
[ Al 30 ] tot
Z : Protonation number

max
: total charge at fully protonated state
Al surf : total number of surface Al atoms
v 0 : initial volume of Al30 in solution (ml)
v : volume of the added base ( ml)
B : concentration of the added base ( M )
[ Al30 ] tot : Total Al30 concentration as calculated for each titration step (mol/L)

For arsenic removal experiments, 1 mL of the PACl solution was replaced by 1 mL of neutral
arsenic stock solution (5 mM As) once the desired pH was held constant for 10 minutes,

- 65 -
PAClAl30 as removal agent for arsenic

resulting in an arsenic concentration of 50
M in the vessel. For kinetic titration experiments,

the pH was kept constant for up to four hours and triplicates were taken after 5, 15, 30, 60,
120, 180 and 240 minutes. Samples were centrifuged at 3.22 103 G and filtered through 0.45

m filters (Infochroma AG). Above pH 6, filters with 0.2
m led to instant clogging. In order
to use the same separation technique for all experiments it was decided to use 0.45
m filters.

4.2.6. Coprecipitation experiments

Coprecipitation experiments were conducted in 330 mL polyethylene terephthalate bottles with

97-98.9 mL aliquots of synthetic groundwater. A volume of 1 mL of 27, 133, and 347
mol/L
arsenite or arsenate solution was added to yield arsenic concentrations of 0.27, 1.33 and 3.47

mol/L. The coprecipitation process was initiated by adding PAClAl30 solution from Locron-S®
stock solutions (5.1 mM, 51.2 mM, and 512.2 mM Altot) to yield Altot concentrations of 0.01 –
10.3 mM. The bottles were placed in a reclined position on a shaker with horizontal movement
at 250 rpm for 30 minutes at room temperature, and the pH was allowed to equilibrate. After 4
hours of settling, the pH was measured and the supernatant solution was filtered through 0.45

m nylon filters and the filtrate was analyzed for As concentrations and cations.
To investigate the arsenic density on Al nanocluster, experiments were carried out in 50 mL
sterile polyropylene tubes with 42.5 mL synthetic water at pH 7.5 ± 0.1. To each batch, 5 mL
of As(III) or As(V) stock solutions (adjusted to pH 7.6) were added to achieve initial arsenic
concentrations of 0.3 to 66.7
mol/L. Finally, 2.5 ml of a Locron stock solution (5.1 mM Altot)
was added to yield total Al concentrations of 255
mol/L.

4.2.7. Field tests

Arsenic removal has been tested on natural groundwater from reducing Pannonian aquifers in
three artesian wells from the towns Avram Iancu, Ciumeghiu and Sepreus (R112, R113, and
R164, notation after Rowland et al., 2011) in the counties of Bihor and Arad in SW-Romania
ca. 100 km south of Oradea, and from three pumped wells in the Nea Triglia geothermal area
of Chalkidiki prefecture, Northern Greece, identified as KL59, KL103, and Pilot. In Romania,
artesian wells have been selected for high arsenic concentrations. 330 mL PET bottles were
filled with 100 ml groundwater, and the PAClAl30 solution was added in two different
concentrations to yield an As:Altot molar ratio of 0.15 and 0.015. The bottle was vigorously
shaken for 5 minutes and left for settling in an upright position over night before the
supernatant solution was filtered (0.45 μm). The samples were taken before and after

- 66 -
 Chapter 4

treatment, immediately acidified by adding 2% 1 M HCl, and stored in the dark at 4 °C until
analysis by HG-AFS.

In the geothermal area of Nea Triglia, Greece, water was collected from each well in 4 L
containers, which were rinsed three times with distilled water before use and filled up to 3.2 L.
Locron-S® was added in two different concentrations: 0.22 g/L ([Altot] = 1.1 mM) and 0.62 g/L
([Altot] = 3.2 mM). The waters were vigorously shaken three times and left to react for four
hours. Samples were collected before and after treatment by filtering through 0.45 μm filters.
For ICP-MS analysis 20 mL of sample was acidified with 1% HNO3 and stored at 4 ºC. For
As(III) and As(tot) analysis 15 mL of water was acidified with 250 μl 1 M HNO3 and stored at
4 °C in a brown PET bottle until analysis with HG AFS and ICP-MS.

4.3. Results and discussion

4.3.1. Characterization of PAClAl30

4.3.1.1. Acid-base properties of PAClAl30

The chemical behavior of PACl in solution is crucial to understand the removal processes and
the best conditions of coagulant application. The initial pH of the PAClAl30 solution is very
similar to that of an Al30 solution with an Al30 concentration of 25 M. Increasing the pH from
4.7 – 6.7 leads to a loss of 18 protons (Z = -18, Figure 4.1a). Aggregation of Al nanoparticles
in a PAClAl30 solution with [Altot] of 15 mM was observed with a laser (
= 632.8 nm) at pH

5.5 (Z-Value = -4). The pH of the point of zero charge (pHPZC) of the PAClAl30 solution at pH
6.7 correlates with the pHPZC of a solution with lower Al13 and Al30 concentrations, and the
obtained values are similar at a surface-specific charge smaller than 0.1 (Figure 4.1b). The
deprotonation of the main compound of the used PACl, Al30, is known to occur in a broad pH
window and takes place at terminal water ligands (Casey et al., 2005). The presented results
indicate that the PAClAl30 solution deprotonates over the same large pH window than Al30 (4.7
– 6.7). Strongest flocculation of the aluminum was found above pHPZC 6.7, where the
electrostatic repulsion was smallest. The processes of aluminum coagulation, floc formation
and the consequent efficiency of precipitation and liquid-solid separation are enhanced with
increasing pH.

- 67 -
PAClAl30 as removal agent for arsenic

Figure 4.1: Titration curves of a 1 hr altered PAClAl30 solution with total aluminium concentrations of
5.1 mM (corresponding to 392
M Al13 or 170
M Al30) in comparison with Al13 (data from Furrer et
al., 1992) and Al30 (data from Casey et al., 2005). a) Z Value. The dashed lines indicate the point of
complete deprotonation of Al13 and Al30, b) Surface-specific charge. The titration until pH 8 was carried
out in 63 min. The pHPZC of the PACl Al30 solution is at 6.7 and corresponds well with the pHPZC of the
25
M Al30 solution.

4.3.1.2. Al floc analysis with 27Al NMR

Solid-state 27Al NMR analysis of Al flocs from a PAClAl30 solution with 15 mM [Altot] showed
one broad peak representing octahedral aluminum (Aloct) centered at ca. 7 ppm (Figure 4.A1,
supplemental material). Two contributions were distinguished for an asymmetrical peak
representing Al(O)4: one at 61 ppm (Al13

-Keggin) and another at 66 - 67.6 ppm. The signal

- 68 -
 Chapter 4

Figure 4.2: The share of Al species from [Altot] = 15 mM in % in a freshly prepared aqueous PAClAl30
solution (empty symbols) and in formed Al flocs (filled symbols). Al speciation of the PAClAl30 solution
was determined by the Ferron method. The indicated mole fractions Al30 (squares), Al13 (upward
triangles) and Almono (empty circles) for the PACl solution are derived from the measured
concentrations for Alpoly, Alb, and Ala after Schönherr et al. (1983, 1987). Al flocs from a PAClAl30
solution with 0.1 M KCl were analyzed by solid state 27Al MAS NMR. Aloct (filled circles) corresponds
to the octahedral aluminum species not present in the structure of Al13 or Al30, Ali (downward triangles)
is an intermediate, octahedral aluminum species.

at 66 – 67.6 ppm falls between the reported chemical shifts for Al30 (68 - 72 ppm, Allouche et
al., 2000) and the Al13
-Keggin (65.5 ppm). Comparison with precipitates of pure Al30

solutions showed the same chemical shift for Al30. It therefore corresponded to an Al30 polymer
that was slightly deformed by aggregation. Al13 remained between 15% and 23% without
notable trend (Figure 4.2). The average of the Al30 share over 24 hours was 48 ± 3%, while
octahedral aluminum not linked to Al13 and Al30 species varied from 27% to 40%. A small
peak at 35 ppm representing 0.4% to 0.8% of total Al, referred to in this text as Ali (Figure
4.2), may be attributed to an intermediate octahedral Al species that was formed during the
transformation of Al30. Compared with the Al species distribution in the aqueous solution,
about 20% to 30% of Al30 was transformed into octahedrally coordinated Al clusters during
aggregation. However, Al13 contents in the PAClAl30 solution were similar to Al13 contents in
the precipitated Al flocs. Therefore, aggregation processes did not affect the Al13 content

- 69 -
PAClAl30 as removal agent for arsenic

(Figure 2). In total, two thirds of the total aluminum in the aggregates were represented by the
Al nanoclusters Al13 and Al30.

4.3.2. Arsenic removal with PAClAl30

The removal of arsenic from aqueous solutions with PAClAl30 includes two processes, which
can occur simultaneously or subsequently: (i) the chemical interaction of As species with the
polynuclear Al. The As(V) species H2AsO4- and HAsO42- and the As(III) species H3AsO3 form
dissolved Al-As complexes already in the acidic pH range, where no aggregation and
precipitation of polynuclear Al can be observed. (ii) The deprotonation of polynuclear Al
complexes is dependent on pH: Al nanoclusters fully deprotonate and aggregate strongly at
near-neutral or higher pH values. The actual removal of arsenic from solution results as a
consequence of aggregation and precipitation of the nanoclusters that might include relevant
amounts of bound arsenic. Since the As(V) species have a higher binding affinity for
aluminum atoms than the As(III) species (Manning and Goldberg, 1997) the former are
removed more efficiently than the latter.

4.3.2.1. The effect of pH

The pH value is a key parameter controlling the deprotonation state and the aggregation of
polynuclear Al species in PACl. Strong aggregation of Al clusters at pH > pHPZC leads to a
pronounced removal of total aluminum from solution (Figure 4.3).
As(V) was completely removed at pH 7 and 8, whereas As(III) removal reached its maximum
of 80% at pH 8 (Figure 3). Although optimum removal pH of both species was achieved at pH
7-8, the interaction of As(III) and As(V) with Al nanoclusters is different. As(V) was removed
in the same proportion than Al over the entire pH range within the margins of error (Figure
4.3). Hence, As(V) is uniformly distributed over all Al surface binding sites, including those of
dissolved Al species at pH 5, 6 and 6.5 (Figure 4.3), and As(V) forms soluble complexes with
Al nanoclusters. Hence, a covalent binding by ligand-exchange reactions is taking place
independently of the charge of Al clusters. The simultaneous elimination of As(V) with Al was
consistent with the precipitation of the As(V)-Al complexes.
As(III) removal at all pH values was significantly lower than As(V) removal. This indicates a
weaker affinity between As(III) and Al nanoclusters. As shown in Figure 4.3, strong
flocculation leads to enhanced As(III) removal at pH 7 and 8. This shows that Al aggregation
is the most important process for the removal of both arsenic species with PAClAl30. Hence, the
pH needs to be kept at the optimum for the removal of Al nanoclusters.

- 70 -
 Chapter 4

Kinetic investigation over 4 hours reaction time showed no strong variation of arsenic removal
(Figure 4.3). Therefore, the reaction of As with Al nanoclusters was completed after at most 5
minutes. Speciation measurements showed no oxidation of As(III).
At optimal pH, 99% of Al was removed by precipitation of Al flocs and filtration. At the total
added Al concentration of 405 mg/L this means that 4 ± 0.6 mg/L remained in solution. Hence,
the maximum contaminant level for aluminum in drinking water of 0.2 mg/L suggested as
indicator parameter by the EC Drinking Water Directive 98/83/EC was exceeded by a factor of
20. For practical application at small-scale the Al removal process therefore needs to be
optimized by using smaller filters or centrifugation.

Figure 4.3: Removal of As(III), As(V) and Al in dependence of pH. As(V) was added as Na2HAsO4
and As(III) was added as NaAsO2 in the concentration of 45
M with samples taken in triplicates after
target pH was reached. Error bars indicate the standard deviation of measurements after 5, 15, 30, 60,
120, 180, and 240 hours. Locron-S® solutions with [Altot] = 15 mM were altered for 30 minutes and had
an ionic strength of 0.1 M KCl. Equilibration pH at t = 0 was 4.8. For As(V), the five titrations with
target pH 5, 6, 6.5, 7, and 8 required 0.53, 4.56, 8.67, 8.09, and 9.47 mL 0.1 M KOH and reached the
target pH values after 3.9, 29.7, 55.9, 18, and 15.7 minutes, respectively. The five As(III) titrations with
target pH 5, 6, 6.5, 7, and 8 required 0.39, 3.82, 8.09, 9.47, and 10.8 mL 0.1 M KOH and reached the
target pH values after 2.7, 13.8, 52, 18, and 15 minutes. As(V) and As(III) removal is increasing with
higher pH and increasing aggregation of the PACl. Arsenate is removed by 100% above pHPZC of
PAClAl30 at 6.7 (0.22 As/coagulated Al), maximum arsenite removal of 80% is achieved at pH 8.

- 71 -
PAClAl30 as removal agent for arsenic

4.3.2.2. The effect of coagulant dosage in water applications

Al30 is more acidic than Al13 due to the presence of acidic
H2O functional groups (Rustad,

2005). To understand how the application of PACl with high Al30 content affects the water
chemistry and the arsenic removal efficiency in natural well water, we used simulated
groundwater with the chemical composition of the Pannonian Basin (Table 4.1) at initial pH of
8 ± 0.4.
Independent from the initial arsenic concentration, 98.6% to 99.4% As(V) were removed with
total aluminum concentration of 1-6 mM (Figure 4b). Aluminum concentrations below 1 mM
showed a scattered Al removal and consequently a scattered As removal. Above 6 mmol Al/L
the pH dropped to 6.5 or lower. Titration experiments revealed that low pH reduced
aggregation and precipitation of aluminum clusters due to charge repulsion (Figure 4.3), and
this was confirmed by Al removal data at high initial Al concentrations with low and high pH
for reference (Figure 4.4c): highest Al removal was achieved when the pH is above pHPZC, and
under these conditions As(V) is removed by > 90%.
As(III) was removed by 35% at most (Figure 4.5b), but it improved in general with increasing
Locron-S® addition due to the increase in available surface, as long as the pH is not too low to
prevent the formation of big aggregates. As(III) removal in titration experiments is about twice
as efficient as in batches with synthetic water where pH was at equilibrium (Figure 4.5a).
PAClAl30 solutions with high Al concentration were more acidic but also showed more
coagulation and aggregation when brought to pH 7-8 than solutions with smaller Al
concentrations. Therefore, As(III) can be removed by up to 80% with high Al concentration at
optimum pH conditions (Figure 4.3).
The adsorption of As(V) on Al was investigated by increasing initial arsenic concentrations to
66.7
mol/L in a solution of 250
M Altot in water of the same composition than used in batch
experiments. The maximum adsorption density for As(V) was determined as 0.2 M/ M Al.
This is a 60% increase compared to alum coagulation (0.12 M As(V)/Al) reported by Edwards
(1994). The As(III) adsorption density is lower by a factor of 100 (Figure 4.6).

- 72 -
 Chapter 4

Figure 4.4: Effect of increasing aluminium concentrations on a) equilibrium pH, b) arsenate removal
with and without Si as 0.34 mM Na2SiO3, the insert zooms to low Al concentrations c) silica removal in
synthetic water (composition see Table 4.1). [As(V)]initial = 0.3 μmol/L, 1.3 μmol/L and 3.47 μmol/L.
The pH decreases with increasing aluminium concentration from pH 8 at low Al concentrations to pH
6.8 at high Al concentrations. The dotted line in a) indicates the pHPZC of PAClAl30, above which Al
aggregation and removal is at maximum.

- 73 -
PAClAl30 as removal agent for arsenic

Figure 4.5: Effect of increasing aluminium concentrations on a) equilibrium pH, b) arsenite removal
with and without Si as 0.34 mM Na2SiO3 and c) silica removal in synthetic water (composition see
Table 4.1). [As(III)]initial = 0.3
mol/L, 1.3
mol/L and 3.47
mol/L. The pH decreased with increasing
aluminium concentration from pH 8 at low Al concentrations to pH 6.8 at high Al concentrations. The
two lines in 5b) show the average removal with and without Si. As(III) removal is enhanced by the
presence of Si.

- 74 -
 Chapter 4

Figure 4.6: Removal of As(III) and As(V) from synthetic water (composition see Table 4.1) by
PAClAl30 (250
mol/L Altot ) with increasing addition of arsenic concentrations (0.3 – 66.7
mol/L) at
pH 7.5 ± 0.1.

4.3.2.3. The effect of silica

Activated silica is used as a coagulation aid for aluminum and ferric-based treatment. It forms
negatively charged colloids that increase the size of aluminum hydroxides during coagulation.
Water of the Pannonian Basin already contained 9.5 mg Si/L in average (Table 4.1). We
conducted batch experiments with and without adding silica to synthetic water and found
initial silica concentrations were reduced after water treatment by up to 50% with 3 mM Altot,
while at the lowest and highest PAClAl30 addition no Si was removed (Figure 4.4c and 4.5c).
Hence, silica removal depended on the effectiveness of aluminum coagulation and separation.
Fan et al. (2003) found an increasing As(V) removal by polyaluminum sulfate and aluminum
sulfate with Si dosages of 4-10 mg Si/L. We have no evidence for As(V) removal being
effected by the presence of Si, because As(V) was already removed by 98%-100% without
silica. However, As(III) removal ranged between 10% and 35% with silica and between 0%-
15% without silica (Figure 4.5b), unless the pH was at or above 7. This suggests that the
presence of Si is favorable to the removal of As(III), either by enhancing the coagulation of
aluminum or by forming complexes with Al and As(III).
Aluminum was removed by nearly 100% in batch experiments (Figure 4.4c and Figure 4.5c).
However, the concentration of Al remaining in solution ranged between 0.03 and 0.8 mg/L
- 75 -
PAClAl30 as removal agent for arsenic

with an average of 0.35 mg/L for waters containing no Si concentrations and 0 to 1.1 mg/L at
Si concentrations of 9.5 mg/L with an average value that is slightly below the drinking water
limit of 0.2 mg/L suggested by the WHO for drinking water (Figure 4.A2, supplemental
material). For high and low Al concentrations, the separation of Al from the aqueous phase is
scattered and can be as low as 60% due to concentration and pH effects as discussed above
(Figure 4.4c). The removal of Al can be enhanced by better separation methods than used in
this study, e.g. smaller filter pores and ultracentrifugation. Centrifugation is neither cost-
effective nor practical for large-scale field operations. It was observed during this study that
filters smaller than 0.45
m clogged easily when filtering Al solutions above pH 6. Therefore,
using several filter steps from 0.45
m to smaller filter pores is a better alternative. Water
treatment facilities in the US using Al2(SO4)3 salts reported similar remaining aluminum
concentrations ranging from 0.01 to 2.7 mg/L, with an overall average of 0.16 mg/L after
treatment (WHO, 1998).

4.3.3. Field validation

In Romania, groundwater from the artesian wells R112, R113, and R164 has As concentrations
in the range of 74 to 185
g/L, mainly present as As(III). Samples R112 and R164 had similar
arsenic concentrations (80
g/L and 74.5
g/L), while R113 groundwater contains 185
g/L
(Table A 4.1, supplemental material). The three wells generally differed in their water
composition, but were within the average groundwater chemistry of the area (Group 1,
Rowland et al., 2011). The well R113 with the highest As concentration had the highest
concentration of TOC (6.1 mg/L), and the groundwater with the lowest As concentration (R
164) had the lowest TOC concentration (0.49 mg/L). The enhanced release of arsenic due to
competition with TOC on sorption sites has been proposed for Hungarian groundwater by
Varsányi and Kovács (2006). Figure 4.7a shows that 33% – 38% of arsenic was removed from
As(III)-rich water with an As-to-Al ratio of 0.14, and 12.6% – 24.1% were removed with an
As-to-Al ratio of 0.014. In this study at three well sites, TOC concentrations of up to 6 mg/L
seemed to have no effect on As(III) removal. H3AsO3 did not change its oxidation state during
treatment (Table 4.A1, supplemental material).
Arsenic removal was tested on water of the geothermal area of Chalkidiki from two open tanks
KL59 and KL103 and from a freshly pumped well (Pilot) with initial arsenic concentrations of
193.2
g/L, 2293
g/L, and 168.8
g/L, respectively, mainly present as As(V). The results
showed that initial As(V) concentrations of 170
g/L were reduced to below the drinking water
guide value with 1 mM Altot, and are further decreased to less than 5
g/L with 3.2 mM. High
As(V) concentrations of up to 2300
g/L can be lowered to 10.8
g/L with the addition of 3.2
- 76 -
 Chapter 4

mM Altot (Figure 4.7b), resulting in a molar ratio of 0.01 M As(V)/ M Al. The As removal data
from these field trials in two different areas confirm laboratory findings despite different water
compositions.
For all investigated wells, measured aluminum concentrations were below the detection limit
of 0.1
g/L after the settling of Al flocs and filtration (0.45
m) of the supernatant solution.
These data are in contrast with the results from batch experiments and a result of the reduction
of boundary effects during up scaling. Initial silica content was in the range or one third above
the concentrations used in lab experiments. For As(III)-rich Romanian groundwater Si was
removed by ca. 50% after treatment, while in the case of As(V)-rich Greek groundwater Si
removal was up to 30% (Table 4.A1, supplemental material). Manganese concentration was
reduced by 25% to 50%. Other metals present in the waters like copper and iron were
eliminated substantially by 40% and 93% - 99%, respectively. Trace elements like uranium
and vanadium showed a tendency to be reduced (Table 4.A1, supplemental material). The
initial pH of the Romanian and Greek groundwater was within the optimal range for removal.
During treatment with PAClAl30 the pH decreased by 0.2-1 units (Table 4.A1, supplemental
material). In Greece, where the water is pumped into open tanks before it is used for irrigation,
the coagulation with polyaluminum chlorides in the form of Locron-S® powder represents an
easy-applicable and effective method to reduce As accumulation in soils due to irrigation
practices.

4.4. Conclusion

A reliable chemical treatment method requires knowledge of the molecular processes under
natural conditions. This study provides insights for an optimal application of PACl with high
Al30 content to remove arsenic from contaminated well water. Our findings give evidence that
the acid-base properties of a polyaluminum coagulant strongly influence the coagulation
processes and removal potential. The protonation state of Al30 governs the formation of Al
flocs, and consequently the removal of both arsenic species, to occur between pH 6.5 and 8.
This is a common pH range for natural groundwater and therefore no pH adjustment has to be
undertaken for precipitation to take place. Analysis of the PAClAl30 solution and the precipitate
obtained from a commercial coagulant powder verified that polyaluminum clusters remain
dominant during flocculation and precipitation. Highly concentrated PAClAl30 (> 5 mM)
solutions result in pH ranges that are not favorable for flocculation and precipitation (< pH 6.5)
and therefore PAClAl30 dosage has to be chosen carefully avoiding unnecessary excess of Al
and decrease of pH.

- 77 -
PAClAl30 as removal agent for arsenic

Figure 4.7: As(V) and As(III) remaining in solution before and after applying PACl Al30 to a) three
artesian groundwater wells R112, R113 and R164 in Western Romania. PACAl30 solution was applied to
each well in the two molar Al:As ratios 70 and 700. b) Three well waters in the geothermal area of
Chalkidiki, Northern Greece. Locron-S® was added in two different concentrations: 1.1 mM [Altot] and
3.2 mM [Altot]. The groundwater chemical composition before and after treatment is listed in the
supplemental information.

- 78 -
 Chapter 4

4.5. Supplemental material

4.5.1. Supplemental figures

Figure 4.A1: Solid-state 27Al NMR spectra of the precipitated flocs of a PAClAl30 solution with [Altot] =
15 mM and 0.1 M KCl after 1 h, 3 h, 6 h, and 24 hours aging.

Figure 4.A2: Al concentrations remaining in water with and without [Si] = 9.5 mg/L after treatment
with PACLAl30, at initial concentrations of 2.8 to 280 mg Al/L. The average of waters with Si present is
slightly underneath the suggested Al drinking water limit of 0.2 mg/L (EC, 1998).

- 79 -
 4.5.2. Supplemental table

PAClAl30 as removal agent for arsenic

Table 4.A1: Chemical groundwater composition of three artesian waters from the Pannonian Basin (R112, R113, and R164) and three
pumped and stored waters from the geothermal area of Nea Triglia, Greece (KL59, KL103, Pilot), before and after treatment with two
different PAClAl30 concentrations (“n.d.”: not detectable , “n.a.”: not available, “-“: no data)
- 80 -
Chapter 5

Adsorption of arsenic on polyaluminum granulate

Jasmin Mertens, Jérôme Rose, Ralf Kägi, Perrine Chaurand, Michael

Plötze, Bernhard Wehrli, and Gerhard Furrer

Submitted for publication in Environmental Science and Technology

- 81 -
Adsorption of arsenic on polyaluminum granulate

Abstract

Polyaluminum granulates (PAG) of 1-3 mm size and with a high content of aluminum
nanoclusters were produced from polyaluminium chloride. The kinetics and efficiencies of
arsenite and arsenate removal from natural water were evaluated. PAG was characterized to be
meso- and macroporous, with a specific surface area of 35 ± 1 m2/g and skeleton density of
1.93 g/cm3. Adsorption experiments were conducted with PAG in deionized water and in
synthetic groundwater. While As(III) sorption followed the Freundlich model, As(V)
adsorption was best described by a Langmuir isotherm with a maximum uptake capacity of
~200
mol/g in synthetic groundwater. The removal of As(III) reached equilibrium within 40
hours, and As(V) was removed almost entirely from solution within 20 hours. Micro X-ray
fluorescence spectroscopy and electron microscopy revealed that As(III) was distributed
uniformly within the grain, whereas As(V) diffused only up to 81
m into PAG. The 5-fold
washing of PAG with deionized water over 60 hours cumulatively desorbed up to 50 % As(III)
and up to 5 % As(V). The results imply that As(V) is adsorbed 3 times faster while being
transported 105 times slower than As(III) in Al hydroxide material.

5.1. Introduction

Dissolution of arsenic-bearing minerals is the main source for elevated arsenic concentrations
in drinking water resources. Trivalent arsenite (H3AsO3) prevails over pentavalent arsenate
species under reducing conditions. As(V) is present under oxidizing conditions as H2AsO4- and
HAsO42- below and above pH 6.8, respectively. As(III) is mainly present as the neutral H3AsO3
below pH 9 (Smedley and Kinniburgh, 2002). Arsenic is classified as carcinogenic, and the
major source of exposure for people is drinking water. To reduce the risk of As-related
diseases, the EU drinking-water directive allows a maximum contaminant level (MCL) of 10

g/L (EC, 1998).
Among various arsenic removal techniques, the processes of coagulation-coprecipitation and
adsorption using iron- or aluminum-based materials are applied most frequently. Coagulation-
coprecipitation systems are most cost-effective and therefore employed in large-scale
operations (EPA, 2002). Especially polyaluminum chlorides (PACl) have received much
attention as coagulants due to their high efficiency at low dosage, low cost and convenient
application (Duan and Gregory, 2003). PACl generally consists of Al monomers, oligomers,
and the nanoclusters Al13 (AlO4Al12(OH)24H2O127+), and Al30 (Al2O8Al28(OH)56(H2O)2618+)

- 82 -
 Chapter 5

(Chen et al., 2006) a large aqueous aluminum hydroxide complex with 2 nm in length. Al13 and
Al30 exhibit high reactivity due to a great number of hydroxide ions and water groups (Furrer et
a., 1992; Casey et al., 2005). PACl has been applied for arsenic removal in lab-scale studies
(Fan et al., 2003; Mertens et al., 2012) and field studies (Tubic et al., 2010).
Compared to coagulants, solid material has the advantage that it can be easily applied in flow-
through water-treatment processes, due to fast separation of the pollutant from the aqueous
solution. Various Fe- and Al-based adsorbents including ferrihydrite (Raven et al., 1998),
hydrous ferric oxide (Wilkie and Hering, 1996), granular ferric hydroxide (Badruzzaman,
2004), Fe2O3 and Al2O3 (Jeong et al., 2007), mesoporous alumina (Kim et al., 2004), activated
alumina (Singh and Pant, 2004), and amorphous aluminum hydroxide (Anderson et al., 1976)
have been tested for As uptake. Adsorption of arsenic onto activated alumina has been rated as
the best available technology, due to low operating costs and high As removal efficiency for
this application (EPA; 2002; Mondal et al., 2006), and activated alumina packed-bed systems
have been used effectively in wellhead and point-of-use water-treatment devices (Lin and Wu,
2001).
The reaction of an adsorbate with the solid-solution interface is crucial for the pollutant’s
mobility control. It includes the transport of the adsorbate to the surface by convection and
molecular diffusion, interaction with the surface, intraparticle diffusion, dehydration and
surface-complex formation (Stumm, 1992). The transport of inorganic ions like arsenic in
granular, porous solids occurs via diffusion in the pore space and at the solid surface. The
sorption capacity and time dependence for sorption are derived mainly from macroscopic
evidence, such as dissolved concentration data. Adsorption of arsenate and arsenite is known to
be relatively fast (Raven et al., 1998; Anderson et al., 1976). Slower sorption processes are
related to surface precipitation or to diffusion into the pore space of particles (Fuller et al.,
1993).
While the adsorption kinetics of both arsenic species on iron-based hydroxide grains are well
documented, limited information is available on the transport of As in aluminum-based porous
sorbents. Adsorption on activated alumina was found to be slow, and was described by a pore
diffusion model with high tortuosity indicating an irregular twisted pore structure (Lin and Wu,
2001). Optimizations of aluminum materials, e.g. mesoporous alumina (Kim et al., 2004),
regarding pore size and pore distribution have improved arsenic uptake and kinetics.
Previous studies estimated the fate of arsenic in porous iron- and aluminum- based solid
materials during transport processes based on the concentration measured in solution (Raven et
al., 1998; Kim et al., 2004; Singh and Pant, 2004; Lin and Wu, 2001, Dutta et al., 2004).

- 83 -
Adsorption of arsenic on polyaluminum granulate

Here we combine analyses of the As fraction remaining in solution with spectroscopic

evidence of arsenate and arsenite transport in a solid Al oxide material. We tested the removal
of arsenic uptake with polyaluminum granulate (PAG), that was formed from polyaluminum
chloride solutions after desalting and neutralization. Since it is not clearly distinguishable if the
solid surfaces were made of polyaluminum oxides or Al oxides, we used the term “Al oxide
surfaces”. The physicochemical properties regarding surface area, porosity and density of the
PAG were evaluated. PAG was applied to remove As(III) and As(V) from water, and
adsorption equilibria and kinetics were described by adsorption isotherms and by a pseudo-
second-order rate model, respectively. We employed micro-X-ray fluorescence (
XRF) and
scanning electron microscopy (SEM) to investigate arsenic transport into the grains. By means
of these techniques, actual transport mechanisms of As(V) and As(III) in porous aluminum
hydroxides could be visualized, and were used to interpret data from aqueous measurements in
more detail.

5.2. Materials and methods

5.2.1. Chemicals

All chemicals (NaAsO2, NaHAsO4·7H2O, CaCO3, NaHCO3, MgCl2, NaOH, disodium citrate)
were reagent grade from Sigma Aldrich and Fluka. High-purity 18 M water (Millipore, US)
was used to prepare stock solutions and synthetic groundwater. Synthetic groundwater
containing 1 mM CaCO3, 8 mM NaHCO3 and 0.6 mM MgCl2 (Table 5.1) was prepared after
Roberts et al. (2004) simulating the composition from As-rich water in the Pannonian Basin,
Eastern Europe. The pH was adjusted with CO2 bubbling through 1 M NaOH to 7.5 ± 0.1.

5.2.2. Preparation of PAG

Polyaluminum chloride (PACl) solutions with high content of Al30 and concentrations of 1.2 M
Altot (= 40 mM Al30) were obtained by dissolving 234.3 g Locron-S powder (Clariant, Muttenz,
Switzerland) in 1L high-purity 18 M water. The granulate material was generated by
coagulation from PACl solutions after four repetitive cycles of aggregation, centrifugation,
washing with ultrapure water to remove the chloride ions, and consecutive air-drying. Dried
material was grinded and pressed into pellets using a pill press at a pressure of 103 kg/cm2.

- 84 -
 Chapter 5

Table 5.1: Chemical composition of synthetic groundwater

pH 7.5 ± 0.1
mg/L mmol/L
Alkalinity - 8.0
Chloride 21.3 0.6
Sodium 184.0 8.0
Calcium 40.0 1.0
Magnesium 15.0 0.6

5.2.3. Characterization of PAG

The specific surface area was determined with the BET method from 11-point N2 adsorption
isotherm data at 77.3 K in the p/p0-range 0.05 - 0.3 (Autosorb 1MP, Quantachrome,
Odelzhausen, Germany). Before analysis, the samples (0.28 g) were outgassed for 15 hours at
150°C. The size distribution and volume of macro- and mesopores were determined using low
(375 kPa) and high (400 MPa) pressure mercury porosimeters (Pascal 140 and Pascal 440,
Porotec, Hofheim, Germany). Skeleton density was measured in triplicates with a pycnometer
using cyclohexane. Randomly oriented powder samples were analyzed to evaluate whether
crystalline precipitates formed inside the granulate material. For that purpose, bulk mineralogy
was determined using a Theta-Theta-X-ray Diffractometer (D8 Advance, Bruker AXS,
Karlsruhe, Germany) with Co-K
radiation (40 kV, 30 mA) operating in step scan mode over
an angular range of 4-70° with 0.02° (2 theta) steps and 4 s count time. Prior to analysis,
aliquot (ca. 1 g) samples from bulk dried granulate were grinded by hand to obtain <20
m
particles. The granulate material was tested for its mechanical stability by a four-step washing
procedure (Figure 5.A1).

5.2.4. Adsorption experiments

Adsorption is most commonly described by the Langmuir and Freundlich isotherms (Raven et
al., 1998; Lin and Wu, 2001; Duddridge and Wainwright, 1981; Reed and Matsumoto, 1993).
The Langmuir adsorption isotherm is obtained by the equation
C KL
q = qmax (5.1)
(1+ C K L )

- 85 -
Adsorption of arsenic on polyaluminum granulate

where q is the quantity of the sorbed compound (mmol g-1), C is the equilibrium concentration
(mmol L-1) of the dissolved compound, and KL is the Langmuir adsorption constant (L mmol-
1
). qmax is the maximum uptake (= saturation) capacity, assuming monolayer adsorption at
uniform adsorption sites.
The Freundlich isotherm, is based on the assumption of multi layer coating of the solid surface
with preferential adsorption at high-energy sites (steep section of the curve) and unlimited
sorption to sites with lower energy with increasing concentration in solution (flattening of the
curve) (Merkel and Planer-Friedrich, 2008). The Freundlich model describes an exponential
relationship between the sorbed and dissolved molecules and is described by the equation

q = C n KF (5.2)

where n and KF are constants. All adsorption experiments were carried out in 50 ml
polyropylene tubes with 45 ml aliquots of deionized water or synthetic groundwater adjusted to
pH 7.5 ± 0.2. To each batch, 5 ml of As(III) or As(V) were added from stock solutions to initial
arsenic concentrations in the range of 0.3 to 3.8 mmol/L to yield a total volume of 50 ml.
Finally, 0.5 g dry Al granulate material was added, to receive a water-to-solid (w/s) ratio of
1/100. Tubes were shaken horizontally at 250 rpm for 20 hours at 22-24 °C. To study the
adsorption mechanism of As(III), an experiment was conducted with 20 hours As(V)
adsorption at initial concentrations of 3.8 mmol/L followed by 20 hours As(III) adsorption at
initial concentrations of 3.8 mmol/L.
To determine the kinetics of the adsorption, batch experiments were carried out as described
above in deionized water with As(III) or As(V) at initial concentrations of 3.8 mM. In order to
understand the effect of capillary forces for the kinetics of As uptake, the experiments lasting
1, 5, and 20 hours were repeated with PAG that was previously water-saturated in deionized
water, shaking at 250 rpm for 15 hours. For each time step three batches were prepared, and
the solution was separated from the granulate material by decantation. The supernatant solution
was filtered through 0.45
m nylon filters (Infochroma AG, Zug, Switzerland), and the filtrate
was analyzed for As(III) and cation concentration. The arsenic removal was calculated from
the difference between the initial and final concentrations in each batch experiment.

5.2.5. Chemical analysis

As(III) was determined with a hydride generation atomic fluorescence spectrometer

(Millennium Merlin/ Millennium Excalibur System, PS Analytical Ltd, Kent, UK). To detect
only As(III), a pH 4.8 disodium citrate buffer (0.5 M) was used (detection limit was 0.7
g/L).

- 86 -
 Chapter 5

Cations (Astot, Al, Ca, Mg, Na) were measured by inductively coupled plasma-mass
spectrometry (ICP-MS, Agilent 7500cx, Agilent US) and by inductively coupled optical-
emission spectrometry (ICP-OES, Vista-MPX, Varian). Standards were prepared by dilution of
single element standards (As) and multi element standards (Al, Ca, Mg, Na) from Merck
(Darmstadt, Germany). The detection limits were 0.5 – 10
g/L for ICP-MS and 50-100
g/L
for ICP-OES.

5.2.6. Desorption experiments

The reversibility of the adsorption was determined using granulate from batch experiments
loaded with initial As(III) or As(V) concentration of 2 mmol/L. Desorption of arsenic was
studied in five steps, and each step involved the exchange of solution with 50 ml arsenic-free
synthetic or deionized water. This procedure was undertaken after 1 h, 5 h, 24 h, 48 h and 60 h
contact with the As-free solution. The decanted solution was filtered through 0.45
m nylon
filters, and the filtrate was analyzed for As(tot) and As(III).

5.2.7. As distribution on individual grains

The spatial distribution of As in individual grains from 20h experiments in deionized water
with 2.6 mmol/L and 3.8 mmol/L initial As(III) and As(V) concentrations was analyzed using
micro X-ray fluorescence (
XRF) and electron microscopy (EM).

5.2.7.1.
XRF analysis

Individual grains from batch experiments with 2.6 mmol/L As(III) or As(V) in deionized water
were air-dried for 48 hours. Further, they were embedded in an araldite resin and dried for 24
hours at room temperature, before cutting them into thin sections of approximately 1 mm
thickness with a diamond wire saw. The chemical maps of the thin section were carried out at
the lab-scale on a microscope (XGT7000 Horiba Jobin Yvon, Horiba Ltd., Kyoto, Japan),
equipped with an X-ray tube producing a finely focused and high-intensity beam with a 10 μm
spot size (Rh X-ray tube, accelerating voltage of 30 kV, current of 1 mA). X-ray emission from
the irradiated sample was detected with a liquid-nitrogen-cooled high-purity Si detector.
Samples were scanned in full vacuum mode with a total counting time of 7
600 s. Elemental
profiles through the grain section were obtained from As and Al chemical maps by integrating
the intensity of As-K and Al-K lines every 6 μm (i.e. each pixel size) along a line (20 pixel
width) from the surface to the core of the grain.

- 87 -
Adsorption of arsenic on polyaluminum granulate

5.2.7.2. Electron microscopy

Individual grains of PAG in contact with 2.6 mmol/L and 3.8 mmol/L As(III) or As(V)
solutions were embedded in resin and polished until the cross-sections of the grains were
exposed. The surface of As-free, bulk PAG grains and uncoated As-loaded grains were
analyzed in low vacuum mode in a scanning electron microscope (SEM; Nova NanoSEM230,
FEI, Hillsboro, USA) operated at an acceleration voltage of 20 kV. The backscattered electron
(BSE) signal was used for imaging. Elemental analysis was performed with the energy
dispersive X-Ray spectroscopy system (EDX, X-MAX 80, Oxford Instruments, Oxfordshire,
UK) attached to the microscope. Regions of different elemental abundances were identified
and interpreted based on elemental distribution maps of C, Cl and As using Cameo+, which is
implemented in the INCA Software package (suite 4.15, Oxford Instruments Analytical Ltd,
UK).

5.3. Results

5.3.1. Physicochemical characteristics of polyaluminum granulate

The granulate material produced from polyaluminum flocs was 1-3 mm in size and had a
skeleton density of 1.93 g/cm3. Following the definitions by IUPAC(1994), the granulate
material had mainly meso- and macropores (Table 5.2) with a bimodal pore-size distribution
and maxima at 40-50 nm (12.8% of total porosity) and 2-4 nm (13.0% of total porosity)
(Figure 5.1). The BET surface area from duplicate measurements was 35 ± 1 m2/g. From a
structural point of view, the PAG samples were X-ray amorphous. The XRD pattern (Figure
5.A2) exhibits the same very broad lines as fort he Locron material. Only traces of gibbsite
were detected additionally. SEM imaging showed a heterogeneous surface structure of PAG
(Figure 5.2). The physicochemical data for the Al granulate are summarized in Table 5.2.

Table 5.2: Physicochemical properties of polyaluminum granulate

Size 1-3 mm
Skeleton density 1.93 g/cm3
BET surface area 35 ± 1 m2/g
Total Porosity* 219 mm3/g
Macropores (> 50 nm) 79.4 mm3/g
Mesopores (1-50 nm) 140 mm3/g
* Approximately 50 % of the pores are in the 25-75 nm range (compare Figure 5.1).

- 88 -
 Chapter 5

Figure 5.1: Pore distribution of meso- and macropores in polyaluminum granulate.

Figure 5.2: Scanning electron microscopy image of the polyaluminum granulate surface.

5.3.2. As(III) and As(V) adsorption

Both, the Langmuir and Freundlich isotherms were fitted to the As(III) and As(V) adsorption
data on PAG, which were acquired in synthetic groundwater and deionized water at pH 7.5 ±
0.1 after 20 hours equilibration time (Table 5.3). The maximum uptake capacity for As(V) was

- 89 -
Adsorption of arsenic on polyaluminum granulate

reached in synthetic groundwater at an adsorbed content of 198
mol/g. In deionized water the
maximum uptake capacity was not reached (Figure 5.3), the highest sorbed concentration was
375
mol/g. As(V) at initial concentrations of up to 2 mmol/L was removed from synthetic
groundwater by 85 – 99.8%. Removal efficiency decreased to 45-70% at the highest
concentrations used. In deionized water As(V) was removed by 91-99.3% (average = 96.2 ±
2.5%). The sorption isotherm for As(III) showed linear sorption in synthetic groundwater.
Nonlinear regression coefficients (R2) indicated that both isotherms described well the arsenic
partitioning between the solid and aqueous phase (Figure 5.4, Table 5.3). As(III) removal
efficiency in synthetic groundwater was 57% at low concentrations (0.3-1.3
mol/L). With
increasing As(III) concentration the removal efficiency decreased to 33% (at highest initial As
concentrations of 2.6 mmol/L). While the Freundlich isotherm represented the linear behavior
of the data in synthetic water very well, the Langmuir isotherm closely matched the As(III)
removal from deionized water (Table 5.3), where As(III) concentrations were reduced by 50-
95% (73.4 ± 13.7%).

Figure 5.3: Langmuir adsorption isotherm for As(V) on polyaluminum granulate (PAG) in synthetic
groundwater and deionized water. Initial arsenic concentrations range between 0.27
M and 4 mM.
Initial pH was at 7.5 ± 0.2, resulting pH was between 5.1 and 5.6 in synthetic groundwater and between
4.1 and 4.5 in deionized water. Dashed and dotted lines represent the model data for synthetic
groundwater, and deionized water, respectively.

- 90 -
 Chapter 5

Table 5.3: Isotherm parameters for adsorption of As(III) and As(V) on PAG in synthetic groundwater
(SW)1 and deionized water (DW).
Langmuir isotherm Freundlich isotherm
KL Qmax R2 KF n R2
L mmol-1 mmol g-1 mmol g-1 (L mmol) -1
As(V)
SW 38.0 0.190 0.896 0.194 0.29 0.873
DW 6.79 0.940 0.975 3.168 0.90 0.975
As(III)
SW 0.129 0.479 1.00 0.053 0.88 1.00
DW 2.69 0.240 0.924 0.164 0.46 0.853
1
For the composition of synthetic groundwater see Table 5.1.

Figure 5.4: Freundlich adsorption isotherms for As(III) on polyaluminum granulate in synthetic and
deionized water. Initial arsenic concentrations ranged between 0.27
M and 4 mM. Initial pH was at 7.5
± 0.2 and final pH values were between 5.1 and 5.6 in synthetic groundwater and between 4.1 and 4.5
in deionized water. Dashed and dotted lines represent the model data for synthetic groundwater and for
deionized water, respectively.

5.3.3. Adsorption kinetics

The adsorption of arsenite and arsenate in deionized water occurred rapidly within the first 5
hours, and then slowed down considerably until 60.2% and 99.8% arsenic uptake was reached

- 91 -
Adsorption of arsenic on polyaluminum granulate

after 20-40 hours for both As(III) and As(V), respectively (Figure 5.5). Results of experiments
with polyaluminum granulates that were previously water saturated showed no difference
compared to non-saturated granulates for the uptake of As(V). However, As(III) adsorption
was less efficient with pre-saturated than with non-saturated PAG and equilibrium was reached
after 5 and 40 hours, respectively (Figure 5.5).
Kinetic studies revealed that the pseudo-second-order rate law represents a useful model for
adsorption processes controlled by chemical reactions (Ho and McKay, 1999). We employed
the relation for pseudo-second-order kinetics (Equation 3), which was also used as a kinetic
model of As sorption on mesoporous alumina by Kim et al. (2004)

t 1 1
= 2
+ t
qt k ad qe qe (5.3)

where kad is the rate constant of the adsorption (g mg-1 min-1) and qt represents the sorbed
arsenic concentration at time t (mg g-1), and qe denotes the sorbed arsenic concentration at
1 1
-1
equilibrium (mg g ). The terms and
k ad qe 2
qe are determined from the linear plot of t
versus t/qt. The initial sorption rate h (in mg g-1 min-1) is defined as

h = k ad qe2 (5.4)

The pseudo-second-order model described the data well with R2 values of 0.99 for both arsenic
species (Figure 5.6). The sorption rate h for As(V) is three times faster compared to the As(III)
sorption rate (Table 5.4). The half life of As(V) and As(III) removal is reached in 3.5 and 4
hours, respectively.

- 92 -
 Chapter 5

Figure 5.5: As(III) and As(V) uptake by water saturated and unsaturated polyaluminum granulate with
time. [As]initial = 3.8 mM.

Figure 5.6: Pseudo-second-order rate model for As(III) and As(V) removal with unsaturated
polyaluminum granulate. qt is calculated from values in Figure 5.5.

- 93 -
Adsorption of arsenic on polyaluminum granulate

Table 5.4: Sorption kinetics (pseudo-second-order)

Adsorbate Uptake Rate constant Initial sorption
concentration at rate
equilibrium
qe kads h
mg g-1 g mg-1 min-1 mg g-1 min-1
As(V) 526.3 2.5 x 10-5 6.87
As(III) 322.6 2.1 x 10-5 2.15

5.3.4. Arsenic desorption

The reversibility of As(III) and As(V) sorption to PAG in water was investigated with PAG
loaded with 2.6 mmol/g As(III) or 3.6 mmol/g As(V) by exchanging the solution (50 ml) with
As-free synthetic groundwater or deionized water after 1, 5, 24, 48 and 60 hours. In the
subsequent desorption experiment with synthetic groundwater, equal amounts of 12% As(III)
were desorbed from the granulate during the first three steps. Desorption was highest at the
first step, declined at the following treatments and reached a cumulative value of nearly 50% of
the adsorbed concentration after 5 desorption steps (Figure 5.7). As(V) remained almost

Figure 5.7: Desorption of As(III) and As(V) from aluminum granulates with arsenic-free synthetic and
deionized water in 5 steps over three days.

- 94 -
 Chapter 5

completely adsorbed with a cumulative removal of only 5% after the complete sequence. The
desorption in deionized water was less extensive for both arsenic species, with 35% and 1%
desorption of As(III) and As(V), respectively (Figure 5.7).

5.3.5. Arsenic distribution in PAG

The analysis of PAG thin sections by
XRF showed that As(V) was concentrated on the
surface layer of the PAG (Figure 5.8), whereas As(III) was distributed uniformly within the
grain (Figure 5.9). A detailed SEM investigation of the rim area of PAG exposed to 3.8
mmol/L As(V) for 1, 5, 20 and 80 hours showed a clearly distinguishable As front, that moved
into the grain with time (Figure 5.10). The intensity of the As (L
) emission line in the As-rich
and the As-poor area showed that no As is present behind the As front (Figure 5.10). The
spatial-temporal advance of the As-front was derived by measuring the thickness of the
arsenic-rich rim (10 measurements on a single grain) for each reported time step and ranged
from 29±9
m after 1 hour up to 73±8
m after 80 hours (Table 5.5). Backscattered electron
(BSE) images of the entire cross section of individual grains revealed the constant thickness of
the As rim, appearing as light grey-white band surrounding the dark gray PAG grain (Figure
5.A3, Supplemental information). The occasional broadening of the As-rich layer can be
explained by a cutting effect due to non-perpendicular polishing of the grain’s surface.
The investigation of the transport of As(III) into the grain after As(V) adsorption on the PAG
rim revealed that As(III) was partly inhibited by previous As(V) sorption. The Al-normalized
As intensities (As-L) from the central part of PAGs were lower for As(III) that was adsorbed
after As(V) than for pure As(III) sorption. No increase in the As intensities were observed in
adsorption experiments where only As(V) was used (Figure 5.A5). As(III) removal from
solution was reduced by 15% when As(V) was adsorbed to the grain before it got in contact
with As(III) solution.

- 95 -
Adsorption of arsenic on polyaluminum granulate

Figure 5.8: As(V) distribution. a) Bicolor (RG) combined image showing arsenic distribution in a
crosscut of PAG collected by
XRF. Arsenic is displayed in red color, aluminium in green color. b) As
and Al profiles along the AB line (20 pixels width). As(V) was adsorbed to the surface layers of the
PAG. No arsenic was found in the inner core of the grain. [As(V)]total = 3.7 mM, 99.6 % of arsenic was
adsorbed, 0.02 % were removed during 24 hours of desorption at pH 6.24.

Table 5.5: As(V) diffusion into polyaluminum granulate

Time t [As]aq [As]ads [As]ads Depth x of As- As flux
rich rim a Fb
h
mol dm-3 mmol dm-3 mmol g-1
m mol cm-2 s-1
As(V)
0 3480 0 0.0 0 -
1 2150 1.3 2.7 38±9 7.4 10-6
5 946 2.5 5.1 52±7 1.6 10-6
20 69 3.4 6.8 63±9 3.2 10-7
80 1 3.5 7.0 73±8 6.2 10-9
As(III)
0 3404 0 0 0 -
20 1660 1744 3.5 1000 4.8 10-1
a
Average of 10 measurements of the As-rich rim on grain images obtained by SEM (Figure 5.A3)
b
(
[As]aq * Vaq/Asurf,tot)/
t

- 96 -
 Chapter 5

Figure 5.9: As(III) distribution. a) Red-green bicolor combined image showing arsenic and aluminum
distribution in a crosscut of polyaluminum granulate collected by
XRF. Arsenic is displayed in red
color, aluminum in green color. As(III) is evenly distributed within the grain. The slight asymmetry of
the two colors is due to an optical effect of a cut area that is not entirely parallel to the surface of the
grain. b) As and Al profiles along the AB line (20 pixels width). The profile shows that the arsenic had
a constant intensity throughout the entire grain. [As(III)]total = 2.7 mM, 57 % of arsenic was adsorbed,
9.3 % were removed during 24 hours of desorption at pH 6.15.

- 97 -
Adsorption of arsenic on polyaluminum granulate

Figure 5.10 SEM images of the rim area of an As(V) loaded grain ([As(V)]aq,initial = 3.8 mM) after 1, 5,
20 and 80 hours adsorption time. Red areas (EDX measurements) represent As adsorbed in the grain,
green areas reflect areas without As. The element spectrum for the As-rich and As-poor areas are given
for 20 hours adsorption, arrows indicate the locations of the As L
intensity peak.

- 98 -
 Chapter 5

5.4. Discussion

5.4.1. Arsenic adsorption

The arsenate adsorption capacity of PAG in naturally buffered water derived from the
Langmuir model is 198
mol As(V)/g PAG. This value is slightly above the reported range for
other aluminum-based sorbents, e.g. activated alumina (Lin and Wu, 2001) or Al2O3 (Jeong et
al., 2007), but about 8.5 x lower compared to alumina produced with uniform pore-size
distribution and high surface area (Table 5.6). Results from
XRF and electron microscopy
revealed that As(V) is adsorbed to a surface layer of 80
m of the grain’s radius under the
given experimental conditions. Thus, As(V) sorption capacity per gram PAG could be
enhanced by reducing the particle radius to 90
m. The amount of As(III) sorbed to PAG was
always less than As(V) adsorption at the same total arsenic concentrations. The As(III)
adsorption isotherm in naturally buffered water was linear. Although both isotherm models
give a correlation coefficient very close to 1.0 when reaching the smallest value for the sum of
the least squares, there is no indication in the data that As(III) sorption follows a monolayer
adsorption model. Adsorption of both arsenic species was found to be higher in deionized
water than in water with a similar composition of a sodium-bicarbonate water type. Analysis of
the water chemistry after the As removal experiments showed that the concentration of calcium
was reduced by 90-97%, while measured magnesium and sodium concentrations were very
similar to initial concentrations (Figure 5.A4). Therefore, it is proposed that calcium in natural
waters competes for sorption sites on PAG with arsenic.

Table 5.6: Characteristica and adsorption kinetics of various Al sorbents

Material Particle Surface Pore Pore Qmax Time to Pseudo-second
size Area size volume equilibrium order rate
constant kads
mm m2 g-1 nm cm3g-1
mol g-1 hours g Al sorbent
(%) mg-1As min-1
Polyaluminum 1-3 35 ± 1 1-10 0.22 As(V): 198 20 h As(V) As(V): 2.5 10-5
granulate1 40-50 (38) As(III): 240 40 h As(III) As(III): 2.1 10-5

Mesoporous n.a. 307 3.5 3900 As(V):1620 5 h As(V) As(V): 1.5 10-2
alumina2 As(III): 630 5 h As(III) As(III): 4.4 10-4
Activated 0.162 118 n.a. (29.3) As(V): 132 170 h As(V) n.a.
Alumina3 0.104 115 n.a. (25.7) As(III): 46 40 h As(III)
0.067 116 n.a. (17.2)
Al2O3 0.005- 0.55 0.0756 0.01 As(V): 170 1 h As(V) As(V): 2.64
(ALO101)4 0.045
n.a.: not available
1
This work, 2 Kim et al. (2004) , 3 Lin and Wu (2001) , 4 Jeong et al. (2007)

- 99 -
Adsorption of arsenic on polyaluminum granulate

5.4.2. Sorption reversibility

The desorption data of As(III) and As(V) implied a different sorption mechanism for the two
arsenic species on PAG. About 50% of As(III) is desorbed whereas As(V) remained adsorbed
to PAG. These findings indicate that about one half of As(III) adsorbed to PAG forms strong
inner-sphere bonds with Al whereas the other half was bound by weak, outer-sphere
complexes. As(V) sorption is dominated by strong inner-sphere complexes. This complies with
absorption spectroscopy findings of As(III) and As(V) binding mechanisms on 
Al2O3 by

Arai et al. (2001).

5.4.3. Adsorption kinetics and intra-particle distribution

The pseudo-second-order rate constants for As(III) and As(V) are in the same order of
magnitude (Table 5.4), with As(V) being removed from the solution slightly faster than
As(III). Even though the kinetics of the two arsenic species were similar, the examination of
PAG cross sections revealed that As(V), which was removed almost entirely from solution,
concentrated at the outermost zones of the granulate (less than 10% of the radius), while
As(III), which was removed to less than 50% from the solution, was distributed uniformly
throughout the entire granulate (Figures 5.8 and 5.9). In the course of 80 hours, the movement
of As(V) from the bulk solution into the grain slowed down considerably (Table 5.5),
principally due to decreasing concentration in the bulk solution. Consequently, the As(V) mass
flux from 50 ml water volume (Vaq) through the total surface area of adsorbent material used in
one batch experiment (Asurf,tot = 25 cm2, under the assumption that the grains have a spherical
shape and the average grain radius is 1 mm) was reduced from initially 7.4 10-6 to finally 6.2
10-9 mol s-1 cm-2 (Table 5.5). However, if the As concentration in the bulk solution had been
kept constant, the As(V) flux may have remained closer to the initial value. An estimate of the
characteristic time
diff for a diffusion front is given by Kalnin and Kotomin (2001)

diff = x2 (2Deff)-1 (5.5)

Here x stands for the characteristic diffusion length and Deff represents the effective diffusion
coefficient of As in the PAG pore space. With a typical As-rich rim of 50 μm after 5 hours
(Table 5.5) and an estimated As(V) Deff ~ 3 10-6 cm2 s-1 (7 10-6 according to Takahashi et al.
(2011), multiplied by 40% porosity of PAG) we obtain a characteristic diffusion time of 4 s.

- 100 -
 Chapter 5

The difference to the reaction time span of 5 hours (slower by a factor 4500) indicates that the
transport process is dominated by effective adsorption in the outermost zone of the granulate
body. Diffusion reaches further into the granulate only after all available adsorption sites near
the surface are saturated (Figure 5.8). Our results indicate, that As(V) uptake could be
accelerated from 20 h to 1-5 hours by reducing the particle radius to 40-50
m (Table 5.5) at
the given experimental conditions.
In contrast to As(V), As(III) was distributed uniformly within the grain after the reaction time
of 20h. The As(III) flux into the grain was calculated to be five magnitudes higher than for
As(V) while the diffusion coefficient is within the same range (Table 5.5). This showed that
H3AsO3 is transported further into the pore space of the granulate material than As(V) due to
weaker interactions with Al oxide surfaces. A characteristic diffusion time at the 1 mm scale
observed for a 20 hour reaction time of As(III) would be around 40 min (Table 5.5). This
indicates that transport and reaction occur at more comparable time-scales than in the case of
As(V) (slower by a factor 600). As the surface sites do not reach saturation and equilibrium
concentrations in the pore space remain high, the As(III) species are transported deeper into the
granulate (Figure 5.9).
In the case when the granulates were water-saturated prior to the exposure to arsenic solution,
the removal of As(III) was reduced by 20 %. This indicates that As(III) transport is partly due
to advective solution transport by capillary forces and not only diffusion. This is confirmed by
the reduced arsenic signal intensity in the core of a grain exposed first to As(V)-rich water and
then As(III)-rich water compared to granulates after As(III) sorption alone (Figure 5.A5).
Therefore, a lower but faster uptake of As(III) has to be expected when applying PAG in flow-
through systems, where the grains are constantly saturated. In opposite, the transport of As(V)
in PAG is not inhibited by pre-saturation. The As signal intensities in Figure 5.A5 also show
that initial As sorption on the PAG surface is not inhibiting As(III) (and consequently As(V))
to move further into the grain. Hence, pore space is not blocked by previous As adsorption.

5.5. Conclusion

PAG has higher removal potential for As from natural well water than activated alumina. The
kinetic model indicated that the sorption rate is slightly faster for As(V) than for As(III), which
can be attributed to a high affinity of As(V) for Al oxide surfaces. Strong binding of As(V) to
PAG resulted in significant retention and a low penetration depth, and consequently in high
As(V) concentrations in the pores close to the surface of PAG. Contrary, H3AsO3 was

- 101 -
Adsorption of arsenic on polyaluminum granulate

distributed within the entire PAG area, but at lower concentrations than As(V) in the near-
surface pores.
Our results indicate that the transport of the two As species in metal sorbents differs due to
different retention behavior. Based on this knowledge, maximum effectiveness of arsenic
removal can be reached by producing Al oxide sorbents with the optimum size for the two
arsenic species.

5.6. Supplemental material

5.6.1. Supplemental figures

Figure 5.A1: Mechanical stability of PAG during washing procedures. White and black triangles
represent duplicate measurements. Washing cycle 1: rinsed 4 times with deionized water, washing cycle
2: shaken for 1 hour at 150 rpm, cycle 3: shaken for 16 hours at 150 rpm, cycle 4: shaken for 72 hours
at 150 rpm; After each procedure, supernatant water was decanted, and granulates were dried for 6 hrs
at 40 °C.

- 102 -
 Chapter 5

Figure 5.A2: XRD pattern of polyaluminum granulate (PAG) compared to its base material, the
polyaluminum chloride Locron and the XRD patterns of the Al minerals corundum
and gibbsite. The measured XRD patterns exhibit the same very broad lines for polyaluminum
granulate as well as the Locron powder, indicating that the base material consists of almost X-ray
amorphous material and did not change its crystallographic composition. Additionally, three small
peaks in the polyaluminum granulate were identified, which might be an indication that probably some
gibbsite was formed during the production process of polyaluminum granulate. The peaks were too
weak, however, to clearly identify a crystalline gibbsite phase.

Figure 5.A3: BSE images of PAG, which was in contact with 3.8 mmol As/L for 1, 5, 20, and 80
hours. Polished thin sections show the regularity of the As-rich rims allowing consistent rim-width
measurements as presented in Table 5.5.

- 103 -
Adsorption of arsenic on polyaluminum granulate

Figure 5.A4: Initial and remaining concentrations of calcium (black), magnesium (red) and sodium
(green) remaining in supernatant water after 20 hours contact time with PAG (results from adsorption
batch experiments). Calcium was removed considerably by 90 - 97%.

Al

As
C L-line

Si S

Figure 5.A5: Normalized spectra of the centre area of three batches of PAGs loaded with As(V),
As(III), and both, As(V) and As(III) during 20 hours. * As(V) was adsorbed for 20 hours before As(III)
was adsorbed for 20 hours

- 104 -
Chapter 6

Arsenate adsorption on Al nanoparticles and Al-

based sorbents during water treatment – an EXAFS
study

Jasmin Mertens, Jérôme Rose, Bernhard Wehrli, and Gerhard Furrer

In preparation for submission to Environmental Science and Technology

- 105 -
Arsenate adsorption on Al-based sorbents – EXAFS

Abstract

To reduce the health risk for millions of people worldwide, arsenic removal techniques have to
become more efficient. We compared different Al-based sorbents during water treatment
applications based on their arsenate adsorption efficiency, and used X-Ray absorption fine
structure spectroscopy (EXAFS) to investigate the structural environment of arsenate. The
atomic structure of arsenic may have implications for the stability of the As-Al surface
complexes and the recycling of the sorbent material. Arsenate at initial concentrations of 0.3
mol/L to 4 mmol/L was adsorbed to Al nanoparticles, polyaluminum chloride with high
content of Al nanoclusters, gibbsite, and polyaluminum granulate. As(V) adsorption capacity
on polyaluminum chloride and polyaluminum granulate in naturally buffered water was with
10.9 mol/m2 and 6.0 mol/m2 about 4x and 2x higher than on gibbsite. The EXAFS results
indicated that the atomic environment of arsenic is very similar for all studied Al sorbents. In
agreement with previous studies on gibbsite and allophane, As(V) formed bidentate-binuclear
complexes in interaction with Al nanoparticles and polyaluminum granulate. A multiple
scattering model was needed to adequately describe the data in the second shell. Quantitative
results showed distances of 1.68-1.69 Å for As(V)-O distance with 4 oxygen atoms around the
As(V) atom, and 3.18-3.21 Å for the As(V)-Al distance. For the multiple scattering paths, a
distance of 3.15-3.23 Å and 3.49 ± 0.1 was found for the As(V)-O-O and A(V)-O-As(V)-O-
As(V) paths, respectively. Al nanoparticles offer higher adsorption capacity for arsenic in
water treatment and exhibit a similar structural geometry as gibbsite.

6.1. Introduction

Well water contaminated by arsenic poses a serious health risk to millions of people
worldwide. To meet the maximum contaminant level for arsenic in drinking water of 10 g/L
(WHO, 1998, EC, 1998) extensive research of water treatment technologies is needed. The
most common arsenic removal processes are based on adsorption, coagulation and
precipitation with various iron- or aluminum-based sorbents (e.g. activated alumina, hydrous
ferric oxides) or coagulants (e.g. AlCl3, polyaluminum chloride, alum and FeCl3 salts) (Bratby,
2006, Chen et al., 2006, Edwards, 1994, Hering et al., 1997, Fan et al., 2003, Tubi et al.,
2010). Coagulation followed by coprecipitation has been identified as one of the most efficient
and cost-effective ex-situ water treatment technologies (EPA, 2000).

- 106 -
 Chapter 6

In the past decade, inorganic nanoparticles have received increasing interest as removal agents
for arsenic from water due to high surface area-to-mass ratio and high surface reactivity
(Auffan et al., 2009). Recent studies revealed that crystallized nano-iron oxide can adsorb
much more As per surface unit than amorphous ferrihydrite (13.6 and 2.4
mol/m2
respectively) even if the specific surface areas are similar (Auffan et al, 2008). This important
difference was discussed as a ‘nano’ effect based on adsorption site and surface structure of the
nanoparticles. In addition, XAS studies showed that As bonding characteristics on nano
maghemite changes with initial arsenic concentration (Auffan et al, 2008).
Arsenic sorption to aluminum minerals revealed a maximum saturation capacity at the surface
from 0.8 to 2.4
mol/m2, e.g. allophane: (0.8
mol/m2, 370 m2/g) (Arai et al., 2005),
-Al2O3
(1.4
mol/m2, 90.1 m2/g) (Arai et al., 2001), and gibbsite (2.45
mol/m2; 13.5 m2/g) (Ladeira
et al., 2001). X-ray absorption spectroscopy (XAS) studies have shown that As(V) forms inner-
sphere surface complexes with a bidentate-binuclear configuration on Al2O3 (Arai et al., 2001)
and gibbsite (Ladeira et al., 2001) surfaces with As(V)-Al distances of 3.11-3.19 Å. The
surface site coordination of two hexagonal rings of a gibbsite layer exhibits an Al(OH)3 surface
with 12 nearest-neighbored terminal water molecules (Wehrli et al., 1990) (Figure 6.1a). The
XAS investigation of arsenic sorption by a drinking water treatment residual from Al2(SO4)3-
flocculation revealed a combination of inner-sphere bonding types with As(V)-Al distances of
2.47 – 3.14 Å, due to mixed surface geometries (Makris et al., 2009).
In comparison to Al salts, Al nanoclusters have been reported to exhibit higher efficiency as
coagulants in water treatment (Bottero et al., 1981, Chen et al., 2006). The two main robust Al
nanoparticles are Al13 (AlO4Al12(OH)24H2O127+) and Al30 (Al2O8Al28(OH)56(H2O)2618+) with 1
nm width and 1 and 2 nm length, respectively. Al13 is a Keggin cluster of four Al trimers
placed around one tetrahedrally coordinated aluminum atom (Figure 6.1b). The polycation Al30
(Al30O8(OH)56(H2O)2618+), whose structure was XRD refined after crystallization (Rowsell and
Nazar, 2000; Allouche et al., 2000), is composed by two -Keggin Al13 clusters, linked
together by four Al(O6)-units (Figure 6.1c). These Al nano complexes stand out due to their
high specific surface charge (+7 for Al13 and +18 for Al30), deprotonation over a wide pH
range (Furrer et al. 1992; Casey et al. 2005), and high reactive surface area depending on pH
and [OH]/[Al] ratio (Bottero and Bersillon 1988, Rakotonarivo et al., 1984). An aqueous Al13
cluster exhibits twelve octahedrally coordinated Al centers, each with one terminal water
molecule (Figure 6.1b). The Al30 complex exhibits 28 surface sites, of which 24 have the same
surface sites than Al13. The four octahedrons that form the linkage between the two Al13

- 107 -
Arsenate adsorption on Al-based sorbents – EXAFS

Figure 6.1.: The structure of aluminum surface sites on three different aluminum structures with
oxygen and hydrogen depicted as black and white circles. (a) two hexa-coordinated rings of gibbsite
(after Wehrli et al., 1990). (b) The 12 octahedral surface sites of the polynuclear Al13 with one
tetrahedral center (after Wehrli et al., 1990, Furrer et al., 2002), (c) the 28 octahedral surface sites of the
nanocluster Al30 formed by 2 Al13 units linked by four octahedrons (after Casey et al., 2005). Note: only
those terminal water ligands are shown that are structural nearest neighbored to another terminal water
ligand, and therefore best suited for the formation of binuclear complexes.

clusters exhibit six terminal water molecules (Figure 6.1c). Aqueous Al nanoclusters with their
12 and 28 surface sites on one Al cluster are supposed to have higher reactivity than gibbsite.

In the present study we investigated the adsorption of As(V) on Al nanoclusters during a

coagulation-coprecipitation treatment process and on Al granulate formed from the precipitates
of polyaluminum chloride, in comparison to gibbsite. The atomic environment of arsenic in a
pure Al30 solution and in polyaluminum chloride with a high content of aluminum nanoclusters
was analyzed by X-ray absorption spectroscopy and was further compared to gibbsite and
solid-state polyaluminum granulate (PAG) as a function of arsenic concentration.

6.2. Materials and methods

6.2.1. Chemicals

All chemicals used in this work were reagent grade from Sigma-Aldrich or Merck, unless
otherwise stated. High-purity 18 M
water was used to prepare stock solutions. As(V)
solutions of 13.3 mM and 5.33 mM As were prepared by dissolving 1.04 g and 416.3 mg
Na2HAsO4 in 250 ml volumetric flasks. The 5.33 mM solution was diluted 1:1 to achieve a
final concentration of 2.66 mM. A 0.02 M NaNO3 solution was prepared by dissolving 1.69 g

- 108 -
 Chapter 6

NaNO3 in 1 L ultrapure water. Buffered water was prepared by dissolving 1 mM CaCO3, 8

mM NaHCO3 and 0.6 mM MgCl2 in 5 L 18 M water after Roberts et al. (2004).

6.2.2. Aluminum sorbents

For this work, four different types of Al-based sorbents were used. Al30 is a solution with an
Al(tot) concentration of 0.1 M with a content of 99% Al30 nanocluster and 1% Al3+ confirmed
by 27Al NMR. A polyaluminum chloride solution (referred to as PAClnano) with Al(tot)
concentration of 0.15 M was obtained by dissolving 14.64 g of the Al salt Locron-S (Clariant)
in 500 ml deionized water. Ferron measurements of the PAClnano solution revealed a
distribution of Al30, Al13 and Al3+ species of 74 %, 25.4 %, and 0.6 % at pH 5, and a share of
about 50 % Al30 and 20 % Al13 remaining in dried precipitates (Mertens et al., 2012). A
gibbsite suspension was prepared by sol-gel route method by Jolivet et al. (2000). It consists of
hydrolysis of Al nitrate up to pH = 6 to form a gel. The gel was aged for three weeks at pH = 4
to form gibbsite with Al(tot) concentration of 0.77 M Al(III). Polyaluminum granulate (PAG)
was obtained by dissolving 234.3 g Locron-S powder (Clariant) in 1 L, resulting in
polyaluminum chloride (PACl) solutions with Al(tot) concentrations of 1.2 M. Granulate was
generated from the coagulation process by neutralizing the acidic PACl solutions to pH 7 after
Mertens et al. (submitted). In order to reduce the concentration of chloride ions in dried
PAClnano precipitates, six consecutive washing cycles of aggregation and rinsing of aggregates
with high purity water were employed before final settling at pH 7 was allowed. After four
repetitive washing cycles consisting of aggregation, centrifugation, and washing with high
purity water, precipitated material was air-dried (5 days) and compressed using a pill press
operated at a pressure of 1 ton/cm2. The surface area of dried PAClnano and PAG was
determined by N2 adsorption and 11 point BET analysis.

6.2.3. Adsorption isotherm experiments

Adsorption experiments with PAClnano or PAG were performed by mixing different amounts of
13.3 mM Na2HAsO4 solution with a PAClnano solution of 0.255 mM Al(tot) or with 0.5 g PAG
in buffered synthetic water in 50 ml sterile polypropylene tubes. The pH was adjusted to pH
7.5 ± 0.2 with 0.1 mol/L NaOH. The sorption isotherm was obtained by varying the initial As
concentration from 0.3
mol/L to 4 mmol/L. The tubes with PAClnano were shaken horizontally
at 150 rpm for 30 min and the formed aggregates (pHPZC = 6.7) were allowed to settle for 5
hours. Batches containing PAG were shaken horizontally at 150 rpm for 20 hours to reach

- 109 -
Arsenate adsorption on Al-based sorbents – EXAFS

equilibrium. The supernatant solution was analyzed for total As concentrations by inductively
coupled plasma-mass spectrometry (ICP-MS, Agilent 7500cx).
EXAFS samples were prepared by adding the amount of 13.3 mM or 2.66 mM As(V) stock
solutions needed to yield initial As concentrations of 0.13 mM or 0.67 mM to 0.01 M NaNO3
electrolyte solution in a 50 ml polyethylene tube. The removal process was started by adding
Al30, PAClnano or gibbsite stock solutions to yield initial Al(tot) concentration of 75 mM. For
As removal with polyaluminum granulate a solution containing an ionic strength of 10 mM,
and an arsenic concentration of 0.26 mM was prepared by mixing 25 ml of 0.02 M NaNO3, 20
ml of ultrapure water and 5 ml of arsenic stock solution with [As] = 2.66 mM in a 50 ml
polyethylene tube. Subsequently, 0.5 g PAG was added. The pH of all solutions was adjusted
to 7.5 ± 0.2 with 0.1 M NaOH at room temperature (22-25°C). The solutions with gibbsite,
Al30 and PAClnano solutions were shaken horizontally during 30 min, after which the solution
was centrifuged for 30 min at 80’000 rpm. The PAG sample was shaken horizontally for 20
hours. The supernatant solutions were replaced by 50 ml 0.01 M NaNO3 for rinsing, shaken
during 30 min and centrifuged at 80’000 rpm for 30 min (gibbsite, Al30, PAClnano), or shaken
for 20 hours (PAG). The supernatant solution was removed and dissolved As was measured by
inductively coupled plasma-optical emission spectrometry (ICP-OES), and the remaining
precipitate was freeze-dried. As-loaded PAG was air-dried at room temperature for 3 days.

6.2.4. X-ray absorption spectroscopy

The dried As-loaded precipitates and granulates were mixed in a 1:5 ratio with
polyvinylpyrrolidone (PVP), and pressed to 20 mm pellets. In addition to As sorption samples,
XAS spectra of an As(V) solution (0.53 M Na2HAsO4) were collected as As(V) reference.
As K-edge (11.867 eV) X-ray absorption near-edge structure (XANES) and extended X-ray
absorption fine structure (EXAFS) data were recorded in fluorescence detection mode on
beamline SAMBA at the French national synchrotron facility (SOLEIL) (Belin et al., 2005),
using an Au foil for internal energy calibration. In order to minimize thermal agitation, all data
were recorded at 77 K using a liquid N2 cryostat. The beamline operated with a crystal silicon
monochromator (220), a beam of 2.75 GeV and a current of 400 mA. Multiple scans (3-9) were
collected for each sample. Sample spectra were first energy calibrated, based on the inflection
point of the Au reference foil. The determination of the E0 was performed using a second-
degree derivative of the spectra obtained from the Au foil (Au K-edge: 11.919 eV).
Subsequently, spectra were background corrected using a two-polynomial fit, and data were
merged and normalized using the ATHENA program (Ravel and Newville, 2005). Fourier

- 110 -
 Chapter 6

transforms were calculated over a k-space range of 2-12.5 Å using the Kaiser-Bessel function.
Final fitting of the spectra was done in ARTEMIS (IFEFFIT 1.2.10) on Fourier-transformed k3
weighted spectra in q space. The FEFF6 code (Newville, 2001) was used to calculate
theoretical amplitude and phase shift functions for As-O and As-Al backscatters for single and
multiple scattering (MS) paths. MS paths were calculated on a symmetric AsO4 tetrahedron
model with all As-O distances of 1.69 Å and all O-As-O angles of 110°. During fitting the
values for the distances r and the Debye-Waller factors
2 of the As-O and As-Al shells as

well as for As-O-O and As-O-As-O MS paths were allowed to vary.

6.3. Results

6.3.1. Surface area

In order to compare the arsenic surface coverage for different Al sorbents employed in this
study, the amount of sorbed arsenic was related to the surface area. For polyaluminum
granulate, the surface area was measured to be 35 ± 1 m2/g. In the case of polyaluminum
chloride solutions, an accurate measure of the surface area is difficult to achieve, because the
Al30 clusters aggregate depending on the pH and the concentration. Coagulation of Al
nanoclusters in a polyaluminum solution was observed to start at pH 5 and to increase with
increasing pH until the pH of zero charge (pHPZC) of 6.7 was reached (Furrer et al., 1992,
Casey et al., 2005). Based on the size of Al30 (1 nm x 2 nm) and a rectangular shape, the
geometric surface area was calculated as 2550 m2/g. This value implies that single Al30 clusters
interact at each of their surface sites with the solution. This is a rather unrealistic case,
particularly above pH 7, where strong aggregation and precipitation are taking place. The BET
surface area of PAClnano, precipitated at pH 7 after four washing cycles to extract chloride from
Al surfaces, yielded to 13.7 ± 0.2 m2/g. This result is unexpectedly low for the small Al entities
in solution, but can be explained by the aggregation at pH 7 and solidification during the
processes of precipitation and centrifugation, by which the flocs became a white gel-like paste
with low porosity. Between these two extreme cases of the single Al30 cluster in solution and
the precipitated aggregates, the surface area needs to be approximated in the aqueous phase.
Rakotonarivo et al. (1984) calculated the specific surface area of aluminum hydroxide gels at
pH 6.5 and 7.5 at different [OH]/[Al] ratios r (0, 2, 2.5) by using the relationship of the
molecular area of long chain alkylsulfonates, the Avogadro number (NA) and the adsorbed
quantity of alkylsulfonates (Qmax) taken at sorption plateaus from adsorption isotherm data.
Resulting surface areas ranged between 540 and 1210 m2/g, and decreased with increasing pH
- 111 -
Arsenate adsorption on Al-based sorbents – EXAFS

and r. The most comparable value to the present study is 540 m2/g for pH 7.5 and and r value
of 2.5. The same approach was used in this study to calculate the surface area with adsorption
isotherm data of the arsenate ion. With an As-O distance of 1.69 Å the molecular surface area
of the AsO4- tetrahedron (SAAsO4) was calculated to be 9.9 Å2, about 2  smaller than for
aliphatic chains. The specific surface area (SSA) of polyaluminum chloride under the specific
experimental conditions was estimated after Rakotonarivo et al. (1984) by the relation:

SSA = Qmax AAsO4 NA (6.1)

With Qmax from the measured adsorption data (0.0154 mol As/g Altot), calculation of the SSA
at pH 7.5 (pH of the adsorption isotherm data) yielded 894 m2/g. This value for the specific
surface area is in the same range but 1.6  larger compared to the value found by Rakotonarivo
et al. (1984).

6.3.2. As(V) adsorption isotherms

The measured adsorbed As concentrations were calculated for the four discussed surface areas
(Table 6.1). The Langmuir Equation was used to describe the data because the data points were
observed to level off, and it was concluded that a sorption maximum had been reached. The
goodness of fit was indicated by the nonlinear correlation coefficient R2 and was 0.97 for the
PAClnano data. The maximum adsorbed As(V) quantity ranged from 2.25
mol/m2 when
assuming single Al30 clusters to 171
mol/m2 when related to the measured BET surface area
of 13.7 m2/g (Figure 6.2, dashed lines for lower and upper boundary). The last value was just
above to the absolute theoretical maximum of arsenic coverage of 160
mol/m2, assuming a
maximum of 100 As atoms per nm2. This high value is another indication that the surface area
measured on the dried precipitate is not representative for the aqueous Al nanoclusters. The
saturation capacity per area using 540 m2/g for the surface of PAClnano was reached at 10.9

mol/m2, and was about 1.8  higher compared to granulate from the same polyaluminum
material (Figure 6.2), and 4  higher compared to gibbsite reported by Arai et al. (2005). This
was also reflected in the As surface coverage (Table 6.1). Interestingly, by calculating the
arsenate surface coverage for PAClnano with the surface area of 894 m2/g, very similar results
(5.92
mol/m2) were achieved compared to arsenate coverage on PAG (6.01
mol/m2) (Table
6.1). Hence, the uncertainty of specific surface determination complicates an accurate
characterization of surface reactivity of the solid supports.

- 112 -
 Chapter 6

Figure 6.2: Adsorption of As(V) per surface PAClnano and PAG at initial pH 7.5 using the best
proposed surface area after Rakotonarivo et al. (1984). Symbols represent experimental data and solid
lines represent Langmuir models for PAClnano (R2 = 0.97) and for PAG (R2 = 0.94). The two dashed
lines represent the upper (factor 15.7 higher) and lower (factor 4.7 lower) boundary for PAClnano data
using the surface area from BET analysis (13.5 m2/g) and geometrical calculations (2550 m2/g),
respectively.

Table 6.1: As(V) adsorption efficiency of the Al-based sorbents used in this work compared to
previously reported data
Al sorbent Size Surface area As(V) surface coverage

m m2/g
mol/g
mol/m2 As/nm2
PAClnano n.a. 13.7a) 2360 172 104
5405 5870 10.9 6.55
894b) 5290 5.92 3.56
2 10-3 2550c) 5750 2.25 1.36
PAG 1000-3000 35.0 ± 1.0 210 6.01 3.62
Gibbsite1 < 500 45.0 33.1 0.73 0.44
Gibbsite2 n.a. 13.5 33.1 2.45 1.50
Amorphous Al(OH)31 < 500 42.9 38.0 0.87 0.53
Allophane3 3 10-3 – 6 10-3 369.7 ± 1.5 352 0.95 0.57
Activated Alumina4 105 115 133 1.15 0.69
n.a.: not available, a) measured on dried PAClnano powder by BET, b) calculated after equation 10 from
Rakotonarivo et al. (1984) with Q0 from As(V) adsorption isotherms for the added amount Altot (in g), c) calculated
for Al30 entities of the dimension 1 nm x 2 nm, 1 Banning et al. (2005), 2 Ladeira et al. (2001), 3Arai et al. (2005), 4
Lin and Wu (2001), 5 Rakotonarivo et al. (1984)

- 113 -
Arsenate adsorption on Al-based sorbents – EXAFS

6.3.3. Surface coverage for EXAFS samples

The adsorption of As(V) to Al nanoclusters, whether as nanoclusters in suspension or as main

components of polyaluminum chloride and of a polyaluminum granulate, was compared to
As(V) sorption to gibbsite, which is well described in the literature (Ladeira et al., 2001,
Banning et al., 2005). The surface area of 540 m2/g after Rakotonarivo et al. (1984) was
employed to calculate surface loading for the polyaluminum chloride and the Al30 sample
(Table 6.2), because organic substances are believed to have a higher sorption affinity for
metal oxide surfaces (Bauer and Blodau, 2006). The loaded PAClnano samples used in EXAFS
analyses contained 0.12
mol/m2 and 0.61
mol/m2 (Table 6.2). These results accounted for
2% and 10% of the maximum sorption capacity listed in Table 6.1, and were named PAClAs2
and PAClAs10, respectively. The Al30 sample, which was exposed to the same initial As(V)
concentration than the sample PAClAs10 (0.66 mM) showed a similar surface coverage with
0.57
mol/m2 (Table 6.2), indicating that pure Al30 solutions are not better regarding the
uptake of arsenic than polyaluminum chloride solutions with a high amount of Al30. PAG
surface coverage was slightly higher than 0.73
mol/m2. The surface area from Banning et al.
(2005) was used to calculate As surface coverage on gibbsite (Table 6.1). Initial As(V)
concentrations of 0.133 mmol/L and 0.667 mmol/L resulted in 1.4 and 3.8
mol/m2. In
comparison to published values for the surface coverage (Banning et al., 2005, Ladeira et al.,
2001) and because As(V) concentrations in solution were decreased by only 50 %, the value of
3.8
mol/m2 was assumed to be close to the maximum uptake capacity (Table 6.2). Hence, the
Al nanocluster derived materials exhibit a higher uptake capacity than gibbsite.

Table 6.2: Sample codes and summary of As(V) adsorption data for EXAFS samples
Sample Al sorbent [As]initial As adsorption Surface coverage
code1) material
mmol/L %
mol/g
mol/m2 %
Al30As10 Al30 0.667 93.2 311.2 0.57 ~ 10
PAClAs22) PAClnano 0.133 97.6 64.9 0.12 2.0
PAClAs102) PAClnano 0.667 99.9 329.5 0.61 10.7
GibAs403) Gibbsite 0.133 99.3 64.0 1.42 36.8
GibAs1003) Gibbsite 0.667 53.9 182.0 3.85 ~100
PAGAs20 Polyaluminu 0.267 95.6 25.5 0.73 20.8
m granulate
1)
The number behind As indicates the As surface coverage in %
2)
“PACl” in the sample code is the abbreviation for the sorbent material polyaluminum chloride with high
amount of Al nanoclusters
3)
“Gib” was used as abbreviation for the sorbent material gibbsite

- 114 -
 Chapter 6

6.3.4. X-ray absorption spectroscopy

EXAFS spectroscopy was used to discern the atomic environment of As(V) adsorbed to the
different samples. For all samples, the energy height and position in the XANES range of the
As K-edge XAS spectra at 11878 eV corresponded to the sharp edge of the As(V) oxidation
state represented by the As(V) reference (data not shown).
The EXAFS oscillations of gibbsite samples showed the same frequency for two different
adsorbed As concentrations (Figure 6.3a). In general, all materials based on Al nanoclusters
showed the same frequency than the gibbsite samples, indicating similar atomic structures
around As. However, PACl samples showed a slight difference in magnitude for two different
As concentrations (Figure 6.3a). The Fourier transformation of the EXAFS signal yielded
information on the interatomic distances, and showed 3-4 peaks for all prepared samples
(Figure 6.3b). The peak at 1.3 Å was attributed to the As-O shell, and the 2-3 small peaks at
1.9-3 Å corresponded to the As-Al shell and multiple scattering of the photoelectron between
the oxygen atoms surrounding the arsenic atom in tetrahedron. For the gibbsite sample, the first
and second shells were isolated, back-transformed, and fitted in order to obtain the optimized
fit parameters of atomic distance and structural disorder for each shell. The first coordination
shell was simulated with one oxygen shell with a coordination number of 4 oxygen atoms
surrounding arsenic. The fitting yielded an atomic distance of 1.69 Å and a Debye-Waller
factor of 0.002-0.003 Å2 for low and high arsenic concentrations (Table 6.3), confirming the
tetrahedral structure of As(V). The second neighboring shell was fitted by comparing the
backscattering amplitude for As(V)-Al, As-O-O, As-O-As-O, and for a combination of all
paths. It was observed that the multiple scattering paths As-O-O- and As-O-As-O were needed
for a good description of the spectra. To verify the influence of the Al path on the fit, the two
shells of the gibbsite spectrum were fitted by a model with the As-O and the multiple scattering
paths, with and without including the Al path. The two model outputs were subtracted from
each other and the difference was compared to the noise level of the spectrum. The noise level
was calculated using the back-Fourier transformed signal for all distances higher than r = 6 Å.
The obtained residual of the As-Al fit is higher than the noise (Figure 6.4), and therefore a
substantial part of the fitting. The model results showed that two Al atoms surround the As
atom at distance R = 3.18 Å (Table 6.2). These findings correspond to the results of Ladeira et
al. (2001).

- 115 -
Arsenate adsorption on Al-based sorbents – EXAFS

Figure 6.3: (a) k2- weighted normalized
functions, and (b) Fourier transformed arsenic K-edge

functions for As(V) adsorbed on gibbsite and Al nanocluster –derived material.

The verified gibbsite model was applied to the back-transformed data of the Al30-based
sorbents, with slight modifications to achieve the best fitting parameters. No major differences
were found for coordination numbers and atomic distances. However, the Debye-Waller
factors for the multiple scattering paths varied between 0.001 and 0.009, inferring differences
in the structural disorder of the multi-scattering pathways of the photoelectron (Figure 6.5,
Table 6.3). These results suggest that the atomic environment of As(V) adsorbed to Al
nanocluster- based removal materials is very similar to the structure of As(V) adsorbed to
gibbsite, and corresponds to an innersphere bidentate-binuclear configuration.

- 116 -
 Chapter 6

Figure 6.4: Back-transformed EXAFS spectrum of the Al distracted gibbsite model and the noise level.
The difference in models with and without the Al shell is clearly above the noise signal.

Applying this binding geometry with the corresponding distances to the chemical structure of
Al13 and Al30, it is apparent that As can only be adsorbed to the nearest-neighbored terminal
water ligands of neighboring Al octahedrons (Figure 6.6), and that this coordination geometry
corresponds to the same as already postulated for gibbsite. On single Al13 and Al30 clusters, 6
and 8 of these preferred sites are available, respectively (compare Figures 6.1b and 6.1c).
Hence, the theoretical adsorption ratio of As to Al is 0.46 and 0.27 for Al13 and Al30,
respectively. Experimentally, a ratio of 0.2 M As(V)/ M Al using PAClnano was confirmed by
Mertens et al. (2011).
To compare these numbers to gibbsite, we looked at an arrangement of stacked gibbsite layers
in a cube with the dimensions of 100 nm
100 nm
100 nm. By employing the density of
gibbsite (2.33 g/cm3), the Avogadro number, and the molecular weight of Al(OH)3 (78 g/mol),
we calculated that 1.8
107 Al atoms are arranged in 106 nm3 of gibbsite (1.8 Al in 1 nm3),
which corresponds to a calculated specific surface area of 25.7 m2/g. Based on structural
considerations derived from Figures 6.1 and 6.6, we calculated 3.43
104 adsorption sites for
As for each cube face (104 nm2). Assuming that the four lateral faces of the cube are actively
involved, while the two basal faces are not suited for adsorption of As, one gibbsite cube with

- 117 -
Arsenate adsorption on Al-based sorbents – EXAFS

Figure 6.5: Fitting of the 1st and 2nd shell on the back-transformed EXAFS data at kw = 3 for As(V)
adsorbed on gibbsite, Al30 nanoparticles, polyaluminum chloride with high Al nanoparticles share, and
polyaluminum granulate. The colored solid lines represent the data and the dashed lines represent the
fit.

106 nm3 bears 1.37
105 adsorption sites for As. Thus the theoretical maximum value of the
ratio As/Al = 0.0076 (1.37
105 As adsorption sites / 1.8
107 Al atoms).
This is in the range of the experimental data for gibbsite removal (Table 6.2), where 0.359
mmol/L As(V) (53.9% of 0.667 mmol/L) were adsorbed to 75 mmol/L Al(tot), resulting in an
experimental ratio of 0.0048 As/ Al. The theoretical maximum As-to-Al ratio is dependent on
the specific surface area of gibbsite. The published values of the specific surface areas of
gibbsite 13.5 m2/g (Ladeira et al., 2001), and 45 m2/g (Banning et al., 2005) yield As/Al ratios
of 0.0040 and 0.0134, and particle sizes of 190 nm and 57 nm, respectively. The gibbsite with
the lower specific surface area (13.5 m2/g from Ladeira et al., 2001) would contradict our
experimental ratio of 0.0048 As/Al. Thus, we conclude that our initial assumption is sound: the
gibbsite particles in our experiment were typically 100 nm in diameter exhibiting a specific
surface area of 25 m2/g.

Table 6.3: Summary of the EXAFS modeling results

Sample1) Atomic shell N±20% R (Å)* E0
2 (Å2)*
(eV)

- 118 -
 Chapter 6

Al30As10 As-O 3.8 1.69 ± 0.01 5.5 0.0021 ± 0.0004

MS
As-O-O 12 3.11 ± 0.009 5.5 0.009 ± 0.0001
As-O-As-OMS 12 3.49 ± 0.1 6.4 0.0017 ± 0.0018
As-Al 1.8* 3.20 ± 0.05 5.0 0.0069 ± 0.0027
PAClAs2 As-O 4 1.68 ± 0.01 6.0 0.0025 ± 0.0004
As-O-OMS 12 3.20 ± 0.1 4.0 0.0089± 0.00003
MS
As-O-As-O 12 3.49± 0.1 5.0 0.0089± 0.00003
As-Al 2* 3.18 ± 0.04 6.0 0.009 ± 0.00003
PAClAs10 As-O 4 1.69 ± 0.01 6.6 0.0021± 0.0004
As-O-OMS 12 3.15 ± 0.05 4.5 0.004± 0.001
MS
As-O-As-O 12 3.49 ± 0.1 4.2 0.001± 0.0003
As-Al 2* 3.21 ± 0.06 7.0 0.0086 ± 0.0051
GibAs40 As-O 4 1.69 ± 0.02 6.0 0.0029 ± 0.0008
MS
As-O-O 12 3.20 ± 0.1 5.0 0.0069± 0.0018
MS
As-O-As-O 12 3.49 ± 0.1 5.5 0.001± 0.0005
As-Al 1.8* 3.17 ± 0.02 4.2 0.008 ± 0.006
GibAs100 As-O 4 1.69 ± 0.02 6.0 0.0028 ± 0.0004
MS
As-O-O 12 3.23 ± 0.13 6.4 0.009± 0.00013
As-O-As-OMS 12 3.49 ± 0.1 7.5 0.009± 0.00013
As-Al 1.9* 3.19 ± 0.04 5.0 0.0091 ± 0.0037
PAGAs20 As-O 4 1.68 ± 0.005 6.2 0.0017 ± 0.0009
MS
As-O-O 12 3.20 ± 0.1 4.0 0.0089± 0.00019
As-O-As-OMS 12 3.49 ± 0.1 4.0 0.009± 0.00016
As-Al 2.3* 3.18 ± 0.03 6.0 0.0076 ± 0.0058
1)
for sample codes see Table 6.2.
N: coordination number, R: interatomic distance

2: Debye-Waller factor (disorder parameter)
E0 : threshold energy
As-O-OMS and As-O-As-OMS: Multiple scattering paths
* allowed to vary during the fit

- 119 -
Arsenate adsorption on Al-based sorbents – EXAFS

Figure 6.6: Structural diagram of polyaluminum oxides (adapted from Casey et al., 2005) and the
binuclear linkage between the arsenate tetrahedron and two Al(O)6 octahedron with an inter-atomic Al-
As distance of 3.20 Å.

6.4. Conclusions

The EXAFS analysis of the atomic environment of As(V) showed that arsenate was in a
bidentate-binuclear configuration when adsorbed to all the investigated Al sorbents gibbsite,
Al30 nanoparticles, polyaluminum chloride, or solid-phase polyaluminum granulate. This direct
chemical bond implies a robust sorption mechanism, by which As(V) is not released readily
from Al surfaces. The As(V) adsorption isotherms per unit area implied that Al nanoclusters
lost half of the sorption sites in a polyaluminum granulate compared to the Al30 coagulant in
suspension. The reactivity of a pure Al30 solution and Al30-rich polyaluminum chloride was
found to be very similar. Therefore, the simple application of PACl with high share of Al30
particles is a better option than producing pure Al30 suspensions. In general, Al nanoclusters
exhibit 100 sorption sites per mol Al more than gibbsite.

- 120 -
Chapter 7

Discussion and conclusions

- 121 -
Discussion and conclusion

7.1 Introduction
Arsenic contamination of groundwater poses a health risk to millions of people worldwide.
Due to the severe and widespread problem, many treatment options and materials have been
developed in the past decades. Still, some deficits remain in conventional treatment materials
regarding pH dependance and competition with other ions in the water. Recent research in this
sector focused on the improvement of purification processes and material efficiency and the
decrease of costs. Nanoparticles have emerged as a promising material with high removal
potential. In order to be applied at a large scale, new technologies need to be implemented to
conventional treatment processes.
The aim of this thesis was the evaluation of aluminum nanoclusters as removal agents for
arsenic. The Al nanoclusters Al30 and Al13 have previously been shown to be very effective in
water treatment due to their high coagulation capability, and Al30 was more effective as
coagulant compared to Al13 and AlCl3 due to the strongest floc formation (Chen et al., 2006).
Al nanoclusters were applied in a coagulation-coprecipitation system as major components of
polyaluminum chloride (Chapter 4), and in an adsorption–based treatment process as
constituents of polyaluminum granulate (Chapter 5). The coverage and interatomic
environment of arsenic on the Al surfaces of the two materials polyaluminum chloride and
polyaluminum granulate was compared to each other and to gibbsite (Chapter 6). The success
of groundwater treatment is highly dependent on the speciation and concentration of arsenic
and the groundwater chemistry. Therefore, arsenic species and content has been analyzed and
linked to the geochemistry of aquifers in the East Pannonian Basin (Chapter 3).

7.2 Arsenic contamination and water chemistry in the East

Pannonian Basin
The arsenic contamination case of the East Pannoninan Basin was the background for the
application of water treatment with Al nanoclusters. As mentioned in Chapter 1 the water
chemistry is an important factor for the choice of an arsenic treatment method. Groundwater
from various wells with depths between 2 and 3000 m has been analyzed in the Hungarian-
Romanian border area, and four different groundwater types were identified: methanogenic
groundwater of the Na-HCO3 type, sulfate-reducing groundwater with Mg/Ca-HCO3 to Na-
HCO3 composition, water from dug wells of the Mg/Ca-HCO3 type, and thermal-saline waters

- 122 -
 Chapter 7

with Na-Cl, Na-HCO3 and Mg/Ca-HCO3 water compositions. The first group was identified to
consistently contain the highest As concentration in the range from 23.4 to 208
g/L.
Therefore, this water composition was chosen as a reference for the preparation of synthetic
groundwater for arsenic-removal experiments.
Isotope analysis showed that arsenic-rich waters dated back from the last ice-age. From South
and Southeast Asia shallow aquifers (<150 m) were reported to bear high arsenic
concentrations in relatively young (Holocene) sediments. Despite the age difference,
characteristics of the source sediments are very similar in both cases, with layers of alternating
mud, silt and sand, some of them with high organic content. Evidence of iron minerals present
in the sediments of the Pannonian Basin as reported by Varsányi and Kovács (2006) and
Viszián (2002) suggest that As mobilization in this area is controlled by redox conditions.
Total carbon analysis revealed that both the type and concentation of TOC control reducing
conditions, and consequently the dissolution of iron minerals by microbial activity. The low As
concentration found in sulfate-rich water lead to the conclusion that H2S resulting from the
reduction of SO42- removes As from solution, either by forming As sulfides or adsorbing As
onto Fe sulfides.

7.3 Efficiency of Al nanoclusters for the removal of arsenic

The removal of As(III) and As(V) from water by Al nanoclusters was the focus of this thesis.
Two filter materials with Al nanocluster components were used: polyaluminum chloride
(PAClAl30) and polyaluminum granulate (PAG). The latter material was produced from highly
concentrated polyaluminum solutions by repetitive cycles of washing and drying. In the section
to follow, the two different treatment materials are compared and recommendations for arsenic
removal for both materials are discussed separately.

7.3.1 Arsenic removal using PAClAl30

The success of the coagulation and precipitation processes highly depend on the degree of
coagulation and the separation of the solid from the liquid phase. The deprotonation and
consequently the aggregation of aluminum nanoclusters is highly pH dependent. We confirmed
that the deprotonation of a polyaluminum chloride solution followed the deprotonation of Al30
and Al13, and reaches the point of zero charge at the same pH (6.7) as reported by Casey et al.
(2001) for Al30 and by Furrer et al. (1996) for Al13 solutions (Chapter 4). The optimum pH 7-8
for arsenic removal coincided with the pH of strongest coagulation of the Al nanoclusters. This

- 123 -
Discussion and conclusion

pH range corresponds to most groundwater pH, and therefore groundwater treatment can be
conducted without pH adjustment in most cases.
The application of PAClAl30 in As-rich sodium bicarbonate water showed that at least 0.4
mmol/L Altot were needed to achieve optimum As removal. The application of more than 6
mmol/L Altot decreased the pH to below the pHPZC. Consequently, Al particles protonate and
inihibit the formation of large aggregates, resulting in less efficient removal of Al and As(V)
due to settling and filtration (0.45
m). Therefore, to ensure the most efficient removal of As,
the concentration of PAClAl30 needs to be in the range of 0.4 to 6 mmol/L Altot. Using these Al
concentrations, all initial As(V) concentrations (20-2300
g/L) could be reduced to the desired
target value of 10
g/L. High As concentrations (> 1000
g/L) in the water require higher Al
concentrations (e.g. 81 mg/L Altot are needed for 2300
g/L As(V)).
The emission of nano-sized Al particles into the environment should be avoided as Al has been
suspected as a possible cause of Alzheimer’s disease (WHO, 1998). Small filters, reverse
osmosis, or a coagulation aid like silica for a better phase separation, as well as strict controls
of Al release have to be incorporated in such a treatment system.

7.3.2 Arsenic removal using polyaluminum granulate (PAG)

Polyaluminum granulate was characterized as meso- to macroporous amorphous Al phase with

the same XRD characteristics as its base material, the polyaluminum powder Locron. In batch
tests it was observed to be highly acidic. Once added to naturally buffered water, the pH
dropped by about one pH unit. The removal of initial As concentrations in the range of 20
g/L
to 300 mg/L showed that the maximum level for drinking water could only be reached for
initial As concentrations of up to 100
g/L for As(V) and up to 20
g/L for As(III), even if the
removal percentage is > 97 % for As(V). PAG is proposed to be efficient for reducing small
As(V) concentrations to below the drinking water limit, and being able to substantially reduce
high initial As contaminations of up to 100 mg/L by > 90% in well water. Above 100 mg/L,
only a 50-70 % As(V) reduction was reached in synthetic water.
Investigation of thin sections of As-loaded aluminum granulate by scanning electron
microscopy give indications for the optimum geometry of adsorption material. It was observed
that up to 80
m of the grain’s radius are used for the adsorption of As(V). This range depends
on the experimental conditions such as arsenic concentration and exposure time. For the
experimental setup used in Chapter 5, this rim corresponds to about 38% of the total grain
volume under the assumption that the granulates have an ellipsoidal form with equatorial radii
of 2 mm and 1 mm, and a polar radius of 1 mm. The adsorption efficiency of As(V) on PAG

- 124 -
 Chapter 7

could be improved by a factor 2.5 by producing spherical grains with a diameter of 0.3 mm,
where approximately 95 % of the material volume could be used for As(V) sorption. Hence,
the uptake capacity of PAG has the potential to reach 475
mol As(V)/g, which is in the range
of nano-TiO2 applied as powder or fixed on a support (470-500
mol/g; Nilchi et al., 2010;
Pena et al., 2005).

7.3.3 Comparison of PAClAl30 and PAG

The arsenic removal processes of the two filter media are very different. The nanoclusters Al30
and Al13 act as strong coagulants in polyaluminum chloride solutions, while the granulate
material is a solid-state adsorbent. The uptake efficiency of the two materials was compared in
Chapter 2 in terms of percent efficiency (Table 2.1) and in Chapter 6 in terms of surface
coverage (Table 6.1) in water with the composition found in East Pannonian groundwater with
high As concentration. Al nanoclusters in the coagulant performed slightly better in As(V)
elimination than Al granulates. In general, As(III) removal was less efficient than As(V)
removal for both materials. This was to be expected because the uncharged arsenic species is
known to have low affinity for aluminum-based removal agents. However, with PAG almost
20% more As(III) was removed from solution. This can be explained by the extremely high Al
concentration in the granulate material (1 M Altot), and the fact that As(III) filled the entire
pore space of the granulate, where it is better trapped and adsorbed than in the polyaluminum
solutions of lower Al concentration (up to 10 mM). In pH titration experiments (Chapter 4) it
was shown that 80 % of As(III) could be removed in Al solutions of 15 mM after
centrifugation. This result also indicated that the amount of As(III) could be substantially
reduced with the coagulation process by applying centrifugation to separate the flocs from
solution and/or using high Al concentrations. However, the goal of an efficient treatment
method is not only the removal of a contaminant but also the limitation of waste and the cost-
effectiveness. The use of high amounts of a coagulant or the implementation of centrifugation
for large water volumes does not meet these aspects. Therefore, As(III) needs to be oxidized to
As(V) prior to the removal process.
With regard to the employed Al concentration, the coagulation process was more efficient than
the adsorption to polyaluminum granulate. This was also reflected in the comparison of
arsenate surface coverage and adsorption isotherms. With polyaluminum chloride about 5870

mol/g Al nanocluster or 10.9
mol/m2 was achieved; with polyaluminum granulate it was 198

mol/g or 5.7
mol/m2. Hence, Al nanoclusters lose some of their surface sites in the dense
granulate material.

- 125 -
Discussion and conclusion

7.3.4 Molecular structure of arsenic during water treatment with Al nanoclusters

To predict the fate of arsenic in a water-treatment process and the efficiency and cost-
effectiveness of water-purification systems it is essential to understand how arsenic interacts
with the adsorbent surfaces. The determination of the arsenate sorption mechanism with
EXAFS (Chapter 6) showed the formation of inner-sphere bidentate binuclear complexes on
the surfaces of Al30, PAClAl30 and polyaluminum granulate at pH 7.5. These results agree with
findings for gibbsite, both in this study and by Ladeira et al. (2001), as well as with the
sorption of arsenate to other Al surfaces, such as Al2O3 (Arai et al., 2001). This direct chemical
bond implies a robust sorption mechanism (Catalano et al., 2008) and As(V) is not released
readily from Al surfaces, which is confirmed by the desorption experiments in Chapter 5.
While the interatomic structure of As(V)-Al complexes could be well defined, the interaction
of As(III) with Al nanoclusters was not investigated on a molecular scale because As(III) in the
EXAFS samples oxidized within the first minutes under the high-energy beam. The high
energy was specific to the SAMBA beamline and is very useful to obtain highly resolved
structural information.

7.4 Recommendations for further research

In this work, one product of polyaluminum chloride was tested, which was confirmed to
develop a high amount of Al30 species once the powder is dissolved in water, and it was shown
that solutions remain stable for weeks. The distribution of Al species in many commercially
available polyaluminum chlorides is often not known, even when applied to large-scale water
treatment systems. In order to understand whether the high amount of Al30 is making a
difference in the efficiency of As removal with polyaluminum chloride, a systematic
comparison study needs to be carried out where the efficiency of arsenic removal is
investigated together with the determination of Al species in different polyaluminum chloride
solutions.
As mentioned in the introduction, the presence of other ions in the contaminated water can
inhibit the removal of arsenic. Within this PhD project, preliminary batch experiments were
conducted to test the interference with phosphate on As(V) and As(III) removal by PAClAl30
(Appendix 3). Removal tests with natural and synthetic well water showed that phosphate
concentrations in Romanian groundwater had no effect on arsenic removal with PAClAl30.
These preliminary results either suggest that an excess of sorption sites is available for arsenic
and phosphate, or that As(III) is removed partly by precipitation and not by adsorption.
- 126 -
 Chapter 7

However, pilot-scale studies with polyaluminum granulate showed adsorption of phosphate

and impedence of As(III) removal (Appendix 2). Further experiments should be conducted to
accurately determine the interaction of phosphate and arsenic with the Al nanocluster surfaces.
X-ray absorption fine structure (EXAFS) spectroscopy is a useful tool to determine the
molecular structure of complexes and to better understand the sorption mechanisms of a
contaminant during water treatment. However, limitations arise when using high-energy
beamlines on oxidation-sensitive atoms, such as As(III). Future XAS investigations of As(III)
on Al surfaces should consider the use of beamlines with lower energy. Another possibility to
investigate As(V) and As(III) complexes on Al in solution is Fourier-transformed Infrared
(FTIR) spectroscopy. In this work, EXAFS samples have been produced only in deionized
water to understand the general binding mechanism. Adsorption and desorption experiments
have shown that the As adsorption capacity in Na-HCO3 water is lower and that As(III) and
As(V) are easier desorbed than in deionized water. This effect is caused by competition with
other ions in solution. X-ray absorption and electron microscopy techniques have provided new
insight on transport processes during this work, but so far these techniques could only be
applied to As sorption in deionized water. As a next step, As sorption to solid-state adsorbents
such as PAG should be investigated in naturally buffered water with SEM and
XRF, and
questions regarding the widening of the As(V) rim in an equilibrium state and the interference
with calcium should be adressed with this highly resolved microscopy techniques.
Field-scale applications were started within this project, both with polyaluminum chloride and
with the polyaluminum granulate. For waters with As(V) as the dominant arsenic species (as it
was the case for Greece), field tests have been shown to be very effective. For highest
efficiency, PAG and PAClAl30 should mainly be applied to As(V)-contaminated waters, and
more pilot tests should be carried out to better meet natural conditions.
To address the feasability of a water-treatment method, the amount of waste and re-usability of
the used products have to be evaluated and further studies have to be conducted in this
direction.

- 127 -
Appendix

Appendix 1: Column Study - Adsorption of arsenate on

polyaluminum granulate

Appendix 2: Case study - Arsenic removal from groundwater

using polyaluminum granulate in a pilot plant, Sepreus, Romania

Appendix 3: Additional data

- 128 -
 Appendix 1

Appendix 1
Column Study: Adsorption of arsenate on polyaluminum
granulate

As preparation for the installation of the pilot treatment reactor in the field (Appendix 2), two
column experiments with different arsenic concentrations and fillings were conducted in the
ETH laboratory in summer 2009 (start: 30th September, 2009, 12 am; end: 19th August, 2009,
11 am). In the following, the experimental setup and the resulting data are presented.

Experimental setup

Table A1.1: Experimental conditions of column experiments (* Polyaluminum granulate)

Parameters Unit Column A Column B
Column
Total filled Height cm 12.3 12
Diameter cm 3.3 3.3
Radius cm 1.65 1.65
Water flow ml/min 50 50
L/d 72 72
m3/d 0.072 0.072
Time needed to pass
column min 1.45 1.28
m/min 0.1
m/d 144
Fill material
1) PAG*-marble mixture
Height cm 8.5 7.5
Area cm2 8.6 8.6
Volume cm3 = ml 72.7 64.1
Al:Marble wt:wt 1:1 9:1
Weight PAG* g 40 54 Figure A 1.1: Setup
Weight marble g 40 6 and filling of
grain size of marble + PAG* mm 1-2 1-2 columns
Density PAG* g/cm3 1.93 1.93
Bed volume cm3 41.5 31.1
Pore volume cm3 29.1 25.7
2) Quartz sand
Height bottom cm 2.5 2.5
Height top cm 1.3 2
Area cm2 8.6 8.6
Volume cm3 = ml 11.1 17.1

As(V) concentration
g/L 400 80

- 129 -
Column study

Results

Figure A 1.2: (a) As(V) concentration and (b) Al concentration and pH in the outflow of column A
(([As]initial = 400
g/L; As-to-marble ratio = 1:1).

- 130 -
 Appendix 1

Figure A 1.3: (a) As(V) concentration and (b) Al concentration and pH in the outflow of column B
([As]initial = 80
g/L; As-to-marble ratio = 1:1).

Conclusions

Under employed fast flow conditions As breakthrough occurs in column A after 1700 bed
volumes (1 day). In column B with higher PAG content the arsenic concentration in the
outflow reaches >10 ppb after 33000 BV (13 days). In both cases As breakthrough follows pH
increase. With the presented setup arsenate concentrations of 80
g/L can be treated in water
with circum-neutral pH.

- 131 -
Case study

Appendix 2
Case study: Arsenic removal from groundwater using
polyaluminum granulate in a pilot plant, Sepreus, Romania

Background

Polyaluminum granulate (PAG) was used successfully for arsenic removal in batch tests
(Chapter 5) and column tests with 80
g/L As concentration (Appendix 1). In this project,
PAG was applied in a pilot plant in collaboration with the village Sepreus, in Southwest
Romania. The well that was chosen for the pilot plant installation was at the town’s border, on
the land of a private farm. The artesian well was used for animals, washing of clothes and - to a
lesser extend - drinking and cooking in nearby households. The physical and chemical
parameters of the water (Table A 2.1) reflected the general water chemistry of the area. In the
following, the setup, results and problems of the treatment unit are documented.

Table A 2.2: Physical and chemical parameters of the groundwater in Sepreus.

Physical Parameter Unit Value
pH - 7.71
Temperature °C 18.7
Eh mV -97
Oxygen mg/L 1.5
Conductivity μS/cm 1617
Depth m 280 - 300
Chemical parameter Unit Value
Na+ mg/L 326
Ca+2 mg/L 30.9
Mg+2 mg/L 13.8
K+ mg/L 1.2
Si4+ mg/L 10.7
Fe2+ μg/L 460
As(tot) μg/L 80
As(III) μg/L 78
Cl- mg/L 160
SO42- mg/L 2.5
NO32- mg/L 0.13
TOC mg C/L 0.49
Alkalinity(-) mM 12.3

- 132 -
 Appendix 2

Filter material

PAG was produced from concentrated Al solutions (the production procedure is described in
the methods section of Chapter 4). In contrary to coagulation systems, granulates from the
same material have the advantage to minimize the possible risk of aluminium contamination
for drinking water during the treatment process.
PAG was mixed with marble of the same grain size (1-2 mm) in a weight-to-weight (wt:wt)
ratio of 9:1 to buffer the pH of the system (Figure A 2.1). The mixture was then packed in a
column of 19 cm height.

Pretreatment step: As(III) oxidation

Arsenic in the Sepreus well water is present in its trivalent oxidation state. Al granulate,
however, adsorbs preferably As(V). Therefore, As(III) needs to be oxidized to As(V) before
getting in contact with the sorbent material. Ozone was chosen as an oxidizer because of its
good oxidation performance and low energy costs in Romania. The ozone generator had to be
connected to an air pump. For the oxidation with ozone, it also had to be accounted for the
presence of total organic carbon (TOC) in the water (see Box A 2.1). The TOC concentration
of the Sepreus well water is 0.49 mg/L (Table A 2.1). Preliminary tests in the laboratory and in
the field showed that 90 – 100 % As(III) was oxidized after three minutes in contact with
ozone. The concentration of ozone needed to account for the arsenic and TOC concentration of
this particular well (Table A 2.1) was calculated to be 84.4
mol/L (Box A 2.1).
The inflowing gas resulted in gas bubbles in the reactor. Gas in the filter column may fill the
pore space of the adsorbent material. As a consequence, preferential flow may establish within
the media column and reduce sorption capacity. To avoid this phenomenon, a column with a
gas outlet was installed between the oxidation column and the filter column to allow the ozone
to leave the unit.

- 133 -
Case study

Box A 2.1: Implemented ozone concentrations in the field treatment unit in Sepreus.

How much ozone is needed?

Oxidation of As(III):
H3AsO3 + O3 => HAsO42- + O2 + 2H+

0.64
g O3/
g As(III) from Ghurye and Clifford (2001)

 0.013
mol O3 are needed for 1
g As(III)

 Amount needed for the oxidation of 80
g As(III)/L is 1.07
mol O3/L

Oxidation of TOC:
CH2O + 2O3 => H2CO3 + 2O2

 83.3
mol O3/L are needed for 40
mol Corg/L

Total O3 needed: 84.4
mol O3/L

The ozone generator Certizon C300 has the best capacity and produces 20
mol O3 in
excess:
300 mg O3/h = 104.2
mol O3/min = 1L O3/min

Reactor design and set-up

The complete treatment reactor configuration consisted of a peristaltic water pump, the
oxidation step (including the oxidation chamber, degassing chamber, the ozone generator and
the air pump), a sediment filter step, and the arsenic filter step with PAG. All this equipment
was fixed to a case, in which also electricity plugs were installed by connection to the public
electricity landlines (Figure A 2.1). Electricity was needed for the ozone generator, the air
pump and the peristaltic pump. The water was pumped from the artesian well to ensure
constant water pressure in the field. The water flow was adjusted to 1 L/min.

- 134 -
 Appendix 2

Figure A 2.1: Complete reactor setup at the Sepreus field site. 1: Peristaltic water pump, 2: Ozone
generator and air pump. Detailed explanation of the different columns is given in Figure A 2.2.

The first column was a mixing chamber for the water and the ozone. For this purpose, column
1 was empty, and water was flowing together with the gas through the column (Figure A 2.2).
The ozone was conducted into the reactor by a Y-connector. A one-way valve, specifically
produced for ozone-water applications, was inserted between the ozone generator and the
ozone inlet into the water flow, to protect the ozone generator from water inflow and
subsequent destruction. Column 2 served as a chamber to separate the gas from the water and
release gas from the system. The gas was collected on top and left the column through an
opening in the top while the water was dropping to the bottom of the column, from where it
was conducted to the following column by an inner tube (Figure A 2.2). The equilibrium
between gas inflow and water inflow is of crucial importance, and the flow of water and gas
had to be maintained constant throughout the operation of the treatment reactor. Column 3 was
installed in order to avoid the inflow of small sediment particles into the material column.
Sediment particles in the water could cause coating of the granulate surface, and inhibit the
sorption efficiency of the filter material. A conventional 5 μm fibre filter, purchased from
www.waterfilters.com, was used for this purpose. In Column 4, the mixture of aluminium
granulates and marble was used for the adsorption of arsenic. On the bottom and on the top of
the column, sand from the region (Bega sand) of 1.5 – 2 mm grain size was used as fill
material. A fine woven fabric kept sand and granulate from slowly being washed out with the
water. The entire bed volume was 3 L or 3 dm3.The water flow was directed from bottom to
top of the column (Figure A 2.3).

- 135 -
Case study

Figure A 2.2: Detailed sketch of complete column setup in the treatment reactor.

Figure A 2.3: Column packed with polyaluminum granulate

For a smooth running of the unit, a steel pipe needed to be implemented to protect the plastic
hose, by which the water was pumped from the well head into the treatment unit, from
violations. The entire setup was inserted in a metal case to protect it from weather conditions
and damage.

Results

The water flow in the unit was started at 1000 ml/min and slowly declined to 900 ml/min.
Inflow and outflow results are similar, but the outflow is slightly below inflow for most days
(Figure A 2.4). This fast flow was generated to set this experiment to the condition of
consumption habits. The flow was changed in the last two days of operation to test how a
slower flow rate affects the chemistry inside the unit.

- 136 -
 Appendix 2

The pH was stable around pH 7.7 ± 0.2, until the sediment filter was installed at 430 hours
after the start of the operation, after which pH was increased during the oxidation step. The
oxygen content of the inflowing water was generally low during the operation. As expected,
oxygen concentration was higher (up to 2.5 mg/L) in the oxidation chamber, but surprisingly
anoxic conditions were present within the PAG filter column between 100 and 350 h.
However, redox potential and conductivity for all samples taken after the different columns
were in the range of the inflowing water (Figures A 2.5 C and D).
After the implementation of the sediment filter, oxygen concentrations and consequently the
redox potential increased drastically. Conductivity and pH showed a slight increase after the
flow was reduced (Figure A 2.5).

Figure A 2.4: Flow rates over 21 days days total run time of the field reactor.

- 137 -
Case study

Figure A 2.5: (a) pH, (b) O2 , (c) redox potential, and (d) conductivity at valves A (influent), B (after
oxidation), C (effluent), and D (after filter) over 21 days total run-time of the field reactor.

- 138 -
 Appendix 2

Arsenic concentrations in the inflow showed that As(III) is the dominant As species (Figure A
2.6). After the oxidation step with ozone, most As(III) was oxidized to As(V), but 100%
oxidation only occurred after the sediment filter was implemented, due to higher oxygen
concentrations measured on the ozone chamber at the same times (Figure A 2.5(b) ). However,
high total As concentrations in the effluent imply that As removal in the last filtration step was
not effective. Only in the first sample, taken one hour after the start of the unit, total As content
in the effluent corresponded entirely to As(V), resulting in 20% As removal (Figure A 2.6). For
all samples taken 24-350 hours after the launch, As(V) was reduced inside the material
column, indicating anoxic conditions (Figure A 2.5(b)). With the installation of the sediment
filter, As in the effluent was at first (400 - 500 h) well oxidized, with a further decrease in
As(V) towards the end of the operation.
Other potential competitors for sorption sites such as iron (Fe) and phosphorus (P) were
measured in the influent, the effluent and after the oxidation step. Results showed that iron and
phosphorus were removed from the water by about 75% and 66%, respectively, in the first 100
hours (Figure A 2.7). After 100 h operation time, the concentrations of both elements increased
until they reached their initial inflow concentrations at around 200 h, indicating that all
sorption sites were occupied. Al concentrations were monitored to asses the health risk by such
a water treatment unit. All results plotted below the Al drinking water limit of 200
g/L
(WHO, 1998), and in most samples Al concentrations in the effluent did not exceed the inflow
values (Figure A 2.7).

- 139 -
Case study

Figure A 2.6: Arsenic total and species concentrations in the influent, after ozonation and in the
effluent of the field treatment unit.

- 140 -
 Appendix 2

Figure A 2.7: Concentrations of As(tot), Fe, P, and Al in the influent, after ozonation and in the
effluent of the field treatment unit.

- 141 -
Case study

Encountered problems

After the first 6 days high turbidity with mainly shale particles of <0.1 mm – 5 mm in size was
observed in the water, probably coming from the aquifer. It is presumed that these particles
clogged the filter material and pressure built up in the reactor setup, causing a slight decrease
in the water outflow and anoxic conditions in the column with the PAG. When the particles
were noticed after the half run time of the experiment, a sediment filter was implemented
between the degassing step and the filter column in order to prevent the Al granulates to be
coated by the clay material. However, this protection came too late, as no recovery of uptake
was induced.

Daily monitoring and adjustments was necessary to ensure the good performance of the ozone
inflow. At some days, the ozone inlet was found blocked by water, possibly due to over-
pressure in the unit. Another reason might have been the breakdown of the public electricity
supply during the operation, as the village had several power cuts lasting from 5 minutes to
several hours.
Arsenic removal was not as effective as expected. An important problem was the incomplete
oxidation of arsenic. The inflowing As(III) was successfully oxidized to As(V) within the
oxidation step chamber, but was again reduced in the column with the polyaluminum filter.
Secondly, P and Fe occupied sorption sites more readily than As, even if present as As(V), and
not enough sites were available for all three elements. In addition, anoxic conditions prevailing
in the column with the filter material caused fouling of the PAG, and iron-rich clay and/or
organic particles might have been the reason for clogging and back-flow within the unit.

Conclusions
The chosen setup of the arsenic treatment unit in this pilot study was not effective for arsenic
removal. Anoxic conditions developed in the reactor and caused As(V) reduction and fouling
of the polyaluminum granulate. Further optimization is needed to improve the mixing of the
gas and the water phase, and to ensure the exclusion of gas bubbles from the system. Pressure
meters should be installed after each column to have a better control on physical conditions
before receiving chemical data, and to react in time so that no backflow is induced. A sediment
filter should be installed before the oxidation step (and not afterwards), in order to retain iron-
or organic-rich clay particles from the aquifer and prevent clogging of the unit becomes less
probable.

- 142 -
 Appendix 3

Appendix 3
Additional data

The arsenic uptake in the presence of phosphorus was tested by batch experiments in 50 ml
tubes filled with 18M
water, shaken horizontally at 150 rpm for 30 minutes.

Figure A 3.1: As(V) removal with PAClnano in the presence of phosphate. The initial concentrations of
both, As(V) and phosphate are in the same range than the concentrations measured in Romanian
groundwaters

Figure A 3.2: As(III) removal with PAClnano in the presence of phosphate. The concentrations of both,
As(III) and phosphate are in the same range than the concentrations measured in Romanian
groundwaters

- 143 -
References

References

Agett J, O'Brian, G.A. (1985) Detailed model for the mobility of arsenic in lacustrine
sediments based on measurements in Lake Ohakuri. Environmental Science and
Technology 19:231-244
Aiuppa A, Avino, R., Brusca, L., Caliro, S., Chiodini, G., Alessandro, W.D., Favara, R.,
Federico, C., Ginevra, W., Inguaggiato, S., Longo, M., Pecoraino, G., Valenza, M.
(2006) Mineral control of arsenic content in thermal waters from volcano-hosted
hydrothermal systems: insights from island of Ischia and Phlegrean Fields (Campanian
Volcanic Province, Italy). Chemical Geology 229:313-330
Aiuppa A, D'Alessandro, W., Federico, C., Palumbo, B., Valenza, M. (2003) The aquatic
geochemistry of arsenic in volcanic groundwaters from southern Italy. Applied
Geochemistry 18 (9):1283-1296
Allouche L, Gérardin, C., Loiseau, T., Férey, G., Taulelle, F. (2000) Al30: A giant Aluminum
Polycation. Angew Chem Int Ed 39 (3):511-514
Amini M, Abbaspour, A.C., Berg, M., Winkel, L., Hug, S., Hoehn, E., Yang, H., Johnson, A.
(2008) Statistical modeling of global geogenic arsenic contamination in groundwater.
Environmental Science and Technology 42:3669-3675
Anderson MA, Ferguson, J.F., Gavis, J. (1976) Arsenate adsorption on amorphous aluminum
hydroxide. Journal of Colloid and Interface Science 54 (3):391-399
Angelone M, Cremisini, C., Piscopo, V., Proposito, M., Spaziani, F. (2009) Influence of
hydrostratigraphy and structural setting on the occurrence in groundwater of the
Cimino-Vico volcanic area (central Italy). Hydrogeology J 17:901-914
Antics M (2000) Computer simulation of geothermal reservoirs in the Pannonian Basin,
Eastern Europe. In: Proc World Geothermal Congress, 2000, Kyushu - Tohoku, Japan,
May 28-June, 2000
Antics M, Rosca, M. (2003) Geothermal development in Romania. Geothermics 32:361-370
Appelo CAJ, Postma, D. (2005) Geochemistry, groundwater and pollution. AA Balkema
Publishers
Appelo CAJ, Van der Weiden, M.J.J., Tournassat, C., Charlet, L. (2002) Surface complexation
of ferrous iron and carbonate on ferrihydrite and the mobilization of arsenic.
Environmental Science Technology 36:3096-3103
Arai Y, Elzinga, E.J., Sparks, D.L. (2001) X-Ray absorption spectroscopic investigation of
arsenite and arsenate adsorption at the aluminum oxide-water interface. Journal of
Colloid and Interface Science 235:80-88
Arai Y, Sparks, D.L., Davis, J.A. (2005) Arsenate adsorption mechanisms at the allophane-
water interface. Environ Sci Technol 39:2537-2544
Arnórsson S, Andrésdóttir, A. (1995) Processes controlling the distribution of boron and
chlorine in natural waters in Iceland. Geochimica et Cosmochimica Acta 59: 4125-4146
Auffan M, Rose, J., Bottero, J-Y., Lowry, G.V., Jolivet, J.-P., Wiesner, M.R. (2009) Towards a
definition of inorganic nanoparticles from an environmental, health and safety
perspective. Nature Nanotechnology 4:634-641
Auffan M, Rose, J., Proux, O., Borschneck, D., Masion, A., Chaurand, P., Hazemann, J.-L.,
Chaneac, C., Jolivet, Wiesner, M.R., Van Geen, A., J.-P, Bottero (2008) Enhanced
adsorption onto maghemites nanoparticles: As(III) as a probe of the surface structure
and heterogeneity. Langmuir 24:3215-3222
Baciu C, Cordos, E., Roman, C., Bodea, C., Costin, D., Senila, M. (2006) Caractere
Hidrogeochimica ale acviferelor arteziene din câmpia crisurilor. AHR Hydrogeologia 7
(1):3-8

- 144 -
 References

Badruzzaman M, Westerhoff, P., Knappe, D. R.U. (2004) Intraparticle diffusion and adsorption
of arsenate onto granular ferric hydroxide (GFH). Water Research 38:4002-4012
Ballantyne JM, Moore, J.N. (1988) Arsenic geochemistry in geothermal systems. Geochimica
et Cosmochimica Acta 42:475-483
Bang S, Johnson, M.D., Korfiatis, G.P., Meng, X. (2005) Chemical reactions between arsenic
and zero-valent iron in water. Water Research 39 (5):763-770
Banning B (2005) Arsenic speciation in As(III)- and As(V)-treated soil using XANES
spectroscopy. Microchimica Acta 151:181-188
Bauer, M., Blodau, C. (2006) Mobilization of arsenic by dissolved organic matter from iron
oxides, soils and sediments. Science of the Total Environment 354:179-190
Belin S, Briois, V, Traverse, A., Idir, M., Moreno, T., Ribbens, M. (2005) SAMBA a new
beamlin at SOLEIL for X-Ray absorption spectroscopy in the 4-40 keV energy range.
Physica Scripta T115:980-983
Bentley R, Chasteen, T.G. (2002) Arsenic curiosa and humanity. Chem Educator 7:51-60
Berg M, Stengel, C., (2008) ARS 25-28 Arsenic Reference Samples, Interlaboratory Quality
Evaluation (IQE). Report to participants, Eawag, Swiss Federal Institute of Aquatic
Science and Technology, March 2008, Dübendorf, Switzerland
Berg M, Tran, H.C., Nguyen, T.C., Pham, H.V., Schertenleib, R., Giger, W. (2001) Arsenic
contamination of groundwater and drinking water in Vietnam: a human health threat.
Environmental Science and Technology 35:2621-2626
Berg M, Trang, P.T.K., Stengel, C., Buschmann, J., Viet, P.H., Dan, N.V., Giger, W.,, Stuben
D (2008) Hydrological and sedimentary controls leading to arsenic contamination of
groundwater in the Hanoi area, Vietnam: the impact of ironarsenic ratios, peat, river
bank deposits and excessive groundwater abstraction. Chemical Geology 249:91-112
Bertram R, Geßner, W., Müller, D., Danner, M. (1994) Charakterization of Al(III) Species in
Basic Aluminium Chloride Flocculants by means of Ferron m,ethod and 27AlNuclear
Magnetic Resonance. Acta hydrochim hydrobiol 22 (6):265-269
Bertram R, Stieber, E., Geßner, W. (1996) Toxizität von Aluminium - Al Spezies in
protolysierten Aluminiumchloridlösungen. Z Umweltchem Ökotox 8 (2):78-82
Beveridge TJ (1988) Role of cellular design in bacterial metal accumulation and
mineralization. Ann Rev Microbiol 43:141-171
Borzsonyi M, Bereczky, A., Rudnai, P., Csanady, M., Horvath, A. (1992) Epidemiological
studies on human subjects to arsenic in drinking water in Southeast Hungary. Arch
Toxicol 66:77-78
Bose P, Sharma, A. (2002) Role of iron in controlling speciation and mobilization of arsenic in
subsurface environment. Water Research 36 (19):4916-4916
Bottero J-Y, Poirier, J.-E., Fiessinger, F. (1981) Partially neutralized aqueous aluminum
chloride solutions identification of aluminum species and relation between the
composition of the solutions and their efficiency as a coagulant. Water science and
technology 13 (1):601-612
Bottero J-Y, Rose, J., Wiesner, M.R. (2006) Nanotechnologies: Tools for sustainability in a
new wave of water treatment processes. Integr Environ Assess Manag 2 (4):391-395
Bottero J-Y, Bersillon, J.L. (1988) Aluminum and Iron(III) Chemistry - Some implications for
organic substance removal. Advances in Chemistry 219:425-442
Bowen GJ (2009) The Online Isotopes in Precipitation Calculator, Version 2.2(07/2008).
http://www.waterisotopesorg
Bratby J (2006) Coagulation and flocculation in water and waster water treatment. IWA
Publishing 2nd ed. (London):407 pp.
Buschmann J, Berg, M. (2009) Impact of sulfate reduction on the scale of arsenic
contamination in groundwater of the Mekong, Bengal and Red River deltas. Applied
Geochemistry 24:1278-1286
- 145 -
References

Buschmann J, Berg, M., Stengel, C., Sampson, M.L. (2007) Arsenic and manganese
contamination of drinking water resources in Cambodia: coincidence of risk areas with
low relief topography. Environ Sci Technol 41: 2146-2152
Buschmann J, Kappeler, A., Lindauer, U., Kistler, D., Berg, M., Sigg, L. (2006) Arsenite and
arsenate binding to dissolved humic acids: influence of pH, type of humic acid, and
aluminum. Environ Sci Technol 40:6015-6020
Casentini B, Nikolaidis, N.P., Hug, S.J. (2011) Arsenic accumulation in irrigated agricultural
soils in Northern Greece. Science of the Total Environment 409(22):4802-4810
Casey WH, Philips, B.L., Furrer, G. (2001) Aqueous aluminum polynuclear complexes and
nanoclusters: A review. Reviews in Mineralogy and Geochemistry 44:167-190
Casey WH, Rustad, J.J., Banerjee, D., Furrer, G. (2005) Large molecules as models for small
particles in aqueous geochemistry research. Journal of Nanoparticle research 7:377-387
Catalano JG, Park, C., Fenter, P., Zhang, Z. (2008) Simultaneous inner- and outer-sphere
arsenate adsorption on corundum and hematite. Geochimica et Cosmochimica Acta
72:1986-2004
Cavar S, Klapec, T., Grubeic, R.J., Valek, M. (2005) High exposure to arsenic from drinking
water at several localities in eastern Croatia. Sci Total Environ 339:277-282
Chakraborti D, Mukerjee, S.C., Pati, S., Sengupta, M.K., Rahman, M.M., Chowdhury, U.K. L,
D., Chanda, C.R., Chakraborti, A.K., Basu, G.K. (2003) Arsenic groundwater
contamination in Middle Ganga Plain, Bihar, India: a future danger? Environ Health
Perspect 111:1194-1201
Chakraborti D, Sengupta, M. K., Rahman, M.M.,Chowdhury, U.K., Lodh, D., Ahamed, S.,
Hossain, Md. A., Basu, G.K., Mukerjee, S.C., Saha, K. C. (2003) Groundwater arsenic
exposure in India. Arsenic Exposure and Health Effects V 108:3-24
Changui C, Stone, W.E.F., Vielvoye, L., Dereppe, J.-M. (1990) Characterization by nuclear
magnetic resonance spectroscopy, Ferron assay , and acidification of partially
neutralized aluminium solutions. J Chem Soc, Dalton Trans 5:1723-1726
Charlet L, Polya, D. (2006) Arsenic in shallow, reducing groundwater in Southern Asia: an
environmental health disaster. Elements 2:91-96
Chen C-J, Hsu, L-I., Tseng, C-H., Hsueh, Y-M., Chiou, H-Y. (1999) Emerging Epidemics of
arseniasis in Asia. Arsenic exposure and health effects In: Abemathy, C.O., Calderon,
R.L. and Chappell, W.R. (eds.). Arsenic Exposure and Health Effects, Chapman &
HaU, London, pp. 113-121
Chen YC, Su, H.J., Guo, Y.L., Hsueh, Y.M., Smith, T.J., Ryan, L.M., Lee, M.S., Christiani,
D.C. (2003) Arsenic methylation and bladder cancer risk in Taiwan. Cancer Causes &
Control 14:303-310
Chen Z, Fan, B., Peng, X., Zhang, Z., Fan, J., Luan, Z. (2006) Evaluation of Al30 polynuclear
species in polyaluminum solutions as coagulant for water treatment. Chemosphere
64:912-918
Christensen TH, Bjerg, P.L., Banwart, S.A., Jakobsen, R., Heron, G., Alrechtsen, H.-J. (2000)
Characterization of redox conditions in groundwater contaminant plumes. J Contam
Hydrol 45:165-241
Clark ID, Fritz, P. (1997) Environmental Isotopes in Hydrogeology. Lewis Publishers. CRC
Press LLC
Coetsiers M, Walraevens, K. (2006) Chemical characterization of the Neogene Aquifer,
Belgium. Hydrol J 14:1556-1568
Cohut I, Bendea, C. (1999) Geothermal Development Opportunities in Romania.
Geothermische Energie 24/25, 7 Jahrgang/Heft , Geothermische Vereingung, Geeste,
Marz/September:8-13
Csalagovitis I (1999) Arsenic-bearing artesian waters of Hungary. Ann Rep Geological
Institute of Hungary 1992-1993/II:85-92
- 146 -
 References

Daniele L (2004) Distribution of arsenic and other minor tarce elements in the groundwater of
Ischia Island (southern Italy). Environmental Geology 46:96-103
Deak J, Stute, M., Rudolph, J., Sonntag, C. (1987) Determination of the flow regime of
Quaternary and Pliocene layers in the Great Hungarian Plain (Hungary) by D, 18O,
14C and noble gas measurements. In Isotope Techniques in Water Resources
Development. In: Proc IAEA Sympos, Vienna, IAEA Proc Series STI/PUB/757):335-
350
Djuric D, Jevtic, G. (2008) Solutions for groundwater management in areas affected by high
arsenic content – Vojvodina case study. In: Proc 2nd Internat Congress on ‘‘Arsenic in
the Environment’’, 21–23 May 2008, Valencia, Spain:159-160
DPHE Ba (2001) Arsenic contamination of groundwater in Bangladesh. In: Kinniburgh, DG,
Smedley, PL (Eds), Final Report British Geological Survey Report WC/00/19, British
Geological Survey, Vol. 2, Keyworth, UK
Duan J, Gregory, J. (2003) Coagulation by hydrolysing metal salts. Advances in Colloid and
Interface Science 100-102:475-502
Duddridge, JE, Wainwright, M. (1981) Heavy metals in river sediments - calculation of metal
adsorption maxima using Langmuir and Freundlich isotherms. Environmental Pollution
(Series B) 2: 387-397.
Duchi V, Minissale, A., Paolieri, M., Prati, F., Valori, A. (1992) Chemical relationships
between discharging fluids in the Siena-Radicofani graben and the deep fluids produced
by the geothermal fields of Mt Amiata, Torre Alfina and Latera (Central Italy).
Geothermics 21:401-413
Dutta PK, Ray, A.K., Sharma, V.K., Millero, F.J. (2004) Adsorption of arsenate and arsenite
on titanium dioxide suspension. Journal of Colloid and Interface Science 278:270-275
EC (1998) Council directive 98/83/EC. Official Journal of the European Communities L330/32
Edwards M (1994) Chemistry of As removal during coagulation and Fe-Mn oxidation.
AWWA:64-78
Emsley J (2001) Nature's building blocks: an A-Z guide to the elements. Oxford University
Press, Oxford:43, 529
EPA (2000) Technologies and costs for removal of arsenic from drinking water. Report EPA
815-R-00-028:284 pp.
EPA (2001) Treatment of arsenic residuals form drinking water removal processes. Report
EPA 600-R-01-033. http://ec.europa.eu/environment/water/water-
urbanwaste/directiv.html
EPA (2002) Arsenic treatment technologies for soil, waste and water. EPA-542-R-02-004
www.epa.gov/tioclu-in.org/arsenic
Fan M, Brown, R.C., Sung,. S.W., Huang, C.-P., Ong, S.K., van Leeuwen, J. (2003)
Comparison of polymeric and conventional coagulants in arsenic(V) removal. Water
Enviroment Research 75 (4):308-313
Flesch GD, Anderson, A.R., Svec, H.J. (1973) A secondary isotopic standard for 6Li/7Li
determinations. IntJ Mass Spectr Ion Phys 12:265-272
Fletcher T, Leonardi, G., Goessler, W., Gurzau, E., Koppova, K., Kumar, R., Rudnai, P.,
Vahter M. (2008) Arsenic in residential drinking water and cancer in Central Europe -
the ASHRAM study. The 3rd Central and Eastern European Conf Health and the
Environment (CEECHE), Cluj-Napoca, Romania, 19–22nd November 2008
Foster AL, Brown Jr., G.E., Tingle, T.N., Parks, G.A. (1998) Quantitative arsenic speciation in
mine tailings using X-ray absorption spectroscopy. American Mineralogist 83:553-568
Fouillac C, Michard, G. (1981) Sodium/lithium ratio in water applied to geothermometry of
geothermal reservoirs. Geothermics 10:55-70
Foundation AR (1998) Amy, G.L., M. Edwards, M. Benjamin, K. Carlson, J. Chwirka, P.
Brandhuber, L. McNeill and F. Vagliasindi, Black, B., Samuel, A.L., Perry, T., Chinn,
- 147 -
References

D., US EPA (eds.), Arsenic Treatability Options and Evaluation of Residuals

Management Issues. Draft Report, April 1998
Foundation AR (2002) AWWA Research Foundation, Black, B., Samuel, A.L., Perry, T.,
Chinn, D., US EPA (eds.). Amer Water Works Association
Frapporti G, Vriend, S.P., van Gaans, P.F.M. (1996) Trace elements in the shallow
groundwater water of the Netherlands. A geochemical and statistical interpretation of
the National Monitoring Network data. Aquat Geochem 2:51-80
Fuller CC, Davis, J.A., Waychunas, G.A. (1993) Surface chemistry of ferrihydrite: Part 2.
Kinetics of arsenate adsorption and coprecipitation. Geochimica et Cosmochimica Acta
57:2271-2282
Furrer G, Ludwig, C., and Schindler, P.W. (1992) On the chemistry of the Keggin Al13
polymer. I. Acid-Base Properties. J Colloid and Interface Science 149 (1):56-67
Furrer, G., Philips, B.L., Ulrich, K.-U., Pöthig, R., Casey, W.H. (2002) The origin of
aluminum flocs in polluted streams. Science 297: 2245-2247
Gábris G, Nádor, A. (2007) Long-term fluvial archives in Hungary: response of the Danube
and Tisza rivers to tectonic movements and climatic changes during the Quaternary: a
review and new synthesis. Quatern. Sci. Rev. 47:847-854
Garcia-Sanchez A, Moyano, A., Mayorga, P. (2005) High arsenic contents in groundwater of
central Spain. Environ Geol 47:847-854
Garnier J-M, Travassac, F., Lenoble, V., Rose, J., Zheng, Y., Hossain, M.S., Chowdhury, S.H.,
Biswas, A.K., Ahmed, K.M., Cheng, Z., van Geen, A. (2010) temporal variations in
arsenic uptake by rice plants in Bangladesh: The role of iron plaque in paddy fields
irrigated with groundwater. Science of the Total Environment 408 (19):4185-4193
Geary DH, Staley, A.W., Muller, P., Magyar, I. (2002) Iterative changes in Lake Pannon
Melanopsis reflect a recurrent theme in gastropod morphological evolution.
Paleobiology 28:208-221
Ghosh M, Bandyopadhyay, M., Mukherjee, A. (2010) Genotoxicity of titanium dioxide (TiO2)
nanoparticles at two trophic levels:plant and human lymphocytes. Chemosphere
81:1253-1262
Ghurye, G., Clifford, D. (2001) Laboratory study on the oxidation of arsenic(III) to arsenic(V).
Report EPA 600/R-01/021, EPA Contract 8C-R311-NAEX:11 pp.
Gilboa-Garber N (1971) Direct spectrophotometric determination of inorganic sulphide in
biological materials and in other complex mixtures. Anal Biochem 43:129-133
Gomez JJ, Lillo, J., Sahun, B. (2006) Naturally occurring arsenic in groundwater and
identification of the geochemical sources in the Duero Cenozoic Basin, Spain. Environ
Geol 50: 1151-1170
Grafe M, Eick, M.J., Grossl, P.R. (2001) Adsorption of arsenate (V) and arsenite (III) on
goethite in the presence and absence of dissolved organic carbon. Soil Sci Soc Am J
65:1680-1687
Gray NF (2005) Water technology - An introduction for environmental scientists and
engineers. 2nd edition, Elsevier, Oxford:645 pp.
Güler C, Thyne, G.D., McCray, J.E., Turner, A.K. (2002) Evaluation of graphical and
multivariate statistical methods for classification of water chemistry data.
Hydrogeology J 18:455-474
Gurzau ES, Gurzau, A.E. (2001) Arsenic in drinking water from groundwater in Transylvania,
Romania: an overview. In: Chappell, W.R., Abernathy, C.O., Calderon, R.L. (eds.),
Arsenic Exposure and Health Effects IV Elsevier:181-184
Habuda-Stani M, Kule, M., Kalajdi, B., Romi K. (2007) Quality of groundwater in
eastern Croatia. The problem of arsenic pollution. Desalination 210:157-162
Harvey CF, Swartz, C.H., Badruzzaman, A.B.M., Keon-Blute, N., Yu, W., Ali, M.A., Jay,, J.
B, R., Niedan, V., Brabander, D., Oates, P.M., Ashfaque, K.N., Islam, S.,, Hemond HF,
- 148 -
 References

Ahmed, M.F. (2002) Arsenic mobility and groundwater extraction in Bangladesh.

Science 298:1602-1606
Henke KR, Hutchison, A. (2009) Arsenic Chemistry. In: Henke, KR (ed), Arsenic
environmental ehemistry, health threats and waste treatment John Whiley & Sons: 575
pp.
Henley RW, Ellis, A.J. (1983) Geothermal systems ancient and modern: a geochemical review.
Earth Sci Rev 19:1-50
Hering JG, Chen, P.-Y., Wilkie, J.A., Elimelech, M. (1997) Arsenic removal from drinking
water during coagulation. Journal of Environmental Engineering 123 (8):800-807
Hering JG, Chen, P.-Y., Wilkie, J.A., Elimelech, M., Liang, S. (1996) Arsenic removal by
ferric chloride. J AWWA 88 (4):155-167
Ho YS, McKay, G. (1999) Pseudo-second order model for sorption processes. Process
Biochemistry 34:451-465
Hopenhayn C (2006) Arsenic in drinking water. Elements 2:103-107
Horneman A, van Gee, A., Kent, D.V., Mathe, P.E., Zheng, Y., Dhar, R.K., O’Connell, S.,,
Hoque MA, Aziz, Z., Shamsudduha, M., Seddique, A.A., Ahmed, K.M. (2004)
Decoupling of As and Fe release to Bangladesh groundwater under reducing
conditions. Part I: evidence from sediment profiles. Geochimica et Cosmochimica Acta
68:3459-3473
Horváth F, Cloetingh, S. (1996) Stress-induced late-stage subsidence anomalies in the
Pannonian basin. Tectonophysics 266:287-300
Huerta-Diaz MA, Tessier, A., Carignan, R. (1988) Geochemistry of trace metals associated
with reduced sulphur in freshwater sediments. Applied Geochemistry 13:213-233
Islam FS, Gault, A.G., Boothman, C., Polya, D.A., Charnock, J. M., Chatterjee, D., Lloyd, J.
(2004) Role of metal-reducing bacteria in arsenic release form bengal delta sediments.
Nature 430:68-71
IUPAC (1994) Recommendation for the characterization of porous solids. Pure & Appl Chem
66 (8):1739-1758
Jekel M, Amy, G.L. (2006) Chapter 11: Arsenic removal during drinking water treatment. In:
Newcombe, G, Dixon, D (eds.) Interface Science in Drinking Water Treatment:196 -
206
Jeong Y, Fan, M., Singh, A., Chuang, C.-L., Saha, B., van Leeuwen, J.H. (2007) Evaluation of
iron oxide and aluminum oxide as potential arsenic(V) adsorbents. Chemical
Engineering and Processing 46:1030-1039
Jimenez C, Mertens, J., Rowland, H.A., Baciu, C., Berg, M., Furrer, G., Hug, S., Cordos, E.
(2009) Groundwater geochemistry and As content in the Eastern Pannonian Basin
(Romania) – PCA analysis. Geochimica et Cosmochimica Acta 73:A596
Jong T, Parry, D.L. (2003) Removal of sulphate and heavy metals by sulphate reducing
bacteria in short-term bench scale upflow anaerobic packed bed reactor runs. Water
Research 37:3370-3389
Juhász A, Tóth, T.M., Ramseyer, K., Matter, A. (2002) Connected fluid evolution in fractured
crystalline basement and overlying sediments, Pannonian Basin, SE Hungary. Chem
Geol 182:91-120
Juhász E, Phillips, L., Müller, P., Ricketts, B., Toth-Makk, Á., Lantos, M., Kovács, L.Ó.
(1999) Late Neogene sedimentary facies and sequences in the Pannonian Basin,
Hungary. In: Durand, B, Jolivet, L, Horváth, F, Séranne, M (Eds), The Mediterranean
Basins: Tertiary Extension within the Alpine Orogen Geol Soc, vol 156 London Special
Publication:335-356
Juhász G, Müller, P., Toth-Makk, Á. (2004) Alluvial architecture and fluvial cycles in
Quaternary deposits in a continental interior basin, E Hungary. Geol Croat 57:171-190

- 149 -
References

Juhász G, Pogácsás, G., Magyar, I., Vakarcs, G. (2007) Tectonic versus climatic control on the
evolution of fluvio-deltaic systems in a lakae basin, Eastern Pannonian Basin. Sed Geol
202:72-95
Kalnin, JR, Kotomin, E.A. (2001) A novel relation for the effective coefficient in
inhomogeneous media. Computer Modelling & New Technologies 5(1): 18-27
Katsoyiannis IA, Katsoyiannis, A.A. (2006) Arsenic and other metal contamination of
groundwater in the industrial area of Thessaloniki Northern Greece. Environ Monitor
Assess 123:393-406
Kim M-J, Nriagu, J., Haack, S. (2000) Carbonate ions and arsenic dissolution by groundwater.
Environmental Science and Technology 34 (15):3094-3100
Kim Y, Kim, C., Choi, I., Rengaraj, S., Yi, J. (2004) Arsenic removal using mesoporous
alumina prepared via a templating method. Environmental Science and Technology
38:924-931
Kirk MF, Holm, T.R., Park, J., Jin, Q., Sanford, R.A., Fouke, B.W., Bethke, C.M. (2004)
Bacterial sulfate reduction limits natural arsenic contamination in groundwater.
Geology 32:953-956
Korim K (1972) Geological aspects of thermal water occurrences in Hungary. Geothermics
1:96-102
Kouras A, Katsoyiannis, I., Voutsa, D. (2007) Distribution of arsenic in groundwater in the
area of Chalkidiki, Northern Greece. Journal of Hazardous Materials 147:890-899
Krapac IG, Dey, W.S., Roy, W.R., Smyth, C.A., Storment, E., Sargent, S.L., Steele, J.D.
(2002) Impacts of swine manure pits on groundwater quality. Environ Pollut 120:475-
492
Ladeira ACQ, Ciminelli, V.S.T., Duarte, H.A., Alves, M.C.M., Ramos, A.Y. (2001)
Mechanism of anion retention from EXAFS and density functinoal calculations:
Arsenic(V) adsorbed on gibbsite. Geochimica et Cosmochimica Acta 65 (8):1212-1217
Lenkey L, Dövényi, P., Horváth, F., Cloetingh, S.A.P.L. (2002) Geothermics of the Pannonian
Basin, and its bearing on the neotectonics. EGU Stephan Mueller Special Publ Ser
3:29-40
Lin T-F, Wu, J.-K. (2001) Adsorption of arsenite and arsenate within activated alumina grains:
equilibrium and kinetics. Water Research 35 (8):2049-2057
Lindberg A-L, Goessler, W., Gurzau, E., Koppova, K., Rudnai, P., Kumar, R., Fletcher, T.,
Leonardi, G., Slotova, K., Gheorghiu, E., Vahter, M. (2006) Arsenic exposure in
Hungary, Romania and Slovakia. J Environ Monitor 8:203-208
Lloyd JR, Oremland, R.S. (2006) Microbial transformations of arsenic in the environment:
from soda lakes to aquifers. Elements 2:85-90
Lovley DR, Goodwin, S. (1988) Hydrogen concentrations as an indicator of the predominant
terminal electron-accepting reactions in aquatic sediments. Geochimica et
Cosmochimica Acta 52:2993-3003
Lovley DR, Klug, M.J. (1986) Model for the distribution of sulphate reduction and
methanogenesis in freshwater sediments. Geochimica et Cosmochimica Acta 50:11-18
Magyar I, Geary, D.H., Müller, P. (1999) Paleogeographic evolution of the Late Miocene Lake
Pannon in Central Europe. Palaeogeog Palaeoclimatol Palaeoecol 147:151-167
Makris KC, Sarkar, D., Parsons, J.G., Datta, R., Gardea-Torresdey, J.L. (2009) X-Ray
absorption spectroscopy as a tool investigating arsenic(III) and arsenic(V) sorption by
an aluminum-based drinking-water treatment residual. Journal of Hazardous Materials
171:980-986
Manning, B, Goldberg, S. (1997) Sorption and stability of arsenic(III) at the clay mineral-water
interface. Environ. Sci. Technol. 31:2005-2011
Marie A, Vengosh, A. (2001) Sources of salinity in ground water from Jericho Area, Jordan
Valley. Groundwater 39:240-248
- 150 -
 References

McArthur JM, Banerjee, D.M., Hudson-Edwards, K.A., Mishra, R., Purohit, R., Ravenscroft,
P., Cronin, A., Howarth, R.J., Chatterjee, A., Talukder, T., Lowry, D., Houghton, S.,
Chadha, D.K. (2004) Natural organic matter in sedimentary basins and its relation to
arsenic in ground water: the example of West Bengal and its worldwide implications.
Applied Geochemistry 19:1255-1293
McArthur JM, Ravenscroft, P., Safiulla, S., Thirlwall, M.F. (2001) Arsenic in groundwater:
testing pollution mechanisms for sedimentary aquifers in Bangladesh. Water Resour
Res 37:109-117
Merkel BJ, Planer-Friedrich, B. (2008) Groundwater Geochemistry - A practical guide to
modeling of natural and contaminated aquatic systems, 2nd ed., Berlin-Heidelberg: 237
pp.
Mertens J (2010) Report on the optimal reactor configurations for the sustained ex-situ
remediation of groundwater contaminated with a range of geogenic elements, including
any publications in refereed journals. Deliverable of the European Commission FP6
Marie Curie Research Training Network AquaTRAIN (Contract No MRTH-CT-2006-
035420; Coordinator DA Polya, University of Manchester, UK) Deliverable 13 (WP3
Deliverable 35): 49 pp.
Mertens J, Casentini, B., Masion, A., Pöthig, R., Wehrli, B., Furrer, G. (2012) Polyaluminum
chloride with high Al30 content as removal agent for arsenic-contaminated well water.
Water Research 46:53-62
Mertens, J, Rose, J., Kägi, R., Chaurand, P., Plötze, M., Wehrli, B., Furrer, G. Adsorption of
arsenic on polyaluminum granulate, Environ. Sci. Technol. (submitted)
Millot R, Guerrot, C., Vigier, N. (2004) Accurate and high precision measurement of lithium
isotopes in two reference materials by MC-ICP-MS. Geostand Gepanal Res 28:53-159
Millot R, Négrel, P. (2007a) Multi-isopotic tracing (d7Li, d11B, 87Sr/86Sr) and chemical
geothermometry: evidence from hydro-geothermal systems in France. Chem Geol
244:664-678
Millot R, Négrel, Ph., Petelet-Giraud, E. (2007b) Multi-isotopic (Li, B, Sr, Nd) approach for
geothermal reservoir characterization in the Limagne Basin (massif Central, France).
Applied Geochemistry 22:2307-2325
Millot R, Scaillet, B., Sanjuan, B. (2010) Lithium isotopes in island arc geothermal systems:
Guadeloupe, Martinique (French West Indies) and experimental approach. Geochimica
et Cosmochimica Acta 74:1852-1871
Mondal P, Majumder, C.B., Mohanty, B. (2006) Laboratory based approaches for arsenic
remediation form contaminated water: Recent developments. Journal of Hazardous
Materials B137:464-479
Moore JN, Fickin, W.H., Johns, C. (1988) Partitioning of arsenic and metals in reducing
sulfidic sediments. Environ Sci Technol 22:432-437
Mrazovac S, Basic, D. (2009) Methane-rich geothermal waters in the Pannonian Basin of
Vojvodina (northern Serbia). Geothermics 38:303-312
Mukherjee A, Fryar, A.E., O'Shea, B.M. (2009) Major occurrences of elevated arsenic in
groundwater and other natural waters. In: Henke, KR (ed), Arsenic environmental
ehemistry, health threats and waste treatment John Whiley & Sons: pp.303-350
Mukhopadhyay D, Sanyal, S.K. (2004) Complexation and release isotherm of arsenic in
arsenic-humic/fulvic equilibrium study. Australian Journal of Soil Research 42 (7):815-
824
Nádor A, Lantos, M., Tóth-Makk, Á., Thamó-Bozsó, E. (2003) Milankovitch-scale multi-
proxy records from fluvial sediments of the last 2.6 Ma, Pannonian Basin, Hungary.
Quatern Sci Rev 22:2157-2175

- 151 -
References

Nádor A, Thamó-Bozsó, E., Magyari, Á., Babinszki, E. (2007) Fluvial responses to tectonics
and climate change during teh Weichselian in the eastern part of the Pannonian Basin
(Hungary). Sedimentary Geology 202:174-192
Newman DK, Beveridge, T.J., Morel, F.M.M. (1997) Precipitation of arsenic trisulfide by
Desulfotomaculum auripigmentum. Appl Environ Microbiol 63:2022-2028
Newville M (2001 ) IFFEFIT: interactive XAFS analysis and FEFF fitting. J Synchrotron Rad
8:322-324
Nickson RT, McArthur, J.M., Burgess, W., Ahmed, K.M., Ravenscroft, P., Rahman, M. (1998)
Arsenic poisoning of Bangladesh groundwater. Nature 395:338
Nickson RT, McArthur, J.M., Ravenscroft, P., Burgess, W.G., Ahmed, M. (2000) Mechanism
of arsenic release to groundwater, Bangladesh and West Bengal. Applied Geochemistry
15 (4):403-413
Nikolaidis NP, Dobbs, G.M., Lackovic, J.A. (2003) Arsenic removal by zero-valent iron: field,
laboratory and modeling studies. Water Research 37 (6):1416-1425
Nikolaidis NP, Tyrovola, K. (2006) Removal of arsenic form groundwater - mechanism,
kinetics, field/pilot and modeling studies. In: Lo, IMC, Lai, KCK (eds) Zero-valent iron
reactive materials for hazardous waster and inorganics removal ASCE Publications,
reston, VA:pp. 151-164 (Chapter 159)
Nilchi A, Rasouli Garmarodi, Janitabar Darzi, S. (2011) Removal of arsenic from aqueous
solutions by an adsorption process with titania-silica binary oxide nanoparticle loaded
polyacrylonitrile polymer. Journal of Applied Polymer Science 119:3495-3503
Nordstrom DK, Ball, J.W., Donahoe, R.J., Whittemore, D. (1989) Groundwater chemistry and
water–rock interactions at Stripa. Geochimica et Cosmochimica Acta 53:1727-1740
O' Day P (2006) Chemistry amd mineralogy of arsenic. Elements 2:77-83
O' Day PA, Vlassopoulos, D., Root, R., Rivera, N. (2004) The influence of sulphur and iron on
dissolved arsenic concentrations in the shallow subsurface under changing redox
conditions. Proc Nat Acad Sci 101:13703–13708
Panno SV, Hackley, K.C., Hwang, H.H., Greenberg, S.E., Krapac, I.G., Landsberger, S.,
O’Kelly D.J. (2006) Characterization and identification of Na–Cl sources in
groundwater. Groundwater 44:176-187
Parker DR, Bertsch, P.M. (1994) Identification and Quantification of the “AI13” tridecameric
polycation using Ferron. Environ Sci Technol 26 (5):908-914
Parker DR, Bertsch, P.M. (1994) Formation of the “AI13” tridecameric polycation under
diverse synthesis conditions. Environ Sci Technol 26 (5):914-921
Pena ME, Korfiatis, G.P., Patel, M., Lippincott, L., Meng, X. (2005) Adsorption of As(V) and
As(III) by nanocrystalline titanium dioxide. Water Research 39:2327-2337
Piazzoli A (2011) Arsenic removal form drinking water using nanoparticles. Term Paper in
Biogeochemistry and Pollutant Dynamics, May 2011: 18 pp.
Polya DA, Gault, A.G., Diebe, N., Feldman, P., Rosenboom, J.W., Gilligan, E., Fredericks, D.,
Milton, A.H., Sampson, M., Rowland, H.A.L., Lythgoe, P.R., Jones, J.C., Middleton,
C., Cooke, D.A. (2005) Arsenic hazard in shallow Cambodian groundwaters. Mineral
Mag 69:807-823
Quicksall AN, Bostick, B.C., Sampson, M.L. (2008) Linking organic matter deposition and
iron mineral transformations to groundwater arsenic levels in the Mekong delta,
Cambodia. Applied Geochemistry 23:3088-3098
Rakotonarivo E, Bottero, J.Y., Oases, J.M. (1984) Study of the adsorption of long chain
alkylsulfonates from aqueous solutions on aluminum hydroxide gels. Colloids and
Surfaces 9:273-292
Rao AR, Srinivas, V.V. (2006) Regionalization of watersheds by hybrid-cluster analysis. J
Hydrol 318:37-56

- 152 -
 References

Ravel B, Newville, M. (2005) Athena, Artemis, Hephaestus: data analysis for X-Ray
absorption spectroscopy using IFEFFIT. J Synchrotron Rad 12:537-541
Raven KP, Jain, A., Loeppert, R. H. (1998) Arsenite and arsenate adsorption on ferrihydrite:
kinetics, equilibrium, and adsorption envelopes. Environmental Science and
Technology 32:344-349
Reed, B.E., Matsumoto, M.R. (1993) Modeling cadmium adsorption by activated carbon using
the Langmuir and Freundlich isotherm expressions. Separation Science and Technology
28(13&14):2179-2195
Reza AHMS, Jean, J.-S., Lee, M.-K., Liu, C.-C., Bundschuh, J., Yang, H.-J., Lee, J.-F., Lee,
Y.-C. (2010) Implications of organic matter on arsenic mobilization into groundwater:
Evidence from northwestern (Chapai-Nawabganj), central (Manikganj) and
southeastern (Chandpur) Bangladesh. Water Research 44:5556-5574
Roberts LC, Hug, S.J., Dittmar, J., Voegelin, A., Saha, G.C., Ali, M.A., Badruzzaman, A.B.M.,
Kretzschmar, R. (2007) Spatial distribution and temporal variability of arsenic in
irrigated rice fields in Bangladesh. 1. Irrigation water. Environ Sci Technol 41:5960-
5966
Roberts LC, Hug, S.J., Ruettimann, T., Billah, MD M., Khan, A. W., Rahman, M.T. (2004)
Arsenic Removal with iron)II) and Iron(III) in water with high silicate and phosphate
concentrations. Environ Sci Technol 38:307-315
Rögl F (1999) Mediterranean and Paratethys. Facts and hypotheses of an Oligocene to
Miocene Paleogeography (short overview). Geol Carpath 50:339-349
Rowland HAL, Boothman, C., Pancost, R., Gault, A.G., Polya, D.A., Lloyd, J.R (2009a) The
role of indigenous microorganisms in the biodegradation of naturally occurring
petroleum, the reduction of iron, and the mobilization of arsenite from West Bengal
aquifer sediments. J Environ Qual 38:1598-1607
Rowland HAL, Jimenez, C., Omoregie, E., Mertens, J., Berg, M., Baciu, C., Hug, S.J. (2009b)
Controls on groundwater geochemistry and arsenic mobilisation processes in aquifers
of Eastern Europe (Pannonian Basin). Geochimica et Cosmochimica Acta 73:A1126
Rowland HAL, Omoregie, E., Millot, R., Jimenez, C., Mertens, J., Baciu, C., Hug, S.J., Berg,
M. (2011) Geochemistry and arsenic behaviour in groundwater resources of the
Pannonian Basin (Hungary and Romania). Applied Geochemistry 26:1-17
Rowsell J, Nazar, L.F (2000) Speciation and thermal transformation in alumina sols: structures
of the polyhydroxyoxoaluminum cluster [Al30O8(OH)56(H2O)26]18+ and its
-Keggin
moieté. J Am Chem Soc 122:3777-3778
Rozanski K (1985) Deuterium and oxygen-18 in European groundwaters – links to
atmospheric circulation in the past. Chem Geol 52:349-363
Rustad JR (2005) Molecular dynamics simulation of the titrtaion of polyoxocations in aqueous
solution. Geochimica et Cosmochimica Acta 69 (18):4397-4410
Sacchi M, Horváth, F. (2002) Towards a new time scale for the Upper Miocene continental
series of the Pannonian Basin (Central Paratethys). EGU Stephan Mueller Special Publ
Ser 3:79-94
Sansone FJ, Popp, B.N., Rust, T.M. (1997) Stable carbon isotopic analysis of lowlevel
methane in water and gas. Anal Chem 69:40-44
Sarazin G, Michard, G., Prevot, F. (1999) A rapid and accurate spectroscopic method for
alkalinity measurements in sea water samples. Water Research 33:290-294
Saunders JA, Lee, M.-K., Uddin, A., Mohammad, S., Wilkin, R.T., Fayek, M., Korte, N.E.
(2005) Natural arsenic contamination of Holocene alluvial aquifers by linked tectonic,
weathering and microbial processes. Geochem Geophys Geosyst 6:Q04006
Schönherr S, Görz, H., Bertram, R., Müller, D., Gessner, W. (1983) Vergleichende
Untersuchungen an unterschiedlich dargestellten basischen Aluminiumchloridlösungen.
Z anorg allg Chem 502:113-122
- 153 -
References

Schönherr, S, Görz, H., Bertram, R. (1987) Zur Anwendung der zeitabhängigen

Komplexbildung mit Ferron für die Charakterisierung basischer Metallkationen. Wiss.
Z. Päd. Hochsch. Potsdam 31:67-74
Singh TS, Pant, K.K. (2004) Equilibrium, kinetics and thermodynamic studies for adsorption
of As(III) on activated alumina. Separation and PurificationTechnology 36:139-147
Smedley PL, Edmunds, W.M. (2002a) Redox patterns and trace-element behaviour in the East
Midlands Triassic Sandstone Aquifer, UK. Groundwater 40:44-58
Smedley PL, Kinniburgh D.G. (2002b) Review of the source, behaviour and distribution of
arsenic in natural waters. Applied Geochemistry 17:517-568
Smedley PL, Zhang, G., Luo, Z. (2003) Mobilization of aresnic and othertrace elements in
fluviolacustrine aquifers of the Huhhot Basin, Inner Mongolia. Applied Geochemistry
18:1453-1477
Smith AH, Lingas, E.O., Rahman, M. (2000) Contamination of drinking-water by arsenic in
Bangladesh: a public health emergency. Bull WHO 78:1093-1103
Sorg TJ, Logsdon, G.S. (1978) Treatment technology to meet the interim primary drinking
water regulations for inorganics: Part 2. J AWWA 7:379-392
Stauder S (2007) Chemistry and treatment of groundwater in the Vojvodina. Water Sci
Technol 7:93-101
Stumm W (1992) Chemistry of the solid-water interface: processes at the mineral-water and
particle-water interface in natural systems. John Wiley & Sons New York:432 pp.
Stute M, Deak, J. (1989) Environmental isotope study (14C, 13C, 18O, D, Noble gases) on
deep groundwater circulation systems in Hungary with reference to Palaeoclimate.
Radiocarbon 31:902-918
Su C, Puls, R.W. (2001) Arsenate and arsenite removal by zerovalent iron: effect of phosphate,
silicate, carbonate, borate, sulfate, chromate, molybdate, and nitrate relative to chlorite.
Environmental Science and Technology 35 (22):4562-4568
Sylvester P, Westerhoff, P., Möller, T., Badruzzaman, M., Boyd, O. (2007) A hybrid sorbent
utilizing nanoparticles of hydrous iron oxide for arsenic removal from drinking water.
Environmental Engineering Science 24 (1):104-112
Takahashi, Y, Sakamitsu, M., Tanaka, M. (2011) Diffusion coefficients of arsenate and
arsenite in water at various pH. Chem. Letters 40:1187-1188
Theron J, Walker, J.A., Cloete, T.E. (2008) Nanotechnology and water treatment: applications
and emerging opportunities. Critical Reviews in Microbiology 34:43-69
Tretner A (2003) Bedeutung und Verhalten von Arsen in der Hydrosphäre. Grundwasser 1:3-
12
Trunko L (1996) Geology of Hungary. Beiträge zur regionalen Geologe der Erde VIII
(Gebruder Borntraeger, Berlin):464
Tubi A, Agbaba, J., Dalmacija, B., Ivancev-Tumbas, I., Dalmacija, M. (2010) Removal of
arsenic and organic matter from groundwater using ferric and alum salts: A case study
of central Banat region (Serbia). Journal of Environmental Science and Health Part A
45:363-369
Tyrovola K, Nikolaidis, N.P., Veranis, N., Kallithrakas-Kontos, N., Koulouridakis, P.E. (2006)
Arsenic removal from geothermal waters with zero-valent iron - Effect of temperature,
phosphate and nitrate. Water Research 40:2375-2386
Ujevi M, Dui, ., Casiot, C., Sipos L., Santo, V., Dadi, ., J. Halami, J. (2010) Occurrence
and geochemistry of arsenic in the groundwater of Eastern Croatia. Applied
Geochemistry 25:1017-1029
Umitsu M (1993) Late quaternary sedimentary environments and landforms in the Ganges
Delta. Sed Geol 83:177-186
Van Geen A, Rose, J., Thoral, S., Garnier, J.M., Zheng, Y., Bottero, J.Y. (2004) Decoupling of
As and fe release to Bangladesh groundwater under reducing conditions. Part II:
- 154 -
 References

Evidence from sediment incubations. Geochimica et Cosmochimica Acta 68 (17):3475-

3486
Varnavas SP, Cronan, D.S. (1988) Arsenic, antimony and bismuth in sediment and waters from
the Santorini hydrothermal field, Greece. Chemical Geology 67 (3-4):295-305
Varsányi I, and Kovács, L.O. (2006a) Arsenic, iron and organic matter in the groundwater in
the Pannonian Basin, Hungary. Applied Geochemistry 21:949-963
Varsányi I, Fodre, Z., Bartha, A. (1991) Arsenic in drinking water and mortality in the
Southern Great Plain, Hungary. Environ Geochem Health 13:14-22
Varsányi I, Kovács, L.Ó (2006b) Arsenic, iron and organic matter in sediments and
groundwater in the Pannonian Basin, Hungary. Appl Geochem 21:949-963
Varsányi I, Kovács, L.Ó. (2009) Origin, chemical and isotopic evolution of formation water in
geopressured zones in the Pannonian Basin, Hungary. Chem Geol 264:187-196
Varsányi I, Kovács, L.Ó., Karpati, Z., Matray, J.M. (2002) Carbon forms in formation waters
from the Pannonian Basin, Hungary. Chem Geol 189:165-182
Varsányi I, Matray, J.-M. Kovács, L.O. (1997) Geochemistry of formation water in the
Pannonian Basin (southeast Hungary). Chemical Geology 140:89-106
Varsányi I, Matray, J.-M. Kovács, L.O. (1999) Hydrogeochemistry in two adjacent areas in the
Pannonian Basin (Southeast-Hungary). Chemical Geology 156:25-39
Vatutsina OM, Soldatov, V.S., Sokolova, V.I., Johann, J., Bidden, M., Weissenbacher, A.
(2007) A new hybrid (polymer/inroganic) fibrous sorbent for arsenic removal from
drinking water. Reactive and Functional Polymers 67 (3):184-201
Veto I, Futo, I., Horvath, I., Szanto, S. (2004) Late and deep fermentative methanogenesis as
reflected in the H–C–O–S isotopy of the methane–water system in deep aquifers of the
Pannonian Basin (SE Hungary). Org Geochem 35:713-723
Viczián I (2002) Mineralogy of pliocene to pleistocene pelitic sediments of the Great
Hungarian Plain. Acta Mineral-Petrograph 43:39-53
Vidovic M, Nikic, Z., Milovanovic, B. (2006) Water quality of the North Banat basal aquifer
system. Geograph Pannon 10:43-46
Warner KL (2001) Arsenic in glacial drift aquifers and the implication for drinking water –
Lower Illinois River Basin. Ground Water 39:433-442
Weerasooriya R, Tobschall, H.J., Wijesekara, H.K.D.K., Arachchige, E.K.I.A.U.K.,
Pathirathne, K.A.S. (2003) On the mechanistic modeling of As(III) adsorption on
gibbsite. Chemosphere 51:1001-1013
Wehrli B, Wieland, E., Furrer, G. (1990) Chemical mechanisms in the dissolution kinetics of
minerals; the aspect of active sites. Aquatic Sciences 52 (1):1-31
Welch AH, Wetjohn, D.B., Helsel, D.R., Wanty, R.B. (2000) Arsenic in groundwater of the
United States: occurrence and geochemistry. Ground Water 38:589-604
WHO (1998) Aluminium in Drinking-water. Guidelines for drinking-water quality, 2nd ed,
Geneva:pp. 14
WHO (2003) Arsenic in drinking water. Background document for the development of the
WHO guidelines for drinking-water, Geneva:pp. 18
Wilkie JA, Hering, J.A. (1996) Adsorption of arsenic onto hydrous ferric oxide: effects of
adsorbate/adsorbent ratios and co-occurring solutes. Colloids and Surfaces A:
Physicochemical and Engineering Aspects 107:97-110
Winkel L, Berg, M., Amini, M., Hug, S.J., and Johnson, A. (2008) Predicting groundwater
arsenic contamination in Southeast Asia from surface parameters. Nature Geoscience
1:536–542
Winkel LHE, Trang, P.T.K., Lan, V.M., Stengel, C., Amini, M., Ha, N.T., Viet, P.H., Berg, M.
(2011) Arsenic pollution of groundwater in Vietnam exacerbated by deep aquifer
exploitation for more than a century. PNAS 108 (4):1246-1251

- 155 -
References

Winston RB (2000) Graphical User Interface for MODFLOW. Version 4: US Geological

Survey Open-File Report 00-315: pp. 27
http://waterusgsgov/nrp/gwsoftware/GW_Chart/GW_Charthtml, Version 12100
Wolthers M, Charlet, L., van der Weiden, C.H., van der Linde, O.R., Rickard, D. (2005)
Arsenic mobility in the ambient sulfidic environment: Sorption of arsenic(V) and
arsenic(III) onto disordered mackinawite. Geochimica et Cosmochimica Acta 69
(14):3483-3492
Xu Y-H, Nakajima, T., Ohki, A. (2002) Adsorption and removal of arsenic(V) from drinking
water by aluminum-loaded Shirasu-zeolite. Journal of Hazardous Materials B92:275-
287
Yamamoto M, Urata, K., Murashige, K., Yamamoto, Y. (1981) Differential determination of
arsenic(III) and arsenic(V), and antimony(III) and antimony(V) by hydride generation-
atomic absorption spectrophotometry, and its application to the determination of these
species in sea water. Spectrochm Acta 36B:671–677
Yu GQ, Sun, D.J., Zheng, Y. (2007) Health effects of exposure to natural arsenic in
groundwater and coal in China: An overview of occurrence. Environ Health Perspect
115:636-642
Zhu X, Zhu, L., Duan, Z.,Qi, R., Li, Y., Lang, Y. (2008) Comparative toxicity of several metal
oxide nanoparticle aqueous suspensions to zebrafish (Danio rerio) early developmental
stage. J Environ Sci Health, Part a: Toxic/Hazard Subst 43 (3):278-284
Zobrist L, Dowdle, P.R., Davis, J.A., Oremland, R.S. (2000) Mobilization of arsenite by
dissimilatory reduction of adsorbed arsenate. Environmental Science and Technology
34:4747-4753

- 156 -
Acknowledgements

- 157 -
Acknowledgements

First of all, I would like to thank my supervisors for giving me the independence in my work to
develop at my pace and for their support in much more than scientific questions this big project
was bringing along. This thesis is the result of a collaboration between four different research
institutions in France, Romania and Switzerland, and the different working styles, point of
views, and scientific as well as cultural backgrounds made this PhD a very enriching
experience. The resulting range of expertise and opinions in the different working groups have
immensely progressed this work. I am very grateful to my supervisors to always have searched
together for solutions and for your availability when I mostly needed it. Particularly, I want to
thank Jérôme for all the helpful discussions, introducing me to the nanoparticle topic, and the
invaluable help in all the synchrotron work and EXAFS analysis, and Calin for taking care of
my housing and administration during my stay in Romania, the immense organisation for all
the field trips, and the geological and geochemical feedbacks. Ein großes „Merci vielmals“
goes to Bernhard for giving me the opportunity to do this project, his availability in all
management questions, and invaluable scientific inputs to basically everything, and to Geri, for
his everlasting patience in teaching me chemical principles, for his support in the chemical lab
work, and for his confidence in me.
In addition to my supervisors, this work profited from the contributions of many other people:
Helen Rowland was a driving force in field work, and I learned a lot about water sampling and
analysis, water treatment setup, and British expressions from her. Without the big help of
Jacky Drese and Anna Diem in the production and characterization of aluminum granulate I
would have never made it to the 1.7 kg of material, it was just enough for this project!
Discussions with Annette Johnson and Michael Berg were always a great source of ideas in the
setup of lab and field experiments. I am very grateful to Stephan Hug for invaluable
discussions and help in sample analysis, and for his direct and cheering-up email response
during water treatment monitoring in Romania. I thank Armand Masion for very helpful
explanations and comments, and 27Al NMR measurements, and Rosemarie Pöthig for her
availability on short notice, her efforts to measure my samples, and to come personally to
Zurich for data discussions. It was a great pleasure to work with Ralf Kägi, who is thanked for
all the motivating discussions and the measurements on scanning electron microscopy. Perrine
Chaurand is gratefully acknowledged for introducing me to
XRF, continuing measurements
when I left CEREGE, and for kindly organizing my personal lab space when I arrived. I would
not have been able to do the synchrotron run without the reliable help of Naresh Kumar and
Emmanuel Doelsch, or the synchrotron analysis without the helpful comments from Mélanie
Auffan.

- 158 -
 Acknowledgements

I am indebted to all the people that helped me with laboratory work and their technical
expertise: Thomas Rüttimann for showing me the HG AFS, ICP-MS, and ICP-OES at Eawag,
Hermann Mönch and Vanessa Sternitzke for support on BET measurements, the granulate
production process and column experiments at Eawag, Kurt Barmettler and Björn Studer for all
material support and ICP-OES analysis at ETH, Caroline Stengel for her support in
interlaboratory quality comparison, Hélène Miche for her help in ICP-OES measurements at
CEREGE, Bela Abraham and all his lab staff from ICIA for repairing my pH electrode and
taking great care of all my laboratory needs during my time in Cluj, Michael Plötze and his
group at IGT for XRD, BET, density and porosity measurements. During field work and
administration in Romania translations from Bogdan Rentea and Cristi Pop were extremely
helpful.
I had the opportunity to interact with scientists from a broad range of disciplines within the
European project AquaTRAIN, which I wish to thank for all inputs and support, particularly
the coordinator, David Polya for steering the train over the hilly road. The good group of ESRs
and ERs was very sustaining and it was always fun to hang out at meetings with Claudia,
Chris, Adrien, Babak, Enoma, Helen, Aimee, Peter, Naresh, Geerke, Geoffrey, Julia, Tony,
Catherine. I am exceptionally grateful for the good collaboration with Cristina and Barbara and
for the very supportive dinners with my mentor Lenny. The science chats in the streets or over
petit-déjeuner in the best bakery of Aix-en-Provence with Raoul were very motivating, and
working together was always a great pleasure.
I highly appreciated all the help I received from other PhD students from ETH and Eawag. I
exceptionally want to thank Vanessa for all the immediate help in column and batch
experiments and countless discussions, and Linda for actually motivating me for this topic and
for always being available in questions about arsenic speciation.
I want to express my greatest gratitute to all the people that made my life easier and socially
comfortable in the different places I have been within these moving PhD years: Merci à Lise,
Vincent, Eva, Alice, Rémi, Nicolas, Julien, Blanche, Marie-Ange, Marie et la groupe des fous
aventeurs Vincent, Nico, Barbara, Jeanne, Victor Hugo, Romain, Aurélie, Angie, Pierre-Yves,
Valérie qui m’ont montré toutes les coigns lointains et jolies autour d’Aix-en-Provence en
escaladant, en baladant, à pied ou en canoo. Raluca, multumesc frumos pentru îmi aratasc
istorie si cultura din orasul tau! I am extremely grateful to the family Grec in Sepreus for
letting me live in their house for one month during my field work in Romania, where I always
found a warming fire in my room and the greatest ciorba when I came home.

- 159 -
Acknowledgements

The good atmosphere in the group of Aquatic Chemistry has immensely contributed to
enjoying my work. Most of all sharing ups and downs, chocolate, coffee, and many
crosscultural dinners with my dear office mates Simone, Tonya, Roland, Nuttakan, Nanina,
Jan, and Stefan, has made the last four years an extraordinary experience. I am very thankful to
Dave for his great interest in my work, our discussions, and his neverending very useful ideas.
The great pool of friends that I found at ETH and Eawag during the last four years created an
extremely inspiring and active environment, many thanks to Fabio, Annemarie, Claire, Yama,
Matthias, Marie-Eve, Dörthe, Krista, Manu, Kay, Ilaria, Charlotte, Laurie, Beate, Anke, Yves,
Peggy, Gabriele, Anna, Alex, Cristina, Wicher, Nima, Golnoosh, Mukta, Valeria, Angela,
Andi, Monica, Rainer, Nazanin, Hector, Michaelis, Bernd, Anja, Claudia, Kerstin, Ahmad,
Erika, Julia, Julie, Benoît, Luis.
Ich möchte mich vor allem bei meinen Eltern und meiner ganzen Familie für die große und
langjährige Unterstützung bedanken, die mich besonders in den letzten paar Monaten durch
liebe Briefe, Gespräche und Emails weiter getragen hat. Last but not least, sono molto grata a
Stefano per avermi risollevata piú di una volta negli ultimi anni. In particolare, per la sicurezza
che mi ha dato durante l'anno nel quale siamo stati separati e per il suo aiuto nel miglioramento
dei processi di modellazione. E alla fine la sua frase di incoraggiamento: "Sú, é quasi finita!" é
diventata realtá.

This PhD thesis is a contribution of the AquaTRAIN Marie-Curie Research training Network
(Contract No. MRTN-CT-2006-035420) funded under the European commission Sixth
Framework Programme (2002-2006) Marie Curie Actions, Human Resources & Mobility
Activity Area – Research Training Networks.
This work was further supported by the Swiss Federal Institute of Aquatic Science and
Technology (Eawag).

- 160 -
 Curriculum Vitae
Jasmin M. Mertens
Born on 1st November, 1983 in Aachen, Germany

Education

09/2007– 08/2011 PhD, ETH Zurich, Switzerland

PhD Thesis: Treatment of arsenic-contaminated water with
polyaluminum nanoclusters

09/2005 – 08/2007 Diplom Geology, RWTH Aachen, Germany

Specializations: organic geochemistry, hydrogeology, sedimentology
Diplom thesis: The long-term fate of organic contaminants in waste
and corresponding seepage water from a deposit landfill: a compound
specific prognosis.

08/2004 – 08/2005 Study abroad, Dept. of Earth and Planetary Science, University
of Hawai'i at Manoa, USA
Specializations: hydrogeology, coastal sedimentology, Matlab

09/2001 – 07/2004 Vordiplom Geology, RWTH Aachen, Germany

Work Experience
03/2010-08/2010 Visiting researcher, European center for environmental
research/ Centre National de la recherche scientifique
(CEREGE/ CNRS), Aix-en-Provence, France
Application of X-ray-based techniques (XRF, EXAFS, XANES) for
the study of intermolecular processes

09/2009 - 02/2010 Visiting researcher, University Babes-Bolyai (UBB), Cluj-

Napoca, Romania
Pilot-scale water treatment planning, installation and monitoring at
an artesian village well

02/2007 – 04/2007 Trainee, Swiss Federal Institute of Aquatic Science and

Technology (Eawag), Switzerland
Development of the design and setup for the extraction of volatile
organic compounds from the water phase at low concentrations in the
section of “water resources and drinking water”

Academic Distinctions
2008 Award for outstanding oral presentation at CEECHE

2007 - 2010 Marie-Curie Fellowship in the EU AquaTRAIN Programme

2004 - 2005 German Academic Exchange Service (DAAD) Scholarship

- 161 -
Publications
Mertens J, Casentini B, Masion, A, Pöthig R, Wehrli B, Furrer G (2012) Polyaluminum
chloride with high Al30 content as removal agent for arsenic-contaminated well water,
Water Research, 46, 53-62

Mertens J (2011) Al nanoclusters in coagulants and granulates: application in arsenic

removal from water. Reviews in Environmental Science and Biotechnology, 10(2), 111-
117

Rowland HAL, Omoregie E, Millot R, Jimenez C, Mertens J, Baciu C, Hug SJ, Berg M
(2011) Geochemistry and arsenic behaviour in groundwater resources of the Pannonian
Basin (Hungary and Romania). Applied Geochemistry 26, 1-17.

Selected conference presentations

Jasmin Mertens, Jérôme Rose, Perrine Chaurand, Ralf Kägi, Benrhard Wehrli, and Gerhard
Furrer (2011) Removal of arsenic from well water by Al nanoclusters, International
conference on chemistry in the environment (ICCE), 11–14 September 2011, Zurich,
Switzerland

Jasmin Mertens, Armand Masion, Rosemarie Pöthig, Bernhard Wehrli, and Gerhard Furrer
(2011) Al nanoclusters: application in water treatment and implications for arsenic
removal, IWA specialist conference on applications of nanotechnology in the water
sector, 15-18 May 2011, Ascona, Switzerland

Jasmin Mertens, Helen A. L. Rowland, Cristina Jimenez, Calin Baciu, Stephan Hug,
Bernhard Wehrli, and Gerhard Furrer (2010) Removal of As from drinking water by
polynuclear aluminium, AquaTRAIN International conference on Geogenic Chemicals
in Groundwater and Soils, 7-8 June 2010, BRGM, Orleans, France

Jasmin Mertens, Helen A. L. Rowland, Cristina Jimenez, Calin Baciu, Stephan Hug,
Bernhard Wehrli, and Gerhard Furrer (2009) Al nanoclusters: A new method for As
removal from water. Goldschmidt, 21-26 June 2009, Davos, Switzerland

Jasmin Mertens, Bernhard Wehrli, Gerhard Furrer (2009) Removal of arsenic from drinking
water with Al nanoclusters. Workshop on integrated remediation approaches: modeling
and monitoring in combination with remediation, 20-21 April 2009, VITO NV,
UGhent, Antwerp, Belgium

Jasmin Mertens, Helen A. L. Rowland, Cristina Jimenez, Bernhard Wehrli, Calin Baciu,
Stephan Hug, Gerhard Furrer (2008) As contamination of groundwaters in the Tisza
River Basin: Treatment of well water with Al nanoclusters. The 3rd Central and Eastern
European Conference on Health and the Environment (CEECHE), 19-22 November
2008, Cluj-Napoca, Romania.

- 162 -