55901413

Development of methods based on ICP-mass spectrometry
for the determination, speciation and isotopic analysis of

metals and oxy-anions in an environmental context
Kristof Tirez
FeCr2O4
CdNO3
0
Hg
11
δ B
PbCrO4
2-
SeO4
112 +
BrO3
- 10 + Cd
B
V Fe Mo Mn As
Mg
Zn Ba 78
Se+
Ni Na
Cl 52
Cr
+
Co
Development of methods based on ICP-mass spectrometry
for the determination, speciation and isotopic analysis of
metals and oxy-anions in an environmental context
Ghent University, Faculty of Sciences, Department of Analytical Chemistry, Atomic

and Mass Spectrometry
Flemish Institute for Technological Research (VITO)
Doctoral dissertation to meet the requirements to take the doctoral exam

Doctor of Science: Chemistry
by
Kristof Tirez
Promoter: Prof. Dr. Frank Vanhaecke

Co-promoter: Dr. Nicole De Brucker
Ghent, 2013
Acknowledgements
This research would not have been possible without the support of a lot of people. I wish to thank
several people who accompanied and guided me along the route.
After a few years, it soon became clear to my parents that no talented sportsman or writer was
created, but rather a playful scientist with a strong tendency to tease his brothers and sister. I started
my master in chemistry almost by accident, but by the time I graduated as environmental chemist, I
had found my interest. My parents supported me all along this learning process, while brothers and
sister underwent my practical jokes. Many thanks!
Herman De Schepper (†) (Flemish Environmental Agency) and Prof. Richard Dams (Institute for
Nuclear Sciences, UGhent) introduced me into environmental chemistry. Under their supervision, I
performed my master of science in environmental sanitation thesis in the laboratory of the Flemish
Environmental Agency, headed by Dirk Quaghebeur (†). During this thesis, Rudy Cautaerts and Ingrid
Temmerman taught me the importance of quality control analysis in a laboratory environment and
the practical skills were instructed to me by Gerda Degent. I remember the good atmosphere and
collegiality in the laboratory and I am grateful I could be part of it. Many thanks!
The first steps in my professional research career were performed in 1996 in the laboratory of Prof. L.
Hosten (†) (Laboratory for Chemical Technology, UGhent) and under the auspices of the same
persons from the Flemish Environmental Agency mentioned before. After the sudden demise of
Herman De Schepper, Rudy Cautaerts became mentor of the project and he introduced me further in
the professional aspects of environmental related issues. I was given the opportunity to learn and
work together with passionate colleagues of the Flemish Environmental Agency dealing with the
implementation of environmental regulations. Many thanks!
In 1997, a call for a PhD dissertation dealing with the use of ICP-MS in environmental related issues
was announced at VITO. A PhD proposal was written in co-operation with Nicole De Brucker (VITO),
Prof. Frank Vanhaecke and Prof. Luc Moens (Department of Analytical Chemistry, UGhent).
Unfortunately, the project was not awarded. However, I did get a position of researcher at the
analytical laboratory of VITO. Nicole De Brucker, Pierre Geuzens and Theo Rymen encouraged me to
develop methods based on ICP-MS for the determination of elements in an environmental context at
VITO and I want to thank them for giving me this opportunity.
In 2000, I became responsible for the inorganic analytical team and I’m (still) very grateful to work
together with these experienced laboratory technicians and devoted researchers: Filip Beutels,
Wilfried Brusten, Jos Broeckx, Suzy Bleyen, Annick Cluyts, Valère Corthouts, Nicole De Brucker, Jef De
Wit, Karlien Duyssens, Ludwig Goetelen, Bart Noten, Ellen Poelmans, Peter Van Bree, Chris Van Hoof,
Mai Wevers and Wendy Wouters. Many thanks!
Since then, I have been involved in numerous environmental related studies giving me the
opportunity to work together with a lot of VITO colleagues and external researchers. The list is long,
and will probably be incomplete if I had to sum up all the professional contacts I have to
acknowledge. Therefore, I will acknowledge the co-authors of the papers as representation for all the
scientific collaborations: Ludo Diels, Sandra Van Roy, Johan Patyn, Holger Scharf, Domenico Calzolari,
Rob Cleven, Monika Kisser, Detlef Lück, Geert Silversmit, Nico Bleux, Elke Adriaenssens, Edward
Roekens, Kelly Servaes, Laszlo Vincze, Patrick Berghmans, Petru Jitaru, Piet Seuntjens, Michael
Berglund, Philip Taylor, David Widory, Emmanuelle Petelet, Agnès Bregnot, Dongmei Xue, Pascal
Boeckx, Jan Bronders, Nele Desmet and Guy Vanermen. Many thanks!
I have always considered team work as a stimulating habitat and I am very grateful to the VITO
management and directors that I got the opportunity to do my professional work with a large input
of my own personality, knowledge and creativity.
Thanks to the European Winter Conferences on Plasma Spectrochemistry and the hospitality of Frank
Vanhaecke, I became integrated in the ICP-MS community. I felt at ease in the relativising scientific
community and remember with great pleasure also the company of Stefaan Van Winckel, Lieve
Balcaen, Prof. Martin Resano and Prof. Jorge Ruiz Encinar at these occasions.
The sporadic encounters with Frank Vanhaecke have finally led to the idea of writing a PhD
dissertation on the research I have been involved so far. I am very grateful to Frank and Nicole, my
promoters, for giving me the time and support needed to accomplish it.
For providing me analytical data of regulatory monitoring studies of trace elements in the Flemish
environment, I would like to thank the following persons from the Flemish Environmental Agency:
Martin Verdievel, Henk Maeckelberghe, Johan Annys, Ilse Theuns, Ward De Cooman, Koen Toté, Elke
Adriaenssens, Bo Van den Bril, Jordy Vercauteren, Ywan De Jonghe and Ralf Eppinger.
The reported analytical work in my PhD dissertation has been performed by the “heavy metal” team
at VITO, which consists (besides me) of Wilfried Brusten, Filip Beutels, Karlien Duyssens and Chris
Vanhoof. The team works on the basis of an intensive theory-practice interaction, i.e., theory is when
you (think to) know everything but nothing works, practice is when everything works but nobody
knows why. Many thanks!
I wish to thank the Public Waste Agency of Flanders (OVAM), in particular Luc Debaene and
Dominique Suys, the Flemish Environmental Agency (VMM), in particular Ward Roekens and Elke
Adriaenssens, the Belgian Federal Public Service (FOD) of Health, Food chain safety and Environment,
in particular Robert Martens, the Environment, Nature and Energy Department (LNE) and the
Institute for Reference Materials and Measurements (IRMM), in particular Thomas Linsinger, for
funding some of the research projects mentioned in this PhD dissertation.
At this point, I also would like to thank my wife, Katrien, for loving and supporting me, for sharing the
good and the lesser moments in my life, for help guiding our children, Nathan, Arne and Charlotte to
their own future and to make us all feel at home. A woman seems always to have one more child
extra in the family to educate, since a man does what he can, but a woman does what a man cannot
(Isabel Allende) ... I cannot thank her enough.
I thank my kids for distracting me from the moment I open my eyes (even sometimes before) and to
incite me to go through the full spectrum of possible emotions on a daily basis.
Finally, I wish to thank the reader, for your interest in my work.
Kristof, december 2013

Scope and contents of the PhD dissertation
“Where is the wisdom we have lost in knowledge?

Where is the knowledge we have lost in information?”
T. S. Eliot (Nobel Prize in Literature, 1948)
In this PhD dissertation, the development of methods based on inductively coupled plasma-mass
spectrometry (ICP-MS) for the determination, speciation and isotopic analysis of metals and oxy-
anions in an environmental context is discussed.
What about heavy metals? Although the use of the term “heavy metals” has been criticized as
ambiguous and pointless, this has not banned the term itself in scientific literature and legislation.
The oldest recorded scientific use dates back to 1936 (related to elemental density), but today, this
term represents an ill-defined umbrella term for various elements (primarily transition metals, but
also some non-metals). In the dissertation, a pragmatic approach is followed on how to deal with the
term “heavy metals”, i.e. the term is frequently used in legislation without a widely accepted
scientific rationale and is simply defined as a set of elements. However, preference will be given to
the use of the terms elements (trace and major), metals and oxy-anions.
The success story of ICP-MS in regulatory environmental monitoring of elements was written in the
stars with the adage: to measure is to know. With its multi-element capability (determination of 80
elements in 3 minutes), excellent sensitivity (µg/L, ng/L), combined with linearity over several orders
of magnitude, ICP-MS was embraced by the environmental analytical community in the 1990s and
this only 10 years after the first description of the ICP-MS technique [1,2].
Around the turn of the millennium, ICP-MS instruments began their march in the Flemish commercial
environmental laboratories and were mainly deployed for the analysis of water and thereafter, for
the analysis of solid matrices. At this moment and with the current (European) regulatory
environmental limits in force, ICP-MS is increasingly being considered as the method of choice. This
market demand resulted in its turn in the development of ICP-MS units as reliable (limited
downtime), robust (no/limited interference from matrix) and high-throughput instrumentation,
ideally suited for research purposes and routine analysis. As an example of the latter, the cost price
on the international market for the determination of 51 elements with ICP-MS in geological survey
after aqua regia digestion amounts to 16,4 € only [3].
With high-throughput analysis and an almost exponentially growing amount of measurement data,
the adage of the regulatory environmental community slightly started to change to: to measure is to
know … but do we understand? Traditionally, inorganic chemical analysis has been used to
investigate the presence (qualitative) and concentration (quantitative) of (trace) elements in
environmental matrices. However, total elemental determination does not provide information on
the physicochemical characteristics, biological activity, toxicity, mobility and bioavailability of
elements in these samples, because such information can only be derived through the determination
of chemical species. This triggered the development of the so-called hyphenated ICP-MS techniques.
Separation techniques such as liquid or gas chromatography were hyphened (or interfaced) with ICP-
MS instruments, making it possible to quantify one or more individual chemical species. This market
demand resulted in the development of commercial interfaces to easily hyphen different separation
techniques (primarily GC, LC) to ICP-MS over the last 10 years.
In summary, since its introduction 30 years ago, ICP-MS conquered a large part of the market of
environmental monitoring of metals and oxy-anions. Riding on the waves of this analytical success
story and starting 15 years ago, the research described in this dissertation is in essence related to the
environmental monitoring adage to measure is to know … but do we understand? This understanding
is both related to the analytical measurement itself, as to the environmental processes studied and
at a given point, this also led sometimes to the insight that “all I know is that I know nothing” (or at
most a subtle part of it).
The year 1980 is not only considered as the starting point of the history of ICP-MS, but it is also the
year in which the environmental policy in Belgium was transferred from the federal level to the
regions (Flemish, Walloon and Brussels Capital Region). Since then, different Flemish environmental
agencies were founded, rolling out the implementation and enforcement of the environmental
legislation related to heavy metals in Flanders. Also in 1980, Lars Ulrich posted an advertisement to
start up the metal band Metallica, a landmark in heavy metal music.
As an introduction to the actual experimental work carried out and described in this PhD
dissertation, the synchronicity in the emergence of environmental legislation on the one hand and
that of ICP-MS in environmental analysis since 1980 on the other will be discussed.
Thereafter, and as part of the to measure is to know adage, examples will be given of regulatory
monitoring in the field of water, air and soil, where the superior sensitivity combined with multi-
element capabilities of ICP-MS have played - and still play - an important role. This is especially the
case for the derivation of ambient background concentrations of elements present in the different
environmental compartments. As metals and oxy-anions are naturally present in our environment,
knowledge of ambient background concentrations is needed to provide baseline data for pollution
and risk assessment studies. The range of background concentration of elements, derived and
compiled in this PhD dissertation, will illustrate a match made in heaven with the sensitivity of ICP-
MS instruments.
The focus of the first part of the dissertation is on the analysis (of acid digests) of environmental
samples by ICP-MS to achieve a multi-elemental determination of elements. Special attention is
given to the impact of European harmonisation efforts in order to prevent environmental pollution
“to stop” at national (or regional) borders.
Subsequently, the scientific papers covering the development of methods based on ICP-MS for the
determination, speciation and isotopic analysis of metals and oxy-anions in an environmental context
are presented with emphasis on the connecting thread of the PhD dissertation, i.e., the relation
between the ICP-MS measurement, the environmental issue and the regulatory context.
The scientific papers have been arranged chronologically in the following chapters (see also
timeline):
• Speciation-Determination of oxy-anions in water and solid samples

o Characterization of inorganic selenium species by ion chromatography with ICP-MS
detection in microbial-treated industrial waste water
o Determination of hexavalent chromium by species-specific isotope dilution mass
spectrometry and ion chromatography-1,5-diphenylcarbazide spectrophotometry
o Validation of a European Standard for the determination of hexavalent chromium in solid
waste material
o Determination of hexavalent chromium in ambient air: a story of method-induced Cr(III)
oxidation?
o Determination of bromate in drinking waters using low pressure liquid chromatography /
ICP-MS
• Tracer experiments with stable isotopes in environmental studies and use of natural isotopic
variation as a proxy
o Full uncertainty calculation on quantitative determination of tracer and natural cadmium
in soil column effluents with ICP-MS
o Boron isotope ratio measurements in WFD monitoring programs: comparison between
double-focusing sector-field ICP (ICP-SFMS) and thermal ionization mass spectrometry
(TIMS)
(2013) Determination of
(2003) Panoramic analysis for monitoring bromate in drinking water
trace metals in natural waters by low pressure LC/ICP-MS
hexavalent chromium by species- (2011) Ultratrace isotopic
specific isotope dilution mass (2003) Determination of analysis of uranium in
spectrometry organotin species environmental and biological
samples
(2000) Determination (2001) Determination

of bromate in drinking of selenium species in (201O) Boron isotope
water waste water ratio in groundwater
(1999) Study of (2011) Determination of

tracer cadmium (2001) Study of uncertainty (2007) Study of
hexavalent chromium in
transport in soil contributions in tracer cadmium hexavalent chromium in
particulate matter
transport in soil soil and waste material
1997 2000 2005 2010
Timeline (1997 – 2013) showing the chronological order of the development of methods based on
inductively coupled plasma-mass spectrometry (ICP-MS) for the determination, speciation and
isotopic analysis of metals and oxy-anions in an environmental context.
References :
A numerical style of references (i.e., [14]) is used throughout the dissertation, the reference list
starts at p. 226. Except for the scientific papers, for which the reference lists are separately
provided at the end of each paper.
Table of Contents
CHAPTER 1 Synchronicity in the emergence of the Flemish environmental legislation and ICP-MS
in environmental analysis since 1980 ________________________________________________ 1
1.1. Some milestones in the environmental legislation of heavy metals in Flanders 1
1.2. Who is implementing the environmental policy in Flanders? 4
1.3. Which methods are used for regulatory environmental monitoring of heavy metals? 7
1.4. The atomic spectrometry timeline in environmental analysis 12
1.4.1. ICP-MS instrumentation ______________________________________________ 12
1.4.2. Atomic spectrometry timeline _________________________________________ 20
1.4.3. Acid digestion of environmental matrices ________________________________ 25
CHAPTER 2 ICP-MS applications in environmental regulatory monitoring ________________ 27

2.1. Perspective on elemental analysis 27
2.2. Ambient background concentration of elements in soil in Flanders 31
2.3. Ambient background concentration of elements in ambient air in Flanders 40
2.3.1. Ambient background concentration of elements in fine particulates in Flanders __ 40
2.3.2. Ambient background concentration of elements in atmospheric deposition _____ 45
2.4. Ambient background concentration of elements in river sediment in Flanders 50
2.5. Ambient background concentration of elements in surface and ground water in Flanders55
2.5.1. Ambient background concentration of elements in surface water in Flanders ____ 56
2.5.2. Ambient background concentration of elements in ground water in Flanders ____ 62
2.6. Summary of ambient background concentrations of elements in Flanders 66
2.7. Summary and conclusion 67
CHAPTER 3 Elemental speciation – Determination of oxy-anions in water and solid samples 69

3.1. Perspective on elemental speciation 69
3.2. Removal of selenate and selenite from waste water 73
3.3. The ban on hexavalent chromium 86
3.3.1. The ban on hexavalent chromium in packaging material _____________________ 87
3.3.2. The ban on hexavalent chromium in electronic waste (RoHS and WEEE) _______ 108
3.3.3. Hexavalent chromium in ambient air ___________________________________ 129
3.4. Oxyhalide by-products in drinking water disinfection 152
CHAPTER 4 Tracer experiments with stable isotopes in environmental studies and use of natural
isotopic variation as a proxy _____________________________________________________ 169
4.1. Perspective on isotopic tracers 169
4.2. Metal leaching from soils contaminated by historic smelter emissions 172
4.3. Sources of nitrate leaching 194
CHAPTER 5 Summary and conclusions ___________________________________________ 220

Synchronicity in the emergence of the Flemish environmental legislation and ICP-MS in environmental
1
analysis since 1980
CHAPTER 1 SYNCHRONICITY IN THE EMERGENCE OF THE FLEMISH

ENVIRONMENTAL LEGISLATION AND ICP-MS IN ENVIRONMENTAL ANALYSIS
SINCE 1980
In the following section, the synchronicity in the emergence of the Flemish and European
environmental legislation dedicated to heavy metals and the emergence of ICP-MS in environmental
analysis since 1980 is described. In addition to a timeline for and some milestones in both topics, also
the interrelation between ICP-MS and (Flemish) environmental legislation is discussed.
1.1. SOME MILESTONES IN THE ENVIRONMENTAL LEGISLATION OF HEAVY METALS IN FLANDERS
“Why has it seemed that the only way to protect the environment is with heavy-handed government
regulation?” Gale Norton (United States Secretary of the Interior, 2001-2006)
The first legislation in Flanders aimed at protecting the environment against the harmful effects of
polluting industries (unhealthy or offensive odour), the oldest "Environmental Law" was established
by the Imperial Decree of October 15th, 1810 [4,5].
By decision of the Regent on February 11th, 1946, Titles I and II of the General Regulations for Labour
(ARAB) were established. The ARAB bundled various laws and decrees relating to occupational safety,
occupational hygiene and environmental protection in a code. Title I provided a classification of
hazardous, unhealthy or nuisance-causing establishments subjected to an environmental license
requirement [4].
In 1977, Belgium got a first written warning from the European Union for failing to implement a
European directive dating from 1975 on waste [6]. This delay was caused by the Belgian state reform
that was in progress. With the Institutional Reform Act of August 8th, 1980, the management of the
environment in Belgium became a task of the regions (Flemish, Walloon and Brussels Capital Region).
The regions were therefore entitled to their own devices to implement a policy on environmental
nuisance [5]. 1980 is also the starting point of the timeline in the emergence of the Flemish
environmental legislation, as shown in Figure 1.
Shortly after the Belgian state reform in 1980 and under pressure of the proliferation of landfills, the
actions of the environmental movement and the exhortations of the European Court, the issue of
waste is discussed on one of the first meetings of the Flemish Council [6]. On July 2nd, 1981, the
decree "concerning the prevention and management of waste-materials" - simply ”Decree on
Waste”- sees the light of day. The Flemish waste policy objectives are: 1) waste prevention, 2)
promotion of reuse, recovery and recycling, 3) organization of the disposal of generated waste. To
reach these goals, the Public Waste Agency of Flanders (OVAM) was founded. In 1982, the Belgian
Nuclear Research Centre (SCK) starts to act as reference laboratory for OVAM.
A new policy on nuisance-causing establishments was announced with the Decree of the Flemish
Council of June 28th, 1985, concerning Environmental Licenses [5]. This environmental decree
entered into force on September 1st, 1991, with the first implementation order: the Order of the
Flemish Government of February 6th, 1991, establishing the Flemish regulations concerning
2 analysis since 1980
environmental Licenses (Title I of VLAREM). This order deals with the various procedures and the list
of establishments and activities that are classified as nuisance.
In 1991, the European Commission (EC) issued a nitrate directive (91/676/EEC) concerning the
protection of water against contamination by nitrates from agricultural sources [7]. To meet the EC
directive, the application of manure and fertilizers has been restricted in Flanders by the Decree on
Manure (1991). The Flemish Land Agency (VLM), founded in 1988, helps farmers to reach these
goals.
The Flemish environmental policy was extremely fragmented and to rationalize and simplify the
Environmental law, an "Interuniversity Commission Review of the Environmental Law" was
established in 1989 (the so-called “Bocken Commission”)[6]. In 1994, this Commission proposes a
report: all sub-regulations should be integrated into a system of licensing and this initiated the
Flemish environmental code.
A subsequent implementation order, the Order of the Flemish Government of June 1st, 1995,
concerning General and Sectoral provisions relating to Environmental Safety (Title II of VLAREM)
included the environmental conditions under which an establishment may be operated [4]. The use
of Best Available Techniques (BAT) and the setting of environmental quality standards were the
starting points in the Flemish environmental regulation.
The Flemish Institute for Technological Research (VITO), founded in 1991, takes over the role of
reference laboratory for OVAM from the Belgian Nuclear Research Centre (SCK) with respect to the
disciplines soil and waste. In 1994 VITO takes over the role of reference laboratory from the Belgian
Institute of Hygiene and Epidemiology (IHE) for the disciplines water and air.
On February 22nd, 1995, the Flemish Parliament approved the Decree for soil remediation [6]. This
soil remediation legislation had also been prepared by the Interuniversity Commission for the
Revision of Environmental Law (“Bocken Commission”). A year later, the implementation decisions
were bundled in the Order of the Flemish Government of March 5th, 1996, concerning the
establishment of the Flemish regulations on Soil Remediation (VLAREBO). These texts created a legal
framework to address soil problems in Flanders. In VLAREBO, standard values were defined for
background values (unpolluted soils) and for soil remediation values, depending on the destination
type of the soil (housing, industry). The most severe soil remediation values form a balance between
a limited enrichment relative to the natural background composition of the soil on the one hand and
the maximum protection of human health and the ecosystem from direct exposure on the other.
On December 17th, 1997, the Flemish Government approved the Decision establishing the Flemish
regulations relating to waste prevention and management (VLAREA)[5]. VLAREA introduced the
concept of secondary raw material for several applications, e.g., as fertilizer, soil conditioner,
construction material or soil. It is assumed that the main impact to the environment by using
secondary raw materials is caused by the leaching of inorganic contaminants (e.g., heavy metals) into
the underlying soil and groundwater. The basic principle is the “marginal soil load”, i.e., a very limited
enrichment of the underlying soil layer by inorganic contaminants.
In 1998, the European Commission (EC) issued Council Directive 98/83/EC on the quality of water
intended for human consumption (adopted as Decision of the Flemish Government in 2002)[8,11].
The directive is intended to protect human health by laying down healthiness and purity
requirements which must be met by drinking water within the European Community. About 340
billion litre of drinking water is produced yearly in Flanders by 15 drinking water suppliers [8]. The
quality of the drinking water in Flanders is controlled by the Flemish Environmental Agency (VMM)
in cooperation with the drinking water suppliers [9].
3
analysis since 1980
In the 1990s, the European Commission (EC) recognized the need for further action to avoid long
term deterioration of the quality and quantity of freshwater resources and it called for a program of
actions to be implemented by the year 2000, aiming at sustainable management and protection of
freshwater resources [10]. These considerations coincided with the request made by the European
institutions to the Commission to come forward with a proposal for a Directive, establishing a
framework for a European water policy; this resulted in the adoption of the Water Framework
Directive (WFD) 2000/60/EC on October 23rd, 2000. The Water Framework Directive has put forward
a challenging legislative framework, establishing "good status" environmental objectives for all
waters – surface, coastal, transitional, and ground waters – to be achieved by the end of 2015.
The EU regulatory Groundwater Framework on the protection of groundwater against pollution

caused by certain dangerous substances that has been developed at the end of the 1970s (Directive
80/68/EEC) will be repealed by December 21th, 2013, under the Water Framework Directive. In the
field of emission limit value legislation (waste water), the Dangerous Substances Directive
(2006/11/EC) and its Daughter Directives on various individual substances were adopted in 2006.
The Flemish soil remediation decree of 2007 replaced the old decree of 1995 [6]. The simpler decree
regulates soil certificates and remediation of contaminated soil. The decree of 1995 mainly had an
effect on companies in operation and on transfer of ownership of (contaminated) grounds.
Abandoned or unused areas that were no longer used because of soil contamination, called
brownfields, remained unaffected. The new decree of 2007 aims to protect the available green
spaces and to upgrade contaminated areas.
In 2008, The European commission adopted a new Air Quality Framework Directive 2008/50/EC, in
which most of the existing legislation related to air quality (including the first air quality framework
directive of 1996) was merged into a single directive with no change to existing air quality objectives
[13]. The Fourth Daughter Directive (Directive 2004/107/EC), including target values for arsenic,
cadmium and nickel in ambient air, entered into force in 2013.
In 2010, the Flemish Government approved, as part of the integral water policy, new environmental
quality targets for surface water and groundwater and environmental quality standards for
sediments and biota, based on the European Water Framework Directive (Directive 2008/105/EC).
This Directive marks an important step in the use of sediments and biota as matrices for assessing
long term trends and chemical status in the integral water policy [14]. In 2011, the associated
Directive 2009/90/EG of the Commission, laying down technical specifications for chemical analysis
and monitoring of water, was adopted as well (e.g., instead of total elemental concentrations,
dissolved elemental concentrations are to be determined in surface water).
In 2011, a decision of the Flemish Government establishing the regulations for recognition related to
Environmental monitoring (VLAREL) is adopted [4]. This decision provides regulations on the
recognition of professional qualifications and technical specifications for chemical analysis and
monitoring of the Environment. Herein, it is mentioned that the Flemish Institute for Technological
Research (VITO) is (re)designated as reference laboratory for Flanders.
In 2011, the Flemish Parliament approved the new Decree on Materials [6]. This represents the final
part of sustainable materials management in Flanders. It implements the European Framework
Directive (EC) 2008/98 for waste management and anchors the sustainable materials management
and replaces the waste decree of 1981. The decree requires that a comprehensive look at the
material chain is essential for a lasting solution to the issue of waste. The objectives of the new
legislation are: 1) environment-friendly and sustainable waste management, 2) reducing the harmful
effects of materials throughout the life cycle and 3) counteracting the depletion of resources.
Parallel to the decree, a new legislation that completely replaces the VLAREA, the Flemish regulations
for the sustainable management of material cycles and waste (VLAREMA), was approved in 2012.
This regulation not only includes the European end of waste approach, but also the successor of the
secondary raw material approach, which is now referred to as raw material. As the handling of waste
has an influence on soil and groundwater quality, a scientific and policy inter-correlation was needed
between the VLARE(M)A (waste) and VLAREBO (soil, groundwater) codes. New risk-based limits for
mineral materials were derived, based on safe concentrations in soil and groundwater [12]. Risk-
based limit values indicate that experimental data obtained using the leaching test method on the
mineral substance (waste), do not exceed safe concentrations in soil and ground water.
In summary, during the last 30 years, different Flemish regulations for the protection of our
environment have been implemented. The last years, this legislation was amended many times,
mostly under the influence of new European environmental legislation : Water Framework
Directive, Ground Water Directive, Air Quality Directive…. a notable absentee in this list is soil. The
soil directive has languished since it was first introduced in 2006. Some member states stymied the
proposed legislation, claiming it was not the EU’s prerogative to dictate policies on land use.
Also, a shift from “environmental protection” to “environmental sustainability” is being introduced

(i.e., Sustainable material management in Flanders). Implementing a more sustainable approach to
the consumption and production of chemicals would not only benefit Europe's environment, but also
reduce the detrimental effects arising in other parts of the world (see also § 3.3, e-waste).
Interesting links :
• The Navigator Environment, Nature and Energy, is a unique instrument that gives the updated
and consolidated version of the Flemish environmental legislation. A clear structure and a
sophisticated search engine allows finding the desired legislation
http://navigator.emis.vito.be/milnav-consult/
• EUR-Lex provides free access to European Union law and other documents considered to be
public. The website is available in the 23 official languages of the European Union : http://eur-
lex.europa.eu/en/tools/about.htm
o Council Directive 91/676/EEC of 12 December 1991 concerning the protection of waters
against pollution caused by nitrates from agricultural sources
o Council Directive 98/83/EC of 3 November 1998 on the quality of water intended for
human consumption
o Directive 2000/60/EC of the European Parliament and of the Council of 23 October 2000
establishing a framework for Community action in the field of water policy
o Directive 2008/50/EC of the European Parliament and of the Council of 21 May 2008 on
ambient air quality and cleaner air for Europe
o Directive 2008/105/EC of the European Parliament and of the Council of 16 December
2008 on environmental quality standards in the field of water policy
o Directive 2008/98/EC of the European Parliament and of the Council of 19 November
2008 on waste
1.2. WHO IS IMPLEMENTING THE ENVIRONMENTAL POLICY IN FLANDERS?
“Nothing is more destructive of respect for the government and the law of the land, than passing laws which
cannot be enforced.”(A. Einstein)
The Flemish Minister for Environment implements the environmental policy. For the preparation and
implementation of the policy, the minister can count on the collaboration of the Environment,
Nature and Energy Department (LNE) and a number of agencies from the Environment, Nature and
Energy policy domain:
5
analysis since 1980
• The Public Waste Agency of Flanders (OVAM) is responsible for waste management and soil
remediation in Flanders. It is a public Flemish Institution, established after the decree of July
2nd, 1981, covering waste management and prevention. Waste removal and soil remediation
were included as well. In 1995, Flanders got a more specific legislation on soil remediation:
the soil remediation decree. The soil remediation decree provides the Flemish government
with an instrument to fight historical as well as recent soil pollution. One of its objectives is
to remediate historical soil pollution within a period of 40 years. The soil remediation decree
equally offers a range of possibilities to prevent new soil pollution or to remediate right
away.
• The Flemish land agency (VLM), founded in 1988, is responsible for the organization and
management of the open space in Flanders. Furthermore, it contributes to shaping rural
policy in Flanders. The VLM contributes to meeting the environmental targets set in the
Nitrates Directive, by actively guiding the farmers, stimulating them to practice sustainable
agriculture and monitoring a correct implementation of the manure legislation. Its field of
activity comprises the rural areas and the peri-urban open space in Flanders.
• The mission of The Flemish Environment Agency (VMM), established in 1991, is to contribute
to the realization of the environmental policy objectives by reporting on the state of the
environment and by preventing, limiting and reversing harmful impacts on water systems
and pollution of the atmosphere, and to the realisation of the integrated water policy
objectives. VMM also produces the Flanders State of the Environment Report (MIRA). MIRA
describes, analyses and evaluates the state of the Flemish environment, discusses the
environmental policy pursued and looks ahead at the future environment.
(organizational chart courtesy of http://www.lne.be/organisatie/organogrammen)
Interesting links :
th
• A special website was devoted to the 30 anniversary of the Public Waste Agency (2011), with a
'virtual exhibition', a timeline, a collection with hundreds of photos, video clips and testimonials:
http://www.30jaarovam.be/index.php/
(1990) Administration
(1988) Flemish Land Environment, Nature,
(1982) SCK acts Agency (VLM) is Land and Water
(1980) Management as reference founded Management
laboratory for (AMINAL) is founded (2006) AMINAL is changed
of the environment
OVAM to the Environment,
in Belgium becomes
(1991) Flemish Institute Nature and Energy
a task for the regions
for Technological Research Department (LNE)
(VITO) is founded
(1985) First (1990) Flemish
Flemish Environment Agency
Minister for (VMM) is founded
(1981) Public Waste Environment (1994) Flemish Reference
Agency of Flanders laboratory for environmental
(OVAM) is founded analysis (VITO) is founded
1980 1985 1990 1995 2000 2005 2010
(1981) Decree concerning the (1996) VLAREBO : Order of the (2002) Decree of on the
Prevention and Management Flemish Government establishing the quality of water intended
of Waste-materials Flemish soil remediation and
(1991) VLAREM I : for human consumption
Order of the Flemish protection regulations
(1985) Decree of the Flemish Government
Council concerning concerning
Environmental (1995) VLAREM II : Order (1997) VLAREA : Order of the
Environmental Licences
Licences of the Flemish Flemish Government for the (2011) Decree concerning
Government concerning establishment of the flemish the sustainable
General and Sectoral management of material (2011) VLAREL : Decision of
(1991) Decree on the regulations relating to waste
provisions relating to cycles and waste the Flemish Government
protection of waters against prevention and management
Environmental Safety establishing the Flemish
pollution by nitrates from (1995) Decree on regulations for recognition
agricultural sources soil remediation related to Environmental
(2003) Decree relating to monitoring
integral water policy
(1995) Decree concerning
general provisions relating
to environmental policy
91/676/EEC 1996/62/EC Air Quality 98/83/EC on the quality of water 2000/60/EC Water 2008/98/EC Waste
Nitrate Directive Framework Directive intended for human consumption Framework Directive Framework Directive
Figure 1: Timeline (1980 – 2012) showing the founding of Flemish agencies related to environmental issues (upper part) and milestones of important
Flemish/European environmental legislation (lower part).
7
analysis since 1980
1.3. WHICH METHODS ARE USED FOR REGULATORY ENVIRONMENTAL MONITORING OF HEAVY METALS?
Given the significance of manmade dispersion of heavy metals in our environment, major legal acts
in the EU (e.g., Water Framework Directive, Air Quality Directive, see §1.1) have included monitoring
programs in order to report on the state of our environment. In this paragraph a brief overview will
be given on the decision-making process leading to the analytical methods to be used for regulatory
monitoring. The emphasis will be on the determination of heavy metals in general and, in particular,
on the use of ICP-MS (see § 1.4.1).
In the Decision of the Flemish Government of June 29th 1994 (and revised in 2011 in VLAREL), it is
mentioned herein that the Flemish Institute for Technological Research (VITO) acts as reference
laboratory for Flanders for the disciplines water, air, soil, waste and manure. Besides, this decision
provides regulations on the recognition of professional qualifications and technical specifications for
chemical analysis and monitoring of the Environment.
In support of the implementation of the environmental policy (see §1.2) and since 1994, VITO started
to provide reference methods for the analysis of the different environmental matrices:
• WAC: compendium for sampling and analysis of water

• LUC: compendium for sampling and analysis of air
• CMA: Compendium for sampling and analysis in support of the material and soil remediation
decrees
• BAM: compendium for sampling and analysis of manure, animal feed and soil in support of
the manure decree
• BOC: Compendium for sampling and analysis in the context of soil protection
The Compendia of Flemish reference methods for environmental analysis can be found via the
following link: http://www.emis.vito.be/referentielaboratorium (last accessed on 20/01/2013).
Especially for heavy metals - and in more general terms, inorganic parameters - the Flemish
reference methods are nowadays based on European standards developed by the European
Committee for Standardization (CEN) [35]. These standards have a unique status since they also are
national standards in each of its 32 European member states. With one common standard in all these
countries and every conflicting national standard withdrawn, comparison between analytical results
becomes more feasible.
The national organisations responsible for drafting and publication of European standards in, e.g.,
France (AFNOR), the Netherlands (NEN), Germany (DIN), work with national committees, consisting
of national experts, where comments can be formulated on European standards. These comments
made by national committees often include much more than just scientific argument (e.g., proposals
for which acid digestion to be used are strongly related to methods that have been historically used
in the country).
At Belgian level, NBN - the Belgian Bureau for Standardisation - is the national organisation
responsible for the drafting and publication of standards. However, since 1980, the management of
the environment in Belgium became a task for the regions (Flemish, Walloon and Brussels Capital
Region). Since then, there has hardly been consultation or discussion between the regions as to
which reference methods to use for environmental analysis.
As the development of European standards with respect to environmental monitoring were strongly
linked to the European Directives, VITO in consultation with LNE and OVAM (see §1.2) started to
consistently introduce European standards as Flemish reference methods since 2002. Besides,
technical committees with Flemish recognised laboratories were organised, to discuss the practical
implementation of the analytical reference methods.
A similar trend has been observed in the Walloon Region and was also recently (2012) initiated by
the Brussels Capital Region. In some way, the European Committee for Standardization has led to a
“rapprochement” in the internal Belgium situation. Still, structural consultation between the regions
in this respect would be desirable. The European Committee for Standardization (CEN) only accepts
comments from the Belgian Bureau for Standardisation (NBN), but often no unanimous comments
are formulated because of the absence of a common Belgian environmental policy.
CEN closely cooperates with its international counterpart, the International Organisation for
Standardization (ISO) in the fields of soil and water. With respect to ICP-MS, the following EN
standards have been published so far (2012) by different CEN technical committees in the context of
Environmental analysis:
CEN/TC 230 - Water analysis

• EN ISO 17294-2:2004 Water quality - Application of inductively coupled plasma-mass
spectrometry (ICP-MS) - Part 2: Determination of 62 elements (ISO 17294-2:2003)
• EN ISO 17294-1:2006 Water quality - Application of inductively coupled plasma-mass
spectrometry (ICP-MS) - Part 1: General guidelines (ISO 17294-1:2004)
CEN/TC 264 - Air quality
• EN 14385:2004 Stationary source emissions. Determination of the total emission of As, Cd,
Cr, Co, Cu, Mn, Ni, Pb, Sb, TI and V
• EN 14902:2005 Ambient air quality. Standard method for the measurement of Pb, Cd, AS,
and Ni in the PM 10 fraction of suspended particulate matter
• EN 15841:2009 Ambient air quality - Standard method for determination of arsenic,
cadmium, lead and nickel in atmospheric deposition
CEN/TC 400 Horizontal standards in the fields of sludge, biowaste and soil
• CEN/TS 16171:2012 Sludge, treated biowaste and soil - Determination of elements using
inductively coupled plasma-mass spectrometry (ICP-MS)
CEN/TC 292 Characterization of waste
• EN 16192:2012 Characterization of waste - Analysis of eluates
(a reference is made to EN ISO 17294-1:2006 and EN ISO 17294-2:2004)
• EN 15192:2006 Characterization of waste and soil - Determination of Chromium(VI) in solid
material by alkaline digestion and ion chromatography with spectrophotometric detection
(a note is made that hyphenated methods with ion chromatographic separation and
detection techniques, such as inductively coupled plasma-mass spectrometry (ICP-MS) may
be used)
The lag in the publication of EN standards related to ICP-MS (starting in 2004) is somewhat in
contrast with the first United States Environmental Protection Agency (US EPA) standard related to
ICP-MS that was already published in 1990 for the determination of trace elements in water (US EPA
200.8). This US EPA procedure was updated in 1994 to address improved sensitivity and detector
modifications and in 2005, guidance on the use of collision/reaction cell technology was introduced.
US EPA also pioneered in 1997 with the first standard method using ion chromatography coupled to
inductively coupled plasma-mass spectrometry (Bromate in Drinking Water, US EPA 321.8) and in
2007 with a standard method for elemental and speciation isotope dilution mass spectrometry (US
9
analysis since 1980
EPA 6800). According to the US National Environmental Methods Index (NEMI), 18 standard methods
for the analysis of water using ICP-MS are available to date.
In Flanders, ICP-MS was introduced for the first time as a Flemish reference method in the
compendium for sampling and analysis in support of the waste and soil remediation decrees in 2006
(CMA, see also Figure 9). The selection of the ICP-MS method as a reference method was the result
of a round robin test that was organized by VITO in commission of OVAM in 2004 to evaluate the
analytical performance of laboratories for the determination of Ba, Cd, Cr, Mo, Sb and Se in eluates
with ICP-AES and ICP-MS [36]. Nineteen laboratories (from Flanders/Belgium and the Netherlands)
with recognized expertise for eluate analysis participated in this round robin test and analyzed four
samples spiked with the elements at relevant concentration levels. Based on these results, it was
concluded that for the elements Mo, Ba, Cr and Cd accurate and reproducible data were obtained,
independent of the technique applied (ICP-MS or ICP-AES). For the determination of the elements Sb
and Se, the ICP-MS technique is the method of choice, especially when concentrations in the lower
μg/L-range have to be determined with minimum measurement uncertainty. The results of this
round robin test were also used as validation data for the introduction of ICP-MS in the EN standard
for the analysis of eluates (EN 16192:2012).
The development of EN standards is governed by the principles of consensus and national

commitment and this process can sometimes take a long time as illustrated in the next example.
Project HORIZONTAL started in 2002 with the aim to develop harmonised European standards in the
field of sludge, soil, and treated biowaste. These standards were needed in order to facilitate the
regulation of these major streams in the multiple decisions related to different uses and disposal
governed by EU Directives. The work started with desk studies to evaluate the feasibility of the
development of a horizontal standard. Drafts of horizontal technical specifications (CEN/TS) were
submitted to CEN/TC 400, a project committee especially devoted to this project. CEN/TC 400
decided in 2010 that the horizontal technical specifications (CEN/TS) from the Project HORIZONTAL
were not sufficiently validated in the course of previous interlaboratory comparisons and should be
upgraded to European standards by generating new reliable validation data [37]. A new validation
was organised in 2013, in which a soil, a sludge and a biowaste material were distributed among
different European laboratories and in which elements were determined after digestion/extraction
of the materials according to the following European standards (EN) and technical specifications
(CEN/TS):
• EN 16173:2012 Sludge, treated biowaste and soil - Digestion of nitric acid soluble fractions of
elements
• EN 16174:2012 Sludge, treated biowaste and soil - Digestion of aqua regia soluble fractions
of elements
inductively coupled plasma-optical emission spectrometry (ICP-OES)
graphite furnace atomic absorption spectrometry (GF-AAS)
• CEN/TS 16175-2:2013 Sludge, treated biowaste and soil - Determination of mercury - Part 2:
Cold vapour atomic fluorescence spectrometry (CV-AFS)
To date (2013), the technical specifications (CEN/TS) for the determination of elements have not
been published as European standard yet and the example illustrates the time lapse in the
development of EN standards. It is worth to mention that along with this new validation study, a
comparison was made between CEN/TS 16171 (ICP-MS) and CEN/TS 16175-2 (CV-AFS) for the
determination of Hg in the digestion solutions. A good correlation was found between both
determination methods at these low levels of mercury and it is to be expected that Hg will be
included in the scope of CEN/TS 16171 as one of the elements determinable by ICP-MS. This will
certainly favour the use of ICP-MS for soil, sludge and biowaste analysis in the future, as no separate
Hg determination (with, e.g., CV-AFS) will be needed. Moreover and under impulse of the national
standardisation organisations of the Scandinavian countries, nitric acid is being promoted as the sole
acid to be used for environmental sample digestion. This would further play into the hands of ICP-MS
as the method of choice for monitoring. However, countries such as Germany and the Netherlands
have historically used aqua regia digestion and it remains to be seen whether they will agree with
this proposal.
For the analysis of elements in water in the context of environmental regulatory monitoring, a study
was performed by VITO in 2006 to evaluate the different analytical methods available for selection as
reference method [38]. It was recommended that ICP-MS should be included in the compendium for
sampling and analysis of water (besides ICP-AES and AAS) and this was done in the revision of the
compendium in 2009. At the same time, the reference methods for the analysis of waste water as
defined in VLAREM II were updated (annex 4.2.5.2, until 2009 only atomic absorption spectrometry
and polarography were defined as reference methods). In 2011, by decision of the Flemish
Government (in line with Directive 2009/90/EG and in support of the European Water Framework
Directive), ICP-MS became the only method of choice for chemical analysis and monitoring of
dissolved elements in surface water (with the exception of Hg, which is not included in EN ISO 17294,
and for which a separate determination with, e.g., CV-AFS is still needed).
For the analysis of elements in air in the context of environmental regulatory monitoring, the first
version of the compendium for sampling and analysis of air (LUC) became available in 2012. In this
compendium, a reference to ICP-MS, ICP-AES and AAS is made. In the same year, the reference
methods for the analysis of elements in emissions as defined in VLAREM II were updated and a
reference to the compendium was included (annex 4.4.2, until 2012 only AAS, ICP-AES, DCP-OES and
XRF were defined as reference methods).
Currently, there is a greater appreciation of the importance of metrology in analytical chemistry with
many environmental laboratories accredited to international standards such as ISO 17025. However,
with increasing standardization and accreditation of professional qualifications, a critical comment
can be formulated that possibly too much emphasis is placed on achieving uniform end results of an
analysis rather than on understanding the analytical measurement within the context of the sample
(obtaining an accurate value).
The assumption of standardization and accreditation in a regulatory context is that only one or a few
constituents of a sample - usually the analytes of interest - play a role in the measurement result that
is obtained. Though, this assumption is usually not true for environmental samples. When applying
standardized procedures in routine, users tend to have as little concern as possible about which
other species in the sample might yield a false response or whether the method is applicable to all
samples in the measurement run. Also policy makers have an ambiguous relation to the concept of
measurement uncertainty and how to deal with it in a regulatory context. This approach of using
standardized procedures has been referred to “the chemical analysis of things as they are not” and
chemists were classified as “Determinators”, while more traditional analytical chemists can be
categorized as “Analysts” [39]. Determinators’ primary interest is how the analytical result pertains
to a particular system and this is also often the assumption in the current routine use of
environmental reference methods.
In some way, the increasing standardization and accreditation of professional qualifications that is
introduced in the environmental regulatory monitoring has led to a false impression that also the
understanding of the analytical measurement within the context of the sample was increased. One
11
analysis since 1980
must be aware that a breakdown of the monitoring chain (i.e., drafting the sampling plan, sampling,
sample preparation, sample analysis, data interpretation) into different entities / persons increases
the risk of losing an overview on the real environmental issue. Unfortunately and under pressure of
economic competition, an inverse relationship is being created in the (commercial) environmental
laboratories between standardization and accreditation on the one hand and understanding and
knowledge of the analytical measurement within the context of the sample on the other. It is to be
expected that with increasing juridical issues in environmental regulatory monitoring, environmental
analysts will become more and more environmental determinators. This is especially true when one
takes into account that, from a regulatory point of view, strict compliance with the reference
measurement methods is considered more important than obtaining an accurate value.
On the other hand, it is observed that with each introduction of new environmental legislation, not
only the number of analytes of interest is increased, but also physicochemical properties of the
sample are considered to acknowledge the issue of (bio)availability (e.g., in WFD, the water quality
criteria for Cd in surface water depends on hardness and pH, see § 2.5), the issue of geochemical
characteristics (e.g., soil remediation standards for heavy metals depends on the pH, clay and organic
carbon content of the soil, see § 2.2) or the issue of toxicity (e.g., hexavalent chromium in waste
leachates, see § 3.3).
Hopefully, this trend will give rise to a better understanding of the analytical measurement within
the context of the sample itself and will lead to joined forces between environmental analysts and
determinators for the correct interpretation of the actual environmental situation.
Another negative side-effect of harmonisation and standardisation of analytical methods is that new
and innovative monitoring / screening methods face difficulties of acceptance and use in a regulatory
context. This is also related to the concept of measurement uncertainty (especially using screening
methods) and how to deal with it in a regulatory context, i.e., the translation of probabilistic data
(analytical data) into boolean data (environmental issue? yes/no). While from a scientific point of
view, it is believed that a combined monitoring methodology using on-line and/or screening
monitoring methods in combination with standardised methods would lead to a more cost-effective
and better understanding of local or regional environmental issues.
Interesting links :
• Institut scientifique de service public (ISSeP) is the reference laboratory for the Walloon region:
http://www.issep.be
• Brussels instituut voor milieubeheer (BIM) is the reference laboratory for the Brussels Capital
Region, http://www.leefmilieubrussel.be/
• The analytical methods used for environmental monitoring can be hazardous to the environment
and to human health. Greener analytical methods use fewer hazardous solvents, use safer
chemicals, prevent waste, and conserve energy during the sample preparation and analysis.
Greenness profiles for many of the monitoring methods are available in the (US) National
Environmental Methods Index (NEMI).The greenness profiles, based on four criteria, give
guidance for selecting a method that has a less negative impact on the environment :
https://www.nemi.gov/
• Extended search engine for standards developed by The European Committee for
Standardization (CEN): http://esearch.cen.eu/esearch/
1.4. THE ATOMIC SPECTROMETRY TIMELINE IN ENVIRONMENTAL ANALYSIS
In the following section (§ 1.4.1), ICP-MS will be discussed with emphasis on the different
components on the one hand and on how the different types of (commercially available) instruments
cope with spectral interferences, the Achilles’ heel of ICP-MS, on the other. In the context of this
dissertation, quadrupole-based ICP-MS units (ELAN 6000 and NEXION 300S, Perkin Elmer) and a
sector-field ICP-MS unit (Element II, Thermo Fisher Scientific) were used (see Figure 2).
Figure 2: (left) Young scientist operating a Perkin Elmer Elan 6000 quadrupole-based ICP-MS instrument (1999);
(right) schematic representation of a sector field ICP-MS instrument (Element II), Thermo Fisher Scientific.
Thereafter (§ 1.4.2), the emergence of ICP-MS in environmental analysis since 1980 and some
milestones of important Flemish/European/US standard methods based on ICP-MS for the
determination of elements in environmental matrices are described (see also § 1.3). Sample
preparation is discussed in § 1.4.3, as this is usually needed in environmental analysis to convert
analytes to a more suitably detectable form or to separate them from the sample matrix.
1.4.1. ICP-MS INSTRUMENTATION
An ICP-MS instrument combines a high-temperature inductively coupled plasma (ICP) ion source
(temperature of around 6000-10000 K) with a mass spectrometer [15-17]. The ICP source converts
the atoms of the elements in the sample into ions; these ions are then separated from one another
according to their mass-to charge ratio and detected by the mass spectrometer (the different
components are visualised in Figure 3).
Atmospheric pressure
nebuliser spray chamber plasma sampling interface
Low pressure (< 2 Pa)

ion lens collision reaction mass spectrometer detector
cell
Figure 3: General scheme of an ICP-MS instrument with indication of different components.

13
analysis since 1980
→ The different components of an ICP-MS instrument [15-17]
The sample is typically introduced into the ICP plasma as an aerosol, either by aspirating a liquid or
dissolved solid sample into a nebulizer-spray chamber or by using a laser to directly convert solid
sample into an aerosol. Once the sample aerosol is introduced into the ICP torch, it is completely
desolvated. The molecules set free out of the solid particles thus formed are atomized and these
gaseous atoms are then ionized towards the end of the plasma. From a physical point of view, a
plasma is nothing else than the fourth state of matter besides solid, liquid and gas (most of the
matter in our universe exists in the state of a plasma)[19]. The ions formed in the ICP are typically
positive ions, M+ or M2+, therefore, elements that prefer to form negative ions, such as Cl, Br, F, etc.,
are very difficult to determine via ICP-MS. The sample matrix may affect the degree of ionization
that will occur in the plasma or give rise to ionic species that may interfere with analyte
determination. Argon is commonly used as plasma gas.
Once the elements in the sample are converted into ions, they are brought into the mass
spectrometer via the interface cones [21]. The interface region in the ICP-MS transmits the ions
travelling in the argon stream at atmospheric pressure into the low pressure region of the mass
spectrometer (<2 Pa). This is done through the intermediate vacuum region created in-between the
two interface cones, the sampler and the skimmer. The sampler and skimmer cones are metal disks
with a small hole (∼1mm) in the centre. The purpose of these cones is to sample the centre portion
of the ion beam coming from the ICP torch. Due to the small diameters of the orifices in the sampler
and skimmer cones, ICP-MS has some limitations as to the amount of total dissolved solids allowed in
the samples. Generally, it is recommended that samples have no more than 0.2% total dissolved
solids (TDS) for best instrument performance and stability. If samples with very high TDS levels are
run, the orifices in the cones will eventually become blocked, causing decreased sensitivity and
detection capability and requiring the system to be shut down for maintenance. This is why many
sample types, including digested soil and rock samples must be diluted before ICP-MS analysis (see
also § 2.2).
The ions from the ICP source are then transported to the mass spectrometer by the electrostatic
lenses in the system. Different types of ICP-MS systems have different types of lens systems. The
simplest employs a single lens, while more complex systems may contain as many as 12 ion lenses.
Each ion optic system is specifically designed to work with the interface and mass spectrometer
design of the instrument.
Once the ions enter the mass spectrometer, they are separated according to their mass-to-charge
ratio. Different types of mass analyzers are being used in commercially available ICP-MS instruments,
i.e., quadrupole mass analyzers, sector-field mass analyzers and time-of-flight (TOF) analyzers.
The most commonly used type of mass spectrometer is the quadrupole mass filter (ICP-QMS). A
quadrupole mass filter consists of 4 rods, AC and DC voltages are applied to the two diagonally
opposed pairs of the rods. The two pair of rods thus obtained, show DC voltages of opposite signs,
while the sinusoidal AC voltages (RF frequency) show a 180° phase difference. The result is that an
electrostatic filter is established that only allows ions of a single mass-to-charge ratio (m/z) to pass
through the setup to the detector at any given time. The quadrupole mass filter is a sequential filter,
with the settings being changed to select another m/z. However, the voltages on the rods can be
switched at a very high rate. This high scanning speed is why the quadrupole ICP-MS is often
considered to have (quasi) simultaneous multi-elemental analysis properties. Very recently, a new
type of quadrupole-based ICP-MS instrument has been introduced onto the market [22]. This
instrument is referred to as a triple quadrupole or an ICP-QQQ set-up by the manufacturer (Agilent),
although the terminology “triple quadrupole” is not entirely correct, as in-between the two
quadrupole analyzers, an octopole-based collision/reaction cell is located. The first quadrupole
prevents all off-mass ions from entering the cell, allowing more controlled and efficient interference
removal in reaction mode (see further, Figure 5), regardless of sample type.
Conventionally, quadrupole ICP-MS instruments provide approximately unit mass resolution. High
mass resolution of sector-field instruments (ICP-SFMS) is a feature which depends on the geometry
of the electric and magnetic sector, and the slit configuration. In this type of instrument, a
combination of a magnetic sector and an electrostatic sector are used to separate and focus the ions.
Such an arrangement is called a double-focusing high resolution mass spectrometer. In ICP-MS,
reverse Nier-Johnson geometry - where the magnetic sector is located before the electrostatic sector
- is commonly used in order to decouple the electric fields in the electrostatic sector from any electric
field originating by the ICP RF generator (e.g., ELEMENT ICP-SFMS, Thermo). The resolution of sector-
field instruments can be changed by adjusting the width of the entrance and exit slits before and
after the MS. Typical ICP-SFMS instruments have resolving powers up to 104 and are operated at pre-
set resolution settings for low, medium or high-resolution to make their operation easier for the
user. Unfortunately, for every 10-fold increase in resolving power, there is a concomitant decrease in
signal intensity. Recently a sector-field mass spectrometer with Mattauch-Herzog geometry became
commercially available (Spectro MS). The advantage of this geometry over the Nier-Johnson
geometry is that the ions of different masses are all focused onto the same plane and offers truly
simultaneous detection of all isotopes (Li to U) as a result of a semiconductor-based multi-channel
detector, albeit at low mass resolution only.
Another method of mass spectrometry is based on the measurement of the time-of-flight of ions of
known energy over a known distance. The time-of-flight mass analyzer (ICP-TOFMS) is a pseudo-
simultaneous detection type device, which can accomplish a full mass spectrum in approximately 30
µs. In this technique, ions are subjected to a pulsed electric field which, ideally, imparts the same
kinetic energy (KE) on all ions in the packet. These ions are then directed to a field-free region (i.e.,
drift tube) where the differences in velocities spatially and temporally separate ions of differing m/z.
The KE attained by each ion will be product of the ion's charge and strength of the electric field.
Once the ions have been separated according their mass-to-charge ratio, they must then be detected
or counted by a suitable detector. The fundamental purpose of the detector is to translate the
number of ions striking the detector into an electrical signal that can be measured and related to the
number of atoms of that element in the sample via the use of calibration standards. Most detectors
use a high negative voltage on the front surface of the detector to attract the positively charged ions
to the detector. Once the ion hits the active surface of the detector, a number of electrons is
released, which then in turn strike the surface of the detector, amplifying the signal. In the past
years, the continuous dynode channel electron multiplier (CEM), which was used on earlier ICP-MS
instruments, has been replaced with discrete dynode type detectors. Discrete dynode detectors
generally have wider linear dynamic ranges than CEMs, which is important in ICP-MS, as the
concentrations analyzed may vary from sub-µg/L to high mg/L. Both detector types can be run in two
modes, pulse-counting and analog, which further extend the instrument's linear dynamic range and
can be used to protect the detector from excessively high signals. A second type of ICP-SFMS
instrument is also available that uses multiple detectors, this type is called multi-collector ICP-SFMS
or MC-ICP-SFMS [16]. These instruments are generally designed and developed for the purpose of
performing high-precision isotope ratio analyses. Since an array of 5-10 detectors can be positioned
around the exit slit of a double-focusing system, the isotopes of a single element can generally all be
determined simultaneously, leading to the technique's high-precision. MC-ICP-MS instruments tend
to use simpler, less expensive Faraday cup type detectors, as these have the ability to deal with the
very high count rates common with magnetic sector instruments and do not suffer from dead time
effects. However, these detectors do not have the flexibility, necessary for quadrupole ICP-MS
instruments.
15
analysis since 1980
Interesting You tube links on the principles of ICP-MS :

• Agilent, ICP-MS (collision cell) : http://www.youtube.com/watch?v=MQqtV2oiC6U, last accessed
on 08/01/2013
• Agilent, ICP Triple Quad : http://www.youtube.com/watch?v=b9BfKcxmltI&feature=relmfu, last
accessed on 08/01/2013
• Perkin Elmer, ICP-MS (collision and reaction cell) : http://www.youtube.com/watch?v=L-
FYh2z9mi0&feature=related, last accessed on 08/01/2013.
• Spectro, MC-ICP-SF-MS : http://www.youtube.com/watch?v=75uMtzyo844, last accessed on
08/01/2013
→ How to cope with spectral interferences? [17]
Spectral interferences are often considered as the Achilles’ heel of ICP-MS. Several strategies on how
to cope with spectral interferences were used in the context of this dissertation and are described in
the following section.
Spectral interferences in ICP-MS are caused when ions generated from the plasma, the sample,
entrained air or a combination thereof carry a nominal mass-to-charge ratio that is identical to that
of the analyte ion. Whether or not an interfering ion signal present at mass m ± ∆m will be separated
from an analytical signal present at mass m depends on the mass difference ∆m and the resolution of
the instrument. Mass resolution R is generally defined as R = m/∆m, where ∆m is the mass difference
necessary to achieve a valley of 10% between two neighbouring peaks of identical intensity at mass
m and m ± ∆m as shown in Figure 4.
Figure 4: Illustration of 10% valley definition of mass resolution (courtesy of [17]).
Quadrupole mass spectrometers as used in ICP-MS provide peak widths between 0.7-1.0 amu. For
most environmental applications, and especially for elements in the mass region 40-80 amu (Cr, Cu,
Ni, Zn, As), this resolution is not sufficient to separate the overlapping signal from a molecular or
isobaric ion from that of the nuclide of interest.
Spectral interferences may be subdivided into several groups, as they can be attributed to the
presence of isobaric atomic ions, multiply charged ions and polyatomic ions of various origins.
Isobaric overlap exists when the signals of nuclides of different elements coincide at the same
nominal mass. Since the mass difference between isobars is in general very small, resolutions
between 104 and 108 are required for their separation. The maximum mass resolution setting of
commercial ICP-SFMS instruments (about 104) is by far insufficient to overcome interferences of this
kind. Therefore, an alternative approach to cope with this problem is used. For each element - with
the only exception of indium - at least one isotope is free from isobaric overlap and can be monitored
instead, but in many cases this will not be the most abundant one. This strategy is often successful. If
more complex samples are analysed, the number of analyte signals disturbed by other kinds of
interferences is drastically increased in the mass range below 100 amu. Since isobaric interferences
are usually easily predictable, they can also be corrected for mathematically by relying on natural
isotopic abundances.
Multiply charged ions will be found in the mass spectrum at a position m/z (m is the nominal mass
and z is the ion charge), i.e., doubly charged ions are for instance found at half of their nominal mass.
Multiply charged ions usually have a lower formation rate in an ICP and occur mainly in the medium
mass range. Doubly charged ions of the main matrix constituents are frequent contributors.
Polyatomic ions cause the most severe spectral interference problems in environmental analysis and
thus, their origin shall briefly be discussed. Polyatomic interferences are less predictable and depend
on the sample composition (analytes and matrix) and the operational parameters of the ICP-MS
system. Therefore, it is difficult and sometimes impossible to mathematically correct for them. These
interferences require mass resolutions of 103 to 104 for their resolution up to mass 70 and sometimes
more than 104 at higher masses. Polyatomic interferences may find their origin in the sample itself,
such as oxide ions which, owing to their high bond strength, have a real chance of ‘surviving’ the
passage through the hot zones of the plasma. They may also arise from the discharge gas,
contaminants, entrained air, the reagents and solvents used and the matrix of the sample. They
might become extremely complex in organic matrices. The formation of cluster ions from the
dominant species in the plasma (Ar, H, O, C, N) is a major source of polyatomic ions that do not only
arise in the plasma, but can also be formed during the extraction process [18]. The formation of
interferences is governed by thermodynamics (exothermic and endothermic processes) and is
possible if the energy which is needed for the process is provided by the plasma or the excess
reaction heat. The product molecule usually has a binding energy which is high enough to survive the
short passage through the plasma and the interface as the exposure time of particles in the plasma is
only about 2 ms.
Proper optimization of the instrument is the very first step towards overcoming interferences.
Nonetheless, there are several strategies to cope with residual interferences (besides the use of high
mass resolution), which will be described shortly in the following sections. The common 40Ar35Cl+
interference - argon from the plasma and chloride from the sample matrix combine to form a
polyatomic species, 40Ar35Cl+, which carries the same nominal mass as the arsenic isotope (m/z 75) -
will be used as an example to demonstrate how the different strategies cope with spectral
interferences.
Mathematical correction procedures
Mathematical correction is usually applied in case of relatively simple situations (isobaric

interferences) or if all other measures fail. In the case of an isobaric interference, another nuclide of
the interfering element is measured and the contribution of the interference to the analyte signal is
calculated relying on the abundances of the isotopes of the interfering element.
The situation is more complicated for polyatomic interferences as their occurrence are not easily
predictable and depends on the operational parameters as well as on the matrix. Therefore, the
formation probability has to be estimated using external matrix-matched solutions or the same ion
has to be observed at a different m/z and its intensity used for further correction.
This is demonstrated via the following example: 75As+ suffers from a spectral interference in Cl-
containing solutions as a result of the occurrence of the 40Ar35Cl+ ion. As a consequence, the following
mathematical strategies can be used: The 40Ar35Cl+ interference is measured along with the 35Cl+ (or
17
analysis since 1980
37
Cl+) signal in pure matrix solutions, containing only the matrix and Cl of increasing concentration,
but no As. The ratio 40ArCl+/35Cl+ is calculated from these solutions and an average ratio is further
used as correction factor. Now, the signal on mass 75 is measured in the solution along with the
signal of 35Cl+ on mass 35 (or 37Cl+ on mass 37). The 35Cl+ (or 37Cl+) signal is now multiplied with the
previously calculated factor, providing an estimate for the interference of 40Ar35Cl+ on mass 75. The
remaining signal is most probably the As signal. Unfortunately, the solution isn't always this simple:
75
As+ also suffers from the polyatomic interference 40Ca35Cl+ in calcium-containing samples (e.g., soil).
Another possibility would be that the 40Ar37Cl+ signal on mass 77 is measured along with the signal on
mass 75. Via the isotopic abundances of Cl, the 40Ar35Cl+ signal can be calculated from the 40Ar37Cl+
signal. Again, the remaining signal is most probably the requested As signal. Nonetheless, we can
observe an additional problem: we also find 77Se+ on mass 77, where we expect our 40Ar37Cl+
interference. Therefore prior to calculating the 40Ar35Cl+ contribution from the 40Ar37Cl+ signal, we
have to subtract the signal of 77Se+. We can estimate the contribution of 77Se+, relying on the isotopic
abundances of Se and using, e.g., the 82Se+ signal intensity. Nonetheless, we have to consider that we
find an isobaric interference of 82Kr+ on mass 82. Not to consider any 40Ar21H2+ interference.
It is obvious that this strategy only leads to satisfying results when a relatively high analyte
concentration and low matrix concentration are occurring. As good practice, it is always
recommended to monitor more than one isotope (if possible), even if the other isotopes are less
abundant.
Sample introduction systems
Among various strategies discussed over many years, the appropriate selection of a sample
introduction system best suited for the analytical problem can already help to overcome much
solvent- and matrix-based spectral interferences. Cooled spray chambers are applied to reduce the
amount of water vapour transferred to the plasma. These systems use heating/cooling devices in
order to first vaporize the solvent and subsequently condense and drain the solvent. Advanced
systems use desolvation systems to significantly reduce interferences. Solvent molecules can pass
through membranes and are removed by a counter stream of Ar, whereas the dried aerosol is
transported to the plasma (oxide formation rates can be reduced by some orders of magnitude
compared to conventional introduction systems) [23]. Depending on the temperature programmes
chosen for electrothermal vaporization, solvent-based interferences can be reduced prior to the
analysis of the trace elements. Dissolution of solid samples and thus interferences caused by mineral
acids can be avoided if laser ablation is chosen as a sample introduction device, even though other
matrix-based interferences can still occur.
Sample/matrix separation
One straight chemical method to avoid formation of interferences is the direct separation of the
analytes of interest from the matrix. Even spectral interferences from samples with a high level of
dissolved salts can be avoided if, e.g., hydride or cold vapour generation is applied for trace/matrix
separations. The analytes of interest can be separated from the interfering matrix by
chromatography as well. This is accomplished either by using a batch process or by direct coupling of
chromatography to ICP-MS (see also § 3.2 )[24].
Cool/cold plasma technique
A number of interfering ions are formed inside the inductively coupled plasma or at the sampling
cone surface in contact with the plasma. Thus, changing the working conditions of the plasma is a
powerful tool, especially for the reduction of argon-based interferences. This approach was realised
in the ‘cool/cold plasma’ technology. Cool/cold plasma conditions are obtained by increasing the
nebulizer gas flow rate or addition of an additional gas to the central channel of the torch, along with
the reduction of the RF-generator power. In this operation mode, the risk of secondary discharges is
high. The breakthrough was achieved once effective plasma shielding was developed by appropriate
induction coil grounding or application of a plasma shield. The cold plasma technique helps to reduce
argon-based spectral interferences such as Ar+, ArC+, ArO+, ArCl+, Ar2+. This can be used for an
improvement of limit of detections (LODs) for elements such as Ca, K, Cr, and Fe, which are difficult
to determine at low levels, e.g., in the electronic industry. These advantages have to be paid for
under the form of many additional spectral interferences coming from water clusters (combinations
with H2O and H3O+) and increased oxide formation [19]. Additionally, a loss in sensitivity for all
elements with first ionization energy above 8 eV have been observed and even worse, a loss of
robust plasma conditions causing elevated matrix effects hampers its application. As a consequence,
this mode is mainly used for relatively pure samples with low salt content.
Collision and reaction cell technology
Among all approaches for overcoming spectral overlap, those based on instrumental improvements
have been most successful and in particular, the application of collision and reaction cells in front of
a low resolution quadrupole mass analyser have changed the whole spectrum of applications in ICP-
QMS. In current cell technologies, a pressurized cell, equipped with an ion guide (which – depending
on the instrument – is an RF-only band-pass quadrupole, a hexapole or an octopole) is used to
overcome a number of prominent spectral interferences by gas phase reactions with a reaction gas
(reaction cell) or by application of a retarding field in combination with a non-reactive collision gas
(collision cell). In the latter case, molecular polyatomic species are expected to lose more kinetic
energy by collisions with a buffer gas then atomic ions. With application of a retarding field, the
polyatomic ions can be selectively discriminated against (kinetic energy discrimination). The KED
mode of operation in the Agilent 7700 series ICP-MS is represented below (figure by courtesy of
Agilent).
Figure 5: The KED mode of operation in the Agilent 7700 series ICP-MS (courtesy of Agilent).
Reaction gas technology is based on the formation of different molecular ions form the interfering
polyatomic species (e.g., 2 ArAr+ + H2 → 2 ArArH+), neutralisation of interferences (e.g., ArCl+ + NH3 →
ArCl + NH3+) or by shifting the analyte of interest to a higher mass (e.g., As+ will be converted to AsO+
and the mass will be shifted from mass 75 to 91 using O2 as reaction gas). Usually, reaction gas
19
analysis since 1980
technology is advantageous to be coupled with an ion filter in order to exclude newly formed
interferences. Whether a reaction will proceed or not can be predicted by thermodynamics,
nevertheless optimization is usually done empirically. Unfortunately, also no single reaction or
collision gas or application of the combination of a buffer gas and a retarding field is able to
overcome all possible interferences from polyatomic species and most often, the interfered element
has to be determined in a separate run to achieve the best sensitivities for as many elements as
possible. Moreover, the matrix dependence of the reaction/collision cell technology is not fully
elucidated. Nonetheless, the technology has no inherent limitations, in contrast to high resolution
ICP-SFMS where the maximum resolution is 104.
Mass resolution
Conventionally, quadrupole ICP-MS instruments provide approximately unit mass resolution. Typical
sector-field ICP-MS instruments have resolving powers up to 104 and are operated at pre-set
resolution settings for low, medium or high-resolution to make their operation easier for the user.
Unfortunately, for every 10-fold increase in resolving power, there is a concomitant decrease in
signal intensity.
Most polyatomic species can be separated from the nuclide of interest by application of high mass
resolution (m/∆m < 104). In most of the relevant examples of spectral interference, a resolution of
about 4000 (most often called medium resolution) would be by far sufficient for separation. Only in
special cases, some polyatomic species are observed, which require a higher mass resolution (4000 <
m/∆m < 10 000). In the example given for 75As+, a resolution of about 8000 is sufficient to overcome
this problem.
+
As+ ArCl As+ ArCl+
Figure 6: Effect of measuring in the different resolution modes (low (~ quadrupole), medium and high
75 + 40 35 +
resolution, respectively) on the separation between the As (green) and Ar Cl (black) signals.
Of course, high mass resolution is not a panacea against all spectral interferences. As mentioned
already, isobaric interferences are far beyond the scope of most commercially available devices. MO+
ions, where oxygen is a constituent of many mineral acids and water, can be resolved from the
corresponding analyte signals in the mass range below 100 and above 140 amu (a resolution of less
than 104 is sufficient for a complete separation). In the mass range between 100 and 140 amu, a
resolution of up to 106 would be necessary. The required resolution exceeds by far the resolution
attainable with commercially available instrumentation. This is a crucial point for, e.g., low level rare
earth element (REE) determination if formation of oxides is not prevented by other means. This
example demonstrates that high mass resolution is a very powerful approach to overcome many, but
not all of the spectral interferences occurring in routine applications.
1.4.2. ATOMIC SPECTROMETRY TIMELINE
One of the most successful analytical plasma sources in emission as well as in mass spectrometry, is the
inductively coupled plasma. The analytical inductively coupled plasma (ICP) was described in 1964 by
Greenfield and later introduced for analytical applications in atomic emission spectroscopy (AES) [17]. The first
commercial ICP-AES instrument was constructed in 1974. In 1978, Houk started to couple an ICP to a
quadrupole-based mass analyzer. First results were already published in 1980, the official birth year of ICP-MS
(see also timeline, Figure 9)[1,25].
In the 1980s, a transition from Flame Atomic Absorption Spectrometry (FAAS) to Inductively Coupled
Plasma-Atomic Emission Spectrometry (ICP-AES) as a workhorse in environmental laboratories was
driven by a commercial necessity for higher sample throughputs with a higher degree of automation.
In the 1990s, many water laboratories shifted to ICP-MS as the technique promised to offer the
throughput capabilities of ICP-AES, coupled with the sensitivity of Electrothermal Atomic Absorption
Spectrometry (ETAAS). The higher detection power offered by ICP-MS, compared to that of ICP-AES,
led to its rapid acceptance by the Environmental and Earth Science Communities [26].
The main trend in water analysis since then has been the gradual decline and replacement of FAAS
with ICP-AES, and the rise and fall of ETAAS as a competitor from ICP-MS. AAS is still heavily used
throughout the world due to its low relative cost and usefulness for cross-checking results from other
instruments. Most of the advances in AAS and ICP-AES now are front-end modifications; advances in
the instruments themselves seem to be mostly increased computerisation with automatic set-up of
conditions. One of the limiting factors for AAS is that it is essentially mono-elemental; attempts have
been made to create fast sequential instruments to speed up analysis or perform simultaneous
multi-element ETAAS (e.g. continuum source atomic absorption spectrometry [27]). Some of the
more important innovations in ICP-AES include the introduction of dual view instruments and the
adoption of Echelle spectrometers coupled with array detectors [28].
However, it is argued in a recent 25-year retrospective spectrometry update that the most important
advance in water analysis has been the development of ICP-MS from a research curiosity into a
robust, precise and sensitive analytical tool that now can detect most elements in fresh and drinking
waters without any or after minor sample pretreatment [26,29]. Routine methods for the elemental
analysis of drinking water were published in 1990 (see §1.3). By the early 1990s, the capability of ICP-
MS as an element-specific HPLC detector was being increasingly exploited for speciation analysis (see
§3.1). Sensitivity improvements were gained using similar strategies to those employed with ICP-AES,
such as the use of ultrasonic nebulisers.
Figure 7 shows the elements traditionally determined by ICP-MS and their approximate Instrumental
Detection Limit (IDL). Care should be taken to note that IDLs are calculated as 3 times the standard
deviation of a blank measurement and represent the best possible detection capability of the
instrument. In real life, the Method Detection Limit (MDL) will generally be (sometimes) higher than
the IDL and will depend upon many factors, including: laboratory and instrument background levels,
sample matrix, sample collection and preparation methods, and operator skill level (see also
monitoring studies, CHAPTER 2). However, the IDL can be used as a general guide as to the relative
capabilities of the ICP-MS technique as compared to other analytical techniques.
21
analysis since 1980
Figure 7: Approximate detection capabilities of the ELAN 6000 quadrupole ICP-MS (Courtesy of PerkinElmer, Inc.).
It should be noted that for several elements including S, Se, B, Si, P, Br, I, K, and Ca, ICP-MS shows
fairly high detection limits. In the case of I and Br, this is due to the fact that very few positive ions
are formed in the ICP plasma for these elements. For elements such as S, Se, P, K, and Ca, isobaric
and molecular interferences from either the sample matrix or plasma species interfere with the
primary isotope. This means that less abundant isotopes with less interference (if available) must be
used for determination of these elements, which will degrade detection capabilities for these
elements.
Sensitivity improvements due to instrument design were obtained when sector-field mass
spectrometers were used with the ICP ion source [17]. Single-collector sector-field ICP-MS
instruments were designed to offer higher sensitivity than quadrupole-based systems. When they
became available commercially in 1989, their use tended to be restricted to facilities where
applications necessitated high mass resolution or enhanced sensitivity. The use of such instruments
has not spread widely to other environmental laboratories, partly due to cost considerations, but
mostly because the needs of these laboratories were satisfied by the performance of quadrupole-
based systems. Spectral interferences remained an issue with quadrupole-based systems and
necessitated the development and frequent checking of computer algorithms to correct for
interferences from major element-containing polyatomic ions, particularly for trace elements with an
m/z of ≤ 80. Many analysts also modified their dissolution protocols, sometimes detrimentally, in
order to produce a final solution in dilute nitric acid because of the contribution of other mineral
acids to the formation of significant polyatomic interferences.
In the 2000 Atomic Spectrometry Update-Environmental analysis of JAAS, the first collision cell
instrument with a hexapole cell was described; however, the boom in the literature describing the
use of these instruments occurred during the period covered by the 2003 review, and by the 2005
review, a themed issue of JAAS focusing on the use of collision and reaction cell techniques in atomic
mass spectrometry has appeared [26]. Nowadays, collision and reaction cell technology is included in
practically all commercially available quadrupole-based ICP-MS instruments.
The dash to ICP-MS has however not consigned FAAS, ETAAS and ICP-AES to museums. For
laboratories undertaking analysis of environmental matrices and focusing attention to a couple of
elements and/or applications only, such techniques remain competitive. This is illustrated in the
figure below, providing an overview of the catalogue prices of the different commercially available
instruments (year 2012, prices by courtesy of Thermo, Agilent, Spectro and Perkin Elmer). The vast
majority (around 90%) of ICP-MS systems sold worldwide use a quadrupole mass spectrometer, with
alternative configurations (e.g., double-focusing sector-field, time-of-flight) making up the remainder
[31].
ICP-SFMS (reverse Nier-Johnson)

triple quadrupole ICP-MS
MC-ICP-SFMS (Mattauch-Herzog)
quadrupole ICP-MS (collision - reaction cell)
ICP-OES
FAAS - ETAAS
0 50 100 150 200 250 300 350 400
catalogue price k€ (year 2012)
Figure 8: Catalogue prices of commercially available atomic spectroscopy instruments (year 2012).
In summary, in the sixties of last century, commercial Atomic Absorption Spectrometry (AAS)
instruments were introduced, followed by inductively coupled plasma-atomic emission spectrometry
instruments (ICP-AES) in the seventies and inductively coupled plasma-mass spectrometry
instruments in the eighties (ICP-MS). It is clear from Figure 10 that since its introduction in 1980, ICP-
MS applications have continuously increased. The papers on ICP-MS as a speciation tool start
appearing in 1986 and represent 20 % of the total published papers on ICP-MS during the last 10
years. The papers on ICP-MS in environmental research also started to appear in 1987 and represent
15 % of the total number of published papers on ICP-MS during the last 10 years. Of all papers
published in the field of ICP-MS in environmental research, approximately ¼ is dedicated to
speciation analysis.
23
analysis since 1980
(2012) Triple quadrupole ICP-MS

(2010) Multi-collector, sector field
(1993) Sector field ICP-MS ICP-MS with Mattauch –Herzog
(1986) First atomic with reverse Nier-Johnson geometry
geometry (2003) Determination of
spectrometry update
organotin species
on Environmental
analysis in JAAS
(2002) Establishment of
(1996) Commercial European Virtual
(1992) Multi-collector, sector (2010) Boron
introduction of Institute for Speciation
field ICP-MS instrumentation isotope ratio in
Collision/reaction cell ICP-MS Analysis (EVISA)
(1983) Introduction of the debuts groundwater
first commercially available
ICP-MS instruments
(1989) First publication (1994) Low power, ‘cool
of Bromate in drinking of selenium species in (2011) Determination of
of F. Vanhaecke on plasmas’ ICP-MS described
(1980) First scientific water waste water hexavalent chromium in
ICP-MS for interference reduction
publication on ICP-MS particulate matter
(1999) Study of
(1985) First European tracer Cadmium (2007) Study of
Winter Conference on transport in soil hexavalent chromium in
Plasma Spectrochemistry soil and waste material
1980 1985 1990 1995 2000 2005 2010
(2006) ICP-MS is adopted as reference

method for the analysis of soil, waste,
(1990) First version of U.S. standard method eluates and groundwater in Flanders
(EPA 200.8) for the determination of trace
elements in water by ICP-MS
(1997) U.S. standard method (EPA
321.8) for the determination of (2009) ICP-MS is adopted as reference method
Bromate in Drinking Waters by Ion for the analysis of water in Flanders
Chromatography Inductively Coupled
Plasma - Mass Spectrometry
(2004) European standard method (EN ISO (2011) ICP-SFMS is adopted as
17294) for the determination of elements in reference method for the analysis of
water by ICP-MS surface water in Flanders
(2009) European standard method (EN 15841)

for the determination of elements in ambient air
by ICP-MS
(2007) U.S. standard method (EPA 6800)

for elemental and speciation isotope (2012) European standard method (EN 16171) for
dilution mass spectrometry the determination of elements in sludge, treated
biowaste and soil by ICP-MS
Figure 9: Timeline (1980 – 2012) showing ICP-MS related topics and research performed during PhD (upper part) and milestones of important Flemish/European/US
standard methods for the determination of elements in Environmental matrices (lower part) [28,29].
1800
1600
1400
number of publications
1200
1000 ICP-MS
ICP-MS speciation
800
ICP-MS environmental
600 ICP-MS environmental speciation
400
200
0
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
Figure 10: Number of published papers/year from 1980 to 2011 on” ICP-MS”,” speciation using ICP-MS”,” ICP-MS in environmental topics” and “speciation using ICP-MS in
environmental topics” (web of science).
25
analysis since 1980
1.4.3. ACID DIGESTION OF ENVIRONMENTAL MATRICES
Sample preparation is usually needed in chemical analysis to convert analytes to a more suitably
detectable form, to separate them from the sample matrix or to concentrate species for trace
analysis [32]. This step is often pointed out as the Achilles’ heel in chemical analysis, in view of the
risks of analyte losses, sample contamination and incomplete sample decomposition or analyte
extraction. In addition to these sources of systematic errors, sample preparation is often the most
time-consuming step in analytical procedures. This contrasts with the impressive development of
modern instrumental techniques that significantly improved detectability and reduced the analysis
time. In spite of this development, sample treatment is still needed to improve selectivity, to
minimize matrix effects and to avoid deleterious effects on the instrument (e.g., clogging of
nebulizers and irreversible retention of substances in chromatographic columns).
For most environmental analysis, acid digestion procedures are employed to completely transfer the
analytes of interest from the matrix into solution, so that they can be introduced into the
determination step (ICP-MS, ICP-AES, AAS) in liquid form. The goal of every digestion process is
therefore the complete dissolution of the analytes and the complete decomposition of the matrix,
while avoiding loss of or contamination with analytes. In this context, wet chemical digestions
utilizing various mineral acids are carried out in either an open system, that is, under atmospheric
pressure, or in closed vessels. Further, the samples may be heated in convection or microwave
ovens. Below, the acids (or combination of acids) typically used for environmental regulatory analysis
are summarized. The cost of the acid increases with the required purity grade. For most
environmental applications, the cost of the acid solution varies between 40 – 150 €/l acid. However,
for ultra-trace element determination the price of the highest purity grade acids vary between 300 to
800 €/l (year 2012, prices by courtesy of Merck, Fisher Scientific and Baker).
use
Nitric acid (HNO3) Oxidizing agent (CH2)n+2 HNO3 → CO2+2NO+2H2O
Frequently mixed with H2O2 or HCl, HF
Forms soluble nitrates with most elements
Hydrochloric acid (HCl) Non-oxidizing acid
Forms soluble chlorides with most elements
Hydrofluoric acid (HF) Non-oxidizing acid
Decomposition of silicates SiO2+6HF→H2SiF6+2H2O (excess required, otherwise
loss of e.g. BF3, SiF4, SeF4).
Complexing required to mask fluorides prior to further use:
H3BO3+4HF→HBF4+3H2O
Hydrogen peroxide (H2O2) In combination with nitric acid, increased oxidation potential 2H2O2→2H2O+O2
-
Reoxidizes NOx into NO3 (and suppresses the formation of the yellow nitrous
oxides typical of nitric acid)
Nitro hydrochloric acid HCl:HNO3 = 3:1
(aqua regia) Forms NOCl and releases chlorine as the active component: 2NOCl→2NO+Cl2
Digestion of precious metals, sulphides
Considering that Aqua regia digestion was said to be “discovered” around 800 AD by the Persian
alchemist Jabir ibn Hayyan (Geber), little has changed the last 30 years in terms of acid mixtures
employed and the degree of total decomposition achieved [26]. Nevertheless, sample preparation
has always been one of the great challenges in environmental analysis, mainly because of the highly
refractory nature of some mineral phases. Beginning in the 1970s, microwave radiation has been
successfully used to speed up heat transfer and to improve both safety conditions and control of
chemical reactions at high pressures and temperatures [32]. In addition to more efficient
decomposition, microwave-assisted procedures have contributed to minimize the consumption of
toxic reagents and, consequently, the amount of residues. Due to improvements in the reaction
flasks, with the development of safer, more resistant materials and designs, it was possible to
perform sample decompositions at higher temperatures and pressures. In these conditions, the
oxidizing power of HNO3 greatly increases and dangerous reagents (e.g., HClO4) are not required, not
even to decompose samples with high carbon content. High-performance closed vessel microwave-
assisted digestion systems are now the norm in many trace elemental laboratories. It must be noted
however that performance of such systems has taken twenty or so years to evolve to the current
level of maturity. Further advances are on the way with the next generation of microwave-assisted
high-pressure autoclave systems, which will offer even superior performance.
Aqua regia leaching of soils remains the standard approach. Here, given that such work is generally
carried out in commercial laboratories, sample throughput is a key issue and as such, hot block
digestion techniques remain dominant. High sample throughputs remain an issue for microwave
digestion systems. The emerging dominance of inductively coupled plasma-mass spectrometry (ICP-
MS) as the technique of choice meant that fusions were no longer fashionable because the high
dissolved solids content of the resulting solution, even after dilution, led to cone blockage. With
improved tolerance to high TDS in ICP-MS, fusion procedures are now, once again, in routine use in
geochemical and environmental laboratories when complete dissolutions are required [30].
At present, awareness of green aspects has led analytical chemists to consider key indicators (e.g.,
operation time, safety, volume/concentration of solvents and energy consumption) when developing
new sample preparation methods. A drawback of sample preparation is the generation of large
amounts of toxic wastes, because classical procedures consume large amounts of acids (for sample
decomposition) and organic solvents (for analyte extraction and sample clean-up). Indeed, sample
treatment is usually the analytical step that yields most residues and demands most energy [32].
These wastes need to be suitably managed and treated in order to avoid environmental
contamination, so taking time and increasing the cost of analysis. In the context of green analytical
chemistry, ultrasound-assisted solid-sample pretreatment and digestion using solely nitric acid is
being increasingly evaluated. The same holds true for speciation analysis, where soft extraction
under mild conditions is generally necessary [33,34].
ICP-MS applications in environmental regulatory monitoring 27
CHAPTER 2 ICP-MS APPLICATIONS IN ENVIRONMENTAL REGULATORY

MONITORING
2.1. PERSPECTIVE ON ELEMENTAL ANALYSIS
Given the importance of clean water, air and soil for human use and environmental quality, it is not
surprising that major chemical pollution issues have been important drivers for research in
environmental chemistry [41]. Large issues driving environmental research since the beginning of the
“environmental era”, starting in 1960s, were nutrient over enrichment of lakes (eutrophication), the
proliferation of pesticides (DDT), acid rain (or acid deposition), atmospheric transport of mercury,
harmful disinfection by-products in drinking water, e-waste disposal, ….
The term heavy metals is commonly used to address adverse effects of elements in our environment.
Although IUPAC (International Union of Pure and Applied Chemistry, i.e., the world authority on
chemical nomenclature) explicitly criticized the use of the term heavy metals as ambiguous and
pointless, this has not banned the term itself in popular and scientific literature [42]. The oldest
recorded scientific use of the term dates back to 1936 (related to elemental density), but today it
represents an ill-defined umbrella term for various elements (primarily transition metals, e.g., Cr, Ni),
but also some non-metals (e.g., Se, As). Because of this ambiguity, the term heavy metals will be
written in italics in the following paragraphs and considered as a term which is simply defined as a
set of elements frequently used in legislation without a widely accepted scientific rationale.
Heavy metals have unique characteristics that should be considered when assessing their
environmental risks. Unlike most organic substances, heavy metals are neither created nor destroyed
by biological or chemical processes. Rather, they are transformed from one chemical form into
another. Because heavy metals are naturally occurring, many organisms have evolved mechanisms to
regulate their accumulation and storage. Moreover, some elements are essential nutrients, so, when
they are not present in sufficient concentrations, growth, survival and reproduction of the organisms
can be affected. Table 1 summarizes the essentiality status for some environmentally relevant
elements [43]. These features, along with the fact that elements naturally occur as inorganic forms in
environmental compartments and are cycled through the biotic components of an ecosystem,
complicate the evaluation of toxicity data for inorganic substances.
Essential Non-essential
Cr, Co, Cu, Fe, Mn, Mo, Ni, Se, Zn As, Sb, Cd, Pb, Hg, Tl, Ag, Sn
Table 1: Essentiality of elements to living organisms.
Generally spoken, manmade dispersion of heavy metals in the environment takes place via
discharges in the air in the form of dust particles, as a result of discharge into the surface water or
soil and via accidental spill. Heavy metals end up on (and in) the soil as a result of atmospheric
deposition, by dumping waste or through the use of fertilizers. Once present in the soil, they can
permeate down into the groundwater. However, binding processes delay the movement of elements
into the soil to a significant extent. They can pollute the surface water as a result of run-off. Heavy
metals in the surface water can quickly settle in the watercourse sediment, where they may be
released into the surface water again over a long period.
28 ICP-MS applications in environmental regulatory monitoring
Given the significance of manmade dispersion of heavy metals in our environment, major legal acts
in the EU (e.g., Water Framework Directive, Air Quality Directive) have included monitoring programs
in order to report on the state of our environment. In the context of risk assessment however, one
must keep in mind that heavy metals are naturally occurring constituents of our environment that
vary in concentrations across geographic regions. In other words, as heavy metals are naturally
present in our environment, knowledge of (regional) ambient background concentrations is needed
to provide baseline data for pollution and risk assessment studies.
Along with the implementation of the regulatory monitoring requirements in the different European
member states, the use of analytical methods developed by the European Committee for
Standardization (CEN) was recommended (see also § 1.3). With one common standard in all these
countries and every conflicting national standard withdrawn, comparison between analytical results
becomes more feasible. A beneficial side-effect, taking into account that since 1980, the
management of the environment in Belgium became a task for the regions, is that the European
environmental regulatory monitoring requirements not only triggered a harmonisation among the
different European member states, but also led to a “rapprochement” in the internal Belgium
situation.
As a case study, the impact of the European environmental regulatory monitoring requirements in
Flanders will be discussed in this chapter (see Figure 11). As far as available, only monitoring studies
using European Standard methods (EN) based on ICP-MS instrumentation were considered.
Ostend Ghent Antwerp
Figure 11: Map situating Flanders with placemarks of some important cities (courtesy of Google Earth).
The question “How low is low and how low do we need to go?” can be adequately answered by ICP-
MS in environmental regulatory monitoring of elements [44](see also § 1.4.2). It is argued that the
risk of sample contamination nowadays represent a far greater challenge than adequate
instrumental limits of detection for regulatory environmental monitoring. In analogy with speciation
analysis, the measurement itself is often no longer the limiting step, but the sampling and sample
pretreatment (e.g., extraction, leaching, digestion) are considered the most critical steps for accurate
trace environmental monitoring. Moreover, it is argued that hindrance to the implementation of the
European monitoring requirements is not the technical feasibility of analysis at these concentration
levels, but rather communication, knowledge exchange and harmonization among laboratories [45].
In the following paragraphs, an inventory of regulatory monitoring studies (courtesy of VMM and
OVAM) on trace elements in the Flemish environment (soil, water, air) is provided. All of these
measurements were carried out with the aim to define Flemish regulatory environmental limits or to
investigate the Flemish environmental situation with respect to European regulatory environmental
limits. For the derivation of ambient background concentrations of heavy metals, only rural (non-
polluted) background locations were selected. Only for soil the monitoring study was performed at
VITO, for the other compartments, the monitoring studies were performed or commissioned by
VMM.
In summary, the focus in this chapter is on the analysis (of acid digests) of environmental samples
using European Standard methods (EN) based on ICP-MS instrumentation. These examples will show
that the sensitivity combined with multi-element capabilities of ICP-MS played (and still plays) an
important role in this context.
Interesting links:
• MIRA - Flanders Environment Report - offers an overview of the state of the environment in
Flanders divided into different environmental themes [40] (since 1994):
http://www.milieurapport.be/
• The European Environment Agency (EEA) coordinates the European environment information
and observation network by providing information on the different environmental issues since
1994: http://www.eea.europa.eu/themes
Statistics :
Box-and-whisker plots were used for visualizing the background concentration ranges that were
determined in the different monitoring studies. Box-and-whisker plots display differences between
populations without making any assumptions of the underlying statistical distribution: they are non-
parametric. The median/quartile/range type of plots used in the dissertation were computed using
Statistica (StatSoft, Inc. (2011), STATISTICA (data analysis software system), version 10. www.statsoft.com)
and are a convenient way of graphically depicting background concentration ranges.
As an example, the median/quartile/range plot of ambient background concentrations of Zn in wet

atmospheric deposition in 75 samples collected at Koksijde during 2010-2011 and analysed after acid
digestion with ICP-MS are visualized below.
The median/quartile/range plot describes the central tendency of the variable (Zn) in terms of the median
of the values (represented by the smallest box in the plot). The spread (variability) in the variable values is
th th
represented by the quartiles (the 25 and 75 percentiles, larger box in the plot, height = H) and the non-
outlier range (the "whiskers" in the plot). The non-outlier range is the range of values that fall below the
upper outlier limit (i.e., + 1.5 * the height of the box, H) and above the lower outlier limit (i.e., -1.5 * the
th
height of the box, H). The upper value of the box in the box plot (i.e., the 75 percentile) is indicated as
th
UBV, the lower value of the box in the box plot (i.e., the 25 percentile) is indicated as LBV.
The spacings between the different parts of the box help to indicate the degree of dispersion (spread) and
skewness in the data, and identify outliers and extremes. Values "far" from the middle of the distribution
are referred to as outliers and extreme values if they meet the conditions specified below.
A data point is considered to be an outlier value if the following conditions hold (the diagram on the right
illustrates the ranges in the box and whisker plots) :
data point value > UBV + 1,5*(UBV - LBV) or data point value < LBV – 1,5*(UBV – LBV).
A data point is considered to be an extreme value if the following conditions hold:

data point value > UBV + 3 *(UBV - LBV) or data point value < LBV – 3 *(UBV - LBV).
2.2. AMBIENT BACKGROUND CONCENTRATION OF ELEMENTS IN SOIL IN FLANDERS
Soil is defined as the top layer of the earth’s crust. It is formed by mineral particles, organic matter,
water, air and living organisms. As interface between the earth, the air and the water, the soil
performs many vital functions: food and other biomass production, storage, filtration and
transformation of many substances including water, carbon, nitrogen. Soil has a role as a habitat and
gene pool, serves as a platform for human activities, landscape and heritage and acts as a provider of
raw materials [46].
Trace metals in surface soils are derived from both parent materials and anthropogenic activities.
Because of the latter, it is often difficult to quantify the natural background concentrations of metals
in soils [47-49]. In fact, it can be argued that natural background no longer exists on this planet
because of human influence, and this is particularly true for densely populated and industrialised
regions like the Flemish region of Belgium. Therefore, the usual, or ambient, concentration of a metal
in soil consists of both a natural geochemical fraction and an anthropogenic fraction [50]. The
anthropogenic fraction refers to moderate diffuse inputs into the soil, not the inputs from local point
sources that generally result in a much elevated concentration.
At the moment, only nine EU member states have specific legislation on soil protection (especially on
contamination) [46]. Different EU policies (for instance on water, waste, agriculture) are contributing
to soil protection. But as these policies have other aims and other scopes of action, they are not
sufficient to ensure an adequate level of protection for all soil in Europe. In anticipation of a
European Soil Framework Directive with the objective to protect soils across the EU, the Flemish
Government pioneered in 1996 with the first order of the Flemish Government establishing the soil
remediation and protection regulations (VLAREBO, 1996). In this context, soil quality reference
values were needed to be developed as a basis to identify and assess soil contamination processes at
regional level. Hence, there was a need to establish levels of trace elements currently found in
common (clean) soils.
For the first soil remediation and protection regulation (VLAREBO, 1996), data on soil characteristics
and total trace element contents of soils in Flanders were compiled from various sources [47,51]. The
derived median ambient background concentration of trace elements (determined as aqua regia
extractable amounts) in Flemish soil are summarized in Table 2 (see column top soil (0-20 cm)
VLAREBO 1996)[47]. Baseline values for a standardized soil (i.e., 10 % clay, 2 % organic material) were
derived as the 90th percentile of the top soil measurements and were included in the VLAREBO
legislation of 1996 (see column baseline values VLAREBO 1996).
In 2002, the Public Waste Agency of Flanders (OVAM), responsible for waste management and soil
remediation in Flanders, decided to harmonise the acid digestion procedure for waste and soil
analysis. For the acid digestion of soil, it was decided to use the same digestion method as for waste,
i.e., Flemish reference method CMA/2/II/A.3 (see also § 1.3), which is based on European standard
EN 13656, developed by CEN/TC 292 (Characterization of waste)[52].
As this reference method included the use of HF (besides HNO3 and HCl), there was a need for
revision of the baseline values for heavy metals as defined in VLAREBO (1996), which were derived as
aqua regia extractable amounts (only HNO3 and HCl). The choice of an acid digestion procedure can
have an effect on the accuracy and precision attainable in the multi-elemental analysis of soil [54,55].
In the case of aqua regia digestion, complete solubilization of refractory minerals is not achieved
(e.g., illustrated for Cr, see Figure 13, where only ~ 80 % is recovered as compared to full recovery
with the supplementary use of HF).
The following paragraphs describe how the (current) baseline values for heavy metals, as defined in
the soil remediation decree of 2007, were derived.
→ Background sampling locations
For the determination of the baseline values for elements according to acid digestion reference
method CMA/2/II/A.3 (EN 13656), 45 top soil samples (0-20 cm and 50-100 cm) and 45 subsoil
samples (tertiary geological period) were collected in Flanders in 2006 (see Figure 12)[53].
Figure 12: Locations of soil sampling sites used for the determination of baseline values (courtesy of Google
Earth).
→ European regulatory monitoring method
The methods used for the determination of elements in soil were based on:
• EN 13656:2003 Characterization of waste – Microwave-assisted digestion with hydrofluoric
(HF), nitric (HNO3) and hydrochloric (HCl) acid mixture for subsequent determination of
elements in waste
• EN 13657:2002 Characterization of waste - Digestion for subsequent determination of aqua
regia soluble portion of elements (EN 16174:2012 Sludge, treated biowaste and soil -
Digestion of aqua regia soluble fractions of elements)
inductively coupled plasma-optical emission spectrometry (ICP-OES)
• CEN/TS 16175-2:2013 Sludge, treated biowaste and soil - Determination of mercury - Part 2:
Cold vapour atomic fluorescence spectrometry (CV-AFS)
The concentrations of As, Cd, Co, Cr, Cu, Hg, Mo, Ni, Pb, Se, V and Zn in these soil samples were
determined after microwave extraction according to the European standard digestion methods EN
13656 (HF/ HCl /HNO3) and EN 13657 (HCl /HNO3).
In short, standard procedure EN 13656 (translated as Flemish reference method CMA/2/II/A.3)

involves sieving of the soil sample with a 2 mm sieve, drying and milling. About 0,5 g of the sample is
weighed in a digestion vessel; 6 ml of 12 M HCl, 2 ml of 16 M HNO3 and 4 ml of 29 M HF are added
and the sample is subsequently digested in a microwave system. At the end, 44 ml of 4% (m/v) boric
acid is added (boric acid masks fluorides prior to further analysis, H3BO3 + 4 HF → HBF4 + 3 H2O), the
digestion solution is transferred to a volumetric flask of 100 ml and made up to volume.
The digestion according to EN 13657 involves the addition of 6 ml of 12 M HCl and 2 ml of 16 M HNO3
(aqua regia). In the case of aqua regia digestion, complete solubilization of refractory minerals is not
achieved (e.g., for Cr, see Figure 13), but trace element concentrations can be determined without
severe matrix interferences in a 1/5 dilution of the digestion solution by ICP-MS.
180
160
Cr (mg/kg) HNO3 /HCl and SF-ICP-MS
140
120
100
80
60
y = 0,8207x - 3,711
R² = 0,9539
40
20
0
0 20 40 60 80 100 120 140 160 180
Cr (mg/kg) HNO3 /HCL/HF and ICP-OES
Figure 13: Comparison between determination of Cr in soil after aqua regia and HNO3/HCl/HF digestion (data
courtesy of OVAM [53]).
In the case of HNO3/HCl/HF digestion and when using ICP-MS, the dilution factor needed to suppress
matrix effects due the high salt content (boric acid) is so large that sensitivity becomes comparable
with a (matrix-)robust technique such as ICP-AES. In ICP-MS procedures, it is generally recommended
to keep the level of total dissolved solids below 0.2 % (2000 mg/l) to minimize deposition of solids in
the sample introduction system, the plasma torch or the cone apertures of the ICP-MS instrument.
For this reason, the determination of elements in the top and subsoil samples following HNO3/HCl
digestion was realized by ICP-SFMS and following HNO3/HCl/HF digestion by ICP-AES.
The ICP-SFMS procedure involves a 1/5 dilution of the aqua regia digestion solution prior to analysis.
The elements As, Se and Zn were measured in high resolution mode, the remainder of the elements
in medium resolution mode. Detection limits were in the order of 10 – 50 µg/kg soil (calculated as 3 x
s on 10 procedure blank solution analyses and corresponding to an actual measurement of 10 – 50
ng/L in the diluted digestion solution) [53].
The trueness of the method was controlled by checking the elemental concentrations with the
reference values for the aqua regia soluble concentration certified reference material BCR-141R
(Trace elements in calcareous loam soil), BCR-144R (Sewage Sludge) and BCR-146R (Sewage sludge).
The recovery for the certified elements (Cr, Co, Ni, Cu, Cd, Pb and Zn) ranged between 94 and 103 %,
which is in line with other reported ICP-MS data on these reference materials [56].
The precision expressed as 95 % confidence interval is dependent on the concentration level in soil
analysis. According to multiple soil analyses within the context of validation of CEN/TS 16171 (Sludge,
treated biowaste and soil - Determination of elements using inductively coupled plasma-mass
spectrometry), a median precision of 7 % in the range > 10 mg/kg, 12 % in the range 1-10 mg/kg and
18 % in the range < 1 mg/kg was derived [57].
→ Background concentration of elements in soil
The median concentration of elements determined in Flemish soil after HNO3/HCl digestion followed
by sector-field ICP-MS analysis and after HNO3/HCl/HF digestion followed by ICP-AES analysis are
summarized in Table 2 for the 45 top soil samples (0-20 cm and 50-100 cm) and 45 subsoil samples
(tertiary geological period) [53]. As soil type has a major influence on the concentrations of elements,
statistical analyses on the soil geochemical data (pH, clay content, organic carbon content, effective
cation exchange capacity) were performed with the aim of estimating local or soil-type specific
ambient background concentrations for the trace elements.
As the reference digestion method includes the use of HF since 2002, baseline values for a
standardized soil (i.e., 10 % clay, 2 % organic material) were derived in Flanders as the 90th percentile
of the top soil measurements using HNO3/HCl/HF digestion and these were included in the VLAREBO
legislation of 2007 (see Table 2)[53]. For the first soil remediation and protection regulation
(VLAREBO, 1996), the baseline values were derived as the 90th percentile on trace elements
determined as aqua regia extractable amounts in Flemish soil [47]. For comparison, also median
values for total concentration of elements in European topsoil (0-20 cm) from the Forum of European
Geological Surveys (FOREGS) Geochemical Baseline Mapping Programme are represented [58].
Top soil Baseline Baseline

Top soil Top soil Subsoil
(0-20 cm) European values values
(0-20 cm) (50-100 cm ) (tertiary)
VLAREBO topsoil VLAREBO VLAREBO
VLAREBO 2007 VLAREBO 2007 VLAREBO 2007
1996 2007 1996
HNO3/HCl HNO3/HCl/HF HNO3/HCl HNO3/HCl HNO3/HCl HNO3/HCl/HF HNO3/HCl HNO3/HCl/HF HNO3/HCl
ICP-SFMS ICP-AES ICP-SFMS ICP-SFMS ICP-AES
mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg
As 7 11 5.9 12 8.7 7.6 7 7 16 19
Cd 0.3 <0.5 0.1 <0.5 0.05 <0.5 0.18 0.15 0.7 0.8
b
Co 5 7 5 6 7 6 0.05 7.8
Cr 25 34 24 40 62 82 24.6 60 62 37
Cu 11 16 5 9 3 5 9.6 13 20 17
a a a a a c
Hg 0.06 0.06 0.03 0.03 0.01 0.17 0.04 0.1 0.55
Mo 0.4 2 0.1 2 0.3 1.5 0.62
Ni 7 9 11 10 12 14 3.5 18 16 9
Pb 26 34 8.25 17 6.1 3.4 21.5 23 31 40
Sb 0.5 <2 0.1 <2 0.3 <2 0.6
Se 0.3 <5 0.1 <5 0.3 <5
V 37 48 26 34 62 70 60
Zn 56 60 22 37 29 34 34.5 52 77 62
th
Table 2: Median ambient background concentration and regulatory baseline values (90 percentile) for trace
elements in soil.
a
cold vapour-AFS was used for the determination of Hg according to CEN/TS 16175-3.
b
for Co a large discrepancy was observed between the median value (0.05 mg/kg) and the average value (0.42 mg/kg)[47].
c
a baseline level of 0.1 was derived in Flemish soils according to ref. 51.
Top soil (0-20 cm) VLAREBO 2007: median elemental concentrations in 45 Flemish top soil samples (0-20 cm) following HNO3/HCl
digestion and analysis by sector-field ICP-MS and HNO3/HCl/HF digestion and analysis by ICP-AES [53].
Top soil (50-100 cm) VLAREBO 2007: median elemental concentrations in 45 Flemish top soil samples (50-100 cm) following
HNO3/HCl digestion and analysis by sector-field ICP-MS and HNO3/HCl/HF digestion and analysis by ICP-AES [53].
Subsoil (tertiary) VLAREBO 2007: median elemental concentrations in 45 Flemish subsoil samples (50-100 cm) following HNO3/HCl
digestion and analysis by sector-field ICP-MS and HNO3/HCl/HF digestion and analysis by ICP-AES [53].
Top soil (0-20 cm) VLAREBO 1996: median elemental concentrations in Flemish top soil samples (0-20 cm) following HNO3/HCl
digestion and analysis by flame atomic absorption [47].
European topsoil (0-20 cm): median values for total concentrations in the European topsoil (0-20 cm)[58].
baseline values (VLAREBO 2007): baseline values (90th percentile) for a standardized soil (i.e., 10 % clay, 2 % organic material) using
HNO3/HCl/HF digestion.
baseline values (VLAREBO 1996): baseline values (90th percentile) for a standardized soil (i.e., 10 % clay, 2 % organic material) using
HNO3/HCl digestion.
A regression and correlation analysis was performed on the data for each element obtained upon
HNO3/HCl digestion and ICP-SFMS analysis on the one hand and HNO3/HCl/HF digestion and ICP-AES
analysis on the other (e.g. see Figure 13). For the elements Mo, Sb, Cd and Se, the majority of the
concentrations were below the quantification limit of ICP-AES. For this reason, no baseline values
were defined for Mo, Sb and Se. For Cd, the baseline value of 0.7 mg/kg was derived as the 90th
percentile of the soil measurements (using HNO3/HCl/HF digestion) with a concentration level above
the detection limit of the ICP-AES method.
With the current soil digestion method (CMA/2/II/A.3, EN 13656), the determination of baseline
values with ICP-AES or ICP-MS is difficult due to severe matrix effects caused by the high salt content
of the acid digestion solution. When using aqua regia digestion, this matrix effect is less pronounced
and soil background concentrations of elements such as Cd, Sb and Se can be easily determined with
the superior sensitivity of ICP-MS as compared to ICP-AES.
As mentioned in § 1.3, it is to be expected that Hg will be included in the revised version of CEN/TS
16171 as one of the elements that can be determined with ICP-MS. This will certainly favour the use
of ICP-MS for soil, sludge and biowaste analysis in the future, as no separate Hg determination (with
e.g., CV-AFS) will be needed. Moreover and under impulse of the Scandinavian countries, nitric acid
is being promoted as the sole acid to be used for environmental sample digestion. This would further
play into the hands of ICP-MS as the method of choice for environmental regulatory monitoring.
The elemental concentrations in Flemish soils following HNO3/HCl digestion and analysis by sector-
field ICP-MS are represented in the figures below.
1,4
1,2
1,0
0,8
mg/kg
Median
0,6 25%-75%
Non-Outlier Range
Outliers
Extremes
0,4
0,2
0,0
Cd (0-20)
Cd (50-100)
Hg (0-20)
Hg (50-100)
Cd subsoil
Hg subsoil
3,0
2,5
2,0
mg/kg
1,5
Median
25%-75%
Non-Outlier Range
Outliers
1,0 Extremes
0,5
0,0
Mo (0-20)
Sb (0-20)
Sb (50-100)
Se (0-20)
Se (50-100)
Mo (50-100)
Se subsoil
Mo subsoil
Sb subsoil
50
40
30
mg/kg
Median
25%-75%
20 Non-Outlier Range
Outliers
Extremes
10
Ni (50-100)
As (0-20)
As (50-100)
Co (0-20)
Co (50-100)
Cu (0-20)
Cu (50-100)
Ni (0-20)
Ni subsoil
As subsoil
Co subsoil
Cu subsoil
200
180
160
140
120
mg/kg
100
Median
25%-75%
Outliers
Extremes
60
40
20
0
Cr (0-20)
Pb (0-20)
Pb (50-100)
Cr (50-100)
V (0-20)
V (50-100)
Zn (0-20)
Zn (50-100)
Cr subsoil
Pb subsoil
V subsoil
Zn subsoil
Figure 14: Elemental ambient background concentration in Flemish soils following HNO3/HCl digestion and
analysis by sector-field ICP-MS (45 top soil samples (0-20 cm and 50-100 cm) and 45 subsoil samples (tertiary
geological period), collected at different locations in 2006, see Figure 12 (data courtesy of OVAM [53]).
As can be derived from the median values of European geochemical baseline values taken from the
FOREGS Geochemical database, the concentrations of Cd in the top soil in Flanders are elevated in
comparison with the European median value (see also Table 2 and § 4.2)[58]. This is also visualized in
the map below, representing a geostatistical analysis of the FOREGS Geochemical database.
Figure 15: Map of Cd in topsoil samples in Europe from the FOREGS Geochemical database (concentrations
mg/kg) (courtesy of Google Earth)[58].
The assessment of risks to human health and the environment, related to exposure of chemicals in
agricultural soil and grazing land soil, is one of the requirements of the REACH regulation
(Registration, Evaluation, Authorisation and Restriction of Chemical substances). In order to provide
harmonised geochemical data of arable land and land under permanent grass cover at the European
scale, the GEMAS (Geochemical Mapping of Agricultural Soils of Europe) project was started in 2007
[59]. This geological survey was performed in 34 European countries, covering an area of
approximately 5.6 million km2 at a sample density of 1 site per 2500 km2, collecting one sample from
arable land (0–20 cm) and land under permanent grass cover (0–10 cm) each. Soil samples were
digested following HNO3/HCl digestion and analysed by ICP-MS (the final report and data set will be
freely available in December 2013)[59]. On average, there is a factor 6 difference in the median
concentrations of the elements among the countries sampled. Several elements (e.g., Ni) show an
even substantially larger difference up to a factor of more than 100. At the continental scale, the
occurrence of ore deposits and geology play the key role in determining the element distribution
patterns. When defining background value for any one element in future European soil legislation,
this will be an important factor to consider.
Interesting links:
• Heavy metals in European soils, a geostatistical analysis of the FOREGS geochemical database
using Google Earth. The maps present results of mapping concentrations of eight critical heavy
metals (arsenic, cadmium, chromium, copper, mercury, nickel, lead and zinc) using the 1588
georeferenced topsoil samples from the FOREGS Geochemical database. The concentrations were
interpolated over the 26 European countries that contributed to the database :
http://eusoils.jrc.ec.europa.eu/library/Data/Foregshmc/index.htm
• GEMAS aims to provide harmonised geochemical data of arable land and land under permanent
grass cover at the European scale (final report and data set will be freely available in December
2013) : http://www.bgs.ac.uk/gbase/GEMAS.html
• The Geochemical atlas of Europe, http://weppi.gtk.fi/publ/foregsatlas/maps_table.php
2.3. AMBIENT BACKGROUND CONCENTRATION OF ELEMENTS IN AMBIENT AIR IN FLANDERS
A clean air supply is essential to our own health and that of the environment. But since the industrial
revolution, the quality of the air we breathe has deteriorated considerably - mainly as a result of
human activities. Increasing industrial activity and energy production, the burning of fossil fuels and
the dramatic rise in traffic on our roads all contribute to air pollution in our cities which, in turn, can
lead to serious health problems [60,64].
2.3.1. AMBIENT BACKGROUND CONCENTRATION OF ELEMENTS IN FINE PARTICULATES IN FLANDERS
The issue of air quality is still a major concern for many European citizens. It is also one of the areas
in which the European Union has been most active. Since the early 1970s, the EU has been working
to improve air quality by controlling emissions of harmful substances into the atmosphere, improving
fuel quality, and by integrating environmental protection requirements into the transport and energy
sectors. As the result of EU legislation, progress has been made in tackling air pollutants such as
sulphur dioxide, lead, nitrogen oxides, carbon monoxide and benzene [65]. However, despite a
reduction in some harmful emissions, air quality continues to cause problems.
Fine particulates present a health risk which is of increasing concern in Flanders. It is therefore
important to identify the sources and causes of this type of pollution. To do so, the Flemish
Environment Agency (VMM) has carried out several particulate matter (PM) chemical
characterization studies, where hot spots and rural background sites were monitored [70,71,72]. The
following paragraphs describe how ambient background concentrations of elements were derived in
ambient air for Flanders.
A monitoring campaign consisting of element characterization in PM10 (particles with an aerodynamic

diameter less than 10 µm) was organized by VMM on different industrial and background locations.
Once a week, PM10 samples were collected during 2008-2009 at the background locations
Moerkerke, Aarschot and Retie and during 2010 at Retie (250 filter samples were collected in total).
Figure 16: Locations of rural background sampling sites for ambient particulate matter and atmospheric
deposition in Flanders (courtesy of Google Earth).
The method used for the determination of elements in PM10 was based on:
• EN 14902:2005 Ambient air quality - Standard method for the measurement of Pb, Cd, As,
and Ni in the PM 10 fraction of suspended particulate matter
Ambient air samples are typically collected for TSP (airborne particles or aerosols less than 100 µm),
PM10 (particles with an aerodynamic diameter less than 10 µm), or PM2.5 (particles with an
aerodynamic diameter less than 2.5 µm) analysis. Samples are collected using an apparatus that
incorporates an inlet, sample collection media, and air-sampling pumps intended to collect air for a
specified time period. Sampling inlet size and flow rate generally define the size of the particles
collected on the sampling medium. During sample collection, air is drawn through the sampling
device and particle matter is collected on filter media. Due to the generally low concentrations of
elements in air, it is critical to evaluate filter media for elemental content. Thus, a “blank” filter must
always be analyzed in conjunction with the air samples collected.
In Flanders, PM10 sampling is performed by sampling of 55 m³ air through a quartz filter (Ø 47mm)
over a period of 24 hours. On return to the lab, the filters are weighed to determine the total PM10
concentration. Subsequently, the filter (or parts of it) is (are) analyzed for trace elemental
composition by ICP-MS after HNO3/H2O2 digestion or elemental composition is measured on the
whole filter directly with X-ray fluorescence spectroscopy (XRF) [74-76].
According to EN 14902, the filter (or parts of it) is transferred into a microwave sample vessel, 8 ml of
16 M HNO3 and 2 ml of 10 M H2O2 are added, while ensuring complete submersion of the filter in the
digestion acid. The microwave is programmed so that the acid mixture reaches approximately 180 °C
within 20 min. Subsequently, the temperature is slowly increased to up to approximately 220 °C and
then held for about 20 min. After the digestion procedure, the vessels are allowed to cool down to
room temperature. The digestion solutions are transferred into labelled volumetric flasks of 100 ml
and diluted to the mark with ultrapure water.
The ICP-(SF)MS procedure involves a direct analysis of the digestion solution. In a validation study of
EN 14902 performed at VITO using ICP-SFMS, the elements As, Se and Zn were measured in high
resolution mode, the remainder of the elements in medium resolution mode [66]. Detection limits
calculated as 3 x s on 24 procedure blank solution analyses were in the order of 5 – 50 ng/filter
(corresponding to ~ 0.01 to 0.1 ng/m³) for Cd, Ag, V, As, Mn, Co, Mo, Cr, Ni, Sn, Sb, Tl and 50-500
ng/filter (corresponding to ~ 0.1 to 1 ng/m³) for Ba, Cu, Pb, Se, Zn, Fe and 5000 ng/filter
(corresponding to ~ 10 ng/m³) for Al. These instrumental detection limits cover possible
contamination during the acid digestion procedure and instrumental sensitivity. In order to verify the
possible contribution of the type of filter, 10 blank Teflon® and 10 blank quartz filters were digested
and analysed. When using Teflon® filters, no significant contribution was observed. However, when
using quartz filters, contamination was observed for Ni (30 ng/filter), Mn (60 ng/filter), Cr (130
ng/filter), Zn (700 ng/filter), Ba (1200 ng/filter) and Fe (1900 ng/filter).
The trueness of procedure EN 14902 has been controlled by checking the recovery of elements in
NIST 2584 - Trace Elements in Indoor Dust [66,67]. EN 14902 stipulates that the average recovery for
Cd and Pb should be between 90-110 % and for As and Ni between 85 – 115%. These requirements
were fulfilled for the average recovery of 30 analyses of NIST 2854 for Cd (98%), As (100%) and Pb
(93%), but not for Ni (66%). The value for Ni in NIST 2854 is an indicative value (not a certified one),
however the average recovery of Cr also amounts to 72% only, indicating that refractory elements
may not completely solubilise using the HNO3/H2O2 digestion and that the supplementary use of,
e.g., HF for complete recovery of Ni from PM10 is needed [68]. The average recovery for other
elements with indicative values amounted to: Mo (80%), V (80%), Mn (83%), Co (87%), Cu (92%), Sb
(29%), Ba (87 %), Zn (99%) and Se (102%)[66].
For comparison purposes, NIST 2854 was also brought into solution via microwave-assisted digestion
using HNO3 and HF (with addition of H3BO3 at the end) and measured by ICP-AES according to EN
14385 [69]. This European Standard specifies a reference method for the determination of the mass
concentration of specific elements in exhaust gases from hazardous and municipal waste
incinerators. The recovery for all elements was in the range 90-110 %, underlining the influence of
the acids used in the digestion procedure [66].
The precision expressed as 95 % confidence interval on PM10 is dependent on the concentration

level. Based on multiple analysis of NIST 2854 (n=30) according to EN 14902, a precision in the range
of 10 to 20 % can be derived for all elements, with the exception of Sb (32%)[66]. These results are in
line with reported performance characteristics of method EN 14902 based on a European
interlaboratory comparison exercise [76].
→ Background concentration of elements in PM10
The results of ambient background concentration of elements in PM10 in Flanders, based on 250
filters measured after acid digestion with ICP-MS according to EN 14902:2005, are summarized
below [71,72,73](data courtesy of VMM). The filters were collected at Moerkerke, Aarschot and
Retie.
3
Median
ng/m³
25%-75%
Non-Outlier Range
Outliers
2 Extremes
0
As Cd Mo Sb
24
22
20
18
16
14
Median
ng/m³
25%-75%
Outliers
10 Extremes
0
Ba Cr Cu Mn Ni Pb Ti V
500
400
300
Median
ng/m³
25%-75%
Non-Outlier Range
Outliers
200 Extremes
100
0
Al Ca Fe K Zn
Figure 17: Ambient background concentration of elements in PM10 in Flanders, determined with ICP-MS after
acid digestion according to EN 14902 (250 PM10 samples collected at 3 different locations in 2009-2010, Figure
16, data courtesy of VMM [71-73]).
The following median values can be derived for Flemish ambient background concentrations of
elements in fine particulates. The yearly median PM10 value at rural background sites in Flanders
amounts to 20 µg/m³ (the annual European limit value for the protection of human health is 40
µg/m³) [71,72].
Al As Ba Ca Cd Cr Cu Fe K Mn Mo Ni Pb Sb Ti V Zn
ng/m³
Median 57 0.61 2.7 160 0.20 2.0 5.0 160 120 4.3 0.44 1.8 8.1 1.1 3.6 2.2 22
AQL 6a 5a 20a 500b
(a) Air Quality Limit measured as contents in PM10, target value in force since 31/12/2012
(b) Air Quality Limit annual mean, measured as contents in PM10, limit value in force since 1/1/2005
Table 3: Median ambient background concentrations of elements in PM10 in Flanders (data courtesy of VMM
[71-73]).
For baseline value determination of elements, the sensitivity of ICP-MS is needed, especially for the
determination of cadmium. Detection limits attainable with EDXRF and WDXRF are in the order of
150 and 50 ng Cd/filter, respectively [62,74,75]. This corresponds to ~ 1 ng/m³ (WDXRF) and is a
factor of 10 higher than the detection limit obtained according to method EN 14902 using ICP-SFMS
(and well above the median background value of 0.2 ng Cd/m³). For the determination of the other
elements for which European Air quality limits have been defined so far (As, Pb and Ni), the
sensitivity of EN 14902 using ICP-SFMS or EDXRF are both fit for purpose. However, in analogy to soil
digestion, the HNO3/H2O2 digestion (EN 14902) may not completely solubilise refractory elements,
e.g., Ni, and the use of HF for complete recovery from PM10 may be needed [68].
According to the Air quality in Europe report 2011 (European Environment Agency), the annual mean
concentrations of Cd and As in PM10 in Belgium seems elevated in comparison with that in other
European member states (see Figure 18)[60]. However it must be noted that the monitoring stations
in Belgium are a mix of traffic, industrial and rural or urban background areas. Differences in
locations of the monitoring stations between member states are of concern and interpretation or
comparison of the chemical status should therefore be done carefully.
Figure 18: Annual mean concentrations of As (left) and Cd (right) in PM10 in 2009 in Europe (figure courtesy of
European Environment Agency [60]).
2.3.2. AMBIENT BACKGROUND CONCENTRATION OF ELEMENTS IN ATMOSPHERIC DEPOSITION IN FLANDERS
Total atmospheric deposition of elements, which is defined as the sum of wet and dry deposition,
can be estimated using wet-only and bulk collectors [63]. The wet-only collector is designed to collect
only sedimenting wet particles, while the bulk collector is designed to collect all sedimenting wet and
dry particles.
As atmospheric mercury exists mainly in the form of elemental mercury vapour (Hg0)(90 to 99%),
besides particle-bound mercury (< 5%) and gaseous divalent mercury (e.g., HgCl2)(<5%), atmospheric
deposition occurs mainly via dry deposition or wash-out of particle-bound and gaseous divalent
mercury. Mercury in the form of elemental vapour (Hg0) has a long atmospheric lifetime, which
makes transport on hemispheric and global scales feasible. In the northern Hemisphere,
anthropogenic emissions have increased the background concentrations of mercury in air by a factor
of 2-3 since before industrialization. Within the context of the compulsory measurement CAMP
(Comprehensive Atmospheric Monitoring Program) within the OSPAR Convention (Convention for
the Protection of the Marine Environment of the North-East Atlantic), weekly measurements of Hg in
wet deposition are carried out at different background locations in Europe [82].
The following paragraphs describe how ambient background concentrations of elements were
derived in atmospheric deposition for Flanders.
In Flanders, total atmospheric deposition of elements for ambient background concentration is

estimated with wet-only collectors at Koksijde on a weekly basis and with bulk collectors at Koksijde
and Bonheiden on a monthly basis (see Figure 16)[61]. Wet deposition of mercury is sampled on a
weekly basis at Koksijde and background concentrations of total gaseous mercury are continuously
monitored at Houtem (Veurne) with an automated mercury analyser [84].
The methods used for the determination of elements in atmospheric deposition were based on:
• EN 15841:2009, Ambient air quality - Standard method for determination of arsenic,
cadmium, lead and nickel in atmospheric deposition
• EN 15853:2010, Ambient air quality - Standard method for the determination of mercury
deposition
• EN 15852:2010, Ambient air quality - Standard method for the determination of total
gaseous mercury
For bulk sampling, VMM is currently in transition to change to the European standard EN 15841,
which slightly differs from the Belgian standard NBN T94-101 (1976)[78]. EN 15841 specifies methods
for sampling wet-only and bulk deposition of As, Cd, Ni and Pb. The samples of the bulk and wet-only
collectors are transferred to the laboratory in the sampling bottle and acidified with nitric acid (to 1
v/v %). A test portion is subsequently microwave-digested by adding suitable volumes of nitric acid
and analysed by graphite furnace atomic absorption spectrometry (GF-AAS) or by ICP-MS. Since
2009, samples are analysed with ICP-MS by VMM [61].
Detection limits calculated as 3 x s on replicate measurements of blank collectors are in the order of
0.05 – 0.2 µg/m².day [62]. Based on duplicate sampling of bulk depositions at Koksijde, a precision
(95 % confidence interval) of better than 20 % is achieved for all elements in 80 % of the monitored
depositions. However, in the other cases, a difference up to a factor of 4 was observed in the
duplicate measurements, illustrating that contamination (e.g., Cu, Zn) represents a challenge for
accurate trace environmental monitoring.
Weekly measurements of Hg in wet deposition at Koksijde are carried out according to EN 15853.
The (wet only) precipitation sample is stabilized with hydrochloric acid and transferred to the
laboratory in the collection vessel. Mercury in the precipitation sample is oxidised using bromine
monochloride and subsequently analysed by an atomic fluorescence spectrometer equipped with a
gold amalgamation system for Hg preconcentration (Leeman Labs, Hydra AF) [82]. The method
detection limit amounts to 0.001 µg/m².day with a precision (95 % confidence interval) of less than
10 %. Based on duplicate sampling, a combined uncertainty (ca. 95% confidence level) of 53 % was
derived [82].
Background concentrations of total gaseous mercury are monitored according to EN 15852 (2010,
Ambient air quality - Standard method for the determination of total gaseous mercury) with an
automated mercury analyser, consisting of a gold amalgamation system coupled to an atomic
absorption spectrometer [84].
→ Background concentration of elements in atmospheric deposition
The results for the bulk and wet-only collectors in 2010-2011 at background locations in Flanders,
analysed after acid digestion with ICP-MS, are summarized in the figures below [61]. The median
values of bulk and wet deposition in 2010-2011 at rural background sites in Flanders are summarized
below.
As Cd Cr Cu Fe Mn Ni Pb Zn
µg/m².day
Median bulk deposition 0.37 0.10 0.94 4.6 160 23 1.1 4.6 21
Median wet deposition 0.21 0.06 0.40 9.0 21 8.1 0.67 1.4 12
Table 4: Median values of bulk and wet deposition at rural background sites in Flanders (data courtesy of VMM
[61,62]).
Especially for cadmium and arsenic, the superior sensitivity of ICP-MS is needed for ambient
background determination of these elements in atmospheric deposition. However, duplicate
sampling of bulk depositions at Koksijde, where differences up to a factor of 4 were observed in 20 %
of the duplicate measurements, underline that the risk of sample contamination (e.g., Cu) may
represent a far greater challenge than adequate instrumental limits of detection for regulatory
environmental monitoring.
Based on the weekly measurements of Hg in wet deposition at Koksijde, the time-weighted average
mercury deposition for the year 2011 at Koksijde amounted to 0.025 µg Hg /m2.day [82]. The median
background value of daily measurement of total gaseous mercury in 2011 amounted to 0.7 ng/m³
[84].
3
µg/m².day
Median
25%-75%
Non-Outlier Range
Outliers
2 Extremes
0
Cd As Cr Ni
100
80
60
µg/m².day
Median
25%-75%
Non-Outlier Range
Outliers
40 Extremes
20
0
Cu Pb Zn Mn
Figure 19: Ambient background concentration of elements in total atmospheric deposition in Flanders,
analysed with ICP-MS after acid digestion (52 samples collected at Koksijde and Bonheiden in 2010-2011, data
courtesy of VMM [61,62]).
3
µg/m².day
Median
25%-75%
Non-Outlier Range
Outliers
2 Extremes
0
Cd As Cr Ni
100
80
60
µg/m².day
Median
25%-75%
Non-Outlier Range
Outliers
40 Extremes
20
0
Cu Pb Zn Mn
Figure 20: Ambient background concentration of elements in wet atmospheric deposition in Flanders, analysed
after acid digestion with ICP-MS (75 samples collected at Koksijde in 2010-2011, data courtesy of VMM
[61,62]).
In Figure 21, the total (dry and wet) deposition of cadmium across Europe in 2008 (g/km².year) is
presented [85]. Cadmium is primarily produced as a by-product from the extraction, smelting and
refining of zinc and other non-ferrous metals (see also § 4.2). It has predominantly been used in
rechargeable nickel-cadmium batteries, although sales to consumers have now been banned in
Europe, with the exception of certain uses. Cadmium is also used in the production of pigments,
coatings and platings. Emissions of cadmium to air arise primarily from combustion processes in
power plants and industry. For comparison, the median background value of bulk deposition for Cd
of 0.1 µg Cd/m².day derived above, corresponds to 36.5 g/km².year.
2
Figure 21: Total (dry and wet) deposition of cadmium across Europe in 2008 (g/km .year)(figure courtesy of
European Monitoring and Evaluation Programme [85]).
Cadmium emissions to air are ultimately deposited onto land or directly into fresh and marine
waters. Atmospheric deposition in urban areas will typically result in cadmium being washed from
impervious surfaces, collected and discharged to a receiving water, either directly or via a
wastewater treatment plant. Combined wet and dry deposition of cadmium across Europe is
variable, generally ranging between 10 and 50 g/km2.year but reaching in excess of 100 g/km2.year in
parts of central and south-eastern Europe (see Figure 21). Across the 32 European Economic Area
member countries, emissions of cadmium to air have declined significantly over recent years — by
about 58 % between 1990 and 2008 [85]. This decline reflects improvements in abatement
technologies at industrial facilities and, in some countries, the closure of older plants as a result of
economic restructuring.
2.4. AMBIENT BACKGROUND CONCENTRATION OF ELEMENTS IN RIVER SEDIMENT IN FLANDERS
Trace element concentrations in soils and sediments are mainly dependent on the nature of the
parent material from which they are derived, but anthropogenic influences can also alter trace
element concentrations in terrestrial and fluvial environments [89]. Consequently, the natural
background concentration of trace elements in sediments (as in soils, see § 2.2) can be used as a
reference value to estimate the contamination level of sediments. In the past, quantitative
relationships between the bedrock and the surrounding soil and/or sediments have been
established, but further development of the soil, weathering, erosion and transport processes can
alter this relationship between the soil and the underlying lithology.
The occurrence of background concentrations of trace elements in soils and sediments is, besides the
lithology, also influenced by their clay and organic matter content [47]. Therefore, clay and organic
matter content are often used to calculate ‘corrected’ background values for trace metal
concentrations in soils and sediments.
In an attempt to establish a regression model to explain trace element concentrations in Flemish

soils, Tack et al. found a rather poor relationship between total heavy metal concentrations in soils
and the clay- and organic matter content. This indicates that the independent variables (clay and
organic matter content) are not sufficient to explain trace element background concentrations in soil
and that the addition of other variables could probably improve the model. To overcome this
drawback of the “clay fraction”, data normalization towards “conservative elements” (Al, Fe, Li …) is
often carried out. In many studies, Fe and/or Al are used as “normaliser” to account for the natural
variability of trace elements in soils and sediments and a positive linear correlation between Fe and
trace element concentrations has been found in several studies [89].
In Europe, stream sediments have been widely used as a sampling medium for geochemical mapping.
Since 1985, the Forum of European Geological surveys (FOREGS) has been working on a geochemical
atlas of Western Europe to deduce background geochemical information and to assess
environmental pollution of floodplain and present-day river systems [90]. With the implementation
of European Directive 2008/105/EQS, the use of sediments for assessing long term trends and
chemical status was introduced in the European integral water policy [14].
Since 2000, the Flemish Environmental Agency is monitoring the sediment quality at 600 locations in
Flanders in both navigable and unnavigable waters, according to the Triad approach. This Triad
sediment quality assessment is an integrated investigation technique based on the analysis of
physico-chemical, biological and ecotoxicological parameters [91]. For the evaluation of the physico-
chemical parameters, a standardization to an organic matter content of 5% and a clay content of 11%
is done for heavy metals and organic contaminants.
The following paragraphs describe how ambient background concentrations of elements were
derived in river sediments for Flanders.
For the determination of median ambient background concentrations of trace elements, only those
locations with no acute impact on benthic biota (= ecotoxicological class 1) and good biological
quality (= presence of benthic macro invertebrates, biological class 1) were selected [92,93]. From
the 228 river sediment locations sampled during the 2010-2011 monitoring campaign (and measured
with ICP-MS by VMM), the selection criteria of no acute impact on benthic biota and good biological
quality were met at only 22 locations and these are visualized below [94].
Figure 22: Locations of river sediment and surface water sampling sites used for the determination of ambient
background values (courtesy of Google Earth).
The methods used for the determination of elements in river sediment were based on:
• EN 16174:2012 Sludge, treated biowaste and soil - Digestion of aqua regia soluble fractions
of elements
• EPA 7473:1998 Mercury in solids and solutions by thermal decomposition, amalgamation,
and atomic absorption spectrophotometry
The determination of elements in river sediments was performed by the Flemish Environmental
Agency (after aqua regia digestion) with ICP-AES until 2009, thereafter ICP-SFMS was used for the
determination [92]. With a turn-over time of 4 years, the 600 locations are sampled [93].
The trueness of procedure CEN/TS 16171 (ICP-MS) after aqua regia digestion was controlled by
checking the recovery of elements spiked to river sediments. The average recovery for the elements
checked, varied between 90-110 %. The precision of procedure CEN/TS 16171 after aqua regia
digestion on river sediments, expressed as 95 % confidence interval, is dependent on the
concentration level and in the range of 14-24 % for all elements [93].
→ Background concentration of elements in river sediment
Ambient background concentrations of elements in Flemish river sediment with no acute impact on
benthic biota and good biological quality are summarized in the figures below. The median elemental
background concentrations analysed by ICP-MS after aqua regia digestion during the 2010-2011
monitoring campaign are summarized in Table 5 [92,93].
Median values of elemental concentrations from the analysis of 105 river sediments in Flanders in
2006 are also summarised in Table 5 [89]. These analyses were performed to investigate the
relationship between heavy metals (As, Cd, Cr, Cu, Hg, Pb, Ni, Sn and Zn) and major elements and the
(geological) location. For comparison purposes, also the median values for aqua regia concentrations
in European stream sediments from The FOREGS Geochemical Baseline Mapping Programme are
represented [58].
10
6
Median
mg/kg
25%-75%
Non-Outlier Range
Outliers
4 Extremes
0
Hg Cd Se Sn
100
80
60
Median
mg/kg
25%-75%
Non-Outlier Range
Outliers
40 Extremes
20
0
Ni Cu As Pb Cr
2,4
2,2
2,0
1,8
1,6
1,4
Median
g/kg
25%-75%
1,2 Non-Outlier Range
Outliers
1,0 Extremes
0,8
0,6
0,4
0,2
0,0
Zn Na Mn P
Figure 23: Ambient background concentration of elements in river sediment in Flanders, analysed with ICP-MS
after aqua regia digestion according to CEN/TS 16171 (samples collected at 22 different locations in 2010-2011,
see Figure 22)(data courtesy of VMM [92,93]).
River sediment River sediment River sediment European river sediments

monitoring monitoring reference values
2010-2011 2006
HNO3/HCl HNO3/HCl HNO3/HCl
ICP-MS ICP-AES
mg/kg mg/kg mg/kg mg/kg
As 17 8 11 6
Cd 0.15 0.6 0.38 0.28
Cr 20 27 17 21
Cu 8.2 16 8 14
a
Hg 0.052 0.1 0.05 0.038
Mn 169 452
Ni 8.2 10 14 16
Pb 17 26 14 14
Se 1.0 1.5
Sn 1.3 1.4 2.25
Zn 125 131 67 36
Table 5: Median ambient background concentration and reference values
of trace elements in Flemish river sediment.
a
thermal decomposition, amalgamation, and atomic absorption spectrophotometry was used for the
determination of Hg according to EPA 7473.
River sediment monitoring 2010-2011: median elemental background concentrations in 22 Flemish river
sediments following HNO3/HCl digestion and analysis by ICP-MS [92,93].
River sediment monitoring 2006: median elemental concentrations in 105 river sediments, sampled and
analysed in 2006 [89].
River sediment reference values: reference values for a standardized river sediment (i.e., 11% clay, 5% organic
matter)[91].
European river sediments: median values (aqua regia) in European stream sediments from The FOREGS
Geochemical Baseline Mapping Programme [58].
For the elements Cd and Se, the superior sensitivity of ICP-SFMS, as compared to ICP-AES, is
necessary to determine background concentrations. An advantage of ICP-MS is that in addition to
trace elements at mg/kg level, also matrix elements at g/kg level can be determined in the same
measurement run (panoramic analysis). The total Fe and Ca contents are relevant parameters to
predict heavy metal concentrations in sediments, besides organic matter and clay content [89]. The
results of these matrix elements at the 22 selected background locations are summarized below.
80
70
60
50
Median
g/kg
40 25%-75%
Non-Outlier Range
Outliers
Extremes
30
20
10
0
Mg K Ca Al Fe
Figure 24: Ambient background concentration of matrix elements in river sediment in Flanders, determined
after aqua regia digestion with ICP-MS according to CEN/TS 16171 (samples collected at 22 different locations
in 2010-2011, see Figure 22) (data courtesy of VMM [92,93]).
2.5. AMBIENT BACKGROUND CONCENTRATION OF ELEMENTS IN SURFACE AND GROUND WATER IN

FLANDERS
Water has been denoted as “blue gold” and this is also recognized in the preamble of the European
Water Framework Directive where it is stated that: “water is not a commercial product like any other,
but rather a heritage, which must be protected, defended and treated as such”[95].
The increasing demand for cleaner rivers and lakes, groundwater and coastal beaches has been
evident for considerable time (since 1975, the need for water has doubled worldwide). This demand
is one of the main reasons why the European Commission has made water protection one of the
priorities of its work [96]. Since the adoption of the Water Framework Directive (WFD) in 2000, the
European Water Policy has undergone a thorough restructuring process. As opposed to earlier water
protection, which focused on single types of water, the area covered by this Directive extends to all
aquatic systems, surface waters (rivers and lakes), groundwater and coastal waters.
Monitoring Programmes play a key role in the practical implementation process of the European
Water Framework Directive. They form the basis for developing effective programmes of measure
that aim at reaching the WFD quality aims. The WFD includes a plan of action for the implementation
in the member states, setting out clear deadlines for each of the requirements, e.g., monitoring
programmes had to be implemented by the end of 2006.
The WFD identified the need to monitor water status throughout the Community on a systematic
and comparable basis. Accordingly, there was a need for comparability between results obtained, not
only by different laboratories, but also at different points in time or place. Under the 6th Framework
Programme, the EU R&D project SWIFT-WFD (Screening methods for Water data InFormaTion in
support of the implementation of the Water Framework Directive) was launched. The main objective
of this project was to support the successful implementation of the WFD, which closely depends on
the quality of monitoring data and its comparability form river basin to river basin. To do so,
interlaboratory studies with classical laboratory-based analyses were organised for parameters
regulated by the WFD for the assessment of laboratory performance in major component and trace
element determination [97].
In commission of the Flemish Environment, Nature and Energy Department (LNE, see § 1.2), VITO
participated in 3 environmental regulatory monitoring intercomparison exercises that were
organised in the period 2004-2006 as one of the 94 laboratories from 21 European member states
[98]. For trace element determination, the majority of participating laboratories used ICP-OES or ICP-
MS; to a lesser extent AAS-based techniques were applied (Flame, graphite furnace or hydride
generation). A general ranking of performances for all the analytes in all the exercises was made,
individuating the more problematic analytes in terms of lack of comparability of results (expressed as
average coefficient of variation (CV) %), lack of capacity to provide satisfactory or at least doubtful
results (expressed as percentage of |Z|-scores > 3), and lack of sensitivity of analytical methods
(expressed in terms of percentage of less than limit of detection results) (see Figure 25). The main
problems were encountered for trace element determination that are on the top of the list (i.e.,
highest difficulty ranking); Unexpectedly, analytes such as total P and nitrate were ranked as more
difficult to determine then most pesticides. The determination of pH caused the least problems in
terms of precision, accuracy and sensitivity.
The results of the European intercomparison exercises on environmental regulatory monitoring

(2004-2006) suggested that the routinely applied methods for the determination of trace elements in
water matrices should be improved to be able to fulfil WFD requirements. Furthermore, matrix
effects need to be taken into account, and specific methods avoiding matrix interferences need to be
implemented and routinely adopted (especially for As, Se and Cr). To date, dealing with matrix and
spectral interferences is still one of the main challenges in the routine adoption of ICP-MS as a
monitoring method for surface water (see also § 1.4.1).
Figure 25: Cumulative difficulty ranking of performance for all analytes monitored (sum of precision (CV%,
blue), accuracy (Z>|3|, purple) and sensitivity (< LoD, yellow) criteria)(figure courtesy of Final report on SWIFT-
WFD PTs results [98]).
2.5.1. AMBIENT BACKGROUND CONCENTRATION OF ELEMENTS IN SURFACE WATER IN FLANDERS
When evaluating toxicity data to derive quality standards for metals, total metal concentrations are
usually not directly related to ecotoxicological effects because many abiotic and biotic processes can
modify the availability of metals, even rendering them unavailable for uptake. This means that the
fraction available for uptake and toxicity may be a very small part of the total amount of metal
present.
Due to several physicochemical processes, metals exist in different chemical forms which might differ
in (bio)availability. Thus, the (bio)availability of metals in both laboratory tests and in the “real”
environment may be affected by several physicochemical parameters such as the pH and hardness of
water and the dissolved organic carbon (DOC) content.
The Water Framework Directive explicitly acknowledges the issues of (bio)availability and naturally
occurring concentrations for metals [43]. Member states may, when assessing the monitoring results
against the environmental quality standard (EQS), take into account:
(a) natural background concentrations for metals and their compounds, if they prevent
compliance with the EQS value; and
(b) hardness, pH or other water quality parameters that affect the bioavailability of metals.
Ideally, the derivation of EQS for metals requires an explicit consideration of (bio)availability using
speciation models or, failing that, to utilise dissolved concentrations instead of total concentrations.
On May 21st, 2010, the Flemish government approved a decree on new environmental quality
standards for a wide range of dissolved elements (operationally defined as filtered through 0,45 µm
membrane filter) in surface water. Today’s distinction between dissolved and particulate phases in
(geo)chemistry can be traced back to the application of 0.5 µm cellulose acetate membrane filters to
ocean waters by Goldberg et al. (1952) [99]. Since then, different studies called for relinquishing the
misleading delineation of particulate and dissolved phases at about 0.45 µm, but this operational
definition has become firmly established by now [100,101]. The following paragraphs describe how
ambient background concentrations of dissolved elements were derived in surface water for
Flanders.
The Flemish Environment Agency started monitoring of dissolved elements in 2007 and
systematically expanded the monitoring network to about 500 measuring points at present [106]. For
the determination of median ambient background concentrations of dissolved trace elements in
Flemish surface water, the same 22 background locations were chosen as for river sediment
(ecotoxicological and biological class 1, see § 2.4 and Figure 22)[92,107].
The methods used for the determination of dissolved elements in surface water were based on:
• EN 17294-1:2006 Water quality - Application of inductively coupled plasma-mass
spectrometry (ICP-MS) - General guidelines
spectrometry (ICP-MS) - Part 2: Determination of 62 elements
• EN 12846:2012 Water quality - Determination of mercury - Method using atomic absorption
spectrometry (AAS) with and without enrichment
On average, the 22 selected background locations were sampled and measured more than 300 times
over the period 2009 – 2012 by 3 different recognized laboratories [107]. For the elements Be, Cu,
Hg, Mo, Se, Sn, Ti and Tl, more than 90 % of the reported values were lower than the quantification
limit of the laboratory, therefore a (median) quantification limit is reported in Table 6. It should be
noted that the quantification limits used by recognized laboratories in environmental regulatory
monitoring are related to the legal requirements and, as such, they are not an indication of the
sensitivity attainable by ICP-MS. For the other elements and on average, less than 40 % of the
reported values were lower than the quantification limit of the laboratory. In order to calculate a
mean concentration of each element for each of the locations, the approach as set by Directive
2009/90/CE (technical specifications for chemical analysis and monitoring of water status) was
followed, i.e., if the amount of an element in a given sample is below the limit of quantification, the
measurement result shall be set to half of the value of the limit of quantification for the calculation
of the mean value.
→ Background concentration of dissolved elements in surface water
Ambient background concentrations of dissolved elements in Flemish surface water, sampled on 22

locations with no acute impact on benthic biota and good biological quality are summarized in the
figures below [92,107].
1,4
1,2
1,0
µg/L (< 0,45 µm)
0,8
Median
25%-75%
Non-Outlier Range
Outliers
0,6
Extremes
0,4
0,2
0,0
Cd Pb U Cr Sb
10
8
µg/L (< 0,45 µm)
6
Median
25%-75%
Non-Outlier Range
Outliers
4 Extremes
0
V Co As Ni
200
180
160
140
µg/L (< 0,45 µm)
120
Median
100 25%-75%
Non-Outlier Range
Outliers
80 Extremes
60
40
20
0
Zn Ba B Mn
Figure 26: Ambient background concentrations of dissolved trace elements (<0,45 µm) in Flemish surface water
as measured by ICP-QMS (average of 300 samples collected at 22 different locations in 2009-2012, see Figure
22) (data courtesy of VMM [92,107]).
The median ambient background concentrations of dissolved trace elements in Flemish surface
waters of the monitoring campaign 2009-2012 are summarized in Table 6 [92,107]. For comparison,
the median values for concentrations of dissolved elements in European stream water from the
FOREGS Geochemical Baseline Mapping Programme and the environmental quality standard values
as defined in VLAREM II Annex 2.3.1 are also represented [58].
Monitoring European EQS values

2009-2012 stream water VLAREM II
(< 0.45 µm) (< 0.45 µm) (< 0.45 µm)
µg/L µg/L µg/L
As 1.2 0.63 3
B 40 16 700
Ba 24 25 60
Be < 0.25 0.009 0.08
Cd 0.027 0.01 0.08-0.25
Co 0.75 0.16 0.5
Cr 0.24 0.38 5
Cu <2 0.88 7
a
Hg <0.01 0.05
Mn 80 16
Mo < 2.5 0.22 340
Ni 2 1.9 20
Pb 0.15 0.093 7.2
Sb 0.31 0.07 100
Se <2 0.34 2
Sn < 0.5 3
Ti <1 0.9 20
Tl < 0.05 0.005 0.2
U 0.21 0.32 1
V 0.71 0.46 4
Zn 9.6 2.7 20
Table 6: Median ambient background concentration and EQS values
of dissolved trace elements in surface water.
a
cold vapour-AFS was used for the determination of Hg according to EN 17852
All derived median concentrations of elements at the background locations are below the EQS values
with an exception for Co. Although the question “How low is low and how low do we need to go?”
can be adequately answered by ICP-MS in environmental regulatory monitoring of dissolved
elements in surface water [44], more than 90 % of the values reported for the elements Be, Cu, Hg,
Mo, Se, Sn, Ti and Tl during the monitoring campaign were lower than the quantification limit of the
laboratory. Regulatory monitoring of trace-level metals commonly relies on infrequent grab sampling
followed by instrumental analysis to determine filtered concentrations. However, limitations of these
methods are often ignored. These primarily concern (i) potential contamination and changes in metal
speciation during sample storage and transport, (ii) levels of suspended/dissolved organic matter or
colloids, and (iii) sample volume, filtration equipment and size of the filter used [102]. As for most
monitoring techniques, results obtained are operationally-defined (the protocol defines the answer),
and while this may facilitate comparability of results acquired across Europe, data generated may
have little relevance to ambient dissolved or bioavailable metal concentrations where metals may be
present in aquatic environments as free ions, complexed with organic and/or inorganic ligands or
sorbed to suspended particles and colloids.
Until now, assessment of the sampling protocol, not commonly included in quality control schemes,
has received little attention, and this may well become critical for implementation of legislation such
as the WFD. Where contaminant levels fluctuate, optimisation of the sampling frequency is of utmost
importance, and 12 equally spaced-sampling events per year may not provide representative
information (e.g., flood events). One possible solution is the use of passive sampling devices to
provide time-weighted average (TWA) pollutant concentrations in water [102,103]. Passive samplers
such as the diffusive gradient in thin film (DGT) device and Chemcatcher accumulate by diffusion a
labile fraction of metals with minimal disturbance of the system and provide more toxicologically
relevant and representative metal concentrations. It is now being recognised that passive samplers
can play a valuable role in monitoring water quality within a legislative framework such as the
European Union’s Water Framework Directive [104,105]. The data from these devices can be used
alongside the results obtained from conventional spot sampling to improve risk assessments and to
provide information required to take decisions on undertaking potentially expensive remedial
actions. It is expected that the aquatic monitoring sector will follow a transition similar to that which
occurred in air monitoring, where data obtained from passive samplers or on-line measurement
devices can be complementary used within a legal framework. But, as stated earlier, the introduction
of harmonised and standardised analytical methods tend to delay the acceptance and use of new
and innovative monitoring / screening methods.
Under the WFD, member states were required to report chemical status for fresh and coastal water
bodies. The map below summarizes the chemical status for European surface waters at river basin
district (RBD) scale in 2010 and describes the relative proportion of water bodies of good status,
those at risk of failing to achieve good status and those of unknown status [85].
Figure 27: WFD chemical status of surface water as of December 2010 (figure courtesy of European
Environment Agency [85]).
The information is based on data drawn from a variety of sources and reflects information as
reported by December 2010. However, despite the improved knowledge base arising from this body
of information reported by the different European member states, very often uncertainty remains as
to whether the observed concentrations of a particular substance pose a risk to aquatic
environments and human health. In this report, the European Environment Agency notes that the
assessment is on-going, and that conclusions remain tentative. This can be illustrated by the example
of the reported excessive levels of metals including cadmium and mercury, which challenge good
chemical status in a number of RBDs. In Sweden, widespread excessive levels of mercury in aquatic
biota mean that all RBDs currently fail to achieve good chemical status. In Sweden, unlike other
countries, biota has been chosen as the matrix for monitoring, making detection more likely than in
the water column, given the propensity of mercury to bioaccumulate. Such inconsistency in reporting
chemical status between member states is of concern and interpretation or comparison of the
chemical status should therefore be done carefully.
This example underlines that harmonization in communication and reporting of our chemical status
is of concern besides the technical feasibility of analysis at these concentration levels and that no
consistent map of the chemical status of European surface waters is yet available [45].
2.5.2. AMBIENT BACKGROUND CONCENTRATION OF ELEMENTS IN GROUND WATER IN FLANDERS
The European Groundwater Directive (2006/118/EC) establishes a regime which sets ground water
quality standards and introduces measures to prevent or limit inputs of pollutants into groundwater.
The directive establishes quality criteria that take account local characteristics and allow for further
improvements to be made based on monitoring data and new scientific knowledge.
Aquifers from the same typology can have strongly different groundwater chemistry and natural
background levels are rather a range of values than single values. Wendland et al. developed a
European aquifer typology map as a tool for comparison of hydrochemical data sets across Europe
and as a basis for referencing hydrogeochemical groundwater composition on a European scale
[108]. Nine major aquifer rock types were distinguished and secondary criteria were used to further
subdivide and refine the aquifer typology. M. Coetsiers et al. attempted to derive natural background
levels for four aquifers in Flanders of the sand and gravel typology as 90th and 97.7th percentile of a
carefully chosen dataset to approach the natural groundwater composition [108]. The range of
natural background levels (90th percentile) in these four aquifers (fluvial Pleistocene and Tertiary
marine) are summarised in Table 7.
In support of the implementation of the Water Framework Directive and the Groundwater Directive,
new groundwater quality standards, threshold values and background levels were adopted in
Flanders in 2010 (VLAREM II, Annex 2.4.1., background levels and threshold values were separately
established per groundwater body). The following paragraphs describe how ambient background
levels of dissolved elements were derived in ground water for Flanders.
The Flanders Subsoil Database (Databank Ondergrond Vlaanderen) includes several groundwater
monitoring networks with the intention of getting a representative picture of the groundwater
quantity and quality of the aquifers in Flanders [86]. In the context of environmental monitoring, the
phreatic groundwater monitoring network and the primary groundwater monitoring network are of
importance.
The phreatic groundwater monitoring network was originally established as a monitoring tool in
function of the European Nitrate Directive (91/676/EEC). Through this monitoring network, the
presence of nitrate-vulnerable areas (the so-called hydrogeological homogeneous zones) was
confirmed and analysed. Nevertheless, since 2007, the entire area of Flanders is nitrate-vulnerable
area due to the widespread contamination problem. Now, the network measures the effects of the
action programs (Manure Action Plans-MAP) taken in the scope of the Flemish Manure decree for a
further improvement of the groundwater quality (see also § 4.3). The phreatic groundwater network
is fully operational since 2004. In the meantime, the network has become multi-functional and forms
part of the monitoring program in the scope of the Water Framework Directive (2000/60/EG) and the
Groundwater Directive (2006/118/EG) as well. To meet the requirements, the network was further
optimized and enlarged. On a half yearly basis, 2107 multi-level wells (with about 5100 filter screens)
in agricultural area and approximately 80 multi-level wells (190 filter screens) in natural background
area are sampled and analyzed.
Figure 28: Geographical distribution of the wells of the phreatic groundwater monitoring network across the
different hydrogeological homogeneous regions (figure courtesy of Databank Ondergrond Vlaanderen [86]).
The primary monitoring network consists of about 435 wells with more than 800 screens, selected in
order to provide representative data of important, mainly deeper, aquifers. The primary
groundwater monitoring network aims primarily at establishing regional groundwater resources and
at following the evolution of the quantity of the groundwater bodies in time. In support of the Water
Framework Directive, the primary groundwater monitoring network is also increasingly used for
groundwater quality research. In this context, it is mainly used to determine the physico-chemical
background levels, but also for detecting anthropogenically induced changes in the quality of the
deeper aquifers [87].
Figure 29: Geographical distribution of the wells of the primary groundwater monitoring network, with the
phreatic groundwater bodies as background (figure courtesy of Databank Ondergrond Vlaanderen [86]).
For the determination of ambient background levels of dissolved elements in the 42 groundwater
bodies defined in Flanders, data from the primary and the phreatic groundwater monitoring network
of 2006 were used [88].
The methods used for the determination of dissolved elements in ground water were based on:
spectrometry (ICP-MS) - General guidelines
spectrometry (ICP-MS) - Part 2: Determination of 62 elements
• EN 11885:2009 Water quality - Determination of selected elements by inductively coupled
plasma-optical emission spectrometry (ICP-OES)
• EN 15586:2003 Water quality - Determination of trace elements using atomic absorption
spectrometry with graphite furnace
• EN 17852:2008 Water quality - Determination of mercury - Method using atomic
fluorescence spectrometry
• EN 12846:2012 Water quality - Determination of mercury - Method using atomic absorption
spectrometry (AAS) with and without enrichment
Different recognized laboratories were involved in the groundwater monitoring campaign of 2006
and for the determination of dissolved trace elements, different analytical techniques were used
(ICP-AES, ICP-MS and GF-AAS)[88].
→ Background concentration of dissolved elements in ground water
The background levels were derived per groundwater body as the 90th percentile of the groundwater
monitoring campaign of 2006. Most of the trace element concentrations determined in more than
2600 groundwater samples were reported as lower than the quantification limit of the laboratory
(inter alia caused by the use of different analytical techniques)[88]. Based on this dataset, the
expected median background concentrations for dissolved trace elements in groundwater are < 0.05
µg/L for Hg, < 0.1 µg/L for Cd, < 1 µg/L for Pb, Cr and Cu, < 10 µg/L for As and Ni and < 20 µg/L for Zn.
For the major elements, a wide range is observed in the background levels in the different
groundwater bodies (e.g., for Na a concentration range from 5 to 5000 mg/L).
The median ambient background concentrations of dissolved elements derived from regulatory
monitoring in Flemish ground water are summarized in Table 7 (column ground water monitoring
2006)[88]. In addition, the range of natural background levels (90th percentile), derived by Coetsiers
et al. in fluvial Pleistocene and Tertiary marine aquifers is given in Table 7 [108]. The range of the
background levels defined for each of the 42 groundwater bodies (as defined in VLAREM II, Annex
2.4.1), is also represented in the table below (these regulatory values represent the 90th percentile of
the measurements performed at background locations per groundwater body).
Ground water
Fluvial Pleistocene / Ground water
Background levels
Tertiary marine aquifers monitoring 2006
(VLAREM II Annex 2.4.1)
Range of Median Range of
th th
90 percentile 90 percentile
(< 0.45 µm) (< 0.45 µm) (< 0.45 µm)
µg/L µg/L µg/L
As 5.6 - 30 < 10 1 - 60
B 0.3 - 1.2
Cd 0.12 - 0.5 < 0.1 0.05 - 1
Cr 0.8 - 4 <1 1 - 37
Cu 5 - 10.2 <1 0.5 - 7
Hg 0.05 - 0.25 < 0.05 0.03 - 0.5
Ni 3.7 - 39 < 10 5 - 60
Pb 0.5 - 5 <1 1 - 18
Sb 0.03 - 1
Zn 26 - 260 < 20 16 - 250
mg/L mg/L mg/L
Al 0.01 - 1.5 < 0.1 0.01 - 0.8
Ca 73 - 200 79 4 - 700
Fe 5 - 44 2.3 0.12 - 50
K 10 - 36 3.5 3 - 200
Mg 12 - 39 10 2 - 800
Na 38 - 320 22 12 - 6000
Mn 0.3 - 1.3 0.26 0.02 - 2.2
th
Table 7: Range of 90 percentile concentrations in Fluvial Pleistocene and Tertiary marine aquifers, median
ambient background concentrations derived from regulatory monitoring (2006) and range of regulatory
background levels of dissolved elements in Flemish ground water (VLAREM II Annex 2.4.1).
The given ranges of ambient background concentrations in ground water point to ICP-MS as the
method of choice, as the superior sensitivity combined with the wide linear dynamic range
(determination of trace and major elements in the same run) favours the use of ICP-MS as compared
to ICP-AES and GF-AAS.
2.6. SUMMARY OF AMBIENT BACKGROUND CONCENTRATIONS OF ELEMENTS IN FLANDERS
As a case study and for the first time, ranges of background concentration of heavy metals were
summarised for all environmental compartments in Flanders based on European Standards on the
one hand and on a selection of (non-polluted) background locations on the other. In a very short
period of time, the majority of the human population has become urban, and it is estimated that by
2050 two out of every three people in the world will live in cities [110]. Urban areas are extremely
important socially, economically, and culturally, but they also have a profound impact on the
environment (i.e., urban geochemistry). The background concentration estimates reported in Table
8, can be considered as an indicator of the current “natural” chemical status of our environment and
can be used to assess future anthropogenic influences within an EU regulatory monitoring context.
soil air water

EN 16174 EN 16174 EN 16174 EN 14902 EN 15841 EN 15841 EN 16174 EN17294 EN17294
CEN/TS 16171 CEN/TS 16171 CEN/TS 16171 CEN/TS 16171
0-20 cm 50-100 cm tertiary PM10 bulk wet river surface ground
deposition deposition sediment water water
mg/kg mg/kg mg/kg ng/m³ µg/m².day µg/m².day mg/kg µg/L µg/L
(< 0.45 µm) (<0.45 µm)
As 7.2 6.1 8.8 0.61 0.37 0.21 17 1.2 < 10
Ba 2.7 24
Cd 0.32 0.10 0.050 0.20 0.098 0.059 0.15 0.027 < 0.1
Co 5.2 4.6 6.6 0.75
Cr 25 24 62 2.0 0.94 0.40 20 0.24 <1
Cu 11 4.7 2.7 5.0 4.6 9.0 8.2 <2 <1
a a a b c d e
Hg 0.062 0.027 0.01 0.7 0.025 0.052 < 0.01 < 0.05
Mn 4.3 23 8.1 169 80
Mo 0.44 0.16 0.39 0.44 < 2.5
Ni 7.4 6.3 12 1.8 1.0 0.67 8.2 2.0 < 10
Pb 26 8.3 6.1 8.1 4.6 1.4 17 0.15
Sb 0.5 0.14 0.32 1.1 0.31
Se 0.28 0.14 0.29 1.0 <2
Sn 1.3 < 0.5
Ti 3.6 <1
V 40 26 62 2.2 0.71
Zn 57 26 29 22 21 12 125 9.6 < 20
Table 8: Estimates of median ambient background concentrations of heavy metals present in the different
environmental compartments in Flanders.
a
CEN/TS 16175-2:2013 Sludge, treated biowaste and soil - Determination of mercury - Part 2: Cold vapour atomic fluorescence
spectrometry (CV-AFS).
b
EN 15852:2010, Ambient air quality - Standard method for the determination of total gaseous mercury.
c
EN 15853:2010, Ambient air quality - Standard method for the determination of mercury deposition.
d
EPA 7473:1998 Mercury in solids and solutions by thermal decomposition, amalgamation, and atomic absorption spectrophotometry.
e
EN 12846:2012 Water quality - Determination of mercury - Method using atomic absorption spectrometry (AAS) with and without
enrichment.
EN 14902:2005 Ambient air quality - Standard method for the measurement of arsenic, cadmium, lead and nickel in the PM 10 fraction of
suspended particulate matter.
EN 15841:2009, Ambient air quality - Standard method for determination of arsenic, cadmium, lead and nickel in atmospheric deposition.
CEN/TS 16171:2012 Sludge, treated biowaste and soil - Determination of elements using inductively coupled plasma-mass spectrometry
(ICP-MS).
EN 16174:2012 Sludge, treated biowaste and soil - Digestion of aqua regia soluble fractions of elements.
EN 17294-1:2006 Water quality - Application of inductively coupled plasma-mass spectrometry (ICP-MS) - General guidelines.
EN 17294-2:2004 Water quality - Application of inductively coupled plasma-mass spectrometry (ICP-MS) - Part 2: Determination of 62
elements.
The range of elemental background concentrations, derived from the different examples of Flemish
regulatory monitoring, illustrate a match made in heaven with the sensitivity of ICP-MS instruments
(see Table 8). This being said, it is also clear that the panoramic analysis capabilities of ICP-MS are not
yet fully exploited in current environmental monitoring studies with respect to its linear dynamic
range and multi-element capabilities [109].
For the determination of elements in soil and river sediment, the HNO3/HCl/HF digestion which is
currently prescribed in Flanders hampers the optimal use of ICP-MS. However, there is a trend in the
European Committee for Standardization (CEN) to promote nitric acid and aqua regia as the sole
acids to be used for environmental sample digestion. Additionally, with the incorporation of Hg in
CEN/TS 16171 as one of the elements determinable by ICP-MS, the use of ICP-MS in future regulatory
monitoring of solid matrices will certainly be favoured. The examples of environmental monitoring
described in this chapter, show that the aqua regia digestion in combination with ICP-MS is a
powerful tool for the determination of ambient background concentrations of major and trace
elements. One must keep in mind that the choice of the acid digestion procedure will have
implications on the measured concentration of some elements, as acid digestion procedures are
prone to operational settings.
In support of the European Water Framework Directive, ICP-MS (EN 17294) became the only method
of choice for chemical analysis and monitoring of dissolved elements in surface water. More
monitoring studies using ICP-MS and with focus on low level determination of elements are needed
to refine the derived estimates of ambient background concentrations (now often reported as < limit
of quantification). However, dealing with matrix and spectral interferences is still one of the main
challenges in the routine adoption of ICP-MS as a monitoring method. For the determination of
concentrations in surface and ground water, it can be argued that the risk of sample contamination
nowadays represents a far greater challenge than the instrumental limits of detection attainable with
ICP-MS.
The same holds true for the determination of elements in fine particulates and atmospheric
deposition: sampling and contamination free acid digestion are considered the most critical steps for
accurate trace environmental monitoring. The European standard methods for the determination of
elements in ambient air (EN 14902 and EN 15841) prescribes nitric acid digestion (in combination
with H2O2 or traces HF), which also favours the use of ICP-MS as sensitive and multi-elemental
determination method.
2.7. SUMMARY AND CONCLUSION
A clean, healthy and diverse environment is a prerequisite for prosperity and well-being for everyone
and this can only be achieved through a common legal framework. This common framework is
needed in order to assure that environmental pollution, no matter where it comes from, is assessed
in the same way across national or even regional borders. At a European level, major legal acts
include the Water Framework Directive, the Waste Framework Directive, the Air Quality Directive
and the upcoming Soil Framework Directive and all of these include monitoring programs to report
on the state of our environment.
European Standards (EN) related to the use of ICP-MS for the determination of elements in the
different environmental compartments have been published starting from 2004 onwards and is still
on-going. As a consequence, these European standard methods have been introduced as reference
methods in the Flemish Environmental legislation and regulatory monitoring. Nowadays, European
standard methods (EN) related to the use of ICP-MS are available for the determination of elements
in all environmental compartments (water, air, soil) and this is undoubtedly the most important
advance in environmental regulatory monitoring of elements the last decade.
Heavy metals have unique characteristics that should be considered when assessing their
environmental risks. Unlike most organic substances, heavy metals are neither created nor destroyed
by biological or chemical processes. Rather, they are transformed from one chemical form into
another. Moreover, in the context of risk assessment, one must keep in mind that heavy metals are
naturally occurring constituents of our environment that vary in concentrations across geographic
regions. In other words, as heavy metals are naturally present in our environment, knowledge of
(regional) ambient background concentrations is needed to provide baseline data for pollution and
risk assessment studies.
Because of the availability of ICP-MS instruments nowadays, it can be argued that the main
hindrance to the implementation of the European monitoring requirements is not the technical
feasibility of analysis at these concentration levels, but rather potential contamination during
sampling, sample storage, sample handling and analysis. Two additional comments with respect to a
harmonised implementation of the European monitoring requirements may be formulated at this
point. First, as for most monitoring techniques, results obtained are operationally-defined. The
protocol, e.g., acid digestion, 0.45 µm filtration, ... defines the answer, and while this may facilitate
comparability of results acquired across Europe, data generated may have little relevance to
bioavailable metal concentration. Second, more harmonization in communication and reporting of
our chemical environmental status between member states is needed to compare regulatory
monitoring data on a European scale, e.g., choice of sampling locations (e.g., rural versus industrial)
and type of sample (e.g., biota versus water column).
As stated earlier, the panoramic analysis capabilities of ICP-MS are not yet fully exploited in
environmental monitoring studies yet with respect to its linear dynamic range and multi-element
capabilities. The determination of heavy metals in environmental monitoring has traditionally
focused on a relative small number of elements (e.g., As, Cd, Pb, Hg), but it is to be expected that
future regulatory monitoring will include more elements (holistic approach) and that the panoramic
analysis capabilities of ICP-MS will be further exploited. In this future context, high mass resolution
ICP-MS instruments operated under clean room conditions will become the method of choice for
regulatory environmental monitoring, as high mass resolution is a universal and powerful approach
to overcome spectral interferences in routine applications. In addition, the labor-intensive steps of
preparation of calibration standards, dilution of samples and the addition of internal standards can
be automated by using an in-line auto-dilution/auto-calibration sample delivery system, which could
make this approach well-suited for the demands of a high-throughput environmental laboratory.
Furthermore, it significantly lowers the risk of contamination of the sample, standards or blanks,
because all these functions are being carried out in-line, with no manual intervention by the analyst.
Elemental speciation – Determination of oxy-anions in water and solid samples 69
CHAPTER 3 ELEMENTAL SPECIATION – DETERMINATION OF OXY-ANIONS IN

WATER AND SOLID SAMPLES
3.1. PERSPECTIVE ON ELEMENTAL SPECIATION
During the second half of the 20th century, it became evident that trace elements play a major role
whenever biological activities, environmental chemistry or material characteristics are discussed and
the determination of trace elements has therefore gained outstanding importance in environmental
and life sciences. Elements, even when present at very low concentrations in biological and
environmental matrices, can exert fundamental influence on ecosystems and the vital functions of
organisms (see also Table 1, p.27)[111].
In this context, analytical chemists increasingly realized that, in general, total concentrations of
chemical elements cannot give information on mobility, bioavailability and the eventual impact of
elements on ecological systems and biological organisms [112].
The chemical nature and quantity of the relevant element species in a matrix, its physical and
chemical association, rather than the total element concentration, is highly responsible for the
mobility, bioavailability and finally the ecotoxicological or toxicological impact of an element.
Elements usually interact as parts of macromolecules (complexes with humic substances, proteins,
enzymes, hormones, etc.), low molecular weight metabolites (metal–peptide-complexes, metal-
carrying metabolites) or according to their oxidation state. Therefore, only the knowledge on
speciation provides sufficient information for adequately assessing whether an element is toxic,
without (known) impact at a specific concentration or essential [112].
While the concept of ‘‘elemental fractionation’’ in the sense of distinguishing fractions of the total
element concentration according to characterization, e.g., as being ‘‘bioavailable’’, developed slowly
around 1960, it was only during the late 1980s, that instrumental elemental analysis reached the
detection power necessary to measure small fractions of trace elements - and only sometimes a
single species - in environmental and biological samples (see also §1.4.2) [112].
In the past years, speciation analysis matured and many important technical and conceptual
developments were published [113]. Separation and detection methods already established were
combined in partly novel ways and modified according to the relevant speciation problems.
Combination and hyphenation of separation techniques and element-selective or molecule-selective
detection systems are now generally established as the basis for speciation analysis. Early
approaches used off-line combinations of HPLC and AAS or even voltammetry. Further instrumental
progress in speciation analysis was based on direct hyphenation of HPLC, GC or capillary
electrophoresis (CE) to plasma-based detectors. Initially, this was realized by coupling to ICP-atomic
emission spectrometry (ICP-AES), which in some cases was even completed with on-line hydride
generation and UV digestion devices, but soon HPLC was coupled to ICP-mass spectrometry (ICP-MS).
Nowadays, the HPLC/ICP-MS coupling constitutes a widespread, sensitive and versatile speciation
instrumentation, also providing multi-element/isotope ratio capabilities. Meanwhile, the use of such
systems is well established, even for low species concentrations, e.g., by the use of sector-field ICP-
MS, or for oxy-anions like As and Se. Detection of the latter elements is critical when using ICP-MS, as
their nuclide(s) suffer(s) from spectral interference. However, with the use of commercially available
ICP-MS instruments using dynamic reaction cells (DRC) or collision cells, even this problem can be
overcome (see also § 1.4.1). For addressing quantitative speciation approaches, species-unspecific
70 Elemental speciation – Determination of oxy-anions in water and solid samples
isotope dilution mass spectrometry (IDMS) was introduced successfully. With the availability of
isotopically labelled species, e.g., by in-house synthesis, even species-specific IDMS was becoming
possible for reliable species quantification [114].
Together with the instrumental developments, however, the awareness grew that the terms and
nomenclature used in the rapidly growing number of speciation papers was inconsistent and that the
quality control with respect to accuracy and traceability of numerous published data was insufficient.
The harmonization of terms related to speciation by giving clear definitions was elaborated by an
expert group under the auspices of IUPAC (International Union of Pure and Applied Chemistry) as
follows: speciation analysis is the analytical activity of identifying and/or measuring the quantities of
one or more individual chemical species in a sample; the chemical species are specific forms of an
element defined as to isotopic composition, electronic or oxidation state, and/or complex or
molecular structure; the speciation of an element is the distribution of an element amongst defined
chemical species in a system [115,142]. Apart from clearly defining speciation terms, they also
distinguished chemical speciation analysis from operationally defined procedures, which were not
considered as real chemical speciation analysis, as predominantly the identification of the single
species is missing.
Another conclusion from the IUPAC expert group was that quality control and quality assurance
gained increasing importance in speciation analysis. Today, unequivocal identification of element
species is a strong demand, although low species concentration, interfering matrix components,
insufficient instrumental selectivity or sensitivity, etc. can still easily impede speciation attempts.
Very important for the development of interdisciplinary cooperation between scientists from
different fields was a series of workshops organized within the framework of the EU-funded network
"Speciation 21" during the years 1998-2000 [119]. The “Speciation 21” network linked scientists in
analytical chemistry, working on the development of methods for the chemical speciation of trace
elements, potential users from industry and representatives of legislative agencies, in the field of
environment, food and occupational health and hygiene.
Since then and as a next logical step, elemental speciation is now also found under the keyword
‘‘metallomics’’ (speciation analysis employed to study metabolic pathways and their probable
changes, e.g., in Se metabolism [121]). Nowadays, orthogonal speciation schemes (e.g., ESI-MS and
HPLC/ICP-MS) and multi-disciplinary approaches are being used increasingly for (complex) species
identification [111].
A multi-disciplinary approach including X-ray absorption near-edge spectroscopy (XANES) was used in
this PhD dissertation for the validation of the determination of hexavalent chromium in ambient air
(see § 3.3.3). XANES has proven to be well-suited for determining elemental speciation because it
can directly and non-destructively analyse particulate matter samples at low concentration levels
[136]. X-ray based techniques, such as XANES, are generally regarded as reference methods for solid
state speciation since they enable the identification and quantification of chromium oxidation states
while being non-destructive. The most employed techniques are (XANES), X-ray Diffraction (XRD) and
X-ray Photoelectron Spectroscopy (XPS)[129]. Whereas XPS and XANES will enable quantification of
the redox status of , e.g., chromium, only the nature of the crystalline phases under which chromium
is present can be revealed by XRD. In a XANES experiment, the absorption of X-ray photons by the
sample is measured as a function of X-ray energy around and above an absorption edge of an
element in the sample. Structural and electronic information can be extracted from the observed fine
structure in the absorption spectrum. This fine structure is divided into X-ray Absorption Near Edge
Structure (XANES, the fine structure around the absorption edge) and Extended X-ray Absorption
Fine Structure (EXAFS, the fine structure above the edge: >50 eV). Analysis of variations in the X-ray
absorption coefficient in the XANES and EXAFS spectral regions of the absorption spectrum provides
structural information concerning the local environment of the element, in terms of oxidation state,
interatomic distances, and the type and number of coordinating ligands. The primary disadvantages
of the XANES direct speciation analysis technique are that the structural information obtained
represents a weighted average over the different chemical species of an element probed in the
sample and the requirement of a synchrotron X-ray source [68].
In summary, speciation is a modern and current analytical discipline where many questions have to
be answered, especially in environmental chemistry and the life sciences [143]. Many analytical
techniques have therefore been developed and corresponding applications have been designed
during the last 20 years, especially focusing on environmental and biological samples. These efforts
often considered ‘traditional’ elemental species such as methyl mercury, hexavalent chromium,
different selenium and arsenic species and also heavy metals bound by biomolecules. In the
following paragraphs, the scientific papers related to the development of methods based on ICP-MS
for the speciation of oxy-anions in an environmental context are introduced with emphasis on the
connecting thread of the PhD dissertation, i.e., the relation between the ICP-MS measurement, the
environmental issue and the regulatory context. The scientific papers have been arranged
chronologically in the following paragraphs (see timeline):
§ 3.2 Removal of selenate and selenite from waste water
K. Tirez, W. Brusten, S. Van Roy, N. De Brucker and L. Diels, Characterization of inorganic selenium
species by ion chromatography with ICP-MS detection in microbial-treated industrial waste water, J.
Anal. At. Spectrom., 2000, 15, 1087-1092.
§ 3.3 The ban on hexavalent chromium
§ 3.3.1 The ban on hexavalent chromium in packaging material
K. Tirez, W. Brusten, A. Cluyts, J. Patyn and N. De Brucker, Determination of hexavalent chromium

by species-specific isotope dilution mass spectrometry and ion chromatography – 1,5 –
diphenylcarbazide spectrophotometry, J. Anal. At. Spectrom., 2003, 18, 922-932.
§ 3.3.2 The ban on hexavalent chromium in electronic waste (RoHS and WEEE)
K. Tirez, H. Scharf, D. Calzolari, R. Cleven, M. Kisser, D. Lueck, Validation of a European Standard for
the determination of hexavalent chromium in solid waste material, J. Environ. Monit., 2007, 9, 749–
759.
§ 3.3.3 Hexavalent chromium in ambient air
K. Tirez, G. Silversmit, N. Bleux, E. Adriaenssens, E. Roekens, K. Servaes, C. Vanhoof, L. Vincze, and P.

Berghmans, Determination of hexavalent chromium in ambient air: a story of method-induced
Cr(III) oxidation?, Atmospheric Environment, 2011, 45, 5332-5341.
§ 3.4 Oxyhalide by-products in drinking water disinfection
K. Tirez, W. Brusten, M. Wevers, F. Beutels and F. Vanhaecke, Determination of bromate in drinking

waters using low pressure liquid chromatography / ICP-MS, J. Anal. At. Spectrom., 2013, 28, 1894 –
1902.
hexavalent chromium by species-
specific isotope dilution mass (2003) Determination of
spectrometry organotin species [139]

(2001) Determination
of bromate in drinking of bromate in drinking
of selenium species in
water by HPLC-ICP-MS water by low pressure
waste water
LC/ICP-MS
(1999) Determination (2007) Study of (2011) Determination of

of selenium species in hexavalent chromium in hexavalent chromium in
yeast extract soil and waste material particulate matter
1997 2000 2005 2010
inductively coupled plasma-mass spectrometry (ICP-MS) for the speciation of metals and oxy-anions
in an environmental context.
Interesting links:
• European Virtual Institute for Speciation Analysis (EVISA), is a service provider in the field of
speciation analysis. EVISA's web portal is the primary source for all those seeking information on
chemical species with respect to analysis, biological activity (toxicity, nutritional value,
metabolism), legislation (laws, rules, standards) and research in related fields.
http://www.speciation.net/
3.2. REMOVAL OF SELENATE AND SELENITE FROM WASTE WATER
It must be noted that two elements - selenium and arsenic - attracted (and still attract) special
attention across all fields of speciation and the number of papers dealing with As- or Se speciation is
still increasing rapidly. This is due to the discovery of new species and their structural elucidation, as
well as to the detrimental health effects of several As species or the protective health effects of some
Se-species.
Selenium plays a major nutritional and biological role in living systems, and speciation analysis of
inorganic and organoselenium compounds to define exact biological roles is a major analytical
challenge [120]. Selenium is an essential nutrient and has been known to be a necessary component
of the human diet for many years (see Table 1, p.27). Selenium is also toxic at levels above the rather
narrow optimum range in the human diet; daily consumption of less than 0.1 mg kg−1 of body weight
will result in selenium deficiency, while levels above 1 mg kg−1 are considered toxic . Selenium is of
interest as it may offer a protective effect against several degenerative diseases. The organic form of
selenium provided by selenium yeast has been shown to differ in bioavailability and metabolism
compared with inorganic (e.g., selenate, selenite) forms of dietary selenium. Dietary
supplementation with selenium yeast has been used to study the effects of selenium status on the
risk of developing cancers or precancerous lesions. The potential beneficial effect of Se in terms of
cancer prevention has led to widespread commercial interest in selenium-containing dietary
supplements, sold in various chemical forms [120].
In 1999 and within the Speciation 21 Thematic network, an intercomparison exercise for selenium
speciation in yeast and arsenic speciation in sea plant (IAEA-140 FUCUS sample) was organised [119].
The participants were encouraged to use different extraction and analytical methods. This was the
starting point of the development of methods based on ICP-MS for the speciation of oxy-anions at
VITO (Figure 30, selenium speciation in hot water yeast extract).
Time (min)
Figure 30: Overlay of a chromatogram of a standard solution (pink) of Se-ethionine (Se-E), Se-methionine (Se-
M), selenocystine (Se-C) and selenocysteine (Se-Ci) with a chromatogram of a hot water extract of yeast (red).
Samples analysed by quadrupole-based ICP-MS (Elan 6000) equipped with a micro-concentric nebuliser (MCN-
100, Model M-2, CETAC; 10 µl sample loop; IC column: IONPAC AS10 microbore, 2 × 250 mm (Dionex); eluent:
-1
100 mM NaOH + 1% EtOH at 200 µl min .
Besides the beneficial properties of selenium species, selenium introduced to the environment can
rapidly bioaccumulate via the food chain, and can be present at significant levels for higher animals.
High selenium levels in wildlife exert known toxic effects, including general tissue damage and
reproductive failure, as well as teratogenic effects [121]. In nature, selenium is most commonly
observed as selenate (+VI), selenite (+IV), or selenide (-II). Though complexed selenium is of low
toxicity, selenate and selenite are considered toxic. These two forms of selenium are generally found
in water, and display bioaccumulation and bioavailability.
Selenium is often associated with sulphur-containing ores such as pyrite, sphalerite and chalcopyrite
[122]. These sulfate/sulphite-containing ores are prevalent in the mining of metals such as copper,
nickel, silver, lead and uranium. Consequently, selenium contamination has become an emerging
issue in such mining processes. Selenium contamination typically occurs in the aqueous stream, and
this stream must be treated prior to discharge.
Removal of oxy-anions such as selenite and selenate from wastewater has been achieved by
filtration, chemical treatment and/or biomediated removal [122,123]. The removal of oxy-anions is
much more difficult than that of the cations, because anions with a similar structure, such as nitrate,
sulphate and phosphate, often co-exist at higher concentrations. Searching effective adsorbents,
reagents and/or novel methods for the removal of the oxy-anions represents an important
environmental remediation issue.
Recently, in a research project initiated by Essencia (Belgian Federation for Chemistry and Life
Sciences industries) and VOKA (Flanders' Chamber of Commerce and Industry), the preventive or
remedial actions that companies can take to ensure compliance with the environmental quality
parameters for Flemish surface water in terms of the oxy-anions of antimony, boron, molybdenum
and selenium were studied [123]. The results of laboratory tests confirmed that selenium is difficult
to remove from industrial waste water by physico-chemical actions. Moreover, as the removal
efficiency of Se (IV) was shown to be higher than that of Se (VI), knowledge of selenium speciation in
waste water is essential to predict removal efficiency and consequently, the feasibility of water
purification techniques to ensure compliance with the environmental quality parameters.
The paper presented hereafter describes the development of an HPLC/ICP-MS method for the
characterisation of selenium species in support of the development of a biologically based
technology for the removal of oxy-anions from heavily contaminated wastewaters (see paper:
Characterization of inorganic selenium species by ion chromatography with ICP-MS detection in
microbial-treated industrial waste water, J. Anal. At. Spectrom., 2000). The efforts to set up an
HPLC/ICP-MS system for the speciation of selenium at that time, were mainly devoted to the
automation and a proper communication interface between the sample injection system, the HPLC
pump and the ICP-MS instrument.
Nowadays, LC/ICP-MS can truly be considered widespread sensitive and versatile speciation
instrumentation (see also § 3.4, where a low pressure LC system was used for the (routine)
determination of bromate in drinking water). Recent approaches to determine Se species, specifically
at ultratrace levels, have also focussed on a robust and work/time-efficient method by online
coupling of a preconcentration (trap) column to an IC-ICP-MS system [124]. Other novel approaches
used ICP tandem mass spectrometry (ICP-MS-MS) with O2 reaction/collision gas to completely
remove severe interferences with the Se speciation originating from the plasma source and the
biological sample matrix [125].
Characterization of inorganic selenium species by ion chromatography with

ICP-MS detection in microbial-treated industrial waste water
Kristof Tirez1, Wilfried Brusten1, Sandra Van Roy1, Nicole De Brucker1 and Ludo Diels1
1
Flemish Institute for Technological Research (VITO), Belgium
Journal of Analytical Atomic Spectrometry, 2000, 15, 1087-1092.

(Reference list of this paper, see p. 84)
Abstract
Microbial selenium reduction was investigated for application in the removal of selenium from
industrial waste water. This paper presents the results of the development of a routine ICP-MS
method for the determination of inorganic selenium species in microbial-treated industrial waste
water. For the speciation of water soluble inorganic selenium species (selenate and selenite) two
different anion exchange columns were compared. The column separation performance for the
determination of selenium was investigated for both inorganic and some organic selenium species.
The increase in sensitivity for selenium after the addition of carbon-containing solutions was
investigated by the addition of increasing amounts of ethanol to the eluent. The optimum addition
level of 2% v/v ethanol resulted in a sensitivity enhancement factor of 5 under the given instrumental
settings of the ICP-MS. The ICP-MS instrument showed good long-term stability (3% over 40 h), as
shown by the analysis of two independent control standards after every 10 samples with addition of
Scientific publication
arsenobetaine as the internal standard. The different selenium species showed the same signal
sensitivity, resulting in a quantification limit of 1 µg/L per species. Over 600 microbial-treated waste
water samples were characterized for selenate and selenite content. A selenium compound present in
some of the samples could not be identified. No evidence of DL-selenocystine and DL-
selenomethionine and DL-selenoethionine could be found. The microbial removal of selenium from
industrial waste water was at first hampered by the presence of nitrate (denitrification) and the
partial reduction of selenate was observed after the complete removal of selenite to the metallic
form.
Introduction
Techniques for the removal of dissolved metals from waters are needed for a variety of applications,
including mining and several environment-related issues. The requirements are driven by regulatory
guidelines for the removal of site-specific pollutants with an overall goal of meeting national surface
water standards. Laboratory tests have demonstrated that microbial reduction of metals is a
promising technique for the removal of metal contaminants from mining process waters.1 Various
metals can be used as electron acceptors by microorganisms, and are subsequently reduced in the
process. The reduced forms of some metals, such as arsenic, chromium and selenium, are less
soluble and can be separated from the waste water. A cost-effective, off-the-shelf technology is not
available for selenium removal from large-volumes of waste water. The microbial selenium reduction
was investigated for application to selenium removal from industrial waste water. This paper
concentrates on the development of a method for the determination of the inorganic selenium
species in the microbially-treated industrial waste water.
Inorganic selenium is most commonly found in four oxidation states (Se6+, Se4+, Se0 and Se2-). Selenate
(SeO42-) and selenite (SeO32- ) are highly water soluble, but elemental selenium (Se0) is much less
soluble in water. The most reduced form, hydrogen selenide (H2Se) occurs as a toxic gas, but is
readily oxidized to elemental selenium in the presence of air. A typical selenium contaminated waste
water contains a mixture of selenate, selenite and elemental selenium. The relative proportions
depend on the oxidation.
In the last few years, the development of analytical methods for selenium species determination has
considerably increased (D'Ulivo et al.2). The determination of selenium species by ICP-MS is
hampered by (i) the high ionization potential of selenium, the degree of ionization is only 30 %
leading to poor sensitivity; and (ii) the several (poly-atomic) interferences on all selenium isotopes.3
To overcome the poor sensitivity carbon containing solutions can be added. The enhancement effect
is dependent on the ionization energy of the analyte and the most pronounced effect is seen on
elements such as selenium and arsenic with ionization energies in the range of 9-11 eV.4
The use of an anion exchange column gives the opportunity to separate not only the selenium
species but also possible interferences. Moreover the use of IC-ICP-MS for the quantitative
determination of total selenium in industrial waste water can be a good alternative to circumvent the
effect of large amounts of interfering anions as Cl-, Br-, SO42- and PO43- on the different selenium
isotopes. The coupling of a micro-concentric nebuliser, operating at low flow rate, favors
chromatographic microbore columns, which have the advantage of good resolving power and small
sample amounts.5 The superior chromatographic detection power of ICP-MS compared to flame
atomic absorption spectrometry has also been demonstrated by Pedersen et al.6 for the speciation of
four selenium compounds.
The aim of this study was to examine the speciation performance of two commercial available anion
exchange columns, the influence of the addition of ethanol to enhance signal intensity and the effect
of interferences on the different selenium isotopes. This should allow us to choose the most suitable
isotope to obtain a reliable method for routine analysis.
Experimental section
Instrumentation
Samples were analysed on an Elan 6000 ICP-MS system (Perkin-Elmer, SCIEX, Thornhill, ON, Canada)
equipped with an MCN-100 micro-concentric nebuliser (Cetac technologies, Omaha, NE, USA) fitted
to a scott-type Ryton double-pass spray chamber.
Auto-sampler
ICP-MS FIAS 400
6-port valve
HPLC pumps
Picture 1: HPLC system coupled to an Elan 6000 ICP-MS system.

During optimization of the instrument settings (Table 1), the MCN-100 was fed by means of a
peristaltic pump. The speciation of the redox species was carried out by anion exchange
chromatography with an ANX 3206 microbore column (Cetac) and an IONPAC AS 10 column (Dionex,
Sunnyvale, CA, USA). Both columns were provided with a guard column to remove dissolved and
particulate contaminants from the sample and eluent before they reached the column. The eluent
was pushed through the column via a series 10 HPLC pump (Perkin-Elmer). A FIAS 400 (Perkin-Elmer)
peristaltic pump provided the filling of a PEEK sample loop. The sample loop was connected to an
automated 6-port valve (LabPRO 6, Rheodyne, California, USA). The measurement system, including
sample uptake, filling of the loop, injection on the column and ICP-MS measurement is
computerized. The integration of the chromatograms was performed with Turbochrom (Perkin-
Elmer). Before analysis the instrument was preconditioned by aspirating the eluent for at least 1
hour.
parameter Value
instrument settings
nebuliser gas flow 0.9 ± 0.05 l/min
RF power 1200 V
analog stage voltage -1950 V
pulse stage voltage 1500 V
lens voltage 5±1V
measured peak width 0.7 amu
dead time detector 40 ns
measurement settings
scan mode peak hopping
dwell time 30 ms
sweeps / reading 5
readings / replicate 1
replicates 1000
76 77 78 82 35 79 75
Isotopes monitored Se, Se, Se, Se, Cl, Br, As
Table 1: Optimised instrumental measurement settings ICP-MS, Elan 6000.
Standard solutions and reagents
For the preparation of all solutions ultra-pure water with a resistivity of 18 MΩ cm-1 obtained from a
Milli-Q water purification system (Millipore, Milford, MA, USA) was used. Stock standard solutions of
2 µg ml-1 were prepared from 100 µg ml-1 selenate and selenite standards (Inorganic Ventures,
Lakewood, NJ, USA). Calibration standards were diluted from these stock solutions. The
concentration level of the standards was 50, 150 and 250 ng ml-1 Se(IV) and 50, 250 and 500 ng ml-1
Se(VI). The standards were prepared in 100 mM NaOH (pro analysi, Merck, Darmstadt, Germany) for
the use of the AS-10 column. A 5 mM ammonium malonate (Inorganic Ventures) solution was used
as eluent for the separation of selenium species on the ANX 3206 column. DL-Selenocystine, DL-
Selenomethionine and DL-Selenoethionine were purchased from Sigma (Bornem, Belgium). Stock
standard solutions of 100 µg ml-1 of each compound in ultra-pure water were stored in the dark at
4°C. A quality control standard solution of 25 ng ml-1 Se(IV) was diluted from a multi-element
standard solution of 100 µg ml-1 Se (10580 ICP Multi Element Standard VI, Merck). This quality
control standard was analysed after every 10 samples. A commercial available 1031 µg kg-1
arsenobetaine reference material solution (European Commission, CRM 626) was used as internal
standard.7 The signal enhancement studies were performed with ethanol (pro analysi, Merck).
Sample treatment
The study for the microbial removal of selenium from industrial effluents was performed with four
selected bacteria. Two belonged to the genus of Pseudomonads and were isolated from a
contaminated industrial site in the UK. Two belonged to the Spinghomonads and were isolated from
a soil of the mining region in Likasi in the Republic of Congo. Pseudomonas fluorescens J3/73b and
J1/15a and Spinghomonas sp. AS44 and AS99 were used. Sand columns (10 cm height and 2.5 cm
diameter) were inoculated with the mentioned bacteria. After growth of the bacteria in the columns,
waste water was circulated through the fixed sand bed. A carbon source, glucose or acetate, was
added to the waste water in a final concentration of 0.1%. Samples were taken as a function of time
for the determination and characterization of the removed selenium species. The samples were
filtered through a 0.45 µm membrane filter (Schleicher&Schuell, Dassel, Germany) to remove
suspended solids and micro organisms and then stored in the dark at 4 °C. Prior to measurement the
samples were diluted 5-fold and a concentrated solution of the eluent was added to match the
appropriate separation conditions.
Results and discussions
Interferences
The selenium isotope masses 76, 77, 78 and 82 were chosen for the study, as the abundance of 74Se
is too small to achieve the required sensitivity (Table 2).8 Although all monitored selenium isotopes
suffer from polyatomic argon interference, only 80Se is unmeasurable because of the large 40Ar2+
interference (Table 2). The isobaric interference of krypton (present as trace element in the argon
supply) and the argon polyatomic interferences result in an increase of the ICP-MS background
signal. This increase depends on the relative abundances of the interfering constituents. The
resulting height of the baseline in the chromatograms depends on the given instrumental settings:
77
Se (70 cps: 40Ar36ArH+) < 82Se (1500 cps: 82Kr+ and 40ArH2+) << 78Se (13500 cps: 78Kr+ and 40Ar38Ar+) <
76
Se (65500 cps: 40Ar36Ar+). Although the baseline levels are stable during measurement, they
generate, to a certain extent, a higher noise, resulting in less accurate results for 78Se and 76Se for low
selenium concentrations.
From the point of view of sensitivity, selenium isotope 82 is preferred to isotope 77, because the
relative abundance is higher (Table 2), resulting in a 13% gain in signal. When choosing a suitable
analytical mass, the possibility of matrix interferences must be considered. In this case however,
when performing ion chromatography, these interferences can be circumvented. The chloride ions,
which interfere as 40Ar37Cl on 77Se, have a different retention time and do not interfere on the
measurement of 77Se. Considering the background level and the abundance, the selenium isotope 82
was selected for quantitative measurement and the 76Se, 77Se and 78Se isotopes for qualitative
purposes.
isotope natural abundance (ref. 8) isobaric poly-atomic interference

(%) interference
74 74
Se 0.89 Ge+ 40
Ar34S+
76 76
Se 9.37 Ge+ 36 40 + 38 38 + 31 14 +
Ar Ar , Ar Ar , P2 N
77 40 37 + 40 36
Se 7.63 - Ar Cl , Ar ArH+
78 78 + 38 40 + 31 16 +
Se 23.77 Kr Ar Ar , P2 O
80 80 + 40
Se 49.61 Kr Ar2+
82 82 + 40
Se 8.73 Kr Ar2H2+,81BrH+, 34S16O3+
Table 2: Relative natural abundances and some interferences on selenium isotopes.
AS-10 column versus ANX 3206 column
Initially the ANX 3206 column was used for the determination of the inorganic selenium species. The
specifications for the eluent concentrations provided by the manufacturers for optimal separation of
selenate and selenite were applied. The optimisation study included different flow rates (100, 200,
300 ml min-1) and sample loop volumes (10, 25, 50 ml). The analysis of the actual industrial waste
water led to the following observations: the samples contained large amounts of chloride and
bromide ions; and, in some treated waste waters, an unknown selenium peak appeared in front of
the selenite (Fig. 1).
Se IV
?
Se VI
2.2
2.0
1.8
1.6 HBr
Signal intensity
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32
time/min
Fig.1 Microbial treated industrial waste water analysed with ANX 3206 column, eluent 5 mM ammonium
-1 82
malonate, flow rate of 200 µl min , monitored isotope Se.
A sample loop volume of 10 ml combined with a flow rate of 200 ml min-1 gave optimal separation of
the selenium species and interfering chlorides and bromides. The long term stability of the ICP-MS
measurement on the quality control standard of 25 ng ml-1 Se(IV) in 5 mM ammonium malonate was
2% (12 h, n=5). The unknown peak could be identified as selenium after comparison of the signal of
the four different selenium isotopes.
In an attempt to identify this unknown selenium species, standard solutions of DL-selenocystine, DL-
selenomethionine and DL-selenoethionine were analysed. Due to a minor resolution of the organic
selenium components, the comparison of retention times of these organic selenium components
with the unknown selenium species in the chromatogram were ambiguous. In order to identify the
unknown peak, the AS-10 column was tested. The same optimization criteria were examined (flow
rate of 200 ml min-1; 10 ml loop), leading to a better resolution of the selenium species (Fig. 2).
Se IV
4.0
Signal intensity
3.0
Se VI
2.0
1.0
-0.0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34
time/min
-1 IV VI
Fig. 2: 50 ng ml Se and Se measured on ANX 3206 (upper), eluent 5 mM ammonium malonate, flow rate
-1 -1
200 µl min ; measured on AS-10 (lower), eluent 100 mM NaOH + 1 % EtOH, flow rate 200 µl min .
Successive spiking of the industrial waste water with DL-selenocystine, DL-selenomethionine and DL-
selenoethionine did not help to positively identify the unknown selenium species. No evidence of DL-
selenoethionine was expected, since it does not occur naturally. Most frequently, reduced selenium
is bound to amino acids and released when the cell dies.9,10 Also methylation reactions were
mentioned by Altringer et al.11 In a study on the aquatic chemistry of selenium by Cooke et al.12 in
surface waters, six dissolved selenium species were identified: selenate and selenite, non-volatile
organic selenides (selenoamino acids and dimethylselenonium ion) and the volatile methylated forms
dimethyl selenide and dimethyl diselenide. As equal sensitivity was observed for the organic and
inorganic selenium species, the concentration of the unidentified selenium species could be
determined and amounted to a maximum of 20% of the total selenium content. On the basis of the
better performance characteristics, the AS-10 column was used for further investigations.
Signal enhancement
Several studies have demonstrated the signal enhancing effect of the addition of carbon to the
solution to be analysed, or in the analysis of carbon containing solutions such as wine and port.3,4,13
The addition of organic compounds modifies the ionisation equilibrium in the plasma. Different
mechanisms for this signal enhancement have been proposed.3 The most pronounced effect is seen,
as stated earlier, on elements such as selenium and arsenic with ionization energies in the range of
9±11 eV.
An overview and study of different carbon-containing solutions is presented by Gammelgaard et al.4

In our study ethanol was used as the enhancement solution. It has been shown that optimum RF
power and nebuliser gas flow depend on the content of ethanol in the solution. Initially the ICP-MS
instrument settings were optimised with a 200 ng ml-1 Se (100 mM NaOH, 2% v/v ethanol) solution
and were kept constant during the experiments. Several samples of 200 ng ml-1 Se solutions in 100
mM NaOH with increasing amounts of ethanol (0.5±5% v/v) were analysed and the net results
(corrected for background contribution) are shown in Fig. 3 for three different isotope masses. In
order to stabilize the signal, the measurement was performed after aspiration of the solution for
approximately 10 min. A continuous decrease (20%) of the signal of the 5% v/v ethanol solution
during the aspiration period was observed. To eliminate memory effects of the ethanol addition,
rinsing of the system for 60 min was needed to reach the basic level. As can be seen in Fig. 3 an
enhancement factor of 5 was achieved in 2% v/v ethanol, which is in agreement with literature data.3
600%
500%
relative intensity
400%
300%
200%
100%
0 1 2 3 4 5
ethanol (% v/v)
-1
Fig. 3: Net relative intensities of 200 ng ml Se 100 mM NaOH solutions in increasing amounts of ethanol
78 77 82
(lines represent Se, Se and Se).
Analytical characteristics
Internal standardisation is used to correct for signal drift, instrument instability, signal suppression or
signal enhancement caused by the matrix. In previous studies the similarity in mass of the internal
standard and the element of interest was a major point of concern for a good correction. The use of
an enhancement reagent assumes an internal standard with a first ionization energy close to that of
selenium. Because speciation studies are performed, one can choose a stable selenium compound
not present in the sample under investigation, e.g., DL-selenomethionine, to act as an internal
standard for the other selenium compounds. A possible cause of error is that the component can
degrade to a selenium compound of interest during the measurement run. To counteract the
possible degradation of the internal standard and the consequent formation of selenium compounds
of interest, the use of a stable organic arsenic compound, e.g., arsenobetaine, can be considered. No
evidence of arsenobetaine (nor DL-selenomethionine) is expected in the treated industrial waste
water. Arsenic is close to selenium regarding its mass and ionization energy and may therefore be
used as an internal standard. Arsenic is mono-isotopic resulting in high sensitivity and it has been
demonstrated that arsenobetaine is a stable compound.7 Therefore 100 ng ml-1 arsenobetaine (CRM
626) was added to every analysed sample to correct for signal instability. After filtration of the
microbial-treated waste water to remove microorganisms, the sample was diluted in 100 mM NaOH
just before measurement and the internal standard was added. For a typical routine measurement
run, including calibration standards, quality control solutions and waste water, the integrated peak
areas of the internal standard in each sample are shown in Fig. 4.
3.5E+06
3.0E+06
2.5E+06
2.0E+06
cps
1.5E+06
1.0E+06
5.0E+05
0.0E+00
0 5 10 15 20 25 30 35 40
time (hours)
-1
Fig. 4: Integrated peak areas of the arsenobetaine internal standard (100 ng ml CRM 626) on 54 consecutive
IC-ICP-MS determinations measured over a period of 40 hours (relative standard deviation 2.8 %).
The IC-ICP-MS results showed good long term stability; a relative standard deviation of 2.8% is
calculated for 54 consecutive measurements over a period of 40 h. The repeatability of the quality
control standards over the same period, after internal standard correction, amounted to 25.2 ± 0.4
ng ml-1 Se(IV) (n=6) and 506.8 ± 6.8 ng ml-1 Se(VI) (n=6). An overview of the simultaneous monitored
isotopes is given in Fig. 5.
As-B
Cl-
Br-
Se IV
Se-
8 Met
As V
6
Signal intensity
Se VI m/z 75
4
m/z 35
m/z 81
2
m/z 77
? HBr
-0 m/z 82
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34
Time (min)
-1 -1
Fig. 5: Industrial treated waste water spiked with 100 ng ml Se DL-Selenomethionine (Se-Met) and 100 ng ml
-1
arsenobetaine (As-B) analysed with AS-10 column, eluent 100 mM NaOH + 2% v/v EtOH, flow rate 200 µl min ,
10 µl sample loop. The unidentified selenium species is marked with a (?).
It can be seen that the interfering chlorides and bromides are well separated from the analytical
species of interest. A limit of quantification (10σ) of 1 ng ml-1 Se was calculated based on the
standard deviation of 5 replicates of a 2 ng ml-1 Se(IV) solution. In the same run a limit of
quantification (10σ) of 0.3 ng ml-1 As was calculated based on the standard deviation of 5 replicates
of a 2 ng ml-1 As(V) solution. The detection limits for the selenium species are in agreement with the
HPLC-ICP-MS results reported by Pedersen et al.6 using 3% methanol as enhancement reagent.
Stability
For accurate determination of selenium species, major concerns are the sampling, storage and
sample pretreatment. At all these stages, volatilization, adsorption, desorption, contamination and
interconversion of species, etc. may occur. Extensive stability studies of selenium species in a variety
of matrices have been reviewed.1,14 In pure aqueous solution at -20°C without acidification, no
variation of selenate and selenite was observed over 12 months. At room temperature the maximum
time of sample storage was found to be between 2 and 9 months depending on the material of the
container. It was also noticed that chlorides tend to stabilise both species. A seawater reference
material certified for selenite was stable over 5 years if stored in a cool place. A ground water
containing selenium was stable for up to 13 months stored at 4 °C in a polyethylene container and
the stability of selenium compounds in purified samples followed the order
selenate>selenomethionine>selenite.1 In other studies, however, losses of selenium and a conversion
of selenite to selenate after four weeks was observed in groundwater samples stored in polyethylene
at 4 °C.14 Long-term stability studies of organic selenium species in aqueous solutions by Olivas et
al.15 showed that the best results were obtained with Pyrex containers stored at 4 °C in the dark. No
significant losses of selenocystine and selenomethionine were observed over one year. Rapid losses
of trimethylselenonium, however, were observed under any conditions.
In this study the industrial waste waters were filtered through a 0.45 µm membrane filter to remove
suspended solids and micro organisms, and stored in Pyrex containers in the dark at 4 °C. Prior to
measurement the samples and calibration solutions were diluted 5-fold in 100 mM NaOH. No
conversion of selenite, DL-selenomethionine and DL-selenoethionine was observed over a period of
12 months. The selenate stock solution showed a 2% conversion to selenite after a period of 6
months. In the chromatogram of freshly prepared DL-selenocystine in 100 mM NaOH two peaks
were observed.
The smallest peak was expected to be selenocysteine, the reduced form of selenocystine. Similar
observation in stock standard solutions of selenocystine has been reported by Emteborg et al.16 Since
the retention time did not overlap with the unidentified selenium species in the industrial waste
water no further investigations were performed (e.g., reduction of DL-selenocystine with
mercaptoethanol as described by Burnell et al.17).
Microbial selenium removal
As this paper concentrated mainly on the development of a routine IC-ICP-MS method, the results of
microbial selenium removal are briefly summarised. For an extensive review of the over 600 analysed
waste water samples we refer to the synthesis report of the coordinating Brite-Euram project.18 It
was observed that selenate and selenite are in competition with nitrate (denitrification) as the
electron acceptor. The industrial waste water contained nitrate at nitrogen concentrations between
10 and 50 mg ml-1, hampering the selenium reduction. The bacteria could reduce the effluent
selenium concentration to concentrations below 1 mg ml-1. The reduction of selenate is observed
after complete removal of selenite. The reduction of selenate to the metallic form was the most
time-consuming step in the removal of selenium from industrial waste water but was never
complete. The selenium removal by Pseudomonas fluorescens J3/73b from a non-ferrous waste
water is shown in Fig. 6. The selenium concentration was reduced from 2.6 mg ml-1 to 0.5 mg ml-1.
Selenite is completely reduced in contrast to selenate. Experiments showed optimal reduction of
selenium under anoxic conditions with glucose as the carbon source.
2.5 Se IV (IC-ICP-MS)
Se VI (IC-ICP-MS)
2
total Se (ICP-AES)
[Se ] µg ml -1
1.5
0.5
0
0 1 2 3 4 5 6
time (days)
Fig. 6: Microbial selenium removal by Pseudomonas fluorescens J3/73b in a non-ferrous industrial waste water.
Conclusion
A robust routine IC-ICP-MS method for the determination of selenate and selenite and some organic
selenium species is presented. To correct for signal drift, instrument instability and signal
enhancement caused by the addition of 2% v/v ethanol to the eluent, an arsenobetaine solution is
used successfully as the internal standard for the analysis of industrial waste water. The addition of
ethanol results in a selenium signal enhancement factor of 5 and a quantification limit of 1 ng ml-1 Se.
The reduction of selenate to the metallic form is the most time-consuming step in the removal of
selenium from industrial waste water. The microbial removal of selenite is, in contrast to selenate,
complete.
Acknowledgements
We wish to thank Magnus Johansson (IRMM, Geel, Belgium) for aid in the instrumental setup and discussions
on the subject, Valère Corthouts (VITO) for the development of an uncloggable MCN and Jos Broeckx (VITO) for
total selenium determinations using ICP-AES. Furthermore we would like to thank Dirk Mestach (Nijlen Correct
Translating Agency) for a critical review.
This work was supported by the Brite-Euram programme of the European Commission in the project (BE96-
3584) ``Development and assessment of biotechnological methods for the removal of metal anions from waste
water (BIROX)''.
References paper, Characterization of inorganic selenium species by ion chromatography with

ICP-MS detection in microbial-treated industrial waste water, J. Anal. At. Spectrom., 2000, 15, 1087-
1092.
1. D.J. Adams, P.B. Altringer and W.D. Gould, Biohydrometallurgical Technologies, Edited by A.E.
Torma, M.L. Apel and C.L. Brierley, The Minerals, Metals & Materials Society, 1993, 755 - 771.
2. A. D'Ulivo, Analyst, 1997, 122, 117R - 144R.

3. B. Gammelgaard and O. Joens, J. Anal. At. Spectrom., 1999, 14, 867 - 874.
4. J. Goossens, F. Vanhaecke, L. Moens and R. Dams, Anal. Chim. Acta, 1993, 280, 137.
5. A. Woller, H. Garraud, J. Boisson, A.M. Dorthé, P. Fodor and O.F.X. Donard, J. Anal. At. Spectrom.,
1998, 13, 141 - 149.
6. G.A. Pedersen and E.H. Larsen, Fresenius' J. Anal. Chem., 1997, 358, 591 – 598.
7. European Commission, BCR Reference Materials, The Certification of the Contents (mass
fractions) of Arsenobetaine in Solution (CRM 626) and of Total Arsenic, Arsenobetaine,
Dimethylarsenic acid in Tuna Fish Tissue (CRM 627), 1997, ISBN 92-828-2536-1.
8. K.J.R. Rosman, P.D.P. Taylor, Pure & Appl. Chem, 1998, 70 (1), 217.
9. T.L. Gerrard, S.N. Telford and H.H. Williams, Journal of Bacteriology, 1974, 3, 1057 -1060.
10. B.A. Silverberg, P.T.S. Wong and Y.K. Chau, Arch Microbiol. 1976, 107, 1- 6.
11. P.B. Altringer, D.M. Larsen and K.R. Gardner, 1989, Biohydrometallurgy CANMET SP89-10, editors
J. Salley, R.G.L. McCready, P.L. Wichlace.
12. T.D. Cooke and K.W. Bruland, Environ. Sci. Technol., 1987, 21, 1214 - 1219.
13. C. Marisa, R. Almeida, M. Tiresa, S.D. Vasconcelos, Anal. Chim. Acta., 1999, 396, 45 - 53.
14. I. Héninger, M. Potin-Gautier, I. de Gregori and H. Pinochet, Fresenius' J. Anal. Chem., 1997, 357,
600 - 610.
15. R.M. Olivas, P. Quevauviller and O.F.X. Donard, Fresenius' J. Anal. Chem., 1998, 360, 512 - 519.
16. H. Emteborg, G. Bordin and A.R. Rodriguez, Analyst, 1998, 123, 245 - 253.
17. J.N. Burnell, J.A. Karle and A. Schrift, J. Inorg. Biochem., 1980, 12, 343 - 351.
18. Brite-Euram programme, European Commission project (BE96-3584), “Development and
assessment of biotechnoligical methods for the removal of metal anions from waste water
(BIROX)”.
3.3. THE BAN ON HEXAVALENT CHROMIUM
Chromium has been used extensively in the chemical industry in applications such as metal plating,
leather tanning, pigment or anti-corrosion coatings [126]. Chromium production processes, as well as
its different uses, are responsible for the release of chromium compounds into the environment (soil,
surface and ground waters).
Chromium can exist in numerous chemical forms with oxidation states ranging from 0 to VI, the
trivalent and hexavalent forms being more likely to occur in the environment due to their higher
stability. An Eh-pH diagram of the Cr-O-H system, depicting the dominant aqueous species and stable
solid phases, is represented in Figure 31 [127]. When assessing the potential toxicity of chromium,
the concept of speciation must be emphasized: Cr(III) is considered as an essential micronutrient in
the human diet and is used as a nutritional supplement for humans and animals (see also Table 1,
p.27). In contrast, Cr(VI) is highly toxic in case of both acute and chronic exposure and its compounds
are regulated through the Dangerous Substances Directive and classified as human carcinogens by
the United States Environmental Protection Agency (US EPA) and the International Agency for
Research on Cancer (IARC).
-10 5
Figure 31: Eh-pH diagram of the system Cr-O-H (Σ [Cr] = 10 , 298.15 K, 10 Pa)(figure courtesy of [127]).
The difference in toxicological properties between Cr(III) and Cr(VI) compounds is closely related to
the chemical characteristics of the species, which also influence the stability, mobility, and
bioavailability of these species in the environment. Several recent regulations have incorporated the
speciation of elements in restrictions imposed on environmental quality (workplace air) and on

products (cement, packaging, toys and anti-corrosion coatings).
Analytical techniques that allow the distinction between and quantification of Cr(III) and Cr(VI)
species are necessary for successful implementation of such regulations. The major challenge
encountered by these techniques is to avoid any change of chromium speciation during the analysis.
When wet-chemistry analytical methods are to be used on solid samples (e.g., soils, waste, etc.), an
extraction step has to be performed. This step can lead to difficulties with regards to the yield of
extraction, meaning the result would be underestimated in case of incomplete extraction. In the
following paragraphs, studies related to the determination of Cr(VI) on solid samples are summarised
in chronological order.
3.3.1. THE BAN ON HEXAVALENT CHROMIUM IN PACKAGING MATERIAL
In 2001, VITO was commissioned by the Belgian Federal Public Service of Health, Food chain safety
and Environment to review the most critical types of packaging on the Belgian market with respect to
the concentration levels of lead, cadmium, mercury and hexavalent chromium. According to EU
directive 94/62/EC on Packaging and Packaging waste, member states shall ensure that the sum of
these four elements shall not exceed 100 mg kg-1.
In the qualitative and quantitative market analysis performed by VITO, some 200 packaging materials
were screened with X-ray fluorescence for total metal content [132]. A selection of 40 critical
packaging materials (total chromium content higher than 100 mg kg-1) were further analysed for
Cr(VI) content. The materials analysed were, among others, glass packaging, plastic packaging,
wrapping papers and metal packaging.
Different extracting solutions have been used for the extraction of Cr(VI) form solid materials. The
method yielding the best results relied on a combination of Na2CO3 and NaOH with continuous
swirling and heating at 95°C. This approach was promoted in 1996 by the US EPA with the reference
US EPA Method 3060A and was adopted as basis of the European standard EN 15192:2006 as well
(see further, §3.3.2).
Once the Cr(VI) is extracted, different quantification methods are available. For the purpose of
screening critical types of packaging on the Belgian market, two different methods for the
determination of Cr(VI) were implemented at VITO. The first method consisted of an ion
chromatography-1,5-diphenylcarbazide spectrophotometry method, in which Cr(VI) in solution
oxidizes 1,5-diphenylcarbazide to give a red/violet complex with chromium, which can be quantified
photometrically at 540 nm. The second and most accurate one is performed using high-performance
liquid chromatography coupled to inductively coupled plasma-mass spectrometry (HPLC/ICP-MS) and
involves isotopic spiking of the sample to enable the correction of species interconversion taking
place during preparation and analysis steps [141] (see paper: Determination of hexavalent chromium
by species-specific isotope dilution mass spectrometry and ion chromatography-1,5-
diphenylcarbazide spectrophotometry, J. Anal. At. Spectrom., 2003).
Determination of hexavalent chromium by species-specific isotope dilution

mass spectrometry and ion chromatography-1,5-diphenylcarbazide
spectrophotometry
Kristof Tirez1, Wilfried Brusten1, Annick Cluyts1, Johan Patyn1 and Nicole De Brucker1
1

Abstract
The accuracy of the determination of hexavalent chromium (Cr6+) in solid matrices remains a
challenging field of effort and improvement. An alkaline digestion (0.5 M NaOH / 0.28 M Na2CO3)
followed by ion chromatography and spectrophotometrical determination by post column
derivatisation with 1,5-diphenylcarbazide (DPC) provides accurate and reproducible results. To gain a
better insight into the species interconversion during the digestion the concept of species-specific
isotope dilution mass spectrometry by adding enriched chromium isotopes that have been chemically
converted into trivalent (Cr3+) and hexavalent chromium was used. For the analysis of the digestion
solutions ion chromatography coupled to inductively coupled plasma – mass spectrometry (IC-ICP-MS)
was peformed. The limit of detection in the digestion solution by the latest technique amounted 0.8
ng ml-1 Cr6+ .The accuracy of an alkaline digestion on solid waste materials was compared with a
water extraction (leaching test industrial waste), indicating that the alkaline digestion is accurate
from point of view of minimal species interconversion on the one hand and maximum amount Cr6+
extracted on the other, while significant oxidation and reduction reactions were observed during
leaching tests with water. In the frame of European Directive 94/62/EC on Packaging and Packaging
waste, the developed analytical method has been successfully used to compare the concentrations of
Cr6+ on 40 packaging materials against the regulatory limit value of 100 mg/kg by weight (sum of
concentration levels of lead, cadmium, mercury and hexavalent chromium).
Introduction
An accurate and precise method for extracting and analyzing hexavalent chromium (Cr6+) in soils,
sediments, waste materials and airborne particulates is needed because of human, ecological and
legal concerns related to Cr6+ in the environment [1]. Cr6+ compounds are known to be toxic and
carcinogenic agents for a variety of organisms. They are mobile in soil/water systems, much more
than trivalent chromium (Cr3+) compounds, because Cr6+ compounds are strong oxidizers and highly
soluble as anionic forms, while Cr3+ compounds tend to form relatively inert precipitates near neutral
pH [2].
The distribution of Cr6+ and Cr3+ strongly depends on pH and potential. According to the Eh-pH phase
diagram, Cr6+ is thermodynamically stable at relatively higher pH and Eh, while Cr3+ is stable at
relatively low pH and Eh [3]. Kinetics also play an important role in the interconversion between Cr6+
and Cr3+ [4].
From the literature published in the last years it can be concluded that there is a good number of
procedures including different analytical techniques for the speciation of Cr6+ and Cr3+ in solution [5-
12]. When it comes to chromium speciation in solid matrices, accuracy of the methods remains a
field of additional effort and improvement. A review on analytical methodologies for chromium
speciation in solid matrices by Marques et al. emphasis the lack of reported recovery by most
authors [2]. Possible analytical biases in the analysis of Cr6+ in soil samples using an alkaline digestion
solution (US EPA 3060A) have been described [1,4,13]. The results indicated that the digestion (US
EPA 3060A) combined with a spectrophotometric detection method (US EPA 7196A) provided
satisfactory performance in quantifying the amount of total Cr6+ [14]. The authors stressed the
necessity however to characterise ancillary parameters, such as oxidation reduction potential (ORP),
pH, TOC, Fe2+ and S2- to make an affirmative determination regarding the capacity of the sample to
contain Cr6+. Reducing conditions, as defined by the Cr Eh-pH phase diagram, the presence of TOC, S2-
or Fe2+, or acidic conditions, singularly or in combination, indicate the potential for a sample to (i)
reduce a laboratory Cr6+ spike or (ii) not sustain the existence of Cr6+ in the sample's native
environment [15,16]. The results also showed that both water soluble (K2CrO4) and insoluble
(BaCrO4, PbCrO4) forms of Cr6+ can be used to obtain satisfactory matrix spike recovery results.
Furthermore method-induced oxidation only occurs with freshly precipitated Cr(OH)3, which is not
likely to be present in environmental samples. Aged Cr(OH)3 as well as Cr2O3 do not oxidize when
subjected to method US EPA 3060A [1]. This viewpoint of analytical biases in the analysis of Cr6+ in
soil samples is subscribed by Huo et al. as well [4]. The author compared speciated isotope dilution
mass spectrometry (SIDMS) with a diphenylcarbazide spectrophotometric method for the
determination of Cr6+ in soil and indicated possible biases during digestion and the possibility to
mathematically correct for species transformation. The principle of SIDMS has been demonstrated in
the patent and literature [10,17].
SIDMS uses the concept of spiking the sample with known amounts of enriched isotopes that have
been chemically converted into the same forms of the species to be analysed. The isotopic spike for
each species has a unique isotope enrichment. In the case of chromium, two isotopic spikes, 53Cr6+
enriched in 53Cr and 50Cr3+ enriched in 50Cr, are prepared for the simultaneous determination of Cr3+
and Cr6+. Environmental samples containing Cr species are spiked with both 53Cr6+ and 50Cr3+. The
species are separated at the end of the manipulation by means of chromatography and different
isotope ratios are measured. In contrast to "classic" chromatography, which is a snapshot in time
recording the state of affairs at the end of the manipulation, any interconversions that occur after
spiking are traceable by SIDMS and can be quantitatively corrected by monitoring isotopes in each
species. The determination of chromium by inductively coupled plasma-mass spectrometry (ICP-MS)
in environmental samples is hampered by polyatomic interferences [12,18]. The combination of
elements that can give rise to interference at the analytical masses of Cr (i.e., 50, 52, 53, 54) are Ar,
C, Ca, Cl, K, O and S, in other words, those normally present in environmental matrices as well as the
plasma gas and in the reagents used for digestion. The use of ion chromatography when coupled to
an ICP-MS results, besides the speciation of the chromium species, in an opportunity to separate
possible interferences from the isotopes of interest.
However for so far, and as stressed by the authors of the SIDMS method, the monitoring and
correction for the oxidation of Cr3+ during the preparatory alkaline digestion was under investigation.
In contrast to solutions where the Cr3+ and Cr6+ content and consequently the interconversion of
species can be determined via SIDMS, the preliminary alkaline digestion of the solid sample followed
by filtration makes a reproducible recovery of the Cr3+ spike almost impossible. The alkaline digestion
at pH 12, used to stabilize the original Cr6+ in the sample, makes the Cr3+ spike to precipitate.
Although acidifying the filtrate afterwards would thermodynamically resolve the added Cr3+ spike,
this is due to the slow kinetic properties of Cr3+ on the one hand and the instability of Cr6+, the
element of concern, under these conditions a tricky task [19]. This statement was confirmed in
preliminary experiments performed in our lab, where the enriched spike solutions 50Cr3+ was added
to the digestion mixture (pH 12) and could not be recovered as soluble chromium species even after
acidification and gentle heating. Coedo et al. reported however a speciation of chromium in
steelmaking solid wastes, where after the alkaline digestion soluble Cr3+ compounds present as the
negatively charged Cr(OH)4- could be determined [20]. The authors present a method to differentiate
Cr6+, soluble Cr3+ as the negatively charged Cr(OH)4- complex and insoluble Cr3+ (coming from Cr2O3
and Cr metal). For the determination of the latter a fusion was performed. In none of the analysed
samples however soluble Cr3+ as the negatively charged Cr(OH)4- complex was found.
The interconversion of species can not be determined via SIDMS in soils as the Cr3+ present may be
chemically different from the added spike. As mentioned previously, Cr2O3 and aged Cr(OH)3
precipitate are resistant to oxidation, free Cr3+ and freshly precipitated Cr(OH)3 are relatively easy to
oxidize. Although the fresh Cr(OH)3 precipitate is not a representative form of Cr3+ in environmental
samples, the possibility of biases caused by oxidised Cr3+ can not be excluded [4].
A review of different digestion methods from soil is given by James et al. [21]. There is at this time a
consensus on the use of an alkaline digestion solution to achieve minimal species interconversion
and maximum extraction efficiency of hexavalent chromium on solid materials (see also §3.3.2).
Due to the impossibility to measure Cr3+ in the alkaline digestion solution the existing calculations for
speciated isotope dilution mass spectrometry on liquid solutions had to be modified in a way that
maximum information can still be obtained from the spiking of both chromium species. A method of
calculation is presented in this paper indicating if conversion from Cr6+ to Cr3+ or Cr3+ to Cr6+ was likely
to occure during the digestion.
The determination of hexavalent chromium is required in legislation of some member states of the
European Union concerning acceptance criteria for industrial waste on landfill and regulatory values
in packaging material. In this paper results are shown of the application of this analytical method of
species-specific isotope dilution mass spectrometry is these two European legislation related topics.
Method of calculation
Due to the impossibility to measure Cr3+ in an alkaline solution, the existing calculations for speciated
isotope dilution mass spectrometry as presented in the US EPA 6800 method and U.S patent
5,414,259 [17,22] had to be modified in a way that maximum information can still be obtained from
the spiking of both chromium species. In these two references is demonstrated that the
concentration of the different species as well as the interconversion can be deconvoluted from the
following equations:
( 50
x Wx +
A x C III 50
)
A sIII C sIII WsIII (1 − α ) + ( 50
x Wx +
A x C VI 50
A sVI C sVI WsVI β )
/ 52 =
III
R 50
( 52
x Wx +
A x C III 52
A sIII C sIII WsIII )(1− α ) + ( 52
x Wx +
A x C VI 52
A sVI C sVI WsVI )β
R III
=
( 53
x Wx +
A x C III 53
)
A sIII C sIII WsIII (1 − α ) + ( 53
x Wx +
A x C VI 53
A sVI C sVI WsVI β )
53 / 52
( 52
x Wx +
A x C III 52
A sIII C sIII WsIII )(1− α ) + ( 52
x Wx +
A x C VI 52
A sVI C sVI WsVI )β
( 50
x Wx +
A x C III 50
) (
A sIII C sIII WsIII α + 50
x Wx +
A x C VI 50
)
A sVI C sVI WsVI (1 − β )
/ 52 =
VI
R 50
( 52
x Wx +
A x C III 52
A sIII C sIII WsIII )α + ( 52
x Wx +
A x C VI 52
A sVI C sVI WsVI )(1 − β)
R VI
=
( 53
x Wx +
A x C III 53
) (
A sIII C sIII WsIII α + 53
x Wx +
A x C VI 53
)
A sVI C sVI WsVI (1 − β)
53 / 52
( 52
x Wx +
A x C III 52
A sIII C sIII WsIII )α + (52
x Wx +
A x C VI 52
A sVI C sVI WsVI )(1 − β)
where,
III
R 50 / 52 is the mass bias corrected measured isotope ratio of 50Cr to 52Cr of Cr3+ in
the spiked sample
50
A x is the atomic fraction of 50Cr in the sample
3+
C III
x is the concentration of Cr in the sample (µmole/g, unknown)
50
A sIII is the atomic fraction of 50Cr in the 50Cr3+ spike
C sIII is the concentration of Cr3+ in the 50Cr3+ spike (µmole/g)
C sVI is the concentration of Cr6+ in the 53Cr6+ spike (µmole/g)
WsIII is the weight of the 50Cr3+ spike (g)
C VI
x is the concentration of Cr6+ in the sample (µmole/g, unknown)
α is the percentage of Cr3+ oxidized to Cr6+ after spiking (unknown)
β is the percentage of Cr6+ reduced to Cr3+ after spiking (unknown)
These formulas are based on a chemical pure isotope enriched spike of both components. In other
words the chromium isotope enriched material contains only one redox species of chromium. To
confirm this assumption both solutions of isotope enriched material were analysed in order to
control their chemical purity. While no detectable 50Cr6+ was present in the 50Cr3+ isotope enriched
spike solution, 2 % of the isotope enriched 53Cr6+ spike was present as 53Cr3+. For this reason the
above mentioned formulas had to be changed in order to correct for this impurity of 53Cr3+ in the
isotope enriched 53Cr6+ spike. This was done by introducing a chemical purity factor (P) in these
equations. This purity factor stands for the fraction Cr6+ in the 53Cr isotope enriched material.
( 53
x Wx +
A x C III A s Cs WsIII + (1 − P) ⋅ 53 A sVICsVI WsVI (1− α ) +
53 III III
) ( 53
x Wx + P ⋅ A s Cs Ws
A x C VI 53 VI VI VI
β )
/ 52 =
III
R 53
( 52
A x C Wx + A C W
III
x
52 III
s
III
s s
III
+ (1 − P) ⋅ A C W
52 VI
s
VI
s s
VI
)(1− α ) + ( 52
A x C Wx + P ⋅ A C W
VI
x
52 VI
s
VI
s s
VI
)β
( 50
x Wx +
A x CIII As Cs WsIII + (1 − P) ⋅ 50 AsVICsVI WsVI (1− α ) +
50 III III
) ( 50
A x CVI 50 VI VI VI
β )
/ 52 =
III
R 50
( 52
A x C Wx + A C W
III
x
52 III
s
III
s s
III
+ (1 − P) ⋅ A C W
52 VI
s
VI
s s
VI
)(1− α ) + ( 52
VI
x
52 VI
s
VI
s s
VI
)β
( 53
x Wx +
A x CIII As Cs WsIII + (1 − P) ⋅ 53 AsVICsVI WsVI α +
53 III III
) ( 53
A x CVI 53 VI VI VI
(1 − β) )
/ 52 =
VI
R 53
( 52
A x C Wx + A C W
III
x
52 III
s
III
s s
III
+ (1 − P) ⋅ A C W
52 VI
s
VI
s s
VI
)α + ( 52
VI
x
52 VI
s
VI
s s
VI
)(1 − β)
( 50
x Wx +
A x C III A s Cs WsIII + (1 − P) ⋅ 50 A sVI CsVI WsVI α +
50 III III
) ( 50
x Wx + P ⋅ A s C s Ws
(1 − β) )
/ 52 =
VI
R 50
( 52
x Wx +
A x C III A s Cs WsIII + (1 − P) ⋅ 52 A sVI CsVI WsVI
52 III III
)α + ( 52
x Wx + P ⋅ A s C s Ws
)(1 − β)
The concentration of Cr3+ and Cr6+ in the sample as well as possible interconversions can be derived
from these equations. However, as mentioned above, due to the alkaline digestion no Cr3+ spike is
recovered and consequently the R 50 III
/ 52 and R 53 / 52 can not be measured. Therefore a new model
III
was conceived based on the ratios R 50

VI
/ 52 and R 53 / 52 ,
VI
P⋅ 50
A sVI C sVI WsVI + 50
A x C VI VI
x Wx + 50
A sIII C sIIIconvert WsIII
/ 52 =
VI
R 50
P⋅ 52
A x C VI VI
x Wx + 52
A sIII C sIIIconvert W
P ⋅ 53 A sVI C sVI WsVI + 53

A x C VI VI
x Wx + 53
A sIII C sIIIconvert WsIII
R VI
53 / 52 =
P⋅ 52
A x C VI VI
x Wx + 52
A sIII C sIIIconvert W
where,
P chemical purity factor, fraction Cr6+ in the 53Cr enriched material.
C sIIIconvert the concentration of Cr3+ in the 50Cr3+ spike converted to Cr6+
These equations take into account the possible conversion of the added isotope enriched Cr3+ spike
to Cr6+ during the digestion and subsequent handling. When Cr3+ oxidation occures, the Cr3+ present
as chemical impurity in the isotope enriched Cr6+ spike will also be (partially) converted and has to be
included into the equation,
C sIIIconvert
P⋅ 50
A sVI C sVI WsVI + 50 A x C VI
x Wx + A s C s convert Ws +
VI 50 III III III
⋅ (1 − P) ⋅ 50
A sVI C sVI WsVI
III
C
/ 52 =
VI s
R 50 III
C
P⋅ A sVI C sVI WsVI + 52 A x C VI
x Wx + A s C s convert Ws + ⋅ (1 − P) ⋅
52 VI 52 III III III s convert 52
A sVI C sVI WsVI
III
C s
III
C
P ⋅ 53 A sVI C sVI WsVI + 53 A x C VI
x Wx + A s C s convert Ws + ⋅ (1 − P) ⋅ 53 A sVI C sVI WsVI
VI 53 III III III s convert
III
C
/ 52 =
VI s
R 53 III
C
P⋅ A sVI C sVI WsVI + 52 A x C VI
x Wx + A s C s convert Ws + ⋅ (1 − P) ⋅
52 VI 52 III III III s convert 52
A sVI C sVI WsVI
III
C s
6+
From these equations the following relative fractions of Cr present in the alkaline digestion can be
derived,
C sVI
F VI
= : the fraction Cr6+ of the enriched 53Cr6+ spike
+C +C
spike VI III VI
C s sconvert x
VI
C
VI
Fnatural = x
: the fraction natural Cr6+ extracted from the sample
C VI
s +C III
sconvert +C VI
x
III
C
III
Fspike = sconvert
: the fraction converted Cr3+ from the enriched 50
Cr3+ spike
C VI
s +C III
sconvert + C xVI
When the total amount of Cr6+ present in the sample, Crtotal VI

, is calculated by means of linear
6+
regression based on natural Cr standard solutions and summation of the integrated peak areas on
all chromium isotopes, the concentration of the different fractions Cr6+ in the alkaline digest can be
derived,
[Cr ]VI
spike = Fspike
VI
⋅ Crtotal
VI
[ ] : the concentration of 53Cr6+ spike
[Cr VI
natural ] = VI
Fnatural [
⋅ Crtotal
VI
] : the concentration of natural Cr6+
[Cr ]III
spike = Fspike
III
⋅ Crtotal
VI
[ ] : the concentration of converted 50Cr3+ spike
In contrast to SIDMS the calculated Cr6+ concentration extracted from the sample is not corrected for
possible reduction reactions encountered during the alkaline digestion. However the actual
concentration of the 53Cr6+ spike in the digest is a measure of the recovery and gives an indication of
conversion from Cr6+ to Cr3+ was likely to occure during the digestion. The actual concentration of the
converted 50Cr3+ spike in the digest gives an indication of the fact that conversion (oxidation) from
Cr3+ to Cr6+ was likely to occure during the digestion. One could correct for these possible reactions
by first determing the amount of Cr3+ in the sample by substracting the total amount of chromium
(after acid digestion) by the amount of Cr6+ in the sample and secondly take into account the
percentage of Cr3+ converted to Cr6+. Because of the different properties of the chemical forms of
Cr3+ regarding this digestion and the slow kinetic properties of Cr3+ it is however questionable to
compare the behaviour of the native Cr3+ with the added 50Cr3+ spike.
Experimental
Instrumentation
Samples were analysed on an Elan 6000 ICP-MS system (Perkin-Elmer, SCIEX, Thornhill, ON, Canada)
equipped with a MCN-100 micro-concentric nebuliser (Cetac technologies, Omaha, NE, USA) fitted to
a Scott-type double pass spray chamber (Perkin-Elmer, see also Picture 1). The speciation of the
redox species was carried out by anion exchange chromatography on an IONPAC AS 11 microbore
column (2 mm i.d.) and an IONPAC AS 10 microbore column (Dionex, Sunnyvale, CA, USA). Both
columns were provided with a guard column to remove dissolved and particulate contaminants from
the sample and eluent before reaching the column. The eluent was pushed through the column via a
HPLC pump model 2250 (BISCHOFF, Leonberg, Germany), all fittings as well as the pump head were
made in PEEK . A FIAS 400 (Perkin-Elmer) peristaltic pump provided the filling of a 25 µl PEEK sample
loop. The sample loop was connected to an automated 6-port valve (LabPRO 6, Rheodyne, California,
USA). The measurement system, including sample uptake, filling of the loop, injection on the column
and ICP-MS measurement is computerized. The integration of the chromatograms was performed
with Turbochrom (Perkin-Elmer). Before analysis the instrument was preconditioned by aspirating
the eluent for at least 1 hour.
Hexavalent chromium was also spectrophotometrically determined after ion chromatography by

post column derivatisation with 1,5-diphenylcarbazide [23].
Picture 2: Ion chromatography instrument equipped with post-column derivatisation.
Although direct spectrophotometric analysis, without ion chromatography, is often applied in water
analysis, several problems have been documented when this method is applied to some types of
samples with complex matrices [4]. Analysis of bulk samples using this methodology suffers from
potential interferences by other coloured species or species forming coloured reaction products with
DPC, such as V, Mo and Hg. Low results and poor recoveries were observed when samples contained
substances that reduce Cr6+ in acidic solution [24]. To avoid the aforementioned interferences the
speciation of the redox species was carried out by anion exchange chromatography on an IONPAC AS
7 column (Dionex, Sunnyvale, CA, USA). The column was provided with a guard column to remove
dissolved and particulate contaminants from the sample and eluent before reaching the column. The
eluent was pushed through the column via an isocratic pump P1000 (Spectro-Physics analytical, San
Jose, California). The post column reagent was added via an isocratic pump P1500 (Spectro-Physics
analytical, San Jose, California). The 20 µl PEEK sample loop was manually filled. Post column
derivatisation of the Cr6+ with diphenylcarbazide was followed by detection of the coloured complex
at 540 nm with a UV-150 detector spectrophotometer (Spectro-Physics analytical).
Some of the alkaline digestions were performed with ultrasound-assisted leaching. The ultrasonic
radiation was applied by means of a Branson 5210 (47 kHz, 185 W, Soest, the Netherlands).
Milli-Q water purification system (Millipore, Milford, MA, USA) was used. A stock solution of 1000
mg l-1 Cr6+ was prepared by dissolving 2.828 g potassium dichromate (pro analysi, Merck, Darmstadt,
Germany) in 1 l. A Stock solution of 1000 mg l-1 Cr3+ was purchased (Merck). A quality control stock
solution of 1000 mg l-1 CrO4- was purchased (Merck).
For the determination of the chemical purity of the enriched chromium isotope solutions on the AS
11 column, 0.14 M NH4NO3 (Merck) was used as eluent (flow rate 0.4 ml min-1). This eluent solution
was acidified to pH 3 with concentrated subboiled nitric acid (Merck). For the determination of Cr6+
in the extracted solid samples on the AS 11 column, 250 mM NaOH (Merck) was used as eluent (flow
rate 0.35 ml min-1). To control the stability of the ICP-MS the eluent was spiked with an internal
standard of 10 ng ml-1 Cesium. This signal, measured on 133Cs, was continuously monitored during
the measurement of the chromatogram. For the determination of Cr6+ in the extracted solid samples
on the AS 7 column, 250 mM (NH4)2SO4 - 49 mM NH4OH (Merck) was used as eluent (flow rate 1 ml
min-1).
Calibration standards were diluted from the stock solutions. The concentration level of the standards
was 50, 150 and 250 ng ml-1 Cr3+ and Cr6+. The standards were prepared in the appropriate eluent
concentration. For the determination of the dead time of the detector a 200 ng ml-1 Cr6+ solution
prepared in 0.14 M NH4NO3 was measured. For the determination of the mass bias a 50 ng ml-1 Cr6+
solution prepared in the appropriate eluent was used.
53
Preparation of Cr6+ isotope enriched solution
Approximately 5 mg of a 53Cr isotope enriched Cr2O3 solid material (supplied by the Institute for
Reference Materials and Measurements (IRMM), Geel, Belgium) was weighed in a beaker. 5 ml of
HClO4 (Merck) was added and the solution was heated during 3 hours until complete digestion. After
cooling to room temperature the solution was diluted to 500 ml with Milli-Q water in a volumetric
flask. The concentration of the stock solution amounted approximately 10 mg l-1. The relative
abundances of the chromium isotopes in the enriched material are summarised in table 1.
53
isotope natural Cr Cr6+ spike 50
Cr3+ spike
50
Cr 4.345 0.04 93.725
52
Cr 83.789 2.63 6224
53
Cr 9.501 97.21 0.042
54
Cr 2.365 0.12 0.010
Table 1. Relative abundances (%) of isotopes of natural chromium [35],
53 6+ 50 3+
Cr enriched spike and Cr enriched spike.
50
Preparation of Cr3+ isotope enriched solution
Approximately 5 mg of a 50Cr isotope enriched Cr2O3 solid material (supplied by the Institute for
Reference Materials and Measurements (IRMM), Geel, Belgium) was weighed in a beaker. 5 ml of
HClO4 (Merck) was added and the solution was heated during 3 hours until complete digestion. After
cooling to room temperature, 2 ml H2O2 was added and the solution was heated again for a 30
minutes. The solution was cooled again and diluted to 500 ml with Milli-Q water in a volumetric flask.
The concentration of the stock solution amounted approximately 10 mg l-1. The relative abundances
of the chromium isotopes in the enriched material are summarised in table 1.
Preliminary alkaline digestions showed significant oxidation of the added Cr3+ spike. This could be
related to the presence of traces of H2O2 in the spike solution. Although in acid medium Cr6+ is
converted to Cr3+ in the presence of H2O2, in alkaline medium oxidation takes place of Cr3+ in the
presence of this chemical [25]. To decompose the traces of H2O2 in the spike solution without altering
the redox species of chromium, the solution was treated in a UV digestion apparatus for 8 hours (UV
digester 705, Metrohm, Herisau, Switzerland).
Alkaline digestion for hexavalent chromium (US EPA method 3060A)[13]
Only a brief summary is given of US EPA method 3060A in order to visualise all components that may
interfere in the latter ICP-MS determination. Approximately 2.5 g of a sample is placed into a 250 ml
digestion glass vessel and spiked with an aliquot of enriched isotope 50Cr3+ and enriched isotope
53 6+
Cr , finally 50 ml of the digestion solution is added. The digestion solution consists of a 0.5 M
NaOH and a 0.28 M Na2CO3 solution. Also approximately 61 mg Mg(NO3)2 is added as well as 0.5 ml
of a 1.0 M K2HPO4 / KH2PO4 buffer (pH 7). The addition of Mg2+ in a phosphate buffer to the alkaline
digestion solution has been showed to suppress the oxidation of Cr3+. In US EPA method 3060A an
addition of 400 mg MgCl2 is prescribed, but due to the ClO interference on chromium isotopes and
the matrix effects observed in the ICP-MS analysis due to the high amount of salts, the added
amount has been changed to 60 mg Mg(NO3)2.
The sample is covered witch watch glass and heated at 363-368 K for at least 60 minutes with
continuous stirring. The contents is quantitatively transferred, after gradually cooling to room
temperature, to the filtration apparatus (polytetrafluoroethylene filter). Special attention must be
paid to the filter media. Reduction of Cr6+ to Cr3+ has been reported on sampling filter media such as
mixed cellulose ester filters. Polyvinyl chloride (PVC), polyvinyl fluoride (PVF) or
polytetrafluoroethylene (PTFE) filter media may be used and it is recommended to treat filters with
base prior to filtration [26]. The filtrate and rinses of the digestion vessel are collected in a 100 ml
volumetric flask and diluted with Milli-Q water. The filtrate was injected on the AS 10 column
without further treatment for the ICP-MS analysis. For the spectrophotometric analysis, an
appropriate amount of the eluent ((NH4)2SO4 , NH4OH) was added to the filtrate.
The digestion can also be performed with ultrasound-assisted leaching. Ultrasonic radiation has
shown to be a powerful aid in the acceleration of extracting a number of analytes of solid samples.
The ultrasound-assisted digestion of hexavalent chromium has been successfully applied to soils and
industrial hygiene samples [27-29].
A 30 day holding time for Cr6+ digestion from solid materials and a 7 day holding time for Cr6+ analysis
once solubilised in the alkaline digestate has been scientifically designated as appropriate for Cr6+
analysis according to this procedure [30].
Cr6+ is easier to reduce in a pH 7 solution than in a pH 12 solution. Although the alkaline digestate
can be stored for future analysis, it is recommended to neutralise the solution prompt before
measurement, as Cr6+ has been observed to reduce during neutralisation. This reduction may be
kinetically slow, but extended storage time could cause significant loss of Cr6+ [4].
Water extraction for hexavalent chromium
For the water extraction, approximately 2.5 g of a sample was placed into a 250 ml digestion vessel
and 50 ml water was added. The sample was covered witch watch glass and heated at 363 -368 K for
at least 60 minutes with continuous stirring. The contents was quantitatively transferred, after
gradually cooling to room temperature, to the filtration apparatus (0.45 µm membrane filter). The
filtrate and rinses of the digestion vessel were collected in a 100 ml volumetric flask and diluted with
water. In case of spiking, the aliquots of enriched material were added on the sample before addition
of the extraction medium.
Calculations
The equations mentioned in the method of calculation paragraph have been solved with the
standard tools of the MATLAB software (version 5.2, Mathworks Inc., Natick, MA). Although the
regression analysis can easily be done by the standard tools of MATLAB, the statistical toolbox has
been used, in order to perform also some t-tests on the intercept.
Results and Discussion

Chemical purity of the enriched isotope material (Dionex AS 11)
The SIDMS equations assume chemical pure isotope enriched spikes of both redox species. A Dionex
AS 11 chromatographic column coupled to a guard column was optimised for the determination of
Cr3+ and Cr6+. The optimisation also included the separation of both chromium species from the
major interferences of concern (table 2).
isotope Isobaric Ar interference matrix interference double charged

50 50 36 14 34 16 100
Cr Ti (5.4) Ar N (0.34) S O (4.20) Mo2+ (9.63)
50 38 12 35 15
V (0.25) Ar C (0.06) Cl N (0.28)
37 13
Cl C (0.27)
52 40
Cr - Ar12C (98.5) 35 16
Cl OH (75.6) 104
Pd2+ (9.3)
36
Ar16O (0.34) 35 17
Cl O (0.03) 104
Ru2+ (18.3)
38 14 36 16
Ar N (0.06) S O (0.02)
53 40
Cr - Ar13C (1.1) 37 16
Cl O (24.2) 106
Cd2+ (1.2)
35 18 106
Cl O (0.15) Pd2+ (27.1)
36 16
S OH (0.02)
54 54 40
Cr Fe (5.80) Ar14N (99.24) 37 16
Cl OH (24.2) 108
Pd2+ (26.7)
38 16
Ar O (0.06)
Table 2. Major interferences on the chromium isotopes measured by ICP-MS (relative abundance).
In figure 1 a chromatogram of the possible interferences and both chromium species is shown. It can
be seen that Cr3+ is retained on the anion exchange column. Cr3+ has a strong tendency to form
hexacoordinated octahedral complexes with ligands such as water, ammonia, halides, sulfates and
organic acids [31]. A possible explanation for the retention of Cr3+ on this column is related to the
presence of the functional groups of the Dionex AG11 column, as proposed by Charles et al. [32].
They are alkanol quaternary ammonium and can interact with the water molecules of the
Cr (H 2 O) 36+ complex leading to increased retention.
-
SO 24−
Cl
Cr3+
CrO 24− CO 32−

Signal intensity
m/z 34
m/z 35
m/z 12
m/z 52
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Time (min)
Figure 1: Retention times of the different alkaline digestion reagents after optimisation of Dionex AS 11 column
-1
(eluent: 0.14 M NH4NO3 (pH 3); flow rate: 0.4 ml min ; sample loop 25 µl). Chromatograms of a mixture
-1 2- -1 - -1 2- -1 3+ -1
containing 20 mg l SO4 (m/z 34), 10 mg l Cl (m/z 35), 50 mg l CO3 (m/z 12), 100 µg l Cr and 100 µg l
6+
Cr .
To confirm the assumption of chemical pure spikes, both solutions of isotope enriched material were
analysed. The solutions were diluted in 0.14 M NH4NO3 and injected on the column (figure 2).
Cr6+
Cr3+
Cr3+ impurity
Signal intensity
m/z 53
m/z 50
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (min)
Figure 2: Chemical purity of isotope enriched materials analysed on Dionex AS 11 column. Upper
53 6+ 50 3+
chromatogram Cr spike, lower chromatogram Cr spike; eluent: 0.14 M NH4NO3 (pH 3); flow rate: 0.4 ml
-1
min ; sample loop 25 µl.
While no detectable 50Cr6+ was present in the 50Cr3+ isotope enriched spike solution, 2 % of the
isotope enriched 53Cr6+ spike was present as 53Cr3+. In order to correct for this impurity of 53Cr3+ in the
isotope enriched 53Cr6+ spike the equations were adjusted. This was done by introducing a chemical
purity factor (P) in the equations as already mentioned in the calculation paragraph. This purity
factor stands for the fraction Cr6+ in the 53Cr isotope enriched material.
Determination of hexavalent chromium in alkaline digestion (Dionex AS 10)
To improve the calculations for the determination of hexavalent chromium in the alkaline digestion
based on IC-ICP-MS measurements, the relative standard deviation on the mass bias was optimised
and the influence of spectral interferences investigated.
Determination of mass bias
The observed isotope ratios deviates from the standard values as a function of the difference in mass
between the two isotopes [33]. The true ratio of isotopes A and B , (A/B)t, can be related to the
measured ratio, (A/B)m, by:
A A
  =   ⋅ (1 + a ⋅ n )
 B  m  B  t
where a is the bias per mass unit, and n is the mass difference between isotopes A and B. Also,
determination of the dead time is critical when ratios of isotopes with different relative abundances
are calculated. The relation between the measurement time parameters, dwell time and number of
sweeps, of the ICP-MS on the standard deviation of the measured ratios was studied (figures 3 and
4).
1,5
Cr50/Cr52
RSD (%) on ratio
1 Cr53/Cr52
0,5
0
0 10 20 30 40 50
dwell time (ms)
50 52 53 52
Figure 3: Relative standard deviation on the ratios Cr/ Cr and Cr/ Cr in function of
-1
the dwell time (number of sweeps 300) measured on a 50 ng ml Cr solution.
1,5
RSD (%) on ratio
1
Cr50/Cr52
Cr53/Cr52
0,5
0
0 100 200 300 400
number of sweeps
50 52 53 52
Figure 4: Relative standard deviation on the ratios Cr/ Cr and Cr/ Cr in function of
-1
the number of sweeps (dwell time 30 ms) measured on a 50 ng ml Cr solution.
Under the given conditions a dwell time of 30 ms per element and 300 number of sweeps resulted in
a satisfactory compromise between total analysis time and standard deviation on the ratios. Under
the given optimised time conditions the mass bias was calculated before every measurement run on
a 0.25 M NaOH solution containing 50 µg l-1 Cr6+. Over a period of two months an average bias per
mass unit, a, of 0.036 ± 0.004 (n = 14) was calculated, i.e., � 50Cr� 52Cr�𝑚 = � 50Cr� 52Cr�𝑡 ∙ �1 + 0.036 ∙
(−2)� and � 53Cr� 52Cr� = � 53Cr� 52Cr� ∙ �1 + 0.036 ∙ (1)�. The optimised instrumental measurement
𝑚 𝑡
settings of the IC-ICP-MS are summarised in table 3.
parameter value
instrument settings
nebuliser gas flow 0.9 ± 0.05 l/min
RF power 1200 W
analog stage voltage -1950 V
lens voltage 5±1V
dwell time 30 ms
sweeps / reading 5
total measurement time 15 minutes
12 31 34 35 50 52 53 54 133
Isotopes monitored C, P, S, Cl, Cr, Cr, Cr, Cr and Cs
Table 3. Optimised instrumental measurement settings ICP-MS.
Analytical characteristics
As the concentration of NaOH in the alkaline digestion amounted 0.25 M, the optimisation of the Cr6+
speciation on the AS 10 chromatographic column was performed with 0.25 M NaOH as eluent.
Neutralising the alkaline digestion before measurement has been shown to give precipitation of lead
chromate at high concentrations. Moreover during neutralisation to a pH of 7.5, reduction of Cr6+
and detection of small amounts of Cr3+ have been observed on soil extracts spiked with Cr6+ solutions
[4]. The flow rate was optimised in order to separate the major interferences from the analyte of
concern. Special attention was paid to the interfernce arising from CO32-, essential in the digestion for
dissolution of BaCrO4 [4]. A chromatogram recorded under optimised conditions can be seen in figure
5.
CrO 24−
2−
Cl-
CO 3
Signal intensity
m/z 52
m/z 35
m/z 12
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5
Time (min)
-1 6+ -1
Figure 5: Alkaline digestion solution spiked with 100 ng ml Cr , eluent 0.25 M NaOH, flow rate 350 µl min ,
sample loop 25 µl.
The stability of the IC-ICP-MS system operating under these conditions (NaOH) has been
demonstrated in a previous paper [34]. A relative standard deviation of 2.8 % was calculated for 54
consecutive measurements of the internal standard over a period of 40 h. To correct for the
instability of the ICP-MS in this study, the eluent was spiked with an internal standard of 10 ng ml-1
Cesium. The signal of the internal standard, measured on 133Cs, was continuously monitored during
the measurement of the chromatogram. The integrated peak areas of the analytes of concern were
corrected for the measured value of the internal standard. The detection limit, calculated as three
times the standard deviation on 10 determinations of a 5 ng ml-1 Cr6+ standard solution, amounted
0.8 ng ml-1 Cr6+.
Alkaline digestion versus water extraction
In the legislation of some of the member states of the European Union the determination of
hexavalent chromium among others is required in leaching tests in order to meet regulatory criteria
for acceptance on landfill. In these leaching tests a waste samples is leached with water in a specific
liquid to solid ratio. In order to study the influence of the extraction medium, 3 waste samples were
extracted in duplo in water and alkaline medium. The samples can be characterised as filter cake,
chromium dust and low carbon pitslag. In each medium the samples were extracted as such and
spiked with an aliquot of enriched isotope 50Cr3+ and enriched isotope 53Cr6+. The digestion solutions
were analysed with ion chromatography diphenylcarbazide (IC-DPC) and ion chromatography
inductively coupled plasma-mass spectrometry (IC-ICP-MS).
As can be seen from the differences in the duplo analysis (results reported in table 4 and 5), the
alkaline digestion is – as expected - more reproducible than the water extraction. Moreover the
extraction efficiency for Cr6+ is higher in the alkaline digestion in comparison with the water
extraction.
The added Cr6+ spike is stable during the alkaline digestion (recovery in all samples > 90%). The
recovery of the Cr6+ spike in the water extraction however varied between 46 to 90 % and also
oxidation of the added Cr3+ spike was observed, indicating species interconversion during the water
extraction. Calculation of the recovery of Cr6+ in case of measurement with IC-DPC is sometimes
misleading as the value is calculated on the difference of two measurements performed on two
different extractions. While in the case of measurement with species-specific isotope dilution mass
spectrometry the calculation is performed on the same extraction. The oxidation of the added Cr3+
spike is negligible during the alkaline digestion (the observed oxidation may be due to traces of H2O2
present in the spike solution).
IC-DPC IC-ICP-MS
identification [Cr ]
VI
total
R [Cr ] [Cr ] [Cr
VI VI
total
VI
natural ] [Cr ]
VI
spike
R [Cr ]
III
spike
Ox. [Cr ]
VI
-1 -1 -1 -1 -1 -1 -1
ng ml % mg kg ng ml ng ml ng ml % ng ml % mg kg
sample1 259 74 260 74
sample1 541 176 110 561 416 138 86 6.1 3.8 120
+spikea
sample1 281 79 269 76
sample1 372 57 60 378 246 127 79 4.6 2.9 69
+spikea
sample2 2155 433 1838 369
sample2+ 3856 532 701 3238 2985 246 77 7 2.2 592
spikeb
sample2 1984 390 2218 436
sample2+ 3915 603 706 3560 3290 266 83 4 1.3 646
spikeb
sample3 240 48 230 46
sample3+ 379 87 43 365 266 74 46 25.2 15.8 52
spikea
sample3 240 47 212 42
sample3+ 391 94 45 384 237 144 90 2.6 1.6 46
spikea
Table 4. Water extraction of three different waste samples (sample 1 (filter cake), sample 2 (chromium dust)
and sample 3 (low carbon pitslag).
a -1 50 3+ -1 53 6+
added amount of spike 160 ng ml Cr + 160 ng ml Cr ;
b -1 50 3+ -1 53 6+
added amount of spike 320 ng ml Cr + 320 ng ml Cr .
IC-DPC IC-ICP-MS
identificatio
n
[Cr ] VI
total
R [Cr ] [Cr ] [Cr
VI VI
total
VI
natural ] [Cr ]
VI
spike
R [Cr ]
III
spike
Ox. [Cr ]
VI
-1 -1 -1 -1 -1 -1 -1
sample1 2334 659 1964 554
sample1 2469 84 649 2241 2078 157 98 3 1.9 584
+spikea
sample1 2242 626 2023 565
sample1 2100 -89 542 2070 1919 151 94 2.5 1.6 536
+spikea
sample2 5056 1024
sample2+ 5530 14 1019
spikeb
8
sample2 5007 993 4745 941
sample2+ 5288 88 973 4876 4570 298 93 8 2.5 895
spikeb
sample3 421 82 394 77
sample3+ 603 11 85 621 450 168 105 3 1.9 87
spikea
4
sample3 411 81 370 72
sample3+ 586 10 83 543 384 157 98 2.5 1.6 75
spikea
9
Table 5. Alkaline extraction of three different waste samples (sample 1 (filter cake), sample 2 (chromium dust)
and sample 3 (low carbon pitslag).
a -1 50 3+ -1 53 6+
added amount of spike 160 ng ml Cr + 160 ng ml Cr ;
b -1 50 3+ -1 53 6+
added amount of spike 320 ng ml Cr + 320 ng ml Cr .
Key to the used symbols in table 4 and 5:
IC-DPC
• [
VI
Crtotal ]
: the amount of Cr6+ present in the extraction solution, ng ml-1
• R, recovery : the theoretical recovery assuming that added spike (e.g. 160 ng ml-1 Cr6+) is
completely recovered. The recovery is calculated as the difference between the total amount
of Cr6+ present in the spiked and unspiked (as such) extraction solution
• [Cr ]: the amount of Cr
VI 6+
extracted from the sample, mg kg-1
IC-ICP-MS
• [ ]
VI
Crtotal : the amount of Cr6+ present in the extraction solution, ng ml-1
• [Cr ] : the amount of extracted natural Cr
VI
natural
6+
present in the extraction solution, ng ml-1
• [Cr ] : the amount of enriched Cr
VI
spike
6+
spike present in the extraction solution, ng ml-1
• R, recovery : The recovery is calculated as the difference between the amount of Cr6+ spike
[ VI
]
present in the extraction solution Crspike and the theoretical added amount of spike (e.g.
-1
160 ng ml )
• [Cr ] : the amount of enriched Cr
III
spike
3+
spike converted to Cr6+, ng ml-1
• Ox., oxidation : The oxidation is calculated as the difference between the amount of
[ ]
converted Cr3+ spike ( Crspike ) and the theoretical added amount of spike (e.g., 160 ng ml-1);
III
• [Cr ]: the amount of Cr

VI 6+
extracted from the sample, mg kg-1
Alkaline digestion on packaging materials
In December 1994 the EEC Council of Ministers passed a directive on Packaging and Packaging waste
94/62/EC. In Article 11 of this directive concentration levels of heavy metals in packaging have been
assigned. Member states shall ensure that the sum of the concentration levels of lead, cadmium,
mercury and hexavalent chromium present in packaging or packaging components shall not exceed
100 mg/kg by weight. In order to overview the most critical types of packaging on the Belgian
market, a qualitative and quantitative market analysis was performed by our institute. In the frame
of this study, some 200 packaging materials were screened with X-Ray Fluorescence for total metal
content. A selection of 40 critical packaging materials (total chromium content higher than 100
mg/kg) were further analysed for Cr6+ content. The materials analysed were glass packaging, plastic
packaging, wrapping papers and metal packaging.
Picture 3: Examples of household packaging: glass packaging, (beer) crates, steel packaging, pieces of plastic
bags and homogenized plastic nets.
For the determination of Cr6+ an alkaline digestion, described as above, was performed. Cr6+ was
analysed in the digestion solutions with IC-DPC and IC-ICP-MS. Cr6+ can be found in chromate-based
paints. Chromate-based paints primers are applied, as a first-coat primer, for corrosion inhibition.
These paint primers generally consist of a paint matrix containing a sparingly soluble chromate salt
such as strontium chromate. The chromate based paints contain an epoxy polymer, a chromate salt,
a combination of organic solvents, dispersion agents and other fillers. To improve extraction
efficiency from paint samples, the use of 1,4 dioxane in a second successive alkaline digestion was
reported. Empirical observation indicated that 1,4-dioxane broke up the paint samples into strands
and filaments. This increased the surface area of the paint available to the digestion solution and
enhanced the extraction of Cr from the paint matrix. However only a slight improvement using 1,4-
dioxane was reported [26]. For this reason no modifications were applied in this study to the alkaline
digestion method.
A good correspondence was found between the total Cr6+ content analysed with IC-DPC and IC-ICP-
MS. The duplo analysis performed on each sample (spiked and unspiked) were in good agreement,
indicating that the alkaline digestion was a reproducible method for extracting Cr6+ from packaging
material. It has to be stressed that this method determines the alkaline soluble Cr6+, which is based
on literature the most complete but may differ from the total Cr6+ content. However it is the author’s
opinion that at this moment this digestion is state of the art from point of view of minimal species
interconversion on the one hand and maximum amount Cr6+ extracted on the other. Some results of
this study are shown in table 6.
IC-DPC IC-ICP-MS
Identif. [Cr ]
VI
total
R
[Cr ] [Cr ] [Cr
VI VI
total
VI
natural ] [Cr ]
VI
spike
R
[Cr ]
III
spike
Ox.
[Cr ]
VI
-1 -1 -1 -1 -1 -1 -1
Orange net 1723 71 1579 65
Orange net +
spikea
1739 16 1699 1600 94 94 4.5 4.5 63
Orange net 71 2.9 62 2.6
Orange net +
spikea
160 89 169 66 101 101 2.0 2.0 2.7
Yellow net 190 8.5 201 9.0
Yellow net +
spikea
318 128 321 216 102 102 1.8 1.8 9.0
Yellow
shopping bag
541 22 495 20
Yellow
shopping bag + 662 121 666 552 108 108 5.2 5.2 22
spikea
Yellow
shopping bag
321 13 342 14
Yellow
shopping bag + 438 117 405 308 93 93 4.6 4.6 12
spikea
Yellow drum 16660 660 16400 650
Yellow drum + b
spikea
15520 1134 16400 16240 116 116 3.1 3.1 640
Yellow drum 8 0.31 6.6 0.3
Yellow drum +
spikea
100 92 104 4 99 99 <1 <1 0.2
Red drum 99 3.8 102 3.9
Red drum +
spikea
202 103 212 110 101 101 1.3 1.3 4.2
a
Table 6. Alkaline extraction on some packaging materials (the key to the used symbols is given in table 4-5).
-1 50 3+ -1 53 6+ b
added amount of spike 100 ng ml Cr + 100 ng ml Cr . this sample was, due to the high concentration
6+ 6+
of natural Cr , analysed in 100 fold diluted. The actual measured enriched spike Cr in this solution amounted
-1
1.16 ng ml . The recovery of 116 % is based on this measurement but has to be seen as an indicative value as
-1
the detection limit is around 1 ng ml .
Of all analysed samples two packaging materials contained concentrations higher than 100 mg /kg. A
yellow metal drum used for industrial packaging contained 650 mg/kg Cr6+ and a gold coloured
wrapping paper contained 210 mg/kg Cr6+. The recovery of the enriched spike Cr6+ measured with IC-
ICP-MS varied between 90 and 110% indicating that during the digestion all Cr6+ compounds were
stable. Except for some analysed steel cans (containing fruit, vegetables, sardine) the enriched spike
Cr6+ was not recovered. The oxidation of the added enriched spike Cr3+ was negligible in all
digestions. The approximately 2 % oxidation occurred as well in the blank solution were enriched Cr3+
was added indicating that this problem was likely to originate from traces of H2O2 present in the
enriched Cr3+ spike solution.
From this study it could be concluded that critical packaging are predominantly coloured materials.
Presence of heavy metals in the sample was not attributed to the bulk sample but in most cases to
the applied colouring agents and pigments (red and yellow pigments, e.g., lead chromate). In table 7
and 8, for each type of respectively household and industrial packaging a summary is made of the
total flow on the Belgian market, the flow of the critical packaging and the heavy metal(s) causing an
exceeding of the limits.
type of packaging critical critical packaging flow (% of 100 mg/kg limit

packaging flow packaging total packaging flow) exceeded due to:
decorated appr. 0.05% Cd, Pb, possibly
glass Cr (VI)
transparent appr. 20% Pb
glass
glass 44.40% brown glass Pb
green glass 24.4% Pb
blue glass Pb
paper/cardboard 18.60% / / /
steel 10.90% / / /
aluminium 1.60% / / /
plastic bottles 8.30% / / /
shopping bags small Pb, Cr (VI)
plastic 10.80% nets small Pb, Cr (VI)
crates appr. 1.2% Pb, Cd
wood Small / / /
beverage carton 2.70% / / /
others:
wrapping foil Small Cr (VI), Pb
others 2.80 % / / /
Table 7: Flows and problem areas per type of household packaging on the Belgian market.
type of packaging critical critical packaging flow 100 mg/kg limit

packaging flow packaging (min-max %) exceeded due to:
(min-max)
paper/cardboard 24-49 % / / /
metal 8-10 % drums almost all metal packaging Pb, Cr (VI)
are drums
plastic 13-21 % pallets unknown possibly Pb
wood 28-43 % / / /
others 2-2.4 % brown glass unknown Pb
Table 8: Flows and problem areas per type of industrial packaging on the Belgian market.
Conclusions
This study has proven that, regarding some of the legal concerns related to Cr6+ in the environment,
the alkaline digestion according to US EPA 3060A followed by ion chromatography - 1,5-
diphenylcarbazide spectrophotometry is suited for routine analysis of hexavalent chromium in solid
matrices. The ability of the species-specific isotope dilution mass spectrometry to calculate the
recovery of the enriched hexavalent chromium spike on the one hand and the oxidation of the Cr III
spike on the other in the digestion solution was a helpful tool to assess the species interconversion
and consequently the validity of the digestion.
This method determines the alkaline soluble Cr6+, It can be stated however that based on literature
this method is one of the most rigorous digestions but still can differ from the total Cr6+ content. The
accuracy of the determination of hexavalent chromium in solid matrices remains a challenging field
of effort and improvement from point of view of minimal species interconversion on the one hand
and maximum amount of Cr6+ extracted on the other.
Acknowledgements
We wish to thank Michael Berglund and Philippe Taylor (IRMM, Geel, Belgium)) for the supply of isotope
enriched chromium, Eduard Temmerman (University of Gent, Belgium) for information regarding the kinetics
involved in the chromium redox species, Wendy Verelst for experimental assistance.
References paper, Determination of hexavalent chromium by species-specific isotope dilution

mass spectrometry and ion chromatography – 1,5 – diphenylcarbazide spectrophotometry, J. Anal.
At. Spectrom., 2003, 18, 922-932.
1. R.J. Vitale, G.R. Mussoline, J.C. Petura, B.R. James, J. Environ. Qual. 23 (1994) 1249.
2. M.J. Marques, A. Salvador, A.E. Morales-Rubio, M de la Guardia, Freasenius J. Anal. Chem. 362
(1998) 239.
3. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, Centre Belge d’Etude de la
Corrosion CEBELCOR, Brussels, 1974.
4. D. Huo, L. Yusheng, H.M. Kingston, Environ. Sci. Technol. 32 (1998) 3418.

5. R. Nusko, K.G. Heumann, Anal. Chim. Acta 286 (1994) 283.
6. J.G. van Raaphorst, H.M. Haremaker, P.A. Deurloo, B. Beemsterboer, Anal. Chim. Acta 286 (1994)
291.
7. M. Derbyshire, A. Lamberty, P.H.E. Gardiner, Anal. Chem. 71 (1999) 4203.
8. C. Whalley, A. Hursthouse, S. Rowlatt, P. Iqbal-Zahid, H. Vaughan, R. Durant, Water, Air and soil
pollution 112 (1999) 389.
9. C. Barnowski, N. Jakubowski, D. Stuewer, J.A.C. Broekaert, J. Anal. At. Spectrom. 12 (1997) 1155.
10. H.M. Kingston, D. Hou, Y. Lu, S. Chalk, Spectrochim. Acta, Part B 53 (1998) 299.
11. A.C. Harzdorf, Intern. J. Environ. Anal. Chem. 29 (1987) 249.
12. N. Violante, F. Petrucci, P. D. Femmine, S. Caroli, Microchemical Journal 59 (1998) 269.
13. SW 846 EPA, Method 3060A: Alkaline digestion for hexavalent chromium, Test Methods for
evaluating Solid waste Physical/Chemical Methods, 1996.
14. SW 846 EPA, Method 7196A: Chromium, Hexavalent (colorimetric), Test Methods for evaluating
Solid waste Physical/Chemical Methods, 1992.
15. N. Kozuh, J. Stupar, B. Gorenc, Environ. Sci. Technol. 34 (2000) 112.
16. M. Pantsar-Kallio, S.P. Reinikainen, M. Oksanen, Anal. Chim. Acta 439 (2001) 9.
17. H.M. Kingston, U.S. Patent 5,414,259, 1995.
18. H. Vanhoe, J. Goossens, L. Moens, R. Dams, J. Anal. At. Spectrom. 9 (1994) 177.
19. C. Kim, Q. Zhou, B. Deng, E.C. Thornton, H. Xu, Environ. Sci. Technol. 35 (2001) 2219.
20. A.G. Coedo, T Dorado, I. Padilla, F.J. Alguacil, J. Anal. At. Spectrom. 15 (2000) 1564.
21. B.R. James, J.C. Petura, R.J. Vitale, G.R. Mussoline, Environ. Sci. Technol. 29 (1995) 2377.
22. SW 846 EPA, Method 6800: Elemental and speciated isotope dilution mass spectrometry, Test
Methods for evaluating Solid waste Physical/Chemical Methods, 1998.
23. SW 846 EPA, Method 7199: Determination of hexavalent chromium in drinking water,
groundwater and industrial wastewater effluents by ion chromatography, Test Methods for
evaluating Solid waste Physical/Chemical Methods, 1996.
24. R. Roehl, M. Alforque, Atomic spectroscopy 11 (1990) 210.
25. M. Pettine, L. Campanella, F. J. Millero, Environ. Sci. Technol. 36 (2002) 901.

26. R.A. Sabty-Daily, K.K. Luk, J.R. Froines, Analyst 127 (2002) 852.
27. J.L. Luque-Garcia, M.D. Luque de Castro, Analyst 127 (2002) 1115.
28. J.M. Boiano, M.E. Wallace, W.K. Sieber, J.H. Groff, J. Wang, K. Ashley, J. Environ. Monit. 2 (2000)
329.
29. J.Wang, K. Ashley, E.R. Keneedy, C. Neumeister, Analyst 122 (1997) 1307.
30. R.J. Vitale, G.R. Mussoline, K.A. Rinehimer, K.L. Moeser, Soil and Sediment contamination 9
(2000) 247.
31. R.F.M.J. Cleven, J.L.M de Boer, A. van der Meulen, Characterization of the Cr (VI) / Cr (III) ratio in
aerosols, Chapter 4, Environmental analytical and physical chemsitry series, Vol. 1,
Environmental particles, Lewis Publishers, Chelsea, 1992.
32. S. Charles, C. Pécheyran, M. Gautier, O.F.X. Donard, poster presented at the 2000 Winter
Conference on Plasma Spectrochemistry, Fort Lauderdale, FL, USA, 2000.
33. G.P. Russ, J.M. Bazan, Spectrochim. Acta 42B (1987) 49.
34. K. Tirez, W. Brusten, S. Van Roy, N. De Brucker, L. Diels, J. Anal. At. Spectrom. 15 (2000) 1087.
35. K.J.R. Rosman, P.D.P. Taylor, Pure & Appl. Chem. 70 (1998) 217.
3.3.2. THE BAN ON HEXAVALENT CHROMIUM IN ELECTRONIC WASTE (ROHS AND WEEE)
Waste of electrical and electronic equipment (WEEE or e-waste) and

the Restriction of Hazardous Substances (RoHS Compliance) has
become an intensely debated global issue [128]. Electronic and
electrical products, such as refrigerators, washing machines, mobile
phones, personal computers, printers, and television sets, are
ubiquitous in the modern society. At the same time, continuing
technological innovation has resulted in early obsolescence of many
electronic/electrical products. The average lifespan (2 yr) of a new
computer, e.g., was in 2005 less than half of that (4.5 yr) in 2000, and
has been continually declining. A combination of increasing ownership
and shortened lifespan has led to rapid growth in the amounts of unwanted and obsolete electronics
(commonly known as e-waste). It was estimated that the global rate of e-waste generation globally is
approximately 40 million tons yr-1.
Numerous studies have revealed that abundant toxic compounds, including, but not limited to, heavy
metals (Cr(VI), Hg, Cd, Pb), polychlororinated biphenyls (PCBs), polybrominated diphenylethers
(PBDEs), and polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs) could be leached into
the ambient environment during e-waste disposal according to primitive methods, posing serious
risks to environmental and human health. Well-managed sanitary landfills have been demonstrated
to be the most efficient option for preventing these toxic compounds from leaching. The relatively
high cost for e-waste disposal in developed countries has driven recycling operations to developing
countries such as China, India, and Pakistan (Figure 32 , left). It is estimated that nearly 80% of all e-
waste generated in developed countries is exported elsewhere.
Figure 32: Map (left) of the main routes of transboundary transport of e-waste and (right) e-waste processing
regions in China (e = e-waste recycling site, courtesy of [128]).
Presently, approximately 70% of e-waste generated worldwide is processed in China every year. The
Guangdong Province of South China and the Zhejiang Province of East China (Figure 32, right) are the
two regions with the most intensive e-waste recycling activity, mainly due to their convenient
locations for import. Primitive recycling operations can release large amounts of toxic compounds,
subjecting local workers to health hazards. Significantly higher levels of chromium, of 1160 ng m-3 in
airborne particles, of 137−477 μg g−1in soil from an e-waste recycling site and of 307 μg g-1 in surface
sediment from the Nanguan River, draining through an e-waste recycling area of Taizhou, were found
(for comparison, see also ambient concentrations in Table 8, p.66)[128].
Human health is seriously threatened by e-waste recycling via direct exposure routes such as
inhalation and dust ingestion. Concern about the high toxicity of Cr(VI) resulted in legislative
demands by the European Union to restrict its usage and to reduce the environmental impact of
disposed (e)-wastes [129].
Since July 2006, electrical and electronic equipment (EEE) brought onto the European Market, has to
be RoHS (Restriction of Hazardous Substances) compliant. This compliancy implies the restriction of
six hazardous substances, i.e., lead, mercury, cadmium, hexavalent chromium, polybrominated
biphenyls (PBB) and polybrominated diphenyl ether (PBDE) in EEE.
To support the Belgian Inspectorate with the enforcement of this legislation, an inspection campaign
was developed by VITO as an exercise to gain experience in the different aspects of RoHS inspection
[133]. A total of 88 electrical and electronic devices were (non-destructively) screened on-site by
VITO and Federal inspectors in different electronic shops with EDXRF. In total 27 products were
purchased at 5 different locations and dismantled in the laboratory for further testing. For detecting
the presence of Cr(VI) in chromate coatings (on screws), a spot-test screening procedure using cotton
sticks impregnated with diphenylcarbazide was used. Based on independent visual ratings, a lower
limit of detection of 20-50 ng Cr (VI) by point sampling of the screening method was estimated. The
spot test represented a good and specific method for screening the presence of Cr(VI) in chromate
coatings.
Picture 4: RoHS inspection campaign, electrical and electronic devices were (non-destructively) screened on-
site in different electronic shops with EDXRF; 27 products were purchased and dismantled for further testing.
From the year 2000 onwards, VITO was commissioned by OVAM to participate in the European
Committee for Standardization - technical committee 292 - Characterization of waste (CEN TC 292,
see also § 1.3) in support of the implementation of the EU landfill Directive. Within this workgroup, a
European document describing the state-of-the-art extraction and determination methods for the
total content of hexavalent chromium in raw waste and other solid materials was prepared (CEN/TR
14589:2003). Thereafter, and in accordance with the recommendations of this technical report, VITO
in coordination with the Federal Institute for Materials Research and Testing (BAM, Germany)
organised an inter-laboratory comparison in 2005-2006 for the validation of a European standard (EN
15192) for the determination of hexavalent chromium in solid material.
Despite the good analytical performance of the wet-chemistry analytical methods, their use for
determination of Cr(VI) in solid materials always requires an extraction step, which may induce
modification of the speciation. Even if the precision of the quantification step is an important
parameter, a complete or quantitative extraction step still is the main challenge to obtain accurate
results.
Different extracting solutions have been used for the extraction of Cr(VI) from solid materials. Based
on a review study of digestion methods from soil published in 1995, the method yielding the best
results relied on a combination of Na2CO3 and NaOH with continuous swirling and heating at 95°C
[134]. This approach was promoted in 1996 by the US EPA with the reference US EPA Method 3060A
and was adopted in 2006 as the basis of the European standard EN 15192:2006 as well (see paper:
Validation of an European standard for the determination of hexavalent chromium in solid material,
J. Environ. Monit., 2007). There is at this time a consensus on the use of an alkaline digestion solution
to achieve minimal species interconversion and maximum extraction efficiency of hexavalent
chromium from solid materials. However and despite the availability of these standard methods, it
remains difficult to determine whether observed differences between extraction methods are due to
interconversion of chromium species in the extracts or to a low extraction yield or to both.
Recently and using X-ray Absorption Near Edge Structure (XANES), it was reported that a
combination of Na2CO3 and NaOH does not completely solubilize all forms of hexavalent chromium,
depending on the matrix of the soil samples [135]. On the same soil sample that was used in the
inter-laboratory comparison for the validation of the European standard EN 15192 (see Table 1a, soil
2, p. 115), only 57% of the total Cr(VI) content measured by XANES was extracted. This work also
showed that the use of spiked soils for extraction studies is not an ideal approach for studying
alkaline extraction methods as this does not necessarily reflect the complexities of real soils. It is
conceivable that XANES spectroscopy can serve as a benchmark for improvements or modifications
to alkaline extraction methods, which might consist of sequential extraction approaches or different
extraction conditions, tailored to the type of matrix being studied. This future work would be of
particular importance in the case of environmental studies dealing with the toxicity of Cr(VI), which
may concern easily extractible Cr(VI) but also more insoluble phases.
The conclusion of the European validation study: “The accuracy of the determination of Cr(VI) in solid
matrices remains a challenging field in terms of maximum extraction efficiency and minimum species
interconversion” was confirmed but the problem was not resolved.
Validation of an European standard for the determination of hexavalent

chromium in solid material
Kristof Tirez,a1 Holger Scharf,b1 Domenico Calzolari,c1 Rob Cleven,d1 Monika Kissere1 and Detlef Lückb2
a
b
Federal Institute for Materials Research and Testing (BAM) Germany
c
FENICE S.p.A., Italy
d
National Institute for Public Health and the Environment (RIVM), The Netherlands
e
Umwelt- und Abfall-Beratung, Austria
1
European Committee for Standardisation, Technical Committee 292 (Characterisation of waste),
Working group 3 ( Parameters related to species soluble in mineral acid or water - Analysis and
digestion methods)
2
International Organisation for Standardisation, Technical Committee 190 (Soil quality),
Subcommittee 3 (Chemical methods and soil characteristics), Working group 1
Journal of Envrionmental Monitoring, 2007, 9, 749-759.

Abstract
A European standard for the determination of Cr(VI) in solid material has been elaborated in the
framework of an international co-operation and finally validated in the course of an interlaboratory
comparison. The procedure is based on the alkaline digestion prescribed by US EPA method 3060A
followed by ion chromatography and determines an operationally defined content of Cr(VI), including
water-soluble and insoluble chromates. A preliminary robustness study was carried out in order to
compare different extraction methodologies and to study the equivalency of different analytical
methods for the determination of Cr(VI) in alkaline extracts of soil and waste materials. During an
interlaboratory validation trial with 19 European laboratories a set of 4 samples (2 soil and 2 waste
samples) was analysed to determine performance characteristics for different combinations of
digestion and detection methods. With the procedures prescribed by the new European standard (EN
15192) acceptable results were obtained for both soil samples and one of the waste samples (sludge).
However, for the second waste sample (fly ash) a large deviation in analytical results was observed.
This indicates that particularly for waste materials a possible occurrence of strong matrix effects has
to be considered and supplementary quality control data are needed in order to assess the validity of
analytical results. The accuracy of the determination of Cr(VI) in solid matrices remains a challenging
field in terms of maximum extraction efficiency and minimum species interconversion.
Introduction
The two oxidation states of chromium present in the environment, i.e., Cr(III) and Cr(VI), are different
in physico-chemical properties. Cr(III) is considered to be a trace element species essential for the
proper functioning of living organisms. Hexavalent chromium (Cr(VI)) compounds are known to be
toxic and carcinogenic agents for a variety of organisms. Toxicity, mobility and bioavailability of
chromium strongly depend on its chemical form. Most Cr(VI) compounds are usually highly soluble,
mobile and bioavailable compared to sparingly soluble Cr(III) species. Because of the ecological
relevance of chromium species in the environment, a number of reviews of analytical methods for
the determination of Cr(VI) have been published.1-5
To limit the release of Cr(VI) into the environment, the European Commission has set-up restrictions
in several directives. These European directives have triggered studies and new analytical protocols
for Cr(VI) determination and Cr speciation in solid matrices: packaging material,15 cement,16 materials
in the automotive industry,17 steelmaking solid waste,18 fly ash,19 paint samples20 and airborne Cr(VI)
compounds.21,22
Within the framework of the Landfill Directive (2003/33/EC) knowledge of the composition of the
waste is the first step in the acceptance procedure for a safe disposal.6 In accordance with the criteria
for acceptance of waste at landfills some European member states have implemented leaching limit
values for Cr(VI). For this reason European standard EN 12506 was developed for the analysis of
Cr(VI) in water eluates from solid waste material.7
Enforcement of Packaging Directive 62/94 has required the development of reliable reference
methods for the determination of heavy metals in glass.8 The International Commission on Glass
developed in this framework a recommended procedure for the determination of Cr(VI) detectable
down to 2 mg Cr(VI)/kg of glass.9
The European Directive 2003/53/EC restricts the use of cement and cement products, that contain,
when hydrated, more than 2 mg/kg of soluble Cr(VI).10 CEN/TC51 has recently proposed a method for
the determination of the water soluble Cr(VI) content of cement.11 Also within the framework of
2002/95/EC on the restriction of the use of certain hazardous substances in electrical and electronic
equipment a method for the determination of Cr(VI) is under development within the International
Electrotechnical Commission.12,13
Several European member states have issued regulatory or reference values for Cr(VI) in soil for risk
assessment purposes. In soil for agricultural use a maximum recommended concentration of 6.6 mg
Cr(VI)/kg dry substance has been proposed in the Flemish Region of Belgium.23 For the most sensitive
type of land use the Swedish guideline value suggest a maximum concentration of 5 and 120 mg/kg
dry substance for Cr(VI) and Cr(III) respectively.24 In Italy the highest permissible Cr(VI)
concentrations in soil are 2 and 15 mg/kg, depending on whether they are exploited for parkland or
industrial uses.25
However for the determination of Cr(VI) in contaminated soil and solid waste material no European
standard is available. For this reason, a state of the art document regarding Cr(VI) speciation was
prepared by Working Group 3 of Technical Committee 292 (Characterisation of waste) of the
European Committee for Standardisation (CEN/TC292/WG3).26 In the document an overview of Cr(VI)
speciation in solid materials has been given. As a result of the survey, it was decided by
CEN/TC292/WG3 in co-ordination with Subcommittee 3 of Technical Committee 190 of the
International Organisation for Standardisation (ISO/TC190/SC3) to propose a common standard for
the determination of hexavalent chromium in solid material based on existing standards.
Picture 5: In- and outdoor meetings of CEN TC 292 WG 3 at Lofoten (Norway, left), Aix-en-Provence (France,
middle) and Bokrijk (Belgium, right).
For the extraction of Cr(VI), US EPA method 3060A and ISO 16740 were considered, whereas for the
determination of Cr(VI) in the extract solution, US EPA 7199, ISO 16740 and ISO 10304-3 were taken
into account.14,27,42,44
As the scope of these standards did not cover solid material (with the exception of US EPA 3060A), a
robustness study had to be carried out in order to optimise the selected methodology and to
evaluate the applicability for soil and especially waste samples. This paper describes the results of
the robustness study and the results of the subsequent interlaboratory validation study on the
determination of Cr(VI) in solid material.
Extraction of Cr(VI) from solid matrices

To quantify Cr(VI) concentrations in a solid matrix and verify their compliance with maximum
contaminant limits, a preliminary extraction of Cr(VI) from the solid is necessary. Extractants should
be able to solubilize all forms of Cr(VI) without inducing changes in the speciation of chromium. The
extraction procedure is not required to be selective for Cr(VI) since the differentiation between
oxidised and reduced Cr species may be obtained by using specific analytical methods.
Initial attention to the determination of Cr(VI) in solid matrices was given to Cr(VI) levels in
workplace atmospheres. In 1977, the National Institute for Occupational Safety and Health (NIOSH)
proposed a method for the determination of Cr(VI) in atmospheric particulate matter. In 1984, the
US EPA proposed a protocol, known as Method 3060, which consisted of digesting 100 g solid sample
in 400 mL hot, alkaline solution of 0.28 M Na2CO3 and 0.5 M NaOH (pH about 12) for 30 – 45 min.28
Vitale et al. tested the accuracy of the hot alkaline digestion of soil samples and investigated the
reasons for occasionally poor Cr(VI) spike recoveries.29 In their study they confirmed the ability of the
method to extract soluble and insoluble Cr(VI) forms and concluded that very poor Cr(VI) spike
recoveries were observed only in strongly reducing samples. They proved that method-induced
oxidation of Cr(III) might occur with freshly precipitated Cr(OH)3. To help interpret recovery data,
they suggested the use of ancillary soil chemical parameters, including redox potential, pH, S2− and
total organic carbon concentrations as indicators of the redox state of samples. In the case of
samples characterised by strong reducing conditions, poor Cr(VI) recoveries should not be attributed
to deficiencies of the used analytical methods without any additional proof. It is more likely that they
are an indication on the potential of samples to reduce Cr(VI) spike and on their inability to sustain
the existence of Cr(VI) in the original sample.
James et al.30 compared five different extractants for Cr(VI) from soil: distilled water (pH 5.7),
phosphate buffer (pH 7.0), carbonate-hydroxide solutions (pH 12) with and without heating, and
hydroxide solutions (pH 13) with sonication. Their findings suggested that a carbonate-hydroxide
solution heated at 85 °C was the most effective extractant for operationally defining Cr(VI) in soil. In
1996 US EPA revised the Method 3060 for extracting Cr(VI) from soil, sludges, sediments and solid
wastes.27 This new method (3060A), that was adopted as basis for the European standard, consisted
of an alkaline digestion at 90 – 95 °C for 60 min according to the findings by James et al.30. This
method consists of placing 2.5 g of a field-moist and homogenised sample in a 250 mL digestion
vessel to which 50 mL of digestion solution (0.28 M Na2CO3/0.5 M NaOH), 400 mg of MgCl2 and 0.5
mL of 1.0 M phosphate buffer (0.5 M K2HPO4/0.5 M KH2PO4) are added. The addition of Mg2+ in a
phosphate buffer to the alkaline extraction solution will prevent risks of Cr(III) oxidation, which
otherwise may lead to Cr(VI) overestimate, particularly in samples with high Cr(III)/Cr(VI) ratios.
Although the US EPA revised Method 3060A gives maximal dissolution of all forms of Cr(VI) in solid
samples while minimizing method-induced oxidation and reduction, species transformation may still
occur. To correct for species transformation in the analysis of Cr(VI) in solid samples, speciated
isotope dilution mass spectrometry can be used as described by Huo and Kingston.31 EPA RCRA
Method 6800 [Speciated Isotope Dilution Mass Spectrometry (SIDMS)] addresses the correction for
such degradations or conversion.32
Cr(III)-Cr(VI) interconversions may take place when reactants, which are able to reduce Cr(VI) or
oxidise Cr(III) are present. The experimental conditions adopted for the extraction of Cr(VI) from solid
matrices control these possible interconversions. Cr(VI) may react with many inorganic reductants
such as Fe(II) and sulphide; a number of organic compounds including carboxylic and hydroxo-
carboxylic acids, aldehydes, phenols, humic acid, etc. are also able to reduce Cr(VI). Humic material
and Fe(II) are common components in soil and sediments and can be easily released from these
solids into strong alkaline solutions. However at pH values higher than 10, the risk of reduction of
Cr(VI) is strongly diminished.
Contrary to the diminished risk of reduction of Cr(VI) with increasing pH, the risk of oxidative
processes converting Cr(III) to Cr(VI) tends to increase with increasing pH. Cr(III) aging is also strongly
and positively affected by increase in pH and temperatures, thus reducing as matter of fact the
potential oxidation of Cr(III). The term aging refers to the change in physical and chemical properties
of freshly precipitated metal hydroxides and metal oxide hydrates as a function of time.33 Aging
processes include changes in the physical state (recrystallization), chemical structure (formation of
new crystalline phases accompanied by changes in atom connectivity), and composition (hydrolytic
polymerization).
Molecular oxygen and manganese oxides are possible oxidants during the digestion of solids. The US
EPA method 3060A took into account the possibility that native Cr(III) in solid matrices could be
oxidised under alkaline conditions and suggested that, in the case the oxidation was suspected, Mg2+
was added to the alkaline extracting solution to suppress the oxidation.27 It was hypothesised that
the suppression was due to Cr(III) coprecipitation with Mg2+ or to sorption of Mg2+ on Mn oxides
rendering them less prone to oxidise Cr(III).29 Mg2+ was also proved to play a strong negative effect
on the rates of oxidation of Cr(III) with H2O2 because of its influence on the aging of Cr(III).34 This
effect was supposed to be due to the formation of a solid phase of the type CrxMg(1 − x)1.5(OH)3 that,
similarly to the mixed solid phase CrxFe(1 − x)1.5(OH)3, controls the solubility of Cr(III).35 This effect of
Mg2+ is probably observed also in the case of the oxidation of Cr(III) with O2 and MnO2 and
substantiates the US EPA choice of adding this ion to suppress the oxidation of Cr(III) during the
alkaline digestion of solids. A similar influence on Cr(III) aging was also proved in the case of
carbonate.34
Based on these considerations concerning the kinetics of Cr(III) oxidation, a value of pH around 10,
high temperature and high concentrations of Mg2+ and carbonate ions would minimise risks of Cr(III)
conversion to Cr(VI) during the digestion of solid samples.
Determination of Cr(VI) in extracts

The photometric diphenylcarbazide (DPC) method is the most common method for determining
Cr(VI) in aqueous solutions.36 Nevertheless, this method suffers from the presence of interfering
compounds, some of which are explicitly reported in published protocols.37,38 In addition to species
containing molybdenum, mercury, iron, vanadium, which give a positive interference, the presence
of reductants able to compete with DPC under acid conditions may lead to underestimation of Cr(VI)
concentrations. Hydrogen peroxide, which reduces Cr(VI) to Cr(III) under acid conditions, is one of
the possible reductants in aqueous solutions. These also include Fe(II), sulphide, sulphite and a
number of organic compounds. However, the presence of effective concentrations of reductants of
Cr(VI) is not common in the analysis of aqueous samples, while it becomes much more probable in
the case of the application of the DPC method to soil and waste extracts.
Strong alkaline conditions favour the dissolution of Fe(III) species and humic-like matter that
interfere in the determination of Cr(VI) by the DPC method. The dissolution of Fe(III) is driven by the
formation of negatively charged Fe(III) hydrolysis products such as Fe(OH)4- while the release of
humic matter is connected with the formation of humates, which are soluble under strong alkaline
conditions.39,40
To separate Cr(VI) from positive interferences ion chromatography (IC) can be used.41 The ion
chromatographic method obviates most of problems due to dilution of the sample with the eluent
stream, passage through a guard column and Cr(VI) separation on an anion exchange column. An ion
chromatography method followed by a post column derivatisation of Cr(VI) by DPC was published as
US EPA 7199 and ISO 16740.14,42 Moreover when applying derivatisation of Cr(VI) with DPC, the use
of the ion chromatography method is advocated to overcome interferences from reductants during
the post column reaction.
An ion chromatography method followed by direct spectrophotometric detection of Cr(VI) at 365 nm
was published as ISO 10304-3.44
Experimental
Collection and preparation of the samples
For the robustness study, three soil samples and three waste samples were selected, homogenised
and divided into subsamples by means of a rotating sample divider into 100 mL brown glass bottles
(containing 50 -100 g of sample). A description of the samples and sample pretreatment is given in
Table 1a.
Code Sample origin Sample pretreatment

Soil 1 sandy soil, collected at a depth from 2 to 9 m A: air-dried, sieved (< 1 mm) and homogenised
from a site with galvanising industry B: air-dried, sieved (< 1mm), ground with a ball
(Belgium) mill to pass (> 95 %) a 125 µm sieve and
homogenised
Soil 2 loamy soil, collected along high way berm of air-dried, sieved (< 1 mm), ground with a ball
E17 at a depth of 1.5 – 2.5 m (Belgium) mill to pass (> 95 %) a 250 µm sieve and
homogenised
Soil 3 sandy soil with a very low content of organic A: air-dried homogenised fraction 0.5 - 2.0 mm
matter, excavated during a remediation B: air-dried fraction 0.5 - 2.0 mm, ground with
campaign on the terrain of a former a ball mill to pass (> 95 %) a 250 µm sieve and
electroplating plant (Germany) homogenised
Waste 1 paint sludge (Italy) air-dried, crushed and sieved to < 2 mm,
further completely ground to particle size <
250 µm and homogenized
Waste 2 fly ash (France) sieved (< 250 µm) without additional drying
and homogenized
Waste 3 filter cake (Belgium) air-dried, sieved (< 1 mm) and homogenised.
Table 1a: Description of the samples and the sample pretreatment.
For indications regarding the major matrix components, all samples were analysed with an Energy
Dispersive - X-Ray Fluorescence (EDXRF) screening method (Table 1b). The high performance X-
LAB2000 (Spectro Analytical Systems, Kleve, Germany) spectrometer was equipped with a 400 W Pd
end-window tube and a Si(Li) detector cooled with liquid nitrogen. For signal optimization the used
targets were Al203 and Pd as Barkla polarizer, an HOPG crystal as Bragg polarizer and a Mo target as
secondary target. The software was also provided with a quantitative programme set up with
geological samples. For laboratory analysis, all samples were analyzed as powders (dried and finely
ground) and placed in XRF sampling cups (Bruker AXS, Karlsruhe, Germany) provided with a 2.5 µm
Mylar foil (Chemplex, Tuckahoe, NY, USA).43 For the determination of the elements C, H, N, S a
Truspec elemental analyser (LECO Instrumente GmbH , Mönchengladbach, Germany) was used.
Element Dimension Soil 1 Soil 2 Soil 3 Waste 1 Waste 2 Waste 3

TIC % 0.29 1.15 1.03 2.62 0.59 0.34
TOC % 0.05 1.14 0.59 0.06 0.10 0.14
H % < 0.1 1.04 - < 0.1 < 0.1 1.45
N % 0.066 0.066 - 0.830 0.047 0.105
S % 0.203 0.444 0.588 1.0 3.77 11.3
Al % 0.7 2.5 1.0 1.8 1.3 < 0.1
Ca % 2.2 9.3 7.3 11 17 19
Fe % 0.46 6.5 1.7 5.8 0.81 8.0
K % 1.0 0.47 0.99 0.21 4.1 0.016
Mn % 0.007 0.062 0.034 0.120 0.036 0.100
Si % 31 15 24 2.6 3.5 0.061
Ti % 0.10 0.20 0.14 0.033 1.30 0.008
As mg/kg 2.2 26 5.4 <2 34 11
Ba mg/kg 298 378 424 993 1220 < 18
Br mg/kg <1 1.2 <1 121 891 <1
Cd mg/kg <3 <3 7 <3 267 <3
Cl mg/kg 626 < 55 129 5160 162000 63

Co mg/kg <8 100 <8 75 18 138
Cr mg/kg 46 8700 972 14800 592 18600
Cu mg/kg 7.6 58 68 6910 1020 350
Hg mg/kg <2 <2 <2 2.5 2.6 <2
Mo mg/kg <2 <2 <2 181 31 735
Ni mg/kg 7.8 457 22 2640 63 11300
P mg/kg 375 970 410 13400 1380 224
Pb mg/kg 7.6 196 109 1810 3700 29
Sb mg/kg <5 6.2 7.6 38 977 5.3
Sn mg/kg <9 <9 27 767 1320 13
Sr mg/kg 67 97 170 396 359 104
Zn mg/kg 14 293 159 2770 15000 55
Zr mg/kg 180 175 150 1350 195 < 11
Table 1b: Matrix composition of the samples.
Analytical methods
The following digestion equipments were evaluated during the robustness study:
• a thermostatically controlled hotplate with a magnetic stirrer (Ikamag RCT, IKA Labortechnik,
Germany);
• a thermostatically controlled heating block with a magnetic stirrer and reflux condenser
(Behrotest DET 6 MAG, Behr Labortechnik, Germany );
• an ultrasonic-assisted leaching system, where the ultrasonic radiation was applied by means
of a Branson 5210 ultrasonic bath (47 kHz, 185 W, The Netherlands).
For the determination of Cr(VI) in solid material the following alkaline digestion procedure was
applied: Approximately 2.5 g of the homogenised sample was placed in an appropriate digestion
vessel. 50 mL of digestion solution (0.28 M Na2CO3/0.5 M NaOH), 400 mg of MgCl2 and 0.5 mL of 1.0
M phosphate buffer (0.5 M K2HPO4/0.5 M KH2PO4) were added to the solid sample. The samples
were heated to (92.5 ± 2.5) °C for at least 60 minutes under continuous stirring (or, in case of
ultrasonic agitation, at 20 or 60°C). After digestion, the vessels were allowed to cool to room
temperature and the solutions were filtered using a 0.45 µm membrane filter (regenerated cellulose
Porafil membrane filter, Macherey-Nagel, Düren, Germany).
The determination of Cr(VI) in the alkaline extracts was performed after ion chromatographic
separation by spectrophotometric detection either at 365 nm (direct UV detection) or at 540 nm
after post-column derivatisation with 1,5-diphenylcarbazide in acid solution according to ISO 16740
and ISO 10304-3.14,44 In parallel, Cr(VI) was determined by species-specific isotope dilution mass
spectrometry according to Tirez et al..15 The total Cr concentration in the alkaline digestion solutions
was measured with ICP-AES after adjusting the pH < 2 with nitric acid.
Results and discussion
Robustness study
Objectives
Based on a survey of procedures in use in the different European member states it could be noticed
that a harmonization of digestion equipment and measurement method was needed.26 The
robustness study was performed in order to study the equivalency of different methods for the
determination of Cr(VI) in waste and soil using alkaline extraction and to produce a set of validation
samples with a homogeneous Cr(VI) load.
Regarding the evaluation of the extraction methodology, the objectives of the robustness study
implied the evaluation of hotplate, heating block and ultrasonic bath as digestion equipment. For the
determination methods of Cr(VI) in the extract solution, the robustness study implied the evaluation
of ion chromatographic separation and spectrophotometric measurement either at 365 nm or at
540 nm after post column derivatisation with 1,5 diphenylcarbazide and direct determination of
chromium in the alkaline extract with inductively coupled plasma-atomic emission spectrometry (ICP-
AES). The robustness study was performed within 3 laboratories (VITO, BAM and FENICE S.p.A.).
Digestion equipment – soil samples
Hotplate digestion was compared with ultrasound-assisted leaching on soil 1A/B and soil 3A/B (see
Table 1a) in two different laboratories. Both laboratories reported differences between hotplate
digestion at 92.5 °C and ultrasound-assisted leaching at respectively 20 °C and 60 °C on both soil
samples. The recovery of Cr(VI) in the ultrasound-assisted leaching at 20 °C and 60°C was less than 50
% and 75 %, respectively, of the recovery of Cr(VI) with hotplate digestion. These results are in
agreement with the comparison of five extraction methods on soils published by James et al..30 The
efficiency of extraction of different Cr(VI) forms (K2CrO4, SrCrO4, BaCrO4 and PbCrO4) using hotplate
digestion and ultrasonic extraction has also been studied in the frame of the determination of Cr(VI)
in workplace air.21 It was found that all these Cr(VI) forms (representing water-soluble, sparingly
soluble and insoluble forms) were completely dissolved when hotplate digestion was employed, but
that less than quantitative recovery for barium chromate was obtained using ultrasonic extraction.
The measured Cr(VI) content in soil 1 after hotplate digestion amounted to 2 mg/kg, while the total
Cr content measured with EDXRF and ICP-AES after digestion with an acid mixture of HNO3/HCl/HF
according to EN13656 45 amounted to 46 mg/kg. The measured Cr(VI) content in soil 3 after hotplate
digestion amounted to 400 mg/kg, while the total Cr content measured with EDXRF and ICP-AES after
digestion with an acid mixture of HNO3/HCl/HF amounted to 800 mg/kg.
The comparison between hotplate and heating block digestion was performed in one laboratory on
soil 3A/B. The results are summarized in Table 2.
Soil 3A Soil 3B
mg Cr(VI)/kg mg Cr(VI)/kg Recovery K2Cr2O7 spike % Recovery PbCrO4 spike %
Heating block 359 ± 10 (n=3) 394 ± 4 (n=9) 99 – 102 100 – 101
Hotplate 352 ± 26 (n=4) 402 ± 7 (n=4) 96 – 100 96 – 98

Table 2: Results of the robustness study comparing hotplate and heating block digestion methods.
Both digestion equipments gave comparable results and quantitative recoveries even for the
insoluble Cr(VI) spike PbCrO4. The grinding of the sample to a particle size of less than 250 µm (soil
3B) was reported to yield a significant better precision between the subsamples. It was also noticed
that the addition of MgCl2 and phosphate buffer suppressed the (minor) oxidation of an added CrCl3
spike (respectively 0,7 and 0.2 % oxidation of 1000 µg Cr(III) added to 50 mL of the alkaline digestion
solution).
The determination of Cr(VI) on soil 1 revealed that approximately 20 % of a soluble Cr(VI) spike (250
µg, added as K2Cr2O7) was not recovered. Duplicate analysis with both heating block and hotplate
showed a (79 ± 2)% recovery. In order to control this reduction process on soil 1, the Cr(VI) in soil
1A/B was also determined by species-specific isotope dilution mass spectrometry. The measured
Cr(VI) content on soil 1 amounted to 5 mg/kg. The used method allows the direct determination of
biases during digestion and confirmed that only 77 % of the added isotopically enriched 53Cr(VI) spike
was recovered in soils 1A/B. Moreover, also a 10 - 15 % oxidation of the isotopically enriched 50Cr(III)
spike was observed on soil 1A/B. This underpins observations that reduction reactions of Cr(VI) by
organic matter or other reducing reagents may occur simultaneously with oxidation of Cr(III).46 The
observed higher Cr(VI) content (5 versus 2 mg/kg) was attributed to incomplete suppression of
oxidation due to addition of only 1/10 of the prescribed amount of MgCl2 to the alkaline digestion
solution. The diminished added amount of MgCl2 was applied in order to reduce possible
interferences of chloride during species-specific isotope dilution mass spectrometry measurement.
The determination of Cr(VI) in soil 2 indicated no particular problems regarding the digestion and
possible oxidation/reduction processes. The measured Cr(VI) content in soil 2 after hotplate
digestion amounted to 2000 mg/kg, while the total Cr content measured with EDXRF amounted to
8700 mg/kg. Soil 1B and 2 were selected as validation samples for the interlaboratory trial.
Digestion equipment – waste samples
The results of the determination of Cr(VI) on the filter cake sample (waste 3) showed large
discrepancies. The addition of Cr(III) spike to control possible oxidation processes revealed that up to
90 % of the spike was oxidised during digestion. The measured Cr(VI) content in waste 3 after
hotplate digestion amounted to 1000 mg/kg, while the total Cr content measured with EDXRF and
ICP-AES after digestion with an acid mixture of HNO3/HCl/HF amounted to 20,000 mg/kg. As stated
above one must consider that when Cr(VI) is stabilised with respect to pH, Cr(III) is in an unstable
environment. The determination of Cr(VI) in a matrix in which the Cr(III) content is more than 20
times that of Cr(VI) remains ambiguous. Prolongation of the digestion time from 1 to 2 and 3 hours
showed an increase in the Cr(VI) content of about 25 % and 90 %, respectively. Also the drying of the
sample affected the Cr(VI) content drastically. Drying of this sample at different temperatures (40,
60, 80, 105 and 200°C) showed an increase of the apparent Cr(VI) content up to 10000 mg Cr(VI)/kg,
suggesting an increase of oxidation potential with drying. In order to confirm the oxidation processes
on waste 3, the Cr(VI) was determined by species-specific isotope dilution mass spectrometry. This
method confirmed that on the air dried sample 20 % of the added isotopically enriched 50Cr(III) spike
was oxidised during digestion, while the recovery of the added isotopically enriched 53Cr(VI) spike
amounted to 95 %.
The determination of Cr(VI) in waste sample 1 indicated no particular problems regarding the
digestion and possible oxidation/reduction processes and was therefore selected as validation
sample for the interlaboratory trial. Waste sample 2 was also considered to be a representative
waste sample and was retained for the validation.
Direct determination of Cr(VI) in the alkaline extract
When it is assumed that no species of chromium other than Cr(VI) are present after digestion then
direct determination will be possible. Based on a study performed by the Danish Hydraulic Institute
(DHI), the concentration of Cr(VI) in the alkaline digestion solution of soils equals the total chromium
concentration and can be determined directly with inductively coupled plasma-atomic emission
spectroscopy (ICP-AES) or atomic absorption spectrometry (AAS).47 Practically, in order to prove that
Cr(VI) is the only soluble form of chromium present in the alkaline digestion solution, spiking of the
sample with Cr(III) followed by the same digestion procedure as applied to the test portion is
required. Due to the high element concentrations, e.g. of sodium, in the alkaline digestion solution
the calibration strategy must be adapted appropriately. In many cases matrix matching of the
calibration solutions and/or dilution of the sample together with addition of internal standards or
using the standard addition method is necessary. The filtration of the alkaline digestion solution
before acidification is essential, as “active” chromium hydroxide, an exclusively hydrogen-bonded,
layered array of Cr(OH)3(H2O)3 units, instantaneously dissolves in acid to form Cr(H2O)63+.33 One
minute after its precipitation in the pH range 8.5-9.7, acidification yields > 99 % of Cr(H2O)63+. When
precipitates of “active” chromium hydroxide are aged in buffered aqueous suspension, the amount
of Cr(H2O)63+ recovered after rapid acid dissolution decreases with time.
The filtration of the alkaline digestion solution is also necessary to prevent possible losses of Cr(VI).
Although completely extracted, for BaCrO4 and PbCrO4 there is a risk that these compounds can
reprecipitate during neutralisation. Huo et al. have studied the processes responsible for possible
losses of Cr(VI).48 They found that CO32- is essential in the dissolution of BaCrO4 and that the
precipitation of BaCO3 greatly decreases the concentration of the free Ba2+, driving the dissolution of
BaCrO4 and releasing CrO42- as a free ion in solution. The precipitated BaCO3 is removed at the
filtration step. When neutralising the solution, CrO42- cannot precipitate because nearly all Ba has
been removed during filtration.
On the other hand, the dissolution of PbCrO4 in the alkaline digestion does not require CO32-, because
the process is driven by the formation of Pb(OH)2 and Pb(OH)n2-n. Although much of the Pb can be
removed after extraction as Pb(OH)2 by filtration, certain complexed Pb species can still remain in the
filtrate. During neutralisation and due to the continuously decreasing concentrations of OH- and CO32-
, the Pb species are converted from Pb(OH)n2-n to PbCrO4. Therefore some of the dissolved PbCrO4
reprecipitates during neutralisation, resulting in the loss of Cr(VI) from solution.
The direct determination of Cr(VI) in the alkaline digestion solution by ICP-AES was evaluated during
the robustness study. For the measurements the digestion solutions were diluted in 10 % (v/v) nitric
acid (ultra pure) and were placed in an ultrasonic bath for at least 15 minutes to degas. The
calibration solutions were prepared in 10 % v/v nitric acid solution (ultra pure). A 10 mg/l rhodium
containing solution of 10 % (v/v) nitric acid was added on-line as internal standard and mixed with
the sample solution before introduction in ICP-AES (20 % internal standard / 80 % sample). The
chromium emission line 205.522 nm was used as analytical detection line after correction with
internal standard. A ratio of 0.99 ± 0.04 between direct determination of Cr(VI) in 22 different
alkaline digestion solutions with ICP-AES versus ion chromatography with detection after post
column derivatisation with 1,5-diphenylcarbazide was observed.
Conclusion robustness study
From the results of the robustness study it was concluded that hotplate and heating block digestions
gave comparable results on all samples if continuously stirring was performed and the temperature
was controlled. In contrary, ultrasonic bath extraction (at 25°C and 60°C) gave significantly lower
recoveries of Cr(VI). The alkaline extracts were analysed with ion chromatography followed with
either direct spectrophotometric detection or detection after post column derivatisation with 1,5-
diphenylcarbazide. No significant differences could be observed between both detection methods.
Direct determination of the total chromium content in the alkaline digestion solution of the different
materials with ICP-AES gave comparable results when dilution and internal standard correction were
performed.
In accordance with the results of the robustness study it was decided within CEN/TC292/WG3 and
ISO/TC190/SC3/WG1 to limit the normative scope of the draft standard for digestion to hotplate and
heating block equipment. The quantification of Cr(VI) in the alkaline digestion solution can be
performed after ion chromatographic separation and either by direct spectrophotometrical
measurement at 365 nm (direct UV detection) or after post-column derivatisation with 1,5-

diphenylcarbazide in acid solution at 540 nm.
It was considered that the processing of Cr(III) and Cr(VI) spikes to estimate the accuracy of the
method is a prerequisite when analysing unknown sample matrices. Spike recovery experiments not
only reveal the quantitative recovery of the Cr(VI) spike, but also assist in the interpretation
regarding the oxidation-reduction potential of the sample. Nevertheless, as could be shown for soil 1,
waste 2 and waste 3, the determination of Cr(VI) in a matrix in which the Cr(III) concentration is more
than 20 times that of Cr(VI) may remain ambiguous.
Interlaboratory trial
Objectives
The samples soil 1B, soil 2, waste 1 and waste 2 were retained from the robustness study and were
considered as an appropriate set of validation samples with a homogeneous Cr(VI) load for the
interlaboratory trial. These four samples and a quality control solution were sent to the participating
laboratories listed in the acknowledgements (soil 1B was renamed soil 1 in the interlaboratory trial).
The quality control sample consisted of a Cr(VI) spiked alkaline digestion solution with a Cr(VI)
concentration of 424 µg/L and was intended to control the quality of the detection method. Data
sets of participants for the soil and waste samples were accepted for further statistical evaluation
only in the case that the results obtained for this control solution did not deviate more than 10%
from the assigned value. For the four test materials, every participating laboratory was obliged to
give results obtained at least by one of the methods proposed to be included into the analytical
protocol of the planned new European standard (Table 3, methods A - D). Additionally, participants
were requested to perform recovery experiments with Cr(III) and Cr(VI) spikes on each of the
delivered samples. The amounts of added spikes had to be chosen according to the contents of Cr(VI)
in the samples under investigation (same order of magnitude).
Data obtained with alternative digestion and detection methods could be reported as well (Table 3,
methods E - O). The average results of the Cr(VI) determinations per method and per sample are
summarised in Figure 1.
soil 1
5
4,5
4
3,5
mg Cr(VI)/kg
3
2,5
2
1,5
1
0,5
0
A B C D E F G H I J K L M N O
2500 soil 2
2000
mg Cr(VI)/kg
1500
1000
500
0
Figure 1a: Overview of accepted Cr(VI) results on soil 1 and soil 2 according to the different combinations of
digestion and determination methods (see Table 3). For methods A-D, the error bar represents the 95%
confidence interval of the mean calculated according to C.I. = ± t95%, n-1*s/√n, with t being a factor derived from
Student´s t-distribution, s the standard deviation and n the number of all accepted individual results.
waste 1
16000
14000
12000
mg Cr(VI)/kg
10000
8000
6000
4000
2000
0
waste 2
25
20
mg Cr(VI)/kg
15
10
0
Figure 1b: Overview of accepted Cr(VI) results on waste 1 and waste 2 according to the different combinations
of digestion and determination methods (see Table 3). For methods A-D, the error bar represents the 95%
confidence interval of the mean calculated according to C.I. = ± t95%, n-1*s/√n, with t being a factor derived from
Student´s t-distribution, s the standard deviation and n the number of all accepted individual results.
Code N Digestion Method of determination

Method A 4 hotplate IC with direct spectrophotometric detection
Method B 9 hotplate IC with spectrophotometric detection after post-column
derivatisation with DPC
Method C 5 heating block IC with direct spectrophotometric detection
Method D 5 heating block IC with spectrophotometric detection after post-column
Method E 1 hotplate ICP-AES
Method F 2 hotplate ET-AAS
Method G 3 hotplate direct spectrophotometric detection
Method H 1 hotplate, water IC with direct conductivity detection
Method I 1 heating block ICP-AES
Method J 1 heating block ET-AAS
Method K 1 heating block IC-ICP-MS
Method L 1 heating block direct spectrophotometric detection
Method M 2 closed vessel micro-wave IC with spectrophotometric detection after post-column
Method N 1 closed vessel micro-wave IC with direct spectrophotometric detection
Method O 1 ultrasonic ICP-AES
Table 3: Combinations of digestion and determination methods used by laboratories participating in the
interlaboratory trial (N = number of laboratories).
Evaluation and interpretation – proposed methods
The data for methods A - D were statistically evaluated according to ISO 5725-252 using commercial
software (Prolab, Quo data GmbH, Dresden, Germany) and are summarized in Table 4.
Table 4: Performance characteristics of an international interlaboratory comparison on Cr(VI)

determination (calculations according to ISO 5725-2).
Sample N n w(Cr(VI)) SR VR Sr Vr R r
[mg/kg] [mg/k [%] [mg/kg [%] [mg/k [mg/kg]
] ] ]
Soil 1 15 45 1.69 0.43 25.19 0.22 13.08 1.18 0.61
Soil 2 19 57 2007 205 10.22 88 4.36 568 242
Waste 1 19 57 11360 1308 11.51 788 6.94 3622 2183
Waste 2 13 39 12.90 8.97 69.55 1.59 12.31 24.85 4.40
N number of accepted laboratories

n number of accepted results
w(Cr(VI)) mass fraction of Cr(VI) calculated from N laboratory means
SR reproducibility standard deviation
Sr repeatability standard deviation
VR relative reproducibility standard deviation
Vr relative repeatability standard deviation
R reproducibility limit
r repeatability limit
In Tables 5 to 8 an overview of the Cr(VI) determination is given per sample and per combination of
digestion and detection method (methods A - D).
Table 5: Data for Cr(VI) determination and spike recoveries on soil 1 (low contaminated topsoil).
Method N n w(Cr(VI)) SDw CVw rec. SDrec.Cr(VI) rec. SDrec.Cr(III)

[mg/kg] [mg/kg] [%] Cr(VI) [%] Cr(III) [%]
[%] [%]
A 3 9 1.75 0.46 26.32 98.0 7.9 3.5 5.1
B 7 21 1.83 0.23 12.61 94.8 11.7 -1.7 12.4
C 2 6 1.58 0.56 35.13 95.5 10.6 3.6 0.8
D 3 9 1.36 0.51 37.25 96.5 2.7 1.1 3.7
N number of accepted laboratories

n number of accepted results
w(Cr(VI)) mass fraction of Cr(VI) calculated from N laboratory means
SDw standard deviation calculated from N laboratory means
CVw coefficient of variation of laboratory means
rec. Cr(VI) mean recovery of Cr(VI) spike
SDrec.Cr(VI) standard deviation of recoveries of Cr(VI) spike

rec. Cr(III) mean recovery of Cr(III) spike detected as Cr(VI)
SDrec.Cr(III) standard deviation of recoveries of Cr(III) spike detected as Cr(VI)
Table 6: Data for Cr(VI) determination and spike recoveries on soil 2 (high contaminated topsoil).

[%] [%]
A 4 12 2010 209 10.41 98.5 5.1 3.0 3.6
B 8 24 2073 102 4.92 99.1 16.9 1.4 10.7
C 4 12 1843 269 14.57 101.2 12.2 4.9 2.8
D 3 9 2044 221 10.82 101.1 10.8 1.3 5.0
Table 7: Data for Cr(VI) determination and spike recoveries on waste 1 (paint sludge).

[%] [%]
A 4 12 10695 838 7.84 96.9 5.5 2.4 2.8
B 8 24 11299 867 7.67 95.5 5.6 -1.7 8.5
C 4 12 11478 1327 11.56 97.9 13.0 1.7 6.0
D 3 9 12249 1796 14.66 96.7 7.6 4.2 3.4
Table 8: Data for Cr(VI) determination and spike recoveries on waste 2 (fly ash).

[%] [%]
A 2 6 11.91 6.16 51.70 67.9 53.9 25.5 26.2
B 5 15 14.09 8.88 63.03 90.3 46.1 13.8 20.3
C 3 9 14.64 10.09 68.93 74.0 38.0 6.6 7.7
D 3 9 9.83 13.08 133.04 49.1 55.8 3.1 7.1
The performance characteristics for Cr(VI) determination in the case of both soils and waste 1 were
acceptable. However, for waste 2 the large relative reproducibility standard deviation suggests
strong matrix effects. This indicates that for unknown matrices supplementary quality control data
are needed in order to assess the validity of the analytical results.
The spike recoveries obtained with methods A - D are good in the case of the two soil samples
(recovery Cr(VI) spike > 95 %, recovery Cr(III) spike < 5 %). Method B yielded for both soils the most
reproducible results. Especially for soil 1 (low contaminated) this can be related to the superior
sensitivity of the detection method (generally, the direct UV detection is less sensitive than the
detection after post-column derivatisation with 1,5-diphenylcarbazide).
For the two waste samples the spike recoveries obtained with methods A - D were considered good
only in the case of the paint sludge (recovery Cr(VI) spike > 95 %, recovery Cr(III) spike < 5 %).
However, for the fly ash sample the spike recovery data are unsatisfactory. The ranges of recoveries
for Cr(VI) and Cr(III) are unrealistically large and can be attributed to the poor reproducibility of the
determination due to the reducing tendency of the sample matrix. The latter was deduced based on
additional tests applying spiking with isotopically enriched chromium species in the digestion
procedure. Based on these results, sample heterogeneity as a major cause of poor recoveries could
be excluded. In this case no valid Cr(VI) content can be reported on the fly ash sample and any test
report should include a remark on the recoveries of the spiked samples. Fly ash matrices generally
contain Al, Si, Fe, Na, Ca, K, S, Zn, Ti and Cl as major components (see Table 1b). It is probable that
oxidation/reduction reactions will occur in this matrix as no universal condition exist which can deal
with the numerous potential mechanisms for redox-chemistry involving chromium. Similar problems
regarding the determination of Cr(VI) in fly ash have been reported by Karwas et al. and Hua et
al..50,51
Evaluation and interpretation – alternative methods
The direct determination of total chromium in the alkaline digestion solutions of the different
materials according to methods E, F, I and J (Table 3) gave results comparable to those obtained with
the selective methods A - D. Given as overall averages of the results obtained with methods E, F, I
and J, the following Cr(VI) contents were found: 1.81 ± 0.06 mg/kg for soil 1, 2070 ± 250 mg/kg for
soil 2 and 11400 ± 1300 mg/kg for waste 1. For soil 1 direct spectrophotometric determination of
Cr(VI) in the alkaline digestion solution after complexation with diphenylcarbazide was obviously
hampered by co-extracted interfering substances and is therefore not recommended (methods G
and L).
The use of a closed vessel micro-wave system (methods M and N) with temperature control and
continuous stirring as alternative heating device was reported by 2 laboratories. The closed vessel
method has been reported for determination of Cr(VI) in fly ash samples and is assumed to prevent
Cr(III) from oxidizing to Cr(VI) to a larger degree compared to hotplate method where the extraction
solution is exposed to ambient air.19 The reported results from one laboratory were in good
agreement with the methods using hotplate and heating block equipment (soil 1: 1.48 mg/kg; soil 2:
2100 mg/kg; waste 1: 10970 mg/kg). The Cr(VI) results reported from the other laboratory were only
in agreement for waste 1 (10180 mg/kg) and in moderate agreement for soil 2 (1620 mg/kg); for soil
1 a four fold higher Cr(VI) content was recovered (6.8 mg/kg). The observed difference for soil 1 may
be explained by partial Cr(III) oxidation (total Cr content of soil 1: 46 mg/kg) as was also noticed
during the robustness study when the addition of MgCl2 during digestion was omitted.
The results reported with the ultrasonic assisted leaching method (method O) confirmed the lower
recoveries that were found in the robustness study. For soil 1 and soil 2, respectively, 40 % and 80 %
of the average contents summarized in Table 4 were found. For waste 1 the reported result (10909
mg/kg) was in agreement with the average mean content (11360 mg/kg). These results are in line
with the results reported by one of the participating laboratories applying hot water extraction
(method H; soil 1: < 1 mg/kg; soil 2: 1007 mg/kg; waste 1: 9770 mg/kg), where also less effective
solubilisation can be expected compared to the alkaline digestion. Comparable results found for
waste 1 can be considered as an indication that for this sample the Cr(VI) was readily available.
Spiking experiments with soluble Cr(VI) (K2Cr2O7) have shown similar recoveries when using hot
water or alkaline digestion, however for recovery of insoluble Cr(VI) species (BaCrO4 and PbCrO4) an
alkaline digestion at elevated temperatures is needed.30 The soluble and exchangeable fractions of
Cr(VI) are useful parameters for estimating the mass fraction of Cr(VI) in soil that may leach to
groundwater or be absorbed by plants and micro-organisms. However, quantifying insoluble forms of
Cr(VI) is pertinent to environmental hazards associated with, e.g., airborne, respirable dust, colloid
and solute movement in groundwater and possible leaching from landfilled waste.
Summary
An European standard for the determination of Cr(VI) in solid material has been elaborated in the
framework of an international co-operation and finally validated in the course of an interlaboratory
comparison.52 The procedure is derived from the alkaline digestion prescribed by US EPA method
3060A and determines an operationally defined content of Cr(VI), including easily and sparingly
soluble chromates. Although the method is one of the most rigorous with respect to extraction
efficiency, an interpretation of obtained results with regard to the potential of the original sample to
sustain the existence of Cr(VI) is needed. Especially, when the Cr(VI) content is less than 5% of the
total chromium amount, the determination of Cr(VI) may remain ambiguous. The accuracy of the
determination of Cr(VI) in solid matrices, and especially in waste, remains a challenging field from the
point of view of minimal species interconversion on the one hand and maximum amount of Cr(VI)
extracted on the other.
Acknowledgements
We wish to thank our colleagues who, under the auspices of the European Committee for Standardisation and
the International Organisation for Standardisation, contributed to this work during the development and
validation of the standard. We wish to thank Luc Debaene from the Public Waste Agency of Flanders for
funding. The following laboratories participated in the validation campaign and are gratefully acknowledged:
• AGES Vienna,Competence Centre of Elements, Wien, Austria
• Analytico Milieu BV, Barneveld, The Netherlands
• ARPA - FVG, Udine, Italy
• ATOTECH Deutschland GmbH, Berlin, Germany
• BAM, Division I.1, Berlin, Germany
• CESI, Piacenza, Italy
• Deutsche Metrohm GmbH & Co. KG, Filderstadt, Germany
• Eurolab Srl, Nichelino To, Italy
• FENICE S.p.A., Centro Servici Ecologici, Cascine Vica Rivoli, Italy

• FISIA ITALIMPIANTI S.p.A., Laboratorio Chimico e Biologico, Genova, Italy
• Forschungs- u. Qualitätszentrum FQZ Brandenburg GmbH, Eisenhüttenstadt, Germany
• Health & Safety Laboratory, Inorganics & Fibres Section, Buxton/Derbyshire, UK
• INERIS, Verneuil en Halatte, France
• Institut Alpha, Wasser- und Umweltanalytik, Ulm, Germany
• Institut Pasteur de Lille, Departement Eaux Environnement, Lille, France
• Metrohm Belgium, Berchem, Belgium
• Syndial - Attivita Diversificate, Centro Igiene e Protezione Ambiente, Ferrara, Italy
• Ultra-Traces Analyses Aquitaine (UT2A), Helioparc Pau-Pyrenees, Pau, France
• VITO, Milieumetingen – MIM, Mol, Belgium
References paper, Validation of a European Standard for the determination of hexavalent

chromium in solid waste material, J. Environ. Monit., 2007, 9, 749–759.
1. R.J Bartlett and B.R. James, 1988. Mobility and Biovailability of Chromium in Soils. In: Chromium
in Natural and Human environments. Nriagu, J.O and Nieboer, E., eds.; Wiley-Interscience: New
York, 267-304.
2. M.J. Marques, A. Salvador, A.E. Morales-Rubio and M. de la Guardia, Fresenius J. Anal. Chem.,
1998, 362, 239-248.
3. J. Kotas and Z. Stasicka, Environmental pollution, 2000, 107, 263-283.
4. M. Pettine and S. Capri, Analytica Chimica Acta, 2005, 540, 231-238.
5. V. Gomez and M.P. Callao, Trends in Analytical Chemistry, 2006, 25, 1006-1015.
6. 2003/33/EC, Official Journal of the European Communities, Council decision of 19 December
2002 establishing criteria and procedures for the acceptance of waste at landfills pursuant to
Article 16 of and Annex II to Directive 1999/31/EC (2003/33/EC).
7. EN 12506 (2003): Characterization of waste - Analysis of eluates - Determination of pH, As, Ba,
Cd, Cl-, Co, Cr, Cr (VI), Cu, Mo, Ni, NO2-, Pb, total S, SO42-, V and Zn.
8. European Directive 62/94 on Packaging and Packaging waste, OJ No L365,31.12.94, 10-23.
9. ICG/TC 2, A collaborative study on the determination of hexavalent chromium in container
glasses, Glass Technology, 2001, 42, 148-152.
10. Directive 2003/53/EC (the 26th amendment on the marketing and use of certain dangerous
substances and prepartions (nonylphenol, nonylphenol ethoxylate and cement)).
11. EN 196-10 (2006): Methods of testing cement - Part 10: Determination of the watersoluble
chromium (V) content of cement.
12. 2002/95/EC DIRECTIVE 2002/95/EC OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 27
January 2003 on the restriction of the use of certain hazardous substances in electrical and
electronic equipment.
13. IEC 62321, 111/54/CDV, IEC Committee 111, Procedures for the determination of levels of six
regulated substances in electrotechnical products.
14. ISO 16740 (2005): Workplace air - Determination of hexavalent chromium in airborne particulate
matter - Method by ion chromatography and spectrophotometric measurement using diphenyl
carbazide.
15. K. Tirez, W. Brusten, A. Cluyts, J. Patyn and N. De Brucker, J. Anal. At. Spectrom., 2003, 18, 922-
932.
16. J. Scancar, R. Miacic, F. Seby and O.F.X. Donard, J. Anal. At. Spectrom., 2005, 20, 871-875.
17. F. Séby, M. Gagean, H. Garraud, A. Catetbon and O.F.X. Donard, Anal. Bioanal.Chem., 2003, 377,
685-694.
18. A.G. Coedo, T. Dorado, I. Padilla and F.J. Alguacil, J. Anal. At. Spectrom., 2000, 15, 1564-1568.
19. H.M.“Skip” Kingston, R. Cain, D. Huo and G.M.M. Rahman, J. Environ. Monit., 2005, 7, 899-905.
20. R.A. Sabty-Daily, K.K. Luk and J.R. Froines, Analyst, 2002, 127, 852-858.
21. K. Ashley, A.M. Howe, M. Demange and O. Nygren, J. Environ. Monit., 2003, 5, 707-716.
22. J.M. Boiano, M.E. Wallace, W.K. Sieber, J.H. Groff, J. Wang and K. Ashley, J. Environ. Monit., 2000,
2, 329 –333.
23. personal communication, J. Bierkens (Flemish Institute for Technological Research) and G. Van
Gestel (Public Waste Agency of Flanders).
24. Swedish Environmental Protection Agency, Guidelines for polluted soils, 2002 (Chapter 2).
25. Decreto Ministeriale n. 471, Gazzetta Ufficiale Supplemento Ordinario N. 293, 1999.
26. CEN/TR 14589 (2003): Characterization of waste - State of the art document – Chromium(VI)
speciation in solid matrices.
27. United States Environmental Protection Agency, Method 3060A, in: Test Methods for Evaluating
Solid Wastes, Physical/Chemical Methods, SW-846, Update, Office of Solid Waste and Emergency
Response, Washington, DC, 1996.
28. United States Environmental Protection Agency, Method 3060, in: Test Methods for Evaluating
Solid Wastes, Physical/Chemical Methods, second ed., SW-846, Office of Solid Waste and
Emergency Response, Washington, DC, 1984.
29. R.J. Vitale, G.R. Mussoline, J.C. Petura and B.R. James, J. Environ. Qual., 1994, 23, 1249-1256.
30. B.R. James, R.J. Vitale, G.R. Mussoline and J.C. Petura, Environ. Sci. Technol., 1995, 29, 2377-
2381.
31. D. Huo and H.M.”Skip” Kingston, Anal. Chem., 2000, 72, 5047-5054.
Solid wastes, Physical/Chemical Methods, SW-846, Draft Update IVA, Office of Solid Waste and
Emergency Response, Washington, DC, 1998.
33. L. Spiccia and W. Marty, Inorg. Chem., 1986, 25, 266-271.
34. M. Pettine, F.J. Millero and T. La Noce, Mar. Chem., 1991, 34, 29-46.
35. B.M. Sass and D. Rai, Inorg. Chem., 1987, 26, 2228-2232.
36. M. Pettine, S. Capri, Analytica Chimica Acta, 2005, 540, 239-246.

37. APHA, AWWA, WEF 20th ed. APHA: Washington DC 1998. 3–65.
38. United States Environmental Protection Agency, Method 7196A, 1992.
39. J. Millero, W. Yao and J. Aicher, Mar. Chem., 1995, 50, 21-39.
40. M.H.B. Hayes, R.S. Swift, R.E. Wardle and J.K. Brown, Geoderma, 1975, 13, 231-245.
41. E.J. Arar and J.D. Pfaff, J. Chromatogr., 1991, 546, 335-340.
Solid Wastes, Physical/Chemical Methods, SW-846, Update, Office of Solid Waste and Emergency
Response, Washington, DC, 1996.
43. C. Vanhoof, V. Corthouts and K. Tirez, J. Environ. Monit., 2004, 6, 344 -350.
44. ISO 10304-3 (1997): Water quality - Determination of dissolved anions by liquid chromatography
of ions - Part 3: Determination of chromate, iodide, sulfite, hiocyanate and thiosulfate.
45. EN 13656 (2003): Characterization of waste - Microwave assisted digestion with hydrofluoric
(HF), nitric (HNO3) and hydrochloric (HCl) acid mixture for subsequent determination of elements
in waste.
46. B.R. James and R.J. Bartlett, J. Environ. Qual., 1983,12, 173-176.
47. L. Rasmussen, I. Albreksten and A. Favrbo, Danish Hydraulic Institute, Analyse af udvaskeligt
Chrom I jord ekstraheret under basiske forhold, Rapport marts 2000, VKI sag: 11898/11, p.53.
48. D. Huo, Y. Lu and H.M. Kingston, Environ. Sci. Technol., 1998, 32, 3418-3423.
49. ISO 5725-2 (1994): Accuracy (trueness and precision) of measurement methods and results - Part
2: Basic methods for the determination of repeatability and reproducibility of a standard
measurement method.
50. C.P. Karwas, J. Environ. Sci. Health, 1995, A30(6), 1223-1235.
51. B. Hua, B. Deng, E.C. Thornton, J. Yang and J. Amonette, Water Air Soil Pollut., 2006, 1 (1), 1-10.
52. EN 15192 (2006): Characterization of waste and soil - Determination of Chromium(VI) in solid
material by alkaline digestion and ion chromatography with spectrophotometric detection.
3.3.3. HEXAVALENT CHROMIUM IN AMBIENT AIR
VITO’s interest for the determination of Cr(VI) started in 1997, with the certification of the Cr(VI) and
total leachable Cr contents in welding dust loaded on a filter (CRM 545) [131]. This project was
carried out under the Standards, Measurements and Testing Programme (SM&T, formerly BCR) of
the European Commission in order to improve and control the quality of measurement of Cr(VI) in
workplace atmosphere.
Besides in workplace atmospheres, elevated concentrations of hexavalent chromium can also be of

concern in ambient air in urban areas close to industrial zones. In 2008, VITO was commissioned by
the Flemish Environment Agency VMM to develop a monitoring method for the determination of the
concentration level of Cr(VI) in ambient air. A multi-disciplinary approach, including X-ray absorption
near-edge spectroscopy (XANES), was used for the validation of the monitoring method. XANES has
proven to be well-suited for determining elemental speciation because it can directly and non-
destructively analyse particulate matter samples at low concentration levels [136].
Thereafter, in 2010, a monitoring campaign was organised at 2 locations in an industrial zone at

Genk-Zuid (Belgium). The results of the validation and the monitoring campaign were published in a
local newspaper (see press cutting Het Belang van Limburg, 04/03/2011) and the scientific literature
(see paper: Determination of hexavalent chromium in ambient air: a story of method-induced Cr(III)
oxidation ?, Atmospheric Environment, 2011).
The data of the monitoring campaign suggested that (a variety of) environmental factors promoted
Cr(VI)-Cr(III) inter-conversion. Recent studies (2013) have further examined the environmental
factors affecting stability of Cr species in ambient PM in basic filter matrix under typical sampling
conditions [137]. Enriched isotope spikes (53Cr(VI) and 50Cr(III)) were used to monitor the Cr(VI)-Cr(III)
inter-conversion. Some factors that were examined included major air pollutants, i.e., SO2, O3 and
NO2, as well as temperature, humidity, and PM type (diesel exhaust particulate matter and
secondary organic aerosol).
The chamber study characterized concurrent, competing Cr reactions that affect speciation during
sampling with basic filters, i.e., one pathway drives Cr(VI) reduction while the other drives Cr(III)
oxidation. A suite of interrelated environmental factors that affect competing Cr speciation was
identified, including SO2, stable reactive oxygen species (ROS), oxidizable PM matrix components,
temperature, and humidity. Cr speciation under basic matrix depends on the oxidative capacity of
PM matrix: reducing components drive Cr(VI) reduction, whereas stable ROS, such as organic
peroxides drive Cr(III) oxidation. In another study, conversion of Cr(III) during sampling could be
attributed to the reaction of Cr(III) with dissolved Mn and also reactions with gaseous oxidants, such
as O3 and particle-bound reactive oxygen species [138]. This combination of factors lead to Cr(VI)
reduction and Cr(III) oxidation with significant day-to-day and seasonal variations in the field
(summarised in Figure 33). The use of basic, NaHCO3-pretreated filter medium will not prevent Cr(VI)
reduction under sampling conditions.
Figure 33: Inter-conversion between Cr(VI) and Cr(III) during air sampling in basic filter medium (courtesy of
[137].
The SIDMS method (see § 3.3.1) would be one promising approach to improve the measurement
accuracy since enriched isotope (53Cr(VI) and 50Cr(III)) can be applied to the samples to allow
simultaneous monitoring of Cr(VI) reduction and Cr(III) oxidation. However, there are still concerns
regarding the use of the SIDMS method. The critical prerequisite for using this method is that the
spiked isotope species should have the same chemical properties as the target component in the
samples such that the spiked species mimic the species in the air samples. Composition of ambient
Cr(III) and its conversion under sampling and extraction conditions need to be examined prior to
Cr(VI) concentration correction via SIDMS method. Huang et al., the author of the paper reporting on
the examination of environmental factors affecting stability of Cr species in ambient PM, states that
work about Cr(III) solubility on conversion will be reported in future publications [137].
Determination of hexavalent chromium in ambient air: a story of method-

induced Cr(III) oxidation ?
Kristof Tirez,a Geert Silversmit,b Nico Bleux,a Elke Adriaensens,c Edward Roekens,c Kelly Servaes,a Chris
Vanhoof,a Laszlo Vincze,b and Patrick Berghmansa
a
b
Ghent University, Department of Analytical Chemistry, Belgium
c
Flemish Environment Agency, Department Air, Belgium
Atmospheric Environment, 2011, 45, 5332 – 5341.

Abstract
The accuracy of the determination of Cr(VI) in ambient particulate matter remains a challenge from
the point of view of minimal Cr species interconversion. Knowledge of this method-induced oxidation
and reduction is particularly relevant for the determination of Cr(VI) in ambient particulate matter, as
the level of observed Cr(III) oxidation (average of 1.7 % in this study) can contribute significantly to
the monitored range of measured Cr(VI) in PM10. For Cr concentrations in PM10 > 10 ng Cr/m³, this
method-induced oxidation could lead to false positive exceeding of an air quality guideline value of
0.2 ng Cr(VI)/m³ in PM10. The median daily Cr(VI) concentration in PM10 measured over a monitoring
period of more than 2 months at two locations close to a stainless steel factory amounted to 0.9 ng
Cr(VI)/m³ and 0.27 ng Cr(VI)/m³. Average daily Cr(VI)/Cr ratios in PM10 of 3.5 % and 2.6 % were
measured at these locations. The described monitoring for the determination of Cr(VI) in ambient air
via alkaline impregnated filters is sensitive (method detection limit of 0.015 ng Cr(VI)/m³) and
reproducible (precision of the method ~25 %). The average Cr(VI) recovery of 75 % strongly indicates
the effects of ambient sampling conditions and ambient particles on the Cr(VI) recoveries. The
stability of the Cr(VI) and the Cr(III) spike on 0.12 M NaHCO3 impregnated filters observed with
XANES, indicates that the alkaline extraction of the filter in combination with the sampled air matrix
is likely to induce the Cr conversions. The XANES spectra shows further that a Cr-spinel is the
predominant component of Cr in ambient air PM10 at the monitored locations.
Introduction
Understanding the concentrations, composition, and sources of atmospheric particulate matter (PM)
is crucial since inhaling these particles can cause a wide array of negative health effects, including
mortality and morbidity due to cardiovascular and pulmonary disease (Werner et al., 2007). There
are many characteristics that can affect the toxicity of PM, including its physical and chemical
properties. While PM contains an enormous variety of organic and inorganic chemical species,
certain classes of these compounds are of particular interest because of their toxicity. For example,
transition metals such as Ni and Cr have been linked to adverse health effects through field and
laboratory studies (Cleven et al., 1992; Stern,1998).
The focus in this work is on Cr, which has two predominant oxidation states in the atmosphere: +3,
which is an essential nutrient in low doses, and +6, which is highly toxic and carcinogenic (Cleven et
al., 1992; Stern,1998; Goldoni et al.,2006; Caglieri et al., 2006).
Airborne Cr has been analyzed by many different techniques including UV−VIS spectroscopy,
inductively coupled plasma−mass spectrometry, atomic absorption spectroscopy, and X-ray
techniques (Unceta et al., 2010; Ashley et al., 2003). All these methods, except the X-ray techniques,
encounter two limitations: only the total Cr content can be determined, or wet extraction/separation
steps are required to distinguish between Cr(VI) and Cr(III). For example, the standard methods of
ISO, US EPA, NIOSH and OSHA for determination of Cr(VI) and Cr(III) in industrial hygiene particle
samples all require an extraction step (Unceta et al., 2010; Ashley et al., 2003; Ashley et al., 2009;
Boiano et al., 2000; Christensen et al., 1999; Dyg et al., 1994; Wang et al., 1997; Steinsberger et al.,
1994). The experimental conditions adopted for the extraction of Cr(VI) from particles significantly
influence the reliability of the final results owing to possible undesired Cr(VI)-Cr(III) interconversions
(Pettine et al., 2005).
From this point of view X-ray absorption near edge structure (XANES) has two main advantages for
analysis of Cr and other metals: (1) no sample preparation (e.g., extraction) is required, and (2) the
technique can distinguish between different compounds of the same metal, including different
oxidation states (e.g., Cr(VI), Cr(III), and Cr(0)) as well as different compounds with the same
oxidation state (e.g., Cr2O3 vs Cr2S3) (Werner et al., 2007). However for prolonged environmental
monitoring purposes, XANES is not a feasible option due to the requirement of a synchrotron X-ray
source.
For this reason, most papers dealing with environmental monitoring of Cr(VI) in particulate matter
fall back on a workable extraction. The aim of this paper is to highlight the relevance of Cr(III)
oxidation observed during a Cr(VI) monitoring campaign in respect to the accurate risk assessment of
hexavalent chromium in ambient air.
Determination of Cr(VI) in particulate matter in ambient air via extraction
The first methods for the determination of hexavalent chromium in aerosols via extraction were
developed to assess the toxicity of aerosols produced by metallurgical and engineering processes,
such as steel making, cutting, grinding and welding. Ever since then the kinetics and thermodynamics
of the Cr(VI)-Cr(III) equilibrium have played a major role for new method developments. In 1983 it is
reported by Gray et al. that in the fumes generated from stainless steels, the hexavalent chromium
content appears to rise to a maximum some time after formation of the aerosols and then partly
decay again (Gray et al., 1983). A few years later Zatka gives a theoretical background to the
observed oxidation of Cr(III) to Cr(VI) during the determination of hexavalent chromium in welding
fumes (Zatka, 1985).
An extensive review on the characterisation of the Cr(VI)/Cr(III) ratio in ambient air aerosols was
published in 1992 by Cleven et al. It is stated that any data on Cr(VI) before 1987 are probably too
high because of chromium species interconversion. A review on different digestion methods from
soil was presented in 1995 by James et al.. Since then, there is a consensus on the use of an alkaline
digestion solution to achieve minimal species interconversion and maximum extraction efficiency of
hexavalent chromium on solid materials. This in turn has lead to many standardised methods using
the alkaline digestion as reviewed by Unceta et al. in 2010.
Sampling of Cr(VI) in particulate matter in ambient air can roughly be divided in procedures using
impingers/denuders and procedures using (impregnated) filters (Cleven et al., 1992). In the case of
the use of impingers, the Cr(VI) determination can be performed directly in the impinger solution. In
the case of collection of PM on a filter, an extraction is needed prior to analysis. For the
determination of Cr(VI) in the extraction solution, different analytical methods have been applied.
For a discussion on possible detection methods we refer to Unceta et al., since most of the possible
interconversions occur during sampling and extraction.
In the first field evaluation studies to determine the environmental levels of airborne hexavalent
chromium using an impinger train, it is reported that the oxidation of Cr(III) to Cr(VI) does not occur
to a significant extent (0.7 %)(Sheehan et al., 1992). In different field studies using the impinger
method for monitoring airborne hexavalent chromium, ratios of Cr(VI) to total chromium from 0 up
to 25 % were observed; however, no assessment of Cr(III) oxidation is reported (Bell et al., 1997;
Finley et al., 1993; Metze et al., 2004). In more recent studies using impinger trains, the species
transformation from Cr(III) to Cr(VI) was evaluated. By Krystek et al. the oxidation of Cr(III) to Cr(VI)
was reported to be less than 2 %, Li et al. reported an average of 13 % conversion of Cr(III) to Cr(VI) in
solutions containing air sample matrix (Krystek et al., 2007; Li et al., 2002). Also an increasing rate of
conversion of Cr(III) with time in an alkaline solution was observed (3 % after 24 h to 10 % after 72
h)(Li et al., 2002).
An elaborate research on the fate of hexavalent chromium in the atmosphere, including sampling on
an alkaline impregnated filter was published by Grohse et al. in 1988. The potential oxidation of
Cr(III) with MnO2 is theoretically discussed and it is recommended that further testing should be
performed. In some field studies using filters to collect the airborne particulate matter, however, no
assessment of Cr(III) oxidation is reported, while ratios of Cr(VI) to total chromium from 20 – 30 %
were observed (Borai et al., 2002; Talebi et al., 2003). In a recent paper Cr(VI) is reported to
constitute up to 50 % of the total chromium in Polish urban aerosols (Swietlik et al., 2011).
For both sampling methods, impingers and filters, automated measurement systems have been
described for continuous measurement of Cr(VI) in airborne particulate matter. However, no
indication is given on the potential oxidation of Cr(III) to Cr(VI) during the measurement (Samanta et
al., 2001; Khlystov et al., 2006; Isakov et al., 2007).
Because of the several directives and recommendations in Europe and the US to limit and regulate
the presence of Cr(VI) in the environment, governmental funded research has been performed to
develop standard operating procedures for the determination of Cr(VI) in airborne particulate matter
(Ashley et al., 2003). The US EPA funded the development of a standard operating procedure for the
analysis of hexavalent chromium at ambient atmospheric levels by Ion Chromatography (IC) using
impregnated filters (SOP CARB 039, 2006). Within the US national monitoring program, more than
1400 measurement were performed over 22 locations in 2005. The average concentration amounted
to 0.044 ng Cr(VI)/m³ and the highest concentration was reported in Washington DC, 2.97 ng
Cr(VI)/m³ (ERG, 2006). In this standard operating procedure the quality control measures imply solely
the control on possible Cr(VI) reduction.
A similar approach was performed by the Austrian Environmental Agency to monitor the Cr(VI)
content in PM10 in the city of Vienna in 2007. The range of Cr(VI) in PM10 amounted 0.04 – 0.23 ng
Cr(VI)/m³ and represented about 1 % of the total Cr content. No indication is given on the potential
oxidation of Cr(III) to Cr(VI) during the procedure (Hagendorfer et al., 2007).
For an evaluation of Cr(VI) in ambient air in the Netherlands near wood preservation plants and at a
regional site, the National Institute of Public Health and the Environment developed an
impinger/denuder system (Mennen et al., 1998). Preliminary tests without a denuder showed that
conversion of Cr(III) exceeded 10 %. With the introduction of a denuder to remove reactive gases,
the conversion from Cr(III) ranged from 0.7 to 3 %.
With the introduction of speciated isotope dilution mass spectrometry (SIDMS), possible species
interconversions occurring during the digestion can be mathematically corrected (Kingston, 1995).
SIDMS uses the concept of spiking the sample with known amounts of enriched isotopes that have
been chemically converted into the same forms of the species to be analysed.
This approach has been reported to be successful, especially to quantify the reduction of Cr(VI),
where the addition of enriched “soluble” (K2Cr2O7) and “insoluble” (PbCrO4) Cr(VI) compounds were
added to quantify the order of magnitude of reduction (Kingston et al., 1998; Huo et al., 1998; Tirez
et al., 2003; Tirez et al., 2007). With respect to the correction for Cr(III) oxidation, the SIDMS method
suffers some additional limitations. The correction is limited by the knowledge of the form in which
Cr(III) is present in the particulate matter. Cr species in the environment include both soluble and
insoluble forms; particularly, most Cr(III) in ambient air is present as insoluble forms and the
conversions are expected to be lower for these insoluble Cr species. Also the surface of the
particulate matter and the way in which Cr(III) is included will influence the solubility of Cr(III) during
the extraction. Moreover, it has been observed that when using the procedure as described in the
SIDMS patent (Kingston, 1995) for the preparation of the enriched Cr(VI) spike, traces of H2O2 in the
spike solution may, after the addition to the sample in an alkaline environment, lead to possible
oxidation of Cr(III)(Tirez et al., 2003). This implies that the matrix of the spike solutions which are
added to the extraction solution may introduce supplementary artefacts.
Using an isotope dilution mass spectrometric method, Nusko et al. reported that the Cr(III)/Cr(VI)
ratio was found to be about 0.3 in aerosol particles. In a study on Cr(VI) in house dust using enriched
Cr spikes, recovery for Cr(III) of 95 ± 10 % and for Cr(VI) of 90 ± 6 % were reported on blanks (Stern et
al., 2010). An average conversion rate of < 5 % in either direction was observed. However spiking of
the sample matrices with the enriched isotopes resulted in inconsistent recovery of both isotopes for
which no reason could be given. In a recent paper by Meng et al. on the further development and
evaluation of a method for hexavalent chromium in ambient air using enriched isotopes, the Cr(III)
conversion was further studied. The conversion from the enriched Cr(III) spike to Cr(VI) in a pH 4
HNO3 extraction solution was on average 2.6 % for blank filters and 5 % for a NIST 1648 particles
coated on a filter. No correction is finally made for Cr(III) conversion because of the expected
difference in oxidation between the enriched Cr(III) spike and the Cr(III) present in the sample (Meng
et al., 2011).
In summary it can be stated that for the determination of Cr(VI) in particulate matter in ambient air
via extraction, there is enough evidence that all methods reported so far in literature suffer to a
greater or lesser extent from method-induced Cr(III) oxidation. This oxidation sets limits to the lower
limit of Cr(VI) detection relative to the total Cr(III) content of the particulate matter.
Determination of Cr(VI) in particulate matter in ambient air via XANES
X-ray based techniques are generally regarded as reference methods for solid state speciation since
they enable the identification and quantification of chromium oxidation states while being non-
destructive. The most employed techniques are X-ray absorption near-edge spectroscopy (XANES), X-
ray Diffraction (XRD) and X-ray photoelectron spectroscopy (XPS) (Unceta et al., 2010). Whereas XPS
and XANES will enable quantification of the redox status of chromium, only the nature of the
crystalline phases under which chromium is present can be reached by XRD. Detection limits of XPS
and XRD are usually about 0.1 wt. %. XPS enables chromium speciation if Cr(VI) represents at least 10
wt. % of total chromium.
A number of studies have been published on chromium speciation with XANES in combustion aerosol
particles and in airborne fine particles (Werner et al., 2007; Huggins et al., 2004; Goodzari et al.,
2008; Wang et al., 2007; Galbreath et al., 2004; Huggins et al., 2000; Galbreath et al., 1994). A great
advantage of XANES is that it can distinguish between different compounds of the same metal,
including different oxidation states (e.g., Cr(VI), Cr(III), and Cr(0)) as well as different compounds with
the same oxidation state (e.g., Cr2O3 vs Cr2S3). However, because of the fitting procedure to
determine the fraction that each Cr species contribute to the total amount of Cr, any species that
accounts for less than ~5 % is within the uncertainty of the resolving power of the fitting program
(Werner et al., 2007). The ranges of Cr(VI) in particulate matter have been reported to range from 6
to 43 % of the total Cr content in PM2.5 from coal combustion (Galbreath et al., 1994). In particulate
matter reference material NIST 1648 and 1650 the oxidation state of chromium is predominantly (>
95 %), if not entirely, in the trivalent state of the chromium (Huggins et al., 2000). In particulate
matter collected close to an iron and steel industrial area, the main chemical component of Cr in PM
was chromite and oxides of trivalent chromium (Wang et al., 2007).
The speciation of chromium in airborne fine particles in combination with atmospheric

transformations have been extensively studied with (micro-)XANES by Werner et al. and Nico et al..
These studies demonstrated that aerosols contain a suite of different Cr(III) species, as well as more
reduced Cr species. The major component was in all cases a Cr-Fe spinel, suggesting that the
chemistry and toxicity of this phase will likely play a key role in the Cr associated health effect of
ambient PM. In the study of the redox dynamics of mixed metal (Mn, Cr and Fe) ultrafine particles,
two reaction pathways, one reductive and one oxidative, were found to be operating simultaneously
during simulated atmospheric aging. The presence of Mn within the particles enhanced the
importance of the oxidative pathway, leading to more net Cr oxidation during aging, implying that
Mn can mediate oxidation by removal of electrons from other particulate matter (Nico et al., 2009).
In summary XANES has proven to be an appropriate speciation method for Cr in particulate matter in
ambient air, however due to the linear fitting process the method is limited to detect fractions of Cr
species that contribute to more than ~ 2 - 5 % of the total amount of Cr.
Materials and Methods
Cr(VI) - Cr(III) measurements via XANES
Preparation of the pure phase Cr compounds
In total, 6 reagent-grade Cr compounds were purchased: Cr2O3 (Alfa Aesar GmbH & Co., Karlsruhe,
Germany); Cr3C2 (Sigma-Aldrich NV/SA, Bornem, Belgium), FeCr2O4 (Norkem, Grootebroek, The
Netherlands), Cr2S3 (Merck, Darmstadt, Germany), Cr2(SO4)3 (Alfa Aesar) and K2Cr2O7 (Merck). These
samples were mixed with boron nitride (BN, an inert and weakly absorbing binder) and pressed into
self-supporting pellets. For metallic Cr, a 3 µm thick Cr foil was used to collect the Cr K edge XANES
spectrum.
Sampling
XANES analysis was performed on the following particulate matter samples:

1. sample X: two 0.12 M NaHCO3 impregnated ashless cellulose filters (grade 40, Whatman
International Ltd., Maidstone, UK) were spiked with 25 µl of a 1000 µg ml-1 Cr(NO3)3.9H2O
solution (Spex Certiprep Inc., Metuchen, USA). Both filters were measured directly via XANES
after spiking. Subsequently one filter was sampled with a Tecora Bravo Plus sampler (TCR
Tecora Srl., Milan, Italy) for 16 h at the European Synchrotron Radiation Facility (ESRF,
Grenoble, France) on 25/06/2009 (sampled volume : 12.8 m³). The other filter was preserved
under argon atmosphere in an exsiccator for 16h. Thereafter both filters were analyzed.
2. sample X2: two 0.12 M NaHCO3 impregnated ashless cellulose filters (Whatman) were spiked
with 25 µl of a 1000 µg ml-1K2Cr2O7 solution (Merck). Both filters were measured directly via
XANES after spiking. Subsequently one filter was sampled with a Tecora Bravo Plus sampler
(TCR Tecora Srl, Milan, Italy) for 16 h at the European Synchrotron Radiation Facility (ESRF,
Grenoble, France) on 26/06/2009 (sampled volume: 12.8 m³). The other filter was preserved
under argon atmosphere in an exsiccator for 16h. Thereafter both filters were analyzed.
3. samples X3 and X4: ambient particulate matter (PM10) collected on a 150 mm diameter PTFE
filter (TE 38.5 µm, Schleicher&Schuell, ‘s-Hertogenbosch, the Netherlands) during 72 h with
Digitel DHA-80 high volume samplers (Digitel Elektronik AG, Hegnau, Switzerland) in a rural
area in Flanders (Mol, Belgium). Sampling was performed on 22/06/2009 and the filters were
stored in a sample holder under ambient conditions until measurement on 25/06/2009.
Picture 6: Digitel DHA-80 high-volume samplers and Partisol speciation sampler in a rural area (Mol, Belgium),
with details of the Partisol speciation sampler filter holder (PM10).
4. Samples X5 and X6: ambient particulate matter (PM10) collected on a 47 mm diameter PTFE
filter (Schleicher&Schuell) with a Partisol speciation sampler (Partisol Plus model 2300 filter
sampler, Thermo Fisher Scientific Inc, Waltham, USA). The sampling was performed at a
monitoring site of the Flemish Environment Agency (Belgium), which is located at 100 m
from a stainless steel factory and in the vicinity of a residential area. Sampling was
performed in the period from 28/04 to 4/05/2009 (sampled air volume ~ 43 m³, 3 days of
continuous sampling on each filter). The filters were stored in a sample holder under
ambient conditions until measurement on 25/06/2009.
Picture 7: Sampling location and a Partisol speciation sampler used for the simultaneous sampling on 4 filters.
XANES data collection
The Cr K edge XANES measurements were performed at the XAS station of the Dutch-Belgian CRG
beamline (DUBBLE BM26A) of the European Synchrotron Radiation Facility (ESRF, Grenoble, France)
(Nikitenko et al., 2008). This XAS instrument is located at a bending magnet port (magnetic field
strength B=0.4 Tesla, critical energy Ec=9.6 keV) of the 6 GeV electron storage ring and is equipped
with a Si(111) double crystal monochromator. The higher harmonics were suppressed by a Si
reflecting strip on a flat mirror behind the monochromator. The intensity of the incoming and
transmitted X-ray beams was measured with ionisation chambers (Oxford Instrument). The
fluorescence XAS spectra were collected with a LN2 cooled energy dispersive (ED) nine channel
monolithic Ge detector. The spectra were normalized by subtracting the pre-edge background and
dividing the absorption data by the edge step at 50 eV above the Cr K edge position. The spectra on
the Cr reference compounds were recorded in transmission mode. The fluorescence mode was used
for the filter samples. The dimensions of the incoming beam were 4 mm horizontally by 0.5 mm
vertically.
Picture 8: XANES measurements at the XAS station of the Dutch-Belgian CRG beamline (DUBBLE BM26A) of the
European Synchrotron Radiation Facility (ESRF, Grenoble, France).
Cr(VI) – Cr(III) sampling and measurements for the monitoring campaign
Preparation of the filters
Ashless cellulose filters (grade 40, 47 mm, 8 µm, Whatman International Ltd., Maidstone, UK) were
selected for sample collection. Single filters were pretreated with 8 ml of 0.12 M NaHCO3 (Alfa Aesar,
Karlsruhe, Germany) in closed petri dishes (55 x 14 mm, VWR, Leuven, Belgium) for 16 h. The filters
were dried under N2 atmosphere in an exsiccator.
Picture 9: Impregnation of the filters.
The addition of Cr(VI) spike on the filter at a level of 20 ng Cr(VI)/filter was performed by using an
automatic pipette of 40 µl of a 0.5 µg Cr(VI) ml-1 K2Cr2O7 stock solution (Merck). The addition of
Cr(III) spike on the filter at a level of 1000 ng Cr(III)/filter was performed by using an automatic
pipette of 40 µl of a 25 µg Cr(III) ml-1 Cr(NO3)3 stock solution (Spex Certiprep Inc., Metuchen, USA).
Picture 10: Addition of Cr(VI) spike on the filter.
For the determination of total Cr, sampling was performed on Teflon filters (Teflon 47 mm, 2.0 µm,
Pall Corporation, Sint-Stevens-Woluwe, Belgium).
Sampling
Ambient particulate matter (PM10) was collected at 24 hour intervals over a period from 04/10/2010
until 09/12/2010 at two locations in the Flemish region of Belgium. Location 1 was situated at 190 m
of a stainless steel factory in the predominantly downward wind direction. Location 2 was situated in
the same wind direction, but 850 m further away. On location 1 sampling was performed each day
from 9 am to 9 am on the next day. On weekdays samples were recovered immediately after
sampling and transported to the laboratory (~ 2 h) in a cooled transport (< 5°C). Sampling in the
weekend was performed automatically, but samples were left on the instrument and collected on
Monday mornings. Sampling of PM10 at location 2 was performed on weekdays only.
Picture 11: Ambient particulate matter (PM10) sampling location 1 (GK11) and location 2 (GK05).
In order to evaluate concentrations and conversions between Cr-species, 5 separate samplings were
performed simultaneously at each location every day. Samples (a)-(d) were taken with a Partisol
speciation sampler. Sample (e) for the determination of total Cr was sampled with a Partisol Plus
model 2025 filter sampler:
a) a 0.12 M NaHCO3 impregnated ashless cellulose filter
b) a 0.12 M NaHCO3 impregnated ashless cellulose filter (duplicate sample of (a))
c) a 0.12 M NaHCO3 impregnated ashless cellulose filter spiked with 20 ng Cr(VI)
d) a 0.12 M NaHCO3 impregnated ashless cellulose filter spiked with 1000 ng Cr(III)
e) a teflon filter (PTFE) for the determination of total Cr
Extraction and Analysis
The filters were analyzed within 24 hours after sampling for filters collected during weekdays and
otherwise kept in a freezer in the laboratory (-18°C) until measurement. For the alkaline extraction of
the filter a temperature controlled ultrasound-assisted leaching system was used (Branson 5210, 47
kHz, 185 W, Soest, the Netherlands). For the determination of Cr(VI) in particulate matter collected
on a filter, the following alkaline digestion procedure was applied. The filter was placed in a 15 ml
tube (VWR), 10 mL of 0.02 M NaHCO3 digestion solution was added. The filters were extracted at 15
°C ± 2.5 °C for 60 minutes. After digestion, the solutions were filtered using a 0.45 µm membrane
filter (PTFE filtercaps, Dionex, California, USA).
The determination of Cr(VI) in the alkaline digestion was performed after ion chromatographic
separation and measured spectrophotometrically at 530 nm after post-column derivatization with
1,5-diphenylcarbazide in acid solution according to SOP CARB 039 (SOP CARB 039, 2006; Dionex
application 179). The ion exchange chromatography system consisted of a Dionex DX-120 system
coupled to an UV-VIS spectrophotometer (Spectrasystem UV 1000, Thermo, Breda, The Netherlands)
and 2 HPLC pumps (Model 426, Alltech, Deerfield, USA). The ion exchange chromatography system
conditions are summarized in table 1. Calibration standards were prepared daily with deionized
water (18.2 MΩ cm-1) using the Milli-Q system (Millipore, Brussels, Belgium) and from a 1000 µg ml-1
K2Cr2O7 standard stock solution (Merck). The laboratory quality controls were prepared from a 1000
µg ml-1 Cr(VI) standard solution (Merck).
Eluent 250 mM (NH4)2SO4 /100 mM NH4OH

Guard column Ionpac NG1 dionex
Pre-column Ionpac AG7 dionex
Analytical column Ionpac AS7 dionex
Measurement wavelength 530 nm
Eluent flow rate 1,5 ml/min
Post column reagent flow rate 0,5 ml/min
Sample volume 1000 µl
Knitted reaction coil 750 µl
Table 1: IC-DPC settings for the determination of Cr(VI) in extraction solutions.
The elemental concentration of Cr on the filters was measured with polarized Energy Dispersive XRF
(EDXRF Spectro, X-lab 2000). The calibration of the XRF spectrometer was performed by an in-house
developed aerosol generation system based on the ultrasonic nebulization of a multi-element
solution and collection of the aerosols on a filter (Vanhoof et al., 2003).

XANES measurements
Stability of Cr(III) and Cr(VI) spiked filters
The transmission XANES spectra on the Cr reference powders pressed in BN pellets are given in figure
1. Comparison with literature shows that the XANES spectrum for Cr2O3 contains two pre-edge peaks
(Wang et al., 2007). However, in our spectrum, the intensity of the second pre-edge peak around
5995 eV is more intense, which suggests that a small fraction of the Cr atoms was oxidized to Cr(VI)
or traces of Cr(VI) were present in the reagent-grade Cr2O3 compound.
Figure 1: Normalized Cr K transmission XANES spectra for the Cr reference compound pressed in BN pellets
(vertically shifted for clarity, bottom shows zoomed view around the absorption edge).
The Cr K fluorescence spectra for the filter spiked with a Cr(III) solution of Cr(NO3)3 (sample X1) is
given in figure 2. The spectrum of the filter stored for 16 h under argon condition is identical to the
freshly loaded filter. The filter that was sampled under ambient air conditions shows a small
broadening of the white line peaks. This points towards a possible broadening of the corresponding
electron levels or a higher structural disorder. However, no indication for oxidation to Cr(VI) is
observed.
Figure 2: Normalized Cr K fluorescence XANES spectra of the 0.12 M NaHCO3 impregnated ashless cellulose
-1
filters spiked with 25 µl of a 1000 µg ml Cr(NO3)3.9H2O solution (sample X1).
The Cr K fluorescence spectra for the freshly loaded Cr(VI) reference solution (sample X2) after 16 h
storage under argon conditions, after 16 h under ambient air and after subsequent 11 h irradiation
with the X-ray beam are given in figure 3.
-1
filters spiked with 25 µl of a 1000 µg ml K2Cr2O7 solution (sample X2).
By comparing the intensity of the Cr(VI) pre-edge with the pre-edge peak intensity of the freshly
prepared Cr(VI) spiked filter, an estimate of the Cr(VI) fraction can be obtained. Both storages under
argon and under ambient air did not significantly change the Cr(VI) fraction. On the freshly prepared
Cr(VI) filter and the filter stored under argon, 2 subsequent XANES scans were recorded with a
collection time of about 18 min each. For both filters, the Cr(VI) fraction in the second scan
decreased to 0.90, showing a reduction under influence of the X-ray irradiation. A second location on
the filter stored under argon was analysed, giving the same result. The sample stored for 16 h under
ambient air has a Cr(VI) fraction of 0.84, showing a partial reduction of the Cr(VI) under ambient air
storage. The time dependence Cr(VI) reduction under X-ray radiation was studied in more detail on
the under ambient air stored filter by recording XANES spectra every 11 min up to a total irradiation
time of 11 hours, a selection of these spectra is given in figure 4. From the time plot of the obtained
Cr(VI) fraction, the standard deviation on the Cr(VI) fraction was estimated to be 0.02. An almost
complete reduction of Cr(VI) on the filter under influence of the X-ray radiation can be observed.
-1
filter spiked with 25 µl of a 1000 µg ml K2Cr2O7 solution (sample X2) during X-ray irradiation over 11h (spectra
were recorded every 11 min, only a selection is shown in this figure).
In summary the XANES data show that the Cr(III) spike and the Cr(VI) spike are stable (> 95%) on the
impregnated filters after 16 h sampling under ambient air conditions. Prolonged X-ray irradiation of
the filter converted the spiked Cr(VI) to Cr(III).
XANES analysis of ambient particulate matter
The Cr K fluorescence XANES spectra for the air filters X3 and X4 sampled in a rural area in Flanders
(Mol, Belgium) are given in figure 5. The main component is identified as FeCr2O4. The observed
noise on the X3 and X4 spectra are due to the low Cr content on the filters, respectively 0.4 and 0.45
µg Cr/filter. The purchased reagent grade FeCr2O4 compound was also measured with XRD. These
results reveal that this compound contains a spinel-structure which is isomorphous with Fe3O4. Based
on the gradient in the diffraction lines, lattice parameters with values of 8.30 Å and 8.27 Å were
calculated. The structure is related to a AlCrFeMgO spinel, suggesting that Cr is infiltrated in a spinel-
structure. This could explain the small differences observed in the XANES spectra between the
reference FeCr2O4 compound and the filter samples.
Figure 5: Normalized Cr K fluorescence XANES spectra for the X3 and X4 ambient air filters (rural area)
compared with the FeCr2O4 reference compound (transmission mode).
The Cr K fluorescence spectra from the X5 and X6 filters are given in figure 6. The main identified
component is the FeCr2O4 form.
Figure 6: Normalized Cr K fluorescence XANES spectra for the X5 and X6 ambient air filters (near anthropogenic
Cr source) compared with the FeCr2O4 reference compound (transmission mode).
The total Cr content of X5 and X6 was respectively 1.8 µg Cr/filter (42 ng Cr/m³) and 4 µg Cr/filter (79
ng Cr/m³). These results are in agreement with literature data suggesting that a Cr-Fe spinel is the
major component of Cr in ambient air PM10 (Werner et al., 2007; Huggins et al., 2000; Wang et al.,
2007).
Monitoring campaign
Laboratory quality controls
With the optimized settings of the ion exchange chromatography system as described in table 1, the
method detection limit (calculated as prescribed in SOP CARB 039) was 0.022 µg Cr(VI) l-1 (this
corresponds to 0.015 ng Cr(VI)/m³). This analytical method is highly sensitive and more cost effective
for monitoring purposes compared to, e.g., IC-ICP-MS (Tirez et al., 2003; Meng et al., 2011). To
control the daily performance of the ion exchange chromatography system, a calibration blank and 3
calibration verification standards were analyzed. To control the daily performance of the extraction,
an extraction blank and two extraction solutions spiked with respectively 2 µg/l Cr(VI) and 100 µg/l
Cr(III) were analyzed. To control the filter impregnation procedure, an impregnated blank filter, an
impregnated filter spiked with 20 ng Cr(VI) and an impregnated filter spiked with 1000 ng Cr(III) were
extracted and analyzed in every measurement run. The results of these quality control
measurements are summarized in table 2.
Quality control value

calibration blank < 0.1 µg Cr(VI)/l (n = 37)

calibration verification standard 0.1 0.103 ± 0.015 µg Cr(VI)/l (2s, n = 35)
calibration verification standard 0.5 0.507 ± 0.018 µg Cr(VI)/l (2s, n = 36)
calibration verification standard 5 5.02 ± 0.25 µg Cr(VI)/l (2s, n = 37)
extraction blank < 0.1 µg Cr(VI)/l (n = 35)
extraction solution + 2 µg/l Cr(VI) 2.02 ± 0.11 µg Cr(VI)/l (2s, n = 35)
extraction solution + 100 µg/l Cr(III) < 0.1 µg Cr(VI)/l (n = 33)
<0.5 µg Cr(VI)/l (n = 3)
impregnated blank filter < 1 ng Cr(VI)/filter (n =32)
<5 ng Cr(VI)/filter (n = 12)
impregnated blank filter + 20 ng Cr(VI) 20.6 ± 5.5 ng Cr(VI)/filter (2s, n = 36)
impregnated blank filter + 1000 ng Cr(III) 4.9 ± 3.9 ng Cr(VI)/filter (2s, n = 34)
Table 2: Quality control measurements during the monitoring campaign.
The extraction showed a good recovery of the Cr(VI) spike 101 ± 5% (95% confidence level, 2s). The
oxidation of the Cr (III) spike added to the alkaline extraction was in 90% of the samples <0.1% (not
detectable) and in 10% of cases <0.5%. The impregnation of the filters was the most critical step with
respect to contamination. The 20 ng Cr (VI) spike on the filter was recovered for 103 ± 27% (95%
confidence level, 2s). The oxidation of the 1000 ng Cr (III) spike on the filter was on average 0.49 ±
0.39% (equivalent to 4.9 ng / filter). Raising the temperature of the ultrasonic assisted extraction
from 15 to 70°C resulted in a 10 fold increase in observed oxidation of the Cr(III) spike.
Field quality controls
Based on the daily duplicate sampling and analysis of sample (a) and (b) at both locations, a precision
of ± 0.26 ng Cr (VI)/m³ (2s, N = 37) could be derived for concentrations < 1 ng Cr (VI)/m³. In the range
> 1 ng Cr (VI)/m³, a relative precision of ± 24 % (2s, N = 30) was derived.
The recovery of Cr(VI) on the sampled impregnated filter (c) spiked with 20 ng Cr (VI) (equivalent to
1.43 ng Cr (VI)/m³ with the average sampling rate of 14 m³/24h) was calculated as: Recovery =
([sample c] - [mean (sample a, sample b)])/1.43 and is summarized in table 3. The average recovery,
calculated for total Cr concentrations in ambient air up to ~ 250 ng / m³, amounted 75 % (N = 87). For
total Cr concentrations in ambient air > 250 ng / m³, the rise in level of Cr(VI) concentration due to
the added Cr(VI) spike (sample c) was less than ~ 20 % of the measured Cr(VI) concentration in the
samples a and b and within the measurement uncertainty. The range of observed Cr(VI) recovery
over the two locations amounted to 33 – 103 % (10-90 % percentile) during the monitoring period.
The difference between the average recovery of the Cr(VI) spiked laboratory quality control (103 %)
and the Cr(VI) spiked field control (75 %) is in line with the literature data reported by Meng et al.
and strongly indicates the effects of ambient sampling conditions and ambient particles on the Cr(VI)
recoveries (Meng et al., 2011). For example, Fe could contribute to the reduction of Cr(VI) to Cr(III)
under typical atmospheric conditions through the following reactions:
Cr(VI) + 3Fe(II) → Cr(III) + 3Fe(III).
N mean st. dev. median 10 percentile 90 percentile

1
Recovery Cr(VI) spike (%) 87 75 39 78 33 103
2
Cr(III) oxidation (%) 58 1.7 1.2 1.4 0.4 3.1
1
calculated as Recovery = ([sample c] - [mean (sample (a),sample(b)])/1.43 for all days where [Cr] < 250 ng Cr/m³.
2
calculated as oxidation = ([sample d] - [mean (sample (a),sample(b)])/71.4 for all days where [Cr] < 100 ng Cr/m³.
Table 3: Summary of the field quality control data of the Cr(III) and Cr(VI) spiked filters.
The oxidation of Cr(III) on the sampled impregnated filter (d) spiked with 1000 ng Cr (III) (equivalent
to 71.4 ng Cr (III)/m³) was calculated as: oxidation = ([sample d] - [mean (sample a, sample b)])/71.4
and is summarized in table 3. The average oxidation (calculated for total Cr concentrations in
ambient air up to ~ 100 ng / m³) amounted 1.7 % (N = 58).
For total Cr concentrations in ambient air > 100 ng / m³, the raise in level of Cr(VI) concentration due
to the oxidation of the Cr(III) spike (sample d) was less than ~ 20 % of the measured Cr(VI)
concentration in the samples a and b, and within the measurement uncertainty. The range of
observed Cr(III) oxidation over the two locations amounted from 0.4 to 3.1 % (10-90 % percentile)
during the monitoring period. The difference between the oxidation of the non-sampled Cr(III) spiked
filter (0.49 %) and the sampled Cr(III) spiked filter (1.7%) is in accordance with literature data and
indicates the importance of the air sampled matrix (Li et al., 2002).
The stability of the Cr(VI) and the Cr(III) spike on impregnated filters observed with XANES indicates
that the extraction of the filter in combination with the sampled air matrix is likely to induce the Cr
conversions.
Monitoring data
The results of the Cr(VI) and total Cr data obtained during the monitoring campaign are represented
in figure 7 and figure 8 and are summarized in table 4.
80 800
ng Cr(VI)/m³
70 700
ng Cr/m³
60 600
50 500
ng Cr(VI)/m³
ng Cr /m³
40 400
30 300
20 200
10 100
0 0
10/okt
12/okt
14/okt
16/okt
18/okt
20/okt
22/okt
24/okt
26/okt
28/okt
30/okt
11/nov
13/nov
15/nov
17/nov
19/nov
21/nov
23/nov
25/nov
27/nov
29/nov
1/dec
3/dec
5/dec
7/dec
9/dec
4/okt
6/okt
8/okt
1/nov
3/nov
5/nov
7/nov
9/nov
Figure 7: 24 h monitoring data of Cr(VI) and total Cr in PM10 at location 1 (near anthropogenic Cr source) over
a period from 04/10/2010 to 09/12/2010.
20 200
18 180
ng Cr(VI)/m³
16 160
ng Cr/m³
14 140
12 120
ng Cr(VI)/m³
ng Cr /m³
10 100
8 80
6 60
4 40
2 20
0 0
10/okt
12/okt
14/okt
16/okt
18/okt
20/okt
22/okt
24/okt
26/okt
28/okt
30/okt
11/nov
13/nov
15/nov
17/nov
19/nov
21/nov
23/nov
25/nov
27/nov
29/nov
1/dec
3/dec
5/dec
7/dec
9/dec
4/okt
6/okt
8/okt
1/nov
3/nov
5/nov
7/nov
9/nov
Figure 8: 24 h monitoring data of Cr(VI) and total Cr in PM10 at location 2 (further from anthropogenic Cr
source) over a period from 04/10/2010 to 09/12/2010.
The Cr(VI) concentrations measured at location 1 (mean of 5.2 ng Cr(VI)/m³; median of 0.9 ng
Cr(VI)/m³) and location 2 (mean of 1.2 ng Cr(VI)/m³; median of 0.27 ng Cr(VI)/m³) are on average
above the air quality guideline value of 0.2 ng Cr(VI)/m³ in PM10 recommended as an annual average
by DEFRA (DEFRA, 2009). The mean daily Cr(VI)/Cr ratios measured at both locations were
respectively 3.5 % and 2.6 %.
10th 90th
N mean median
percentile percentile
ng Cr(VI)/m³ 52 5.2 0.9 0.09 14

Location 1 ng Cr/m³ 64 96 33 3.7 274
near anthropogenic Cr source
daily Cr(VI)/Cr ratio
52 3.5 3.0 0.9 6.7
(%)
ng Cr(VI)/m³ 28 1.2 0.27 0.07 3.5
Location 2
ng Cr/m³ 60 34 11 2.9 102
further from anthropogenic Cr
source daily Cr(VI)/Cr ratio
26 2.6 1.6 0.9 5.3
(%)
Table 4: Summary of the monitoring data of Cr(VI) and total Cr in PM10 at the two locations.
Highest concentrations of total Cr and Cr(VI) were observed when the wind was blowing from the
industrial plant towards the monitoring sites. As expected, the highest concentrations were
measured on the location close to the plant (location 1). However, in the light of the range of
observed Cr(III) oxidation over the two locations (0.4 - 3.1 %), these results have to be interpreted
carefully. It is assumed that the measured Cr(VI) is a combination of method-induced Cr(III) oxidation
and actual Cr(VI) present in the PM10 fraction. Cr(III) oxidation depends on many factors and may be
induced in the air, on the filter and during the extraction. Basically, only the oxidation due to capture
of the PM10 fraction on the filter and due to the extraction are considered as method-induced
artefacts. However, this method-induced oxidation sets limits to the Cr(VI) detection relative to the
total Cr content of the particulate matter.
Correction for the Cr(III) conversion is ambiguous because of the expected difference in oxidation
between the added Cr(III) spike and the Cr(III) present in the sample (most likely Cr spinel) (Meng et
al., 2011). With an average oxidation of 1.7 %, this would mean that for total Cr concentrations
above ~ 10 ng Cr/m³, an air quality guideline value of 0.2 ng Cr(VI)/m³ in PM10 would fall within the
uncertainty on the Cr(VI) determination method. Daily PM10 ambient air quality measurements by
the Flemish Environmental Agency on 20 monitoring stations close to industrial areas in 2009 show
that median values are in 80 % of the monitoring stations between 1.2 – 8.3 ng Cr/m³; However, in
20 % of the monitoring stations, median Cr concentrations in PM10 > 10 ng Cr/m³ were measured
(VMM, 2009). The level of Cr(VI)-Cr(III) interconversions observed at one location may not be
extrapolated to another, due to potential differences in Cr sources in combination with differences in
air matrix. With respect to the interactions between Cr-Mn as possible oxidative pathway, it can be
mentioned that the Mn concentration measured in PM10 at both locations fluctuated in a same way
than the levels of Cr. At the highest Cr concentrations in PM10 observed during the monitoring
campaign, the measured Mn concentrations were between 1/2 to 1/3 of the Cr concentration.
The air quality guideline value of 0.2 ng Cr(VI)/m³ in PM10 proposed by DEFRA was not derived from
epidemiological studies at Cr(VI) ambient concentration levels, but derived by dividing the Lowest
Observed Adverse Effect Level over a 40 year working lifetime by a factor of 1000 (DEFRA, 2009).
Within this process of deriving toxicity based guideline values, the accurate measurement of Cr(VI) is
crucial. Any enforcement of Cr(VI) regulatory values for ambient air in the future will require
supplementary efforts with respect to method standardization.
Conclusion and practical implications

The described monitoring method for Cr(VI) in ambient air via impregnated filters is sensitive
(method detection limit of 0.015 ng Cr(VI)/m³) and reproducible (precision of the method ~25 %).
The Cr(VI) recoveries (average 75 %) strongly indicate the effects of ambient sampling conditions and
ambient particles on the Cr(VI) recoveries. The method – and in more general terms all methods via
extraction – is susceptible to method-induced Cr(III) oxidation. Knowledge of the method-induced
oxidation is particularly relevant for the determination of Cr(VI) in ambient particulate matter as the
level of observed oxidation (average 1.7 %) is within the range of the level of measured Cr(VI)
(average daily Cr(VI)/Cr ratios of 3.5 % respectively 2.6 %). For Cr concentrations in PM10 > 10 ng
Cr/m³, this method-induced oxidation could lead to false positive exceeding of the air quality
guideline value of 0.2 ng Cr(VI)/m³ in PM10.
The stability of the Cr(VI) and the Cr(III) spike on impregnated filters observed with XANES indicates
that the alkaline extraction of the filter in combination with the sampled air matrix is likely to induce
the Cr conversions.
In order to assess the toxicity of Cr containing PM10 in general, we encourage that new research and
monitoring programs in the field of determination of the Cr(VI) content in ambient particulate matter
would report on Cr(VI) and Cr(III) interconversions.
Acknowledgements
We wish to thank Wim Bras and Sergey Nikitenko for assistance at the ESRF DUBBLE beamline. We thank
Annick Cluyts, Ellen Poelmans, Karlien Duyssens, Wilfried Brusten, Jef Daems, Jo Van Laer, Bart Noten for aid in
the experimental and sampling work. We thank the Flemish Environment Agency for authorization to sample at
their monitoring site. The monitoring campaign was commissioned, financed and steered by the Flemish
Environment Agency. G. Silversmit is supported by a postdoctoral fellowship from the Research Foundation-
Flanders (FWO-Vlaanderen, Belgium). FWO-NWO is thanked for making the beamtime at the DUBBLE beamline
available.
References paper, Determination of hexavalent chromium in ambient air: a story of method-

induced Cr(III) oxidation?, Atmospheric Environment, 2011, 45, 5332-5341.
1. Ashley, K., Howe, A.M., Demange, M., Nygren, O., 2003. Sampling and analysis considerations for
the determination of hexavalent chromium in workplace air. Journal of Environmental
Monitoring 5, 707-716.
2. Ashley, K, Applegate, G.T., Marcy, A.D., Drake, P.L., Pierce, P.A., Carabin, N. and Demange M.,
2009, Evaluation of sequential extraction procedures for soluble and insoluble hexavalent
chromium compounds in workplace air samples, Journal of Environmental Monitoring 11, 318-
325.
3. Bell, R.W., Hipfner, J.C., 1997. Airborne hexavalent chromium in Southwestern Ontario. Journal
of the Air & Waste Management Association 47, 905-910.
4. Boiano, J.M., Wallace, M.E., Sieber, W.K., Groff, J.H., Wang, J., Ashley, K., 2000. Comparison of
three sampling and analytical methods for the determination of airborne hexavalent chromium.
Journal of Environmental Monitoring 2, 329-333.
5. Borai, E.H., El-Sofany, E.A., Abdel-Halim, A.S., 2002. Speciation of hexavalent chromium in
atmospheric particulate samples by selective extraction and ion chromatographic determination.
Trends in Analytical Chemistry 21, 741-745.
6. Caglieiri, A., Goldoni, M., Acampa, O., Andreoli, R., Vettori, M.V., et al. 2006. The effect of inhaled
chromium on different exhaled breath condensate biomarkers among chrome-plating workers.
Environmental Health Perspectives 114 (4), 542-546.
7. Christensen, J.M., Byrialsen, K., Vercoutere, K., Cornelis, R., Quevauviller, Ph., 1999. Certification
of Cr(VI) and total leachable Cr contents in welding dust loaded on a filter (CRM 545). Fresenius
Journal of Analytical Chemistry 363, 28-32.
8. Cleven, R.F.M.J., de Boer, J.L.M., 1992. Chapter 4 – Characterization of the Cr(VI)/Cr(III) ratio in
aerosols. In: Environmental analytical and physical chemistry series – Environmental particles,
Buffle, J. & van Leeuwen, H.P. (eds.), vol. 1, chapter 4, pp. 145-186 (ISBN 0-87371-589-6).
9. Department for Environment, Food and Rural Affairs (DEFRA), Scottish Executive, National
Assembly of Wales, Department of the Environment in Northern Ireland (2009), guidelines for
metals and metalloids in ambient air for the protection of human health, PB13263,
http://www.lungsatwork.org.uk/pubsimg/Metals_and_Metalloids.pdf.
10. Dionex application Update 179, Sensitive determination of hexavalent chromium in drinking
water, http://www.dionex.com/en-us/webdocs/110230-AU179-IC-Hexavalent-Chromium-
DrinkingWater-04Mar2011-LPN2772.pdf.
11. Dyg, S., Anglov, T., Christensen, J.M., 1994. Analytica Chimica Acta 286, 273-282.
12. Eastern Research Group (ERG), Swift, J., Howell, M., Tedder, D., 2006. Collection and analysis of
hexavalent chromium in ambient air, Morrisville, USA.
13. Finley, B., Fehling, K., Falerios, M., Paustenbach, D., 1993. Field validation for sampling and
analysis of airborne hexavalent chromium. Applied Occupational and Environmental Hygiene 8
(3), 191-200.
14. Galbreath, K.C., Zygarlicke, C.J., 2004. Formation and chemical speciation of arsenic-, chromium-,
and nickel-bearing coal combustion PM2.5. Fuel Processing Technology 85, 701-726.
15. Galbreath, K.C., Zygarlicke, C.J., Toman, D.L., Huggins, F.E., Huffman, G.P., 1994. Nickel and
chromium speciation of residual oil combustion ash. Combustion Science and Technology 134,
243-262.
16. Goldoni, M., Caglieiri, A., Poli, D., Vettori, M.V., Corradi, M., et al. 2006. Determination of
hexavalent chromium in exhaled breath condensate and environmental air among chrome
plating workers. Analytica Chimica Acta 562, 229-235.
17. Goodzari, F., Huggins, F.E., Sanei, H., 2008. Assessment of elements, speciation of As, Cr, Ni and
emitted Hg for a Canadian power plant burning bituminous coal. International Journal of Coal
Geology 74, 1-12.
18. Gray, C.N., Goldstone, A., Dare, P.R.M., Hewitt, P.J., 1983. The evolution of hexavalent chromium
in metallic aerosols. American Industrial Hygiene Association 44 (6), 384-388.
19. Grohse, P.M., Gutknecht, W.F., Hodson, L., Wilson, B.M., 1988. The fate of hexavalent chromium
in the atmosphere. Research Triangle Institute, California.
20. H.M. Kingston, U.S. Patent 5,414,259, 1995.
21. Hagendorfer, H., Uhl, M., 2007. Speciation of chromium in particulate matter (PM10):
development of a routine procedure, impact of transport and toxicological relevance,
Umwelbundesamt, Vienna, REP-0111, 29 p.
22. Huggins, F.E., Huffman, G.P., Linak, W.P., Miller, C.A., 2004. Quantifying hazardous species in
particulate matter derived from fossil fuel combustion. Environmental Science & Technology 38
(6), 1836-1842.
23. Huggins, F.E., Huffman, G.P., Robertson, J.D., 2000. Speciation of elements in NIST particulate
matter SRMs 1648 and 1650. Journal of Hazardous Materials 74, 1 -23.
24. Huo, D., Yusheng, L., Kingston, H.M., 1998. Determination and correction of analytical biases and
study on chemical mechanisms in the analysis of Cr(VI) in soil samples using EPA protocols.
Environmental Science & Technology 32, 3418-3423.
25. Isakov, V., Touma, J.S., Khlystov, A., 2007. A method of assessing air toxics concentrations in
urban areas using mobile platform measurements. Journal of the Air & Waste Management
Association 57, 1286-1295.
26. James, B.R., Petura, J.C., Vitale, R.J., Mussoline, G.R., 1995. Hexavalent chromium extraction from
soils: a comparison of five methods. Environmental Science & Technology 29 (9), 2377-2381.
27. Khlystov, A., Ma, Y., 2006. An on-line instrument for mobile measurements of the spatial
variability of hexavalent and trivalent chromium in urban air. Atmospheric Environment 40,
8088-8093.
28. Kingston, H.M. “Skip”, Hou, D., Lu, Y., Chalk, S., 1998. Accuracy in species analysis: speciated
isotope dilution mass spectrometry (SIDMS) exemplified by the evaluation of chromium species.
Spectrochimica Acta Part B 53 (2), 299-309.
29. Krystek, P., Ritsema, R., 2007. Monitoring of chromium species and 11 selected metals in
emission and immission of airborne environment. International Journal of Mass Spectrometry
265, 23-29.
30. Li, Y., Pradhan, N.K., Foley, R., Low, G.K.C., 2002. Selective determination of airborne hexavalent
chromium using inductively coupled plasma mass spectrometry. Talanta 57, 1143-1153.
31. Meng, Q., Fan, Z., Buckley, B., Lin, L., Huang, L., et al. 2011. Development and evaluation of a
method for hexavalent chromium in ambient air using IC-ICP-MS. Atmospheric Environment 45,
2021-2027.
32. Mennen, M.G., Koot, W., Putten, E.M., Ritsma, R., Piso, S., Knol, T., Fortezza, F., Kliest, J.J.G.,
1998. Hexavalent chromium in ambient air in the Netherlands. Results of measurements near
word preservation plants and at a regional site. RIVM report no. 723101031.
33. Metze, D., Herzog, H., Gosciniak, B., Gladtke, D., Jakubowski, N., 2004. Determination of Cr(VI) in
ambient airborne particulate matter by a species-preserving scrubber-sampling technique.
Analytical and Bioanalytical Chemistry 378, 123-128.
34. Nico, P.S., Kumfer, B.M., Kennedy, I.M., Anastasio, C., 2009. Redox dynamics of mixed metal (Mn,
Cr, and Fe) ultrafine particles. Aerosol Science and Technology 43, 60-70.
35. Nico, P.S., Zasoki, R.J., 2000. Importance of Mn(III) availability on the rate of Cr(III) oxidation on
δ-MnO². Environmental Science & Technology 34 (16), 3363-3367.

36. Nikitenko, S., Beale, A.M., van der Eerden, A.M.J., Jacques, S.D.M., Leynaud, O., et al. 2008.
Implementation of a combined SAX/WAXS/QEXAFS set-up for time-resolved in situ experiments.
Journal of Synchrotron Radiation 15, 632-640.
37. Nusko, R., Heumann, K.G., 1997. Cr(III)/Cr(VI) speciation in aerosol particles by extractive
separation and thermal ionization isotope dilution mass spectrometry. Fresenius Journal
Analytical Chemistry 357, 1050-1055.
38. Pettine, M., Capri, S., 2005. Digestion treatments and risks of Cr(III)-Cr(VI) interconversions
during Cr(VI) determination in soils and sediments – a review. Analytica Chimica Acta 540, 231-
238.
39. Samanta, G., Boring, C.B., Dasgupta, P.K., 2001. Continuous automated measurement of
hexavalent chromium in airborne particulate matter. Analytical Chemistry 73, 2034-2040.
40. Sheehan, P., Ricks, R., Ripple, S., Paustenbach, D., 1992. American Industrial Hygiene Association
53 (1), 57-68.
41. SOP CARB 039, Standard Operating Procedure for the analysis of hexavalent chromium at
ambient atmospheric levels by Ion Chromatography, 1993 (revision 2002, 2006), California
Environmental Protection Agency, California Air Resources Board.
42. Steinsberger, S., DeWees, W.G., Bell, A.C., Finke, R.A., Knoll, J.E., Midgett, M.R., 1994.
Development and validation of the US EPA hexavalent chromium source emissions measurement
method. Fuel Processing Technology 39, 265-276.
43. Stern, A.H., Yu, C.H., Black, K., Lin, L., Lioy, P.J., et al. 2010. Hexavalent chromium in house dust –
a comparison between an area with historic contamination from chromate production and
background locations. Science of the Total Environment 408, 4993-4998.
44. Stern, R.M., 1988. Indicators of matrix dependent variation in effective dose for chromium and
nickel aerosol exposures. The Science of the Total Environment 71, 301-307.
45. Swietlik, R., Molik, A., Molenda, M., Trojanowska, M., Siwiec, J., 2011. Chromium(III/VI)
speciation in urban aerosol. Atmospheric Environment 45, 1364-1368.
46. Talebi, S.M., 2003. Determination of total and hexavalent chromium concentrations in the
atmosphere of the city of Isfahan. Environmental Research 92, 54-56.
47. Tirez, K., Brusten, W., Cluyts, A., Patyn, J., De Brucker, N., 2003. Determination of hexavalent
chromium by species-specific isotope dilution mass spectrometry and ion chromatography-1,5-
diphenylcarbazide spectrophotometry. Journal of Analytical Atomic Spectrometry 18 (8), 922-
932.
48. Tirez, K., Scharf, H., Calzolari, D., Cleven, R., Kisser, M., Lueck, D., 2007. Validation of a European
standard for the determination of hexavalent chromium in solid material. Journal of
Environmental Monitoring 9, 749–759.
49. Unceta, N., Séby, F., Malherbe, J., Donard, O.F.X., 2010. Chromium speciation in solid matrices
and regulation: a review. Analytical and Bioanalytical Chemistry 397, 1097-1111.
50. Vanhoof, C., Chen, H., Berghmans, P., Corthouts, V., De Brucker, N., Tirez, K., 2003. A risk
assessment study of heavy metals in ambient air by WD-XRF spectrometry using aerosol-
generated filter standards. X-Ray Spectrometry 32 (2), 129–138.
51. Vlaamse Milieumaatschappij (VMM, Flemish Environment Agency), Annual immission monitoring
network, calendar year 2009, http://www.vmm.be/pub/jaarverslag-luchtkwaliteit-in-het-
vlaamse-gewest-2009/luchtkwaliteit_rapport_2009_TW.pdf.
52. Wang, J., Ashley, K., Kennedy, E.R., Neumeister, C., 1997. Determination of hexavalent chromium
in industrial hygiene samples using ultrasonic extraction and flow injection analysis. Analyst 122,
1307-1312.
53. Wang, Y., Li, A., Zhan, Y., Wei, L., Li, Y., et al. 2007. Speciation of elements in atmospheric
particulate matter by XANES. Journal of Radioanalytical and Nuclear Chemistry 273 (1), 247-251.
54. Werner, M., Nico, P., Guo, B., Kennedy, I., Anastasio, C., 2006. Laboratory study of simulated
atmospheric transformations of chromium in ultrafine combustion aerosol particles. Aerosol
Science and Technology 40, 545-556.
55. Werner, M.L., Nico, P.S., Marcus, M.A., Anastasio, C., 2007. Use of micro-XANES to speciate
chromium in airborne fine particles in the Sacramento Valley. Environmental Science &
Technology 41 (14), 4919-4924.
56. Zatka, V.J., 1985. Speciation of hexavalent chromium in welding fumes interference by air
oxidation of chromium. American Industrial Hygiene Association 46 (6), 327-331.
3.4. OXYHALIDE BY-PRODUCTS IN DRINKING WATER DISINFECTION
The first documented drinking water treatment can be found in Egyptian hieroglyphics, describing
procedures to purify water [116]. The basic principles were the same then as they are today: boiling,
chemical treatment and filtration. The importance of drinking water quality and its influence on
human health was known, but the specific contaminations would not be identified for centuries. This
situation was changed in the 19th century, when chlorine was introduced as a chemical disinfectant
for water treatment. The introduction of chlorination to drinking water was followed by remarkable
reductions in cholera, dysentery and typhoid worldwide.
Nowadays, water treatment by disinfection of drinking water is generally considered a major public
health achievement of the 20th century. Consequently, the identification of water contaminants
shifted from microbiological to chemical. The number of chemicals determined in drinking water has
grown exponentially. However, from the hundreds of them present, only very few have been studied
or have documented proof of their health effects.
Nearly half of the parameters monitored are being measured for operational reasons (e.g., iron,
ammonium, pH, chloride, dissolved organic carbon) and/or for reasons of customer satisfaction (e.g.,
colour, taste, hardness). Of the health-related compounds, a number of metals and small groups of
organic compounds and pesticides are being measured on a regular base in a majority of countries.
During the 1970s, it was discovered that chlorination of drinking water produced carcinogens, such
as trihalomethanes, haloacetic acids and other. Since then, environmental regulatory agencies, as
well as drinking water treatment technologists, have been carrying out extensive research for
alternative disinfection methods that minimize the production of by-products with significant health
risks. Ozonation has emerged as one of the most promising alternatives to chlorination.
In the last decade, the use of ozone in the treatment of drinking water to improve taste, odour and
reduce the presence of organic micropollutants has been spreading. In the early 1980s, it became
obvious that application of ozonation in drinking water treatment did not only result in the formation
of oxygenated compounds but, in bromide-containing water, brominated organic compounds as well
as bromate are formed.
Bromate is the most important inorganic oxyhalide by-product, the concentration of which has to be
controlled in drinking water. Bromate is formed when water containing bromide is ozonated. From
the theoretical and practical point of view, it can be seen that bromate formation can be influenced
by many parameters, such as ozone dose, water pH, temperature and indigenous concentration of
bromide. Furthermore, other subjects of interest and topics of advanced research are chlorite and
chlorate [117].
Bromate has been identified as an animal and possible human carcinogen. The International Agency
for Research on Cancer (IARC) classified bromate in group B-2 (the agent is possibly carcinogenic to
humans). The United States Environmental Protection Agency (US EPA), as well as the Commission of
the European Communities, has issued rules that require public water supplies to control previously
unregulated microorganisms and cancer-causing disinfection by-products in final treatment drinking
water [11]. According to these regulations, Maximum Admissible Level (MAL) is 10 μg/L for bromate
and 1000 μg/L for chlorite.
The maximum admissible level for bromate has been primarily based on current analytical capability
(not on toxicological considerations - the target concentration for bromate in drinking water is zero);
thus, there is a need for on-going development and refinement of analytical technologies in order to
permit rapid and reliable determinations at the sub-μg/L level.
Global and national agencies are continually striving to monitor bromate, chlorite and chlorate levels
in drinking water to establish appropriate regulatory limits. In Flanders, the control of drinking water
quality is organized in 81 different drinking water supply zones [118]. A drinking water supply zone is
a geographically defined area within which water intended for human consumption comes from one
or more sources and can be considered as being approximately of uniform quality.
Figure 34: Overview of the 81 drinking water supply zones defined in Flanders [118].
For the 81 different drinking water supply zones in Flanders, approximately 1400 bromate control
analyses of water collected from the tap are performed on a yearly basis in the audit monitoring
program imposed by the European Directive [118]. For the chemical parameters, most exceedings of
maximum admissible concentration in drinking water are observed for lead (69 out of 7900 analysis,
year 2010). In general, the drinking water distributed in Flanders largely complies with the quality
requirements set by the European Directive.
In support of routine determination of bromate in regulatory monitoring programs, the benefits of a

user-friendly and low cost low pressure LC/ICP-MS set-up as compared to the more traditional
HPLC/ICP-MS hyphenation was studied (see paper: Determination of bromate in drinking waters
using low pressure liquid chromatography / ICP-MS, J. Anal. At. Spectrom., 2013).
Determination of bromate in drinking waters using low pressure liquid

chromatography / ICP-MS
Kristof Tirez*a,b, Wilfried Brustena, Filip Beutelsa, Mai Weversa and Frank Vanhaeckeb
a
b
Ghent University, Department of Analytical Chemistry, Belgium
Journal of Analytical Atomic Spectrometry, 2013, 28, 1894 – 1902.

Abstract
This paper describes a user-friendly method for bromate determination that can be implemented
easily on any inductively coupled plasma - mass spectrometer present in drinking water laboratories.
The method is applying low pressure liquid chromatography coupled to an ICP – quadrupole mass
spectrometry instrument (ICP-QMS) or an ICP – sector field mass spectrometry instrument (ICP-SFMS)
and is compared to that relying on high performance liquid chromatography (HPLC) coupled to an
ICP-QMS instrument. The low pressure LC/ICP-MS method uses a low-pressure delivery six-port valve
and a 5 cm anion exchange column, which allows a fully resolved separation of bromate in 13 min
and achieves a limit of quantification of 0.2 µg bromate L-1. The low pressure LC system is small, easy
to install and its operation is fully integrated within the ICP-MS software. The method allows fit-for-
purpose assessment of bromate, potentially present as a Br-containing disinfection by-product in
drinking water, and meets all performance characteristic requirements set by the European Council
for the monitoring of the quality of water intended for human consumption. A median bromate
concentration of 0.5 µg L-1 was obtained for 80 tap water samples collected during regulatory
monitoring campaigns from 2009 until 2012 and covering different water supply areas in the Flemish
region of Belgium.
Introduction
Water treatment by disinfection processes is considered a major public health achievement of the
twentieth century.1 In the 1970s, however, it was discovered that chlorination of drinking water
produced carcinogens, such as trihalomethanes and haloacetic acids.2 Since 1974, more than 500
disinfection by-products have been identified in drinking water. Since that time, environmental
regulatory agencies and drinking water treatment technologists have performed extensive research
on alternative disinfection methods that reduce the generation of organic by-products with
significant health risks.
Ozonation has emerged as one of the most promising alternatives to chlorination.1 In the early
1980s, it became obvious that the application of ozonation in drinking water treatment not only
resulted in the formation of oxygenated compounds, but also in the formation of brominated organic
compounds and bromate. Bromate has been identified as an animal, and possibly human,
carcinogen.3,4 The International Agency for Research on Cancer (IARC) has classified bromate in group
B-2 (the agent is possibly carcinogenic for humans).
A Maximum Admissible Concentration (MAC) of 10 µg L−1 bromate in drinking waters is

recommended by the US EPA,5 the Council of the European Union (in Directive 98/83)6 and the
WHO7.
As the European Council recommends limits of detection (LoDs) for the determination of bromate in
drinking water of less than 2.5 µg L−1, this has necessitated the further development of more
sensitive and/or alternative analysis approaches in the past 20 years. Most techniques currently
available for bromate determination use ion chromatography (IC) as the underlying separation
mechanism.3,4 The methods for bromate determination using IC can be generally divided into: (a)
direct methods (suppressed conductivity detection)8, (b) indirect methods (UV/Vis detection after
post-column derivatization)9 and (c) hyphenated techniques (ICP–MS and MS detection)10-14.
For all of the determination methods mentioned above, standardised methods that can achieve LoDs
of less than 2.5 µg L−1 are available nowadays. For a critical comparison of available ISO, US EPA and
other methods concerning ion chromatography determination of bromate, we refer to Hautman et
al. and Michalski.1,2
The selection of the method used for bromate determination is guided by, among other, the
availability of appropriate equipment, ease of use and expenditures. For the determination of trace
elements in drinking water, most - if not all - of the drinking water laboratories are nowadays
equipped with ICP-MS instrumentation. In spite of the excellent sensitivity, wide linear dynamic
range and multi-element capability of ICP-MS, for hyphenated techniques such as HPLC/ICP-MS, the
ease of use and expenditures are limiting factors for common use in routine drinking water
laboratories. With the introduction of low pressure chromatography, a fast analytical method for
speciation with ICP-MS detection becomes more feasible.15 This low pressure LC system is small, easy
to install and its operation is fully integrated within the ICP-MS software. Moreover, as the low
pressure LC system makes use of an autosampler commonly used for elemental analysis, the system
can be implemented easily on any inductively coupled plasma - mass spectrometer present in
drinking water laboratories.
The work in this paper describes and compares the capabilities and limitations of HPLC/ICP-MS and
low pressure LC coupled to an ICP – quadrupole mass spectrometry instrument (ICP-QMS) and
coupled to an ICP – sector field mass spectrometry instrument (ICP-SFMS) for the determination of
bromate in drinking water.
Experimental section
Instrumentation
HPLC/ICP-MS
Samples were analysed using a quadrupole-based ICP-MS instrument (Elan 6000, Perkin-Elmer,
SCIEX, ON, Canada), equipped with an MCN-100 micro-concentric nebuliser (Cetac technologies, NE,
USA) fitted onto a cyclonic spray chamber for sample introduction. During optimization of the
instrument settings, solution was fed to the MCN-100 by means of a peristaltic pump. The speciation
was carried out with a 100x2 mm P1 AAD 1 anion exchange DMEA microbore column (Institut für
Anorganische Chemie, Universität Hannover, Germany). The separation column consisted of a high-
capacity and high-performance PS/DVB-anion-exchanger functionalized with 2-(dimethylamino)-
ethanol, which has been developed by Seubert et al..16 The eluent was pushed through the column at
a flow rate of 400 µL min-1 via a series 10 HPLC pump (Perkin-Elmer). A FIAS 400 (Perkin-Elmer)
peristaltic pump was used for filling the 1000 µL sample loop. The sample loop was connected to an
automated 6-port valve (LabPRO 6, Rheodyne, CA, USA). The operation of the entire measurement
system, including sample uptake, filling of the loop, injection onto the column and ICP-QMS
measurement, is computerized. Integration of the chromatograms (peak areas) was accomplished
with Turbochrom software (Perkin-Elmer). Before analysis, the instrument was preconditioned by
aspirating the eluent for at least 1 hour.
Low pressure LC/ICP-QMS
Samples were analysed using a quadrupole-based ICP-MS instrument (Nexion 300S, Perkin Elmer),
equipped with a µ-Flow nebulizer (PFA-400, Elemental Scientific Inc or ESI, NE, USA) coupled to a 50
mL baffled cyclonic spray chamber (Perkin Elmer). The speciation was carried out using a SC-DX
chromFAST system (ESI) using a 50 x 4 mm anion exchange column (CF-Se-01, ESI).
The low pressure mobile phase was delivered to the column using an integrated FAST Valve and
Precision Micro Peripump (FAST DXi) (see figure 1, right). The FAST system consists of a six-port
injection valve and a vacuum pump connected to a 2500 µL sample loop. The adjustable injection
time and pump rate control the amount of sample injected from the 2500 µL sample loop. The flow
of eluent was delivered by a peristaltic pump (controlled by the ICP-MS software) using green/orange
flared tubing (i.d. 0.38 mm). The operation of the entire measurement system, including sample
uptake, filling of the loop, injection onto the column and ICP-MS measurement, is computerized.
Integration of the chromatograms (peak areas) was accomplished with periSPEC Peak Area Finder
Software (ESI).
Low pressure LC/ICP-SFMS
Samples were analysed using a single-collector double-focusing sector field ICP-MS instrument
(ELEMENT II, ThermoScientific, Germany), equipped with a µ-Flow nebulizer coupled to a 50 mL

baffled cyclonic spray chamber (Elemental Scientific Inc or ESI, NE, USA). The speciation was carried
out using a SC-DX chromFAST system (ESI) using a 50 x 4 mm anion exchange column (CF-Cr-01, ESI).
The low pressure mobile phase was delivered to the column using the peristaltic pump of the ICP-
SFMS unit (see figure 1, left).
Peristaltic pump Spray chamber Low pressure LC column
Figure 1: Low pressure LC/ICP-MS system: (left) ESI FAST system (in frame) consisting of a six-port injection
valve and a vacuum pump connected to a sample loop; the system is installed on the sample holder of the ICP-
SFMS instrument (ELEMENT II, ThermoScientific); the flow of the eluent is delivered by the peristaltic pump
(controlled by the ICP-SFMS software); (right) FAST Valve and Precision Micro Peripump (FAST DXi) integrated
in the ICP-QMS instrument (Nexion 300S).
For the preparation of all solutions, ultra-pure water with a resistivity of 18 MΩ cm-1 obtained from a
Milli-Q water purification system (Millipore, MA, USA) was used.
Stock standard solutions were prepared from 1 g L-1 bromide and bromate standard solutions
purchased from Merck (Darmstadt, Germany). A quality control standard solution of 5 µg L-1 BrO3-
was prepared from KBrO3 purchased from Spex CertiPrep Inc. (New Jersey, USA). Bromoacetic acid
was purchased from Merck, dibromoacetic acid and bromochloroacetic acid were purchased from
Sigma Aldrich (Steinheim, Germany).
The eluent was prepared by dissolving NH4NO3 (Merck) in Milli-Q water. In case of low pressure
LC/ICP-QMS, Ge (Spex CertiPrep Inc., NJ, USA) was added to the eluent (final concentration: 5 µg L-1)
to optimise the ICP-QMS instrument settings and to monitor signal drift during measurement. In case
of low pressure LC/ICP-SFMS, Cs (Spex CertiPrep NJ, USA) was added to the eluent (final
concentration: 10 µg L-1).
Sample treatment
The samples were filtered through a 0.45 µm membrane filter (Schleicher & Schuell, Germany) to
remove suspended solids and stored in the dark at 4 °C until analysis.
Results and Discussion
HPLC/ICP-QMS
The HPLC/ICP-QMS optimisation was performed in 1999 and the protocol involved the use of a
quadrupole-based ICP-MS instrument. The efforts to set up an HPLC/ICP-QMS system at that time
were mainly devoted to the automation and a proper communication interface between the sample
injection system, HPLC pump and ICP-QMS instrument.17
Optimization of ICP-QMS
Because the quadrupole-based ICP-QMS instrument used at that time (Elan 6000) was not equipped
with a collision/reaction cell to remove interferences poly-atomic interferences, an evaluation of the
potentially occurring interferences was necessary. The bromide isotopes at masses 79 and 81 have
approximately the same relative abundance, but the signal of the latter suffers more from
interference by a poly-atomic argon-containing ion. This interference results in an increase of the
background signal; under the given instrumental settings, the height of the baseline in the
chromatograms is: 79Br (250 cps: 40Ar38ArH+) < 81Br (14500 cps: 40Ar40ArH+). Although the baseline
levels are stable during measurement, to a certain extent they generate a higher noise, resulting in
less accurate results and a deterioration of the LoD. When selecting (a) suitable mass-to-charge
ratio(s) (m/z) for monitoring, also the possibility of matrix-borne spectral interferences needs to be
taken into account. In this case however, these interferences can be circumvented by adjusting
eluent concentration and flow rate of the IC separation. Considering both the background levels and
the isotopic abundances, 79Br was selected for quantitative measurements.
Optimisation of HPLC conditions
Ammonium nitrate, which has the same elemental composition as nitric acid, is considered an
optimal eluent in HPLC/ICP-QMS due to its following features: thermal volatility, hence no significant
salt deposition on the cones; no extra background interferences compared to HNO3 in ICP-QMS; high
concentration tolerance by ICP-QMS without causing pronounced non-spectral interference;
possibility to adjust its pH value from acidic to slightly alkaline without impairing the ionic strength;
no complexing capacity of NH4+ with anions or NO3- with metals and no potential precipitation of
analytes (unlike a carbonate mobile phase, which can induce the precipitation of CaCO3).
The chromatographic anion exchange column offers the possibility of compressing bromate into a
smaller zone at the top of the column when the sample is transferred from the sample loop onto the
separation column. This can be seen in figure 2, where prior matrix matching of the sample with 100
mM NH4NO3 leads to peak broadening. The peak area is equal in both cases but the peak shape is
superior when the samples are injected without addition of eluent. This effect has been referred to
by Nowak et al. as the relaunch effect.11 During the transfer of bromate and the other matrix anions
from the sample loop onto the analytical column, the top of the column is mainly converted into a Cl-
or SO42- form (major anions in drinking water). Because the anions have a much lower affinity to
quaternary ammonium functional groups than nitrate (present in the eluent), the bromate is
compressed into a smaller zone at the column top than it would have been when the column top was
in the NO3- form. Therefore, the resulting bromate peak in water has a much better peak shape than
in the case of addition of NH4NO3 eluent to the sample, because in this case the column top would be
in NO3- form.
BrO3-
7.0
6.0
Signal intensity
5.0
4.0
3.0
2.0
1.0 (b)
-0.0
(a)
0 1 2 3 4 5 6 7
min 8 9 10 11 12 13 14
-1
Figure 2: Compressing effect at the top of the column, measurement with HPLC/ICP-QMS of (a, blue) 10 µg L
- -1 - -1
of BrO3 in water and (b, red) 10 µg L of BrO3 in 100 mM NH4NO3 solution (flow rate 400 µL min , eluent 100
mM NH4NO3).
Analytical characteristics and interferences
The limit of quantification (LoQ) was calculated based on 7 consecutive measurements of a 100 ng L-1
bromate and 100 ng L-1 bromide standard solution. The LoQ calculated as six times the standard
deviation on the 7 replicates amounted to 40 ng bromate L-1 and 100 ng bromide L-1. The peak of
monobromoacetic acid, which can occur as another brominated disinfection by-product, overlaps
slightly with the bromide peak, but is well separated from bromate, the compound of interest. This is
shown in figure 3 and is in agreement with the data reported by Nowak and Seubert, using the same
micobore column.11
BrO3- Br-
35
30
25
Signal intensity
20
15
C2H3BrO2
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
min
-1 - -1 - -1
Figure 3: Standard solution containing 10 µg L of BrO3 , 10 µg L of Br and 10 µg L of C2H3BrO2
-1
(monobromoacetic acid) measured with HPLC/ICP-QMS (flow rate 400 µL min , eluent 100 mM NH4NO3).
Low pressure LC/ICP-SFMS and LC/ICP-QMS
The low pressure LC method uses a commercially available low-pressure flow injection switching
valve system that is easy to use, is easily integrated and has compatible software that works well
with both the ICP-MS and autosampler. The low pressure LC system is small and can be easily
installed between the ICP-MS and the autosampler, or in case of the ICP-QMS instrument used in this
work, even integrated in the instrument (see figure 1, right). A comparative study for the
determination of bromate in drinking water was made between the low pressure LC system coupled
to a sector field ICP-MS unit (ICP-SFMS, ELEMENT II) on the one hand and a quadrupole-based
instrument (ICP-QMS, Nexion 300S) on the other.
Optimisation of low pressure LC conditions
For optimization, the same mobile phase as described above for the HPLC/ICP-QMS protocol was
used and the flow rate was adapted to approximately 170 µL min-1 (10 rotations per minute on the
peristaltic pump). A fixed 100 µL sample loop was connected to the six-port injection valve. A low
pressure LC/ICP-SFMS chromatogram is shown in figure 4. A calibration line up to 50 µg L-1 of BrO3-
and Br- was constructed and it was observed that the sensitivity (slope) expressed as 79Br-signal
versus µg Br L-1 is ~ 15 % less for 79Br-bromide as compared to 79Br-bromate and this for both the ICP-
QMS and the ICP-SFMS instrument. This is assumed to be an effect induced by the NH4NO3 eluent in
combination with the nebulizer system (µ-Flow nebulizer coupled to a baffled cyclonic spray
chamber). This difference was also observed when bromide and bromate standard solutions
(obtained from different suppliers) were analysed directly without LC. The matrix of the eluent in
combination with the nebulisation system have been shown to have an effect on 79Br+ signal.18,19
Kahen et al. hypothesized that gaseous HBr(g) formed during nebulisation diffuses away from the
droplets and particles and can be removed from the injector gas stream by dissolution in the
condensed low temperature eluent on the walls and subsequent drainage.18 Difference in sensitivity
between As species has also been observed when using desolvation systems as part of the
nebulisation system in HPLC/ICP-QMS.20-22,25 For accurate measurements, independent calibrations
for each Br species are needed.
1,5E+05
1,4E+05
1,3E+05
1,2E+05
BrO3- Br-
1,1E+05
cps 79 Br
1,0E+05
9,0E+04
8,0E+04
7,0E+04
6,0E+04
5,0E+04
0 1 2 3 4 5 6 7 8
min
-1 - -1 -
Figure 4: Standard solution containing 2.5 µg L of BrO3 and 2.5 µg L of Br measured with low pressure
79 -1
LC/ICP-SFMS ( Br, 100 µL sample loop, flow rate of 170 µL min , eluent 100 mM NH4NO3).
The known chromatographic interferences for the determination of bromate in drinking water via
ICP-MS detection are bromoacetic acid, dibromoacetic acid, bromochloroacetic acid, bromide and
the phosphate and/or sulphate matrix.14 Under the chromatographic conditions mentioned above
(flow rate of 170 µL min-1, eluent 100 mM NH4NO3), complete overlap of the peaks of
monobromoacetic acid and bromate was observed (the two most difficult brominated disinfection
by-products to resolve due to their close pKa values).10 This feature is also observed in the
chromatograms presented by Guo et al. and in different US EPA methods using IC.1,10,14 To
chromatographically resolve the co-elution of monobromoacetic acid and bromate, the conditions
were optimized to a flow rate of approximately 450 µL min-1 (25 rpm) using 2 mM NH4NO3 as eluent.
While the monobromoacetic acid and bromate peak were well separated, the bromide peak eluted
only after 26 minutes (see figure 5).
1,6E+06
-
Br
1,4E+06
1,2E+06
1,0E+06
8,0E+05
cps
6,0E+05
4,0E+05 -
BrO3
-
Br
2,0E+05
C2H3BrO2
0,0E+00
0 5 10 15 20 25 30
min
-1 -1 - -1 -
Figure 5: Standard solution containing 10 µg L of C2H3BrO2, 50 µg L of BrO3 and 50 µg L of Br measured
79 -1
with low pressure LC/ICP-SFMS ( Br, 100 µL sample loop, flow rate of 450 µL min ). Dotted line: eluent 2 mM
NH4NO3 ; full line: eluent 2 mM NH4NO3 and secondary injection of 375 µL of 100 mM NH4NO3 after 620 s.
For comparison, the HPLC/ICP-QMS method developed by Guo et al. allows the determination of
bromate and bromoacetic acids in 46 minutes.10 The US EPA method 321.8 allows the determination
of bromate in 12 minutes.1,14 In order to speed up the analysis, a sample loop of 2500 µL was
connected to the injection valve and the ESI FAST system was programmed with a ‘gradient’ mobile
phase using 2 mM and 100 mM NH4NO3. In principle, the 2 mM is being constantly delivered at a
flow rate of approximately 450 µL min-1 (25 rpm). First, the sample solution is loaded in the 2500 µL
sample loop using the vacuum pump and thereafter injected during 15 seconds (which corresponds
to an injection volume of ~ 113 µL). Second, the sample loop is filled with the 100 mM NH4NO3
solution and 620 seconds after the injection of the sample solution, the 100 mM NH4NO3 solution is
injected during 50 seconds (which corresponds to an injection volume of ~ 375 µL). The effect of
using a ‘gradient’ mobile phase using low pressure LC/ICP-SFMS is shown in figure 5. By using this
automated second injection as gradient mobile phase, the chromatographic run was shortened to 13
minutes.
To resolve co-elution of bromate, bromide, bromoacetic acid, dibromoacetic acid, bromochloroacetic

acid, phosphate and sulphate (matrix), the low pressure LC conditions were further optimized to a
flow rate of approximately 340µL min-1 using 5 mM NH4NO3 as eluent. First, the sample is loaded in a
1000 µL sample loop using the vacuum pump and injected during 180 seconds (~ 1000 µL). Second,
the sample loop is filled with 50 mM NH4NO3 solution and fully injected 1200 seconds after the first
injection (~ 1000 µL). A low pressure LC/ICP-QMS chromatogram of a standard solution containing
possible interferences is shown in figure 6 (74Ge, which has similar mass to 79Br, was used as internal
standard).23 Retention time shifts in ion chromatography are possible due to weak eluent strengths
and high ionic strength matrices. For phosphate concentrations up to 100 mg L−1 no interferences
were observed, however for sulphate concentrations above 200 mg L−1 peak broadening was
observed for bromate.
6
1 8
0,9
0,8 7
0,7
1,2
79 Br/ 74 Ge
0,6
0,5 5
0,4
0,3 3 4
0,2
0,1
0
0 5 10 15 20 25
min
79 74
Figure 6: Low pressure LC/ICP-QMS chromatogram ( Br/ Ge signal ratio) for a standard solution containing
-1 - -1 - -1 -1 -
(1) 100 mg L of PO4 ; (2) 100 mg L of SO4 ; (3) 5 µg Br L of C2H3BrO2; (4) 5 µg Br L of BrO3 ; (5) secondary
-1 -1 -1 -
injection of 50 mM NH4NO3; (6) 5 µg Br L of C2H2ClBrO2; (7) 5 µg Br L of C2H2Br2O2; (8)5 µg Br L of Br ; 1000
-1
µL sample loop, flow rate of 340 µL min , 5 mM NH4NO3 and secondary injection after 1200 s of 1000 µL of 50
mM NH4NO3.
The flexibility of low pressure LC/ICP-SFMS can be used to develop fit-for-purpose methods. In the
case only bromate is the analyte of interest, the 27 minutes lasting chromatographic run needed to
separate bromide from dibromoacetic acid (figure 6) can be shortened to 13 minutes (figure 5). As
the Maximum Admissible Concentration (MAC) of 10 µg L−1 bromate in drinking waters is rarely
exceeded, one could further reduce the chromatographic run time to 8 minutes (see figure 4). In this
case, monobromoacetic acid and bromate will co-elute and could lead to a false positive exceeding of
the MAC. However, exceedings are rarely observed (see further) and – if needed – these samples
could be easily re-analysed using a ‘gradient’ mobile phase to resolve the co-elution of
monobromoacetic acid.
According to the European drinking water directive, the performance characteristics of the method
of analysis must, as a minimum, be capable of measuring bromate with a trueness of 25 %, a
precision of 25 % and a limit of detection of 2.5 µg L−1.
The flexibility of low pressure LC/ICP-MS can be used to develop a tailor-made method with
predefined sensitivity by simply changing the sample injection time. This is demonstrated by injecting
a solution of 10 µg L-1 BrO3- and 10 µg L-1 Br- from the 2500 µL sample loop during 15 (113 µL), 30
(225 µL), 60 (450 µL), 120 (900 µL) and 240 (1800 µL) seconds. A correlation coefficient of R² = 0.9995
was derived between the measured peak area of bromate and the injection time. An overlay of the
chromatograms, time shifted for the different durations of the sample injection time (e.g.,
chromatogram for a sample injection time 120 s was shifted 105 s to the left), is shown in Figure 7.
This feature can also be used to instrumentally “dilute” a sample out of calibration range by simply
changing the injection time.
1,0E+06
BrO3- Br-
9,0E+05
8,0E+05
7,0E+05 15 sec - 113 µL

30 sec - 225 µL
6,0E+05
60 sec - 450 µL
5,0E+05
cps
120 sec - 900 µL

240 sec - 1800 µL
4,0E+05
3,0E+05
2,0E+05
1,0E+05
0,0E+00
0 2 4 6 8 10 12 14
min
Figure 7: Effect of sample injection time (sample loop volume) on peak area; low pressure LC/ICP-SFMS
79 -1 - -1 -
measurement ( Br) of a standard solution containing 10 µg L of BrO3 and 10 µg L of Br , eluent 2 mM
NH4NO3 and secondary injection after 620 s of 375 µL of 100 mM NH4NO3 (for the overlay of the
chromatograms, a time shift to compensate for the difference in sample injection duration was made).
The LoQ attainable with low pressure LC/ICP-SFMS was calculated based on 7 consecutive
measurements of a 0.5 µg L-1 bromate-spiked tap water using a 900 µL sample loop
(chromatographic conditions according to figure 7). An average concentration of 0.52 µg L-1 bromate
was measured, and the LoQ, calculated as six times the standard deviation on the 7 replicates,
amounted to 0.17 µg L-1 bromate.
The LoQ attainable with low pressure LC/ICP-QMS was calculated based on 7 consecutive
measurements of a 0.5 µg L-1 bromate-spiked tap water using a 1000 µL sample loop
(chromatographic conditions according to figure 6). An average concentration of 0.51 µg L-1 bromate
was measured, and the LoQ, calculated as six times the standard deviation on the 7 replicates,
amounted to 0.19 µg L-1 bromate.
Notwithstanding the relative high first ionisation energy of Br (11.81 eV), resulting in a low degree of
ionisation in the plasma, the calculated LoQ of both systems is similar and fit-for-purpose for the
determination of bromate in drinking water.26 The integrated peak area of a 0.5 µg L-1 bromate spike
solution (1000 µL sample loop) using ICP-QMS amounts to ~ 60000 cps and is a factor of ~ 10 lower
as compared to ICP-SFMS (in low resolution mode). Despite the superior sensitivity of ICP-SFMS, the
chromatographic baseline noise (low pressure LC system), the broadening of the bromate peak
(related to the strength of the eluent) and the manual integration of the chromatograms (peak areas)
at these low levels using the periSPEC Peak Area Finder Software (ESI) are likely to contribute to the
standard deviation from which the LoQ is derived.
A drinking water sample was spiked with 12 µg L-1 bromate and analysed 4 times via low pressure
LC/ICP-SFMS (chromatographic conditions according to figure 7 and using a 900 µL sample loop). An
average concentration of 11.9 ± 0.2 (st. dev.) µg L-1 bromate was measured (recovery of 99 %). The
standard deviation calculated on 7 determinations of quality control samples containing 0.5 and 5 µg
L-1 bromate, processed in different measurement runs, amounted to 0.502 ± 0.028 µg L-1 and 4.92 ±
0.24 µg L-1 bromate, respectively.
A drinking water sample was spiked with 2.5 µg L-1 bromate and analysed 4 times via low pressure
LC/ICP-QMS (according to chromatographic conditions described in figure 6). An average
concentration of 2.52 ± 0.09 (st. dev.) µg L-1 bromate was measured (recovery of 101 %). The
standard deviation calculated on 4 determinations of quality control samples containing 1 and 5 µg L-
1
bromate, processed in different measurement runs, amounted to 0.999 ± 0.021 µg L-1 and 4.93 ±
0.03 µg L-1 bromate, respectively.
The performance characteristics in terms of trueness, precision and LoQ of low pressure LC/ICP-SFMS
and LC/ICP-QMS are amply sufficient to meet the values specified in the European drinking water
directive. According to these required performance characteristics for regulatory monitoring, there is
still room left for further reduction in running costs, e.g., by using Ge as internal standard to be relied
on for Br quantification and thus omitting the daily external calibration procedure (Eickhorst et al.
showed that the Br-to-Ge ratio was nearly constant over 4 months and almost independent of the
ICP–MS instrument settings).24 Moreover, next to Br determination, ICP-MS also offers the possibility
of simultaneous Cl and I monitoring in the samples under investigation.12,24
Bromate analysis in audit monitoring program
The European Directive not only sets quality limits to water intended for human consumption, but
also defines a check and audit monitoring program for the Member States. For the 81 different
drinking water supply zones in the Flemish region of Belgium, approximately 1400 bromate
determinations in water collected from the tap are performed on a yearly basis in the audit
monitoring program.27 Besides these legal monitoring requirements, the 17 different drinking water
companies carry out supplementary bromate determinations at the level of the water production,
water towers and main supplies. All these results are reported to the Flemish Environment Agency
for an evaluation of the drinking water quality. Of the 1439 analyses of tap waters performed during
the audit monitoring program in 2009, only 1 result (11.4 µg L-1) exceeded the Maximum Admissible
Concentration for bromate.27 Of the 1477 tap waters analyzed during the audit monitoring program
in 2010, not a single exceeding was observed for bromate.28
As a supplementary validation and verification of the results of the drinking water companies, a
yearly analysis of about 20 tap waters randomly collected over the different drinking water supply
zones in the Flemish region is carried out by our laboratory. The results of these bromate
determinations of the last 4 years are summarised in figure 8. For the years 2009-2010, the analyses
were performed with HPLC/ICP-QMS, for the years 2011-2012, the analyses were performed with
low pressure LC/ICP-SFMS.
5
3
µg/l Bomate
Median
25%-75%
Non-Outlier Range
Outliers
2 Extremes
0
2009 2010 2011 2012
Figure 8: Box and Whisker plots of bromate concentrations yearly determined in 20 tap waters collected
randomly over different water supply zones in the Flemish region of Belgium (2009-2010 analysis with
HPLC/ICP-QMS; 2011-2012 analysis with low pressure LC/ICP-SFMS).
Partcipation in an interlaboratory trial for the determination of bromate in drinking water is

performed on a yearly basis. When using HPLC/ICP-MS the bias in the interlaboratory trial amounted
4.2 and 1.4%, when using low pressure LC/ICP-SFMS the bias amounted -3.1 and -0.3% (average of 9
participating laboratories).
The overall median value in Flemish tap water is 0.5 µg L−1 bromate and almost all measured
concentrations are below the recommended LoD of 2.5 µg L−1. In 2011, one result (24 µg L−1, extreme
value not shown in figure 8) exceeded the Maximum Admissible Concentration (10 µg L−1). The
maximum admissible level for bromate has been primarily based on current analytical capabilities
(not on toxicological considerations - the target concentration for bromate in drinking water is zero);
thus, there is a need for ongoing development and refinement of analytical technologies in order to
permit rapid and reliable determinations at the sub µg l−1level.
Conclusion
Routine determination of bromate in regulatory monitoring programs is needed to monitor the level
of brominated disinfection by-products in drinking water. The benefits of low pressure LC/ICP-MS as
compared to HPLC/ICP-MS are the ease of use and the availability of a low-cost system for fast and
routine speciation analysis, which can be implemented and automated on any ICP-MS
instrumentation easily. The methodology/technology described here shows that the complexity (and
cost) of the instrumentation required for such analysis has been considerably reduced, without
compromising performance characteristics in terms of LoQ and reliability of daily analysis results.
From the random sampling, organised as a supplementary validation and verification of the results of
the drinking water companies, a median bromate concentration in Flemish tap water of 0.5 µg L−1
was derived. One result exceeding the Maximum Admissible Concentration (10 µg L−1) for bromate
was observed on a total of 80 samples. Global and national agencies are striving to monitor bromate
in drinking water in order to establish appropriate regulatory limits. Depending on the results of
further research, a risk model could possibly provide a more definitive guideline value for oxyhalides
in drinking water. For these reasons, there is a need for further improvement of the existing methods
in terms of sensitivity, cost and reliability. Unquestionably, the analysis of inorganic oxyhalide
disinfection by-products in drinking water will continue to lead to ongoing improvement in water
quality in the future.
Acknowledgements:
We thank Jef De Wit and Bart Noten for aid in the organisation and sampling work. The monitoring campaign
was commissioned, financed and steered by the Flemish Environment Agency. The development of the low
pressure LC/ICP-SFMS method was financed by the Environment, Nature and Energy Department of the
Flemish Government.
References paper, Determination of bromate in drinking waters using low pressure liquid
chromatography / ICP-MS, J. Anal. At. Spectrom., 2013, 28, 1894 – 1902.
1. D. P. Hautman, D.J. Munch, C. Frebisa, H. P. Wagner and B. V. Pepich, (2001) J. Chromatogr. A,

2001, 920, 221.
2. R. Michalski, Polish Journal of Environmental studies, 2005, 14, 257.
3. K. C. Thompson, J. L. Guinamant, V. Ingrand, A.R. Elwaer, C. W. McLeod, F. Schmitz, De Swaef
and, Ph. Quevauviller, J. Environ. Monit., 2000, 2, 416.
4. R. Butler, L. Lytton, A. R. Godley, I. E. Tothill and E. Cartmell, J. Environ. Monit., 2005, 7, 999.
5. US EPA, Stage 1: Disinfectants and Disinfection By-Products Rule. A Quick Reference Guide, EPA
816-F-01-010, 1998.
6. Council Directive 98/83/EC of 3 November 1998 on the quality of water intended for human
consumption Official Journal of the European Communities Legislation, OOPEC, 1998, Vol. 41,
Part 330 32-54.
7. World Health Organization (WHO), Revision of the WHO Guidelines for Drinking Water Quality,
fourth edition, WHO, Geneva, 2011, http://www.who.int/water_sanitation_health/publications/
2011/dwq_guidelines/en/index.html, last accessed on 29/10/2013.
8. G. Schminke and A. Seubert, J. Chromatogr. A, 2000, 890, 295.
9. C. A. Delcomyn, H. S. Weinberg and P. C. Singer, J. Chromatogr. A, 2001, 920, 213.
10. Z. X. Guo, Q. Cai, C. Yu and Z. Yang, J. Anal. At. Spectrom., 2003, 18, 1396.
11. M. Nowak and A. Seubert, Anal. Chim. Acta, 1998, 359, 193.
12. H. Shi and C. Adams, Talanta, 2009, 79, 523.
13. Seubert, G. Schminke, M. Nowak, W. Ahrer and W. Buchberger, J. Chromatogr. A, 2000, 884, 191.
14. USEPA Method 321.8, Determination of Bromate in Drinking Waters by Ion Chromatography
Inductively Coupled Plasma - Mass Spectrometry, Revision 1.0 December 1997.
15. J. Morton and E. Leese, Anal Bioanal Chem., 2011, 399, 1781.
16. Seubert, M. Nowak, Fresenius J Anal Chem, 1998, 360, 777.
17. K. Tirez, V. Corthouts and N. De Brucker, Trace determination of bromate in drinking waters by
ion chromatography coupled to ICP-MS, poster presented at the International Ion
Chromatography Symposium, 2000, Nice, France.
18. K. Kahen, K. Jorabchi and A. Montaser, J. Anal. At. Spectrom., 2006, 21, 588.
19. T. D. B. Lyon, P. A. Robin, W. S. Watson and D. Littlejohn, J. At. Anal. Spectrom., 2005, 20, 757.
20. K. Falk and H. Emons, J. Anal. At. Spectrom., 2000, 15, 643.
21. C. A. Ponce de Leon, M. Montes-Bayon, J. A. Caruso, J. Chromatogr. A, 2002, 974, 1.
22. Dutoit and S.A. Pergantis, J. Anal. At. Spectrom., 2001, 16, 575.
23. F. Vanhaecke, H. Vanhoe, R. Dams and C. Vandecasteele, Talanta, 1992, 39,737.
24. T. Eickhorst and A. Seubert, J. Chromatogr. A, 2004, 1050, 103.
25. Q. Xie, R. Kerrich, E. Irving, K. Liber and F. Abou-Shakra, J. At. Anal. Spectrom., 2002, 17, 1037.
26. R. S. Houk, Anal. Chem., 1986, 58, 97.
27. Flemish government agency, results of the quality controls of the in Flanders region distributed
water intended for human consumption, 2009, http://www.vmm.be/, Accessed 20 Sept 2012.
water intended for human consumption, 2010, http://www.vmm.be/. Accessed 20 Sept 2012.
In this chapter, speciation applications have been designed to determine chromate in solid materials,
selenate and selenite in waste water and bromate in drinking water. All of the applications were
driven by European legislation, i.e., the restriction of hexavalent chromium in electrical and
electronic equipment brought onto the European market, removal of oxy-anions from wastewater to
ensure compliance with European environmental quality parameters and the control of the
maximum level of carcinogenic disinfection by-products in European drinking water monitoring
programs.
Referring to the quote by T.S. Eliot in the introduction (Where is the wisdom we have lost in
knowledge? Where is the knowledge we have lost in information?), one must keep in mind that in the
age of too much information, knowledge of and profound insight in the intrinsic quality of the
analytical data becomes even more valuable when it comes to interpretation of the actual chemical
status of our environment. The latter is especially illustrated in the conclusions formulated in the
three different papers dealing with the determination of Cr(VI) in solid matrices, where research in
the chromium species interconversion led to new analytical challenges and multi-disciplinary
methods to be used. Quality control and quality assurance programs carried out along these
speciation research projects were of utmost importance to derive the level of oxidation of Cr(III) and
to assess the toxicity of Cr-containing solid materials. Correction for the species interconversion
observed remains ambiguous when the chemical form of the species present in the sample (e.g.,
FeCr2O4) is different from the added spike (e.g., Cr(NO3)3.9H2O). On the other hand, knowledge of the
method-induced interconversion is particularly relevant for the determination of Cr(VI) in ambient
particulate matter (PM10) as the level of observed Cr(III) oxidation is within the range of the level of
monitored Cr(VI) and could lead to false positives exceeding of the air quality guideline value.
While the relevance of elemental speciation in regulatory monitoring is not at stake, why is
elemental speciation not done routinely with LC/ICP-MS in analytical laboratories? First, European
Directives and legislation about elemental speciation are to a great extent still missing with the
exception of those for a few species (e.g., Cr(VI), organotin compounds, methylmercury) and
therefore, routine laboratories lack the incentive to invest in the technologies. Second, for the
determination of some species, e.g., bromate, alternative cheap methods (as compared to LC/ICP-
MS) are available for monitoring at the concentration level of the regulatory limit value. Third, the
assumption of monitoring in a regulatory context is that only one or a few constituents of a sample -
usually the analytes of interest - play a role in the measurement that is obtained. When applying
standardized procedures in routine, users tend to have as little concern as possible about what other
species in the sample might yield a false response or whether the method is applicable to all samples
in the measurement run. However, and especially for solid samples, elemental speciation analysis is
not considered as a course of action to always be followed and interpreted in the same way. These
restrictions have often led to the current pragmatic approach in elemental speciation in regulatory
context, i.e., only in case the total concentration is above the limit value, an elemental speciation
analysis is performed.
Speciation analysis nowadays is a modern and important analytical discipline, in which many
questions have to be answered, especially in environmental chemistry and the life sciences. With the
availability of commercial interfaces to easily hyphen different separation techniques (primarily LC,
GC) to ICP-MS, one may forget the exciting research in the past aiming at, e.g., coupling GC via a self-
made heated transfer line to an ICP-MS unit [139]. Also, the efforts to set up an HPLC/ICP-MS system
for the speciation of selenium in the year 2000, were mainly devoted to the automation and a proper
communication interface between the sample injection system, the HPLC pump and the ICP-MS
instrument. Ten years later and with the availability of a low pressure LC system, speciation analysis
can be implemented easily on any ICP-MS instrument and this was demonstrated for the (routine)
determination of bromate in drinking water. Nowadays, LC/ICP-MS can truly be considered

widespread sensitive and versatile speciation instrumentation. Recent approaches to determine Se
species, specifically at ultratrace levels, have also focussed on a robust and work/time efficient
method by online coupling of a preconcentration (trap) column to an IC-ICP-MS system [124]. Other
novel approaches used ICP tandem mass spectrometry (ICP-MS-MS) with O2 reaction/collision gas to
completely remove severe interferences with the Se speciation originating from the plasma source
and the biological sample matrix [125].
Speciation is also increasingly involved in multi-disciplinary approaches using so-called orthogonal

speciation schemes (e.g., ESI-MS and HPLC/ICP-MS) for (complex) species identification and/or
quantification (e.g., metallomics, pharmaceutical substances)[111]. Recent advances in this field,
includes front-end modifications (e.g., micro and capillary nebulizers) to adapt the ICP-MS
instrument to the low flow rates (~ µl min-1) and/or high organic content solutions delivered by the
LC system. Also here, alternative instrumental approaches are needed in the search for lower
detection limits. One such approach is based on the high resolving power provided by double
focusing ICP-MS instruments (e.g., in case of phosphoproteomics to separate the 31P+ and 32S+ peaks
from those of polyatomic ions). A less expensive alternative is the use of reaction cells in ICP-QMS
(e.g., chemically resolve 31P+ and 32S+ from polyatomic ions by their oxidation to 31P16O+ and 32S16O+
after preferential reaction with oxygen gas supplied to the cell)[140].
Tracer experiments with stable isotopes in environmental studies and use of natural isotopic variation as a
169
proxy
CHAPTER 4 TRACER EXPERIMENTS WITH STABLE ISOTOPES IN

ENVIRONMENTAL STUDIES AND USE OF NATURAL ISOTOPIC VARIATION AS
A PROXY
4.1. PERSPECTIVE ON ISOTOPIC TRACERS
As a first approximation, it can be stated that all elements have an isotopic composition that is
invariant in nature. This is the result of thorough mixing of most nuclides prior to the formation of
the solar system some 4.6 billion years ago [144]. Although isotopic abundances are assumed to be
fairly constant in nature, variations do occur. These variations may result from the decay of naturally
occurring and long-lived radionuclides (e.g., 238U -> 208Pb) or natural isotope fractionation effects
[145]. Different isotopes of one and the same element display the same number of electrons and
thus show to a very large extent the same chemical behaviour. Nevertheless, small differences in
their physicochemical behaviour exist as a result of the difference in the number of neutrons in their
nucleus, or in other words, their mass. Due to this relative difference in mass, different isotopes of
the same element may take part to a slightly different extent in physical processes or in
(bio)chemical reactions with natural mass fractionation as a consequence.
The capability to determine isotope abundances is an important asset of mass spectrometry.

Inorganic mass spectrometry recently celebrated its 100th anniversary [146]. In 1910, Thomson
demonstrated the existence of isotopes of chemical elements with the example of two abundant
stable neon isotopes (20Ne and 22Ne). Thomson's student Francis William Aston continued the
research, building the first full functional mass spectrometer (reported in 1919), which rapidly
allowed him to identify no fewer than 212 of the 288 naturally occurring isotopes (or primordial
isotopes found on Earth that have existed in their current form since before Earth was formed). In
1921, F. W. Aston became a fellow of the Royal Society and received a Nobel Prize in Chemistry in the
same year “for his discovery, by means of his mass spectrograph, of isotopes, in a large number of
non-radioactive elements, and for his enunciation of the whole-number rule". Updates on the isotopic
composition of the elements as determined by isotope ratio mass spectrometry are reported on a
regular basis by the Commission on Isotopic Abundances and Atomic Weights (CIAAW) of the
International Union of Pure and Applied Chemistry (IUPAC) [147]. The update involves a critical
evaluation of the published literature and for each element, data from the "best measurement" of
the isotope abundances in a single sample, along with a set of representative isotope abundances
and uncertainties that accommodate known variations in normal terrestrial materials are reported.
The precise and accurate determination of isotope ratios is required for different application fields,
such as: geochronological dating; provenance determination of, among other, objects of art and/or
archeological or forensic relevance and agricultural products of animal or plant origin; the use of
isotope ratios determined in natural chronological archives (e.g., speleothem or coral) to serve as a
paleoproxy, e.g., for temperature or pH; determining isotope ratios of radioactive elements in the
nuclear industry and for radioactive waste control; tracer experiments using highly enriched stable
isotopes or long-lived radionuclides in environmental, biological or medical studies and the isotope
dilution technique as a potentially primary method for the determination of element concentrations
at trace and ultratrace levels.
In general, an important requirement for a technique used for studying (the sometimes extremely
small) variations in the isotopic composition of an element is the possibility to perform analyses with
170
proxy
a very high precision and accuracy. While (gas source) isotope ratio mass spectrometry (IR-MS) is
traditionally used for studying the isotopic composition of the “light” elements (e.g., H, C, N, O and
S), for a long time, thermal ionization mass spectrometry (TIMS) has been considered as the only
technique that provided a sufficient precision to detect subtle variations in the isotopic composition
of the “heavier” elements. Compared to TIMS, the first ICP-MS instruments offered a less demanding
sample preparation and were relatively easy to operate. But on the other hand, they were also
characterized by a rather poor isotope ratio precision ( ~ 0.2 to 0.5 %, while for TIMS precisions down
to 0.005 % are attainable), a more pronounced bias and spectral overlap of the analyte signals with
those of interfering ions [144,146]. Through the years, various modifications have led to a significant
improvement in the isotope ratio precision attainable with single-collector ICP-mass spectrometers
(down to 0.05% or slightly better). The first time that the isotope ratio precision typical for TIMS was
approached, was in 1992, with the introduction of multi-collector (MC) sector-field ICP-mass
spectrometers that are equipped with an array of Faraday collectors [144,145]. In contrast to the
single-collector instruments, with which the analyte signals are monitored sequentially, MC-ICP-MS
instruments are able to monitor the intensity of several ion beams simultaneously, as a result of
which short term variations in signal intensity are affecting all isotopes to the same extent, such that
these cause (practically) no detrimental effect on the isotope ratio precision. In this way, with the
most recent MC instruments, isotope ratio precision down to 0.002 % RSD can be obtained.
In the context of this dissertation, only single-collector ICP-mass spectrometers have been used for
isotope ratio measurements. With the attainable precision of this type of instrumentation, a variety
of environment-related issues can be studied using (i) compounds in which a particular element
shows an isotopic composition that is artificially made sufficiently different from the corresponding
natural one or using (ii) elements for which the natural variations in the isotopic composition are
quite pronounced (e.g., B, Sr or Pb).
Single-collector ICP-mass spectrometers are very well suited for performing isotope dilution
measurements for the determination of element (or species) concentrations at trace and ultratrace
levels using artificially enriched stable isotopes [148]. As an example, the species-specific isotope
dilution method for the determination of hexavalent chromium is discussed in § 3.3. Another
example of the good accuracy and precision attainable with isotope dilution, are the results of the
stability testing of human blood reference materials (BCR CRMs 634, 635 and 636), also performed
with the Elan 6000 quadrupole-based ICP-MS unit. The average results of 24 independent analyses of
each reference material are given hereunder (uncertainty expressed as a 95 % confidence interval).
Pb Cd
µg/L µg/L
BCR 634 – Certified 46 ± 5 1.4 ± 0.4
BCR 634 – measured 51 ± 2 1.27 ± 0.04
BCR 635 – Certified 210 ± 24 6.6 ± 0.6
BCR 635 – measured 209 ± 13 6.5 ± 0.5
BCR 636 – Certified 520 ± 50 11.6 ± 0.6
BCR 636 – measured 512 ± 8 11.6 ± 0.2
In the context of bio-monitoring and with increasing awareness of the possible link between
environmental exposure to certain metals and neuro-degenerative diseases, (ultra)trace
determination of elements in human body fluids are increasingly being performed with ICP-MS. The
analysis by ICP-MS can, however, be hampered by the existence of spectral (e.g., MoO+ on Cd+) and
non-spectral interferences and care must be taken with ambient low-level determination of trace
elements in human body fluids (e.g., problems with the accurate detemination of Cd in blood were
reported in the first Flemish bio-monitoring program, 2002–2006)[149].
171
proxy
The focus in this dissertation, however, is on tracer experiments with stable isotopes in
environmental studies and on the use of natural isotopic variation as a proxy. In the following
paragraphs, the scientific papers related to the development of methods based on ICP-MS for the
isotopic analysis of elements in an environmental context are introduced with emphasis on the
connecting thread of the PhD dissertation, i.e., the relation between the ICP-MS measurement, the
environmental issue and the regulatory context. The scientific papers have been arranged
chronologically in the following paragraphs (see timeline):
§ 4.2 Metal leaching from soils contaminated by historic smelter emissions
K. Tirez, P. Seuntjens and N. De Brucker, Full uncertainty calculation on quantitative determination

of tracer and natural cadmium in soil column effluents with ICP-MS, J. Anal. At. Spectrom., 1999, 14,
1475–1484.
§ 4.3 Sources of nitrate leaching
K. Tirez, W. Brusten, D. Widory, E. Petelet, A. Bregnot, D. Xue, P. Boeckx and J. Bronders, Boron
isotope ratio (δ11B) measurements in WFD monitoring programs: comparison between double-
focusing sector-field ICP (ICP-SFMS) and thermal ionization mass spectrometry (TIMS) , J. Anal. At.
Spectrom., 2010, 25, 964-974.
(2001) Study of uncertainty (2011) Ultratrace isotopic

contributions in tracer cadmium analysis of uranium in
transport in soil environmental and biological
samples [23]
(1999) Study of
tracer cadmium
transport in soil (2010) Boron isotope
ratio in groundwater
1997 2000 2005 2010
inductively coupled plasma-mass spectrometry (ICP-MS) for the determination and isotopic analysis
of metals and oxy-anions in an environmental context.
Interesting links:
• Isotopic Analysis, Fundamentals and Applications Using ICP-MS, F. Vanhaecke and P. Degryse,
2012, ISBN 978-3-527-32896-3 - Wiley-VCH, Weinheim, 529 p.
172
proxy
4.2. METAL LEACHING FROM SOILS CONTAMINATED BY HISTORIC SMELTER EMISSIONS
The transport of heavy metals in soils depends on water flow and physicochemical processes, such as
sorption and immobilisation due to mineral precipitation. Slow water flow and sorption result in
residence times of several decades or centuries in soils [151]. Many efforts have been spent to model
and evaluate heavy metal transport in soils at small scale, i.e., from the lab scale columns up to the
field scale, but only few assess metal leaching from large diffusely contaminated areas [152,153].
From a land management point of view, there is a clear need for models that allow predicting
concentrations of metals leaching to groundwater in a spatially variable context. Such an instrument
allows the land manager to plan measures, to prioritize actions, and to evaluate different scenarios
of soil and land reclamation and rehabilitation.
A prominent example of regional groundwater pollution is the Kempen area at the Dutch-Belgian
border. An area of approximately 700 km² is polluted with Cd and Zn due to historical non-ferrous
smelting activities [154-156].
Figure 35: Map of the affected area with the location of the 4 smelter sites (A, B, C and D) and the 4 classes of
wind direction [155].
Historical emissions of heavy metals due to non-ferrous ore processing, starting at the end of the 19th
century, caused metals to leach into groundwater in vulnerable areas, mainly consisting of acid sandy
soils and shallow groundwater tables.
In commission of OVAM, a spatially resolved model for leaching of Cd into groundwater was
developed for the area [155]. This model uses land-use maps, historical information on emissions,
atmospheric deposition data and soil data. Cd transport through the unsaturated zone can be
simulated taking into account differences in soil type and land-use. This results in spatially resolved
information on amount and timing of metals leaching into the groundwater at a regional scale. The
model allows evaluating management scenarios on the long term and reveals general patterns and
changes over longer periods of time (see maps below).
173
proxy
Figure 36: Map representing concentrations of Cd leaving the unsaturated zone in 2005 and map representing
expected concentrations of Cd leaving the unsaturated zone in 2100 under the remediation measures scenario
[155].
In order to validate the sorption models and to provide basic transport parameters for the leaching
of Cd into groundwater, laboratory-scale migration experiments have been performed. For this
purpose, soil migration experiments, in which a tracer (stable 111Cd isotope) was added on top of a
contaminated soil and the column effluents were fractionally collected, were performed. The level of
contaminant (natural) and tracer (111Cd) cadmium in the soil column effluents was simultaneously
determined by ICP-MS. For the calculation of the uncertainty on the quantitative determination of
tracer and natural cadmium, a method was proposed, emphasising a practical, realistic approach of
estimating uncertainties based on statistical assumptions (see paper: Full uncertainty calculation on
quantitative determination of tracer (111Cd) cadmium and natural cadmium in soil column effluents
with ICP-MS, J. Anal. At. Spectrom., 1999). In a subsequent paper (not presented in this dissertation),
the total uncertainty budget was used as a method evaluation criterion for the determination of
tracer (111Cd) cadmium and indigenous cadmium in soil column effluents with ICP-MS (K. Tirez, M.
Berglund, P. Seuntjens and N. De Brucker, J. Anal. At. Spectrom., 2001, 16, 307-314).
174
proxy
Full uncertainty calculation on quantitative determination of tracer (111Cd)

cadmium and natural cadmium in soil column effluents with ICP-MS
Kristof Tirez1, Piet Seuntjens1 and Nicole De Brucker1

1

Abstract
In laboratory-scale migration experiments a tracer (stable 111Cd isotope) is added on top of a

contaminated soil and the column effluents are fractionally collected. This paper describes an
analytical method for the simultaneous low level quantitative determination with inductively coupled
plasma-mass spectrometry (ICP-MS) of contaminant (natural) and tracer (111Cd) cadmium in soil
column effluents by adaptation of the isotope dilution equation.
For the calculation of the full uncertainty on the quantitative determination of tracer and natural
cadmium, a method has been proposed emphasising the practical, realistic approach of estimating
uncertainties based on statistical assumptions. The GUM Workbench program, the computations
of which follow the rules of the 'ISO guide to the expression of uncertainty in measurement', was
used. At low concentrations of cadmium (< 0.2 ng ml-1 Cd) the uncertainty due to counting statistics is
the major source of uncertainty. At higher concentrations the signal instability of the ICP-MS,
partially due to the clogging of the sampling cone by calcium salts present in the leachant (CaCl2),
forms the largest contribution to the total uncertainty. For concentrations of 0.5 ng ml-1 Cd and higher
the expanded uncertainty U amounts to ± 4 % (coverage factor 2, 95 % probability).
Introduction
For the protection of our groundwater resources the understanding of the transport of pollutants in
soil and groundwater is essential. In the "Kempen" Region of Belgium and the Netherlands, a large
area of sandy soils is contaminated by a diffuse atmospheric deposition of heavy metals. A low pH
and low organic carbon content in the acidic sandy soils enhance transport of mobile heavy metals
from the soil surface through the unsaturated zone into the groundwater.1 -4 After 100 years of non-
ferro industrial activity in the region, metals are slowly leaking from soil into the groundwater. Solute
transport models are used to estimate the future risk of groundwater contamination and
consequently support soil quality remediation.
During the last decade considerable research effort has been expended on clarifying transport
processes simultaneously occurring in the soil.5-8 Laboratory-scale migration experiments are
frequently used to validate sorption models and to provide basic transport parameters. For the study
of heavy metal transport mechanisms, stable isotopes, having nearly the same physicochemical
properties as the pollutant of concern, can be used as a tracer. In these experiments a concentrated
solution of metal tracer is brought on top of a soil column and leached out subsequently.
In this work, the stable 111Cd isotope was used to study the transport of cadmium in sandy soils. Both
the pollutant (cadmium) and the added stable isotope tracer (111Cd) are determined simultaneously
in the soil column effluent. These measurements allow to discern between transport properties of
cadmium residing for a prolonged period (years) in the soil and tracer spiked onto the column (days).
In this way, the effect of contaminant aging on transport can be assessed.
175
proxy
The sensitivity of inductively coupled plasma-mass spectrometry (ICP-MS) combined with the
possibility of different isotope measurements allows the determination of both cadmium fractions in
effluents at very low concentrations. The isotope dilution technique has been widely used to obtain
accurate and precise ICP-MS results. However, the isotope dilution formula cannot be used as such.
In the experimental setup (figure 1), a known amount of stable isotope tracer is added on top of the
soil column but is collected in different leached fractions. The exact amount of tracer in the different
fractions is unknown and forms part of the investigation. Therefore, an external calibration has to be
used in order to quantify the amount of tracer and natural cadmium in each collected fraction.
t=0 t=x t=y t=z
eluent
Tracer
cadmium
Natural
cadmium
[Cd] indigenous cadmium

tracer 111Cd
time
Figure 1: Laboratory-scale migration experiment set-up. Acrylic columns (Ø : 5,6 cm, 10 cm long) filled with
undisturbed sandy soil. The concentration of cadmium on the y-axis is the total cadmium concentration in the
soil column effluents of the indigenous (•) and added tracer (∆) cadmium.
176
proxy
As the analytical results obtained are used for modelling groundwater contamination, the estimation
of the total uncertainty is required. To quantify the combined uncertainty on the ICP-MS results, a
commercial available software program (GUM Workbench) was used. The analysis and
computations of this program follow the rules of the 'ISO guide to the expression of uncertainty in
measurement'9.
Method of calculation
Determination of tracer cadmium (Cdt) and natural cadmium (Cdn) in soil column effluents
For the simultaneous quantitative determination of tracer cadmium (Cdt) and natural cadmium (Cdn)
in soil column effluents, a calculation method was deduced from the isotope dilution equation. This
approach is based on the calculation of the partial fraction of both natural and tracer cadmium on
isotope 111. The concentration is derived from a calibration line obtained with natural cadmium
standards. 111Cd is the major abundant tracer isotope (table 1). The cadmium isotope ratios versus
111
Cd isotope are given in table 2.
10 a
isotope natural cadmium tracer cadmium
An ± U (%) At ± U (%)
106
Cd 1.25 ± 0.06 < 0.005
108
Cd 0.89 ± 0.03 < 0.005
110
Cd 12.49 ± 0.18 0.425 ± 0.020
111
Cd 12.80 ± 0.12 95.740 ± 0.021
112
Cd 24.13 ± 0.21 2.0809 ± 0.0047
113
Cd 12.22 ± 0.12 0.5570 ± 0.0027
114
Cd 28.73 ± 0.42 1.0411 ± 0.0042
116
Cd 7.49 ± 0.18 0.4560 ± 0.0021
a
Certified spike isotope reference material IRMM-621
Table 1: The relative abundances (± uncertainties) of isotopes for tracer and natural cadmium.
111 10 a
isotope ratio versus Cd natural cadmium tracer cadmium
Rn ± U Rt ± U
106 111
Cd / Cd 0.0977 ± 0.0048 < 0.00005
108 111
Cd / Cd 0.0695 ± 0.0024 < 0.00005
110 111
Cd / Cd 0.976 ± 0.017 0.00444 ± 0.00021
112 111
Cd / Cd 1.885 ± 0.024 0.02174 ± 0.00005
113 111
Cd / Cd 0.955 ± 0.013 0.005818 ± 0.000028
114 111
Cd / Cd 2.245 ± 0.039 0.010875 ± 0.000044
116 111
Cd / Cd 0.585 ± 0.026 0.001629 ± 0.000022
a
Certified spike isotope reference material IRMM-621
Table 2: Cadmium isotope ratios versus 111Cd (± uncertainties) for natural and tracer cadmium.
The calculation of the partial fraction natural cadmium on isotope 111 (111fn) is based on the
assumption that the total amount of the measured signal on mass 111 originates from the sum of
natural cadmium (111Cdn) and tracer cadmium (111Cdt). The same assumption can be made for the
measured signal on mass 113. The ratio of the ICP-MS signals of mass 113 to 111 for the sample (Rs)
is given by:
177
proxy
 mol 113 Cd n + mol 113 Cd t

R s =
 mol 111Cd n + mol 111Cd t (1)
mol 111Cd + mol 111Cd = mol 111
 n t Cd s
The fraction of natural cadmium 111 isotope (111Cdn) in the total cadmium 111 isotope signal (111Cds)
is equal to:
mol 111
Cd n mol 111
Cd n Rt − Rs
111
fn = = = (2)
mol 111
Cd t + mol 111
Cd n mol 111
Cd s Rt − Rn
where Rt and Rn are the ratios of the abundances of isotope 113 to 111 for, respectively, tracer and
natural cadmium (see table 2). Rs is the ratio of the signals of mass 113 to 111 for the sample
measured with ICP-MS.
A calibration line on isotope 111 is set up with two natural cadmium calibration standards and the
concentration of natural cadmium in the column effluents can be quantified by:
 111 f n ⋅ I s − I 1 
[Cd n ] =   ⋅ ([Cd 2 ] − [Cd 1 ]) + [Cd 1 ] (3)
 I 2 − I 1 
where Is, I1 and I2 are the ICP-MS signals on isotope 111 for, respectively, the sample, calibration
standard 1 and calibration standard 2. [Cd1] and [Cd2] are the concentrations of the calibration
standards. The concentration of tracer cadmium present in the column effluent can be deduced
from:
111
An Rs − R n
[Cd t ] = ⋅ ⋅ [Cd ] (4)
111
At Rt − Rn
111
An Rs − R n
111
Ct = ⋅ (5)
111
At Rt − Rn
where 111An and 111At are the relative abundances of 111Cd isotope for, respectively, natural and
tracer cadmium (table 1). The concentration of tracer cadmium in the column effluents is given by:
 111 C t ⋅ I s − I 1 
[Cd t ] =   ⋅ ([Cd 2 ] − [Cd 1 ]) + [Cd 1 ] (6)
 I 2 − I 1 
Experimental
Soil history and soil treatment
Undisturbed acrylic soil columns (diameter 5.6 cm; length 10 cm) were taken from different horizons
in a sandy soil at a distance of about 3 km from a metallurgical plant. The experimental site is located
in an old dune heath area with solitary pine trees. The soil was classified as a mesic typic Humod
according to Soil Survey Staff (1996). The columns were capped and transferred to the laboratory.
178
proxy
The soil columns were placed on a glass filter (air entry pressure 2 kPa). At the bottom of the column
a negative pressure was created using a vacuum device (KnF Verder vacuum controller). At the top of
the column solute is applied using a peristaltic pump (minicartridge, Cole Parmer) at a flow rate of 60
cm.day-1. A rotating lid and a thin gravel layer ensures a homogeneous distribution of solute over the
soil surface, as shown by a dye tracing experiment with methylene blue. The effluent is cycled
through a conductivity cell (Shott, 2ml) before being collected by a fraction collector (LKB-Bromma,
2211 Superrac). The conductivity cell is connected to a conductivity measurement device (KM 200,
Sensortechnik Meinsberg) for a continuous registration of electrical conductivity on a computer.
Milli-Q water purification system (Millipore, Milford, MA, USA) was used. HNO3 solutions were
prepared by diluting subboiled 65 % HNO3 (pro analysi, Merck, Darmstadt, Germany). An 1 M CaCl2
stock solution was prepared by dissolving 14.702 g CaCl2.2H2O (pro analysi ,Merck) in 100 ml of
water. All solutions were stored in polyethylene bottles.
For the determination of the mass bias a 25 ng ml-1 Cd solution (2 % v/v HNO3) was prepared from a
Cd stock solution of 1000 µg ml-1 (Merck). For the determination of the dead time a 200 ng ml-1 Rh
solution (2 % v/v HNO3) was used (Perkin Elmer, Norwalk, CT).
Calibration standards were diluted from a Cd stock solution of 1000 µg ml-1 (Merck). The
concentration levels of the standards were 0.5 and 7.5 ng ml-1 Cd and a calibration blank was used.
The standards were prepared in a 1 mM CaCl2 and 2 % v/v HNO3 matrix.
The internal standard 100 ng ml-1 Rh solution (2% v/v HNO3) was prepared from a stock solution of
1000 µg ml-1 Rh (Merck). The choice of the internal standard was based on the similarity between the
masses of the analysed cadmium isotopes (111 and 113 amu) and the rhodium mono isotope mass
(103 amu). For the preparation of the control standard solution of 1 ng ml-1 Cd (2% v/v HNO3, 1 mM
CaCl2) a stock solution of 10000 µg ml-1 Cd (SPEX, New Jersey, USA) was used. The rinsing solution
consisted of a 2 % v/v HNO3 solution.
Instrumentation
Samples were analysed by ICP-MS with a Perkin-Elmer SCIEX (Thornhill, ON, Canada) Elan 6000
instrument using an on line packed reaction coil for addition of internal standard solution. A gem
tipped cross-flow nebuliser (Perkin-Elmer) interfaced with a Perkin-Elmer (Ryton) Scott double pass
spray chamber, using a 2.0 mm id alumina injector tube, comprised the sample introduction system.
The on line addition of the internal standard causes a dilution of the sample by 20 %. A peristaltic
pump controlled the sample liquid uptake rate (± 1.0 ml min-1) with 0.75 mm id tubing for the sample
and 0.19 mm id for the internal standard. The instrumental measurement settings were optimised
and are shown in table 3.
parameter value
instrument settings
nebuliser gas flow 0.88 ±0.02 l/min
ICP RF power 1050 V
analog stage voltage -2137.5 V
lens voltage 7.25 ± 0.5 V
oxide level (CeO) <3%
179
proxy
2+
double charged level (Ba ) <3%
dwell time 25 ms
sweeps / reading 350
readings / replicate 1
replicates 5
sample introduction
peristaltic pump 24 rpm
∅ sample tubing 0.75 mm
on line addition internal standard

∅ internal standard tubing 0.19 mm
packed reaction coil (Dionex) 500 µl
3-way liquid mixing tees, PEEK (Dionex) 0.79 mm (I.D. 10/32)
Table 3: Optimised instrumental measurement settings ICP-MS, Perkin Elmer, Sciex, Elan 6000.
Before analysis, the instrument was preconditioned by aspirating a 1 mM CaCl2 solution (2 % v/v
HNO3) for at least 2 hours. To fulfil the instrument optimisation criteria a plasma set up solution
(Perkin Elmer, Norwalk, CT) was used. For the determination of the optimum lens voltage a 25 ng
ml-1 Cd solution (2% v/v HNO3, 1 mM CaCl2) was prepared from a Cd stock solution of 1000 µg ml-1
(Merck, Darmstadt, Germany).
Sample treatment
At the start of the column experiment a 1 mM CaCl2 leachant solution replaces the original soil
solution. Afterwards, 10 ml of a 111Cd tracer solution of 500 ng ml-1 Cd (IRMM, Geel, Belgium) is
added on top of the soil column (figure 1). The leachant solution (1mM CaCl2) is added on top of the
soil column at a rate of 1,44 l day-1. The fractions are collected over 4 hours (± 250 ml), acidified with
5 ml HNO3 and stored in polyethylene bottles.
Quantification of combined uncertainty on the analytical measurement
The GUM Workbench program was used for the quantification of combined uncertainty on the
analytical measurement (Metrodata GmbH, Germany, http://home.t-online.de/home/metrodata/).
The analysis and computations follow the rules of the 'ISO guide to the expression of uncertainty in
measurement.9 and the EAL-R2 document of the ‘European cooperation for Accreditation of
Laboratories’.11
The program supports a systematic procedure in building an uncertainty analysis as requested in the
EAL document. Starting with the mathematical equation which models the physical relationship of
quantities in the measurement, the data needed for the analysis, like the standard uncertainty or the
distribution of values, is interactively requested. A fitting classification of the input values according
to the available information controls the analysis process.
180
proxy
Results and Discussions

Combined uncertainty calculation on [Cdn]
The described method for the calculation of the combined uncertainty on the quantitative
determination of tracer and natural cadmium emphasises the practical, realistic approach of
estimating uncertainties based on statistical assumptions. The uncertainty estimation can roughly be
divided into four steps: specification (relation between the element of concern and the parameters),
identification of uncertainty sources, quantifying the uncertainty components and calculating the
combined uncertainty.
The ISO guide9 states that the uncertainty on a component can be grouped into two categories
according to the way in which their numerical value is estimated:
• type A : those which are evaluated by statistical methods;
• type B : those which are evaluated by other means.
The difficulty in the estimation of combined uncertainty lies in the realistic quantification of the type
B uncertainty.
In the calculation of the combined uncertainty on the concentration of natural cadmium in soil
column effluents the output quantity, denoted by Y (e.g. [Cdn]), depends upon a number of input
quantities Xi (i = 1, 2, ...., N) (e.g., equation 3) according to a functional relationship
Y = f (X1 , X 2 ,...., X N ) (7)
An estimate of the measurand Y, the output estimate denoted by y, is obtained using input estimates
xi for the values of the input quantities Xi :
y = f ( x 1 , x 2 ,...., x N ) (8)
For not correlated input quantities the square of the standard uncertainty associated
with the output estimate y is given by
N
u 2 ( y) = ∑ u i2 ( y) (9)
i =1
The quantity ui (y) (i = 1,2, ..., N) is the contribution to the standard uncertainty associated with the
output estimate y resulting from the standard uncertainty associated with the input estimate xi
u i ( y) = c i ⋅ u ( x i ) (10)
where ci is the sensitivity coefficient associated with the input estimate xi, i.e., the partial derivative
of the model function f with respect to xi,
∂f ∂f
ci = = X1i = x1 ....X N = x N
(11)
∂x i ∂X i
The following experimental steps are involved in the calculation of the combined uncertainty on the
concentration of natural cadmium in soil column effluents:
1. collection of soil column effluents and sample treatment
2. preparation of calibration cadmium standards
3. measurement of samples by ICP-MS
4. calculation of concentration
181
proxy
step 1: sample treatment
The soil column fractions are collected over 4 hours (± 250 ml), acidified with 5 ml HNO3 (65% ) and
stored at 277 K in polyethylene bottles. The samples are, due to the acidification, diluted by a factor
of approximately 1.02. For each fraction the collected volume is weighed and 5 ml HNO3 is added
with a pipette. The calibration certificate of the balance establishes a reproducibility of 200 ± 0.0002
g (95% confidence). This quantity needs to be divided by 1.96 to give the component of uncertainty
as a standard deviation. The quality control log shows a standard deviation of ± 0.0003 g for a check
weighing of 200 g. The two components are then combined to give an uncertainty:
2
 0.2 
u balance =   + 0.3 2 = 0.32 mg
 1.96 
The density of water at 293 K is 0.99821 g ml-1. For the conversion of the weight of the sample to a
volume a possible temperature variation of ± 3 K (with 95 % confidence) is taken into account. Taking
the coefficient of volume expansion of water as 2.1 x 10-4 K-1, this gives a 95% confidence interval for
the calculation of a volume of ± 3 x 2.1 x 10-4. Dividing this by 1.96 to obtain the standard deviation
gives a value of ± 3.6 x 10-4 for the uncertainty due to incomplete temperature control. The sample
volume is calculated as follows:
Vsample = m sample ⋅ 1
δ water
The uncertainty on the weighing is combined with the uncertainty on the conversion to volume:
2 2
u Vsample  0.32   0.00036 
=   +   = 0.00036
Vsample  200000   0.99821 
resulting in an uncertainty on a volume of 250 ml of ± 250 x 0.00036 = ± 0.09 ml. The technical data
of the 5 ml pipette states an accuracy of 0.14 % on a volume of 5 ml. The standard deviation on
different repeated pipette operations is ± 0.018 ml, resulting in a combined uncertainty on the acid
volume of
2
 0.0014 ⋅ 5 
u Vacid =   + (0.018)2 = 0.018 ml
 
 3 
For a 250 ml sample the dilution factor, d, is equal to (see also table 4)
(Vsamle + Vacid )
d= = 1.02000 ± 0.00007
Vsample
182
proxy
model equation :
(Vsamle + Vacid )
d=
Vsample
ρ water
quantity value standard df sensitivity uncertainty uncertainty

uncertainty coefficient contribution contribution
u (xi) ci ui (y) (%)
Vsample 250 ml 0.0802
msample 249.5525 g 3 E-03 50 -80.1 E-06 -240 E-09 0.0
ρwater 0.998210 g/ml 360 E-06 50 0.02 7.21 E-06 1.0
Vacid 5.000 0.018 50 4.0 E-03 72.0 E-06 99
result :
dilution factor d = 1.02000 ± 0.00014, where the reported uncertainty is calculated using a coverage
factor of 2 (95%).
Table 4: Combined uncertainty on step 1 'treatment of sample'.
For a correct calculation of natural cadmium, the model stated in equation 3 has to be multiplied by
this factor
  111 f n ⋅ I s − I 1  
[Cd n ] =   ⋅ ([Cd 2 ] − [Cd 1 ]) + [Cd 1 ] ⋅ d (3_a)
 I − I  
 2 1 
step 2: preparation of the calibration standards
Two calibration solutions, [Cd1] = 0.5 ng ml-1 and [Cd2] = 7.5 ng ml-1, were prepared from a
commercial stock solution of cadmium. The concentration of cadmium in the commercial solution
was stated to be 1004 ± 2 mg l-1. The dilution scheme is shown in figure 2.
1004
f2 0.000 [Cd1]
1
f3 0.007 [Cd2]
-1
Figure 2: dilution scheme for the preparation of cadmium calibration standards (concentrations in ng ml ).
To calculate the uncertainties associated with each of the concentrations, it is necessary to

determine the uncertainty on the stock solution, estimate the uncertainties in the dilution chain and
combine the two sources of uncertainty. The 1:1000 dilution (f1) was performed using a 100 µl
pipette (V0.1) and 100 ml volumetric flask (V100). The 1:2000 dilution (f2) was be performed using a 50
µl pipette (V0.05) and 100 ml volumetric flask. The 1:133 dilution (f3) was to be performed using a 750
µl pipette (V0.75) and 100 ml volumetric flask. Contributions due to imperfect repeatability and
variation within specification limits must be determined and combined for each type of pipette and
glassware used. The uncertainty arising from imperfect repeatability is estimated from the standard
deviation of a number of fill and weigh operations. The uncertainty arising from variation within
183
proxy
specification is obtained by dividing the manufacturers specification by 3 to convert from

rectangular limits to standard deviation.11 Table 5 summarises the data so obtained.
description std dev (ml) uv (ml) uv/V

Spec./ 3 (ml)
V0.1 pipette of 0.100 ml 0.00056 0.00006 0.00057 0.0057
V0.05 pipette of 0.050 ml 0.00040 0.00003 0.00040 0.0080
V0.75 pipette of 0.750 ml 0.0028 0.00043 0.0028 0.0038
V100 flask of 100 ml 0.051 0.06 0.08 0.0008
Table 5: Uncertainties on the different pipettes used for the preparation of the calibration standards.
To obtain the uncertainties for each dilution, each step is treated as a dilution factor made up from
the initial and final volume.
V100 V100 V100

f1 = ; f2 = ; f3 =
V0.1 V0.05 V0.75
The calculation of the combined uncertainty on the lowest standard is summarised in table 6.
model equation :
[Cd stock ]
[Cd 1 ] =
f1 ⋅ f 2
V100 V100
f1 = ; f2 =
V0.1 V0.05
quantity value standard df sensitivity uncertainty uncertainty

-1
[Cdstock] 1004000 ng ml 1160
50 500 E-09 578 E-06 1.3
f1 1000.0 9.7
f2 2000 22
V100 100.00 ml 0.078
50 -0.010 -7.83 E-06 2.4
V0.1 0.1000 ml 0.0006
50 5.02 2.86 E-03 32.3
V0.05 0.0500 ml 0.0004
50 10.0 4.02 E-03 63.9
result :
-1
[Cd1] = 0.502 ± 0.010 ng ml , where the reported uncertainty is calculated using a coverage factor of 2
(95%).
Table 6: Combined uncertainty on step 2 'preparation of the calibration standard'.
The same model can be used for the calculation of the combined uncertainty on the upper
calibration standard [Cd2] = 7.53 ± 0.11 ng ml-1 (coverage factor 2, 95%). So far, the uncertainty
arising from unknown constant temperature offset during the dilution has not been considered. The
effect can readily be calculated from the known coefficient of expansion as in step 1. Any offset
during the dilution process will affect [Cd1] and [Cd2] by the same factor. This factor, fK, is
incorporated in equation 3_a and applies directly to [Cdn] :
184
proxy
  111 f n ⋅ I s − I 1  
[Cd n ] =   ⋅ ([Cd 2 ] − [Cd 1 ]) + [Cd 1 ] ⋅ d ⋅ f K (3_b)
 I − I  
 2 1 
step 3: ICP-MS measurement
3.1 spectral and non-spectral interferences
The analytical measurements with ICP-MS can be hampered by interference of different kinds
(isobaric, oxides, poly-atomic, or doubly charged interferences). Study of the possible interference is
necessary before starting the analysis. Because all the measured samples (soil column effluents)
have more or less the same composition as the leachant, a solution consisting of 1 mM CaCl2 (2 %
v/v HNO3) was analysed for a first evaluation of the interferences. From all the literature cited
interferences in the cadmium mass range,12-15 only the expected occurring poly-atomic interferences
and all the isobaric interferences are listed in table 7.
element mass interference % abundance

strontium
84 16
100 Sr O 0.55
84 17
101 Sr O -
84 18 86 16
102 Sr O, Sr O 9.84
86 17 87 16
103 Sr O, Sr O 6.99
86 18 87 17 88 16
104 Sr O, Sr O, Sr O 82.40
87 18 88 17
105 Sr O, Sr O 0.05
88 18
106 Sr O 0.17
lead
204 ++
102 Pb 1.4
206 ++
103 Pb 24.1
207 ++
103,5 Pb 22.1
208 ++
104 Pb 52.4
% relative abundance
tin
112
112 Sn 0.97
114
114 Sn 0.65
116
116 Sn 14.53
indium
113
113 In 4.3
palladium
106
106 Pd 27.33
108
108 Pd 26.46
110
110 Pd 11.72
Table 7: Possible occurring interferences for the determination of cadmium in mass range 100 -116 u.
A mass scan of the matrix solution pointed out that approximately 0.2 ng ml-1 Sn, 0.1 ng ml-1 Pb, and
15 ng ml-1 Sr are present in this solution. Since no evidence of the presence of indium was observed
(and expected), the isobaric correction equation for indium on the cadmium isotope 113 was not
used. The non spectral interferences were counteracted by the on line addition of internal standard
(rhodium) as stated under Instrumentation.
185
proxy
3.2 uncertainty on ICP-MS signal, uI
The uncertainties on all the parameters for the calculation of fn, [Cdn] and [Cdt] are known with the
exception of the measured ratio (Rs). The uncertainty on the measured isotope ratio is of course, to
a certain extent, dependent on the concentration of the cadmium in the sample. According to
Poisson statistics,16 the noise or standard deviation of a signal measured using pulse counting can be
estimated as the square root of the signal. This means precision should improve with increasing
signal intensity. However, the aim of the uncertainty calculation, as described in this paper, is not a
theoretical calculation of the dependence of the standard deviation on the instrument parameters
and element concentration. Although those calculations are useful to understand the origin of the
standard deviations,17 the emphasis lies in the estimation of the magnitude of the uncertainty and
the impact in the total uncertainty budget. In order to obtain a robust estimate of the uncertainty on
an ICP-MS signal, the dependency of the signal versus its standard deviation was examined for all the
cadmium measurements in the concentration range of 0 to 7.5 ng ml-1 Cd (n = 425, figure 3).
relative standard deviation in function of signal 111Cd (cps)

10.0
9.0
8.0
7.0
6.0
RSD %
5.0
4.0
3.0
2.0
1.0
0.0
0 10000 20000 30000 40000 50000
111
Cd signal (cps)
Figure 3: Relationship between relative standard deviation (%, 5 replicates) and signal intensity (cps) measured
111
on Cd for different concentrations of cadmium in the sample.
The square root of the signals is proportional to the standard deviation (s∼ I ) up to approximately
10000 cps, with a factor of 0.4 (figure 4). Above this point a proportional relationship s ∼ I was
observed, with a factor of 0.0047 (relative standard deviation of 0.47 %). Under the given ICP-MS
conditions a signal of 1 ng ml-1 Cd gives approximately 1000 cps on 111Cd.
186
proxy
relationship between standard deviation and signal intensity

(111Cd < 10.000 cps)
2
1.8
standard deviation /
1.6
1.4
1.2
sqrt(y)
1
0.8
0.6
0.4
0.2
0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
111
Cd signal (cps)
Figure 4: Relationship between square root of signal intensity (< 10.000 cps) and standard deviation (5
111
replicates) measured on Cd for different concentrations of cadmium. The proportional factor is equal to
0.408 ± 0.145 (standard deviation on 304 samples).
This approach was used to predict the uncertainty on the signal (standard deviation of the mean of 5
replicates) of the cadmium isotopes:
I
I < 10000 cps, u I = 0.4 ⋅ (12)
5
I
I > 10000 cps, u I = 0.0047 ⋅ (13)
5
These results are in agreement with the results found by E. R. Denoyer et al.16
This uncertainty, uI, reflects the repeatability of the signal during the measurement period but not
the stability of the ICP-MS instrument during the analysis of the different samples. Because a
calibration is performed every ten samples, the stability of the signal has to be quantified during that
time period. The differences between the signal intensity (after correction with an internal standard,
Rh) before and after the measurement of ten samples are compared for each individual calibration
standard. The average value of the calculated differences in signal intensities for the standards (0.5
and 7.5 ng ml-1 Cd) was 2.6 ± 1.2 % (eight measurement runs).
The clogging of the sampling cone by calcium salts present in the soil column effluents, causes a
linear drift of the signal intensity resulting in a mean bias drift of 1.3% (2.6 %/2) across the
measurement run. This bias is considered by multiplying the measured intensity of the sample, Is,
with a drift factor IRh = 1 ± 0.013 in the model equation.
  111 f n ⋅ I s ⋅ I Rh − I 1  
[Cd n ] =   ⋅ ([Cd 2 ] − [Cd 1 ]) + [Cd 1 ] ⋅ d ⋅ f K (3_c)
 I 2 − I1  
 
187
proxy
3.3 isotope ratio measurement
The observed isotope ratios deviates from the true values as a function of the difference in mass
between the two isotopes.18 The true ratio of isotopes A and B , (A/B)t, can be related to the
measured ratio, (A/B)m, by :
A A
  =   ⋅ (1 + a ⋅ n ) (14)
 B  m  B  t
where a is the bias per mass unit, and n is the mass difference between isotopes A and B. Also,
determination of the dead time is critical when ratios of isotopes with different relative abundances
are calculated. Ten independent measurements of a 25 ng ml-1 Cd (2 % v/v HNO3) solution resulted in
a mass bias per mass unit value of 0.0161 ± 0.0007 (standard deviation) based on cadmium isotope
ratio 113 to 111.
For isotope dilution measurements, the error magnification factor has to be taken into account for
the estimation of the added spike concentration.19 For an optimum isotope dilution measurement
(lowest uncertainty), the measured ratio has to be situated in the minimum of the error
magnification factor curve. However, for column effluents the isotope ratios differ in each collected
fraction and a compromise has to be found for the amount of tracer cadmium to be added.
step 4: calculation of concentration natural cadmium
Owing to the restricted working area of the ICP-MS instrument (0.5 to 7.5 ng ml-1), the linearity of the
calibration line can be assumed. The combined uncertainty on the concentration natural cadmium in
soil column effluent can be calculated from equation 3_c and for one collected soil column effluent
the calculations are summarised in table 8.
model equation :
  111 f n ⋅ I s ⋅ I Rh − I 1  
[Cd n ] =   ⋅ ([Cd 2 ] − [Cd 1 ]) + [Cd 1 ] ⋅ d ⋅ f K
 I 2 − I1  
 
with
Rt − Rs
111
fn =
Rt − Rn
[Cd stock ]
[Cd 1 ] =
f1 ⋅ f 2
[Cd stock ]
[Cd 2 ] =
f1 ⋅ f 3
V100 V100 V100
f1 = ; f2 = ; f3 =
V0.1 V0.05 V0.75
(Vsamle + Vacid )
d=
Vsample
ρ water
quantity value standard df type sensitivity uncertainty uncertainty
188
proxy

111
fn 0.058401 887 E-06
Is 17196.8 cps 36.2 5 A -4.26 E-06 -1.54 E-04 0.01
IRh 1 0.013 16 A 0.698 9.08 E-03 34.8
I1 748.0 cps 4.9 5 A -6.78 E-04 -3.32 E-03 4.6
I2 11059.0 cps 18.9 5 A -1.73 E-05 -3.27 E-04 0.05
-1
[Cd1] 7.530 ng ml 0.053
-1
[Cd2] 0.502 ng ml 0.005
d 1.02 72.4 E-06
fK 1.00 360 E-06 50 B 0.69 2.48 E-04 0.03
Rt 5.818 E-03 28.2 E-06
Rs 0.061232 366 E-06
Rn 09547 0.0130
113
At 0.5570 2.70 E-03 50 B -0.124 -3.35 E-04 0.05
111
At 95.74 0.0210 50 B 7.21 E-04 1.51 E-05 0.00
113
An 12.22 0.12 50 B -0.0575 -6.90 E-03 20.1
111
An 12.80 0.12 50 B 0.0549 6.59 E-03 18.3
-
[Cdstock] 1004000ng ml 1150 50 B 6.87 E-07 7.91 E-04 0.26
1
113
Is 1086.9 cps 5.9 5 A 7.10 E-04 4.19 E-03 7.4
a 0.0161 700 E-06 10 A -1.49 -1.05 E-03 0.46
n 2.0
Vsample 250.0000 ml 0.0902
Vacid 5.0000 ml 0.0180 50 B 2.71 E-03 4.87 E-05 0.00

msample 249.55250 g 3 E-03 50 B -5.42 E-05 -1.63 E-07 0.00
ρwater .998210 g/ml 0.154 50 B 0.0136 4.88 E-06 0.00
f1 1000.0 9.7
f2 2000 22
f3 133.333 0.509
V100 100.00 ml 0.078 50 B -0.0138 -1.08 E-03 0.5
V0.75 0.7500 ml 0.0028 50 B 0.255 7.13 E-04 0.21
V0.1 0.1000 ml 0.0006 50 B 6.9 3.93 E-03 6.5
V0.05 0.0500 ml 0.0004 50 B 9.99 3.99 E-03 6.7
result :
-1
[Cdn] = 0.690 ± 0.031 ng ml , where the reported uncertainty is calculated using a coverage factor of 2
(95%).
Table 8: Combined uncertainty on concentration natural cadmium in soil column effluents.
The leached concentrations of tracer and natural cadmium for one soil column experiment are
shown in figure 5 and summarised in table 9, with the indication of the fraction that was used as an
example for the uncertainty calculation in table 8.
189
proxy
time (hours) [Cdn] ± U [Cdt] ± U

-1 -1
(ng ml ) (ng ml )
20 6.078 ± 0.248 0.025 ± 0.027
24 4.100 ± 0.167 0.010 ± 0.021
28 4.913 ± 0.200 1.032 ± 0.038
32 2.090 ± 0.086 5.943 ± 0.218
36 1.772 ± 0.073 3.864 ± 0.142
40 1.478 ± 0.062 3.180 ± 0.117
44 0.998 ± 0.043 2.211 ± 0.081
48 0.690 ± 0.031 1.497 ± 0.054
49.2 0.519 ± 0.025 1.127 ± 0.042
65.25 0.348 ± 0.019 0.683 ± 0.027
69.25 0.268 ± 0.017 0.342 ± 0.017
73.25 0.212 ± 0.015 0.210 ± 0.014
77.25 0.238 ± 0.016 0.166 ± 0.014
81.25 0.217 ± 0.015 0.133 ± 0.013
85.25 0.183 ± 0.014 0.091 ± 0.013
89.25 0.188 ± 0.015 0.076 ± 0.013
93.25 0.163 ± 0.014 0.053 ± 0.013
94.05 0.174 ± 0.014 0.049 ± 0.013
98.05 0.162 ± 0.014 0.039 ± 0.013
102.05 0.161 ± 0.014 0.029 ± 0.013
106.05 0.143 ± 0.014 0.023 ± 0.013
110.05 0.139 ± 0.014 0.016 ± 0.013
112.45 0.138 ± 0.014 0.015 ± 0.013
116.45 0.124 ± 0.013 0.010 ± 0.013
Table 9: Concentrations of natural and tracer cadmium (± U, 95 % confidence intervals) in soil column effluents
(figure 5). The fraction that was used for the example in tables 8 and 10 is indicated by bold type.
The calculation of the combined uncertainty on the concentration tracer cadmium is analogous to
natural cadmium and for the same soil column effluent as above the budget was calculated in table
10.
5
-1
[Cd] in ng ml
4 natural cadmium
1.479 ± 0.054
tracer cadmium
3
2
0.690 ± 0.031
0
15 35 55 75 95 115
time (hours)
Figure 5: Example of tracer and natural cadmium in soil column effluents ([Cd] ± 2*u[Cd], 95 % confidence
intervals), the calculation of the uncertainties on the indicated points are summarised in table 8 and 10.
190
proxy
model equation :
  111 C t ⋅ I s ⋅ I Rh − I 1  
[Cd t ] =   ⋅ ([Cd 2 ] − [Cd 1 ]) + [Cd 1 ] ⋅ d ⋅ f K
 I 2 − I1  
 
with
111
An Rs − R n
111
Ct = ⋅
111
At Rt − Rn
other quantity equations see table 8.
quantity value standard df type sensitivity uncertainty uncertainty

111
Ct 0.12589 1.11 E-03
Is 17196.8 cps 36.2 5 A 9.35 E-05 3.38 E-03 1.65
IRh 1 0.013 16 A 1.51 0.0196 55.48
I1 748.0 cps 4.9 5 A -6.00 E-04 -2.93 E-03 1.24
I2 11059.0 cps 18.9 5 A -9.55 E-05 -1.81 E-03 0.47
-1
[Cd1] 7.530 ng ml 0.053
-1
[Cd2] 0.502 ng ml 0.005
d 1.02 72.4 E-06
fK 1.00 360 E-06 50 B 1.5 5.39 E-04 0.04
Rt 5.818 E-03 28.2 E-06
Rs 0.061232 366 E-06

Rn 09547 0.0130
113
At 0.5570 2.70 E-03 50 B 0.0166 4.47 E-05 0.00
111
At 95.74 0.0210 50 B -0.0158 -3.32 E-04 0.02
113
An 12.22 0.12 50 B 7.69 E-03 9.22 E-04 0.12
111
An 12.80 0.12 50 B 0.11 0.0132 25.16
-
[Cdstock] 1004000ng ml 1150 50 B 1.49 E-06 1.71 E-03 0.42
1
113
Is 1086.9 cps 5.9 5 A -9.49 E-05 -5.60 E-04 0.05
a 0.0161 700 E-06 10 A 0.2 1.40 E-04 0.00
n 2.0
Vsample 250.0000 ml 0.0902
Vacid 5.0000 ml 0.0180 50 B 5.87 E-03 1.06 E-04 0.00
msample 249.55250 g 3 E-03 50 B -1.18 E-04 -3.53 E-07 0.00
ρwater .998210 g/ml 0.154 50 B 0.0294 1.06 E-05 0.00
f1 1000.0 9.7
f2 2000 22
f3 133.333 0.509
V100 100.00 ml 0.078 50 B -0.0299 -2.34 E-03 0.79
V0.75 0.7500 ml 0.0028 50 B 1.41 3.94 E-03 2.24
V0.1 0.1000 ml 0.0006 50 B 15 8.53 E-03 10.51
V0.05 0.0500 ml 0.0004 50 B 8.83 3.53 E-03 1.8
result :
-1
[Cdt] = 1.497 ± 0.054 ng ml , where the reported uncertainty is calculated using a coverage factor of
2.1 (95%).
Table 10: Combined uncertainty on concentration tracer cadmium in soil column effluents.
191
proxy
Each collected fraction is, as earlier stated, a unique combination of tracer and natural cadmium at a
certain concentration level. In figure 6 the expanded uncertainty (U = 2*u[Cd], 95 % confidence
intervals) on [Cdn] and [Cdt] for a total cadmium concentration level of 1 ng ml-1 was calculated in
function of the fraction natural cadmium 111 isotope (111fn).
14%
U / [Cdn] (95 % confidence)
V0.75
12% V0.05
10% V0.1
113Is
8%
111An
6% 113An
I2
4%
I1
2% IRh
0%
0.01
0.99
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
111
fn
Figure 6: The major contributions to the expanded uncertainty (U) on the concentration of natural cadmium for
-1 111
a total cadmium concentration level of 1 ng ml in function of the fraction natural cadmium 111 isotope ( fn).
Only the major contributions (> 99%) to the uncertainty are shown. The same example is given for
the expanded uncertainty on tracer cadmium in figure 7.
40%
35% V0.75
U / [Cdt] (95 % confidence)
V0.05
30%
V0.1
25% 113Is
111An
20%
Rt
15% I2
10% I1
IRh
5%
111Is
0%
0.01
0.05
0.15
0.25
0.35
0.45
0.55
0.65
0.75
0.85
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
111
fn
Figure 7: The major contributions to the expanded uncertainty (U) on the concentration of tracer cadmium for
-1 111
a total cadmium concentration level of 1 ng ml in function of the fraction natural cadmium 111 isotope ( fn).
192
proxy
For a 111fn value of about 0.2, which means that the fraction of natural cadmium on isotope 111 is 20
%, the expanded uncertainty on the concentration natural cadmium is approximately 4 %. At this
point the concentration of natural cadmium is 0.650 ng ml-1 (figure 8).
1.000
[Cd] ± U (95 % confidence)
0.800
0.600
0.400 natural cadmium

tracer cadmium
0.200
0.000
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
111
fn
Figure 8: The concentration of tracer and natural cadmium (± U) for a total cadmium concentration level of 1 ng
-1 111
ml in function of the fraction natural cadmium 111 isotope ( fn).
The impact of the uncertainty on the analytical measurement on estimating transport parameters
has been studied for the same example as presented in figure 5. The equilibrium convection-
dispersion model was fitted to the experimental data in order to predict the retardation (r). The
retardation determines the velocity at which an adsorbing component moves through a soil profile.
This parameter could be considered as an indication for the time needed for the contaminant to
reach the groundwater. The retardation of natural (rn) and tracer cadmium (rt) was obtained for
three different sets of experimental data. The first set consisted of the mean analytical
measurements ([Cd]), the second set included an alternate choice of the same analytical
measurements increased or decreased by their expanded uncertainty ([Cd] +/- U). The third set was a
combination of the mean analytical measurements ([Cd]), the same data increased ([Cd] + U) and
decreased ([Cd] - U) by their expanded uncertainty. The results are briefly summarised in table 11
together with a 95 % confidence interval of the estimated retardation. The difference in the
calculated retardation between the three different data sets is small compared to the extent of the
95 % confidence interval within one data set. Without going into detail on modelling aspects this
uncertainty on the retardation is mainly due to fitting of the equilibrium convection-dispersion
equation to the data.
data set rt (Cdt) 95 % confidence intervals rn (Cdn) 95 % confidence intervals

[Cd] 10.77 8.55 - 12.99 17.28 15.62 - 18.96
alternate [Cd] +/- U 10.75 8.47 - 13.03 16.94 15.44 - 18.43
[Cd], [Cd] +U, [Cd] - U 10.77 8.69 - 11.84 17.28 15.60 - 18.97
Table 11: Fitting of different experimental data sets with the equilibrium convection-dispersion model in order
to predict the retardation.
193
proxy
Conclusions
A calculation method for the simultaneous determination of tracer and natural cadmium with ICP-MS
in soil column effluents is described. A full uncertainty calculation method and an overview of the
uncertainty sources are given for the quantitative determination of tracer and natural cadmium with
ICP-MS. Although each collected fraction of the column effluent is a unique combination of tracer
and natural cadmium at a certain concentration level, the following general conclusions can be
formulated. At low concentrations of cadmium (< 0.2 ng ml-1 Cd), the uncertainty due to counting
statistics is the major source of uncertainty. At higher concentrations the signal instability, partially
due to the clogging of the sampling cone by calcium salts present in the leachant (CaCl2), forms the
largest contribution to the total uncertainty on the analytical measurement (4%, 95 % confidence
intervals). The full uncertainty on the analytical measurement was negligible for the estimation of
solute transport parameters.
Acknowledgements:
We wish to thank P. Van Bree (VITO), M. Berglund (IRMM) and C. Quetel (IRMM) for many helpful discussions
on the subject and F. Vanhaecke (RUG), L. Moens (RUG) and D. Mestach (AKZO NOBEL) for a critical review.
References paper, Full uncertainty calculation on quantitative determination of tracer and

natural cadmium in soil column effluents with ICP-MS, J. Anal. At. Spectrom., 1999, 14, 1475–1484.
1. Van der Zee, S.E.A.T.M. and van Riemsdijk, W.H., Wat. Resour. Res., 1987, 23, 11, 2059.
2. Wilkens, B.J. , Ph.D. Thesis, University of Utrecht, the Netherlands, 1995.
3. Boekhold, A.E. Ph.D. Thesis, Wageningen Agricultural University, The Netherlands, 1992.
4. Temminghoff, E.J.M., Ph.D. Thesis, Wageningen Agricultural University, the Netherlands, 1998.
5. Buchter, B., Hinz, C., Gfeller, M. and Flühler, H., Soil Sci. Soc. Am. J. ,1996, 60, 716.
6. Brusseau, M.L, Wat. Resour. Res., 1991, 27,4, 589.
7. Borkovec, M., Bürgisser, C.S., Cernik, M., Glattli, U., Sticher, H.,Transp. in Porous Media ,1996, 25,
193.
8. Selim, H.M., Buchter, B., Hinz. C., Ma L., Soil Sci. Soc. Am. J., 1992, 56, 1004.
9. ISO, Guide to the Expression of Uncertainty in Measurement, First Edition 1995, ISBN 92-67-
10188-9.
10. Rosman, K.J.R. and Taylor, P.D.P., Pure & Appl. Chem, 1998, 70 (1), 217.
11. Eurachem, Quantifying Uncertainty in Analytical Measurement, First Edition 1995.
12. McLaren, J.W., Beauchemin, D. and Berman S.S., Anal. Chem., 1987, 59, 610.
13. Beary, E.S., and Paulsen, P.J., Anal. Chem, 1993, 65, 1602.
14. Graham, S.M., and Robert, R.V.O., Talanta, 1994, 41, 1369.
15. Way, T.W., and Wiedmeyer, R.H., At. Spectrosc., 1998, 19 (5), 150.
16. Denoyer, E.R., At. Spectrosc., 1994, 15, 7.
17. Prudnikov E.D. and Barnes R.M., J. Anal. At. Spectrom, 1999, 14, 27.
18. Russ III, G.P. and Bazan, J.M., Spectrochim. Acta, 1987, 42B, 49.
19. van Heuzen, A.A., Hoekstra, T. and van Wingerden, B, J. Anal. At. Spectrom., 1989, 4, 483.
194
proxy
4.3. SOURCES OF NITRATE LEACHING
Fertiliser use is a major source of nitrate leaching into ground and surface water and constitutes an
important problem in many European Union (EU) member states [157]. To reduce leaching, the EU
Nitrate directive (91/676/EC) focuses on the pollution by nitrates from agriculture. Member states
are obliged to identify waters in which the nitrate concentration is above, or at risk of exceeding the
50 mg/l NO3- norm. Agricultural areas within the catchment area for these waters are then
designated as a nitrate-vulnerable zone, in which member states are obliged to draw up an action
programme and a code for good agricultural practise, involving, for example, fertilisation standards.
The Nitrate directive focuses primarily on the use of animal manure as fertiliser, because this practice
is difficult to manage. In Flanders, difficulties in predicting the exact nutrient availability and uptake
led to the imposition of a precautionary fertilisation standard for nitrate-vulnerable zones of 170 kg
N ha−1 year−1. Although the Nitrate directive has been in force now for 20 years, many EU member
states still experience difficulty in complying with its requirements and achieving environmentally
beneficial outcomes. In some European catchment areas, the nitrate concentrations are still
increasing.
The necessity for environmental management of surface and groundwater quality with respect to
nitrate (NO3-) contamination is therefore well agreed upon nowadays [158]. The current most
common approach to do so, is mainly based on monitoring NO3- concentration in water and soils at
various stations and times. Management programmes based on such a concentration approach can
be very demanding. As an example, in Flanders, the on-going NO3- management programme uses
circa 800 monitoring stations for surface water and circa 2 100 for groundwater. Water samples from
these locations are added to circa 28 000 soils samples from agricultural land to produce about 100
000 "classical" physical and chemical data points annually. Although this allows evaluating the water
quality status, this monitoring strategy has not been successful at providing a clear understanding of
the sources of NO3- pollution. Indeed, there is ample evidence (e.g., EC Implementation of Council
Directive 91/676/) showing that despite a high density of monitoring stations, and associated high
cost, N-species concentration data alone do not permit to establish adequately the nature and
respective contribution of NO3- sources responsible for water contamination. As a consequence, it is
extremely difficult to design and adjust environmental management plans and to evaluate their
effectiveness on the long run.
Due to a wide variety of potential sources, nitrogen (N) can enter the soil under several chemical
forms: nitrate (NO3-), nitrite (NO2-), ammonium (NH4+), organic nitrogen compounds, etc. … N cannot
be considered conservative because it is biologically modified through, for example, nitrification and
denitrification reactions that cause isotope fractionation, and ultimately modifies the isotopic
compositions of the dissolved N species. Since oxygen (O2) is usually available in unsaturated soils,
most of the N that reaches the groundwater has undergone nitrification and is present in its NO3-
oxidized form.
From the sole concentration measurements, it is generally hard to distinguish to what extent
different sources (mineral fertilisers, animal's manure or sewage effluents) are contributing to the
NO3- level observed in water. It requires an extended dataset of spatially and temporally staggered
concentration measurements, as well as detailed information on potential N sources in the
surrounding area, hydrology, soil characteristics, … Even when abundant data is available, budgeting
the different sources based on the sole concentration measurements remains difficult and highly
uncertain. Moreover, such a characterisation of the observed NO3- pollution might become very
expensive and time-consuming in areas exposed to multiple possible N sources.
195
proxy
Given the complexity of the biogeochemical nitrogen cycle, both the nitrogen isotope composition
(δ15N) and oxygen isotope composition (δ18O) of the NO3- molecule have been widely used to i) trace
its natural and anthropogenic sources, ii) identify transformation processes (e.g., denitrification,
nitrification and biological N2 fixation) and iii) assess the N budget in water bodies. Mixing processes,
which are generally controlling the input of NO3- pollution in water, lead to a modification of the
isotopic compositions (both δ15N & δ18O) of the dissolved NO3-. This mixing effect can be amplified by
the superimposition of transformation reactions such as natural denitrification. Therefore,
discriminating multiple NO3- sources based on the δ15N and δ18O isotope composition of NO3- alone
can sometimes become tricky, especially when secondary conversion processes occur.
Due to its ubiquitous nature, boron (B) commonly exists in groundwater as a minor constituent [159].
The large range of B isotope ratios observed in nature implies that significant contrasts between B
sources in groundwater are possible. Previous studies on the B isotope ratio as tracer of human
impact on water resources focused on the identification of waste water and sewage dominated by
synthetic B products, as well as on the impact of fly ash leachate. Boron, also presents the great
advantage of not being affected by NO3- conversion processes (i.e., natural denitrification does not
modify the isotopic composition of B; δ11B). Therefore, the B isotope ratio can be used to improve
the identification of sources of NO3- when NO3- transformation processes are involved.
Recent studies have shown that a multi-isotope approach, based on the combined use of δ15N and
δ18O of NO3- and δ11B is more successful to trace the origin of NO3- contamination compared to the
sole monitoring of corresponding concentrations [159].
A European Life demonstration project was carried out for the Alsace aquifer, which is located in the
Upper Rhine basin, between the German, French and Swiss borders. The aim of this ISONITRATE
project (http://isonitrate.brgm.fr/, last accessed on 20/01/2013) was to demonstrate to policy
makers and implementers that:
1. A water quality monitoring network, operated over years and integrating isotope data (δ15N,
δ18O & δ11B), inherently contains far more information than NO3- concentrations alone (in
terms of identifying pollution sources and assessing their respective contributions to the
pollution pressure). It is technologically feasible, and generates environmental benefits.
2. Such an isotope ratio monitoring methodology greatly enhances the understanding of NO3-
pollution in water and leads to more effective planning and design of environmental
management measures with respect to NO3- pollution in river basins and water bodies
defined by the Water Frame Directive (WFD).
The analytical part of the ISONITRATE project was carried out by the Bureau de Recherches
Géologiques et Minières (BRGM, France), the Department of Applied Analytical and Physical
Chemistry of Ghent University (ISOFYS, Belgium) and VITO. Based on both land use and field
knowledge from local authorities, four sampling sites were chosen in the Alsace aquifer. Ground and
surface water samples were taken 12 times between October 2007 and December 2008 and
analysed for the δ15N and δ18O values of NO3- and the δ11B value (see paper: Boron isotope ratio
(δ11B) measurements in WFD monitoring programs: comparison between double-focusing sector
field ICP (ICP-SFMS) and thermal ionization mass spectrometry (TIMS), J. Anal. At. Spectrom., 2010).
An International Workshop “Towards new methods to manage nitrate pollution within the Water
Framework Directive” was organised in Paris (10-11/12/2009) to present and discuss new
approaches to manage nitrate pollution.
196
proxy
Picture 12: International Workshop “Towards new methods to manage nitrate pollution within the Water
Framework Directive” (Paris, UNESCO building, 10-11/12/2009).
The outcome of the workshop resulted in a guideline for policy makers to identify nitrate polluters
and in which the use of compound-specific nitrogen (δ15N,), oxygen (δ18O), and bulk boron (δ11B)
isotope ratios to identify sources of nitrate-contaminated waters was outlined [160].
197
proxy
Boron isotope ratio (δ11B) measurements in WFD monitoring programs:

comparison between double-focusing sector-field ICP (ICP-SFMS) and thermal
ionization mass spectrometry (TIMS)
Kristof Tirez1, Wilfried Brusten1, David Widory2, Emmanuelle Petelet2, Agnès Bregnot2,
Dongmei Xue3, Pascal Boeckx3 and Jan Bronders1
1
2
Bureau de recherches géologiques et minières (BRGM), France
Ghent University, Faculty of Bioscience Engineering, Belgium
3

Abstract
The aim of our research was to compare δ11B measurements performed with thermal ionization mass
spectrometry (TIMS) and sector-field inductively coupled plasma-mass spectrometry (ICP-SFMS) and
evaluate the feasibility of implementing stable isotope methods in European water Framework
directive (WFD) monitoring programs. The comparison was based on δ11B measurements of 192
ground- and surface water samples and 15 leachates of nitrate pollution source materials (organic
and mineral fertilisers). The precision of δ11B measurements attainable with ICP-SFMS, 2σ = ± 2.6 ‰;
(n=192), is as expected lower than the precision achieved by TIMS, 2σ = ± 0.3 ‰ (n=183). However
the ease of use, rapidity and availability of ICP-SFMS on the one hand and the observed variability in
δ11B in ground- and surface water on the other (from -3.4 to +37 ‰), demonstrates that using ICP-
SFMS as an isotopic screening method would promote the use of isotopic methodology for WFD
monitoring.
Based on the results of the different case studies it is shown that retrieving precise information on the
identification of pollution sources from δ11B values requires reaching the best analytical precision and
accuracy possible. Hence, the superior precision of TIMS advantages tracing of nitrate pollution
sources. However for some cases, e.g., trying to decipher contributions between sources with really
distinct δ11B signatures (e.g., manure and sewage effluent), ICP-SFMS results lead to the same
conclusions and can therefore be used as a first approachable screening method for the
determination of δ11B in WFD monitoring programs.
Introduction
With the increasing precision of state-of-art mass spectrometry instruments in determining isotope
ratios, interest in isotopic fingerprinting techniques for a variety of elements is increasing. The
natural observed variations in isotope ratios can be used, among others, in i) identification of
archaeological artefacts, ii) food authentication, iii) provenance studies and tracing of sources of
contamination.
There is a considerable interest in determining variations of boron isotope ratios (11B/10B) in

geochemistry because of the wide natural range of 11B/10B ratios in rocks, sediments and waters.
Boron isotopes have been successfully used for tracing sources of anthropogenic input into ground-
and surface water,1-5 rainwater and deposition,6-8 freshwater lakes,9 landfill percolates10 and even
anthropogenic emissions in the atmosphere.11
198
proxy
A variety of mass spectrometric techniques are used for isotope ratio measurements, depending on
the demands of the specific application. The fundamentals and the use of plasma source mass
spectrometry for isotope ratio measurements have been reviewed by several authors,12-13 especially
since the introduction of multi-collector ICP-MS.14
The classical method for measuring the boron isotopic composition, δ11B, is thermal ionization mass
spectrometer (TIMS). It yields the highest degree of accuracy and precision (± 0.3‰). It is now well
documented that various ICP-MS techniques are also widely used for measuring isotope ratios, with
excellent results for multi-collector ICP-MS (MC-ICP-MS: δ11B ± 0.2‰), satisfactory results for double-
focusing sector-field ICP-MS (ICP-SFMS: δ11B ± 2‰) and poor results using quadrupole-based ICP-MS
(Q-ICP-MS: δ11B ± 15‰).15-24
As boron isotope ratios are increasingly being applied in geochemistry, the comparison of isotopic
measurements across different instrument types and techniques with respect to the demands of the
application is of concern. Recently, an inter-laboratory comparison of boron isotope measurements
was performed in order to address the correct reporting and comparison of isotopic measurements
across different instrument types and techniques.15 Kasemann et al. reported a comparison of
isotopic measurements with respect to boron isotope composition of marine carbonates to
reconstruct seawater pH values and atmospheric pCO2. 16
To distinguish nitrate sources, trace them in water and quantify their respective contributions,
research showed great added value of using a multi-isotopes approach including boron (δ15N-NO3,
δ18O-NO3 and δ11B).4 Nitrate contamination in water is a worldwide environmental problem and is of
special concern in the European Water Framework Directive.25 Mean nitrate concentrations in
groundwaters in Europe are above background levels but do not exceed the limit of 50 mg L-1 as NO3.
On average, groundwaters in western Europe have the highest nitrate concentration, due to the
most intensive agricultural practices, twice as high as in eastern Europe, where agriculture is less
intense. In the EU, it is estimated that mineral fertilisers account for almost 50 % of nitrogen inputs
into agricultural soils and manure for 40 %. The rate of percolation is often slow and excess nitrogen
levels may be the effects of pollution on the surface up to 40 years ago, depending on the
hydrogeological conditions. There are also other sources of nitrate, including treated sewage
effluents, which may also contribute to nitrate pollution in some rivers.
However chemical data alone, currently used in the different types of monitoring programs defined
in the Water Framework Directive, do not permit to establish unambiguously the type, location and
contribution of different sources of nitrate in a river basin. In particular, differentiating urban and
agricultural origin is difficult, even by increasing the number of monitoring stations or samples. This
information is nevertheless critical in defining correct measures to reduce the nitrate contamination.
Within the frame of the European Life ISONITRATE project, the aim of our research was to compare
δ11B measurements via TIMS and ICP-SFMS for tracing nitrate sources and evaluate the feasibility of
implementing stable isotope methods in European WFD monitoring programs.26 Nitrate
concentrations of more than 1000 groundwater monitoring stations across European countries are
reported to the European Environment Agency.25 However on a local scale, the frequency of quality
monitoring of surface and groundwaters with respect to nutrients is much larger. In Flanders alone,
more than 2000 groundwater and 4000 surface monitoring stations control on a regular basis the
nitrate content.27 Because of these high frequency monitoring requirements, the WFD
implementation has triggered the use of screening methodologies in particular for the detection of
accidental pollution or the control of water bodies at risk.28
For implementation and application of isotope techniques in WFD monitoring programs the ease and
feasibility of measurement methods is of primary concern. The peculiar advantages of sector-field
199
proxy
based ICP-MS include high sample throughput, low analysis cost, instrument robustness, sensitivity
and simple sample preparation. Moreover, these type of instruments are already available and
implemented in WFD monitoring laboratories to analyse the content of different contaminants in
ground- and surface water (Cd, Pb, ...). The more precise techniques (TIMS, MC-ICP-MS), on the other
hand, often require labour-intensive sample preparation, such as chemical purification of the analyte
and expensive equipment. The nitrate source tracking potential based on δ11B measurements with
TIMS and ICP-SFMS was evaluated in four different cases.
Natural variations of δ11B
A synthesis of boron isotope variations in nature has been reported amongst others by Barth.29 The
ratio between the two stable isotopes of boron, 11B and 10B, is usually referred under the δ11B
notation, given by equation 1:
 ( 11
B 10
)
B sample 
δ B = 
11
− 1 ⋅ 10 3 (equation 1)
 (
11
B 10
B)NIST 951 
where NIST SRM 951 (boric acid) is the accepted international reference material, with 11B/10B =
4.04362 ± 0.00137. The relative large mass difference between the two stable isotopes of boron
leads to a wide range of boron isotope variations in natural samples.30 Natural waters, such as
seawater, river water, rainwater, groundwater, brines, geothermal fluids and fumaroles’ condensates
encompass a range of δ11B of nearly 76 ‰.3 The lowest δ11B values at -16 ‰ are reported for
groundwater from the Artesian Basin in Australia, the most enriched reservoirs measured, to date,
are saline groundwater in Israel and brines from the Dead Sea and Australian salt lakes with δ11B
values up to + 60 ‰.
The dominant boron species in aquatic systems are B(OH)3 and B(OH)4-, which are in isotopic
equilibrium as shown in equation 2:2
11
B(OH) 3 + 10
B(OH) 4− ⇔ 10
B(OH) 3 + 11
B(OH) −4 (equation 2)
The calculated equilibrium constant for this reaction is 0.981 at 25°C. This implies that 10B is
preferentially present in the tetrahedral species, while 11B is enriched in the trigonal species. The
B(OH)4- species are preferably adsorbed by soil and minerals, leading to an enrichment of 10B in the
solid phase (fractionated by 30-40 ‰) when boron is incorporated from aquatic systems by
heterogeneous exchange, and a concomitant enrichment of 11B in the residual fluids. In contrast,
leaching of clay minerals (e.g., desorption) or extraction of fluid inclusions in crystalline rocks result
in low δ11B in the residual fluids. In aqueous solutions, the equilibrium between B(OH)3 and B(OH)4- is
pH-dependent (equation 3):
B(OH) 3 + H 2 O ⇔ B(OH) −4 + H + (equation 3)
At high pH values (pH > 11), B(OH)4- dominates, while B(OH)3 is the dominant form at pH < 7. An
equilibrium isotope fractionation can, therefore, only be expected if the aquatic system has a pH
between 7 and 11.
200
proxy
Anthropogenic influence of δ11B in ground and surface water
With respect to nitrate groundwater contamination from intensive agriculture, the main sources to
distinguish are mineral fertilizers, organic fertilizers (animal manure) and sewage. While Boron
concentrations in natural groundwater and surface water are generally low (< 0.05 mg L-1), the
contaminant sources are enriched in boron (0.1 – 1.5 mg L-1 in sewage effluent, > 1 mg L-1 in liquid
manure, up to 22 mg L-1 in mineral fertilizer leachate).31-32 Consequently the boron isotope
composition is sensitive to mixing of pristine and contaminated waters. Moreover the isotopic
composition of boron, as a nitrate co-migrant, is not affected by denitrification and can therefore be
used as a conservative tracer of mixing processes. 31-32 Figure 1 summarizes the boron isotope
composition ranges for the main anthropogenic sources. A synthesis of using coupled nitrogen and
boron (measured with TIMS) isotopes for tracing the sources of nitrate in groundwater has been
reported by Widory et al..4 A recent review of stable isotope methods for nitrate source
identification (δ15N-NO3, δ18O-NO3) was presented by Xue et al..33 In this paper it is shown that
especially in the case of differentiation between manure and sewage the δ15N-NO3 and δ18O-NO3
approach alone does not allow clear differentiation of the sources.
11
Figure 1: Literature overview of δ B values of main NO3 pollution sources in combination with data from the
ISONITRATE project [4].
A summary of boron isotope ratios and concentration data for the main nitrate contaminant sources
are given below.
Sewage water
The first studies of B isotopes as tracers of human impact on water resources have focussed on the
identification of wastewater and sewage dominated by synthetic B products. Sodium perborate
(either monohydrate NaBO 3 ⋅ H 2 O or tetrahydrate NaBO 3 ⋅ 4H 2 O ) is widely used as a
bleaching agent in a variety of domestic and industrial cleaning products. The raw materials are
mainly from large non-marine evaporate deposits in the USA (e.g., Boron, Searles Lake) and western
201
proxy
Turkey (e.g., Kirka) which account for almost 90 % of world production of sodium perborate. During
end use of perborate-enriched detergents and cleaning products, the anthropogenic water soluble
boron compounds are discharged with domestic aqueous effluents into sewage treatment plants,
where little or no boron is removed during conventional processing of the wastewater. Hence the
anthropogenic boron load is almost entirely released into the aqueous environment by entering a
receiving surface water system where further dilution occurs. Boron concentrations in secondary
effluents typically range from 0.1 to 1.5 mg L-1. Co-variations observed between B concentrations of
freshwaters and P concentrations or anionic detergent structures support the fact that sodium
perborate is to be considered as the major anthropogenic source of boron. In natural borate
minerals, δ11B ranges from -5.4 to 10.2 ‰ for Na-borates, from -16 to -1.1 ‰ for Na/Ca borates and
from -21.9 to -4.9 ‰ for Ca-borates. The rather narrow range in δ11B of Na-borate minerals allows
an isotope approach to distinguish a specific anthropogenic source of boron (mainly from industrial
perborate, the dominant use of mined boron) in a given natural aquatic system, characterized by a
distinct local background δ11B. The boron isotopic signature for a series of industrial sodium
perborate monohydrate and tetrahydrate products manufactured in Europe (Germany) were
reported by Barth.16 Sodium perborate monohydrate and tetrahydrate samples were characterized
by δ11B values ranging from -3.9 to +0.9 ‰ and -4.8 to +0.5‰, respectively. The total range in δ11B
values (-4.8 to +0.9‰) overlaps with the ranges of δ11B reported for non-marine Na-borate minerals
and commercial borax from the USA (-1.3 to +10.2 ‰) and Turkey (-5.4 to -1.7 ‰).
Fertilizers
Boron isotope signatures of inputs related to agriculture (e.g., hog manure, cattle feedlot runoff,
synthetic fertilizers) and the combination of N and B isotopes were first used in 1997 to distinguish
NO3- anthropogenic inputs to both ground- and surface water.31 For B to fulfil this role, however,
manure and fertilizers must contain detectable B with distinctive isotopic compositions.
Komer et al. reported averaged boron concentrations in liquid hog manure of 2.9 mg L-l (n=7), 31 for
cattle manure and poultry manure concentrations of respectively 1.8 and 13.4 mg kg-1 were
reported.32 In this study it was also shown that boron concentrations moderately correlated with
potassium, a soluble element that occurs mostly in urine, but not with phosphor, an element that is
mostly in faeces. These correlations are consistent with boron residing mainly in the urine
component of manure. Boron concentrations in fertilizers (on a dry weight basis) ranged from below
detection limit for some brands of ammonium nitrate and urea up to 382 mg kg-1 in magnesium
sulphate.31
However, the amounts of boron added to cultivated fields with fertilizers depend on the application
rates of the specific fertilizer and its boron contents. As an example, it was calculated by Komer et al.
that for typical liquid manure application rates a 0.28-0.42 kg B ha-1 is added, while for N mineral
fertilizers 0.05 kg B ha-1 and in case of some brands of urea or ammonium nitrate no detectable
boron was added.
In conclusion, boron isotopes can be used as tracers for discerning distinct solute sources in natural
waters since (i) boron is highly soluble in aqueous solutions, and therefore an ubiquitous minor or
trace constituent in nearly all water types, (ii) the boron isotopic composition is controlled by several
known parameters among which the solute source compositions and isotope fractionation processes
related to adsorption/desorption, mineral precipitation and dissolution, and volatilization are the
most relevant and (iii) the relative large mass difference between the two stable isotopes of boron
leads to a wide range of variations of boron isotope compositions in the nature.
However considering the concentration ratio of nitrate and boron in the different nitrate sources,
boron isotopes are mainly useful for tracing or discerning organic fertilizer (manure) and sewage
202
proxy
effluent (washing detergents). This is especially relevant considering the impossibility to clearly
differentiate between these two sources with the δ15N-NO3 and δ18O-NO3 approach alone.
Experimental
Instrumentation
Two distinct types of mass spectrometers were used to measure and compare δ11B values:
1. a single-collector double-focusing ICP-SFMS (ELEMENT II, ThermoFisher, Germany)
2. a single-collector thermal ionization mass spectrometer TIMS (MAT261, Finnigan®, Germany)
Water Samples
The ISONITRATE demonstration project relied upon a survey of 12 sampling campaigns over a 15
months period (October 2007 – December 2008).26 The pilot site is located in the Alsace region
(France), it is part of the Upper Rhine basin between the German-French-Swiss border near Basel in
the south and the mouth of the river Nahe near Bingen in the north. The site is flanked by the low
mountain ranges of the Vosges and the Pfälzerwald in the west as well as of the Black forest and the
Odenwald in the east.
Picture 13: Locations of the four sampling sites on the trans-boundary Alsace aquifer [158].
203
proxy
Because of an intensive agricultural land use, viniculture and the presence of industries and mining
activities in the Upper Rhine Valley, water on this pilot site is strongly impacted by anthropogenic
inputs. The average nitrate concentration of the groundwater in the Upper Rhine Valley is just under
30 mg L-1 nitrate and herewith indicates a substantial pollution of the groundwater.
Within the Alsace aquifer four distinct scenarios were selected:
i) Natural case (B1-2, boron corresponding to the local, natural recharge);
ii) Simple case (A1-4, the source of anthropogenic boron is unique);
iii) Complex case (C1-5, multiple distinct sources of boron involved);
iv) Denitrification case (D1-5,nitrate is reduced but should not affect the boron isotopic
budget).
Picture 14: Survey of 12 sampling campaigns over a 15 months period (October 2007 – December 2008).
204
proxy
The description of the collected water samples is given in table 1. All samples for δ11B measurement
were collected in polyethylene (PE) bottles and stored at about 5°C in a refrigerator until they were
analysed. Other parameters that were monitored are pH, Eh, EC, O2, T°, NO3, NH4, P, TOC, Ca, Mg, Cl,
Zn, B, and alkalinity.
code type
A1 Surface water
source of
A2 Surface water
Single
NO3
A3 Groundwater
A4 Groundwater
B1 Surface water
Unpolluted water
B2 Groundwater
C1 Surface water
sources of NO3 C2 Surface water
Multiple
C3 Groundwater
C4 Groundwater
C5 Groundwater
D1 Groundwater
Denitrification
D2 Groundwater
Natural
D3 Groundwater
D4 Groundwater
D5 Groundwater
Table 1: Description of ground - and surface water samples. Four different environmental contexts:
i) single source of NO3, ii) “unpolluted” water, iii) multiple sources of NO3, iv) natural denitrification.
Source material samples
After identification, local sources of anthropogenic inputs were sampled from farms or farming
cooperative and sewage stations. For the natural case, no pollution sources were sampled as this site
represents the natural/uncontaminated reference. For the simple case, which was supposed to be
impacted by a single pollution source consisting of mineral fertilizers used for viticulture, the
sampling of the main fertilizers was done in a local agricultural marketing cooperative (Pfaffenheim,
France). The fertilizers sampled are representative for usage in this specific basin. It thus appears that
even if mineral fertilizers are the dominant products, there is also a non negligible use of organic
fertilizers.
The complex case is located in the eastern part of the Sundgau in an area dominated by farming
(cows, horses), agriculture (maize, wheat, sugar beet, rape), direct waste water inputs to surface
water were also identified from detached houses which are not connected to a water treatment
plant. Most of the waste water of this area are collected and treated in the waste water treatment
plant of Sierentz located a few km south-east from the basin. The effluents of this waste water
treatment plant were sampled in February 2008. The solid residues of the water treatment are dried
in the Sierentz station to be used as fertilizers (dried mud) and were also sampled in February 2008.
The main livestock farming consist of cows and horses. Three samples of manures were sampled
directly from farms of this basin (April 2008) as well as a dunghill liquid effluent. Sampling of the
main fertilizers was done in a local agricultural marketing cooperative (Sierentz). The sampled
fertilizers are representative of the products used in this specific basin. The denitrification case is
located in the German part of the Alsace plain, the site mainly consists of vineyards. The main
mineral fertilizers used in this region were sampled in a local agricultural marketing cooperative (2
samples). A dunghill located nearby the D5 piezometer was also sampled. The source materials were
extracted with milli-Q water (L/S = 10) and the δ11B measurements were performed on the filtrated
leachates.
205
proxy
Reagents and standard solutions
Boron standard solutions were prepared form a 10 g L-1 commercially available standard (Spex
CertiPrep Inc., Metuchen, NJ). De-ionised water was purified by a Millipore Milli-Q system. The δ11B
vales were calculated based on standard reference material NIST 951a Boric acid. All solutions were
gravimetrically prepared in polypropylene bottles.
Optimisation ICP-SFMS
Many factors may affect precision and accuracy of isotope ratio determination by ICP-SFMS, among
which sensitivity, spectroscopic interferences, mass discrimination and dead time correction.
Sensitivity
Using the instrument settings given in table 2, a sensitivity of 100,000 counts per second (cps) per µg
B L-1 for 11B was obtained using Ni cones. In order to achieve better precision on the isotope ratio,
samples containing boron concentrations > 25 µg L-1 were diluted so that both isotopes were
measured in counting mode. The instrument is equipped with a perfluoroalkoxy copolymer resin
(PFA) nebulizer and spray chamber in order to reduce background level. For pure water, a reading of
~ 10,000 cps on 11B was obtained (~ 0.1 µg B L-1). Washing periods of 1 h are insufficient to reduce
the memory effect below 1%, even when using a range of different solvents (mannitol, ammonia).23
In our study, after 4 min rinsing with pure water, the signal intensity on 11B typically dropped to < 2 %
of the analyte signal intensity typical for the natural surface and groundwater samples. The blank
can, thus, be considered as negligible.
Nebulizer type PFA micro flow
Spray chamber PFA Scott type
RF power (W) 1050
-1
Cooling gas flow rate (L min ) 15
-1
Auxiliary gas flow rate (L min ) 1
-1
Nebulizer gas flow rate (L min ) 0.95
-1
Solution uptake rate (ml min ) 0.7
Ion extraction lens (V) -2000
Focus lens (V) -1140
Mass resolution (m/∆m) 300 (low)
Uptake time 4 min
Analysis time 4 min 15 sec
Rinsing time 4 min
Scan type E-scan
Detection mode Counting
10
Mass range 10.012-10.013 ( B)
11
11.008-11.010 ( B)
Mass window (%) 5
Settling time (s) 0.001
10
Sample time (s) 0.02 ( B)
11
0.005 ( B)
Samples per peak 100
Runs 10
Passes 200
Integration type Average
Dead time (ns) 10
11
Table 2: Optimised instrument settings for measurement of δ B with ICP-SFMS.
206
proxy
Spectroscopic interferences
In order to reduce the spectroscopic interference of 40Ar4+ on the 10B+ peak, the ratio frequency
power setting was reduced and the auxiliary and nebuliser gas flow optimised.1 The use of a
magnetic sector ICP-MS at a low resolution mode yields flat topped peaks. Higher signal intensity
coupled to the flat tops of the peaks at lower resolution is used for precise isotope ratio
measurements.34 In the low resolution mode, the peak width (located at 5 % peak height) for 10B is
0.033 amu (9.979–10.046 amu), and the instrument is set to divide this peak into 100 measurement
samples. The mass window for precise isotope ratio measurement is set to a narrow range of 5%,
meaning that the scanning range of the instrument around the boron peak represents ± 2.5% peak
width of the accurate mass and is centred in the flat top region 10.012-10.013.
Mass discrimination
The space charge effect is assumed to have the strongest influence on the total mass discrimination
in an ICP-MS.13 After the positively charged ion beam leaves the skimmer cone, the mutual repulsion
of ion limits the total number of ions which are transmitted by the optics. If an ion beam consists of
light and heavy ions, the light ions are deflected more extensively than the heavy ions, whereas the
heavy ions preferably remain in the central ion beam. The total mass bias can be experimentally
determined by the mass discrimination factor fMD:
fMD = Rtrue/Rmeasured (equation 4)

where R is the isotope ratio of the light over the heavy isotope. Due to the large percentage mass
difference between the two stable boron isotopes and the relatively low degree of ionization of
boron in the argon plasma of an ICP-MS instrument, the mass discrimination is significant. Based on
the alternate NIST SRM 951 measurements during the measurement run, the range of mass
discrimination per mass unit, MD, ranged between 7.5 an 10.5 %. Correction for mass discrimination
is performed by bracketing samples with NIST SRM 951, the average 11B/10B ratio of NIST SRM 951
measured before and after each sample is used to calculate the δ11B value of the bracketed sample,
which is the recommended routine procedure for δ11B analysis.15 The correction for matrix-induced
mass discrimination was also investigated by Gäbler et al. by analyzing seawater. They found, with
instrumental settings similar to ours, comparable results for δ11B measurements by both ICP-SFMS
and NTIMS.1
Dead time
Boron isotope ratios were calculated from dead-time corrected intensities. Because of the 4 fold
difference in natural abundance of B isotopes, a dead time correction in counting mode is necessary
to obtain concentration-independent and accurate values of δ11B. The dead time is iteratively
deduced from the measurement of the 235U/238U isotope ratio in 0.4, 0.6, 0.8 and 1 µg U L-1 standard
solutions according to the manufacturer instructions. The optimized dead time obtained was 10 ns.
In conclusion, using the typical optimized instrumental settings as summarized in table 2, single δ11B
measurement with ICP-SFMS can be performed without sample pre-treatment within 12 minutes
(before starting a series of boron measurements, rinsing the ICP-SFMS instrument with Milli-Q water
is recommended). For the routine measurement of δ11B with ICP-SFMS, the following procedure was
used on a set of 16 samples per campaign.
The instrument is tuned to maximum sensitivity for boron (NIST SRM 951 solution of 25 µg L-1), by
tuning ion lenses and adjusting the nebulizer gas flow rate. Subsequently, in order to reduce the
207
proxy
spectroscopic interference of 40Ar4+ on the 10B+ peak, the auxiliary and nebuliser gas flow are further
optimised.
Samples are analysed by bracketing with NIST SRM 951 standards, the average 11B/10B ratio of NIST
SRM 951 measured before and after each sample is used to calculate the δ11B value of the bracketed
sample, which is the recommended routine procedure for δ11B analysis.15
Boron concentration measurements
When using ICP-SFMS an advantage, compared to TIMS, is that in one single analysis both boron
concentrations and its isotope composition can be obtained simultaneously. For the determination
of the boron concentration in the water samples, an external calibration line where Be was added
on-line as internal standard was used. In the frame of this project it was evaluated if the ICP-SFMS
method optimised for isotopic analysis can be used for the quantification of the concentration. In
this case, the average of the 11B signal (cps) of NIST SRM 951 before and after the measurement of
the 11B signal (cps) of the sample was used to calculate the B concentration of the sample. The
concentration of Boron in the prepared NIST SRM 951 standard amounted 27.5 µg L-1.
Optimisation TIMS
For Boron isotope compositions (δ11B) in water, sample volume is determined to ultimately yield a
quantity of 6 to 10 µg of boron. Samples then undergo a two-steps chemical purification using
Amberlite IRA-743 selective resin (method adapted from Gaillardet et al.).35 First, the sample (pH ~7)
is loaded on a Teflon PFA® column filled with 1ml resin, previously cleaned with ultrapure water and
2N ultrapure NaOH. After cleaning again the resin with water and NaOH, the purified boron is
collected with 15 ml of sub-boiled HCl 2N. After neutralisation of the HCl by Superpur NH4OH (20%),
the purified boron is loaded again on a small 100µl resin Teflon PFA® column. Boron is collected with
2ml of HCl 2N. An aliquot corresponding to 2µg of boron is then evaporated below 70°C with
mannitol in order to avoid boron loss during evaporation.36 The dry sample is loaded onto a tantalum
(Ta) single filament with graphite (C), mannitol (C6H8(OH)6) and cesium (Cs). δ11B are then determined
by measuring the Cs2BO2+ ion.37,38 The analysis is run in dynamic mode by switching between
masses 308 and 309. Each analysis corresponds to 10 blocks of 10 ratios. Samples are always run
twice. Total boron blank is less than 10 ng corresponding to a maximum contribution of 0.2%, which
is negligible. Seawater (IAEA-B1) is purified regularly in the same way in order to check for a possible
chemical fractionation due to an uncompleted recovery of boron, and to evaluate the accuracy and
reproducibility of the overall procedure.39-41 Reproducibility was obtained by repeated
measurements of the NBS951, accuracy and reproducibility are controlled with the analysis of the
IAEA-B1 seawater standard.
Statistical evaluation
The Bland-Altman technique was used to assess agreement between two measurement methods.
This technique was conducted in this study to compare results obtained via the TIMS and ICP-SFMS
for the determination of δ11B. The average ( d ) and standard deviation ( sd ) of the difference (d)
between the measurement results of two methods on the samples were computed. If the differences
are normally distributed, and 95% of the differences lie between d – 1.96 sd and d + 1.96 sd
(termed “95% limits of agreement”), the two analytical methods can be used interchangeably.
208
proxy
The implementation of δ11B and boron concentration measurements in large scale WFD monitoring
programs requires high sample throughput. The discussion on the nitrate source tracking potential is
therefore focussed on the attainable analytical performances of δ11B measurement with a ICP-SFMS
instrument operating without additional sample pretreatment on the one hand versus a TIMS
instrument including labour intensive matrix separation on the other. The interpretation on the use
of the multi-isotopes approach (δ15N-NO3, δ18O-NO3 and δ11B) to distinguish nitrate sources will be
presented elsewhere.26
The average results of the characterisation of the water samples during the monitoring campaign (12
sampling campaigns over a 15 months period) are summarised in the tables below.
Eh Eh Cond (µs/cm) Cond (µs/cm) O2

T (°C) pH
(mV) (mV NHE) à T°C à 25°C (% saturation)
A1 10.4 8.0 89 317 542 778 85
A2 10.4 8.0 143 352 394 567 89
A3 13.1 7.0 143 352 463 624 66
A4 13.2 7.4 204 423 464 611 60
B1 8.2 7.4 157 368 37 62 91
B2 8.1 6.9 194 405 24 38 84
C1 8.9 8.0 194 403 578 804 79
C2 8.9 8.2 179 389 576 827 83
C3 11.7 7.1 125 343 634 898 51

C4 11.8 7.1 148 364 588 838 72
C5 11.9 7.4 126 336 508 720 44
D1 8.1 5.0 265 481 28 43 87
D2 12.0 6.1 207 425 344 481 75
D3 13.5 7.0 96 307 594 780 55
D4 10.9 6.2 104 316 132 177 18
D5 12.5 7.1 146 361 503 691 7
Ca Mg Na K NO3- Total N Cl- SO42- DOC DIC Dry cont. Ash Cont.
mg/l mg/l mg/l mg/l mg N/l mg N/l mg/l mg/l mg/l mg/l mg/l mg/l
A1 114 34 10 1.9 1.5 1.7 20 70 2.6 76 485 349
A2 81 22 7.9 2.8 1.5 2.2 14 47 1.9 58 362 267
A3 81 15 15 18 10 12 29 53 0.8 49 422 292
A4 100 16 8.3 1.6 20 23 39 36 0.7 44 474 258
B1 4.9 1.2 2.5 0.8 0.6 0.9 3.6 3.4 1.6 3.4 48 29
B2 4.3 1.1 1.5 0.3 0.7 1.0 1.2 3.2 0.9 3.1 35 16
C1 130 23 9.0 2.3 7.8 8.1 35 32 3.1 79 568 403
C2 131 24 12 3 8.0 8.6 41 39 3 77 560 389
C3 147 22 12 3.2 18 20 37 34 2.5 81 597 395
C4 139 22 12 3.7 13 15 29 39 2.3 81 541 378
C5 112 20 11 3.7 3.9 4.0 38 33 1.2 71 461 333
D1 2.4 0.6 1 1.7 1.8 2.0 1.7 5.4 0.9 1.5 29 16
D2 59 8.8 13 3.2 17 20 23 100 0.8 11 382 229
D3 120 18 15 1.4 9.0 9.8 24 45 1.7 76 491 363
D4 25 3.5 2.0 0.8 0.7 1.1 1.3 7.5 6.7 21 134 93
D5 126 10 7.0 1.0 7.9 7.0 29 37 0.9 66 467 329
Table 3a: Average results of of the characterisation of the water samples during the monitoring campaign .
209
proxy
Boron concentration levels
An overview of the measured concentrations of boron and nitrate over the different locations is
given in figure 2 and table 3b. The Nitrate Directive (91/676/EEC) aims to control nitrogen pollution
and requires member states to identify groundwaters that contain more than 50 mg L-1 nitrate or
could contain more than 50 mg L-1 nitrate if preventative measures are not taken.
Figure 2: General overview of the coupled variations of nitrate and boron concentrations for the different case
studies. Empty symbols represent surface water samples.
Ground waters A4, D2 and C3 are well above the current limit of nitrate, A3, C4, D3 and 5 are around
the limit and C5, D1 and B1 are below the limit.
[B] µg L-1 2σ min max [NO3] mg N L

-1
2σ min max NO3/B
A1 30 15 22 50 1.5 0.5 0.9 1.8 50
A2 23 10 16 31 1.5 0.9 0.9 2.2 70
A3 70 19 60 89 10.1 1.2 9.5 11.2 140
A4 8.7 3.2 7 12 20.0 2.2 18.0 21.6 2300
B1 4.4 2.0 3 7 0.65 0.2 0.49 0.86 150
B2 3.3 2.2 2 5 0.70 0.2 0.50 0.86 210
210
proxy
C1 25 12 18 40 7.8 2.5 5.0 9.4 320

C2 32 13 25 44 8.0 2.6 5.4 9.3 250
C3 43 32 32 92 17.3 4.4 12.0 21.0 400
C4 46 7 41 51 12.2 4.5 7.2 14.0 260
C5 39 8 35 48 3.8 2.3 2.5 5.8 100
D1 3.7 3.4 1 6 1.8 0.2 1.6 1.9 480
D2 19 8 14 26 17.0 2.4 15.0 18.8 900
D3 46 8 40 57 9.0 1.2 8.3 10.1 200
D4 8.2 4.8 5 11 0.7 0.9 0.1 1.6 80
D5 11 4 9 15 8.0 3.1 4.9 10.1 720
Table 3b: Average boron (B) and nitrate (NO3) concentrations.
The average B concentration for the 16 sites ranged from 3 to 70 µg L-1 (table 4). The variability
during the 15 months sampling campaign per location ranged from ± 2 (B1) to ± 32 (C3) µg B L-1 (2σ,
n=12). For the natural case, the average concentration was 3.8 ± 1 µg B L-1 (2σ, n=12).
Time of sampling (month/year)

-1
[B] µg L 10/ 11/ 12/ 01/ 02/ 03/ 04/ 06/ 08/ 10/ 11/ 12/ mean 2*σ
07 07 07 08 08 08 08 08 08 08 08 08 µg L-1
A1 22 26 28 28 24 29 25 37 28 50 34 34 30 15
A2 23 21 21 19 18 18 16 28 24 25 31 29 23 10
A3 70 71 63 66 60 62 60 65 79 82 74 89 70 19
A4 7.7 8.3 7.5 9 7.2 7 7.1 8.6 12 9.9 8.5 11 8.7 3.2
B1 5.2 4.2 4.6 4.5 4.4 4.6 2.7 4.8 4.3 6.5 3.2 3.5 4.4 2.0
B2 3.7 5 4.3 3.7 3.4 3.8 1.5 2.4 2.1 4.7 2.5 2.3 3.3 2.2
C1 19 25 25 21 18 21 19 23 26 31 28 40 25 12
C2 31 42 30 26 25 28 25 31 44 39 33 32 32 13
C3 34 35 32 37 36 38 38 40 42 44 47 92 43 32
C4 42 44 42 46 45 48 47 41 51 50 47 51 46 7
C5 37 41 36 41 42 42 40 35 35 37 36 48 39 8
D1 4.5 6.2 5.7 5.1 4.7 0.6 2.6 2.1 3.6 3.1 2.9 3.7 3.4
D2 15 23 26 16 15 16 14 19 17 19 23 24 19 8
D3 43 45 40 44 44 45 43 45 48 49 46 57 46 8
D4 9.4 8.1 5.3 10 5.2 5.1 5 9.4 11 11 9.5 9.9 8.2 4.8
D5 11 10 9.2 8.5 9.6 9.3 8.7 12 13 13 13 15 11 4
Table 4: Boron concentrations measured by ICP-SFMS.
A comparison between the B concentration in 72 water samples measured with ICP-SFMS using an
external calibration line on the one hand and the bracketed NIST SRM 951 on the other hand showed
a correlation coefficient of 0.99 (y = 1.02 x + 0.12). The average measurement difference of boron
concentration was 0.4 µg L-1 and the 95 % limits of agreement according to the Bland-Altman
technique amounted –3.4 and 4.3 µg L-1.
The large variation in boron concentration levels observed between the different collected samples
in combination with (in most cases) low variability per sample makes it possible to use the data for
interpretation. However, there is no general correlation between the boron and nitrate
concentrations, e.g., the most contaminated nitrate groundwater A4 has a concentration in boron
similar to the natural background. This is in line with the remarks by Komer et al. that the amounts of
211
proxy
boron added to cultivated fields with fertilizers depend on the application rates of the specific
fertilizer and its boron contents.31
When using boron as a co-migrant tracer of nitrate, it is important to consider the extracted Ntotal/B
concentration ratio from the different pollution sources. In general the following order in the Ntotal/B
ratio can be derived (table 5): mineral fertilizer (ammonium nitrate, urea): 1/106; mineral fertilizer
(NPK): 1/103; organic fertilizer: 1/102; sewage effluent: 1/10. In the case of A4, high nitrate
concentration in combination with low boron concentration could indicate the use of mineral
fertilizers.
NH4 NO3 NO2 Ntot B Ntot/B

Name Source Type Sampling mg N L-1 mg N L-1 mg N L-1 mg N L-1 µg L-1
location
NPK 14-7-17 Mineral fertilizer 7600 5800 <0.15 13000 13900 940
source of
NPK 15-5-20 Mineral fertilizer 10000 6300 <0.15 15000 14400 1040
Single
NO3
Fumeterre Organic/mineral fertilizer 20 45 0.8 160 380 420
Orgaveg Organic/mineral fertilizer 270 0.005 - 630 1300 490
Ammonium nitrate 27% Mineral fertilizer 14000 5300 <0.15 24000 20 1200000
NPK 13-13-21 Mineral fertilizer - - - - - -
Multiple sources of NO3
NPK 18-46 Mineral fertilizer - - - - - -

Urea 46% Mineral fertilizer 220 3.9 - 43000 70 614000
Cow manure Organic fertilizer 12 1.5 0.8 44 150 300
Cow manure Organic fertilizer 16 1.3 74 120 290 410
Cow manure - liquid Organic fertilizer 0.5 2.5 <0.15 11 130 80
Horse manure Organic fertilizer 0.5 0.23 <0.15 30 220 140
WWTP - dry mud WWTP - dry mud 5.4 230 <0.15 810 220 3700
WWTP effluent WWTP-effluent 0.9 1.4 <0.15 1.5 130 12
Ammonium nitrate 27% Mineral fertilizer 17000 13000 <0.15 27000 30 900000
Denitrific
Natural
ation
NPK 14-8-13 Mineral fertilizer 9800 3500 - 13000 2700 4800

Cow manure Organic fertilizer 1.1 2.5 0.2 60 410 150
Table 5: Chemical characterisation of the extracts of collected nitrate pollution sources (L/S = 10, the WWTP
effluent was measured directly).
Comparison of δ11B measurement by ICP-SFMS and TIMS on water samples
The average δ11B values varied over the 16 sampling sites from -3.4 (C5) to 37‰ (D2) (table 7). The
δ11B variability during the 15 months sampling campaign per location ranged from ± 1.0 to ± 15 ‰
(2σ, n=12). These data show that single sampling and analysis will not lead in all cases to a correct
interpretation of the isotope ratios and this has to be considered when implementing in monitoring
programs. On the other hand, the observed large variation in δ11B values allows clear discernible
distinction between samples.
An estimate of the analytical performance characteristics of the δ11B measurement by ICP-SFMS was
derived based on the measurements of the bracketed NIST SRM 951 samples. In a typical
measurement run the 16 samples of the sampling campaign were bracketed in between the
measurement of 17 NIST SRM 951 samples. Per measurement run 15 δ11B values of NIST SRM 951
were derived, e.g., δ11B of NIST SRM 951 (3rd position) was calculated using the average 11B/10B ratio
from NIST SRM 951 (2nd position) and NIST SRM 951 (4th position). The average of the in such a way
calculated δ11B values of NIST SRM 951 (theoretical value = 0) in the course of the project amounted
-0.096 ± 2.6 ‰ (2σ, n=192). The analytical performance is in agreement with previous reported
precision values of δ11B measured by ICP-SFMS.1,15
212
proxy
For TIMS measurement, reproducibility was obtained by repeated measurements of the NBS951 and
the accuracy was controlled with the analysis of the IAEA-B1 seawater standard (δ11B = 38.6 ± 1.7‰).
The 11B/10B ratio of replicate analyses of the NBS951 boric acid standard (after oxygen correction)
was 4.05045±0.00130 (2σ, n = 183). The reproducibility of the δ11B determination was ±0.32‰ (2σ).
The mean value obtained on δ11B of seawater was = 39.21 ± 0.31 ‰ (2σ) (n = 20).
The individual δ11B values measured with ICP-SFMS and TIMS on the water samples are summarized
in respectively table 6 and table 7.
11
δ B 10/07 11/07 12/07 01/08 02/08 03/08 04/08 06/08 08/08 10/08 11/08 12/08 mean 2*σ
A1 6.6 5.4 5 4.3 3.9 3.3 4.5 0.8 6.4 0.6 4.8 4.4 4.2 3.8
A2 4.4 4.4 8.3 7.9 7.7 5.4 7.9 3.4 2.6 2.5 8.6 6.4 5.8 4.6
A3 14.8 14.7 16 18.1 16.9 16.3 19.3 18.5 19 14 15.6 17.3 17 3.5
A4 8.5 7.9 10.7 6.5 7.3 8.3 8.4 9.6 9 8.2 9 9.5 8.6 2.2
B1 12.7 15.3 17 16.1 16.6 19.3 16 14.3 12.8 16.1 17.8 18 16 4.0
B2 22.7 24.9 27.6 26.9 27 24.9 25 30.8 31 27.9 28.2 25.1 27 5.0
C1 -1.2 4.3 4.4 4.9 1 2.8 -2.2 -0.9 1.3 3.7 4.5 3 2.1 5.0
C2 3.9 5.4 5.5 7.1 4.5 3.9 8.9 4.2 5.1 2.4 5.3 6.7 5.2 3.4
C3 4.1 7 4.1 5.3 5.5 4.9 10.2 5.9 8.1 6.1 9.9 10.6 6.8 4.7
C4 3 2.3 -0.6 0.3 0.5 -1.9 -0.6 3.6 2.7 -0.5 2.5 0.7 1.0 3.5
C5 -0.3 -1.1 -2.3 -1.8 -2 -3.9 -1.3 -3.9 2 -3.4 -0.5 0.9 -1.5 3.7
D1 27.2 - 25.3 26 25.7 27.3 27.9 29.4 29.5 26.9 27.3 31.7 28 3.8
D2 41.5 28.3 23.8 38.9 37.2 35.2 40.2 45.4 42.8 43.4 35 39.7 38 12.6
D3 12.1 14.9 13.9 15 14.6 13.7 14 14.8 17 13.8 14 14.4 14 2.3

D4 6.9 8.5 13.6 14.2 13.4 9.5 12.9 10.2 8.1 9.4 11.8 12.9 11 4.9
D5 8.6 4.5 6.8 6.7 6.1 6.8 9.3 8.3 8.4 5.8 6.4 9.4 7.3 3.0
11
Table 6: δ B values (in ‰ vs. NBS951) measured with ICP-SFMS during ISONITRATE sampling campaign.
11
δ B 10/07 11/07 12/07 01/08 02/08 03/08 04/08 06/08 08/08 10/08 11/08 mean 2*σ
A1 3.9 3.3 3.3 3.8 3.1 3.1 1.6 2.3 3.1 4.0 3.9 3.2 1.4
A2 3.4 4.3 6.4 6.3 6.1 6.1 5.2 3.9 3.4 5.3 6.1 5.1 2.4
A3 16.7 16.4 16.3 17.7 17.2 17.2 17.7 17.4 15.6 17.7 16.9 17 1.3
A4 9.0 8.8 8.0 8.8 9.2 9.8 8.9 10.1 7.1 10.1 10.6 9.1 2.0
B1 12.4 18.2 18.6 18.8 19.3 21.5 17.2 16.0 11.9 21.7 22.3 18 6.9
B2 28.6 30.9 31.7 28.3 31.0 29.7 31.1 34.3 31.9 34.1 34.6 32 4.3
C1 -0.4 3.3 2.9 3.2 1.1 1.9 1.0 1.9 1.1 4.7 3.6 2.2 2.9
C2 4.1 4.6 3.7 3.4 2.6 3.8 3.4 2.6 4.5 4.3 3.0 3.6 1.4
C3 4.2 3.8 4.0 4.2 4.5 4.1 4.1 4.2 4.7 6.0 6.4 4.6 1.7
C4 -1.3 -1.6 -1.5 -1.7 -1.4 -2.1 -2.0 -0.4 -1.4 0.1 0.2 -1.2 1.6
C5 -2.8 -3.4 -3.7 -4.1 -4.2 -5.3 -5.3 -3.9 -3.0 -1.0 -1.1 -3.4 2.9
D1 - - 27.5 26.7 29.8 30.2 25.4 30.6 30.9 31.2 32.4 29 4.7
D2 42.0 27.4 20.8 38.9 40.4 36.1 39.4 44.1 44.1 44.5 33.1 37 15.1
D3 13.3 13.2 13.5 13.3 13.0 13.5 13.5 13.8 14.0 14.7 14.9 14 1.2
D4 - 11.2 14.7 12.6 11.2 11.7 11.6 10.4 9.9 12.8 14.1 12 3.1
D5 6.5 6.5 6.4 7.1 7.1 6.6 7.2 7.7 7.3 7.8 7.5 7.1 1.0
11
Table 7: δ B values (in ‰ vs. NBS951) measured with TIMS during ISONITRATE sampling campaign.
There is a positive linear relation between the δ11B values measured by the two methods (figure 3, y
= 0.90x + 0.86) with a high correlation coefficient (r = 0.98).
213
proxy
11
Figure 3: Relation between δ B determined by both TIMS and ICP-SFMS for all water samples. The solid line
represents the 1:1 line (y=x), calculated linear regression equation is y=0.90x+0.86 (r=0.98, n=176).
The average difference in δ11B values measured by TIMS and ICP-SFMS is -0.3‰ (figure 4) and there
is no tendency for the difference to vary with variation of isotope ratios. The Kolmogorov-Smirnov
normality test showed that differences of the δ11B as determined by TIMS and ICP-SFMS (p = 0.23)
were normally distributed. The limits of agreement within which 95% of the differences are expected
are calculated according to the Bland-Altman technique as -5.4 and +4.8‰.
11
Figure 4: Bland-Altman comparison of the TIMS and ICP-SFMS δ B determinations of water samples collected
during the ISONITRATE project. The solid line represents the average difference between both methods, while
the dashed lines represent the 95% limits of agreement.
However, for the 3 locations with a boron concentration less than 5 µg L-1 (B1, B2 and D1), the
difference in δ11B values measured by TIMS and ICP-SFMS are significantly larger (see tables 6 and 7),
214
proxy
which indicates that there is a tendency for the difference to vary with (low) boron concentration.
Further study of the δ11B value measured by ICP-SFMS is needed to attribute these observed
differences to the influence of the matrix and/or the boron concentration level.
The consequences of the difference in analytical precision of δ11B measurements with TIMS and ICP-
SFMS with respect to the interpretation to distinguish nitrate sources, is discussed for the complex
case (C1-C5) and the denitrification case (D1-D5).
Complex case
In figure 5 (left), the δ11B values measured with TIMS versus 1/B for the water samples collected at
the complex case are compared to the ranges measured for local nitrate pollution sources.
Figure 5: Comparison of the identification of sources of nitrate pollution by coupling the reciprocal of B content
11
and δ B values of the water samples from the multiple sources case study. The boxes on the left represent the
11 11
range of the measured δ B values of the different nitrate pollution sources. δ B vales are determined by TIMS
(left) and ICP-SFMS (right) (error bars represent the 95 % confidence limits).
Based on the TIMS results, surface waters (C1 and C2) together with groundwater C3 yield similar
δ11B values, slightly positive (δ11B = 2 to 6‰), whereas both C4 and C5 (groundwaters) were depleted
in δ11B. When plotting δ11B versus 1/B, three different scenarios appear: (i) samples C1, C2 and C3
define a negative trend according to δ11B = -129.8 * (1/B) + 7.9 (n = 33, R²=0.75).
This trend characterises potentially the input of an organic fertilizer (with a δ11B around 8 ‰), in
agreement with the local measured manures and the δ15N-NO3 and δ18O-NO3 data;26 (ii) C4 samples
present a very homogeneous signature over the hydrological cycle.
This sampling site is affected by mineral fertilizer/wastewater, with wastewater being probably more
realistic considering the depleted δ11B of the pollution source (δ11B~ -6 ‰) and the characterization
of the local sources (table 8); (iii) C5 samples present the most negative δ11B signatures, in
agreement with the signatures of the WWTP effluents (locally identified).4
215
proxy
11 11
δ B δ B
B
(TIMS) (ICP-SFMS)
Sampling -1
Name Type µg L ‰ ‰
location
NPK 14-7-17 Mineral fertilizer 13900 0.2 0.8
source of
NPK 15-5-20 Mineral fertilizer 14400 0.4 1.6
Single
NO3
Fumeterre Organic/mineral fertilizer 380 9 13.6
Orgaveg Organic/mineral fertilizer 1300 22.6 20.8
Ammonium nitrate 27% Mineral fertilizer 20 -1.4 -1
NPK 13-13-21 Mineral fertilizer - 24.6 -
Multiple sources of NO3

NPK 18-46 Mineral fertilizer - 14.1 -
Urea 46% Mineral fertilizer 70 20.6 20.6
Cow manure Organic fertilizer 150 10.9 11.1
Cow manure Organic fertilizer 290 6.2 7.8
Cow manure - liquid Organic fertilizer 130 5.8 5.7
Horse manure Organic fertilizer 220 17.4 14
WWTP - dry mud WWTP - dry mud 220 -3.5 -4.2
WWTP effluent WWTP-effluent 130 -2.8 -0.3
Ammonium nitrate 27% Mineral fertilizer 30 - -2.7
Denitrification
NPK 14-8-13 Mineral fertilizer 2700 - 7.2

Natural
Cow manure Organic fertilizer 410 8 8.2
11
Table 8: Comparison of δ B (in ‰ vs. NBS951) values measured by both TIMS and ICP-SFMS on leachates from
NO3 sources.
Based on the ICP-SFMS measurements as shown in figure 5 (right) a negative trend according to δ11B
= -195.8 * (1/B) + 10.8 (n = 33) with a worse correlation (R² = 0.34) is found. The results are visually
more scattered and there is overlap between the C3, C4 and C5 measurement areas. In contrast to
the TIMS data, no trend to a distinct potential nitrate pollution source can be characterized.
These findings illustrate that end-users have to keep in mind that retrieving precise information on
the identification of pollution sources from δ11B values requires reaching the best precision and
accuracy possible.
Denitrification case
On the other hand, as presented in figure 6 for the denitrification case, both analytical methods
come to a same conclusion when trying to decipher contributions between sources with really
distinct δ11B signatures: (i) the signature of D1 values centred around 30‰, close to the value
expected for boron from natural rainwater origin; (ii) D2 (i.e., the samples with the highest NO3
concentrations, presumed to represent the closest agreement with the input of pollution source),
displays large δ11B variations (from 20 to more than 40‰) together with large variation of boron
concentrations (14 to 27 µg L-1). D2 reaches values that are higher (> 40‰) than all the measured
pollution sources. In the light of the available results it is not possible to conclude if other nitrate
sources are present or other processes occur (i.e., interactions with clay minerals); (iii) D3 to D5
display intermediate boron isotopic compositions (5<δ11B<15‰) that may correspond to the mixing
of different pollutions sources including mineral fertilizer, manure and sewage.
216
proxy
Figure 6: Comparison of the identification of sources of nitrate pollution by coupling the reciprocal of B content
11
and δ B values of the water samples from the natural denitrification case study. The boxes on the left
11 11
represent the range of the measured δ B values of the different nitrate pollution sources. δ B vales are
determined by TIMS (left) and ICP-SFMS (right) (error bars represent the 95 % confidence limits).
Comparison δ11B measurement ICP-SFMS and TIMS on source material

Table 8 shows comparable results between δ11B values measured with TIMS and ICP-SFMS on the
leachates of the different pollution sources (correlation coefficient of 0.96, y = 0.88 x + 1.10; due to
limited results (n = 13) no further Bland-Altman statistics were performed). The extracted boron and
nitrate contents (table 5) are in line with previous reported results on manure and mineral
fertilizers.31,32 These results show the method robustness and the ease of using one single ICP-SFMS
method for δ11B measurements of both water samples and leachates of the source materials.
An overview of literature data of δ11B in combination with the data of source materials collected
during the ISONITRATE project are summarised in figure 1. The measured δ11B values are in
agreement with previous published data, but it should be stressed that for interpretation of the
multi-isotopic approach accurate and precise δ11B data of local source materials need to be included
in the monitoring program. As an example the mineral fertiliser NPK 13-13-21, collected in the
multiple sources of nitrate case, showed based on literature unexpected high δ11B value (NPK 13-13-
21 : δ11B = + 24.6; table 8). This may indicate that the origin of boron in this specific mineral fertiliser
material is different than previously measured mineral fertilisers.
Conclusions
During the last decade, the number of isotope systems currently being explored in new investigations
and (routine) application fields has exploded. As both the number of techniques being developed and
the number of laboratories making these measurements increase, it is important to evaluate the fit-
for-purpose of measurement techniques for (routine) application.
An evaluation of boron isotope compositions measured in parallel by both ICP-SFMS and TIMS was
performed in the ISONITRATE project. Based on the results of the different case studies it is shown that
end-users have to keep in mind that retrieving precise information on the identification of pollution
217
proxy
sources from δ11B values requires reaching the best analytical precision and accuracy possible.
However for some cases, e.g., trying to decipher contributions between sources with really distinct
δ11B signatures, ICP-SFMS has shown to come to the same conclusions. The ease of use, rapidity and
availability of ICP-SFMS on the one hand and the observed variability in δ11B in ground- and surface
water on the other demonstrates that using ICP-SFMS as isotopic screening method would promote
the use of isotopic methodology in WFD monitoring programs.
Acknowledgements
This work is supported by the EU LIFE Project (Number LIFE06 ENV/F/000158).
References paper, Boron isotope ratio (δ11B) measurements in WFD monitoring programs:
comparison between double- focusing sector-field ICP (ICP-SFMS) and thermal ionization mass
spectrometry (TIMS) , J. Anal. At. Spectrom., 2010, 25, 964-974.
1. H.E. Gäbler and A. Bahr, Chem. Geol., 1999, 156, 323.

2. S. Eisenhut, K.G. Heumann and A. Vengosh, Fresenius J. Anal. Chem., 1996, 354, 903.
3. S. Barth, Wat. Res., 1998, 32, 685.
4. D. Widory, E. Petelet-Giraud, P. Négrel and R. Ladouche, Environ. Sci. Technol., 2005, 39, 539.
5. B. Chetelat and J. Gaillardet, Environ. Sci. Technol., 2005, 39, 2486.
6. J.D. Mather and N.C. Porteous, Applied Geochemistry, 2001, 16, 821.
7. R.L. Bassett, P.M. Buszka, G.R. Davidson, D. Chong-Diaz, Environ. Sci. Technol, 1995, 29, 2915.
8. N. Demuth and K.G. Heumann, J. Anal. At. Spectrom., 1999, 14, 1449.
9. A. Vengosh, S. Barth, K.G. Heumann and S. Eisenhut, Acta hydrochim. hydrobiol., 1999, 27, 416.
10. S. Eisenhut and K.G. Heumann, Fresenius J Anal. Chem., 1997, 359, 375.
11. B. Chetelat, J. Gaillardet and R. Freydier, Applied Geochemistry, 2009, 24, 810.
12. F. Vanhaecke, L. Balcaen, P. Taylor, Inductively Coupled plasma spectrometry and its application,
second edition, 2006, 160-215.
13. K.G. Heumann, S.M. Gallus, G. Rädlinger and J. Vogl, J. Anal. At. Spectrom., 1998, 13, 1001.
14. M. E. Weiser and J.B. Schwieters, Int. J. Mass Spectrom., 2005, 242, 97.
15. J. Aggarwal et al., J. Anal. At. Spectrom., 2009, 24, 825.
16. S.A. Kasemann, D.N. Schmidt, J. Bijma and G.L. Foster, Chem. Geol., 2009, 260, 138.
17. J.R. Encinar, J.I. Garcia Alonso, A. Sanz-Medel, S. Main, P.J. Turner, J. Anal. At. Spectrom., 2001,
16, 315.
18. S.F. Boulyga, I. Segal, T. I. Platzner, L. Halicz and J.S. Becker, Int. J. Mass Spectrom. , 2002, 218,
245.
19. D.C. Gregoire, Anal. Chem., 1987, 59, 2479.
20. D.C. Gregoire, J. Anal. At. Spectrom., 1990, 5, 623.
21. A. Al-Ammar, E. Reitznerova and R.M. Barnes, Spectrochimica acta Part B, 2000, 55, 1861.
22. P.P. Coetzee and F. Vanhaecke, Anal Bioanal. Chem, 2005, 383, 977.
23. J.K. Aggarwal, D. Sheppard, K. Mezger and E. Pernicka, Chem. Geol., 2003, 199, 331.
24. J.K. Aggarwal, K. Mezger, E. Pernicka and A. Meixner, Int. J. Mass Spectrom., 2004, 232, 259.
25. http://www.eea.europa.eu/themes/water/status-and-monitoring/water-framework-directive-
and-monitoring.
26. ISONITRATE, http://isonitrate.brgm.fr/index.htm.
27. http://www.vmm.be/water/toestand-watersystemen/waar-meten-we-het-water.
28. I.J. Allan, G.A. Mills, B. Vrana, J. Knutsson, A. Holmberg, N. Guiges, S. Laschi, AM Fouillac and R.
Greenwood, Trends in Analytcial Chemistry, 2006, 25, 704.
29. S. Barth, Geol. Rundsch., 1993, 82, 640.
30. T.B. Coplen et al., Pure Appl. Chem., 2002, 74, 1987.
31. S.C. Komor, J. Environ. Qual., 1997, 26, 1212.
218
proxy
32. D. Widory, W. Kloppmann, L. Chery, J. Bonnin, H. Rochdi and J.L. Guinamant, Journal of
Contaminant Hydrology, 2004, 72, 165.
33. D. Xue, J. Botte, B. de Baets, F. Accoe, A. Nestler, P. Taylor, O. Van Cleemput, M. Berglund and P.
Boeckx, Water Research, 2009, 43, 1159.
34. F. Vanhaecke, L. Moens and R. Dams, Anal. Chem., 1996, 68, 567.
35. J. Gaillardet J., CJ Allègre, Earth and Planetary Science Letters, 1995, 136, 665.
36. T. Ishikawa, E. Nakamura, Analytical Chemistry, 1990, 62, 2612.
37. A.J. Spivack and J.M. Edmond, Analytical Chemistry, 1986, 58, 31.
38. A.J. Spivack, M.R. Palmer and J.M. Edmond, Geochimica et Cosmochimica Acta, 1987, 51, 1939.
39. R. Gonfiantini, S. Tonarini, M. Gröning, A. Adorni-Braccesi, A.S.-Ammar, M. Astner, S. Bächler,
R.M. Barnes, R.L. Bassett, A. Cocherie, A. Deyhle, A. Dini, G. Ferrara, J. Gaillardet, J. Grimm, C.
Guerrot, U. Krähenbühl, G. Layne, D. Lemarchand, A. Meixner, D.J. Northington, M. Pennisi, E.
Reitznerová, I Rodushkin, N. Sugiura, R. Surberg, S. Tonn, M. Wiedenbeck, S. Wunderli, Y. Xiao, T.
Zack, Geostandards Newsletter: The Journal of Geostandards and Geoanalysis, 2003, 27, 41.
40. S. Tonarini, M. Pennisi, A. Adorni-Braccesi, A. Dini, G. Ferrara , R. Gonfiantini, M. Wiedenbeck and
M. Gröning, Geostandards Newsletter: The Journal of Geostandards and Geoanalysis, 2003, 27,
421.
41. D. Lemarchand, J. Gaillardet, C. Göpel and G. Manhès, Chemical Geology, 2002, 182, 232.
219
proxy
The versatility of ICP-MS instruments in investigative environmental monitoring was demonstrated in

this chapter, in which the possibility to determine isotope ratios in different environment-related
application fields is discussed. Isotope ratio measurement, especially in the case of isotope dilution
used as a primary method for the determination of trace element concentrations, has always
strongly been linked to measurement uncertainty and metrology.
In the context of this dissertation, single-collector ICP-mass spectrometers were used for isotope
ratio measurements. With the attainable precision of a quadrupole-based ICP-MS instrument, tracer
studies have been successfully performed to study the cadmium transport from the soil surface
through the unsaturated zone into the groundwater. Also a full uncertainty calculation was
performed according to the principles of the ISO Guide to the Expression of Uncertainty in
Measurement, giving an overview of the uncertainty sources in the quantitative determination of
tracer and natural cadmium by ICP-MS. The significant contribution of the signal instability, caused
by the clogging of the sampling cone by calcium salts, to the total uncertainty can be reduced
nowadays. Recently, some instrument manufacturers offer not one type, but a series of sampling
cones, more dedicated to the type of analysis, e.g., highest sensitivity versus higher matrix tolerance.
In the context of environmental forensics and to enhance the understanding of NO3- pollution in
groundwater, a single-collector sector-field ICP-MS instrument was used for the determination of
δ11B, considered as a co-migrant of nitrate, in ground and surface water. In Flanders, more than 2000
groundwater and 4000 surface monitoring stations control the nitrate content on a regular basis.
Because of these high frequency monitoring requirements, the Water Framework Directive
accidental pollution or the control of water bodies at risk. The observed variability in δ11B in ground-
and surface water (δ 11B values ranging from – 3 to + 32 ‰ vs. NBS951), demonstrates that using ICP-
SFMS as an isotopic screening method could promote the use of isotopic methodology in WFD
monitoring programs. When interpreting isotopic data and supplementary to the measurement
uncertainty, a degree of uncertainty exists due to (temporal) variability of nitrate sources, variation
in water flow regimes, mixing of nitrate sources and denitrification. Still, the introduction of a more
holistic approach in the current NO3- monitoring by using isotopic methodology, could be beneficial
and complementary in investigative monitoring of water bodies at risk.
During the past decade, multi-collector inductively coupled plasma-mass spectrometry (MC-ICP-MS)
has replaced TIMS to a large extent for accurate and high-precision determination of isotope ratios
and the technique is now deployed in various contexts [161]. The success of MC-ICP-MS is not only a
result of its higher sample throughput (compared to TIMS), but also its slightly higher matrix
tolerance and its capability to analyze elements with a high ionization energy (e.g., B (8.3 eV)).
However, MC-ICP-MS suffers from an important disadvantage, i.e., instrumental mass discrimination,
indicating that the heavier of any two isotopes is transported more efficiently than the lighter one.
Preferably, “matrix-free” solutions of the target element are deployed. For samples, the target
element is typically isolated from the concomitant matrix via chromatography. Current purification
protocols require manually feeding separation columns, a process that can be time-consuming.
However, also here automated, low-pressure ion exchange chromatography systems that can
process samples in unattended operation are commercially available (i.e., prepFAST-MC™).
220 Summary and conclusions
CHAPTER 5 SUMMARY AND CONCLUSIONS
The analytical methods described in this dissertation endorse that ICP-MS units are nowadays
reliable, robust and high-throughput instrumentation, ideally suited for the determination,
speciation and isotopic analysis of metals and oxy-anions in an environmental context.
The versatility of ICP-MS units as (i) environmental regulatory monitoring tool for the determination
of major and trace elements, (ii) elemental-specific detector in the context of hyphenation with
different separation techniques and (iii) mass spectrometer for the determination of isotope ratios
are summarised below.
→ ICP-MS applications in environmental regulatory monitoring
A clean, healthy and diverse environment is a prerequisite for prosperity and well-being for everyone
and this can only be achieved through a common framework. This common framework is needed in
order to assure that environmental pollution, no matter where it comes from, is assessed in the
same way across national or even regional borders. At a European level, major legal acts include the
Water Framework Directive, the Waste Framework Directive, the Air Quality Directive and the
upcoming Soil Framework Directive and all of these include monitoring programs to report on the
state of our environment. However, the best legal provisions will not be successful if they are not
implemented and this is the prime responsibility of the member states.
Since 1980, the management of the environment became a task for the regions (Flemish, Walloon
and Brussels Capital Region) in Belgium. The last 30 years, different Flemish regulations for the
protection of our environment have been implemented. The last years, this legislation was amended
many times, mostly under the influence of new European environmental legislation. Currently, a shift
from “environmental protection” to “environmental sustainability” is being introduced.
Implementing a more sustainable approach to the consumption and production of chemicals would
not only benefit Europe's environment but also reduce the detrimental effects arising in other parts
of the world (e.g., caused by electronic waste).
As 1980 is also the official birth year of ICP-MS, the link between the emergence of environmental
legislation dedicated to heavy metals and the emergence of ICP-MS in environmental analysis is
discussed in the first chapter of this dissertation.
Along with the implementation of regulatory monitoring requirements in the different European
member states, the use of analytical methods developed by the European Committee for
Standardization (CEN) was recommended. With common analytical European standards (EN) in all
these countries and every conflicting national standard withdrawn, comparison between analytical
results should become more feasible.
A beneficial side-effect, taking into account that since 1980 the management of the environment in
Belgium became a task for the regions, is that the European environmental regulatory monitoring
requirements not only triggered a harmonisation in the different European member states but also
led to a “rapprochement” in the internal Belgian situation. With a witticism, one could say that the
Belgian regions were forced by the different EU directives to unify the decision-making process in
Summary and conclusions 221
environmental issues. Still, structural consultation between the regions in this respect would be
desirable.
The introduction of ICP-MS under impulse of standards developed by the European Committee for
Standardization (CEN) is in environmental regulatory monitoring of elements undoubtedly the most
important advance of the last decade. Because of the availability of ICP-MS instruments nowadays, it
can be argued that the main hindrance to the implementation of the European monitoring
requirements is not the technical feasibility of analysis at these concentration levels, but rather
potential contamination during sampling, sample storage, sample handling and analysis. Two
additional comments with respect to a harmonised implementation of the European monitoring
requirements may be formulated at this point. First, as for most monitoring techniques, results
obtained are operationally-defined. The protocol, e.g., acid digestion, 0.45 µm filtration, ... defines
the answer, and while this may facilitate comparability of results acquired across Europe, data
generated may have little relevance to bioavailable metal concentration. Second, more
harmonization in communication and reporting of our chemical environmental status between
member states is needed to compare regulatory monitoring data on a European scale, e.g., choice of
sampling locations (rural versus industrial) and type of sample (biota versus water column).
As a case study and for the first time, ranges of background concentration of heavy metals were
summarised for all environmental compartments in Flanders based on European Standards on the
one hand and on a selection of (non-polluted) background locations on the other. The focus in
chapter 2 of the dissertation is on the analysis (of acid digests) of environmental samples by ICP-MS,
used for multi-elemental determination of elements. The range of background concentrations,
derived from existing Flemish regulatory monitoring databases, illustrate a match made in heaven
with the sensitivity of ICP-MS instruments.
The range of background concentrations of elements, derived and compiled in this PhD dissertation,
can be considered as an indicator of the current “natural” chemical status of the Flemish
environment and can be used to assess future anthropogenic influences within an EU regulatory
monitoring context. However, as heavy metals do not disappear, but rather disperse from one
environmental compartment to another, the chemical nature and quantity of the relevant element
species in a matrix, its physical and chemical association, rather than the (pseudo) total element
concentration is governing the mobility, bioavailability and, finally, the ecotoxicological or
toxicological impact of that element. Therefore, only knowledge on speciation provides sufficient
information for environmental risk assessment and this represents the subject of the analytical
research papers described in chapter 3.
→ Determination of oxy-anions in water and solid samples
In this dissertation, speciation applications have been designed to determine chromate in solid
materials, selenate and selenite in waste water and bromate in drinking water. All of the applications
were driven by European legislation, i.e., the restriction of hexavalent chromium in electrical and
electronic equipment brought onto the European market, removal of oxy-anions from wastewater to
ensure compliance with European environmental quality parameters and the control of the
maximum level of carcinogenic disinfection by-products in European drinking water monitoring
programs.
The efforts to set up an HPLC/ICP-MS system in the year 2000 for the speciation of selenium were
mainly devoted to the automation and a proper communication interface between the sample
injection system, the HPLC pump and the ICP-MS instrument. Ten years later and with the availability
of a low pressure LC system, speciation analysis can be implemented easily on any ICP-MS
instrument and this was demonstrated for the determination of bromate in drinking water.
Nowadays, LC/ICP-MS can truly be considered widespread sensitive and versatile speciation
instrumentation.
Speciation is also increasingly being involved in multi-disciplinary approaches using so-called

orthogonal speciation schemes (e.g., ESI-MS and HPLC/ICP-MS) for (complex) species identification
and/or quantification (e.g., metallomics, pharmaceutical substances). Recent advances in this field,
includes front-end modifications (e.g., micro nebulizers) to adapt the ICP-MS instrument to the low
flow rates (~ µl min-1) and/or high organic content solutions delivered by the LC system and the use
of reaction cells to chemically resolve polyatomic interferences from the element of interest (e.g.,
SeO+).
Quality control and quality assurance programs carried out along these speciation research projects
are of utmost importance, e.g., control of the level of oxidation of Cr(III). Correction for the species
interconversion observed remains ambiguous when the chemical form of the species present in the
sample (e.g., FeCr2O4) is different from the added spike (e.g., Cr(NO3)3.9H2O). On the other hand,
knowledge of the method-induced interconversion is particularly relevant for the determination of
Cr(VI) in ambient particulate matter (PM10) as the level of observed Cr(III) oxidation is within the
range of the level of monitored Cr(VI) and could lead to false positives exceeding of the air quality
guideline value. It is conceivable that XANES spectroscopy can serve as a benchmark for
improvements or modifications to wet chemistry extraction methods.
While the relevance of elemental speciation in regulatory monitoring is not at stake, why is
elemental speciation not done routinely with LC/ICP-MS in analytical laboratories? First, European
Directives and legislation about elemental speciation are to a great extent still missing with the
exception of those for a few species (e.g., Cr(VI), organotin compounds, methylmercury) and
therefore, routine laboratories lack the incentive to invest in the technologies. Second, for the
determination of some species, e.g., bromate, alternative cheap methods (as compared to LC/ICP-
MS) are available for monitoring at the concentration level of the regulatory limit value. Third, the
assumption of monitoring in a regulatory context is that only one or a few constituents of a sample -
usually the analytes of interest - play a role in the measurement that is obtained. When applying
standardized procedures in routine, users tend to have as little concern as possible about what other
species in the sample might yield a false response or whether the method is applicable to all samples
in the measurement run. However, and especially for solid samples, elemental speciation analysis is
not considered as a course of actions to always be followed and interpreted in the same way. These
restrictions have often led to the current pragmatic approach in elemental speciation in regulatory
context, i.e., only in case the total concentration is above the limit value, an elemental speciation
analysis is performed.
→ Tracer experiments with stable isotopes in environmental studies and use of natural
isotopic variation as a proxy
The versatility of ICP-MS instruments in investigative environmental monitoring is further

demonstrated in chapter 4, in which the possibility to determine isotope ratios in different
environment-related application fields is discussed. Isotope ratio measurement, especially in the case
of isotope dilution used as a primary method for the determination of trace element concentrations,
has always strongly been linked to measurement uncertainty and metrology. Nowadays, the level of
required precision and accuracy will guide to the technique to be used (quadrupole based ICP-MS,
ICP-SFMS or MC-ICP-MS) or conversely the available instrumentation will tell which environmental
applications are within reach. During the past decade, MC-ICP-MS has replaced TIMS to a large
extent for accurate and high-precision determination of isotope ratios and the technique is now
deployed in various contexts (e.g., geochronology, study of a large variety of geochemical

phenomena and provenance determination).
In the context of this dissertation, single-collector ICP-mass spectrometers were used for isotope
ratio measurements. With the attainable precision of a quadrupole-based ICP-MS instrument, tracer
studies have been successfully performed to study the cadmium transport from the soil surface
through the unsaturated zone into the groundwater. Also a full uncertainty calculation was
performed according to the principles of the ISO Guide to the Expression of Uncertainty in
Measurement, giving an overview of the uncertainty sources in the quantitative determination of
tracer and natural cadmium by ICP-MS.
While metrology has always been an asset to analytical chemistry, end-users of analytical data still
have ambiguous relations to the concept of measurement uncertainty and how to deal with it. Often,
so-called significantly higher levels have been reported in (bio-)monitoring studies, without
considering the issue of uncertainty. Reporting estimates of measurement uncertainty is the prime
responsibility of the analyst and should trigger the end-user to use and interpret this information
properly. Also, the translation and interpretation of probabilistic data (analytical data) into boolean
data (yes/no) is still a main challenge in an environmental regulatory context.
In the context of environmental forensics and to enhance the understanding of NO3- pollution in
groundwater, a single-collector sector-field ICP-MS instrument was used for the determination of
δ11B, considered as a co-migrant of nitrate, in ground and surface water. In Flanders, more than 2000
groundwater and 4000 surface monitoring stations control the nitrate content on a regular basis.
Because of these high frequency monitoring requirements, the Water Framework Directive
accidental pollution or the control of water bodies at risk. The observed variability in δ11B in ground-
and surface water (δ 11B values ranging from – 3 to + 32 ‰ vs. NBS951), demonstrates that using ICP-
SFMS as an isotopic screening method could promote the use of isotopic methodology in WFD
monitoring programs. When interpreting isotopic data and supplementary to the measurement
uncertainty, a degree of uncertainty exists due to (temporal) variability of nitrate sources, variation
in water flow regimes, mixing of nitrate sources and denitrification. Still, the introduction of a more
holistic approach in the current NO3- monitoring by using isotopic methodology, could be beneficial
and complementary in investigative monitoring of water bodies at risk.
→ Recommendations for future research
Since its introduction in 1980, ICP-MS evolved from a lab-built instrument to a family of commercially
available analytical techniques, ranging from single collector quadrupole mass filter units (ICP-QMS)
to single-collector and multi-collector sector-field based ICP-MS instruments with high mass
resolution capabilities. The introduction of ICP-MS not only triggered new research opportunities in
the field of environmental, geochemical and life sciences, but also initiated further improvements in
the analytical workflow and front-end modifications. The following recommendations are delineated
for future research and activities linked to ICP-MS in an environmental context:
• The panoramic analysis capabilities of ICP-MS are not yet fully exploited with respect to its
linear dynamic range and multi-element capabilities. The determination of heavy metals in
environmental monitoring has traditionally focused on a relative small number of elements
(e.g., As, Cd, Pb, Hg), but it is to be expected that future regulatory monitoring will include
more elements (holistic approach) and that the panoramic analysis capabilities of ICP-MS will
be further exploited. This is also being stimulated in sustainable material management, for
which knowledge on the presence of (economically) critical elements (e.g., rare earth and
platina-group elements) in (contaminated) solid materials are of potential interest.
• New or innovative approaches in the (on-line)-automation of the labor-intensive steps of

sample digestion, preparation of calibration standards, dilution of samples and the addition
of internal standards. Further improvements in the analytical workflow (e.g., contamination),
front-end modifications (e.g., on-line sample preparation automation) and sample
introduction systems are needed. Automation is not only considered as a time-saving
operation (cost price), but also as a time-gaining opportunity for the analyst to process
(expert judgement) the raw data (quality).
• ICP-MS is increasingly being involved in multi-disciplinary approaches for (complex) species
identification and/or quantification. Further research in the field of speciation includes front-
end modifications (e.g., micro and capillary nebulizers) to adapt the ICP-MS instrument to
the low flow rates and/or high organic content solutions delivered by LC systems. With
respect to possible species interconversion, multi-disciplinary approaches, e.g., using XANES
are needed to evaluate the performance of wet chemistry extraction methods for the
determination of species in solid samples. To separate, detect and quantitate engineered
nanoparticles (ENPs) – emerging contaminants – in environmental matrices, field-flow
fractionation, a size-separation technique, shows a great deal of promise when coupled to
ICP-MS.
• Further instrumental innovation, with focus on reliable and robust ICP-MS instrumentation,
in combination with blue economy principles by, e.g., reducing/eliminating combustible gas
usage, power consumption (create solutions that are both environmentally beneficial and
which have financial benefits). Additionally, innovation in ultra-fast electronics can provide
advantages in the collection of more data per unit of time (e.g., single-particle inductively
coupled plasma–mass spectrometry).
• New developments in (cheap) clean room facilities and workflows (including sampling),
combined with higher purity demands for reagents and recipients are needed to fully exploit
the instrumental detection limits attainable by ICP-MS.
• Optimising the balance between analytical management and analytical quality in a high-
throughput laboratory. An inverse relationship is being created in high-throughput
environmental laboratories between standardization and accreditation on the one hand and
understanding and knowledge of the analytical measurement within the context of the
sample on the other. Where is the wisdom we have lost in knowledge? Where is the
knowledge we have lost in information? is changing to Where is the time to acknowledge the
information ?
• Communication and knowledge exchange among laboratories/scientists and policy-makers
on the translation of the analytical regulatory needs and the scientific
capabilities/boundaries. Especially in the case of passive samplers to become a standard
inventory in the toolbox for regulatory monitoring, the remaining obstacles to a more
widespread adoption lay in communicating the knowledge produced by the scientific
community to the intended audience of policy makers, managers and operational staff, who
administrate and execute regulatory monitoring programs.
The (inorganic) analytical team at Vito anno 2000.
The analytical team at Vito anno 2013.
The biennial meetings at the European Winter Conference with Martin Resano and Frank Vanhaecke
(Zaragoza, Spain, 2011).
226 References
References
For the scientific papers, the reference lists are provided separately at the end of each paper:
• p. 84, Characterization of inorganic selenium species by ion chromatography with ICP-MS detection in
microbial-treated industrial waste water.
• p. 106, Determination of hexavalent chromium by species-specific isotope dilution mass spectrometry
and ion chromatography-1,5-diphenylcarbazide spectrophotometry.
• p. 127, Validation of a European Standard for the determination of hexavalent chromium in solid waste
material.
• p. 148, Determination of hexavalent chromium in ambient air: a story of method-induced Cr(III)
oxidation?
• p. 165, Determination of bromate in drinking waters using low pressure liquid chromatography / ICP-
MS.
• p. 193, Full uncertainty calculation on quantitative determination of tracer and natural cadmium in soil
column effluents with ICP-MS.
• p. 217, Boron isotope ratio measurements in WFD monitoring programs: comparison between double-
focusing sector-field ICP (ICP-SFMS) and thermal ionization mass spectrometry (TIMS).
1. R.S. Houk, V.A. Fassel, G.D. Flesch, H.J. Svec, AL. Gray and C.E. Taylor, Inductively Coupled Argon
Plasma as an Ion Source for Mass Spectrometric Determination of Trace Elements, Anal Chem.,
1980, 52, 2283-2289.
2. A. R. Date and A. L. Gray, Plasma Source Mass Spectrometry Using an Inductively Coupled Plasma
and a High Resolution Quadrupole Mass Filter, Analyst, 1981, 106, 1255-1267.
3. ALS Environmental, http://www.alsglobal.com/en/Our-Services/Life-Sciences/Environmental, last
accessed on 20/01/2013.
4. The Environment, Nature and Energy Department, http://www.lne.be/themas/vergunningen/
regelgeving/historiek/historiek, last accessed on 20/01/2013.
5. Environment, Nature and Energy Department, http://www.lne.be/themas/vergunningen/vlarem-
en-de-milieuvergunning, last accessed on 20/01/2013.
6. The Public Waste Agency of Flanders, http://www.30jaarovam.be/index.php/tijdlijn, last accessed
on 20/01/2013.
7. The Flemish land agency http://www.vlm.be/algemeen/overdevlm/historiek/Pages/default.aspx,
last accessed on 20/01/2013.
8. E. Buyst, K. Lowyck, A. Soete, Kroniek van de Vlaamse Milieumaatschappij, 1991-2011,
http://www.vmm.be/pub/kroniek-van-de-vlaamse-milieumaatschappij-1991-2011#view=Fit, last
9. VMM, Watermeter 2011, Drinking water production and supply in figures, D/2012/6871/007, 63p.
10. History of groundwater EU legislative development, http://ec.europa.eu/environment/water/
water-framework/groundwater.html, last accessed on 20/01/2013.
11. Council Directive 98/83/EC of 3 November 1998 on the quality of water intended for human
consumption, http://ec.europa.eu/environment/water/water-drink/index_en.html, last accessed
on 20/01/2013.
12. K. Broos, I. Joris, J. Vos, M. Quaghebeur, 2011, VITO, Gemeenschappelijk normenkader secundaire
grondstoffen en uitgegraven bodem: haalbaarheidsstudie en normvoorstellen - synthesenota
klankbordgroep, Studie uitgevoerd in opdracht van de OVAM, 2011/RMA/R/0034.
13. European Commission, Air Quality - Existing Legislation, http://ec.europa.eu/environment/air/
quality/legislation/existing_leg.htm, last accessed on 20/01/2013.
14. M. Carere, V. Dulio, G. Hanke, S. Polesello, Guidance for sediment and biota monitoring under the
Common Implementation Strategy for the Water Framework Directive, TrAC Trend. Anal. Chem.,
2012, 36, 15-24.
References 227
15. A. Montaser, Inductively Coupled Plasma Mass Spectrometry, 1998, ISBN 978-0-471-18620-5 - John
Wiley & Sons, 964 p.
16. F. Vanhaecke and P. Degryse, Isotopic Analysis, Fundamentals and Applications Using ICP-MS,
Edition June 2012, ISBN 978-3-527-32896-3 - Wiley-VCH, Weinheim.
17. N. Jakubowski, T. Prohaska, L. Rottmann and F. Vanhaecke, Inductively coupled plasma- and glow
discharge plasma-sector field mass spectrometry, part I, Tutorial: Fundamentals and
instrumentation, J. Anal. At. Spectrom., 2011, 26, 693-726.
18. S. M. McIntyre, J. Wisnewski Ferguson, T.M. Witte, R.S. Houk, Measurement of gas kinetic
temperatures for polyatomic ions in inductively coupled plasma-mass spectrometry: Validation and
refinements, Spectrochim. Acta B, 2011, 66, 248–254.
19. N. Jakubowski, Analytical plasma ion sources for elemental mass spectrometry: where are we
coming from – where are we going to ? J. Anal. At. Spectrom., 2008, 23, 673-684.
20. Agilent Technologies, Inc. 2011, 5990-8572EN, www.agilent.com/chem/RunsOnAir (last accessed
on 20/01/2013), Agilent 4100 microwave plasma atomic emission spectrometer.
21. H. Niu, R.S. Houk, Fundamental aspects of ion extraction in inductively coupled plasma - mass
spectrometry, Spectrochim. Acta B, 1996, 51, 779-815.
22. L. Balcaen, G. Woods, M. Resano and F. Vanhaecke, Accurate determination of S in organic matrices
using isotope dilution ICP-MS/MS, J. Anal. At. Spectrom., 2013, 28, 33-39.
23. K. Tirez, W. Brusten, A direct ultratrace isotopic analysis of uranium in environmental and biological
samples by SFICP-MS for environmental forensics, poster presented at the European Winter
Conference on plasma spectrochemistry, 2011, Zaragosa, Spain.
24. K. Tirez, W. Brusten, S. Van Roy, N. De Brucker and L. Diels, Characterization of inorganic selenium
species by ion chromatography with ICP-MS detection in microbial-treated industrial waste water,
J. Anal. At. Spectrom., 2000, 15, 1087-1092.
25. D. Potter, A commercial perspective on the growth and development of the quadrupole ICP-MS
market, J. Anal. At. Spectrom., 2008, 23, 690–693.
26. O. Butler, H. Evans, A. Fisher, S. Hill, C. Harrington, A. Taylor, M. West and A. Ellis, Atomic
Spectrometry Updates: A 25-year retrospective, J. Anal. At. Spectrom., 2010, 25, 1546–1566.
27. M. Resano and E. García-Ruiz, High-resolution continuum source graphite furnace atomic
absorption spectrometry: Is it as good as it sounds? A critical review, Anal. Bioanal. Chem., 2011,
399, 323–330.
28. http://www.pnnl.gov/science/journals/pdf/jaas_animated.pdf, last accessed on 20/01/2013.
29. D. W. Koppenaal, G. M. Hieftje, Metallomics The Future of Atomic Spectroscopy ?, J. Anal. At.
Spectrom., 2007, 22, 111.
30. O.T. Butler, W.R. L. Cairns, J.M. Cook and C.M. Davidson, Atomic spectrometry update.
Environmental analysis, J. Anal. At. Spectrom., 2012, 27, 187–221.
31. E. McCurdy, N. Yamada and N. Sugiyama, Agilent’s New 8800 Triple Quadrupole ICP-MS: Hardware
and Technology, Agilent ICP-MS journal, March 2012, issue 49, 2.
32. D.L. Rocha, A.D. Batista, F.R.P. Rocha, G.L. Donati, J.A. Nobrega, Greening sample preparation in
inorganic analysis, TrAC Trend. Anal. Chem., 2013, 45, 79-92.
33. C. Bendicho, I. De La Calle, F. Pena, M. Costas, N. Cabaleiro, I. Lavilla, Ultrasound-assisted
pretreatment of solid samples in the context of green analytical chemistry, TrAC Trend. Anal.
Chem., 2012, 31, 50-60.
34. C. Bendicho, I. Lavilla, F. Pena-Pereira and V. Romero, Green Chemistry in analytical atomic
spectrometry : a review, J. Anal. At. Spectrom., 2012, 27, 1831-1857.
35. K. Tirez, G. Vanermen, S. Bleyen, R. Weltens en S. Kreps, 2006, Actualisatie van prestatiekenmerken
en referentiemeetmethoden voor de controle op de naleving van emissiegrenswaarden (Vlarem II)
en voor de controle van de kwaliteit van voor menselijke consumptie bestemd water,
2006/MIM/R/093bis, studie uitgevoerd in opdracht van LNE afdeling Milieu-inspectie.
36. C. Vanhoof en K. Tirez, 2005, Afvalstoffen op stortplaatsen: Bepalen van analysemethoden voor
nieuwe parameters, Studie uitgevoerd in opdracht van de OVAM, 2005/MIM/R/060.
228 References
37. Personal communication Holger Scharf, Federal Institute for Materials Research and Testing (BAM),
Germany.
38. C. Vanhoof, F. Beutels, W. Brusten, B. Noten, V. Corthouts en K. Tirez, 2006, Best Beschikbare
Technieken voor de bepaling van zware metalen in water in het kader van de erkenningen,
2006/MIM/R/026, studie uitgevoerd in opdracht van LNE afdeling Milieu-vergunningen,
http://www.emis.vito.be/onderzoeksrapporten-2, last accessed on 20/01/2013.
39. J.T. Shelley adn G.M. Hieftje, Ambient mass spectrometry: Approaching the chemical analysis of
things as they are, J. Anal. At. Spectrom., 2011, 26, 2153-2159.
40. MIRA, het Milieurapport Vlaanderen, http://www.milieurapport.be/en/home/, last accessed on
20/01/2013.
41. P.L. Brezonik, W.A. Arnold, Water Chemistry: Fifty Years of Change and Progress, Environ. Sci.
Technol., 2012, 46, 5650-5657.
42. R. Hübner, K.B. Astin and J.H. Herbert, “Heavy metal” – time to move on from semantics to
pragmatics ?, J. Environ. Monit., 2010, 12, 1511-1514.
43. European Communities, 2011, Common Implementation Strategy for the Water Framework
Directive (2000/60/EC), Technical Report - 2011 – 055, Guidance Document No. 27, Technical
Guidance For Deriving Environmental Quality Standards, ISBN : 978-92-79-16228-2.
44. W. Shotyk and M. Krachler, Determination of trace element concentrations in natural fresh waters:
How low is “low” and how low do we need to go ?, J. Environ. Monit., 2009, 11, 1747-1753.
45. G. Hanke, S. Polesello, G. Mariani, S. Comero, J. Wollgast, R. Loos, J. Castro-Jimenez, L. Patrolecco,
S. Valsecchi, M. Rusconi, N. Ademollo and G. Bidoglio, Chemical-monitoring on-site exercises to
harmonize analytical methods for priority substances in the European Union, TrAC Trend. Anal.
Chem., 2012, 36, 25-35.
46. European Commission, Environment, Soil : http://ec.europa.eu/environment/soil/index_en.htm,
last accessed on 20/01/2013.
47. F.M.G.Tack, M.G.Verloo, L. Vanmechelen, E. Van Ranst, Baseline concentration levels of trace
elements as a function of clay and organic carbon contents in soils in Flanders (Belgium), Sci. Total
Environ., 1997, 201, 113-123.
48. F.J. Zhao, S.P. McGrath, G. Merrington, Estimates of ambient background concentrations of trace
metals in soils for risk assessment, Environ. Pollut., 2007, 148, 221-229.
49. C. Micó, M. Peris, L. Recatalá, J. Sánchez, Baseline values for heavy metals in agricultural soils in an
European Mediterranean region, Sci. Total Environ., 2007, 378, 13–17.
50. ISO 19258: 2005 International Organisation for Standardisation (ISO), 2005. Soil quality: Guidance
on the determination of background values.
51. F.M.G. Tack, T. Vanhaesebroeck, M.G. Verloo, K. Van Rompaeyb, E. Van Ranst, Mercury baseline
levels in Flemish soils (Belgium), Environ. Pollut., 2005, 134, 173-179.
52. C. Vanhoof, K.Tirez, Extractie/destructiemethoden voor de bepaling van zware metalen in vaste
monsters, Studie uitgevoerd in opdracht van de OVAM, 2003/MIM/R/160, 45p.
53. P. Seuntjens, J. Bierkens, J. Patyn, K. Tirez, D. Wilczek, R. Smolders, D. Lagrou, 2006, Herziening
achtergrondwaarden zware metalen in bodem, studie uitgevoerd in opdracht van OVAM.
54. C. Vanhoof en K.Tirez, Extractie/destructiemethoden voor de bepaling van zware metalen in vaste
monsters, 2002, Studie uitgevoerd in opdracht van de OVAM, 2002/MIM/R/116.
55. E. Engström, A. Stenberg, D.C. Baxter, D. Malinovsky,I. Mäkinen, S. Pönnid and I. Rodushkin, Effects
of sample preparation and calibration strategy on accuracy and precision in the multi-elemental
analysis of soil by sector-field ICP-MS, J. Anal. At. Spectrom,. 2004, 19, 858-866.
56. F. Vanhaecke, H. Vanhoe, L. Moens, R. Dams, Quadrupole ICP-MS mass spectrometry for the
determination of the aqua regia soluble content of trace metals in soil and sludge candidate
reference materials, Bull. Soc. Chim. Belg., 1995, 104 (11), 653-661.
57. C. Vanhoof, Groep AN en K. Tirez, VITO – Milieumetingen, Project Horizontal- Anorganische
resultaten, Studie uitgevoerd in opdracht van de OVAM, 2007/MIM/R/182, 111p.
58. The Geochemical atlas of Europe, http://weppi.gtk.fi/publ/foregsatlas/maps_table.php, last
References 229
59. Geochemical Mapping of Agricultural Soils of Europe, GEMAS, http://www.bgs.ac.uk/gbase/

GEMAS.html, last accessed on 29/08/2013.
60. European Environment Agency, Technical report No 12/2011, Air quality in Europe - 2011 report,
http://www.eea.europa.eu/publications/air-quality-in-europe-2011, last accessed on 20/01/2013.
61. Data obtained by courtesy of Elke Adriaenssens, Vlaamse Milieumaatschappij - afdeling Afdeling
Lucht, Milieu en Communicatie - Dienst lucht.
62. Data obtained by courtesy of Bo Van den Bril, Vlaamse Milieumaatschappij - afdeling Afdeling
Lucht, Milieu en Communicatie - Dienst lucht.
63. VMM, Methode voor de monsterneming en analyse van zware metalen in zwevend en neervallend
stof, http://www.vmm.be/lucht/meetnetten/bemonstering-en-analyse/monsterneming-en-
analyse-zware-metalen, last accessed on 20/01/2013.
64. European Commission, Environment-Air http://ec.europa.eu/environment/air/index_en.htm, last
65. VMM, Grens- en richtwaarden luchtkwaliteit, http://www.vmm.be/lucht/beleid-en-
instrumenten/grens-en-richtwaarden-luchtkwaliteit, last accessed on 20/01/2013.
66. K. Tirez, F. Beutels, B. Noten, W. Brusten, C. Vanhoof, Chemische karakterisering van zwevend en
depositiestof in binnen en buiten omgeving, 2008/MIM/R/007, VITO, 102 p.
67. K. Tirez, Chemische karakterisering van zwevend en depositie stof in binnen en buiten omgeving,
presented on studiedag elementanalyse, 25/10/2007, KVCV Sectie Analytische Chemie,
Grobbendonk, Belgium.
68. K.Tirez, G. Silversmit, L. Vincze, K. Servaes, C. Vanhoof, N. Bleux and P. Berghmans, Speciation and
fractionation of nickel in airborne particulate matter : comparison between selective leaching and
XAS spectroscopy, J. Anal. At. Spectrom., 2011, 26, 517 -527.
69. EN 14385:2004 Stationary source emissions - Determination of the total emission of As, Cd, Cr, Co,
Cu, Mn, Ni, Pb, Sb, TI and V.
70. Chemkar PM10 ‘hotspots’, Chemische karakterisatie van fijn stof in Vlaanderen, 2006-2007,
Afdeling Lucht, Milieu en Communicatie, VMM, D/2009/6871/015, http://www.vmm.be/
publicaties/2009/CK_PM10_TW.pdf/view?searchterm=chemkar, last accessed on 20/01/2013.
71. Chemkar PM10 ‘hotspots’, Chemische karakterisatie van fijn stof in Vlaanderen, 2008-2009,
Afdeling Lucht, Milieu en Communicatie, VMM, D/2010/6871/041, http://www.vmm.be/pub/
rapport-chemkar-pm10/view?searchterm=chemkar, last accessed on 20/01/2013.
72. Chemkar PM10 ‘hotspots’, Chemische karakterisatie van fijn stof in Vlaanderen, 2010, Afdeling
Lucht, Milieu en Communicatie, VMM, D/2011/6871/032, http://www.vmm.be/pub/chemkar-
pm10-chemische-karakterisering-van-fijn-stof-in-vlaanderen-2010/view?searchterm=chemkar, last
73. Personal communication Jordy Vercauteren, VMM, Afdeling Lucht Milieu en Communicatie, dienst
lucht.
74. C. Vanhoof, V. Corthouts and N. De Brucker, An improved aerosol generation system for the
preparation of XRF calibration filters, Adv. X-Ray Anal., 2000, 43, 449-455.
75. C. Vanhoof, H. Chen, P. Berghmans, V. Corthouts, N. De Brucker and K. Tirez, A risk assessment
study of heavy metals in ambient air by WD-XRF spectrometry using aerosol-generated filter
standards, X-Ray Spectrom., 2003, 32, 129-138.
76. M. Gerboles, D. Buzica, R.J.C. Brown, R.E. Yardley, A. Hanus-Illnar, M. Salfinger, B. Vallant, E.
Adriaenssens, N. Claeys, E. Roekens, K. Sega, J. Jurasović, S. Rychlik, E. Rabinak, G. Tanet, R.
Passarella, V. Pedroni, V. Karlsson, L. Alleman, U. Pfeffer, D. Gladtke, A. Olschewski, B. O’Leary, M.
O’Dwyer, D. Pockeviciute, J. Biel-Ćwikowska, J. Turšič, Interlaboratory comparison exercise for the
determination of As, Cd, Ni and Pb in PM10 in Europe, Atmos. Environ., 2011, 45(20), 3488-3499.
77. National Environmental Methods Index, https://www.nemi.gov/, last accessed on 20/01/2013.
78. NBN T94-101:1976 Meting van de atmosferische stofuitval door middel van neerslagcollectoren.
79. I. Gelaude, Bepaling van totaal kwik en kwikspecies in milieu- en biologische monsters, Proefschrift
voorgelegd tot het behalen van de graad van Doctor in de Wetenschappen: Scheikunde,
Universiteit Gent, Faculteit Wetenschappen, Laboratorium voor Analytische Scheikunde, Instituut
230 References
voor Nucleaire Wetenschappen, 2002-2003, http://lib.ugent.be/fulltxt/RUG01/000/792/930/

RUG01-000792930_2010_0001_AC.pdf, last accessed on 20/01/2013.
80. http://ec.europa.eu/environment/air/pdf/pp_mercury_xsum.pdf, last accessed on 20/01/2013.
81. EN 15853:2010 Ambient air quality - Standard method for the determination of mercury
deposition.
82. K. Tirez, W. Swaans, K. Duyssens, Kwik in natte depositie voor CAMP-OSPAR meetstation te
Koksijde, Studie uitgevoerd in opdracht van VMM, 2012/MANT/R/006, 2012, 54 p.
83. EN 15852:2010 Ambient air quality - Standard method for the determination of total gaseous
mercury.
84. Personal communication Koen Toté, VMM, Afdeling Lucht.
85. European Environment Agency, Hazardous substances in Europe's fresh and marine waters - An
overview, 2011, Technical report No 8/2011, ISBN 978-92-9213-214-9, http://www.eea.europa.eu/
publications/hazardous-substances-in-europes-fresh, last accessed on 20/01/2013.
86. Databank Ondergrond Vlaanderen, https://dov.vlaanderen.be, last accessed on 02/02/2013.
87. VMM (2013), zware metalen in het groundwater in Vlaanderen, D/2013/6871/013, Vlaamse
Milieumaatschappij, Aalst, 96 p.
88. Personal communication Ywan De Jonghe and Ralf Eppinger, VMM, Dienst Grondwaterbeheer.
89. V. De Saedeleer, V. Cappuyns, W. De Cooman, R. Swennen, Influence of major elements on heavy
metal composition of river sediments, Geol. Belg., 2010, 13, 257 -268.
90. B. Bölviken , J. Bogen , A. Demetriades , W. De Vos , J. Ebbing, R. Hindel, M. Langedal, J. Locutura, P.
O’ Connor, RT Ottesen, E. Pulkkinen, R. Salminen, O. Schermann, R. Swennen, J. Van der Sluys & T.
Volden, Regional geochemical mapping of Western Europe towards the year 2000, J. Geochem.
Explor., 1996, 56, 141-166.
91. W. De Cooman and L. Detemmerman, Waterbodemonderzoek in Vlaanderen : Hoe gaan we te werk
en wat zijn de eerste meetresultaten? (Sediment Quality Assessment in Flanders: how do we go
about it and what are measuring results?), Water, 2003, 6, 1-7.
92. Personal communication Ward De Cooman, VMM, Teamverantwoordelijke Waterbodemmeetnet.
93. Data obtained by courtesy of Johan Annys, Vlaamse Milieumaatschappij - afdeling rapportering
water, Teamverantwoordelijke anorganische analysen.
94. VMM, Het waterbodemmeetnet VMM bestaat uit een fysisch-chemische, ecotoxicologische en een
biologische beoordeling, http://www.vmm.be/water/kwaliteit-oppervlaktewater/meetnetten-en-
kwaliteitsborging/meetnetten/?searchterm=meetnet, last accessed on 20/01/2013.
95. M. Ricci, I. Kourtchev, H Emons, Chemical water monitoring under the Water Framework Directive
with Certified Reference Materials, TrAC Trend. Anal. Chem., 2012, 36, 47-57.
96. L. Lucentini, P. Pettine, E. Stottmeister, E. Menichini, Chemical analysis of the quality of water for
human consumption: Proposal for the revision of the performance requirements in the Drinking
Water Directive 98/83/EC, TrAC Trend. Anal. Chem., 2013, 45, 37-47.
97. C. Brunori, I. Polyi, C. Pellegrino, M. Ricci, O. Bercaru, F. Ulberth, A. Sahuquillo, E. Rosenberg, Y.
Madrid, R. Morabito, The SWIFT-WFD proficiency testing campaigns in support of implementing the
EU water framework directive, TrAC Trend. Anal. Chem., 2007, 26, 993-1004.
98. B. Brunori and I. Ipolyi, Additional report N° 2 (WP 3.2): Final report on SWIFT-WFD PTs results,
Contract n° SSPI-CT-2003-502492, http://www.ecologic.eu/1142, last accessed on 20/01/2013.
99. E.D. Goldberg, M. Baker and D.L. Fox, Microfiltration in oceanographic research: I. Marine sampling
with the molecular filter, J. Mar. Res., 1952, 11 (2), 194-204.
100. O. Gustafsson and PM Gschwend, Aquatic colloids: Concepts, definitions, and current challenges,
Limnology and aceanography, 1997, 42 (3), 519-528.
101. A.J. Horowitz, K.R. Lum, J.R. Garbarino, G.E.M. Hall, C. Lemieux and C.R. Demas, Problems
associated with using filtration to define dissolved trace element concentrations in natural water
samples, Environ. Sci. Technol., 1996, 30, 954-963.
102. I.J. Allan, J.Knutsson, N. Guigues, G.A. Mills, A.M. Fouillac and Richard Greenwood, Chemcatcher
and DGT passive sampling devices for regulatory monitoring of trace metals in surface water, J.
Environ. Monit., 2008, 10, 821–829.
References 231
103. K. Tirez, passive sampling of dissolved elements in surface water, presented at International
symposium on “In-situ Environmental Monitoring and Policy, The Application of Sensors and
Passive Samplers", May 26-27, 2011, U Ghent, Belgium.
104. G. A. Mills, R. Greenwood, B. Vrana, I. J. Allan and T. Ocelka, Measurement of environmental
pollutants using passive sampling devices – a commentary on the current state of the art, J. Environ.
Monit., 2011, 13, 2979–2982.
105. C. Miege, N. Mazzella, S. Schiavone, A. Dabrin, C. Berho,J-P. Ghestem, C. Gonzalez, J.-L. Gonzalez, B.
Lalere, S. Lardy-Fontan, B. Lepot, D. Munaron, C. Tixier, A. Togola, M. Coquery, An in situ
intercomparison exercise on passive samplers for monitoring metals, polycyclic aromatic
hydrocarbons and pesticides in surface waters, 2012, TrAC Trend. Anal. Chem, 36, 128-143.
106. VMM, Metalen in het oppervlaktewater en waterbodem, www.vmm.be/pub/jaarrapport-water-
2010/at_download/file, last accessed on 20/01/2013.
107. Personal communications Martin Verdievel and Ilse Theuns, VMM, Afdeling Rapportering Water
and Henk Maeckelberghe, VMM, Staf Algemeen-Directeur, Dienst Stroomgebiedbeheer.
108. M. Coetsiers, P. Blaser, K. Martens, K. Walraevens, Natural background levels and treshold values
for groundwater in fluvial Pleistocene and Tertiary marine aquifers in Flanders, Belgium, Environ.
Geol., 2009, 57, 1155-1168.
109. P. Jitaru, K. Tirez and N. De Brucker, Panoramic analysis for monitoring trace metals in natural
waters by ICP-MS, Atomic Spectroscopy, 2003, 24 (1), 1–10.
110. Population Division of the Department of Economic and Social Affairs of the United Nations, World
Urbanization Prospects, the 2011 Revision, http://esa.un.org/unup/index.html, last accessed on
31/08/2013.
111. B. Michalke, Perspective on element speciation, J. Environ. Monit., 2009, 11, 1754–1756.
112. R. Cornelis, J.A. Caruso, H. Crews, K.G. Heumann,Handbook of Elemental Speciation, Handbook of
Elemental Speciation II: Species in the Environment, Food, Medicine and Occupational Health,
Wiley, ISBN: 978-0-470-85598-0, 2005 ,788 p.
113. F. Vanhaecke, L. Moens, Recent trends in trace element determination and speciation using
inductively coupled plasma - mass spectrometry, Fresenius J Anal Chem., 1999, 364, 440-451.
114. K.G. Heumann, Isotope-dilution ICP-MS for trace element determination and speciation: from a
reference method to a routine method?, Anal. Bioanal. Chem., 2004, 378, 318-329.
115. L. Ebdon, L. Pitts, R. Cornelis, H. Crews, O.F.X. Donard and Ph. Quevauviller, Trace elemental
speciation for environment, food and health, The Royal Society of Chemistry, 2001, ISBN 0-85404-
459-0, 391 p.
116. R. Michalski, Review : Inorganic Oxyhalide By-Products in Drinking Water and Ion Chromatographic
Determination Methods, Pol. J. Environ. Stud., 2005, 14 (3), 257-268.
117. D.P. Hautman, D.J. Munch, C. Frebis, H.P. Wagner, B.V. Pepich, Review of the methods of the US
Environmental Protection Agency for bromate determination and validation of Method 317.0 for
disinfection by-product anions and low-level bromated, J. Chromatogr. A, 2001, 920, 221-229.
water intended for human consumption, 2010, http://www.vmm.be/, last accessed on 20/01/2013.
119. R. Cornelis et al., The EU network on trace element speciation in full swing, TrAC Trend. Anal.
Chem., 2000, 19, 210-214.
120. C. B'Hymer, J.A. Caruso, Selenium speciation analysis using inductively coupled plasma-mass
spectrometry, J. Chromatogr. A, 2006, 1114 (1), 1-20.
121. Y. Ogra and Y. Anan, Selenometabolomics: Identifciation of selenometabolites and specification of
their biological significance by complementary use of elemental and molecular mass spectrometry,
J. Anal. At. Spectrom., 2009, 24, 1477-1488.
122. L. Moore and A. Mahmoudkhani, Methods for removing selenium from aqueous systems,
Proceedings Tailings and Mine Waste, Vancouver, Canada, November 6 to 9, 2011.
123. P. Cauwenberg, K. Tirez and S. Van Ermen, Antibomose, Study performed by Vito in commission of
Essencia & VOKA, 2010/SCT/R/106, 2010, 239 p.
232 References
124. M. Lenz, G.H. Floor, H. Winkel, G. Román-Ross, P.F. Corvini, Online preconcentration-IC-ICP-MS for
selenium quantification and speciation at ultratraces, Environ Sci Technol., 2012, 46 (21), 11988-94.
125. Y.Anan, Y. Hatakeyama, M. Tokumoto, Y Ogra, Chromatographic behavior of Selenoproteins in rat
serum detected by Inductively Coupled Plasma Mass Spectrometry, Analytical Sciences, 2013, 29
(8), 787-792.
126. N. Unceta, F. Séby, J. Malherbe, O.F.X. Donard, Chromium speciation in solid matrices and
regulation: a review, Anal. Bioanal. Chem., 2010, 397, 1097-1111.
127. Atlas of Eh-pH diagrams, Intercomparison of thermodynamic databases, Geological Survey of Japan
Open File Report No.419, National Institute of Advanced Industrial Science and Technology,
Research Center for Deep Geological Environments, May 2005, http://www.fssm.ucam.ac.ma/
biblioadmin/opac_css/chimie/Atlas_Eh-pH_diagrams.pdf, last accessed on 28/10/2013.
128. K. Zhang, J.L. Schnoor and E.Y. Zeng, E-waste recycling: Where does it go from here ?, Environ. Sci.
Technol., 2012, 46, 10861-10867.
129. N. Unceta, F. Séby, J. Malherbe, O.F.X. Donard, Chromium speciation in solid matrices and
regulation: a review, Anal. Bioanal. Chem., 2010, 397, 1097-1111.
130. B. Novotnik, T. Zuliani, J. Scancar and R. Milacic, The determination of Cr(VI) in corrosion protection
coatings by speciated isotope dilution ICP-MS, J. Anal. At. Spectrom., 2012, 27, 1484.
131. K. Vercoutere, R. Cornelis, L. Mees, and P. Quevauviller, Certification of the contents of the
chromium(III) and chromium(VI) species and total chromium in a lyophilised solution (CRM 544),
Analyst, 1998, 123 (5), 965-969.
132. N. De Brucker, J. Theunis, V. Timmermans, K. Tirez, M. Van Holderbeke and C. Vanhoof, Heavy
Metals in Packaging on the Belgian market, Study commissioned by the Belgian Federal
Department of the Environment, 2001/MIM/R/110, VITO Integrated Environmental Studies and
Environmental Measurements, 2001, 90p.
133. C. Vanhoof, K. Tirez, B. Noten, G. Vanermen and C. Lassen, Support to a RoHS inspection campaign
in Belgium, Study accomplished under the authority of Federal Public Service of Public Health,Food
Chain Safety and Environment, 2007/MIM/R/061, VITO/COWI, 2007.
134. B.R. James, J.C. Petura, R.J. Vitale, G.R. Mussoline, Hexavalent chromium extraction from soils: a
comparison of five methods. Environ. Sci. Technol., 1995, 29 (9), 2377-2381.
135. J. Malherbe, M.P. Isaure, F. Seby, RP. Watson, P. Rodriguez-Gonzalez, P.E. Stutzman, C.W. Davis, C.
Maurizio, N. Unceta, J.R. Sieber, S.E. Long and O.F.X. Donard, Evaluation of Hexavalent Chromium
Extraction Method EPA Method 3060A for Soils Using XANES Spectroscopy, Environ. Sci. Technol.,
2011, 45 (24), 10492-10500.
136. K.Tirez, G. Silversmit, L. Vincze, K. Servaes, C. Vanhoof, N. Bleux and P. Berghmans, Speciation and
fractionation of nickel in airborne particulate matter : comparison between selective leaching and
XAS spectroscopy, J. Anal. At. Spectrom., 2011, 26, 517 -527.
137. L. Huang, Z.T. Fan, C.H. Yu, P.K. Hopke, P.J. Lioy, B. Buckley, L. Lin, and Y. Ma, Inter-conversion of
Chromium Species During Air Sampling: Effects of O3, NO2, SO2, Particle Matrices, Temperature and
Humidity, Environ. Sci. Technol., 2013, 47 (9), 4408–4415.
138. M.A. Torkmahalleh, L. Lin, T.M. Holsen, D.H. Rasmussen, P.K. Hopke, Cr speciation changes in the
presence of ozone and reactive oxygen species at low relative humidity, Atmospheric Environment,
2013, 71, 92-94.
139. K. Tirez, G. Vanermen, W. Brusten, V. Corthouts, A. Borburgh and L. Debaene, Speciation of
organotins: comparison of GC-LRMS, GC-HRMS and GC-ICP-MS, European Winter Conference on
plasma spectrochemistry, 2003, Garmisch-Partenkirchen, Germany.
140. D. Fernández, N. Sugishama, J. R. Encinar and A. Sanz-Medel, Triple Quad ICPMS (ICPQQQ) as a New
Tool for Absolute Quantitative Proteomics and Phosphoproteomics, Anal. Chem., 2012, 84,
5851−5857.
141. F. Vanhaecke, S. Saverwyns, G. De Wannemacker, L. Moens, R. Dams, Comparison of the
application of higher mass resolution and cool plasma condtions to avoid spectral interferences in
Cr(III)/Cr(VI) speciation by means of high-perfromance liquid chromatography – inductively coupled
plasma - mass spectrometry, Anal. Chim. Acta, 2000, 419, 55-64.
References 233
142. R. Clough, L.R. Drennan-Harris, C.F. Harrington, S.J. Hill and J.F. Tyson, Atomic spectrometry update:
Elemental speciation, J. Anal. At. Spectrom., 2012, 27, 1185–1224.
143. T. Hoffmann and K.G. Heumann, Elemental Speciation, Anal. Bioanal. Chem., 2012, 404, 2125–
2126.
144. L. Balcaen, L. Moens and F. Vanhaecke, Determination of isotope ratios of metals (and metalloids)
by means of inductively coupled plasma-mass spectrometry for provenancing purposes - A review,
Spectochim. Acta B, 2010, 65, 769-786.
145. F. Vanhaecke, L. Balcaen, D. Malinovsky, Use of single-collector and multi-collector ICP-mass
spectrometry for isotopic analysis, J. Anal. At. Spectrom., 2009, 24, 863-886.
146. J.S. Becker, State-of-the-art and progress in precise and accurate isotope ratiomeasurements by
ICP-MS and LA-ICP-MS, , J. Anal. At. Spectrom., 2002, 17, 1172–1185.
147. M. Berglund, M.E. Wieser, Isotopic compositions of the elements 2009 (IUPAC Technical Report),
Pure Appl. Chem., 2011, 83, 397-410.
148. J. Vogl and W. Pritzkow, Isotope Dilution Mass Spectrometry - A Primary Method of Measurement
and Its Role for RM Certification, MAPAN-JMSI, 2010, 25, 135-164.
149. J. Vrijens, P. Couck, C. Schroijen, W. Baeyens and M. Leermakers, Spectral interferences in the
analysis of cadmium in human blood by ICP-MS: comparison between high resolution sector-field
ICP-MS and quadrupole ICP-MS, J. Anal. At. Spectrom., 2011, 26, 1819-1826.
150. K. Tirez, W. Brusten, A direct ultratrace isotopic analysis of uranium in environmental and biological
samples by SFICP-MS for environmental forensics, European Winter Conference on plasma
spectrochemistry, 2011, Zaragosa, Spain.
151. P. Seuntjens, D. Mallants, J. Šimůnek, J. Patyn, D. Jacques, Sensitivity analysis of physical and
chemical properties affecting field-scale cadmium transport in a heterogeneous soil profile, J.
Hydrol., 2002, 264, 185-200.
152. P. Seuntjens, K. Tirez, J. Simunek, M.T. van Genuchten, C. Cornelis, P. Geuzens, Aging effects on
cadmium transport in undisturbed contaminated sandy soil columns, J. Env. Qual., 2001, 30, 1040–
1050.
153. P. Seuntjens, Field-scale cadmium transport in a heterogeneous layered soil, Water Air Soil Poll.,
2002, 140, 401-423.
154. B. Van der Grift, J. Griffioen, Modelling assessment of regional groundwater contamination due to
historic smelter emissions of heavy metals, J. Cont. Hydrol., 2008, 96(1-4), 48-68.
155. I. Joris, J. Bronders, B. van der Grift, P. Seuntjens, Model-based scenario analysis of the impact of
remediation measures on metal leaching from soils contaminated by historic smelter emissions,
2012, submitted Sci. Total Environ..
156. P. Seuntjens, D. Mallants, C. Cornelis, P. Geuzens, Non equilibrium cadmium leaching in layered
sandy soils, Soil Sci., 2001, 166(8), 507-519.
157. B. Van der Straeten, J. Buysse, S. Nolte, L. Lauwers, D. Claeys, G. Van Huylenbroeck, The effect of EU
derogation strategies on the compliance costs of the nitrate directive, Sci. Total Environ., 2012, 421-
422, 94–101.
158. D. Widory, E. Petelet-Giraud, A. Bregnot, J. Bronders, K. Tirez and P. Boeckx, Improving the
management of nitrate pollution in water by the use of isotope monitoring: the δ15N, δ18O & δ11B
triptych, Isot. Environ. Healt S., 2012; DOI:10.1080/10256016.2012.666540.
159. D. Widory, E. Petelet-Giraud, P. Negrel, B. Ladouche, Tracking the sources of nitrate in
groundwater using coupled nitrogen and boron isotopes: A synthesis, Environ. Sci. Technol., 2005,
39, 539-548.
160. J. Bronders, K. Tirez, N. Desmet, D. Widory, E. Petelet-Giraud, A. Bregnot, P. Boeckx, The use of
compound-specific nitrogen (δ15N,), oxygen (δ18O), and bulk boron (δ11B) isotope ratios to identify
sources of nitrate-contaminated waters: A guideline to identify polluters, Environmental Forensics,
2012, 13, 32–38.
161. V. Devulder, L. Lobo, K. Van Hoecke, P. Degryse, F. Vanhaecke, Common analyte internal
standardization as a tool for correction for mass discrimination in multi-collector inductively
coupled plasma-mass spectrometry, Spectrochimica Acta Part B, 2013, 89, 20–29.
courtesy of Tagxedo

55901413

Uploaded by

55901413

Uploaded by

Development of methods based on ICP-mass spectrometry

for the determination, speciation and isotopic analysis of

Ghent University, Faculty of Sciences, Department of Analytical Chemistry, Atomic

Flemish Institute for Technological Research (VITO)

Doctoral dissertation to meet the requirements to take the doctoral exam

Promoter: Prof. Dr. Frank Vanhaecke

Finally, I wish to thank the reader, for your interest in my work.

Kristof, december 2013

“Where is the wisdom we have lost in knowledge?

• Speciation-Determination of oxy-anions in water and solid samples

(2000) Determination (2001) Determination

(1999) Study of (2011) Determination of

1997 2000 2005 2010

CHAPTER 2 ICP-MS applications in environmental regulatory monitoring ________________ 27

CHAPTER 3 Elemental speciation – Determination of oxy-anions in water and solid samples 69

CHAPTER 5 Summary and conclusions ___________________________________________ 220

CHAPTER 1 SYNCHRONICITY IN THE EMERGENCE OF THE FLEMISH

1.1. SOME MILESTONES IN THE ENVIRONMENTAL LEGISLATION OF HEAVY METALS IN FLANDERS

The EU regulatory Groundwater Framework on the protection of groundwater against pollution

Also, a shift from “environmental protection” to “environmental sustainability” is being introduced

1.2. WHO IS IMPLEMENTING THE ENVIRONMENTAL POLICY IN FLANDERS?

(organizational chart courtesy of http://www.lne.be/organisatie/organogrammen)

1980 1985 1990 1995 2000 2005 2010

• WAC: compendium for sampling and analysis of water

CEN/TC 230 - Water analysis

The development of EN standards is governed by the principles of consensus and national

1.4. THE ATOMIC SPECTROMETRY TIMELINE IN ENVIRONMENTAL ANALYSIS

1.4.1. ICP-MS INSTRUMENTATION

nebuliser spray chamber plasma sampling interface

Low pressure (< 2 Pa)

Figure 3: General scheme of an ICP-MS instrument with indication of different components.

→ The different components of an ICP-MS instrument [15-17]

Interesting You tube links on the principles of ICP-MS :

→ How to cope with spectral interferences? [17]

Figure 4: Illustration of 10% valley definition of mass resolution (courtesy of [17]).

Mathematical correction procedures

Mathematical correction is usually applied in case of relatively simple situations (isobaric

Sample introduction systems

Cool/cold plasma technique

Collision and reaction cell technology

1.4.2. ATOMIC SPECTROMETRY TIMELINE

ICP-SFMS (reverse Nier-Johnson)

(2012) Triple quadrupole ICP-MS

1980 1985 1990 1995 2000 2005 2010

(2006) ICP-MS is adopted as reference

(2009) European standard method (EN 15841)

(2007) U.S. standard method (EPA 6800)

1.4.3. ACID DIGESTION OF ENVIRONMENTAL MATRICES

CHAPTER 2 ICP-MS APPLICATIONS IN ENVIRONMENTAL REGULATORY

2.1. PERSPECTIVE ON ELEMENTAL ANALYSIS

Ostend Ghent Antwerp

As an example, the median/quartile/range plot of ambient background concentrations of Zn in wet

A data point is considered to be an extreme value if the following conditions hold:

2.2. AMBIENT BACKGROUND CONCENTRATION OF ELEMENTS IN SOIL IN FLANDERS

→ Background sampling locations

→ European regulatory monitoring method

In short, standard procedure EN 13656 (translated as Flemish reference method CMA/2/II/A.3)

→ Background concentration of elements in soil

Top soil Baseline Baseline

2.3. AMBIENT BACKGROUND CONCENTRATION OF ELEMENTS IN AMBIENT AIR IN FLANDERS

2.3.1. AMBIENT BACKGROUND CONCENTRATION OF ELEMENTS IN FINE PARTICULATES IN FLANDERS

→ Background sampling locations

A monitoring campaign consisting of element characterization in PM10 (particles with an aerodynamic

→ European regulatory monitoring method

The precision expressed as 95 % confidence interval on PM10 is dependent on the concentration

→ Background concentration of elements in PM10

2.3.2. AMBIENT BACKGROUND CONCENTRATION OF ELEMENTS IN ATMOSPHERIC DEPOSITION IN FLANDERS

→ Background sampling locations

In Flanders, total atmospheric deposition of elements for ambient background concentration is

→ European regulatory monitoring method

→ Background concentration of elements in atmospheric deposition

2.4. AMBIENT BACKGROUND CONCENTRATION OF ELEMENTS IN RIVER SEDIMENT IN FLANDERS

In an attempt to establish a regression model to explain trace element concentrations in Flemish

→ Background sampling locations

→ European regulatory monitoring method

→ Background concentration of elements in river sediment