Environmental and health aspects of corrosion – importance of chemical speciation

Yolanda Hedberg

Licentiate thesis

Division of Surface and Corrosion Science School of Chemical Science and Engineering Royal Institute of Technology SE-100 44 Stockholm, Sweden

Stockholm, 2010

I

This licentiate thesis will, with the permission of Kungliga Tekniska Högskolan, Stockholm, be presented and defended at a public dissertation on Thursday October 28, 2010, at 10:00 in conference room 3 at the Institute for Surface Chemistry (YKI), Drottning Kristinas väg 49A, 114 28 Stockholm.

Environmental and health aspects of corrosion – importance of chemical speciation

TRITA-CHE Report 2010:32 ISSN 1654-1081 ISBN 978-91-7415-716-1

KTH Royal Institute of Technology School of Chemical Science and Engineering Department of Chemistry Division of Surface and Corrosion Science Drottning Kristinas väg 51 SE-100 44 Stockholm Sweden

Denna avhandling är skyddad enligt upphovsrättslagen. Alla rättigheter förbehålls. Copyright © 2010 Yolanda Hedberg. All rights reserved. No part of this thesis may be reproduced by any means without permission from the author. The following items are printed with permission: Paper I: © 2009 Elsevier Ltd. Paper II: © 2010 American Chemical Society Paper IV: © 2009 SETAC Paper V: © 2010 SETAC Paper VI: © 2010 Hedberg et al; licensee BioMed Central Ltd.

II

Abstract During the last decades, the interest in corrosion of metals and alloys from an environmental and health perspective has increased rapidly as a consequence of stricter environmental and human exposure legislations, their extensive use as implant materials and an increasing understanding related to occupational and/or daily exposure to airborne particles. Corrosion-induced metal release, however, needs to be understood in detail and to include knowledge related to chemical speciation, i.e. the oxidation state, complexation and chemical form of released metals, parameters of high importance when considering toxicity. In this licentiate work, corrosion-induced metal runoff from roofing materials (copper, zinc, and chromium(III)-, and chromium(VI) surface treated galvanized steel) has been investigated from an environmental perspective with focus on chemical speciation of released metals (Papers III). From these papers it was evident that the total concentration measured in the runoff water is not sufficient for any environmental risk assessment. The environmental fate including changes in chemical speciation and hence metal precipitation has to be considered. For example, it was shown that the copper concentration decreased by three orders of magnitude already in the internal drainage system of a shopping centre with a copper roof, to a concentration lower than storm water collected from a nearby parking space (Paper I). Also, speciation measurements can explain corrosion, metal release and surface processes of chromium surface treated galvanized steel at different sites (urban and marine). Any environmental risk assessment has to be done by considering all metal species released, and compared with ecotoxic values. For example, when most chromium(VI) (the most toxic species) was released, significantly less zinc was released at the same time which decreased the overall ecotoxicity of the runoff water significantly (Paper II). When assessing environmental risks by standard laboratory tests, it is important to understand all mechanisms which are possibly influenced by individual experimental parameters and which often are different for different test substances. Some metals released, as seen in the case of iron, may precipitate with time and be pH-, solution- and buffering dependent. This behavior can lead to strongly underestimated measured metal concentrations (Paper III). When particles of metals or alloys are to be investigated (Papers III-VI), it is essential to conduct a thorough particle characterization, since the surface III

properties cannot be defined. In addition, the surface properties (oxide layer properties) change with varying particle size (Paper VI) and with other experimental parameters such as dispersion (Paper VI). All iron-, and chromium-based particles investigated (Papers III-VI) revealed large differences between alloy particles and pure metals. Particles of pure iron and nickel released significantly more metals compared with particles of the investigated alloys, whereas particles of pure chromium released less metals compared with the alloys. Particles of stainless steel (AISI 316L), ferro-chromium and ferro-silicon-chromium released very low amounts of metals (Papers III-VI). The released quantity increased with increased acidity (Papers III-VI) and also in the presence of complexing agents (ongoing research). The manufacturing process is of high importance, as observed for stainless steel particles when compared with a side product from stainless steel production with similar composition that released significantly more metals (Paper III). Particles of metal oxides, i.e. chromium(III)oxide and iron(II,III)oxide, released very low amounts of metals due to their thermodynamic stability. Ongoing research activities focus on the specific influence of complexing agents and proteins on the metal release process from massive sheet and particles of metals and alloys. The applicability and the possibility to use different analytical tools are investigated and elaborated for small-sized particles. A detailed understanding of the correlation between material and particle characteristics, the metal release process, the chemical speciation in interaction with proteins and/or cells, and the particle/cell interaction is essential to enable any correlation between material/particle characteristics and toxicity. The aim of this licentiate summary is – in contrast to the six included scientific papers – to explain the importance of chemical speciation for corrosion processes from a health and environmental perspective in a popular way to reach a broad non-academic audience. The summary is hence written as a guidance document for stakeholders and the regulatory community working with environmental and health risk assessment.

IV

Populärvetenskaplig sammanfattning Vanligtvis när man tänker på korrosion, alltså metallers oftast långsamma ytförändring, så frågar man sig hur den omgivande miljön påverkar materialet. Motsatsfrågan, nämligen hur metallerna påverkar miljön och människors hälsa, har varit utgångspunkt för studierna inom ramen för denna licentiatavhandling. De senaste årtiondena har politiker, beslutsfattare, miljö- och hälsovetare intresserat sig mer och mer för den eventuella påverkan och de effekter som metaller som frigörs från utomhuskonstruktioner, metallpartiklar eller metalliska implantat kan ha på miljön och på människors hälsa. För att kunna besvara dessa frågor måste man först förstå hur och varför metaller frigörs. Detta kan ske via olika processer, kemiska, eller fysikaliska (t.ex. genom nötning), eller som en följd av korrosion. En av de viktigaste parametrarna är stabiliteten och skyddsegenskaperna hos de naturligt bildade oxiderna på metallerna eller legeringarna. Om oxiden är väldigt stabil korroderar metallen mycket långsamt och en liten mängd metall frigörs generellt till omgivningen. Ingen metall/legering är dock inert, även material med högt korrosionsmotstånd frigör metaller i olika hög grad. Det är väl känt att oxidskiktens stabilitet skiljer sig mellan olika material. Tyvärr så blir de också påverkade av bland annat framställningsprocessen, lagring, transport, och många andra parametrar som gör att två metaller med lika sammansättning kan ha oxidskikt med helt olika stabilitet och därmed frigöra olika mängder metall. Detta gör det svårt att jämföra metallfrigörelse från både lika och olika metaller/legeringar. Ännu svårare är det att jämföra metallpartiklar, vars ytor inte kan slipas eller poleras, och därmed är svåra att definiera. Det är därför väldigt viktigt att känna till alla olika parametrar och hur de påverkar metallfrigörningsprocessen. En stor del av detta licentiatarbete har därför fokuserat på att karakterisera de undersökta materialen och partiklarna korrelerat till mängden frigjord metall. Nästa steg är att förstå hur den metall som frigörs påverkar hälsan och/eller miljön. I de första två artiklarna har metallfrigörningen från olika takmaterial undersökts. Dessa studier visade tydligt att inte bara mängden av frigjorda metaller, utan också deras kemiska form och oxidationstillstånd, är viktig kunskap för att förstå hur metallerna kan spridas och växelverka med miljön samt hur de frigjorda metallerna kan förklara yt- och korrosionsprocesser. I de sista fyra artiklarna har den mängd och form av frigjorda metaller från metall- och legeringspartiklar undersökts. Partiklarna har bl. a. exponerats för konstgjorda kroppsvätskor för att simulera förhållanden då partiklarna andas in, sväljs, eller kommer i V

ögon- och hudkontakt. Resultaten visade tydligt att båda järn- och nickelpartiklar frigör relativt stora mängder metaller medan de flesta legeringspartiklar eller krompartiklar (vilka alla har en väldigt stabil ytoxid) frigör väldigt små mängder metaller. Toxikologiska tester som gjordes på partiklarna visade att de flesta var icke toxiska, medan nickelpartiklarna visade hög toxicitet. Många frågor kvarstår dock hur man kan korrelera toxicitet av metallpartiklar med deras materialegenskaper och partikelegenskaper, studier som forsätter inom ramen för doktorsarbetet.

VI

Populärwissenschaftliche Zusammenfassung Spätestens seit in der EU die REACH Verordnung in Kraft ist, die besagt, dass ab Dezember 2010 (für metallische Produkte) alle Produzenten für die nötigen Daten für eine Umweltklassifizierung ihrer Produkte aufkommen müssen, ist die Forschung im Bereich der Umwelt- oder Gesundheitsbeeinträchtigung durch metallische Produkte entscheidend angestiegen. In Schweden wird außerdem seit langem versucht, die Metallabschwemmung von Metalldächern zu minimieren. Außerdem wird immer mehr Aufmerksamkeit auf metallische Bestandteile in Partikeln in der Luft, die zum Beispiel vom Autoverkehr oder der U-Bahn kommen können, gelegt. Allen diesen Beispielen ist gemeinsam, dass der zugrundeliegende Prozess der Metallfreisetzung die Korrosion ist, also das langsame Zerfallen von Metall. Hierbei handelt es sich um einen elektrochemischen Prozess, dessen Geschwindigkeit stark davon abhängt, wie stabil die so genannte Oxidschicht, die das Metall umgibt, ist. Je nach Material, Herstellung, Lagerung, und – bei Partikeln – Größe der Partikel, ist diese Oxidschicht stabiler oder weniger stabil. Bei einer stabilen Oxidschicht wird allgemein nur wenig Metall in die Umwelt oder den umgebenden Organismus freigesetzt. Ist sie sehr unstabil, so löst sich das Metall sehr schnell auf und kann zu akuter Toxizität führen. Dies ist zum Beispiel bei reinen Eisen- oder Nickelpartikeln der Fall, die sich in simulierten Körperflüssigkeiten sehr schnell komplett auflösen und zu entsprechenden toxischen Reaktionen führen können. Legierungspartikel wie rostfreier Stahl dagegen lösen sich wegen ihrer stabilen Oxidschicht so gut wie gar nicht auf und führen daher auch zu keiner direkten toxischen Reaktion, z.B. Entzündung. Wenn man Metallpartikel untersucht, dann hat man den Nachteil, dass man ihre Oberfläche (die so wichtige Oxidschicht) nicht beeinflussen kann. Unterschiede in Lagerzeiten, Herstellung u.s.w. können zu unterschiedlich stabilen Oxidschichten derselben Partikel führen. Das heißt, dass es schwierig ist, Partikel verschiedenen Materials oder verschiedener Größe zu vergleichen, wenn sich andere Parameter auch unterscheiden. Deshalb ist es wichtig, alle Parameter genau zu kennen und die Partikel sehr genau zu untersuchen (charakterisieren), damit man Ergebnisse vergleichen kann. Auch der Einfluss von anderen experimentellen Parametern auf die Metallfreisetzung muss genauestens bekannt sein. Sowohl im Körper als auch in der Umwelt ist es wichtig, dass man nicht nur weiß wie viel Metall freigesetzt wird von z.B. eingeatmeten VII

Metallpartikeln im Körper oder Metalldächern in der Umwelt, sondern auch in welcher Form (z.B. Oxidationsstufe oder Komplexierung) das Metall vorliegt (so genannte Speziation). Diese entscheidet nämlich, ob das gelöste Metall z.B. im Regenwasser schon sehr schnell ausfällt oder nicht. Im Körper ist die Speziation dagegen entscheidend für die eventuelle Entgiftung und die Intensität der Metallfreisetzung. Das Beachten der Speziation, aber auch das genaue Kennen des zu untersuchenden Materiales (z.B. die Partikelkennwerte) und der beeinflussenden experimentellen Parameter, sind notwendig für jegliche Risikeneinschätzung (z.B. beim Setzen von Grenzwerten) im Bereich der Gesundheit und Umwelt.

VIII

List of papers This thesis is a summary of the following papers: I.

Storm water runoff measurements of copper from a naturally patinated roof and from a parking space. Aspects on environmental fate and chemical speciation. I. Odnevall Wallinder, Y. Hedberg, P. Dromberg Water Research 43, 5031-38 (2009)

II.

Chromium(III) and chromium(VI) surface treated galvanized steel for outdoor constructions – environmental aspects. D. Lindström, Y. Hedberg, I. Odnevall Wallinder Environmental Science and Technology 44, 4322-27 (2010)

III.

Transformation/dissolution studies on the release of iron and chromium from particles of alloys compared with their pure metals and selected metal oxides. Y. Hedberg, I. Odnevall Wallinder Submitted for publication

IV.

Bioaccessibility studies of ferro-chromium alloy particles for a simulated inhalation scenario. A comparative study with the pure metals and stainless steel. K. Midander, A. de Frutos, Y. Hedberg, G. Darrie, I. Odnevall Wallinder Integrated Environmental Assessment and Management 6(3), 441-55 (2010)

V.

Particles, sweat and tears. A comparative study on bioaccessibility of ferrochromium alloy and stainless steel particles, the pure metals and their metal oxides in simulated skin and eye contact. Y. Hedberg, K. Midander, I. Odnevall Wallinder Integrated Environmental Assessment and Management 6(3), 456-68 (2010)

VI.

Bioaccessibility, biovailability and cytotoxicity of iron and chromium based particles – an inhalation scenario. Y. Hedberg, J. Gustafsson, L. Möller, H.L. Karlsson, I. Odnevall Wallinder Particle and Fibre Toxicology 7(23), (2010)

IX

Author’s contribution to the papers The author’s contribution to the papers is summarized below: I.

II.

III.

IV.

V.

VI.

All experimental work, i.e. sample preparation for the metal analysis, metal analysis, and metal speciation (polarography), involved in drafting the manuscript, revision and proof reading. I. Odnevall Wallinder organized the work, designed the experiments, and drafted the manuscript. P. Dromberg was the responsible person for organizing and sampling at Stockholm Vatten AB. All speciation measurements and comparative chromium analysis with polarography and GF-AAS. Major part in interpretation and correlation of the results, drafting the manuscript, submission, revision, and proof reading. D. Lindström carried out and interpreted the zinc runoff measurements, performed surface analytical techniques (XRD, FTIR, SEM/EDS), and corrosion rate tests, and was involved in drafting of the manuscript. I. Odnevall Wallinder was the supervisor and carried out the XPS measurements. All experimental work (exposures, metal analysis, speciation, particle characterization), interpretation, and drafting and submitting, with I. Odnevall Wallinder as supervisor. Tao Jiang and Jon Brunk did minor parts of the experimental work. All comparative exposures and metal analysis for the pure metals and stainless steel. All speciation measurements. Interpretation, large part in drafting and revision. The time-dependent exposures for the ferrochromium alloys were carried out by A. de Frutos. The major part of metal analysis for the bioaccessibility tests, particle characterization, and drafting of the manuscript was carried out by K. Midander. G. Darrie was the industry partner, helped with necessary information on the particles, and was involved in drafting and proof reading. Tao Jiang conducted a minor part in exposures and metal analysis. I. Odnevall Wallinder conducted the XPS analysis and was the main supervisor. All experiments (exposures, metal analysis, particle characterization). Major part in drafting of the manuscript, submission, revision, and proof reading with K. Midander and I. Odnevall Wallinder as supervisors. Carried out the speciation measurements, part of the exposures, metal analyses, and particle characterization, and organized and designed the paper. Major part in drafting, submitting, revision, and proof reading. J. Gustafsson carried out all the toxicological tests and was involved in drafting of the manuscript. Tao Jiang and Klara Midander did parts of the exposures, metal analyses, and particle characterization. I. Odnevall Wallinder carried out the XPS analysis. I. Odnevall Wallinder, H.L. Karlsson and L. Möller were the supervisors.

X

Work not included in this thesis Reports Acute 7 days transformation/dissolution test (100 mg/L, pH 6.0) for chromium(III)oxide Y. Hedberg, S. Hosseinpour, and I. Odnevall Wallinder, commissioned by the International Chromium Development Association, May 2010. Transformation dissolution protocol testing of a CuZn20Ni18 alloy. J. Brunk, Y. Hedberg, I. Odnevall Wallinder, commissioned by European Copper Institute, January 2010. Chronic 28 days transformation/dissolution test (1 mg/L, pH 6.0) for chromium(III)oxide as a worst case scenario for chromium containing alloys and pure chromium particles. Y. Hedberg and I. Odnevall Wallinder, commissioned by the International Chromium Development Association, December 2009. Bioaccessibility of antimony released from four different antimony compounds in synthetic biological media. Y. Hedberg, T. Jiang, I. Odnevall Wallinder, commissioned by the International Antimony Association, June 2010. Bioaccessibility of zirconium and titanium released from different zircon and rutile products in synthetic biological media. Y. Hedberg and I. Odnevall Wallinder, commissioned by QIT Fer et Titane / RTIT Centre de Technologie, Canada, October 2009. Bioaccessibility of molybdenum released from pure molybdenum and different molybdenum compounds in synthetic biological media. Y. Hedberg and I. Odnevall Wallinder commissioned by the International Molybdenum Association (IMOA), August 2009. Transformation dissolution protocol testing of a CuZn20Ni18 alloy. Project no: G-EU-07-02d. S. Jafarzadeh, G. Herting, Y. Ullmann, I. Odnevall Wallinder commissioned by the European Copper Institute, February 2008. XI

Conference proceedings Bioaccessibility of ferro-chromium and ferro-silicon-chromium particles compared to pure metals and stainless steel – aspects of human exposure. K. Midander, A. de Frutos, Y. Hedberg, G. Darrie, I. Odnevall Wallinder, The 12th international ferro alloy congress (INFACON XII), Sustainable future, Helsinki, Finland, June 6-9, 2010. Elaboration of a metal release test for massive metal sheet – Effect of different parameters on the copper release rate. Y. Ullmann, S. Jafarzadeh, G. Herting, I. Odnevall Wallinder, Proceedings for the 17th International Corrosion Congress, Oct. 2008 in Las Vegas, US. Presentations Essential factors assessing the extent of metals released from particles of pure metals, alloys and compounds during the OECD transformation/dissolution test. Oral presentation, 15th Nordic Corrosion Congress, May 2010, Stockholm, Sweden. Metal release from iron- and chromium-based particles in artificial sweat and artificial tear fluid. Oral presentation, 5th Kurt-Schwabe-Symposium, May 2009, Erlangen, Germany. Elaboration of a metal release test for massive metal sheet – Effect of different parameters on the copper release rate. Poster presentation, 17th International Corrosion Congress, October 2008, Las Vegas, US. Popular science publications Die Bindekapazität von Entwässerungssystemen für Kupfer von Kupferdächern - Vergleich von Regenwasserkupferkonzentrationen in einem Kupferdachentwässerungssystem und einem Parkplatz. Y. Hedberg, P. Dromberg, I. Odnevall Wallinder, Wasser, Luft und Boden, 3/2010, 22-23, 2010. Vart tar den koppar som frigörs från koppartak vägen? I. Odnevall Wallinder and Y. Ullmann, Bygg & Teknik 4/09, 28, 2009. XII

Preface This licentiate thesis reflects the first part of my future doctoral thesis. It describes the initial phase of my Ph.D. studies with the main focus on chemical speciation. It is thought to be an understandable explanation of the importance of chemical speciation in corrosion-induced metal release processes, taking place in the environment and in the human body. Also, the difference between metal particles and massive metal from a corrosion, metal release, and toxicity perspective is discussed and explained. This thesis is a result of close research collaboration with the copper, zinc (galvanizers), ferrochromium, and stainless steel industry, as well as policy makers. The interest in the diffuse dispersion of metals into the environment and possible adverse health effects, induced by metal containing particles, has increased significantly during the last years among the public, the policy and decision makers. In December 2007, REACH, an EU regulation on the use of chemicals (also including metals), came into force, which has been the driving force for many studies during the last years. My research activities started in October 2007 (Studienarbeit, somehow similar to Bachelor thesis) working with a standardized test simulating metal release from poorly soluble chemicals, including metals, into surface and marine waters (OECD transformation/dissolution test). These studies continued in 2008 with my master thesis on both environmental and health effects induced by iron- and chromium-based particles from a surface perspective. My Ph.D. studies started in 2009. Despite the urgent need of the industry and legislators for accurate and quantitative data, these studies opened up the opportunity for in-depth studies of many phenomena and generate new knowledge of ongoing interactions between organisms and corroding metals or metal particles. Many of these in-depth studies are ongoing and not included within the scope of this licentiate thesis. This thesis is a first step for the understanding of the possible correlation between corrosion processes, metal release, chemical speciation, and protein and cell interaction (toxicity) of metal particles or released metal. This has been accomplished through a highly interdisciplinary approach combining knowledge including materials and surface characteristics, corrosion mechanisms, metal release, chemical speciation, toxicology and risk assessment. XIII

This summary of the first six papers is at the same time an opportunity to clarify the importance of chemical speciation for environmental and health aspects of corrosion. The summary aims to explain and discuss aspects of chemical speciation and bioavailability, in a popular way, for a broad audience with (non-)academic background, e.g. policy makers, lawyers, environmentalists, ecotoxicologists, toxicologists, and other important stakeholders and legislators. Detailed scientific information is given in the scientific papers of this thesis.

XIV

Table of contents Motivation __________________________________________________ 1  Importance of chemical speciation of released metals on the environment and on human health _________________________________________ 4  Chemical speciation of chromium and copper _____________________ 7  Measurements and estimations of chemical speciation ______________ 8  Techniques used ____________________________________________ 10  Differences between massive metals and metal particles ____________ 12  Differences between pure metals and alloys (Papers III-VI) _________ 14  Environmental risk assessment – possible measurements and obstacles (Papers I-III) ______________________________________________ 18  Chemical speciation and the entire interacting system of released metals and species _____________________________________________________________ 18  Experimental parameters of high importance for environmental risk assessment 22  Material and surface conditions ________________________________________ 22  Outdoor exposures __________________________________________________ 22  Acid cleaning and handling ___________________________________________ 23  Laboratory exposures – OECD transformation/dissolution tests _______________ 23 

Human health risk assessment for metal particles – important parameters for in-vitro and in-vivo testing (Papers IV-VI) ____________________ 25  Main conclusions ___________________________________________ 29  Acknowledgements/Danksagung _______________________________ 30  References _________________________________________________ 31 

XV

XVI

Motivation A completely new research area from a corrosion perspective was initiated in the 1990’s at the division of Corrosion Science at KTH by Prof. Inger Odnevall Wallinder. From then on, the question of how the material influences the environment, rather than the other way around, was investigated with a surface perspective. These research efforts have resulted in a research team with senior scientists and PhD-students, and the generation of a large number of publications and several academic theses. Outside of this group, in-depth investigations of corrosion-induced metal release with a material and (eco-)toxicological perspective are still scarce in the scientific literature. The first aspects investigated were focused to understand the runoff process at atmospheric conditions, in contrast to the corrosion process, the amount of metals that were released into the environment, and to understand the environmental fate of metals including aspects on chemical speciation (some selected: 1-16). Continuing this work, aspects on chemical speciation of metal runoff have been investigated within this thesis (Papers I, II). In the beginning of 2000, research activities were also directed towards corrosion-induced metal release aspects that potentially may adversely influence human health, e.g. by released metals from massive metals and alloys in food related applications 17-24. These activities were followed by parallel investigations of similar aspects for particles of metals and alloys exposed in different synthetic biological fluids simulating e.g. the possible inhalation of airborne particles in the environment or at working environments (e.g. powder metallurgy), and the elaboration of a reproducible metal release test for particles of different material and size 24. Such a test was urgently needed by the industry to ensure accurate and quantitative metal release (bioaccessibility) data to be generated for commercial products and raw material, to allow read-across possibilities with other materials, and to avoid animal testing as far as possible. Data was to be combined with available, and possibly newly generated, toxicity data in a sophisticated way into risk assessment dossiers for each group of materials before the deadline of December 2010 (for metals and alloys imported and/or manufactured in an annual amount exceeding 1 ton/year), set by the EU legislative REACH (Registration, Evaluation, Authorisation and Restriction of Chemicals, 25-26). REACH obligates all producers within the EU or exporting into EU to prove that their products are safe for use 1

from an environmental and health perspective: “No data no market”. This fact combined with a genuine industrial interest for an in-depth understanding of their materials were the major driving forces for a the implementation of a large interdisciplinary project, commissioned by the International Chromium Development Association (ICDA), where KTH was responsible for generating bioaccessibility data combined with material and surface characterization of ferrochromium and ferrosiliconchromium alloys, and the Finnish Institute of Occupational Health (FIOH) was responsible for the generation of risk assessment dossiers (e.g. Papers IIIVI). In parallel, other research activities were conducted at KTH commissioned by other international metal associations generating bioaccessibility data from a material and surface perspective for different metals, alloys and compounds, e.g. chromium, silver, tin, molybdenum, and ferrosilicon alloys. Results from these investigations are not included within the framework of this thesis. Many fundamental questions regarding the influence of experimental parameters on generated results have gradually emerged along the research path of these activities, e.g. the real interaction between proteins and particles, or released metals, the interaction of particles, of varying size or properties, and cultivated human cells, and the influence of different chemical species on the metal release of particles. To answer some of these aspects, a interdisciplinary research team, the “Stockholm Particle Group” (SPG), was initiated in 2007, with closely collaborating scientists at KTH (Surface and Corrosion Science), Karolinska Institutet (Analytical Toxicology) and Stockholm University (Aerosol Science). This team has further been strengthened by a close collaboration with experts on proteins at the Div. Surface and Corrosion Science and YKI – Institute for Surface Chemistry. Since particle characteristics have a key role for most observed effects, close collaboration with experts at international academia and research institutes has also been established. This, more fundamental, work is however only to a minor extent included in this licentiate thesis, since many studies are ongoing. In all, any correlation between material and particle characteristics, bioaccessibility, bioavailability, in-vitro toxicity, and in-vivo toxicity, has been found difficult, since many influencing factors simply have not been possible to study so far and due to the fact that many parameters show synergistic effects. The characteristics of the particles or massive material is 2

a key factor that influences their potential toxicity. However, the challenge is to understand which, how and why individual parameters influence toxicity and potential synergistic effects. Except for detailed understanding of particle characteristics, also the chemical speciation of the metals released is of high importance (a parameter that depends on particle characteristics) for any induced toxicity 27-29. The overall vision of ongoing research activities within this thesis and future studies is to be able to predict and correlate toxicity induced by metal containing particles with specific particle characteristics, as schematically illustrated in Figure 1. Such correlations would save costs and avoid a large number of animal testing. We are today, however, still far away from achieving this universal goal. Even though many questions still remain, some knowledge gaps are answered within the framework of this licentiate thesis. relevance

??

?

particles of metals and alloys

bioaccessibility bioavailability toxicity particle characterisation

protein interaction

cell interaction

Figure 1: Schematic illustration of the research approach conducted to further understand the correlation between toxicity and particle characteristics.

It should be underlined, that this summary does not aim to be an allincluding detailed scientific summary, but aims to provide a short, popular and simplified insight into this field, as explained in the preface. Detailed scientific information is given in the scientific paper(s).

3

Importance of chemical speciation of released metals on the environment and on human health Chemical speciation of metals describes, generally spoken, the actual chemical form of the metal, including their complexed species and oxidation states, e.g. Cu(H2O)62+ (short: Cu2+), or strong complexes such as copper complexed to different organic species, Cu-DOM (dissolved organic matter) 29. It should be underlined that the chemical speciation, the metal concentration and the metal itself determine if the metal-organism interaction is healthy or not. When the metal is of too low concentration, or not in a bioavailable form, the organism shows symptoms of deficiency 2931 . If the metal is a non-essential metal or the metal concentration is too high, the organism shows symptoms of chronic or acute toxicity. In contrast to chemical speciation, the total metal concentration describes the total content of metal without any consideration of complexation or oxidation state. While it is relatively easy and common to measure the total metal concentration, the speciation of metals has to be taken into account for any environmental or human risk assessment. The reasons are several. First of all, any organism reacts differently to different chemical forms of a metal. For example, bioactive chromium, which is easily taken up by the cell, is a complex of Cr(III) linked to amino acids or other biomolecules, while inorganic Cr(III) is nearly non-bioavailable. Chemical speciation of chromium is hence crucial for any toxicological consideration (see e.g. 2935 ). Chemical speciation in the environment is mostly governed by precipitation processes, or in other words immobilization or retention of metals. This is schematically illustrated in Figure 2. When metals are released from external metal constructions (metal runoff), their speciation is mostly as active (free) metal ions, e.g. Cu2+ (so called labile fraction) at the immediate release situation. After interaction with for example concrete pavement and organic species, the labile fraction becomes more and more complexed and precipitates (becomes immobile). Within this thesis, it has been shown that the total copper concentration, released from a copper roof decreased by a factor of 1000 already after interaction with the internal drainage system of the building (Paper I). The chemical speciation of metals that reach the environment with sensitive water organisms (e.g. in a lake) is of high importance for the aquatic ecotoxicity as well as the content of organic species in the lake, the pH, and other species. The corrosioninduced runoff process and the environmental fate of released metals have 4

been investigated extensively, mainly for copper, zinc, chromium and nickel 1-5, 7, 9, 12, 14, 16. Metals in the form of e.g. inhaled metal particles or a metallic implant in contact with the human body induce also other effects that need to be taken into consideration, as illustrated in Figure 3. The metal (particle) itself in interaction with cells can cause toxicity, e.g. oxidative stress 28, either chronic (more pronounced for insoluble particles) or acute toxicity, e.g. by releasing metals. The reducing and very organic environment, including many proteins, changes the speciation of the released metal. This, in turn, shifts the thermodynamic equilibrium and results in more metals being released. Hence, metal particles can be dissolved, or release metals, significantly faster in the human body compared with an inorganic environment of similar ionic strength and pH (ongoing research). Finally, the individual ability of the human body to complex and reduce potentially dangerous (too high concentrations) metals interacting with the human body, is of high importance for potential toxicity. To summarize,  Chemical speciation of metals determines the extent of (eco-)toxicity.  Chemical speciation of metals released from outdoor constructions often immobilizes the metal by precipitation.  Metal release from e.g. metal particles or implant materials within humans is enhanced by complexation with organic species (mostly proteins).  Complexation and the reducing capacity of fluids within the human body are crucial processes that act as defense mechanisms against metal-induced toxicity.

5

1 Metal runoff:  High percentage of  labile fraction

2 3

4

Immobilization of metal due to  complexation

Interaction with  drainage system,  concrete, etc.: High complexed  (inactive) fraction

Interaction with water organisms and the  environment (very small fraction of the  original metal runoff): pH and chemical speciation important for ecotoxicity

Figure 2: Importance of chemical speciation, especially complexation, during the environmental fate of corrosion-induced metal runoff [original drawing by B. Bahar]. 1 The human body is exposed to  metal, e.g. metal  particles inhaled

5 When only dissolved metal is left or present, the ability of the body to reduce and/or complex the  metal is of high importance for the  individual toxicity.

3 The reducing environment and  presence of many complexing agents  changes immediately the  chemical speciation and  concentration of  released Men+

Men+X Men+ Men+

2a The particles are very passive (insoluble) (possible chronic toxicity)

Men+

4 2b The particles release metals (possible acute toxicity)

Even more metal is  released due to the  decreased concentration of free metal ions

Figure 3: Importance of chemical speciation during metal release and toxicological processes in the human body [original drawing by J. Midander]. 6

Chemical speciation of chromium and copper Since the chemical speciation primarily has pH, absence /  been determined for chromium and copper pH, oxidizing destruction of  species within this licentiate thesis (Cu: Paper I, Cr: complexing agents Papers II-VI), their possible speciation and (eco-)toxicity will be shortly discussed. For all metals, complexation is important for oxidation the bioavailability. Generally, the commonly denoted labile fraction, including free metal Cr6+  Cr3+  Cr3+X ions (e.g. Cu2+, Cr3+) and weak metal complexes (e.g. with hydroxide or halides) is reduction highly reactive and/or bioavailable. This is not the case for metals that are not easily taken up by an organism, e.g. Cr3+. For Cr3+, only certain organic complexes (bioactive chromium) allow pH, presence of  pH, reducing complexing absorption by the organism. In general, strong species agents complexes, mostly with organic species, are less reactive, less bioavailable (with important Figure 4: Schematic illustration exceptions!) and precipitate more easily. of chromium speciation. Copper ions can have oxidation state I or II, where II is the normal case in the presence of oxygen 32, 36. While humans, especially children and women, often have deficiency of copper, copper can be toxic to sensitive water organisms at relatively low concentrations 32, 36. The reactivity, toxicity, and bioavailability of Cr3+ and Cr6+ are very different, at least for humans. Figure 4 illustrates schematically the most important speciation of chromium. Cr6+, which normally exists as chromate or dichromate, easily penetrates the cell membranes due to an uptake mechanism intended for sulfates and phosphates 37-38. However, while the difference between Cr6+ and Cr3+ toxicity is large for humans (factor 1000) 37-39 , it is small for sensitive water organisms both in fresh and sea water 4041 . It is important to mention that Cr6+ normally does not form any complex with organic species, but is first reduced to Cr3+. Under normal fresh water or atmospheric conditions, Cr6+ is not stable. Cr6+ is only stable at pH exceeding 7 and in the presence of oxygen (e.g. in sea water) 29, 42-43. In contrast, Cr3+ is more stable at low pH. The probability of forming strong complexes decreases rapidly for Cr3+ when the pH decreases to less than 4 29 . An oxidizing or reducing environment is in addition important for the 7

oxidation state of chromium (e.g. the presence of the oxidizing agent H2O2 generated as a result of inflammation). The presence of complexing agents, e.g. proteins, and their stability (depending on pH, UV irradiation, ionic strength, and other present species) determines the quantity and strength of organic chromium complexes formed.

Measurements and estimations of chemical speciation The large importance of chemical speciation in environmental science and toxicology is evident. However, chemical speciation measurements are not straightforward since they are strongly solution (matrix) and metal dependent, often very time-consuming and also expensive. Therefore, most risk assessments are today based on data of total metal concentrations (measured mainly by atomic absorption spectroscopy and/or inductively coupled plasma mass / optical emission spectroscopy). When the environment of interest, e.g. a fresh water lake, is known in terms of pH, ionic strength, alkalinity, cations and anions, organic matter etc, the speciation of the metal of interest can be calculated and predicted by using different models such as MINTEQ2A 44, and be correlated with ecotoxic values for a certain organism and metal by using the biotic ligand model, BLM 44. However, any modeling of metal speciation can be erroneous due to too few input parameters, or lack of crucial environmental information 33. For instance, the biotic ligand model should only be used in the pH range of 6 to 8 44, conditions seldom relevant for Swedish lakes 30. The chemical speciation of metals can be determined by different techniques. For some metals, e.g. copper, an ion selective electrode is available which allows the measurement of free metal ions in contrast to the total concentration. Also, photometric techniques and chromatography can be used for speciation measurements. In this licentiate thesis, mostly stripping voltammetry (polarography), an electrochemical technique, has been used. All methods have in common that the complex interaction between different species, complexing agents, buffering agents, and other supporting chemicals has to be taken into account, and be understood. At the same time, this interaction is often a limitation for each technique. Since speciation is very sensitive to the water chemistry, pH value and/or other experimental parameters, the original speciation can be changed during the measurement or its preparation. The selection of storage and sample vessel 8

material (low adsorption of metals at the original pH is necessary) of the samples is of high importance, since the chemical speciation of metals in aqueous solution in many cases also is time-dependent. With voltammetry (in this case polarography), only the so-called labile fraction (free metal ions and weakly complexed metal) is measured. The total concentration can, in addition, be measured by destroying organic complexes with e.g. UV irradiation and/or acid and hydrogen peroxide. With polarography (with a mercury drop working electrode), copper, zinc, cadmium, and lead can directly be measured by anodic stripping voltammetry since they form amalgams with the working electrode (used in Paper I). For all other metals (e.g. chromium) or organic species of interest, supporting complexing agents are needed. Depending on the choice of any complexing agent, any supporting chemicals, the pH, and the electrochemical parameters, a large variety of different methods can be applied to analyze specific metal species (e.g. certain complexes or oxidation states) at different concentrations. For chromium, DTPA (diethylenetriaminepentaacetate) was used as a complexing agent, nitrate as a magnifier, and sodium acetate as buffer. The pH had to be kept at 6.2 ± 0.1 to enable a sufficient sensitivity of the measurements. The method had to be adjusted and elaborated (in terms of deposition time, amount of supporting electrolyte, and preparation of the sample) for each solution matrix separately (see Papers II-VI). The principle of this method is given in e.g. 45-47.

9

Techniques used Detailed information on all equipment, the cleaning procedures, exposure conditions, and instruments (incubators, centrifuges etc.) both at outdoor conditions (Papers I-II), and at laboratory conditions (Papers III-VI), is given in the papers. In Table 1, all techniques used within the different papers are listed together with their abbreviations. It should be underlined that many techniques have been used in close collaboration with different colleagues, indicated in the last column of the table. Table 1: Summary of techniques used within this licentiate thesis. If not indicated differently, all colleagues at the Div. Surface and Corrosion Science, KTH. Analysis performed

Metal analysis in solution / chemical speciation

Abbreviation

Name

Used for Papers or nonpublished Papers IIVI

Measured by/at

DPAdCSV

Differential pulse adsorptive cathodic stripping voltammetry

DPASV

Differential pulse anodic stripping voltammetry

Paper I

YH

AAS

Atomic absorption spectroscopy (flame mode)

Papers I-VI

YH

GF-AAS

Graphite furnace atomic absorption spectroscopy

Papers I-VI

YH

ICP-OES

Inductively coupled plasma optical emission spectroscopy

Papers IIIIV

Dep. Geological Sciences, Stockholm University

MALDITOF-MS

Matrix assisted laser desorption / ionization – time of flight – mass spectroscopy

HPLC

High performance liquid chromatography

10

Paper II

Y. Hedberg (YH)

J. Jacksén (Div. Analytical chemistry, KTH) D. Lindström (DL)

XPS SEM/EDS

Surface analysis

I. Odnevall Wallinder YH, DL, K. Midander (KM)

Transmission electron microscopy

M. Tornberg (Swerea Kimab AB)

XRD FTIR

X-ray diffraction Paper II Fourier transform infrared Paper II spectroscopy

P. Qiu, DL D. Persson (Swerea Kimab AB), DL

EBSD

Electron backscatter diffraction

O. Karlsson (Swerea Kimab AB)

CRM

Confocal Raman microspectroscopy Potentiodynamic polarization, Open-circuit potential

J. Hedberg, YH

Quartz crystal microbalance (adsorption of proteins on metal surfaces)

M. Lundin

QCM

Zeta-potential measurement LD

Toxicological tests

Papers IIVI Papers IIVI

TEM

Electrochemical measurements

Particle specific characterization

X-ray photoelectron spectroscopy Scanning electron microscopy / energy dispersive spectroscopy

YH, C. Baldizzone (CB)

Paper VI

YH, CB

Laser diffraction (particle size distribution in solution)

Paper IIIVI

YH, KM

BET

Brunauer Emmet Teller method (specific surface area)

Paper III VI

Kanthal AB

PCCS

Photon-Cross-CorrelationSpectroscopy

Cytotoxicity assay

Cell death

Paper VI

J. Gustafsson (JG)

Comet assay

DNA damage

Paper VI

JG

Hemolysis assay

Lysis of erythrocytes

Paper VI

JG

11

YH

Differences between massive metals and metal particles For any corrosion, metal release, or toxicity studies with metals or alloys, surface characteristics are crucial for the outcome. In particular, the stability, conductivity, crystallographic structure, adhesion to the bulk, and composition of the surface oxide are all decisive properties for its protection from further reaction (dissolution/release of metals) 48-49. The surface oxide stability is also influenced by its composition, heat treatments, metal manufacturing processes as well as by transport, storage (ageing) (see e.g Paper II, 21, 50), and exposure conditions. When comparing different materials, it is hence essential to achieve as similar surface conditions as possible, e.g. by an identical surface finish process (grinding, polishing, cleaning). This is, however, unfortunately not possible for particles. The surface oxide on metal particles is hence often different from the surface oxide on massive metal with defined surface conditions. Differences in storage and transport conditions, humidity, manufacturing processes etc. make it difficult or impossible to compare particles of the same material from a surface and corrosion perspective. Particle size and size distribution strongly influence the surface oxide (Paper VI), the extent of metal release (Papers III, IV, V, VI, 51-52), and the toxicological response 28, 52-53. Other particle characteristics such as surface charge and lateral distribution of the surface charge can be different for differently sized particles (and compared with massive metal), as indicated in Paper VI, factors that may influence the toxicological response. Particle agglomeration is another particle-specific parameter that is strongly dependent on the surface size, the medium, the material, surface charge, and magnetic properties. Agglomeration strongly influences the extent of metal release and probably also subsequent toxicity. For example, different particle loadings (mass or number of particles loaded per volume medium) have been observed to influence the extent of metal release 24, 51. It should be underlined that several parameters, e.g. the surface oxide stability, surface charge, agglomeration etc. can be influenced by specific particle properties such as particle size. Many parameters such as agglomeration 24, 51, 54 and surface charge (Paper VI) are in addition influenced by certain experimental conditions, e.g. the particle loading, sonication, and agitation. It is therefore generally essential to characterize the particles as thorough as possible to enable any comparison with other studies. The morphology (SEM, Papers III-VI), bulk composition (EDS, Papers III-VI), surface oxide 12

composition (XPS, Papers III-VI), size and size distribution (LD, Papers III-VI) including information on agglomeration in solution (LD, Papers IIIVI), and specific surface area (BET, Papers III-VI) have been characterized for all particles investigated. However, specific information on e.g individual phases (EBSD and AFM, ongoing), surface charge (zeta potential, Paper VI and ongoing), electrochemical passivity (electrochemical techniques, ongoing), identification of the exact chemical composition of the surface oxide (CRM, ongoing) is more difficult to obtain, but all the same very important. Figure 5 illustrates some of the most important particle characteristics and surface oxide properties that influence the metal release process, the corrosion process, and hence any concomitant toxicity. It should be underlined that all these characteristics are not only influenced by the particle history, i.e. manufacturing, treatments, storage, transport, and ageing, but also by many experimental parameters.

Particle characteristics Size distribution Particle size

Oxide layer properties Pores and defects

Morphology Agglomeration

Surface atoms /  bulk atoms ratio Shape

Adhesion  and  structure

Bulk and oxide phases

Oxide conductivity Inclusions and  impurities

• Electrochemical passivity • Surface charge

Bulk and surface composition

Figure 5: Important particle characteristics influencing the surface oxide properties and hence metal release, corrosion, and toxicity.

13

Differences between pure metals and alloys (Papers III-VI) As explained in the previous section, surface oxide properties are crucial for any corrosion and also metal release process, and hence for any concomitant toxicological response induced by the metal/alloy of interest. Specially designed alloys such as stainless steels, optimize the properties of the surface oxide, and at the same time other properties such as mechanical properties, compared with their respective pure alloying elements (in this case mainly Fe, Cr, Ni) 48, 55-58. Under the United Nations Globally Harmonized System of Classification and Labeling of Chemicals 59, alloys are explicitly addressed as simple mixtures of metals, with hazard identification and classification based on the intrinsic properties of the individual metal constituents from which they are derived. This simplified view is in the case of an alloy such as stainless steel, highly erroneous. From a metal release perspective, this is evident when comparing released amounts of chromium, iron and nickel from stainless steel grade AISI 316L with corresponding released amounts of metals from the pure alloying metals in different kind of media 8, 17. These studies indicate that stainless steel grade 316L, containing 17.2 wt% chromium and 10.7 wt% nickel actually behaves as a chromium-based material containing 0.02 wt% iron and 0.005 wt% nickel from a metal release perspective, both well below the lowest threshold concentration for hazard classification 19. REACH (Registration, Evaluation, Authorisation and Restriction of Chemicals, 25-26) recognizes alloys as special preparations: “Alloys are preparations under REACH albeit special ones where the properties of the preparation do not always simply match the properties of the components”; “When assessing the risk of the use of one or more substances incorporated into a special preparation (for instance alloys), the way the constituent substances are bonded into the chemical matrix should be taken into account.” 60. Although it is, from a scientific (corrosion science) perspective, evident that the extent of metal released from a pure metal is very different compared with an alloy, no data was available for metal and alloy particles of ironand chromium-based materials, and many alloy particles were hence in danger of erroneous classification based on their bulk compositions. The international ferrochromium and stainless steel industry were early aware of this fact and initiated together with KTH several research projects starting 14

in the early 2000’s to obtain sufficient data for their products in due time to fulfill their REACH deadline - December 2010 - “No data no market”. In Papers III-VI, metal release (and toxicity, Paper VI) from pure metal particles and different ferrochromium alloys and stainless steel (316L) particles has been compared in several different synthetic biological fluids (simulating inhalation and subsequent ingestion, Papers IV and VI, skin and eye contact, Paper V), and fresh water (Paper III). When comparing the iron release from pure iron particles (Fe), iron(II,III)oxide (Fe3O4), and different alloy particles in fresh water (Paper III, Figure 6), it is obvious that i) pure Fe particles release significantly more iron (more than 1000-fold) compared both with stainless steel particles (316L) and with ferrochromium particles (FeCr), or when compared with the thermodynamically stable iron(II,III)oxide, and ii) that there are large differences between different kind of alloys. SP alloy particles, a side product from stainless steel production, release significantly more iron (and also other metals, with exception of chromium) compared to FeCr and 316L particles, designed for a relatively high corrosion resistance. The SP particles have fairly similar composition compared with 316L particles but a completely different particle history (manufacturing process), intended use and application.

15

Figure 6: Released concentration of iron after 24 hours of immersion at pH 6.0 in the OECD transformation/dissolution test (fresh water simulating) of particles of pure Fe, iron(II,III)oxide, and three different alloys (stainless steel 316L, a ferrochromium alloy FeCr, and a side product from stainless steel production SP). The inset graph shows the concentration at a 1000-fold higher magnification.

In contrast, comparing the chromium release from pure metal particles and alloy particles, the alloy particles are generally releasing more chromium compared with the pure metal (Papers III-VI). This is due to the very stable chromium(III)oxide on pure chromium. This behavior is also obvious from metal release data in artificial lysosomal fluid (ALF, pH 4.5), simulating lung conditions at inflammatory conditions, after 168 hours of exposure, for particles of pure Fe, Cr, Ni and 316L (Paper VI, Figure 7). While the nickel and iron particles dissolve nearly completely (a number of 1 corresponds to 100% of dissolution), at least for the fine particles, particles of both Cr and 316L dissolve less than 2% of the total mass. The finest chromium particles, Cr fine, dissolve only to a very low extent, less than 0.1% (mostly iron is released from Cr particles, an impurity in these Cr particles).

16

Figure 7: Released total (Fe+Cr+Ni) amount of metals per amount of particles loaded [µg/µg] after immersion in artificial lysosomal fluid (simulating lung conditions at inflammatory conditions, pH 4.5) for 168 hours, for particles of pure iron (coarse and fine), nickel, chromium (coarse and fine) metal, and stainless steel 316L (coarse and fine).

Results in other artificial biological fluids (Gastric fluid, pH 1.5; artificial sweat, pH 6.5; phosphate buffered saline, pH 7.4; Gamble’s solution, pH 7.4; artificial tear fluid, pH 8.0) emphasize similar results, when pure metal particles and alloy particles are compared (Papers IV-V). Particles of metal oxides, Fe3O4 and Cr2O3, release very low amounts of metals when compared with both particles of pure metals and alloys and when normalized to the surface area (Paper III, V, VI). To summarize,  Particles of pure iron and nickel release significantly (orders of magnitude) more metal compared with alloy particles with high corrosion resistance (stainless steel or ferrochromium alloys).  Particles of pure chromium, with a very stable chromium(III)oxide layer, release significant less metals compared with alloy particles.  Metal release is selective. Chromium is generally the least released metal while iron is released preferentially from particles of stainless steel and ferrochromium.  Depending on the surface oxide stability, the composition, and the manufacturing and particle history, there can be large differences in terms of metal release between different kinds of alloys. 17

Environmental risk assessment – possible measurements and obstacles (Papers I-III) Nowadays, environmental risk assessment is often based on few data and many estimations and extrapolations. Even if many studies were performed to investigate metal release from e.g. outdoor constructions at real conditions, data obtained would not be sufficient enough to make an accurate environmental risk assessment for all products available on the market and for all constructions exposed at different environmental conditions. The reason is the fact that research takes significantly longer time to conduct than it takes for many new products to enter on the global market. Another reason is the fact that the corrosion related metal release research (e.g. comparing runoff rates with corrosion rates) did not really start before the 1990s. There are many important factors that are crucial to understand when generating and interpreting data for environmental risk assessment and to allow data extrapolation and read-across with other materials. Some of these factors are discussed below.

Chemical speciation and the entire interacting system of released metals and species In Paper I (illustrated in Figure 8), the total copper concentration in runoff water from a 2330 m² sized copper roof on a shopping centre in Farsta, a suburb of Stockholm, Sweden, was measured during several individual rain events in 2006 and 2008 after interaction with the internal drainage system (different downspouts made of cast iron and concrete ending up in one vertical 30 m long drainage pipe, and in addition one more sampling point after additional 50 m of concrete pipe). The average copper concentration was approximately 1000 times lower compared with copper released at the immediate release situation from the copper roof (average data for copper runoff, Paper I, 9). A relatively high percentage of copper was determined to be complexed in the runoff water. The copper concentration was even lower than what was measured in rainwater dewatering a nearby parking space. The results clearly illustrate that released copper was complexed and precipitated already during this short transport within the internal drainage system of the shopping centre. It should be underlined that both the environmental fate (previously investigated in e.g. 1, 3-5, 7) and the chemical speciation (bioavailability and 18

ability to form complexes and precipitates) have to be taken into account for accurate risk assessments related e.g. to corrosion-induced metal dispersion from external constructions. Unfortunately, there are still examples of highly erroneous risk assessments and source calculations, since both the chemical speciation and other highly important experimental factors, as discussed later in this section, have not been taken into account or even considered 61.

Figure 8: Comparison of measured total copper (Cu) concentration at the immediate release situation from the copper roof and after interaction with the internal drainage system of a shopping centre (inset graph). The bioavailable copper fraction (labile fraction) is also displayed in comparison with the total copper concentration after interaction with the internal drainage system. Copper concentrations (total and bioavailable) of rain water dewatering a nearby parking space are presented for comparison (inset graph).

In Paper II, environmental aspects from a metal release and corrosion perspective have been investigated for bare zinc sheet (Zn) compared with chromium(III)- (Zn-CrIII) and chromium(VI) (Zn-CrVI) surface treated galvanized steel exposed at un-sheltered outdoor conditions at an urban site (Stockholm, Sweden) and marine site (Brest, France) up to 5 years of exposure. From an environmental point of view, it becomes obvious that the entire system including the release of zinc, chromium(III), and chromium(VI) has to be taken into account, as well as all main influencing factors on the extent of metal release (runoff). A simplified illustration of 19

general observations from these studies is given in Figure 9. For sensitive water organisms, zinc, Cr(III), and Cr(VI) are of similar ecotoxicity, with a predicted no effect concentration (PNEC) of 7.8 µg Zn L-1 62, 4.7 µg Cr(III) L-1, and 3.4 µg Cr(VI) L-1 41. This means that zinc is only about half as toxic compared with Cr(VI) for sensitive water organisms (the difference is several orders of magnitude for humans). The presence of the Zn-Cr(VI) surface treatment reduces the zinc release significantly at the same time as very small amounts of Cr(VI) are released. In all, its presence (as long as intact) reduces any potential ecotoxicity from this runoff water, see Figure 9. However, when considering human exposure e.g. during manufacturing, handling, and transport, the Zn-Cr(VI) may pose adverse effects (not investigated in this study).

Figure 9: Zinc, Cr(III), and Cr(VI) released per area and rainfall quantity during the first 2 days of exposure at the urban site, Stockholm, Sweden, from bare zinc sheet (Zn), chromium(III) surface treated (Zn-Cr(III))-, and from chromium(VI) surface treated (Zn-Cr(VI)) galvanized steel.

As presented in Paper II, based on the chromium speciation measurements, the total system of corrosion and runoff processes was possible to understand. Significantly less zinc was released when Cr(VI) was released (mobile Cr(VI)).Upon transformation of mobile Cr(VI) into Cr(III) or its immobilisation, the extent of corrosion and also zinc runoff increased. The more uniform and thicker chromium(III)oxide surface treatment on galvanized steel, Zn-Cr(III), revealed an improved long term corrosion resistance and resulted in less zinc runoff compared with Zn-Cr(VI) after the first few days of exposure. The chromium speciation was hence crucial 20

for the total runoff process and influenced by several parameters, especially time and exposure site. In Paper III, the OECD transformation/dissolution (T/D) test, a standardized test simulating environmental aquatic exposure, was performed for several iron-, and chromium-based particles. In parallel, different experimental parameters of high importance for the outcome of the test were critically assessed and suggestions for a more reliable and reproducible test given. While there was no complexation of chromium or iron, released in very low amounts, there was a strong precipitation of soluble iron at higher concentrations over time, see Figure 10. Such effects for iron are expected when considering the iron-water chemistry 63-66. However, since the standard test is thought to assess the total released metal fraction, timedependent, solution buffering-dependent, and also solution salt concentration- (varied in the test) dependent effects on measured metal concentrations are in addition necessary to understand and take into account when using data for hazard assessment. By changing parameters, allowed to be changed within the protocol limits, e.g. buffering, time, or salt concentration, generated results can easily be tuned in one direction or another.

Figure 10: Measured iron concentration, mgL-1, released from iron metal particles (100 mg/L) exposed during a 7-day OECD transformation/dissolution test at pH 6.0 (low salt concentration in the medium), and pH 8.0 (low and high salt concentration).

To summarize, chemical speciation, i.e. metal oxidation state, potential complexation, and subsequent precipitation of released metals, has to be taken into account when assessing environmental risks and hazards based on data generated via outdoor exposures or through standardized tests. 21

Experimental parameters of high importance for environmental risk assessment Material and surface conditions As discussed earlier, the surface oxide stability is crucial for any metal release or corrosion process for metals or alloys as massive sheet or as particles. The surface condition of the material of interest should therefore either be thoroughly characterized, or defined, e.g. by a standardized grinding/polishing/cleaning procedure. However, even if standardized, the resulting surface finish can vary significantly 19, 21. When investigating metal or alloy particles, not possible to polish, detailed particle and surface oxide characterization is essential to allow any fair comparison of generated results (Paper III). Beside studies on the surface oxide composition (XPS), bulk composition and morphology of the particles (SEM/EDS), specific surface area (BET), particle size distribution in solution (LD) were conducted for all particles investigated (Papers III-VI, Table). Measurements of surface charge - zeta potential (Paper VI and ongoing), phases and microstructure (EBSD, ongoing AFM studies), surface oxide stability (electrochemical measurements, ongoing), and chemical surface oxide composition (CRM) are currently applied to further generate essential characteristics of the different particles investigated. Such an approach is in our minds essential to allow further understanding of the relationship between particle and surface characteristics, metal release and toxicological effects. Outdoor exposures When outdoor exposure data is used to assess corrosion-induced metal release/runoff to be used within the framework of environmental risk assessments (Papers I-II), several exposure parameters of large importance for the runoff process are important to consider:  Inclination  Orientation  Rain frequency, intensity, amount  Rain composition and pH (site-specific, e.g. containing chlorides at marine sites)  Pollutants  Temperature Other parameters such as surface condition and material, acid cleaning procedure, runoff water collection, and metal analysis etc., are of high 22

importance. All these parameters have been discussed in detail and used for the elaboration of predictive models for copper runoff at urban and rural sites in 11, 13, 67-68. Acid cleaning and handling The risk for sample contamination has to be minimized when very low metal concentrations (ppb or sub-ppb) are to be analyzed (Papers I-VI), or when working with metals, such as zinc, iron, and silicon, that often also exists as contaminants in glassware, ambient laboratory air etc. For non-acidified solutions to be analyzed on metal concentration, polymer vessels are required for collection and storage to avoid adsorption of the metal at the vessel walls. All vessels and used equipment need to be acidcleaned in a proper way, e.g. 24 hours 10% HNO3 and afterwards four times rinsed with ultrapure water to ensure no effects from contamination (Papers I-VI). Very often during environmental risk assessment studies, e.g. ecotoxicological studies, membrane filters are used to separate particles from the solution and to measure the “dissolved fraction” of metals. Ongoing studies indicate though that membrane filters can be highly contaminated and that the acid-cleaning procedure and the metal of investigation influence the measured metal concentration (mostly evident at low metal concentrations). Generated results imply that the “dissolved fraction” is not really measured by analyzing the total metal concentration after membrane filtration (ongoing). Laboratory exposures – OECD transformation/dissolution tests In Paper III, the following parameters were found to significantly influence the metal release process: Already defined by the protocol  Temperature  Time  Particle loading  Dissolved oxygen  Illumination  pH

Needs to be defined in a better way  Dilution  Buffering system  Filtration / solid/liquid separation  Agitation  Solution composition (varies with varying pH)  Particle characterization 23

In all, a general understanding of the total system (material, particles, corrosion, speciation, water chemistry, experimental parameters etc.) is essential to be able to design any reproducible and reliable standard test suitable for any kind of material that should be tested in a comparative way.

24

Human health risk assessment for metal particles – important parameters for in-vitro and in-vivo testing (Papers IV-VI) Countless in-vitro and in-vivo toxicological tests have been performed for a large number of metals, metal solutions, or metal or alloy particles (some selected: 39, 69-89). However, limited efforts have been conducted to assess primary reasons behind these toxicological results, such as individual influencing material- and/or experimental parameters (e.g. storage). For instance, significant differences in cytotoxicity were observed for artificially aged and non-aged particles of a CoCr alloy and particles of stainless steel (316L) 90. Another study has recently indicated that sonication of particle dispersions, a standard procedure to obtain homogeneous dispersions in solution prior to any toxicological test, has a significant influence on the extent of metal release and the toxicity induced52, 54. In Paper VI, it is in addition indicated (more studies ongoing) that the sonication procedure also influences the surface charge (zeta potential) of the particles, an effect that seems to be particle size dependent. It is well established that toxicity is highly particle size dependent 27-28, 52-53. Recent findings have shown that this is only partially related to an increased extent of released metals with decreasing particle size 52. Another aspect that often is forgotten is that the entire particle and material, i.e. the surface oxide properties, also change with decreasing particle size. These findings become obvious when comparing metal release rates (normalized to the surface area) from coarse and fine particles of pure metals in e.g. artificial lysosomal fluid, Figure 11. Pure metal particles, Cr and Fe in this case, revealed a decreasing metal release rate with decreasing particle size, most probably as a result of an increased extent of particle agglomeration for small sized particles and shifts in thermodynamic equilibrium in the closed systems used (when more metal is released (as in the case of the smaller particles), the driving force for further dissolution decreases). The release rate of iron was also decreasing with decreasing particle size for Fe metal particles (the release of Cr is shown in Figure 11 from Fe particles and is due to natural impurities). Similar effects were observed for Cu metal particles of different size exposed in phosphate buffered saline (PBS, pH 7.4) 52. However, for alloy or alloy dust particles (generated in a standardized dustiness test) opposite trends were evident. 25

Smaller-sized particles released more metal per surface area. The protecting surface oxide properties change with decreasing particle size. Recent on-going studies strongly support this hypothesis and indicate that also bulk properties can change with decreasing particle size.

Figure 11: Release rates of chromium, µg cm-2 h-1, from different particles of pure metals (Fe, Cr), chromium(III)oxide, and different alloys and alloy dust particles, after 168 hours of exposure in artificial lysosomal fluid (ALF, pH 4.5), for coarse and fine particles (except Cr2O3 and the dust particles), respectively.

It is therefore essential that any metal or alloy particles to be investigated are thoroughly characterized, that all material, corrosion and metal release processes involved are fully understood, and that the experimental parameters influencing the particle properties are investigated. Any toxicological tests conducted without any particle or material knowledge from different perspectives are of low value. This fact has also recently been noticed by far-sighted toxicologists. 28, 91-94 In Paper VI (see Figure 12), the total concentration of metals released into artificial lysosomal fluid (bioaccessibility), the chemical speciation (only 26

for chromium, important factor for bioavailability), and the toxicological response in different assays (toxicity) have been investigated and correlated, for well characterized iron- and chromium-based alloy particles of two different sizes; stainless steel (AISI 316L), a ferrochromium alloy, and a ferrosiliconchromium alloy. For comparison, particles of dust (generated via a standardized dustiness test to simulate occupational exposure conditions), particles of pure metals, and metal oxides were investigated in parallel. Only the finer sized particles were investigated on toxicity due to their relevance for the inhalation exposure route of interest. Most toxicological responses were not significantly higher or even lower when compared with the control (without particles). One exception was stainless steel (316L) particles, which gave a positive toxicological response in one of three tests conducted (the DNA damage test). It is still unclear why these particles were toxic towards this endpoint and not the other alloy particles (releasing about the same amounts of iron or chromium). However, the small size of the 316L particles could be one explanation (however, even smaller sized particles of other materials were investigated that did not result in any toxicity). Release of nickel could be another explanation, although its extent was relatively low. Recent results indicate that structural differences of the smaller sized particles compared with the larger sized particles could be a reason for the observed toxicity and/or the release of manganese. The smallest sized particles investigated, Cr2O3, did not release any significant amounts of chromium, but were still highly toxic in one of the three tests (hemolysis). This test is sensitive to the surface charge of the particles, a surface property that for Cr2O3 changed from negative values to positive values during the sonication procedure. Surface charge could be another reason for the observed toxicity (DNA damage) of small-sized particles of stainless steel. However, further studies are required (and ongoing) to prove this hypothesis.

27

Paper VI Active (free) chromium ions or complexes in interaction with an organism

Toxicity

low toxicity Fine 316L: DNA damage. Cr2O3: hemolysis

Active (free) chromium ions or complexes in interaction with cells Active (free) chromium ion concentration of a certain oxidation state

Bioavailability

Cr(III) (no Cr(VI)), highly complexed in ALF

Active (free) chromium ions

Total chromium concentration

Bioaccessibility

Coarse particles: < 20µg/L total Cr Fine particles: < 80 µg/L total Cr

Figure 12: Compilation of chromium results on bioaccessibility (metal release in artificial lysosomal fluid (ALF, pH 4.5) for 168 hours), chemical speciation (important for bioavailability), and toxicity (in-vitro) presented in Paper VI on coarse- and fine-sized iron- and chromium-based particles (stainless steel, ferrochromium, ferrosiliconchromium, dust, pure metals, and oxides).

In all, it seems possible but difficult to correlate particle characteristics and metal release with toxicity. More fundamental studies are needed (and partially ongoing) to answer the many questions.

28

Main conclusions Chemical speciation of metals, i.e. the oxidation state and ligands (complexation), is crucial for any toxicity or ecotoxicity considerations. Chemical speciation is influenced by the material characteristics, surface properties, corrosion-induced metal release process, and to a large extent by the environment (UV irradiation, pH, and other species). It is essential to consider chemical speciation when assessing potential environmental risks induced by metal runoff from external buildings and constructions, or when conducting laboratory tests to predict dispersion of metals (Papers I-III). Since the surface properties of metals are crucial for the corrosion and the metal release process, particles that cannot have a defined surface finish have to be thoroughly characterized to enable any fair comparison between metal release or toxicity tests of different particles. The influence of specific particle parameters, and their synergistic effects, on the metal release process and any induced toxicity has to be investigated (Papers III-VI). Particles of alloys behave completely different when compared to their individual metal components. Generally, particles of alloys such as stainless steel or ferrochromium, release significantly less metals and are significantly less acute toxic compared with e.g. nickel metal particles. However, chromium metal particles release the lowest extent of metals due to a stable surface chromium(III)oxide. The extent of metal release depends on parameters such as the type of alloy, prevailing exposure conditions, surface oxide characteristics and stability, manufacturing process and particle history (Papers III-VI). Many experimental parameters that influence results generated in metal release and in-vitro toxicity tests are not yet fully understood. For example, the OECD transformation/dissolution test allows a variety of degrees of freedom of some important parameters, whereas other important parameters are not defined at all. As a consequence, a reproducible and reliable standard test applicable for all different kinds of material is not possible to accomplish (Paper III). Other experimental procedures commonly used for different laboratory tests, such as sonication to disperse particles in solution prior to in-vitro toxicity tests, can influence the particle characteristics and hence the toxicological outcome (Paper VI). Further studies are required to provide a better understanding of the influence of such parameters on the experimental outcome.

29

Acknowledgements/Danksagung I want to thank:  My supervisor Prof. Inger Odnevall Wallinder for great supervising, discussions, and being an incredible nice and invaluable boss.  Cusanuswerk not only for the complete financial support but also for the personal support, the valuable discussions, meetings, and all the nice fellow Ph.D. students; not only did you make me self-confident but also feeling responsible for this world.  For even more financial support and informative meetings: o Paper I: European Copper Institute, Environmental Billion Fund of the city of Stockholm, Sweden, Stockholm Vatten VA AB, Sweden. o Paper II: Nordic Galvanizers Association, Sweden, Rheinzink, Germany, Saferoad, Norway, SSAB, Sweden. o Paper III-V: The International Chromium Development Association, Paris, France. o Paper VI: Swedish Research Council supporting the Stockholm Particle group (SPG), an operative network between three universities in Stockholm, the Karolinska Institutet, the Royal Institute of Technology, and Stockholm University. o Polarography instrument: The Swedish Steel Producers’ Association and the Carl Trygger Foundation. o XPS instrument: Knut and Alice Wallenberg foundation.  All my co-authors, Klara Midander for supervising me initially and working closely afterwards, Johanna Gustafsson, David Lindström, Pia Dromberg, Grant Darrie, Hanna Karlsson, Lennart Möller, and Alfredo de Frutos. Special thanks also to Gunilla Herting for all support.  Dr. Kuria Ndungu (SU, Sweden), Prof. Paul Linhardt (TU Wien, Austria), and Prof. Sannakaisa Virtanen (FAU, Germany) for scientific support.  My future co-authors, Tao Jiang, Maria Lundin, Eva Blomberg, Malin Tornberg, Oskar Karlsson, Peter Szakalos, Jonas Hedberg, Yi Liu, and Jon Brunk.  Our master thesis students, Yi Liu, Claudio Baldizzone, and our summer students Ashkan Rezagholi, Nassim Al Maliki, and Rasmus Karlsson.  My room colleagues for the nicest room in our division.  All my colleagues at our division for a nice atmosphere.  My husband Jonas for invaluable support both at home and at work.  My Swedish family, especially Åsa and Lasse.  My German family: Ihr habt die Saat gesät und mich immer unterstützt. Vielen Dank! 30

References 1. Bahar, B.; Herting, G.; Odnevall Wallinder, I.; Hakkila, K.; Leygraf, C.; Virta, M., The interaction between concrete pavement and corrosion-induced copper runoff from buildings. Environmental Monitoring and Assessment 2008, 140, (1-3), 175-189. 2. Bertling, S.; Degryse, F.; Odnevall Wallinder, I.; Smolders, E.; Leygraf, C., Model studies of corrosion-induced copper runoff fate in soil. Environmental Toxicology and Chemistry 2006, 25, (3), 683-691. 3. Bertling, S.; Odnevall Wallinder, I.; Leygraf, C., In The capacity of limestone to immobilize copper in runoff water. A laboratory investigation, 15th International Corrosion Conference, Granada, Spain, 2002. 4. Bertling, S.; Odnevall Wallinder, I.; Leygraf, C.; Berggren, D., In Immobilization of copper in runoff water from roofing materials by limestone, soil and concrete., 15th International Corrosion Conference, Granada, Spain, 2002. 5. Bertling, S.; Odnevall Wallinder, I.; Kleja, D. B.; Leygraf, C., Long-term corrosioninduced copper runoff from natural and artificial patina and its environmental impact. Environmental Toxicology and Chemistry 2006, 25, (3), 891-898. 6. Bertling, S.; Odnevall Wallinder, I.; Leygraf, C.; Berggren, D., Environmental effects of zinc runoff from roofing materials - A new multidisciplinary approach. Outdoor Atmospheric Corrosion 2002, 1421, 200-215. 7. Bertling, S.; Odnevall Wallinder, I.; Leygraf, C.; Kleja, D. B., Occurrence and fate of corrosion-induced zinc in runoff water from external structures. Science of the Total Environment 2006, 367, (2-3), 908-923. 8. Herting, G.; Odnevall Wallinder, I.; Leygraf, C., A comparison of release rates of Cr, Ni, and Fe from stainless steel alloys and the pure metals exposed to simulated rain events. Journal of The Electrochemical Society 2005, 152, (1), B23-B29. 9. Karlen, C.; Odnevall Wallinder, I.; Heijerick, D.; Leygraf, C., Runoff rates, chemical speciation and bioavailability of copper released from naturally patinated copper. Environmental Pollution 2002, 120, (3), 691-700. 10. Karlen, C.; Odnevall Wallinder, I.; Heijerick, D.; Leygraf, C.; Janssen, C. R., Runoff rates and ecotoxicity of zinc induced by atmospheric corrosion. Science of the Total Environment 2001, 277, (1-3), 169-180. 11. Odnevall Wallinder, I.; Bahar, B.; Leygraf, C.; Tidblad, J., Modelling and mapping of copper runoff for Europe. Journal of Environmental Monitoring 2007, 9, (1), 66-73. 12. Odnevall Wallinder, I.; Bertling, S.; Kleja, D. B.; Leygraf, C., Corrosion-induced release and environmental interaction of chromium, nickel and iron from stainless steel. Water Air and Soil Pollution 2006, 170, (1-4), 17-35. 13. Odnevall Wallinder, I.; Bertling, S.; Zhang, X. Y.; Leygraf, C., Predictive models of copper runoff from external structures. Journal of Environmental Monitoring 2004, 6, (8), 704-712. 14. Odnevall Wallinder, I.; Leygraf, C.; Karlen, C.; Heijerick, D.; Janssen, C. R., Atmospheric corrosion of zinc-based materials: runoff rates, chemical speciation and ecotoxicity effects. Corrosion Science 2001, 43, (5), 809-816. 15. Odnevall Wallinder, I.; Lu, J. S.; Bertling, S.; Leygraf, C., Release rates of chromium and nickel from 304 and 316 stainless steel during urban atmospheric exposure - a combined field and laboratory study. Corrosion Science 2002, 44, (10), 2303-2319. 31

16. Odnevall Wallinder, I.; Bertling, S.; Leygraf, C., Environmental interaction of copper and zinc released from building materials as a result of atmospheric corrosion. Metall 2004, 58, (11), 717-720. 17. Herting, G.; Odnevall Wallinder, I.; Leygraf, C., Metal release rate from AISI 316L stainless steel and pure Fe, Cr and Ni into a synthetic biological medium- a comparison. Journal of Environmental Monitoring 2008, 10, 1092-1098. 18. Herting, G.; Odnevall Wallinder, I.; Leygraf, C., Corrosion-induced release of the main alloying constituents of manganese–chromium stainless steels in different media. Journal of Environmental Monitoring 2008, 10, 1084-1091. 19. Herting, G. Bioaccessibility of stainless steels - Importance of bulk and surface features. Doctoral Thesis, Royal Institute of Technology (KTH), Stockholm, Sweden, 2008. 20. Herting, G.; Lindström, D.; Odnevall Wallinder, I.; Leygraf, C., Multi-analytical investigation of stainless steel grade AISI 420 in simulated food contact. Journal of Food Engineering 2009, 93, (1), 23-31. 21. Herting, G.; Odnevall Wallinder, I.; Leygraf, C., Factors that influence the release of metals from stainless steels exposed to physiological media. Corrosion Science 2006, 48, (8), 2120-2132. 22. Herting, G.; Odnevall Wallinder, I.; Leygraf, C., Metal release from various grades of stainless steel exposed to synthetic body fluids. Corrosion Science, Progress in Corrosion Research in Commemoration of Centenary of Birth of Professor Go Okamoto, International Symposium on "Progress in Corrosion Research" 2007, 49, (1), 103-111. 23. Herting, G.; Odnevall Wallinder, I.; Leygraf, C., Corrosion-induced release of chromium and iron from ferritic stainless steel grade AISI 430 in simulated food contact. Journal of Food Engineering 2008, 87, (2), 291-300. 24. Midander, K.; Pan, J.; Leygraf, C., Elaboration of a test method for the study of metal release from stainless steel particles in artificial biological media. Corrosion Science 2006, 48, (9), 2855-2866. 25. EU REACH in brief; European Commission Environment Directorate General: 2007. 26. EC, Regulation (EC) No 1907/2006 of the European Parliament and of the Council of 18 December 2006 concerning the Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH). In Council, E., Ed. Official Journal of the European Union, 2006; pp L396:L136/133-L136/280. 27. Mossman, B.; Borm, P.; Castranova, V.; Costa, D.; Donaldson, K.; Kleeberger, S., Mechanisms of action of inhaled fibers, particles and nanoparticles in lung and cardiovascular diseases. Particle and Fibre Toxicology 2007, 4, (1), 4. 28. Oberdörster, G.; Oberdörster, E.; Oberdörster, J., Nanotoxicology: An emerging discipline evolving from studies of ultrafine particles. Environmental Health Perspectives 2005, 113, (7), 823-839. 29. Reeder, R. J.; Schoonen, M. A. A.; Lanzirotti, A., Metal Speciation and Its Role in Bioaccessibility and Bioavailability. Reviews in Mineralogy and Geochemistry 2006, 64, (1), 59-113. 30. Landner, L.; Reuther, R., Metals in Society and in the Environment. Springer Netherlands: 2004; Vol. 8. 31. Merian, E., Metals and Their compounds in the environment - Occurrence, Analysis and Biological Relevance. VCH: Weinheim, 1991. 32

32. Landner, L.; Lindeström, L., Copper in Society and in the Environment. 2nd ed.; Swedish Environmental Research Group (MFG): Västerås, Sweden, 1999. 33. Tessier, A.; Turner, D. R., Metal speciation and bioavailability in aquatic systems. John Wiley & Sons Ltd.: Buffins Lane, England, 1995; Vol. 3. 34. Worms, I.; Simon, D. F.; Hassler, C. S.; Wilkinson, K. J., Bioavailability of trace metals to aquatic microorganisms: importance of chemical, biological and physical processes on biouptake. Biochimie 2006, 88, (11), 1721-1731. 35. Plette, A. C. C.; Nederlof, M. M.; Temminghoff, E. J. M.; Riemsdijk, W. H. v., Bioavailability of heavy metals in terrestrial and aquatic systems: A quantitative approach. Environmental Toxicology and Chemistry 1999, 18, (9), 1882-1890. 36. Scheinberg, H., chapter II.9 Copper. Merian, E., Metals and Their compounds in the environment - Occurrence, Analysis and Biological Relevance. VCH: Weinheim, 1991; p 894-908. 37. Cohen, M. D.; Kargacin, B.; Klein, C. B.; Costa, M., Mechanisms of Chromium Carcinogenicity and Toxicity. Critical Reviews in Toxicology 1993, 23, (3), 255 - 281. 38. Costa, M., Toxicity and Carcinogenicity of Cr(VI) in Animal Models and Humans. Critical Reviews in Toxicology 1997, 27, (5), 431 - 442. 39. Hansen, M. B.; Rydin, S.; Menné, T.; Johansen, J. D., Quantitative aspects of contact allergy to chromium and exposure to chrome-tanned leather. Contact Dermatitis 2002, 47, (3), 127-134. 40. Vignati, D. A. L.; Dominik, J.; Beye, M. L.; Pettine, M.; Ferrari, B. J. D., Chromium(VI) is more toxic than chromium(III) to freshwater algae: A paradigm to revise? Ecotoxicology and Environmental Safety 2010, 73(5), 743-749. 41. CSTEE Opinion on the results of the Risk Assessment of: Chromium trioxide (CAS No. 1333-82-0), Sodium chromate (CAS No. 7775-11-3), Sodium dichromate (CAS No. 10588-01-9), Ammonium dichromate (CAS No. 7789-09-5), Potassium dichromate (CAS No. 7778-50-9); Scientific Committee on Toxicity, Ecotoxicity and the Environment: 2003. 42. Gauglhofer, J.; Bianchi, V., chapter II.7 Chromium. In: Merian, E., Metals and Their compounds in the environment - Occurrence, Analysis and Biological Relevance. VCH: Weinheim, 1991; p 853-878. 43. Rollinson, C. L., The Chemistry of CHROMIUM, MOLYBDENUM and TUNGSTEN. Robert Maxwell, M.C.: Oxford, UK, 1975. 44. Allison, J. D.; Brown, D. S., MINTEQA2/PRODEFA2, A Geochemical Assessment Model for Environmental Systems: Version 3.0 User's Manual. 1991. 45. Bobrowski, A.; Królicka, A.; Zarebski, J., Characteristics of Voltammetric Determination and Speciation of Chromium - A Review. Electroanalysis 2009, 21, (13), 1449-1458. 46. Jugade, R.; Joshi, A. P., Highly Sensitive Adsorptive Stripping Voltammetric Method for the Ultra-trace Determination of Chromium(VI). Analytical Sciences 2006, 22, (4), 571-574. 47. Scholz, F.; Lange, B.; Draheim, M.; Pelzer, J., The catalytic adsorptive stripping voltammetric determination of chromium with DTPA and nitrate. Fresenius' Journal of Analytical Chemistry 1990, 338, (5), 627-629. 48. Fehlner, F. P., Michael, J.G., chapter 4: Thin Oxide Film Formation on Metals. In: Marcus, D., Oudar, J. Corrosion Mechanisms in Theory and Practice. Marcel Dekker, Inc.: Paris, France, 1995. 33

49. Leygraf, C., Atmospheric corrosion. In: Marcus, D., Oudar, J. Corrosion Mechanisms in Theory and Practice. 2nd ed.; Marcel Dekker Inc.: Paris, France, 2002; p 529-562. 50. Ullmann, Y. Elaboration of a metal release test for massive metal sheet - effects of different parameters on the copper release rate. Studienarbeit, KTH, Stockholm, 2008. 51. Midander, K.; Pan, J.; Odnevall Wallinder, I.; Leygraf, C., Metal release from stainless steel particles in vitro—influence of particle size. Journal of Environmental Monitoring 2007, 9, 74-81. 52. Midander, K.; Cronholm, P.; Karlsson, H. L.; Elihn, K.; Möller, L.; Leygraf, C.; Odnevall Wallinder, I., Surface Characteristics, Copper Release, and Toxicity of Nanoand Micrometer-Sized Copper and Copper(II) Oxide Particles: A Cross-Disciplinary Study. Small 2009, 5, (3), 389-399. 53. Roduner, E., Size matters: Why nanomaterials are different. Chemical Society Reviews 2006, 35, 583-592. 54. Cronholm, P.; Midander, K.; Karlsson, H. L.; Elihn, K.; Odnevall Wallinder, I.; Möller, L., Effect of sonication and serum proteins on copper release from copper nanoparticles and the toxicity towards lung epithelial cells Under revision for Nanotoxicology 2010. 55. Chandler, K. A., Hudson, J.C., 3.1 Iron and Steel. In: Shreir, L.L., Jarman, R.A., Burstein, G.T., Corrosion Metal/Environment Reactions. 3rd ed.; Butterworth Heinemann: 1994; Vol. 1. 56. MacDougall, B., Graham, M.J., chapter 5: Growth and Stability of Passive Films. In: Marcus, D., Oudar, J. Corrosion Mechanisms in Theory and Practice. Marcel Dekker, Inc.: Paris, France, 1995. 57. Sedriks, A. J., Corrosion of Stainless Steels. 2nd ed.; John Wiley & Sons, Inc.: New York, 1996. 58. Shreir, L. L., 1.4 Corrosion in Aqueous Solutions. In: Shreir, L.L., Jarman, R.A., Burstein, G.T., Corrosion Metal/Environment Reactions. 3rd ed.; Butterworth Heinemann: 1994; Vol. 1. 59. UN, The Globally Harmonized System of Classification and Labelling of Chemicals, ST/SG/AC, 10/30/Rev. 1. In Nations, U., Ed. 2005. 60. EC, Dangerous Substances Directive 67/548/EEC Annex I. In Council, E., Ed. 2006. 61. Cui, Q. Tracing Copper from society to the aquatic environment - Model development and case studies in Stockholm. Royal Institute of Technology (KTH), Stockholm, Sweden, 2010. 62. Bodar, C. W. M.; Pronk, M. E. J.; Sijm, D. T. H. M., The european union risk assessment on zinc and zinc compounds: The process and the facts. Integrated Environmental Assessment and Management 2005, 1, (4), 301-319. 63. Keddam, M., Anodic Dissolution. In: Marcus, D., Oudar, J. Corrosion Mechanisms in Theory and Practice. Marcel Dekker, Inc.: New York, 1995; Vol. 1. 64. Liu, X.; Millero, F. J., The solubility of iron in seawater. Marine Chemistry 2002, 77, (1), 43-54. 65. Shayegani, M.; Ghorbani, M.; Afshar, A.; Rahmaniyan, M., Modelling of carbon dioxide corrosion of steel with iron carbonate precipitation. Corrosion Engineering, Science and Technology 2009, 44, (2), 128-136(9). 66. Skeaff, J. M. H.; David, J.; King, P., A new Approach to the Hazard Classification of Alloys Based on Transformation/Dissolution. Integrated Environmental Assessment and Management 2008, 4, (no. 1), 75-93. 34

67. He, W.; Odnevall Wallinder, I.; Leygraf, C., A laboratory study of copper and zinc runoff during first flush and steady-state conditions. Corrosion Science 2001, 43, (1), 127-146. 68. Odnevall Wallinder, I.; Verbiest, P.; He, W.; Leygraf, C., The influence of patina age and pollutant levels on the runoff rate of zinc from roofing materials. Corrosion Science 1998, 40, (11), 1977-1982. 69. Antonini, J. M.; Lewis, A. B.; Roberts, J. R.; Whaley, D. A., Pulmonary effects of welding fumes: Review of worker and experimental animal studies. American Journal of Industrial Medicine 2003, 43, (4), 350-360. 70. Apopa, P.; Qian, Y.; Shao, R.; Guo, N.; Schwegler-Berry, D.; Pacurari, M.; Porter, D.; Shi, X.; Vallyathan, V.; Castranova, V.; Flynn, D., Iron oxide nanoparticles induce human microvascular endothelial cell permeability through reactive oxygen species production and microtubule remodeling. Particle and Fibre Toxicology 2009, 6, (1), 1. 71. Baker, R. S. U.; Arlauskas, A.; Tandon, R. K.; Crisp, P. T.; Ellis, J., Toxic and genotoxic action of electric-arc welding fumes on cultured mammalian cells. Journal of Applied Toxicology 1986, 6, (5), 357-362. 72. Botta, C.; Iarmarcovai, G.; Chaspoul, F.; Sari-Minodier, I.; Pompili, J.; Orsière, T.; Bergé-Lefranc, J.-L.; Botta, A.; Gallice, P.; Méo, M. D., Assessment of occupational exposure to welding fumes by inductively coupled plasma-mass spectroscopy and by the alkaline Comet assay. Environmental and Molecular Mutagenesis 2006, 47, (4), 284-295. 73. Derelanko, M.; Rinehart, W.; Hilaski, R.; Thompson, R.; Loser, E., Thirteen-week subchronic rat inhalation toxicity study with a recovery phase of trivalent chromium compounds, chromic oxide, and basic chromium sulfate. Toxicol. Sci. 1999, 52, (2), 278-288. 74. Deschamps, F.; Moulin, J. J.; Wild, P.; Labriffe, H.; Haguenoer, J. M., Mortality study among workers producing chromate pigments in France. International Archives of Occupational and Environmental Health 1995, 67, (3), 147-152. 75. Friess, S. L., Carcinogenic risk assessment criteria associated with inhalation of airborne particulates containing chromium(VI/III). Science of The Total Environment, The Chromium Paradox in Modern Life 1989, 86, (1-2), 109-112. 76. Flint, G. N.; Carter, S. V.; Fairman, B., Skin allergy from exposure to alloys of chromium. Contact Dermatitis 1998, 39, (6), 315-316. 77. Hanawa, T.; Kaga, M.; Itoh, Y.; Echizenya, T.; Oguchi, H.; Ota, M., Cytotoxicities of oxides, phosphates and sulphides of metals. Biomaterials 1992, 13, (1), 20-24. 78. Hansen, K.; Stern, R. M., Welding fumes and chromium compounds in cell transformation assays. Journal of Applied Toxicology 1985, 5, (5), 306-314. 79. Huvinen, M.; Makitie, A.; Jarventaus, H.; Wolff, H.; Stjernvall, T.; Hovi, A.; Hirvonen, A.; Ranta, R.; Nurminen, M.; Norppa, H., Nasal cell micronuclei, cytology and clinical symptoms in stainless steel production workers exposed to chromium. Mutagenesis 2002, 17, (5), 425-429. 80. Huvinen, M.; Oksanen, L.; Kalliomäki, K.; Kalliomäki, P.-L.; Moilanen, M., Estimation of individual dust exposure by magnetopneumography in stainless steel production. Science of The Total Environment, Biological Monitoring in Occupational and Environmental Health 1997, 199, (1-2), 133-139. 35

81. Hägerström, H., Paulsson, M., Edsman K, Evaluation of mucoadhesion for two polyelectrolyte gels in simulated physiological conditions using a rheological method. European Journal of Pharmaceutical Sciences 2000, 9, (301-309). 82. ICDA Health Risk Assessment Report for Metallic Chromium and Trivalent Chromium; International Chromium Development Association: 2006. 83. Karlsson, H. L.; Cronholm, P.; Gustafsson, J.; Möller, L., Copper Oxide Nanoparticles Are Highly Toxic: A Comparison between Metal Oxide Nanoparticles and Carbon Nanotubes. Chemical Research in Toxicology 2008, 21, (9), 1726-1732. 84. Karlsson, H. L.; Gustafsson, J.; Cronholm, P.; Möller, L., Size-dependent toxicity of metal oxide particles--A comparison between nano- and micrometer size. Toxicology Letters 2009, 188, (2), 112-118. 85. Karlsson, H. L.; Holgersson, Å. S.; Möller, L., Mechanisms Related to the Genotoxicity of Particles in the Subway and from Other Sources. Chemical Research in Toxicology 2008, 21, (3), 726-731. 86. Karlsson, H. L.; Nilsson, L.; Möller, L., Subway Particles Are More Genotoxic than Street Particles and Induce Oxidative Stress in Cultured Human Lung Cells. Chemical Research in Toxicology 2004, 18, (1), 19-23. 87. Lanone, S.; Rogerieux, F.; Geys, J.; Dupont, A.; Maillot-Marechal, E.; Boczkowski, J.; Lacroix, G.; Hoet, P., Comparative toxicity of 24 manufactured nanoparticles in human alveolar epithelial and macrophage cell lines. Particle and Fibre Toxicology 2009, 6, (1), 14. 88. Merritt, K.; Brown, S. A., Release of hexavalent chromium from corrosion of stainless steel and cobalt - chromium alloys. Journal of Biomedical Materials Research 1995, 29, (5), 627-633. 89. Stern, R. M.; Pigott, G. H.; Abraham, J. L., Fibrogenic potential of welding fumes. Journal of Applied Toxicology 1983, 3, (1), 18-30. 90. Haynes, D. R.; Crotti, T. N.; Haywood, M. R., Corrosion of and changes in biological effects of cobalt chrome alloy and 316L stainless steel prosthetic particles with age. Journal of Biomedical Materials Research 2000, 49, (2), 167-175. 91. Powers, K. W.; Brown, S. C.; Krishna, V. B.; Wasdo, S. C.; Moudgil, B. M.; Roberts, S. M., Research Strategies for Safety Evaluation of Nanomaterials. Part VI. Characterization of Nanoscale Particles for Toxicological Evaluation. Toxicol. Sci. 2006, 90, (2), 296-303. 92. Powers, K. W.; Palazuelos, M.; Moudgil, B. M.; Roberts, S. M., Characterization of the size, shape, and state of dispersion of nanoparticles for toxicological studies. Nanotoxicology 2007, 1, (1), 42 - 51. 93. Stone, V.; Nowack, B.; Baun, A.; van den Brink, N.; von der Kammer, F.; Dusinska, M.; Handy, R.; Hankin, S.; Hassellöv, M.; Joner, E.; Fernandes, T. F., Nanomaterials for environmental studies: Classification, reference material issues, and strategies for physico-chemical characterisation. Science of the Total Environment 2010, 408, (7), 1745-1754. 94. Warheit, D. B., How Meaningful are the Results of Nanotoxicity Studies in the Absence of Adequate Material Characterization? Toxicol. Sci. 2008, 101, (2), 183-185.

36