Surface reactivity, stability, and mobility of metal nanoparticles in aqueous solutions – Influence of natural organic matter and implications on particle dispersion preparation

Sulena Pradhan

Licentiate thesis

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

Stockholm, 2017 This licentiate thesis will, with the permission of Kungliga Tekniska Högskolan, Stockholm, be presented at a licentiate seminar on September 21, 2017, at 10 a.m. in E 31 Lindstedtsvägen 3, KTH Campus, Stockholm, Sweden.

Surface reactivity, stability, and mobility of metal nanoparticles in aqueous solutions

– Influence of natural organic matter and implications on particle dispersion preparation

TRITA-CHE Report 2017:33 ISSN 1654-1081 ISBN 1654 - 1081

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

The following item is printed with permission:

Paper I: Journal of Nanoparticle Research, open access

Denna uppsats är skyddad enligt upphovsrättslagen. Alla rättigheter förbehålls.

Copyright © Sulena Pradhan 2017. All rights reserved. No part of this thesis may be reproduced by any means without permission from the author. LIST OF SCIENTIFIC PUBLICATIONS

Publication included in this thesis Paper I. Pradhan, Sulena; Hedberg, Jonas; Blomberg, Eva; Wold, Susanna; Odnevall Wallinder, Inger. Effect of sonication on particle dispersion, administered dose and metal release of non-functionalized, non-inert metal nanoparticles, Journal of Nanoparticle Research, 2016, 18, 285 Paper II. Pradhan, Sulena; Hedberg, Jonas; Rosenqvist, Jörgen; Jonsson, Caroline M. ; Wold, Susanna; Blomberg Eva; Odnevall Wallinder, Inger. Influence of humic acid and dihydroxy benzoic acid on the agglomeration, sedimentation and dissolution of copper, manganese, aluminum and silica nanoparticles. Manuscript

Author’s contribution Paper I. Experimental work, data analysis and writing the first draft of the publication Paper II. Major experimental work except for ATR–FTIR, data analysis except for ATR–FTIR and silica NP characterization and writing the first draft of the publication

Other publications not within the framework of this thesis Hedberg, Yolanda; Pradhan, Sulena; Cappellini, Francesca; Karlsson, Maria-Elisa; Blomberg, Eva; Karlsson, Hanna; Odnevall Wallinder, Inger; Hedberg, Jonas. Electrochemical surface oxide characteristics of metal nanoparticles (Mn, Cu and Al) and the relation to toxicity, Electrochimica Acta, 2016, 212, 360-371

Hedberg, Yolanda ; Pettersson, Maria ; Pradhan, Sulena ; Odnevall Wallinder, Inger ; Rutland, Mark ; Persson, Cecilia. Can cobalt (II) and chromium (III) ions released from joint prostheses influence the friction coefficient? ACS Biomaterials Science & Engineering, 2016, 1(8), 617- 620.

Pradhan, Sulena; Hedberg, Jonas; Wold, Susanna; Blomberg, Eva; Odnevall Wallinder, Inger. Implementation of the standard operation procedure used and stipulated in the NanoReg EU project on preparation of nanoparticle dispersion for further analysis., Technical Report, Mistra Environmental Nanosafety, Sept. 2015

Ferraz, Natalia; Strömme, Maria; Fellström, Bengt; Pradhan, Sulena; Nyholm Leif; Mihranyan, Albert. In vitro and in vivo toxicity of rinsed and aged nanocellulose-polypyrrole Composites, Journal of Biomedical Materials Research Part A, 2012, 100, 2128-2138

Johansson, Erik; Pradhan, Sulena; Wang, Engang; Unger, Eva; Hagfeldt, Anders; Mats R, Efficient infiltration of low molecular weight polymer in nanoporous TiO2. Chemical Physics Letters, 2010, 502, 225-230 CONFERENCES AND WORKSHOPS

1. Poster - Can organic degradation products stabilize metal nanoparticles in solution? Sulena Pradhan, Jonas Hedberg, Eva Blomberg, Inger Odnevall Wallinder, Susanna Wold nd 2 Nanosafety forum for young scientist, September 15-16, 2016, Visby, Sweden.

2. Mistra meetings – November 11-12, 2015, Gothenburg, Sweden. November 24-25, 2016, Stockholm, Sweden.

3. KTH PhD day - Pitch presentation titled “When small meets smaller” th September 5 , 2016, Lidingö, Sweden. ABSTRACT

The growing development of nanotechnology has resulted in an increased use of nanoparticles (NPs) in various applications ranging from medicine, military, to daily consumer products. There is a concern that NPs can be dispersed into the environment in various ways, for example to air and water during manufacture, use, incineration or recycling of products and thus pose a risk to health and the environment. Risk assessments of NPs are hence necessary. One property of NPs, which may make them very useful and at the same time potentially harmful, is their small size (in nanometer range) and hence high surface area per NP mass.

This study forms part of the National Mistra Environmental Nanosafety Research Program. The program provides an interdisciplinary platform for researchers from e.g. nanoscience, medicine, chemistry, material science, life cycle analysis, and social science. Specific aspects of this program involve characterization of NPs in different environmental settings, toxicity studies of aquatic organisms, integrated risk assessment of NPs, and societal dimensions of nanosafety. The contribution of this thesis within the program includes studies of stability and mobility of metal NPs and their extent of transformation/dissolution upon environmental interaction. Environmental risk assessments of NPs require a detailed understanding of how they change in terms of physical and chemical properties (charge, size, and surface oxide composition), important aspects for their stability, mobility, and reactivity in the environment. Generated data is highly relevant for the other activities of the Mistra Environmental Nanosafety program, e.g. to gain an improved understanding and design of particle dispersions and ecotoxicity studies, as any environmental interaction will result in the transformation/dissolution of the NPs and change the surface chemistry (e.g. adsorption of natural organic matter, changes in surface oxide properties), aspects that largely influence their speciation and potential toxicity.

Common sonication protocols exist to prepare particle dispersions for different in vitro studies. The influence of key parameters stipulated by these protocols on the particle size, transformation/dissolution, and extent of sedimentation was investigated for bare metal NPs. Improved knowledge on these aspects is crucial for design and interpretation of results of NP- related investigations. Reactive metal NPs such as Cu and Mn NPs started to dissolve and release metals already during the probe sonication step of the stock solution, and that the presence of bovine serum albumin (often added as a stabilizing agent) enhanced this process. Even though prolonged sonication time i.e. increased delivered acoustic energy, reduced the size of formed agglomerates, sedimentation was still significant. As a consequence, administered doses from pipetted stock solutions were significantly lower (30-70%) than the nominal doses. The main reason behind the significant extent of agglomeration, with concomitant sedimentation, is related to the strong van der Waals forces prevailing between metal NPs. It is hence essential to determine the administrated dose of metallic NPs in e.g. nanotoxicological testing.

Interactions between metallic NPs and natural organic matter (NOM) were studied in terms of stability, mobility and metal dissolution in order to mimic a potential exposure scenario. NOM was represented by humic acid (HA), a main component of organic matter in the environment, and by dihydroxybenzoic acid (DHBA), a small degradation product of NOM. Sedimentation of the Cu, and the Al NPs were slower in the presence of NOM in freshwater compared with freshwater only, whereas the effect of NOM was small for the Mn NPs. Stabilization was related to surface adsorption of NOM, which increased the steric repulsion between the particles, and in the case of HA also increased the magnitude of the zeta potential (resulting in increased electrostatic repulsion). Slight initial increase in particle stability was observed in freshwater containing DHBA, but after 24 h, sedimentation of the NPs was comparable to the conditions in freshwater only. The presence of HA (at a concentration of 20 mg/L) was found to stabilize the NPs in freshwater for more than 24 h. However, both the lower and higher HA concentration (2 and 40 mg/L) resulted in agglomeration of the Cu and Al NPs already within a few hours. Mn NPs were more stable in terms of sedimentation in freshwater at all three humic acid concentrations. This concludes that the concentration and type of NOM largely influence the stability of the studied metal NPs in solution. In contrast, SiO2 NPs were not influenced by the presence of NOM in terms of stability, most probably predominantly related to smaller attractive van der Waals forces and larger electrostatic repulsion (due to higher surface charge) compared with the metal NPs.

Metal release from the Cu and Al NPs was enhanced in the presence of NOM, whereas no significant influence was observed for the Mn NPs. All metal NPs were dissolved relatively fast; 10% or more of the particle mass was dissolved within 24 h. Speciation predictions revealed rapid complexation between released Cu and Al in solution and NOM, reducing the bioavailability, whereas less complexation was evident for released Mn (as ions). In all, rapid agglomeration and sedimentation imply that any risks associated with the environmental dispersion of these metal NPs will be limited to the vicinity of their source. Mn NPs, having lower sedimentation rates than the Cu and Al NPs, and lack of solution complexation of released ions will likely have a relatively higher probability to be mobile and transported to other aquatic settings. SAMMANFATTNING

Nanoteknologins utveckling har resulterat i en ökad användning av nanopartiklar (NP) i olika applikationer, t.ex. i läkemedel, militär utrustning och dagliga konsumentprodukter. Det finns dock en oro för att NP kan spridas till miljön på olika sätt, t.ex. till luft och vatten vid tillverkning, användning, förbränning eller återvinning av produkter och därmed utgöra en miljö- och hälsorisk. Riskbedömningar av NP är därför nödvändiga att genomföra. En egenskap hos NP, som både gör dem mycket användbara, och samtidigt potentiellt skadliga, är deras storlek (i nanometer-nivå) och därmed stora ytarea.

Denna studie utgör en del av det nationella forskningsprogrammet Mistra Environmental Nanosafety, en tvärvetenskaplig plattform för forskare inom nanovetenskap, miljövetenskap, medicin, ykemi, materialvetenskap, livscykelanalys och samhällsvetenskap. Specifika aspekter som studeras i detta program omfattar bl.a. karakterisering av NP och hur deras egenskaper förändras i olika vattenmiljöer, toxicitetsstudier av vattenlevande organismer, integrerad riskbedömning av NPs och samhälleliga dimensioner av nanosäkerhet. Denna avhandling bidrar till forskningen genom ingående studier av stabiliteten, mobiliteten och upplösningshastigheten av metalliska NP i kontakt med olika vattensystem. Miljöriskbedömningar av NP kräver en detaljerad förståelse av förändring av deras fysiskaliska och kemiska egenskaper (t.ex. laddning, storlek, ytoxid-sammansättning), vilka påverkar deras stabilitet, rörlighet och reaktivitet i miljön. Erhållna forskningsdata används även inom ramen för övriga verksamheter i forskningsprogrammet, t.ex. för en bättre förståelse och design av de ekotoxikologiska undersökningarna, eftersom växelverkan med miljön kan resultera i en total eller delvis upplösning av NPs samt att förändringar i NPs ytkemi (t.ex. genom adsorption av naturligt organiskt material, förändringar i ytoxidens egenskaper) i högsta grad påverkar deras reaktivitet och potentiellt toxiska egenskaper.

Inom nanotoxikologin finns olika protokoll att följa för att förbereda stamlösningar av NP genom sonikering. Kunskap saknas dock till stor del om hur dessa riktlinjer påverkar storlek, omvandling/upplösning, och grad av agglomeration och sedimentation hos reaktiva metalliska NP. Dessa aspekter har undersökts i denna avhandling då de är av avgörande betydelse för experimentell design och tolkning av resultat av NP-relaterade undersökningar. Resultaten visar på snabb partikelagglomeration, sedimentation och frisättning av metaller från de reaktiva partiklarna redan under prepareringen av stamlösningarna genom sonikering (ultraljudsbehandling). Andelen frisatta metaller ökade dessutom i omfattning genom tillsats av bovine serum albumin, ett protein som ofta tillsätts för att stabilisera partiklar i lösning. Även om en förlängd ultraljudsbehandling, dvs. en större tillförd mängd akustisk energi, minskade storleken på bildade agglomerat var sedimentationen fortfarande betydande. Dessa snabba processer medför att de partikeldoser som från stamlösningen sedan tillförs t.ex. till celler i nanotoxikologiska studier kan vara betydligt lägre (30-70% för Cu, Al, Mn NP i denna studie) än de nominella doserna. Den främsta orsaken bakom den betydande omfattningen av agglomeration och efterföljande sedimentation är relaterad till starka van der Waals-krafter som råder mellan metalliska NPs. Resultaten visar tydligt att mätningar av tillförda doser måste kvantifieras för olika system.

Växelverkan mellan metalliska NP och naturligt organiskt material (NOM) studerades med avseende på stabilitet, rörlighet och metallfrisättning för att efterlikna ett potentiellt exponeringsscenario där partiklarna kommer i kontakt med t.ex. sötvatten. NOM representerades i denna studie av humussyra (HA), en huvudkomponent i organiskt material i miljön, samt av dihydroxybenzosyra (DHBA), som utgör små nedbrytningsprodukter av NOM. Sedimentationen av agglomerat av Cu och Al NP var långsammare i närvaro av NOM jämfört med syntetiskt ytvatten utan NOM. Den ökade partikelstabiliteten var för dessa partiklar relaterad till ytadsorptionen av NOM, vilket ökade de sterisk krafterna mellan partiklarna och också partiklarnas ytladdning. Motsvarande effekt för Mn NPs var dock liten trots NOM- adsorption. En liten initial ökning av partiklarnas stabilitet observerades i ytvatten som innehöll DHBA, men efter 24 h var sedimentation jämförbar med förhållandena i ytvatten utan NOM. HA i en koncentration av 20 mg/L, realistisk för verkliga förhållanden, stabiliserade partiklarna längre än 24 h. En både lägre och högre HA koncentration (2 och 40 mg/L) resulterade i agglomeration av både Cu NP och Al NP redan inom några timmar medan Mn NP var mer stabila för alla koncentrationer. Resultaten visar att både koncentrationen och typen av NOM påverkar stabiliteten och mobiliteten av de metalliska NP i lösning. NP av SiO2 påverkades däremot inte alls av närvaron av NOM med avseende på stabilitet, troligen huvudsakligen beroende på lägre attraktiva van der Waals krafter och större elektrostatisk repulsion (på grund av högre ytladdning) jämfört med de metalliska partiklarna.

Andelen frisatta metaller från Cu och Al NP ökade i närvaro av NOM, medan ingen signifikant effekt observerades för Mn NP. Samtliga NP upplöstes relativt snabbt, 10% eller mer av partiklarnas massa inom 24 h. Beräkningar av specieringen av mängden frisatta metaller i lösning visade på snabb komplexering mellan frisatt Cu och Al i ytvatten och NOM, vilket pekar på låg biotillgänglighet, medan en lägre grad av komplexering med NOM predikterades för frisläppt Mn. På grund av snabb agglomeration och sedimentation av de metalliska NPs undersökta i denna studie förväntas eventuella risker i samband en spridningen av dessa metalliska NPs begränsas till källans närhet. Mn NP, med lägre sedimentationshastighet än Cu och Al NP, samt brist på komplexering av frisatta joner, har troligen en relativt högre sannolikhet att vara mobila och kunna transporteras i miljön. CONTENTS

1. AIM OF THIS WORK ...... 1

1.1 Motivation ...... 1

1.2 Aim ...... 2

2. INTRODUCTION ...... 2

2.1 Nanoparticles and nanomaterials...... 2

2.2 Environmental concern ...... 2

2.3 Risk assessment of NPs ...... 3

3. BACKGROUND ...... 4

3.1 Choice of NPs ...... 4

3.2 Relevant NPs characteristics from an environmental perceptive ...... 5

3.3 Selected phenomena influencing NP fate in the environment ...... 6

3.3.1 Colloidal stability ...... 6

3.3.2 Metal release/ dissolution ...... 7

4. EXPERIMENTAL APPROACH ...... 8

4.1 Sample preparation ...... 8

4.2 Characterizations of NPs ...... 9

4.2.1 Size distribution ...... 9

4.2.2 Zeta potential of particles in dispersion ...... 9

4.2.3 Dissolution/metal release ...... 11

4.2.4 Adsorption studies of organic matter ...... 11

5. KEY FINDINGS (Papers I and II)...... 12

5.1 Sonication of NP suspensions largely influence the particle size and extent of metal release (Paper I) ...... 12

5.2 Rapid sedimentation of NPs in the stock solution results in significantly lower administered doses than the nominal doses of sonicated metal NP dispersions – an effect of strong attractive van der Waals forces (Paper I) ...... 15

5.3 Metals are released from Cu, Al and Mn NPs already during the sonication step and the presence of biomolecules can enhance this process (Paper I) ...... 16

5.4 The extent of particle agglomeration of the Cu, Al and Mn NPs depends on the type of surface adsorbed NOM (Paper II) ...... 17 5.5 NOM delayed the sedimentation of the Cu and Al NPs, in particular pronounced in the presence of HA, whereas no significant effect was observed for the Mn NPs (Paper II). ... 19

5.6 The release of metals from the Cu, and Al NPs was enhanced in the presence of NOM, whereas no significant effect was observed for the Mn NPs despite NOM adsorption. Released Cu and Al readily formed strong NOM-complexes of low bioavailability in solution whereas released Mn remained as free ions (Paper II)...... 20

6. MAIN MESSAGES ...... 22

7. FUTURE WORK ...... 23

8. ACKNOWLEDGMENTS ...... 24

9. REFERENCES ...... 25 LIST OF ABBREVIATIONS

AAS Atomic absorption spectroscopy

ATR-FTIR Attenuated total reflectance Fourier transform infrared spectroscopy

DHBA Dihydroxybenzoic acid

DLS Dynamic Light Scattering

DLVO Derjaguin, Landau, Verwey and Overbeek

HA Humic acid

NOM Natural organic matter

PCCS Photon cross correlation spectroscopy

XPS X-ray photoelectron spectroscopy 1. AIM OF THIS WORK

1.1 Motivation The growing development and innovation of nanoparticles (NPs) for different applications rapidly results in an increasing number of NP-based products [1] with metal NPs being an important group of NPs [2, 3]. There is a concern that NPs (including metal NPs) may be released into the environment during production, use, or disposal of these products. Other important sources for metallic NPs include traffic emissions and wear particles from e.g. tires and brake pads, reported as major respiratory pollutants in urban areas [4]. Due to their small size, it has become important to assess possible risks of their environmental dispersion, as they in some cases may be harmful to humans or aquatic organisms. Recent modern analytical techniques have enabled investigations of different properties of NPs in the environment. Nevertheless, there are still knowledge gaps concerning changes in physico-chemical characteristics and transformation of metal NPs upon environmental entry [5].

The concern for adverse effects induced by NPs has resulted in the development of different national and international research projects aiming for in-depth studies of different aspects including e.g. NP toxicity, transformation/dissolution, environmental fate, and hazard assessment. This thesis forms part of the national Mistra Environmental Nanosafety Program initiated by the Swedish Foundation for Strategic Environmental Research (MISTRA). The framework consists of different academic groups from Chalmers University of Technology, Karolinska Institutet, KTH Royal Institute of Technology, Gothenburg University, and Lund University, together with industrial partner Akzo Nobel. The program provides an interdisciplinary platform for researchers from e.g. nanoscience, medicine, surface chemistry, material science, life cycle analysis, and social science. The program focuses on characterization of NPs in different environmental aquatic media, toxicity studies in the food chain using different aquatic organisms, integrated risk assessment of NPs, and societal dimensions of nanosafety. More specifically, the program is divided into five closely integrated work packages (WP); 1) exposure, fate and life cycle assessments; 2) the molecular nano- interface; 3) integrated hazard assessments; 4) social dimensions of nanosafety, and 5) nanotechnological solutions to environmental problems. Three case studies are overarching these work packages; i) emissions of NPs from automobile applications to road run-off water, ii) systematic studies of commercially available nanomaterials, and iii) future nanomaterials. The Mistra Environmental Nanosafety Program supports and contributes to the Swedish environmental quality objective “A non-toxic environment” with the objective to hand over an environment with reduced chemical risks to the next generation.

This thesis has been performed within the framework of WP 1 in the Mistra Environmental Nanosafety program, with implications for case study 1, in which transformations of metal NPs have been investigated. Data and knowledge are generated related to metal NP agglomeration, sedimentation, transformation/dissolution, and changes in surface chemistry and speciation of metal NPs in environmental-relevant media such as freshwater. These aspects are highly important and integrated within the framework of the other work packages of the program, e.g. used to understand and design relevant experiments on ecotoxicity, metal dissolution, and changes in surface chemistry (adsorption of natural organic matter, surface oxide properties). Another example is the generation of rates of agglomeration, sedimentation, and dissolution, essential data in modelling of the fate and risks of NPs in the environment [6, 7]. In addition, thorough work has been done to assess the influence of probe sonication, the

1 most common method used for preparation of NP dispersions, on the particle and surface characteristics, dissolution properties and agglomerate size. These aspects are crucial as they largely influence e.g. nanotoxicological interpretations. This work also relates to the 2030 agenda made by UN on global sustainability goals, Sep.2015 set to protect the planet from different man-made obstacles for a peaceful living. Among the 17 sustainable goals, life below water, life on land, and clean water and sanitation focus on minimizing the use of hazardous chemicals. Since NPs and nanomaterials are the new age chemical substances, knowledge about their potential risks on the environment is a very important for a safe planet and to achieve the set goals.

1.2 Aim The aim of this work was to study metal NPs including mainly copper (Cu), aluminum (Al) and manganese (Mn) related to the following specific objectives:

1. To test and assess how the key parameters of an existing dispersion protocol for sonication of NPs influence the stability/mobility, dissolution/transformation and particle characteristics of reactive metal NPs, and propose relevant sonication settings for reactive metal NPs. 2. To obtain an improved understanding of agglomeration, sedimentation and dissolution processes of metal NPs in freshwater, as well as providing dissolution and sedimentation rates. 3. To investigate how different types of natural organic matter (NOM) interact with metal NPs from an adsorption, agglomeration, sedimentation and transformation/dissolution perspective, and if possible deduce possible NP-specific interactions with NOM.

2. INTRODUCTION

2.1 Nanoparticles and nanomaterials The history of nanotechnology dates back to 600 BC when carbon nanotubes were used in Wootz steel [8]. Earlier use of nanotechnology was also related to painting or art, until the discovery of the modern electron microscope in 1931 [9]. Nanotechnology then became a sophisticated science for materials in the scale of 1-100 nm. For NPs, the surface properties are in general more important for the physical and chemical properties compared to the corresponding bulk material characteristics [10]. For example, NPs used in photovoltaic cells have higher absorbance capacity than if used within continuous sheets of bulk materials [11]. NPs and nanomaterials (NMs) possess large surface areas due to a large number of surface atoms, a property that is useful in different industrial applications. Today, NPs are used in various societal sectors such as medicine, military, astronomy, and also in consumer products such as self-cleaning glass, cosmetics, detergents and antibacterial clothing [12].

2.2 Environmental concern The growing use of NPs has raised concern over their potential harmful effects on human health and the environment [13, 14]. This in turn has warranted thorough scientific investigations in areas of research such as nanotoxicology. The assessment of human risks of NPs in a closed environment with a direct exposure can sometimes be relatively straight forward when compared with conditions in an open environment. In the former case, concentrations and physico-chemical properties of NPs can be monitored with a defined set up [15], whereas the latter situation is much more complex when NPs are released to the

2 environment via e.g. industrial waste or disposal and recycling of NP-products. One aspect of this complexity is the fact that NPs undergo different transformations due to their interactions with various kinds of organic and inorganic constituents present in the environment, which can alter their mobility, toxicity, and particle properties [16, 17]. For example, Ag NPs are sulfidized upon interaction with inorganic or biogenic sul de present in sub-aquatic sediments. This changes their agglomeration behavior, surface chemistry and dissolution properties, and therefore their fate and potential toxicity [18].

Natural organic matter (NOM), [19] present in the environment is a complex matrix of organic molecules formed by the breakdown of terrestrial plants and by products of aquatic plants including phytoplankton (algae, fungi, etc.). NOM is important for the fate of NPs as it influences their agglomeration behavior by adsorbing and changing the surface properties [20]. The molecular weight and size of the NOM depend on the source of extraction. NOM has been classified as humic (hydrophilic) and non-humic (hydrophobic) substances where the former includes high molecular weight organic constituents. Humic substances are further subdivided into three categories:

1. Humic acid (HA), which precipitates out of solution at pH < 2. 2. Humin, which is insoluble at all pH. 3. Fulvic acid, which is soluble at all pH.

The organic matter investigated in this study is HA, which is the main component of humic substances. It contains aliphatic-, aromatic-, carboxylic acid, amine-, alcohol, and carbonyl- functional groups, to name a few. Dihydroxybenzoic acid (DHBA) is another organic molecule studied within this work together with HA, and represents smaller sized degradation products in NOM, schematically shown in Figure 1.

Figure 1. Structure of studied isomers of DHBA and of humic acid

2.3 Risk assessment of NPs NPs have been proposed to be categorized into four groups to organize the prediction of their fate and toxicity: carbon NMs, metal oxide and metal NPs, quantum dots, and nano-polymers [21, 22]. Aitken et al.[23] report that not all NPs are hazardous and within a given category as NPs differ from each other in their properties. Hence a general risk assessment protocol may not be applicable in predicting their fate [23]. In some cases it is important to perform individual risk assessments [24-27]. From a regulatory standpoint, there is work ongoing to adjust NPs into the regulatory framework, e.g. in REACH - Registration, Evaluation, Authorization and Restriction of Chemicals, implemented in the EU [28].

3 Conceptual models have been established to identify critical processes related to environmental risks of NPs, including: (i) the form, route and mass of NMs entering the environment; (ii) transformation, affinity for surfaces and fate in the environment; (iii) transport and geographic distribution in the environment; (iv) bioavailability; (v) toxic responses of individual organisms to exposure; (vi) effects on ecological structure; and (vii) changes in ecological/biogeochemical function in the environment [29]. NP characterization has in this context been identified as a key component for making reliable risk assessments [30]. This includes characterization of the NP suspensions using different experimental techniques and monitoring of their dissolution/transformation [31]. Thorough characterization improves interpretation and quality of e.g. ecotoxicity data, by for example aiding in making proper design decisions based on NP characteristics such as size and dissolution. Hazard identification and risk assessments can be done taking into account the above mentioned physical and chemical properties of NPs together with toxicity result [32-35]. Production volumes and transport of the NPs in the environment can be used in the risk assessments [2, 36]. Bioavailability is another important parameter for NPs, with one definition being “a relative measure of that fraction of the total ambient metal that an organism takes up when encountering or processing environmental media, summed across all possible sources, including water and food” [37]. The bioavailable concentration is very different from the total environmental concentration [38].

As mentioned earlier, NPs can interact with various organic and inorganic components upon release to the environment, which sometimes results in the formation of larger clusters of NPs, denoted agglomerates. The mobility of these agglomerates can be hindered in the environment due to their sedimentation [39]. Mobile NPs that are dispersed into aqueous systems and remain in the environmental medium can be transported further in the environment [40]. The mobility of NPs is thus one an important parameter as longer residence times in solution may enable further transport to other environmental settings of different characteristics. Dissolution is another important process to consider while studying bioavailability [41].

3. BACKGROUND

3.1 Choice of NPs The choice of the metal NPs (Cu, Mn, and Al NPs) of the work in this thesis was made based on differences in electrochemical and surface oxide characteristics. Cu NPs undergo relatively rapid dissolution in cell media at near neutral pHs [42] compared to Al NPs, which have a more protective surface oxide that efficiently hinders corrosion and dissolution at certain conditions [43]. Al spontaneously forms a passive surface oxide in air that remains in water of neutral and mild acidic pHs. At alkaline or very acidic conditions, it may be readily dissolved [44]. The surface oxides of Mn NPs and Cu NPs are non-passive though efficient barriers in many conditions. As an example, the extent of metal release from in cell media follows the sequence; Cu NPs>Mn NPs>Al NPs [45]. However, the metal release process is highly material and solution specific.

Cu NPs are e.g. widely used in electronics, metallic inks and textiles [46-48]. The environmental concentration of engineered Cu NPs in Asian waters river water has as an example been estimated to 0.06 mg/L, which according to Chen et al. is a high concentration for aquatic microorganisms [49, 50]. However, the Cu NP concentration in wastewater effluents in other investigations in central Europe has been estimated to be much lower than this concentration (<0.1 —g/L) [33]. There is hence a large uncertainty concerning

4 environmental concentrations of Cu NPs (and of all metal NPs). Although there are studies on effects of NOM on Cu NP dissolution and changes in chemical and physical properties [51-53], there is still a knowledge gap concerning the interaction mechanisms between Cu NPs and NOM. Similarly, Mn and Al NPs are used in various applications in the society, [54, 55] but lack predictions of their environmental concentrations. Possible risks induced by the use of such NPs cannot be ignored [54, 56]. The lack of existing knowledge about the interaction between NOM and these metal NPs was another reason for their investigations in this thesis.

3.2 Relevant NPs characteristics from an environmental perceptive There are several important key parameters of NPs to be able to predict their environmental fate in freshwater environments. Such parameters include e.g. their size and surface properties (e.g. surface charge, surface oxide composition). The size is important as it influences the tendency to form agglomerates, subsequent sedimentation and metal release [57]. Besides, biological uptake and subsequent effects, also largely size-dependent effects have been reported, e.g. by Ivask et al [58] and De Jong et al [59]. Smaller sized gold and silver NPs proved to be more toxic and more easily transported in the studied mammalian bodies than corresponding larger sizes particles.

Surface properties are other important factors to consider while assessing NP fate [60, 61]. Metal NPs spontaneously form surface oxides in contact with ambient air, with the exception of noble metals such as gold and platinum. The characteristics of this surface oxide changes depending on the chemical environment. Dissolution and adsorption are for instance influenced by the properties of the surface oxide [62, 63]. For instance, the surface oxide on Al NPs, Al2O3, is passive in nature in pure water at a near neutral pH, making the Al NPs less reactive compared with the Cu NPs which undergo faster dissolution due to a less protective oxide. The oxide layers of the metal NPs investigated in this thesis have been thoroughly characterized elsewhere by means of different surface analytical and electrochemical techniques [45]. The composition of the surface oxides of the Al, Mn and Cu NPs are schematically illustrated in Figure 2.

Mn2O3/MnO2 CuO

Al2O3 MnO Cu2O

Figure 2. Schematic illustration of surface oxides on the metal NPs investigated in this thesis (adapted from [44]) The surface charge is besides the surface oxide composition and the particle size very important parameters that govern the environmental interaction of NPs [64]. Metal NPs acquire a surface charge when exposed to an aqueous solution. The charges on the surface originate e.g. due to protonation or deprotonation of surface hydroxyl groups. The surface charge is important as it determines the repulsive electrostatic force between particles and influences the adsorption characteristics of different ligands (briefly described in section 4.3) [65].

5 3.3 Selected phenomena influencing NP fate in the environment

3.3.1 Colloidal stability The classical DLVO (Derjaguin, Landau, Verwey, and Overbrook) theory for colloidal stability can help to estimate the stability of NPs in aqueous solutions. This theory describes the colloidal stability by considering the balance (summation) between two forces: the attractive van der Waals forces and the repulsive electrostatic double-layer forces between particles. The van der Waals force is always attractive between similar particles in any media and is due to interactions between induced and/or permanent dipoles. The van der Waals force depends on the size of the particles and the material properties such as the dielectric constant, which is expressed by the Hamaker constant. Metals NPs are conductive and polarizable and therefore they have high dielectric constants and refractive indices. This leads to a significantly higher Hamaker constant for metals than for other materials. Hence, metal NPs in aqueous solution without any surface modifications have a high tendency to agglomerate and precipitate (sediment)due to the strong van der Waal force [66].

On the other hand, charged NPs will develop a repulsive force, due to overlap of the ion clouds, which give rise to an increase in the osmotic pressure. Hence, NPs with the same polarity of the surface charge will repel each other. In aqueous solutions, NPs acquire surface charges due to dissociation of surface oxide groups and this will result in the redistribution of ions in the aqueous solution, Figure 3. These surface groups attract oppositely charged ions (counter ions) from the aqueous solution and ions with similar charge (co-ions) will be depleted to make a neutral system [67]. The concentration of counter ions will decrease and the concentration of opposite ions will increase with the distance from the surface until equilibrium is reached out in the bulk solution. This distribution of ions surrounding the particles is called the electrical double layer and the potential in this layer will decay exponentially with the distance. The repulsive electrostatic double-layer force (if strong enough) between particles improves the colloidal stability.

Although DLVO theory has been used for explaining the colloidal stability, the validity of this theory for NPs is not so clear. It is because NPs does not always possess spherical shape as mentioned in the classic DLVO theory. Besides, as mentioned earlier, NPs possess more surface atoms than the bulk and so dielectric properties of NPs differs from the bulk materials [68]. These dielectric properties plays an important role in DLVO force calculations [69]. In addition, the DLVO theory only considers the electrostatic double-layer repulsion and the attractive van der Waals forces. However, other forces such as steric repulsion cannot be ignored, e.g. when organic matter is adsorbed to the NPs. Steric repulsion, which originates from volume restrictions and inter-penetration effects of the adsorbed molecules, can hence also play an important role in particle agglomeration. When two particles with adsorbed organic molecules such as DHBA or HA are brought close to each other, the particles can repel each other due to overlapping/compression of the adsorbed layers. This repulsion contributes to the repulsive forces between particles. Furthermore, the attractive van der Walls force will be lowered for particles coated with an organic layer (to some extent depending on layer characteristics) since the Hamaker constant for the organic layer is significantly lower compared to the metal NP. Figure 3 illustrates these different forces acting between particles in an aqueous medium.

6

Figure 3. Schematic illustration of electrostatic-, van der Waals- and repulsive steric forces between particles in solution

3.3.2 Metal release/ dissolution Metal release/dissolution means that metals(ions) leave the NP and migrate to the bulk solution as illustrated in Figure 4 [70] This process is governed by controlling factors such as the surface oxide composition and characteristics [71], the solubility of oxide constituents, electrochemical properties [72], solution chemistry, and concentration gradients of metal ions between the particle surface and bulk solution. The extent of metal release depends further on material- and surface properties as well as ligand-interactions and size [73]. For instance, Zn NPs undergoes rapid dissolution whereas titanium dioxide and cerium oxide NPs hardly dissolve in pure water at neutral pH. The presence of different ligands in solution (e.g. counter ions and organic matter) may either promote dissolution by ligand-induced processes, or hinder dissolution due to the formation of a protective layer [74].

Figure 4. Illustration of dissolution/metal release processes from the surface of metal NPs

Metal release/dissolution data provides information on whether the NPs remain as particles or becomes dissolved and released into solution and, from speciation measurements/predictions, whether released metals form strong complexes in solutions of low bioavailability or remain as free ions or labile complexes of high bioavailability. These aspects are thus important to consider when assessing the fate of NPs in the environment.

7 4. EXPERIMENTAL APPROACH

4.1 Sample preparation Cu, Al and Mn NPs were investigated in this thesis. ZnO and SiO2 NPs were in some studies investigated for comparison. Dispersions of the NPs were prepared from stock solution with particle loadings of 2.56 g/L or 1 g/L in ultrapure water using probe sonication for 15 min or 3 min [75]. After sonication, the NP dispersions were diluted to 0.1 g/L in different exposure media. The exposure media were 1 m M sodium perchlorate, NaClO4 (Paper I) and synthetic freshwater (Paper II, ionic strength 1.2 m M) with DHBA and HA in concentrations given in Table 1. NaClO4 was used due to its relative low impact on corrosion, compared with e.g. sodium chloride, NaCl.

Table 1. Concentrations of model NOM in synthetic freshwater, and composition of the freshwater

Organic matter Concentration in synthetic freshwater

2,3 DHBA 0.1 mM (15.41 mg/L)

1 mM (154 mg/L)

3,4 DHBA 0.1 mM (15.41 mg/L)

1 mM (154.1 mg/L)

Humic acid 2 mg/L

20 mg/L

40 mg/L

Freshwater components Molar concentration (M)

-5 NaHCO3 7.5·10

-6 KCl 7.8·10

-4 CaCl2·H2O 2·10

-5 MgSO4·7H2O 5·10

The concentrations of HA and DHBA were chosen according to the range of environmental concentrations of humic acid.[76]

8 4.2 Characterizations of NPs

4.2.1 Size distribution As mentioned above, the NP size is an important parameter for the transformation of NPs in the environment as it e.g. influences sedimentation and dissolution. Therefore the evolution of particle size over time was investigated in the media described in Table 1.

Changes in size of the NPs and their agglomerates in solution were investigated by means of photon cross correlation spectroscopy (PCCS), which is based on the principles of dynamic light scattering (DLS). A laser beam illuminates the particle suspension and the scattered light intensity is measured with a detector, Figure 5. Observed fluctuations in the intensity of the light are the result particle movement (diffusion) in the solution. In PCCS, two lasers are used instead of one, as used in conventional DLS. The simultaneous irradiation of the two lasers gives rise to two speckles that are detected in two detectors [77]. Speckles with the same scattering vector are only taken into consideration for further analysis and multiple scattering is hence less of a problem.

Figure 5. Schematic illustration of DLS analysis of NP dispersion

The auto-correlator connected to the photodetector then compares the intensity of the scattered light at time t with the intensity to a time later and the fluctuations are quantified by a second order correlation function. The software uses a non-linear least square fitting algorithm to fit the measured correlation function to solve the correlation function decay ȗ. The decay is used to calculate the diffusion constant D, which in turn is used to calculate the particles hydrodynamic radius (r) by using Stokes- Einstein equation, eq.1:[78]

퐫=퐤퐓ퟔ훑훈퐃 (eq. 1)

Where k = Boltzmann’s constant T= temperature in K η = solvent viscosity

4.2.2 Zeta potential of particles in dispersion As explained earlier in section 1.4, the surface charge is an important property of NPs. In aqueous solutions, all particles will be charged and surrounded by a layer of oppositely charged ions, forming an inner strongly bonded layer (the Stern layer). Outside this layer, there is the diffuse (electrical) double layer, which consists of ions loosely bound to the particles and ions with similar charge as the particles. The distribution of these ions changes with the distance from the surface until equilibrium is reached. When the particles move in

9 the solution, the counter ions closest to the surface will move with the particle and the rest of the ions in the electrical double layer will be more stationary in the solution creating a plane of shear (slipping plane). The zeta potential is the potential measured at the slipping plane, Figure 6 [79]. A higher magnitude of the zeta potential corresponds to a more highly charged surface, which implies improved colloidal stability due to repulsion between charged surfaces of similar polarity [79].

Figure 6. Schematic illustration of the diffuse double layer of a NP and the position of the slipping plane where the zeta potential is determined

The zeta potential is estimated by irradiating a laser through the sample at the same time as an electric field is applied to it. The particles then move due to their interaction with the electric field, a phenomenon known as electrophoresis. The particle velocity, V, is determined from the Doppler shift from the movement of the particles due to the applied electrical field, and from this the electrophoretic mobility is calculated. The zeta potential, ȗ, is then estimated from the electrophoretic mobility by using Henry’s equation (eq. 2).

μ=2휀휁푓(휅푎)3휂 (eq. 2)

Where μ = mobility 휀= dielectric constant ζ = zeta potential k = Boltzmann’s constant

-1 f(ka) is Henry’s function where ț[nm ] is the Debye parameter, and its reciprocal is the Debye screening length of electric double layer, and a is the particle radius. ka thus is the particle radius to the double layer screening thickness.

Two approximations are frequently used to estimate the value of f(ka), and they are schematically presented in Figure 7. For values of ka<<1, the value of f(ka) is approximated to 1. This is called the Hückel approximation. This approximation is hence valid for a situation when the screening length of the double layer is thicker than the particle radius, and thus represents a situation with relatively small NPs (<<100 nm) and/or an electrolyte of low ionic strength (mM range or lower). For example, the Hückel approximation is valid for a 20 nm NP in a solution of 1 mM ionic strength [80]. The Smoluchowski approximation assigns a

10 value of 1.5 for f(ka)and is used when the screening length of the electrical double layer is much smaller than the particle radius (ka>>1). This hence corresponds to a situation with larger particle sizes and/or higher ionic strengths compared with the situation when the Hückel approximation is valid. In this study, the Smoluchowski approximation has been used for zeta potential measurements in freshwater without NOM added, whereas the Hückel approximation has been used to calculate their zeta potential in freshwater containing HA or DHBA, as the particles sizes are smaller than 100 nm and the ionic strength approx. 2 mM in these media [80].

Figure 7. Schematic illustration of particle size vs. thickness of the electrochemical double layer using the Smoluchowski and the Hückel approximations, respectively

4.2.3 Dissolution/metal release The dissolution/metal release from the metal NPs was determined by means of atomic absorption spectroscopy (AAS). AAS was chosen because it can determine metal concentrations in solution down to a few —g/L. Dissolution in this context relates to chemical dissolution of metals whereas metal release in addition involves electrochemical and ligand-induced processes.

The technique relies on a solution that has been atomized and that absorbs incident electromagnetic radiation. Light is incident from a narrow line source, a hollow cathode lamp, which emits light of specific wavelengths depending on the specific element of interest. The incident electromagnetic radiation passes through the atoms, absorbing photons from the incident light. The detector at the other end shows the reduction in intensity of incident light. The value of absorbance is then used to calculate the concentration of the specific metal in the sample by using Beer-Lambert law. Two different modes of AAS have been used in this study:

1. Flame AAS for metal concentrations in the mg/L range. 2. Graphite furnace AAS for trace concentrations of metals (—g/L).

The concentration of the measured metal is compared with the amount of light absorbed by known standard concentrations of the same metal in order to obtain the actual metal concentration of the sample [81].

4.2.4 Adsorption studies of organic matter Attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR) was used to study the adsorption of NOM on the NPs. A benefit of this technique is that it can be

11 employed in situ and is hence able to investigate the adsorption of molecules in solution, e.g. upon NOM interactions with the NPs.

ATR-FTIR measures the absorption of infrared light by the samples. An IR beam is irradiated to an optically dense crystal with a high refractive index in a certain angle. The evanescent wave, created due to total internal reflectance, then interacts with the sample positioned above the crystal leading to absorption of infrared light. The evanescent wave will thus get attenuated due to absorption of certain vibrational transitions in the samples. The recorded attenuated IR beam is then used to generate an IR spectrum [82].

5. KEY FINDINGS (Papers I and II)

5.1 Sonication of NP suspensions largely influences the particle size and extent of metal release (Paper I). There are several methods to disperse NPs in an aqueous solution, e.g. using sonication probes or bath sonicators [83, 84]. It is essential to have a dispersion method that is reproducible and well-documented protocol. It is furthermore very important to have knowledge on how it influences the NP properties, in terms of e.g. particle size, surface oxide composition and dissolution [85, 86]. A dispersion method has been designed by Taurozzi et al. [83] to prepare NP dispersions in a way that could produce reproducible results in different laboratories. This protocol emphasized that the delivered acoustic energy of the sonication procedure should be determined and reported. Taurozzi et al. [87] Further more reported that different sonication methods disperse NPs to different extent.

Manual shaking, vortex-blending, bath sonication and probe sonication were in this study tested for dispersing Cu-, Mn-, Al-, and ZnO NPs [75]. Probe sonication was found to be the most efficient method to prepare dispersions of the studied NPs in terms of smallest particles size and lowest degree of sedimentation during sonication [75]. These findings were in line with previous investigations [88-90]. A particle loading of 2.56 g/L is common practice for stock solutions. With this concentration of NPs, large agglomerates were observed when dispersed and sonicated at standardized settings using a probe sonicator for the Cu-, Mn-, Al-, and ZnO NPs (7.35 W of delivered acoustic energy). The underlying reason is mainly that such a high particle loading results in a high collision frequency of the NPs, which in turn results in pronounced agglomeration, Figure 8. The stock solution concentration was therefore reduced from 2.56 g/L to 1 g/L in order to have less sedimentation and agglomeration in the sonicated stock solutions.

12

Figure 8. Schematic illustration of the influence of NP loading in the stock solution on their extent of sedimentation and agglomeration

The extent of particle de-agglomeration and sedimentation was also seen to depend on the amount of delivered acoustic energy during sonication, and on the properties of NPs. 1 mM NaClO4 was used for the investigations of NP agglomeration in order to use a relatively non- corrosive electrolyte, thereby minimize the effect of metal NP corrosion. The results showed that the agglomerate size in the Mn NP and the Al NP stock solutions that were sonicated for approx. 3 min (182 s) in 1 mM NaClO4 were relatively larger compared to suspensions sonicated for approx. 15 min (882 s), Figure 9. However, the Cu NPs that were sonicated for the longer time period (882 s) did not disperse the agglomerates significantly better, with relatively similar particle sizes as observed in 1 mM NaClO4. It has previously been suggested that different amounts of delivered acoustic energy are needed to disperse different NPs, which is connected to e.g. differences in effective density and surface chemistry of the NPs [91].

Figure 9. Schematic illustration of NP agglomerates formed in Cu-, Mn- and Al NP stock solutions(1 g/L, ultrapure water) sonicated for 3 min and 15 min followed by dilution to approx. 0. 1 g/L in 1 mM NaClO 4 No significant differences in the amount of released Cu in solution (Cu able to pass through a membrane with a pore size of 20 nm) over time were observed for Cu NPs in 1 mM NaClO4 between solutions sonicated for 3 and 15 min, Figure 10. The total amount of released Cu after 24 h could be under estimated, as the amount of soluble Cu in solution was close to the theoretical saturation point (4 mg/L) [75], and the amount of Cu, possibly precipitated or

13 bonded to large complexes (< 20 nm) was not considered during analysis. For the Mn NPs on the other hand, as can be seen in Figure 10, the longer sonication time enabled de- agglomeration to relatively smaller sizes. This resulted in a relatively larger specific surface area and thus more Mn release compared with release in the shorter sonication time [75].

0.25 4 24 0.25 4 24

Figure 10. Release of Cu (left) and Mn (right) in 1 mM NaClO4 (after membrane filtration with a pore size of 20 nm) from Cu NP suspensions (0.1 g/L) probe-sonicated for 3 and 15 min, respectively

As earlier mentioned, high particle loadings of the stock solutions (2.56 g/L vs. 1g/L) resulted in increased agglomeration of the NPs. This influenced was clearly noticed for Mn NPs as shown in Figure 11, where less Mn was released for the highest particle loading (2.56 g/L) compared with the lower loading (1 g/L) of the stock solution. The surface area of large agglomerates is less compared to a relatively more dispersed system, and thus the amount of released metals becomes less. Differences between Cu NPs in stock solutions of different loadings were not clearly visible as the release of Cu was close to its saturation point of 4 mg/L, see previous discussions.

0.25 4 24

Figure 11. Released amount of Mn from Mn NP suspensions (1 g/L and 2.56 g/L stock solutions) in 1 mM NaClO4 after probe-sonication for 15 min, at a particle loading of 0.1g/L

14 It is concluded that the dispersity of the NPs depends on the stock solution concentration and the delivered acoustic energy due to sonication, together with NP properties. Metal release/dissolution of NPs in turn depends on the dispersity of NPs and the NP characteristics such as surface oxide characteristics.

5.2 Rapid sedimentation of NPs in the stock solution results in significantly lower administered doses than the nominal doses of sonicated metal NP dispersions – an effect of strong attractive van der Waals forces (Paper I). NP investigations (e.g. nanotoxicity studies) entail reproducible and well-characterized NP dispersions that can be repeated and used in subsequent studies. The pipetted concentration, or dose, taken from a stock solution is normally expected to be same in all experiments and equal to the nominal concentration of the stock solution. This may however not be the case if there is significant agglomeration in the stock solutions.

The pipetted concentrations of NPs from sonicated stock solutions were investigated to quantify the possible error in the administrated dose induced as a result of NP sedimentation during preparation of particle suspensions. The administered (real) doses transferred from the sonicated stock solutions were for all metal NPs found to be significantly lower, in the range of 30-70%, than the nominal doses, Figure 12. Observed differences between the nominal and the administered dose were independent on particle loading and sonication time (delivered acoustic energy).

Figure 12. Schematic illustration of the difference between the nominal and the added (administrated) concentration of NPs pipetted from a stock solution (2.56 g/L or 1 g/L) probe- sonicated for 15 min or 3 min in ultrapure water

Observed differences in doses are predominantly related to the strong van der Waals attractive forces between the metal NPs, which result in rapid agglomeration and concomitant sedimentation of the NPs already in the stock solutions. This is also evident from DLVO calculations showing that the magnitude of the van der Waals attraction is dominant over the electrostatic repulsion for all metal NPs [75]. High van der Waals attractions are related to their high Hamaker constants, which depend on the dielectric constant and electronic structure of the metals [92]. Effective Hamaker constants for a core-shell geometry (i.e. metal core with a surface oxide) were calculated to investigate the effect of the surface oxide on the metal NPs in the DLVO calculations. Even though the surface oxide thickness reduced the Hamaker constant, the reduction was relatively small for oxide thicknesses relevant for the metal NPs

15 of this study. This confirms a dominance of the van der Waals forces over the electrostatic forces. The trend in nominal dose between the NPs was found to be material specific (Figure 6, Paper I). The difference was highest for the Mn NPs for which the administered dose was less than 40% of the nominal dose. This means that each metal NP has to be investigated individually for determining the real administered dose. Effective density differences between metal NPs also play a role for particle sedimentation and should be considered.

It is concluded that probe-sonicated NPs form large agglomerates already in ultrapure water during the sonication step of the stock solution and undergoes rapid sedimentation. It is hence essential to monitor the actual concentration of transferred doses in e.g. nanotoxicity studies in order to avoid misinterpretations.

5.3 Metals are released from Cu, Al and Mn NPs already during the sonication step and the presence of biomolecules can enhance this process (Paper I). Sonication of metal NP suspensions evidently alters the properties of the NPs, as described in terms of size and extent of agglomeration. Even the surface oxide characteristics become influenced by the sonication process, properties that in turn can influence the extent of metal release from the NPs. Since metal NPs have been disclosed to dissolve and release metals already during the sonication step, it is vital to know if, and to what extent, this occurs. If the metal NPs dissolve to a large extent already in the stock solution, the added dose will both comprise metal NPs and dissolved metal species. This has of course large implications on e.g. nanotoxicity studies where NP-specific mechanisms are being investigated. The effect of sonication on NP dissolution was investigated in which NP dispersions (Cu, Mn, Al, ZnO) were filtered through a 20 nm-sized pore membrane directly after sonication (15 min) followed by immediate analysis by means of AAS, Figure 13. This measured metal concentration includes both NPs sized less than 20 nm and ionic metal species released during sonication. However, as shown in Figure 2 in Paper I, the NPs were much larger sized than 20 nm, which means that the fraction sized < 20 nm in Figure 13 mainly represents dissolved metal species.

Figure 13. Released metals in solution without BSA (left) and with BSA (right) after membrane filtration with a pore size < 20 nm in stock solutions (1 g/L) of NPs of Cu, Mn, Al and ZnO probe-

16 sonicated for15 min. The asterisk for Al NP indicates that the concentrations of Al could not be quantified due to low concentrations (<0.1% of added Al)

As seen in Figure 13, metals were released from the NPs already during the sonication step of the stock solution. Observed differences in metal release between the NPs is due to differences in surface oxide characteristics of the metal NPs (Cu, Mn, Al). Most metals were released from the Mn NPs followed by the Cu NPs. The passive surface oxide of Al efficiently hindered the release of Al. The ZnO NPs were chemically dissolved (no surface oxide) relatively fast. For the Cu NPs, the total measured Cu concentration in solution was limited by saturation with concomitant precipitation.

The addition of bovine serum albumin (BSA) to the stock solutions prior to sonication for Cu and Mn NPs further increased the release of Cu from the Cu NPs to 4% and 32% respectively as shown in Figure 13. This is because BSA adsorbs to the NPs with a concomitant weakening of metal-oxygen bonds due to its surface coordination. This result in a ligand exchange with the surface oxide that either enhances the release and/or changes the surface pH. BSA can also enhance the release of metals due to complexation of dissolved metal species in solution driving these processes forward. BSA has been used in several toxicity studies to stabilize NPs during sample preparation [93-97].

The effect of using stabilizing agents such as BSA on the extent of metal release/NP dissolution should hence be taken into consideration during e.g. toxicity investigations of NPs. Results from such toxic assessments can otherwise be misleading as the reported toxicity could be due to released metal ions rather than metal NPs. This is e.g. important since metal NPs and released ions and formed complexes have different affinities and abilities to interact with cell membranes [98, 99]. The presence of adsorbed BSA, a bio-corona, will also influence subsequent cell- interactions.

5.4 The extent of particle agglomeration of the Cu, Al and Mn NPs depends on the type of surface adsorbed NOM (Paper II). If NPs are dispersed into the environment, their interaction with NOM is one key factor that will influence their agglomeration behavior since, if adsorbed; it will change the surface properties of the NPs. It is hence important to study the interaction between NPs and NOM, including different kinds of NOM, in order to enable predictions and improve the current understanding on how metal NPs will change their size once dispersed in e.g. in a freshwater setting.

Figure 14 shows the size distributions of Cu, Mn, and Al NPs approx. 15 min after being sonicated in ultrapure water and diluted to freshwater containing 1 mM 2,3 DHBA, 1 mM 3,4 DHBA, or 20 mg/L HA. It is evident that both DHBA and HA in freshwater, at least initially, counteracted agglomeration compared with freshwater only. The NPs agglomerated and sedimented in less than 30 min upon exposure in freshwater without any organic component. This is, as discussed in section 5.3, predominantly due to high van der Waals forces between the metal NPs, and manifested by the large agglomerate sizes observed in Figure 14.

17

Figure 14. Size distributions of Cu, Mn and Al NPs in synthetic freshwater with and without NOM (1 mM 2,3 DHBA, 3,4 DHBA or 20 mg/L HA), measured approx. 15 min after solution sonication

The results show that HA in a concentration of 20 mg/L kept the NP agglomerate size close to the primary sizes of the Cu and Al NPs (see Paper II). This reduction in agglomerate size upon the addition of HA, also observed for the DHBA monomers, Figure 14, is related to the electrostatic and steric repulsion induced by adsorbed HA or DHBA. Surface adsorption of DHBA or HA to the NPs was confirmed by the ATR-FTIR measurements (Figure 2, Paper II). The observed size distributions of the NPs in freshwater containing NOM depended on its concentration in freshwater (Figure 7, Paper II). For example, size distributions of the different metal NPs are presented in Figure 15 for different concentrations of HA.

Figure 15. Relative differences in size of metal NP agglomerates in synthetic freshwater only compared with freshwater containing HA of different concentration (pH 6.2)

The size of NP agglomerates formed in freshwater containing 20 mg/L HA was the smallest among all three concentrations of HA. Cu and Al NPs formed relatively large agglomerates in 2 mg/L and 40 mg/L HA compared to 20 mg/L HA, as also shown in Figure 7, Paper II. This may be connected to the ability of HA to cross-link at high concentrations (40 mg/L) with increased agglomeration as a result, see Ghosh et al.[100]. At low HA concentrations (2 mg/L), the steric repulsion may be insufficient (too thin layer of adsorbed HA) to overcome and screen the van der Waals forces.

Observed changes in surface charge were expected based on the adsorption of DHBA and HA. Differences in surface charge, represented in terms of the measured zeta potential, are shown for the different metal NPs and NOM in Figure 16.

18

Figure 16. Differences in zeta potential of the metal NPs (0.1 g/L) in synthetic freshwater (pH 6.2) only and when containing either 1 mM 2,3 DHBA, 3,4 DHBA or 20 mg/L HA) measured approx. 15 min after the sonication

The zeta potential became more negative in freshwater in the presence of NOM, which indicates adsorption of NOM to the NPs. The magnitude of the zeta potential depended on the type of NOM and also on the orientation of functional groups linked to the particle surface. For example, the Mn NPs were more charged in freshwater containing 1 mM 2,3 DHBA compared with freshwater with 1 mM 3,4 DHBA. All NPs became more charged, and hence more stable in solution in the presence of HA compared with DHBA (briefly discussed in section 5.4).

The size of the agglomerates depends on the concentration of DHBA (Figure 7, Paper II). For example, agglomerates of Cu NPs in freshwater containing 0.1 mM 2,3 DHBA or 0.1 mM 3,4 DHBA were larger compared to corresponding sizes in freshwater containing 1 mM of the DHBA monomer.

In all, it is evident that the understanding of the fate of metal NPs in the environment requires knowledge also on concentrations and types (e.g. size, functional groups) of NOM and on its influence on other phenomena including e.g. dissolution, agglomeration, and sedimentation of metal NPs, as discussed in sections 5.4 -5.6.

5.5 NOM delayed the sedimentation of the Cu and Al NPs, in particular pronounced in the presence of HA, whereas no significant effect was observed for the Mn NPs (Paper II). NOM is present in any natural water and is a complex mixture of different organic molecules, and its composition and properties varies depending on different geographical areas. This variation gives it e.g. altered properties in terms of metal complexation. It is hence important to look at different kinds of NOM when assessing its influence on the metal NP stability.

NP size distributions were monitored for 24 h in freshwater containing 20 mg/L HA to study agglomeration and sedimentation. As described in section 5.3 above, HA in concentration 20 mg/L was found to disperse NPs close to their primary size. The dispersity of the NP suspensions in 20 mg/L HA was retained up to 24 h as shown in Figure 17. However, for the higher HA concentration (40 mg/L) rapid sedimentation of the NPs took place within 4 h. At the lowest concentration of HA investigated (2 mg/L), the Cu and Al NPs underwent complete sedimentation within 6 h whereas the Mn NPs tended to be more stable with much lower sedimentation velocities.

19

Figure 17. Size distributions of NPs in synthetic freshwater with 20 mg/L humic acid over 24 h

In DHBA, Cu and Al NPs started to sediment after 4 h with complete sedimentation after 6 h. Metal NPs in freshwater of lower (0.1 mM) DHBA concentrations sedimented faster compared with corresponding conditions with the higher DHBA concentration (1 mM). This implies that even small organic molecules like DHBA postpone the agglomeration of these metal NPs when compared to freshwater only. In contrast, the Mn NPs remained stable in the presence of both DHBA and in HA up to 24 h.

The sedimentation velocity of the metal NPs was calculated by taking into account the time of complete sedimentation of the NPs in fresh water with and without NOM (Table 4, Paper II). These numbers can be used in fate modelling to predict their environmental mobility. In general, the numbers imply limited mobility due to rapid sedimentation, with the exception of conditions with HA when sedimentation in many cases was significantly delayed. In contrast to the metal NPs, SiO2 NPs were not influenced by either the presence of DHBA or HA and readily remained stable up to 21 days (Figure 9, Paper II). One factor contributing to this may be due to its high charge of SiO2 together with significantly lower Hamaker constant compared with the metal NPs and hence smaller attractive forces between the NPs [101].

The dependence of the type and concentration of NOM on NP sedimentation was strong for the metal NPs. Smaller sized DHBA provided less protection against agglomeration, showing that it, in this context, cannot function as a model for the full, complex NOM, represented by HA.

5.6 The release of metals from the Cu, and Al NPs was enhanced in the presence of NOM, whereas no significant effect was observed for the Mn NPs despite NOM adsorption. Released Cu and Al readily formed strong NOM-complexes of low bioavailability in solution whereas released Mn remained as free ions (Paper II).

The interaction between NOM and metal NPs also influenced the extent of particle dissolution/metal release from the NPs. This is a complex interplay between many different phenomena including corrosion, surface oxide-NOM interactions, solution chemistry, and changes in surface pH, to name a few aspects. Dissolution/release rates can be used for environmental fate modelling of NPs in different aquatic settings and be used to improve the mechanistic understanding of how NOM, e.g. due to surface complexation or de-agglomeration, influences the propensity of metal NPs to release metals.

20 As an example, the extent of Cu release from the Cu NPs was enhanced in freshwater containing DHBA or HA compared with fresh water only. This is schematically illustrated in Figure 18. The same situation was observed for the Al NPs. In contrast, no large effects of NOM were observed for the Mn NPs, and scarce complexation with NOM was predicted for released Mn ions, remaining mainly as free ions in solution. Further details are given in Figure 10, Paper II.

Figure 18. Schematic illustration of the release of Cu from Cu NPs in synthetic freshwater only and in freshwater with either the DHBA monomer or HA. Speciation predictions of released Cu in solution are illustrated

This enhancement in metal release can partially be due to surface complexation of the DHBA and HA molecules to the surface oxide of the NPs as seen using ATR-FTIR (Paper II), conditions that weaken the metal-oxygen bonds and thus increase the extent of dissolution/metal release [102, 103]. Corrosion of NPs due to a reduced surface pH upon adsorption of NOM could be another possible reason for an increased amount of released metals/particle dissolution. This will also influence the metal concentration in solution, seen especially for Al and Cu that both have relatively low solubility in freshwater. More Cu in solution was released from the Cu NPs into freshwater containing the 2,3 DHBA molecule than if containing the 3,4 DHBA molecule. This could be related to the formation of a binuclear complex with the 3,4 DHBA molecule that decelerates dissolution compared to the 2,3 DHBA molecule (Paper II) with which Cu forms mononuclear complexes.[104] The Cu solubility was higher in freshwater containing 3,4 DHBA compared with 2,3 DHBA, which means that the enhanced extent of released metals into solution observed with the former is not because of a relatively higher solubility of Cu in this medium. Conversely, the release of Mn from the Mn NPs was not significantly enhanced by the introduction of DHBA (Figure 10, Paper II), although DHBA did adsorb to the Mn NP surface, as deduced by ATR-FTIR (Figure 2, Paper II). Similar observations have previously been reported where Mn release was seen to be unaffected by cell media containing proteins [42]. This concludes that the adsorption of organic molecules to the surface is probably not the rate limiting step in the dissolution process of the Mn NPs of this study.

In all, the results show that the dissolution/metal release rates of the metal NPs were relatively high in freshwater in the presence of DHBA or HA, with 50% of the particle mass estimated to be dissolved within 7-60 h, based on a first order dissolution rate equation (see Paper II). This implies that the influence of NOM will reduce the mobility of these metal NPs in freshwater settings and, that a large fraction of the metal NPs will dissolve. This implies that released Cu and Al from the Cu and Al NPs predominantly will form non-bioavailable

21 complexes in solution. Released Mn will not form complexes to the same extent and most probably mainly remain as free ions.

6. MAIN MESSAGES

It was concluded that sonication plays a vital role in size and dissolution of NPs as described in Section 5.1-5.3 and schematically illustrated in Figure 19. Sonication parameters thus should be standardized for individual NPs depending on their surface characteristics and kind of intended investigation. In addition, characterization of surface properties and the administered dose of NPs (transferred from stock solution) are very important aspects when investigating effects of metal NPs.

Figure 19. Schematic illustration of the effect of sonication on the dispersion of NPs metal NPs in suspension

The study further illustrates that NOM, such as DHBA and HA, adsorbs to the NPs and increases their colloidal stability by electrostatic and steric repulsion, though the strong van der Waals forces of metal NPs dominate (Sections 5.4-5.6). However, the extent of de- agglomeration and stability depends on the type of NOM and its concentration in freshwater as shown schematically in Figure 20.

22

Figure 20. Schematic illustration of the effect of NOM in freshwater on the extent of agglomeration, sedimentation, dissolution/metal release and speciation for metal NPs

Even though all concentrations and types of NOM to different extent enhanced the particle stability of the NPs of this study, in particular for synthetic freshwater containing 20 mg/L HA, the results imply a possibility for an increased mobility whereas an increased NP dissolution in the presence of NOM will limit their potential transport.

7. FUTURE WORK

A few proposals on future research directions related to this work.

1. Detailed study on NP-NOM interactions. Further detailed investigations of NP-NOM interaction and their chemistry will provide knowledge about their properties such as stability and dissolution in environment.

2. Effect of sonication on organic matter stabilization. It would be relevant to study the agglomerate sizes and stability of NPs when they are exposed to organic molecules without sonication. In the actual environments, where particles will get exposed without sonication, it would be interesting to assess the concentration of organic matter required to stabilize NPs.

3. Behavior of NPs in different compartments of the aqueous environment. As pH and ionic strength vary in different aqueous settings in the environment, it would be interesting and relevant to study the effect of these conditions on NP stability and mobility.

4. Combinatory effects of fulvic and humic acid. Besides HA, Fulvic acid (FA) is the other important fraction of humic substances present in environment. It has relatively lower molecular mass than HA but contains more hydrophobic groups such as alkanes and aromatic hydrocarbons. Since HA contains more hydrophilic groups (hydroxyl and carboxyl), it is easier for HA to transfer from the water phase to the surface of

23 NPs [105]. This leads to the assumption that FA may result in less stable NPs aggregates than HA, as observed for TiO2[106]. However for some NPs such as Ag, FA tends to stabilize the particles more than HA. Interactions between FA and metal NPs would thus be interesting and relevant to look into in the context of NP risk assessment.

8. ACKNOWLEDGMENTS

I would like to thank everyone who contributed in some way to the work described in this licentiate thesis. It has been a learning experience for me on both a scientific and a personal level.

First and foremost, I am thankful for the funding provided by Mistra. I feel lucky to be a part of Mistra programme- Mistra environmental nanosafety. Many thanks and appreciation for all my supervisors for their support, motivation, patience and continuous optimism. This work would hardly have been completed without their guidance. My deep gratitude to Inger Odnevall Wallinder for giving me the opportunity to perform in this project. I would like to thank her for the guidance, useful critiques, advice and assistance in keeping my progress on schedule which have always push me positively. Many thanks to Susanna Wold for countless interesting discussions on the project and other fun chatting. I would like to thank Eva Blomberg for sharing her expertise in DLVO theory and colloidal chemistry. I would particularly single out Jonas Hedberg for all the guidance and cooperation during these two years. Special thanks to Jörgen Rosenquist for all the discussion while in Göteborg and patience to answer my queries through long emails. Many thanks to all the co-authors from Karolinska Institutet in the publication included in this thesis for the valuable insight to this project.

I would like to thank Yolanda Hedberg for the help during the initial days in the group. I always look you as a best role model for a scientist, mentor and teacher. Many thanks to Gunilla Herting for all discussions about AAS and interesting lunch room discussion. I would also like to take the opportunity to thank Mark Rutland to review my thesis during his summer vacation.

I thank master students Frederik Mathiason, Sara Isaksson and Maria-Elisa Karlsson (Shorty) for the help in my experiments during sunny summer in Sweden. Ex PhD student and a dear friend Sara Skoglund for teaching me the use of our very dear instrument - PCCS beside continuous support and encouragement. Thank you to Eva Luna at RISE for helping me in Zetasizer measurements at any time of my need. Thank you to Nazanin Alipour, PhD student at the polymer department for the cooperation on the sonicator standardization.

I would like to thank all my former and present friends at surface and corrosion division for listening, offering me advices and supporting me through the entire process. Special thanks to Laetita (Thank you for the help during thesis writing), Adrian, Cem, Laetita, Krishnan, Akansha, Tingru, Gen, Min, Marie, Fan, Erik, Georgia, Angelica, Neda, Golrokh, Zahra for lighting up the situation to a fun mode every day.

Last but not the least, my gratitude to my three member family in Sweden and the big fat family back in India for their endless love, support and patience so that I could dedicate precious family time and focus in the development of my research interest.

24 9. REFERENCES

1. Hansen, S.F., L.R. Heggelund, P.R. Besora, A. Mackevica, A. Boldrin, and A. Baun, Nanoproducts–what is actually available to European consumers? Environmental Science: Nano, 2016. 3(1): p. 169-180. 2. Sun, T.Y., D.M. Mitrano, N.A. Bornhöft, M. Scheringer, K. Hungerbühler, and B. Nowack, Envisioning Nano Release Dynamics in a Changing World: Using Dynamic Probabilistic Modeling to Assess Future Environmental Emissions of Engineered Nanomaterials. Environmental Science & Technology, 2017. 51(5): p. 2854-2863. 3. Piccinno, F., F. Gottschalk, S. Seeger, and B. Nowack, Industrial production quantities and uses of ten engineered nanomaterials in Europe and the world. Journal of Nanoparticle Research, 2012. 14(9): p. 1-11. 4. Dall'Osto, M., A. Thorpe, D. Beddows, R. Harrison, J.F. Barlow, T. Dunbar, P. Williams, and H. Coe, Remarkable dynamics of nanoparticles in the urban atmosphere. Atmospheric Chemistry and Physics, 2011. 11(13): p. 6623-6637. 5. Grassian, V.H., A.J. Haes, I.A. Mudunkotuwa, P. Demokritou, A.B. Kane, C.J. Murphy, J.E. Hutchison, J.A. Isaacs, Y.-S. Jun, and B. Karn, NanoEHS–defining fundamental science needs: no easy feat when the simple itself is complex. Environmental Science: Nano, 2016. 3(1): p. 15-27. 6. Hendren, C.O., G.V. Lowry, J.M. Unrine, and M.R. Wiesner, A functional assay-based strategy for nanomaterial risk forecasting. Science of The Total Environment, 2015. 536: p. 1029-1037. 7. Arvidsson, R., S. Molander, B.A. Sandén, and M. Hassellöv, Challenges in exposure modeling of nanoparticles in aquatic environments. Human and ecological risk assessment, 2011. 17(1): p. 245-262. 8. Sanderson, K., Sharpest cut from nanotube sword. Nature news, 2006. 10.1038. 9. Yao, N. and Z.L. Wang, Handbook of microscopy for nanotechnology. 2005: Springer. 10. Guo, D., G. Xie, and J. Luo, Mechanical properties of nanoparticles: basics and applications. Journal of Physics D: Applied Physics, 2013. 47(1): p. 013001. 11. Jacoby, M., The future of low-cost solar cells. Chemical and Engineering News, 2016. 94(18): p. 30-35. 12. Buzea, C., I.I. Pacheco, and K. Robbie, Nanomaterials and nanoparticles: Sources and toxicity. Biointerphases, 2007. 2(4): p. MR17-MR71. 13. Li, Y., Y. Zhang, and B. Yan, Nanotoxicity overview: nano-threat to susceptible populations. International journal of molecular sciences, 2014. 15(3): p. 3671-3697. 14. Sahu, S.C. and D.A. Casciano, Nanotoxicity: from in vivo and in vitro models to health risks. 2009: John Wiley & Sons. 15. Ramachandran, G., Assessing nanoparticle risks to human health. 2016: William Andrew. 16. Yeganeh, B., C.M. Kull, M.S. Hull, and L.C. Marr, Characterization of airborne particles during production of carbonaceous nanomaterials. Environmental science & technology, 2008. 42(12): p. 4600-4606. 17. Lowry, G.V., K.B. Gregory, S.C. Apte, and J.R. Lead, Transformations of nanomaterials in the environment. 2012, ACS Publications. 18. Lowry, G.V., B.P. Espinasse, A.R. Badireddy, C.J. Richardson, B.C. Reinsch, L.D. Bryant, A.J. Bone, A. Deonarine, S. Chae, and M. Therezien, Long-term

25 transformation and fate of manufactured Ag nanoparticles in a simulated large scale freshwater emergent wetland. Environmental science & technology, 2012. 46(13): p. 7027-7036. 19. Lowry, G.V., K.B. Gregory, S.C. Apte, and J.R. Lead, Transformations of Nanomaterials in the Environment. Environmental Science & Technology, 2012. 46(13): p. 6893-6899. 20. Wang, Z., L. Zhang, J. Zhao, and B. Xing, Environmental processes and toxicity of metallic nanoparticles in aquatic systems as affected by natural organic matter. Environmental Science: Nano, 2016. 3(2): p. 240-255. 21. Ju-Nam, Y. and J.R. Lead, Manufactured nanoparticles: an overview of their chemistry, interactions and potential environmental implications. Science of the total environment, 2008. 400(1): p. 396-414. 22. Stone, V., B. Nowack, A. Baun, N. van den Brink, F. von der Kammer, M. Dusinska, R. Handy, S. Hankin, M. Hassellöv, and E. Joner, Nanomaterials for environmental studies: classification, reference material issues, and strategies for physico-chemical characterisation. Science of the Total Environment, 2010. 408(7): p. 1745-1754. 23. Aitken, R., P. Borm, K. Donaldson, G. Ichihara, S. Loft, F. Marano, A. Maynard, G. Oberdörster, H. Stamm, and V. Stone, Editorial. Nanotoxicology, 2009. 3(4): p. 263- 264. 24. Handy, R.D. and B.J. Shaw, Toxic effects of nanoparticles and nanomaterials: implications for public health, risk assessment and the public perception of nanotechnology. Health, Risk & Society, 2007. 9(2): p. 125-144. 25. Renn, O., M. Roco, and E. Litten, Nanotechnology risk governance. 2006: International Risk Governance Council Geneva. 26. Linkov, I., F.K. Satterstrom, J. Steevens, E. Ferguson, and R.C. Pleus, Multi-criteria decision analysis and environmental risk assessment for nanomaterials. Journal of Nanoparticle Research, 2007. 9(4): p. 543-554. 27. Linkov, I., J. Steevens, G. Adlakha-Hutcheon, E. Bennett, M. Chappell, V. Colvin, J.M. Davis, T. Davis, A. Elder, and S.F. Hansen, Emerging methods and tools for environmental risk assessment, decision-making, and policy for nanomaterials: summary of NATO Advanced Research Workshop. Journal of Nanoparticle Research, 2009. 11(3): p. 513-527. 28. Hansen, S.F., S.N. Sørensen, L.M. Skjolding, N.B. Hartmann, and A. Baun, Revising REACH guidance on information requirements and chemical safety assessment for engineered nanomaterials for aquatic ecotoxicity endpoints: recommendations from the EnvNano project. Environmental Sciences Europe, 2017. 29(1): p. 14. 29. Luoma, S., F.R. Khan, and M.-N. Croteau, Bioavailability and Bioaccumulation of Metal-Based Engineered Nanomaterials in Aquatic Environments, in Frontiers of Nanoscience. 2014, Elsevier Science. 30. Lynch, I., K. Dawson, J. Lead, and E. Valsami-Jones, Macromolecular Coronas and their importance in Nanotoxicology and Nanoecotoxicology. Frontiers of Nanoscience. Elsevier, Amsterdam, 2014. 31. Hartmann, N.B., M. Ågerstrand, H.-C.H. Lützhøft, and A. Baun, NanoCRED: A transparent framework to assess the regulatory adequacy of ecotoxicity data for nanomaterials–Relevance and reliability revisited. NanoImpact, 2017. 6: p. 81-89. 32. Kuempel, E.D., C.L. Geraci, and P.A. Schulte, Risk assessment and risk management of nanomaterials in the workplace: translating research to practice. Annals of occupational hygiene, 2012. 56(5): p. 491-505. 33. Pu, Y., F. Tang, P.-M. Adam, B. Laratte, and R.E. Ionescu, Fate and Characterization Factors of Nanoparticles in Seventeen Subcontinental Freshwaters: A Case Study on

26 Copper Nanoparticles. Environmental Science & Technology, 2016. 50(17): p. 9370- 9379. 34. Jacobs, R., J.A. Meesters, C.J. ter Braak, D. van de Meent, and H. van der Voet, Combining exposure and effect modeling into an integrated probabilistic environmental risk assessment for nanoparticles. Environmental toxicology and chemistry, 2016. 35(12): p. 2958-2967. 35. Richarz, A.-N., J.C. Madden, R.L.M. Robinson, à. LubiĔski, E. Mokshina, P. Urbaszek, V.E. Kuz, T. Puzyn, and M.T. Cronin, Development of computational models for the prediction of the toxicity of nanomaterials. Perspectives in Science, 2015. 3(1): p. 27-29.

36. Sun, T., D.M. Mitrano, N.A. Bornhöft, M. Scheringer, K. Hungerbuehler, and B. Nowack, Envisioning nano release dynamics in a changing world: using dynamic probabilistic modelling to assess future environmental emissions of engineered nanoparticles. Environmental Science & Technology, 2017. 37. Luoma, S.N., Silver nanotechnologies and the environment. The Project on Emerging Nanotechnologies Report, 2008. 15. 38. Judy, J.D. and P.M. Bertsch, Bioavailability, toxicity, and fate of manufactured nanomaterials in terrestrial ecosystems. Advances in agronomy, 2014. 123: p. 1-64. 39. Hedberg, Y., J. Gustafsson, H.L. Karlsson, L. Möller, and I. Odnevall Wallinder, Bioaccessibility, bioavailability and toxicity of commercially relevant iron-and chromium-based particles: in vitro studies with an inhalation perspective. Particle and fibre toxicology, 2010. 7(1): p. 23. 40. Sáez, A.E. and J.C. Baygents, Environmental transport phenomena. 2014: CRC Press. 41. Handy, R.D., R. Owen, and E. Valsami-Jones, The ecotoxicology of nanoparticles and nanomaterials: current status, knowledge gaps, challenges, and future needs. Ecotoxicology, 2008. 17(5): p. 315-325. 42. Hedberg, Y., S. Pradhan, F. Cappellini, M.-E. Karlsson, E. Blomberg, H. Karlsson, I. Odnevall Wallinder, and J. Hedberg, Electrochemical surface oxide characteristics of metal nanoparticles (Mn, Cu and Al) and the relation to toxicity. Electrochimica Acta, 2016. 212: p. 360-371. 43. Akaighe, N., R.I. MacCuspie, D.A. Navarro, D.S. Aga, S. Banerjee, M. Sohn, and V.K. Sharma, Humic acid-induced silver nanoparticle formation under environmentally relevant conditions. Environmental science & technology, 2011. 45(9): p. 3895-3901.

44. DiBari, G.A., The Electrochemical and Corrosion Behavior of Aluminum. 1970, DTIC Document. 45. Hedberg, Y.S., S. Pradhan, F. Cappellini, M.-E. Karlsson, E. Blomberg, H. Karlsson, I. Odnevall Wallinder, and J.F. Hedberg, Electrochemical surface oxide characteristics of metal nanoparticles (Mn, Cu and Al) and the relation to toxicity. Electrochimica Acta, 2016. 212: p. 360-371. 46. Li, Y., J. Fu, R. Chen, M. Huang, B. Gao, K. Huo, L. Wang, and P.K. Chu, Core–shell TiC/C nanofiber arrays decorated with copper nanoparticles for high performance non- enzymatic glucose sensing. Sensors and Actuators B: Chemical, 2014. 192: p. 474-479.

47. Lee, J.H., J. Hong, J.H. Kim, D. Song, Y.S. Kang, and S.W. Kang, Surface tuned copper nanoparticles by 1-methyl-3-octylimidazolium tetrafluoroborate and its applications to facilitated CO 2 transport. Chemical Engineering Journal, 2014. 235: p. 252-256.

48. Kubota, S., T. Morioka, M. Takesue, H. Hayashi, M. Watanabe, and R.L. Smith, Continuous supercritical hydrothermal synthesis of dispersible zero-valent copper

27 nanoparticles for ink applications in printed electronics. The Journal of Supercritical Fluids, 2014. 86: p. 33-40. 49. Chen, H., Y. Chen, X. Zheng, X. Li, and J. Luo, How does the entering of copper nanoparticles into biological wastewater treatment system affect sludge treatment for VFA production. Water research, 2014. 63: p. 125-134. 50. Chio, C.-P., W.-Y. Chen, W.-C. Chou, N.-H. Hsieh, M.-P. Ling, and C.-M. Liao, Assessing the potential risks to zebrafish posed by environmentally relevant copper and silver nanoparticles. Science of The Total Environment, 2012. 420: p. 111-118. 51. Wang, L.-F., N. Habibul, D.-Q. He, W.-W. Li, X. Zhang, H. Jiang, and H.-Q. Yu, Copper release from copper nanoparticles in the presence of natural organic matter. Water Research, 2015. 68: p. 12-23. 52. De Schamphelaere, K.A., F.M. Vasconcelos, F.M. Tack, H.E. Allen, and C.R. Janssen, Effect of dissolved organic matter source on acute copper toxicity to Daphnia magna. Environmental Toxicology and Chemistry, 2004. 23(5): p. 1248-1255. 53. Aiken, G.R., H. Hsu-Kim, and J.N. Ryan, Influence of dissolved organic matter on the environmental fate of metals, nanoparticles, and colloids. Environmental Science and Technology-Columbus, 2011. 45(8): p. 3196. 54. Liu, X., C. Chen, Y. Zhao, and B. Jia, A review on the synthesis of manganese oxide nanomaterials and their applications on lithium-ion batteries. Journal of Nanomaterials, 2013. 2013(http://dx.doi.org/10.1155/2013/736375). 55. Mukherjee, A., I. Mohammed Sadiq, T. Prathna, and N. Chandrasekaran, Antimicrobial activity of aluminium oxide nanoparticles for potential clinical applications. Science against microbial pathogens: communicating current research and technological advances, 2011. 1: p. 245-251. 56. Zhen, Z. and J. Xie, Development of manganese-based nanoparticles as contrast probes for magnetic resonance imaging. Theranostics, 2012. 2(1): p. 45-54. 57. Hedberg, Y. and K. Midander, Size matters: Mechanism of metal release from 316L stainless steel particles is governed by size-dependent properties of the surface oxide. Materials Letters, 2014. 122: p. 223-226. 58. Ivask, A., I. Kurvet, K. Kasemets, I. Blinova, V. Aruoja, S. Suppi, H. Vija, A. Käkinen, T. Titma, and M. Heinlaan, Size-dependent toxicity of silver nanoparticles to bacteria, yeast, algae, crustaceans and mammalian cells in vitro. PloS one, 2014. 9(7): p. e102108. 59. De Jong, W.H., W.I. Hagens, P. Krystek, M.C. Burger, A.J. Sips, and R.E. Geertsma, Particle size-dependent organ distribution of gold nanoparticles after intravenous administration. Biomaterials, 2008. 29(12): p. 1912-1919. 60. Bantz, C., O. Koshkina, T. Lang, H.-J. Galla, C.J. Kirkpatrick, R.H. Stauber, and M. Maskos, The surface properties of nanoparticles determine the agglomeration state and the size of the particles under physiological conditions. Beilstein journal of nanotechnology, 2014. 5(1): p. 1774-1786. 61. Verma, A. and F. Stellacci, Effect of surface properties on nanoparticle–cell interactions. Small, 2010. 6(1): p. 12-21. 62. Ma, Y. and P.B. Balbuena, Surface properties and dissolution trends of Pt3M alloys in the presence of adsorbates. The Journal of Physical Chemistry C, 2008. 112(37): p. 14520-14528. 63. Karlsson, H.L., P. Cronholm, Y. Hedberg, M. Tornberg, L. De Battice, S. Svedhem, and I. Odnevall Wallinder, Cell membrane damage and protein interaction induced by copper containing nanoparticles—Importance of the metal release process. Toxicology, 2013. 313(1): p. 59-69.

28 64. Hedberg, Y., M. Norell, P. Linhardt, H. Bergqvist, and I. Odnevall Wallinder, Influence of surface oxide characteristics and speciation on corrosion, electrochemical properties and metal release of atomized 316L stainless steel powders. International Journal of Electrochemical Science, 2012. 7(12): p. 11655-11677.

65. Dunphy Guzman, K.A., M.P. Finnegan, and J.F. Banfield, Influence of surface potential on aggregation and transport of titania nanoparticles. Environmental Science & Technology, 2006. 40(24): p. 7688-7693. 66. Min, Y., M. Akbulut, K. Kristiansen, Y. Golan, and J. Israelachvili, The role of interparticle and external forces in nanoparticle assembly. Nature materials, 2008. 7(7): p. 527-538. 67. Pfeiffer, C., C. Rehbock, D. Hühn, C. Carrillo-Carrion, D.J. de Aberasturi, V. Merk, S. Barcikowski, and W.J. Parak, Interaction of colloidal nanoparticles with their local environment: the (ionic) nanoenvironment around nanoparticles is different from bulk and determines the physico-chemical properties of the nanoparticles. Journal of The Royal Society Interface, 2014. 11(96): p. 20130931. 68. Banerjee, A. and M.K. Harbola, Calculation of van der Waals coefficients in hydrodynamic approach to time-dependent density functional theory. The Journal of Chemical Physics, 2002. 117(17): p. 7845-7851. 69. Gatica, S.M., M.W. Cole, and D. Velegol, Designing van der Waals forces between nanocolloids. Nano letters, 2005. 5(1): p. 169-173. 70. Borm, P., F.C. Klaessig, T.D. Landry, B. Moudgil, J. Pauluhn, K. Thomas, R. Trottier, and S. Wood, Research strategies for safety evaluation of nanomaterials, part V: role of dissolution in biological fate and effects of nanoscale particles. Toxicological Sciences, 2006. 90(1): p. 23-32. 71. Hedberg, Y., M.-E. Karlsson, E. Blomberg, I. Odneval Wallinder, and J. Hedberg, Correlation between surface physicochemical properties and the release of iron from stainless steel AISI 304 in biological media. Colloids and Surfaces B: Biointerfaces, 2014. 122: p. 216-222. 72. Hedberg, Y.S. and I. Odnevall Wallinder, Metal release from stainless steel in biological environments: A review. Biointerphases, 2016. 11(1): p. 018901. 73. Hedberg, Y., J. Hedberg, and I. Odnevall Wallinder, Particle characteristics and metal release from natural rutile (TiO2) and zircon particles in synthetic body fluids. Journal of Biomaterials and Nanobiotechnology, 2012. 3(1): p. 37-49. 74. Xing, B., N. Senesi, and C.D. Vecitis, Engineered Nanoparticles and the Environment: Biophysicochemical Processes and Toxicity. Vol. 4. 2016: John Wiley & Sons. 75. Pradhan, S., J. Hedberg, E. Blomberg, S. Wold, and I. Odnevall Wallinder, Effect of sonication on particle dispersion, administered dose and metal release of non- functionalized, non-inert metal nanoparticles. Journal of Nanoparticle Research, 2016. 18(9): p. 285. 76. Stevenson, F.J., Humus chemistry: genesis, composition, reactions. 1994: John Wiley & Sons. 77. Pusey, P., Suppression of multiple scattering by photon cross-correlation techniques. Current opinion in colloid & interface science, 1999. 4(3): p. 177-185. 78. Brown, W., Dynamic light scattering: the method and some applications. Vol. 49. 1993: Oxford University Press, USA. 79. Shaw, D.J. and B. Costello, Introduction to colloid and surface chemistry: Butterworth- Heinemann, Oxford, 1991, ISBN 0 7506 1182 0, 306 pp,£ 14.95. 1993, Elsevier.

29 80. Doane, T.L., C.-H. Chuang, R.J. Hill, and C. Burda, Nanoparticle ȗ-potentials. Accounts of chemical research, 2011. 45(3): p. 317-326. 81. Welz, B., Atomic absorption spectroscopy. 1976: Verlag Chemie. 82. Wojciechowski, C., N. Dupuy, C. Ta, J.-P. Huvenne, and P. Legrand, Quantitative analysis of water-soluble vitamins by ATR-FTIR spectroscopy. Food Chemistry, 1998. 63(1): p. 133-140. 83. Taurozzi, J.S., V.A. Hackley, and M.R. Wiesner, Ultrasonic dispersion of nanoparticles for environmental, health and safety assessment-issues and recommendations. Nanotoxicology, 2011. 5(4): p. 711-729. 84. Cohen, J.M., G.M. DeLoid, and P. Demokritou, A critical review of in vitro dosimetry for engineered nanomaterials. Nanomedicine, 2015. 10(19): p. 3015-3032. 85. Son, J., J. Vavra, and V.E. Forbes, Effects of water quality parameters on agglomeration and dissolution of copper oxide nanoparticles (CuO-NPs) using a central composite circumscribed design. Science of The Total Environment, 2015. 521: p. 183-190.

86. Suslick, K.S. and D.J. Casadonte, Heterogeneous sonocatalysis with nickel powder. Journal of the American Chemical Society, 1987. 109(11): p. 3459-3461. 87. Taurozzi, J.S., V.A. Hackley, and M. Wiesner, Preparation of nanoparticle dispersions from powdered material using ultrasonic disruption. NIST special publication, 2012. 1200(2). 88. Lorimer, J., T. Mason, T. Cuthbert, and E. Brookfield, Effect of ultrasound on the degradation of aqueous native dextran. Ultrasonics Sonochemistry, 1995. 2(1): p. S55- S57. 89. Wang, J., J. Wu, Z. Zhang, X. Zhang, Z. Pan, L. Wang, and L. Xu, Sonocatalytic damage of bovine serum albumin (BSA) in the presence of nanometer anatase titanium dioxide (TiO 2). Ultrasound in medicine & biology, 2006. 32(1): p. 147-152. 90. Hielscher, T., Ultrasonic production of nano-size dispersions and emulsions. arXiv preprint arXiv:0708.1831, 2007. 91. Cohen, J., G. DeLoid, G. Pyrgiotakis, and P. Demokritou, Interactions of engineered nanomaterials in physiological media and implications for in vitro dosimetry. Nanotoxicology, 2013. 7(4): p. 417-431. 92. Hotze, E.M., T. Phenrat, and G.V. Lowry, Nanoparticle aggregation: challenges to understanding transport and reactivity in the environment. Journal of environmental quality, 2010. 39(6): p. 1909-1924. 93. Ji, Z., X. Jin, S. George, T. Xia, H. Meng, X. Wang, E. Suarez, H. Zhang, E.M. Hoek, and H. Godwin, Dispersion and stability optimization of TiO2 nanoparticles in cell culture media. Environmental science & technology, 2010. 44(19): p. 7309. 94. Jun, J.Y., H.H. Nguyen, H.S. Chun, B.-C. Kang, and S. Ko, Preparation of size- controlled bovine serum albumin (BSA) nanoparticles by a modified desolvation method. Food Chemistry, 2011. 127(4): p. 1892-1898. 95. Bihari, P., M. Vippola, S. Schultes, M. Praetner, A.G. Khandoga, C.A. Reichel, C. Coester, T. Tuomi, M. Rehberg, and F. Krombach, Optimized dispersion of nanoparticles for biological in vitro and in vivo studies. Particle and fibre toxicology, 2008. 5(1): p. 14. 96. MacCuspie, R.I., A.J. Allen, and V.A. Hackley, Dispersion stabilization of silver nanoparticles in synthetic lung fluid studied under in situ conditions. Nanotoxicology, 2011. 5(2): p. 140-156. 97. Moore, C., H. Monton, R. O'Kennedy, D. Williams, C. Nogues, C. Crean, and V. Gubala, Controlling colloidal stability of silica nanoparticles during bioconjugation

30 reactions with proteins and improving their longer-term stability, handling and storage. Journal of Materials Chemistry B, 2015. 3(10): p. 2043-2055. 98. Yu, S.-j., Y.-g. Yin, and J.-f. Liu, Silver nanoparticles in the environment. Environmental Science: Processes & Impacts, 2013. 15(1): p. 78-92. 99. Williams, M., Ions, not particles, make silver toxic to bacteria. Rice University. Accessed February, 2013. 12: p. 2013. 100. Ghosh, S., H. Mashayekhi, B. Pan, P. Bhowmik, and B. Xing, Colloidal behavior of aluminum oxide nanoparticles as affected by pH and natural organic matter. Langmuir, 2008. 24(21): p. 12385-12391. 101. Valmacco, V., G. Trefalt, P. Maroni, and M. Borkovec, Direct force measurements between silica particles in aqueous solutions of ionic liquids containing 1-butyl-3- methylimidazolium (BMIM). Physical Chemistry Chemical Physics, 2015. 17(25): p. 16553-16559. 102. Harter, R.D. and R. Naidu, Role of metal-organic complexation in metal sorption by soils. Advances in Agronomy, 1995. 55: p. 219-263. 103. Town, R.M. and H.P. van Leeuwen, Metal ion–humic acid nanoparticle interactions: role of both complexation and condensation mechanisms. Physical Chemistry Chemical Physics, 2016. 18(27): p. 18024-18032. 104. Stumm, W., The inner-sphere surface complex. 1995, ACS Publications. 105. Leelamanie, D. and J. Karube, Effects of hydrophobic and hydrophilic organic matter on the water repellency of model sandy soils. Soil Science and Plant Nutrition, 2009. 55(4): p. 462-467. 106. Luo, M., Y. Huang, M. Zhu, Y.. Tang, T. Ren, J. Ren, H. Wang, and F. Li, Properties of different natural organic matter influence the adsorption and aggregation behavior of TiO2 nanoparticles. Journal of Saudi Chemical Society, 2016, https://doi.org/10.1016/j.jscs.2016.01.007.

31