Ecotoxicological assessment of ZnO nanoparticles to Folsomia … dissertation.pdf · Project CP-FP 247739 (2010-2014) under the 7th Framework Programme of the European Commission

Ecotoxicological assessment of ZnO nanoparticles to Folsomia candida

P.L. Waalewijn-Kool

This Ph.D. thesis focuses on the ecotoxicity and bioavailability of ZnO nanoparticles (ZnO-NP) for a soil-dwelling organism, the springtail Folsomia candida. Different fate and effect studies were performed in natural soils to unravel the contribution of particulate and dissolved Zn to ZnO-NP toxicity. This study shows that the release of toxic Zn2+ ions from ZnO-NP continues for at least one year, but that this does not lead to increased toxicity. This research suggests that ZnO-NP can be evaluated using the current risk assessment of Zn. The studies performed during this Ph.D. project were part of the European project NanoFATE (Nanoparticle Fate Assessment and Toxicity in the Environment).

Eco

toxico

log

ical assessment o

f ZnO nano

particles to

Folsomia cand

ida

P.L. Waalew

ijn-Kool


Pauline Lydia Waalewijn-Kool

This research was conducted in the context of NanoFATE, Collaborative Project CP-FP 247739 (2010-2014) under the 7th Framework Programme of the European Commission (FP7-NMP-ENV-2009, Theme 4).

Thesis 2013-06 of the Department of Ecolocial Science, VU University Amsterdam, The Netherlands

Layout and printing: Off-Page, www.offpage.nl

ISBN: 978-94-6182-324-3

VRIJE UNIVERSITEIT


ACADEMISCH PROEFSCHRIFTter verkrijging van de graad Doctor aan

de Vrije Universiteit Amsterdam,op gezag van de rector magnificusprof.dr. F.A. van der Duyn Schouten

in het openbaar te verdedigenten overstaan van de promotiecommissie

van de Faculteit der Aard- en Levenswetenschappenop donderdag 24 oktober 2013 om 11.45 uur

in de aula van de universiteit,De Boelelaan 1105

doorPauline Lydia Waalewijn-Kool

geboren te Leiderdorp

promotor: prof.dr. N.M. van Straalencopromotor: dr.ir. C.A.M. van Gestel

“Ik heb u tot een keurmeester van metalen gemaakt, zodat u het erts kunt testen en de waarde ervan kunt bepalen.” Jeremia 6:27.

Het Boek, oktober 2009

Table of contents

Chapter 1 General introduction 9

Chapter 2 Effect of different spiking procedures on the distribution 25 and toxicity of ZnO nanoparticles in soil

Chapter 3 Chronic toxicity of ZnO nanoparticles, non-nano ZnO and 39 ZnCl2 to Folsomia candida (Collembola) in relation to bioavailability in soil Supporting Information 52

Chapter 4 Sorption, dissolution and pH determine the long-term 57 equilibration and toxicity of coated and uncoated ZnO nanoparticles in soil Supporting Information 69

Chapter 5 The effect of pH on the toxicity of ZnO nanoparticles to 84 Folsomia candida in amended field soilSupporting Information 98

Chapter 6 Effect of soil properties on the toxicity of ZnO nanoparticles 107 to Folsomia candida in a comparison of four natural soilsSupporting Information 120

Chapter 7 Summary and discussion 129

Bibliography 145

Appendix I Summary of ecotoxicity studies with ZnO nanoparticles 156(ZnO-NP) and terrestrial invertebrates

Appendix II Letter to the Editor. Metal-based nanoparticles in soil: 158New research themes should not ignore old rules and theories

Samenvatting 163

Acknowledgements 167

Curriculum vitae 169

SENSE Certificate 171

1General introduction

1Scope

Nano is a prefix meaning extremely small. When quantifiable, it translates to one-billionth. The term “nano-technology” was first used by Norio Taniguchi in 1974 and refers to the engineering of functional systems at the molecular scale (Maynard, 2006). Nowadays different types of zinc oxide nanoparticles (ZnO-NP) can be purchased on the commercial market with the potential of being released in the environment (Gottschalk and Nowack, 2011). Although the size-dependent properties make engineered nanoparticles (ENPs) desirable for many applications, their release into the environment and implications for the environmental fate and effects need to be investigated (Klaine et al., 2008). This thesis focuses on how ZnO-NP reacts in natural soils and how toxic these particles are for a soil dwelling organism, Folsomia candida. In this Chapter I introduce ZnO-NP as one of the most widely used ENPs and their route into the soil environment. Then, I describe the fate processes that tend to occur when ZnO-NP are in the soil and how these processes relate to soil properties. An overview of available terrestrial studies is shown, indicating adverse effects of ZnO-NP to soil organisms. And an outline of the thesis is given in which the investigated aspects of the fate and effects assessment of ZnO-NP are addressed.

2 Engineered nanoparticles

ENPs exhibit novel properties with specific and improved functionality and increased efficiency (Navarro et al., 2008). The physico-chemical properties of ENPs are attributable to their small size (surface area and size distribution), chemical composition (purity, crystallinity, electronic properties), surface structure (surface reactivity, surface groups, inorganic or organic coatings), solubility, shape, and aggregation (Nel et al., 2006; Hassellöv et al., 2008). The surface area-to-volume ratio is a function of particle size, with smaller particles having a larger surface area per weight than larger particles (Bottero et al., 2011). Their higher surface to volume ratio makes ENPs potentially more reactive than larger particles. Approximately 35-40% of the atoms are localized at the surface of a 10 nm nanoparticle compared to less than 20% for a particle larger than 30 nm (Auffan et al., 2009).

ENPs are used in a large variety of industrial and consumer products, such as medical devices, pharmaceuticals, cosmetics, electronics, textiles, food packaging, fuel catalysts, biosensors and for environmental remediation. The use of ENPs has grown dramatically in the last decade and will continue growing in the coming years (Aitken et al., 2006; Royal Society, 2004). The commercially useful properties of ENPs have resulted in a large range of different ENPs. They can be classified based on their chemical composition into carbon-containing (organic) and metal-based (inorganic)

11

1

General introduction

ENPs (Nowack et al., 2007; Peralta-Videa et al., 2011). Inorganic ENPs include metal

and metal oxide nanoparticles such as silver (Ag), which are widely incorporated into

products for their antibacterial and antifungal properties; titanium oxide (TiO2) and

zinc oxide (ZnO), which are used in sunscreens for their ultraviolet (UV) absorbance

and reflecting capability; and cerium oxide (CeO2), which is used in fuel additives as a

catalyst. Mixtures of different phases are also manufactured (e.g. cadmium selenide).

ENPs can be produced by a huge range of procedures which can be grouped into

top-down and bottom-up strategies (Ju-Nam and Lead, 2008). Top-down approaches

are defined as those by which ENPs are directly generated from bulk materials using

physical methods such as milling, attrition or repeated quenching. Bottom-up strategies

involve molecular components as starting materials linked with chemical reactions,

nucleation and growth processes (Christian et al., 2008). ENPs can be produced with

different shapes or structures; they can be spherical, tubular or irregularly shaped and

can have a functionalized surface or a coating (Nowack et al., 2007).

ENPs are defined as manufactured particles with sizes smaller than 100 nm in one

or more dimensions. The cut-off value of 100 nm is considered the working definition

of the International Organization for Standardization (ISO, 2008) and used by many

researchers working with ENPs nowadays (e.g. Handy et al., 2008; Heinlaan et al., 2008;

Unrine et al., 2010). As ENPs rarely have one size after release into the environment,

it is more realistic to consider ENPs with a distribution of particle sizes around the

nanoscale (Handy et al., 2008). The European Union recommends considering particle

size distributions in the definition of ENPs instead of one particle size only. This

means that when 50% of the total number of particles in a certain material has one

or more dimensions smaller than 100 nm, all material is considered as nanomaterial.

The definition of the European Union for nanomaterials reads: “Nanomaterial means a

natural, incidental or manufactured material containing particles, in an unbound state

or as an aggregate or as an agglomerate and where, for 50 % or more of the particles in

the number size distribution, one or more external dimensions is in the size range 1-100

nm.” (EU, 2011). The definition of ENPs is still under debate among environmental

scientists to establish a scientific and practical distinction between “nano” and their

“bulk” material or “non-nano” counterparts (Auffan et al., 2009b; Bleeker et al., 2012).

ZnO nanoparticles

ZnO-NP are among the most commonly used metal oxide nanoparticles. Piccinno et al.

(2012) reported that ZnO-NP was produced in global quantities between 100 and 1000

t/year, with two out of fifty European producers manufacturing ZnO-NP above 100 t/year

(max 100000 t/year). Different morphological structures of ZnO-NP can be manufactured,

such as nanorods, nanocrystals, nanobelts and nanowires (Wang et al., 2004). The

characteristics of ZnO-NP lend them great potential for UV-absorbing applications,

because they mobilize electrons within their atomic structure while absorbing UV radiation

(Wolf et al., 2003). ZnO-NP exhibit a wide band gap (3.37 eV), which means that electrons

12

Chapter 1

are more mobile, and high exciton binding energy (60 meV) (Wang et al., 2004; Song et al., 2011a). These physical properties in ZnO-NP make them able to absorb almost the whole UV spectrum (200-400 nm) and can ensure efficient excitonic emission.

ZnO-NP is incorporated in UV and antimicrobial coatings (Piccinno et al., 2012). Such particles are also used in personal care products such as anti-ageing creams, hand creams and sunscreens (Gulson et al., 2010; Pinnell et al., 2000). Sunscreens containing physical UV-blockers, such as ZnO and titanium dioxide (TiO2) provide effective protection against UV-induced DNA damage in cells (Cayrol et al., 1999). In normal pigment size ranges (150-300 nm for TiO2 and 200-400 nm for ZnO) these particles reflect and scatter visible light, making the sunscreens appear white. However, as particle sizes decrease to nano dimensions (typically between 20 and 150 nm for TiO2 and 40-100 nm for ZnO) they absorb and scatter UV radiation, making the sunscreens appear transparent on skin and thus more aesthetically pleasing (Osmond and McCall, 2010).

Available information suggests that the nanoparticles of zinc oxide do not enter the body through the skin (Gamer et al., 2005; Song et al., 2011a). It also suggests that the amount of ionized zinc, which may be released from zinc oxide nanoparticles and enter the body through the skin, is likely to be insignificantly small. ZnO-NP can be considered safe for use on the skin of humans and are approved as sunscreen ingredient in the majority of countries around the world. However, from an environmental risk assessment point of view, insufficient assessments are available for regulators to adequately control potential hazard associated with the environment (Nowack and Bucheli, 2007). Potential release and dissolution of ZnO-NP poses a risk for organisms living in soil, sediments and water (Navarro et al., 2008). As more ZnO-NP will be manufactured and commercially available, the fate and effects of ZnO-NP on our environment need to be investigated.

3 Environmental exposure of ZnO nanoparticles

ZnO-NP are likely to enter the environment due to their increased use and disposal. The main entry into the environment is through the wastewater system, although there is also potential for wash-off during swimming, entry into landfills and incineration of discarded bottles. ZnO-NP that are released in wastewater will partition between final effluent (discharged to rivers or sea) and sewage sludge (spread to land or sea). Soil is a potential sink for ZnO-NP when sewage sludge is used for land application (e.g. fertilizer). Nowadays, sludge application on land is applied in most European countries, although in The Netherlands and Switzerland this is currently prohibited. Once in the environment, ZnO-NP are likely to come into contact with a range of aquatic and terrestrial species, causing potential hazard to organisms living there.

Concentrations of ZnO-NP have rarely been measured in the environment and transformations of ZnO-NP may also lead to increased levels of other Zn forms in water and soil (Lowry et al., 2012; Nowack et al., 2012). It is extremely difficult to determine

13

1


increased environmental levels of anthropogenic Zn levels from ZnO-NP usage (Stone et al., 2010). Modelling data has been published on the potential release of ZnO-NP in the environment, with current estimates of ZnO-NP concentrations in the UK can reach up to 76 µg/l in water and 3194 µg/kg in soil (Tiede et al., 2009). Gottschalk et al. (2009) estimated predicted environmental concentrations (PEC) of 430 ng/l for treated wastewater and a PEC of 10 ng/l for surface water. For soils an annual increase of ZnO-NP concentration of 0.093 µg/kg/y was estimated and this increase was 3.25 µg/kg/y in case of sludge application on land (Gottschalk et al., 2009). For UK a realistic worse-case scenario of 22.6 µg/kg soil was estimated for agricultural land with sludge application (data provided by V. Keller, Centre for Ecology and Hydrology, UK).

4 ENP characterization

Increased usage of ZnO-NP may lead to increased environmental exposure levels in future (Scown et al., 2010). Various techniques are available to characterize ENPs in environmental media. Each method has its limitation in applicable size and concentration ranges. For chemists it is an enormous challenge to measure samples for environmentally relevant concentrations (ng/l, pg/l), because detection limits are not sufficiently low (Jiang et al., 2009; Tiede et al., 2009). Environmental samples also contain a high background of natural and unintentionally produced nanoparticles (Hassellöv et al., 2008). Appropriate characterization techniques for ENPs in soil do not exist (Lead and Wilkinson, 2006).

For relatively simple analyses, such as the aggregation state of ENPs in aqueous solution, it is possible to use dynamic light scattering (DLS) or microscopy-based techniques such as scanning and transmission electron microscopy (SEM/TEM). DLS measurements are based on the relationship between a particle’s diffusion coefficient in solution and its size-dependent Brownian motion, i.e. the random moving of particles by collisions with water molecules. Particle motion is calculated by time-integrated measurements of incidences of light scattering within a sample. From this the particle size and size distributions can be derived. Electron microscopy represents an important technique for directing viewing ENPs at their original domain sizes. Primary particle size, shape and size distributions can be determined directly from SEM and TEM images using digital processing or by scoring images manually. TEM allows for the highest magnification of ENPs sized structures but solely from a two-dimensional perspective, and is best applied to electron dense materials, such as ZnO-NP (Handy et al., 2008). In SEM the interaction of the electron beam with the particle surface are scanned over the sample. Due to high depth of field in SEM a three-dimensional appearance can be obtained. Most of the drawbacks from the use of EM arise from artefacts due to drying and analysing under ultra-high vacuum. Also aggregation and distortion are complicating factors for interpreting images.

14

Chapter 1

Ultrafiltration (UF) is a membrane filtration in which hydrostatic pressure forces a

liquid against a semi permeable membrane. Suspended solids and solutes of high

molecular weight are retained, while water and low molecular weight solutes pass

through the membrane. This separation process is used in industry and research for

purifying and concentrating macromolecular (103 - 106 Da) solutions, especially protein

solutions. Nowadays it also applied in nanotoxicology to obtain particle-free extracts

(Hassellöv et al., 2008).

5 Environmental fate of metal-based ENPs in soil

The physico-chemical characteristics of metal-based ENPs are believed to determine

their environmental fate in soil and subsequently their bioavailability and toxicity (Crane

et al., 2008). Understanding the processes that control the stability of ENPs is central to

understanding their fate and bioavailability (Darlington et al., 2009). Once into the soil, ENPs

lose their pristine particle state and complex transformation processes occur. Environmental

fate processes that are most likely to occur in soils are aggregation/agglomeration, sorption

to surfaces and dissolution to the ionic metal (Tourinho et al., 2012).

Soil represents a relatively complex medium for the understanding of the physico-

chemical behaviour of ENPs. In comparison with aqueous solutions, in which aggregation

behaviour can be understood largely in terms of particle stability, soils present a solid

phase with which ENPs may interact, as well as with the dissolved (porewater) phase.

Environmental transformation processes of metal-based ENPs in soils are presented

in Figure 1 as a conceptual illustration. All scenarios are likely to occur in the soil

environment and this means that organisms may be exposed to different forms and

sizes of ENPs. ENPs can be present in the soil as pristine nanoparticles, as aggregates/

agglomerates of ENPs, as dissolved ENPs or as free metal ions. The binding of free

ENPs or aggregated ENPs to natural organic matter or other particles present in soils

is defined as sorption. The dissolved ENPs and free metal ions in the pore water

are considered bioavailable to soil organisms and are therefore important forms of

ENP emission under investigation in environmental research. In the pore water surface

modification, complexation and precipitation may occur due to interaction with other

organic and inorganic components of the soil.

Aggregation and agglomerationParticle aggregation refers to formation of clusters in a colloidal suspension and leads

to destabilization of colloidal systems (Verwey and Overbeek, 1948). In solution, the

interactive forces between the surfaces of two colloid particles are described as the

sum of attractive van der Waals forces and repulsive electrostatic forces. Aggregation

is defined as the association of primary particles by strong bonding, whereas

agglomeration is defined as association by weak bonding caused by van der Waals

15

1


forces (EU, 2011). Aggregated ENPs will have a lower surface charge than individual particles because of the lowering of the electrostatic repulsive forces between particles.

Surface coating Once particle aggregates have formed in solutions, they will grow in size and settle to the bottom of a test container. To obtain stable ENPs solutions, it is therefore necessary to stabilize the dispersion of ENPs by providing a barrier (“coating”) to close approach of two particles (Christian et al., 2008, Ju-Nam and Lead, 2012). Coated nanoparticles typically have very different surface charges compared to uncoated nanoparticles, which can modify the environmental fate processes of ENPs in soil. The stability of ENP coatings is important in determining how long the particles maintain the primary engineered surface properties. A surface coating of ENPs may degrade or change over time so it may also be necessary to consider the fate and effects of coated ENPs. In addition, the coating itself may attract and bind specific chemicals from the environment. A coating is likely to influence aggregation and dissolution of the core material of the ENP, because such a layer may prevent the release of metal ions (Gottschalk et al., 2009). Nowadays, the majority of the ENPs are produced with

Dissolution

Ionic forms

Sorption

Soil

Pore water Solid

Aggregation/agglomeration

Com

plex

atio

nPr

ecip

itatio

n

Inte

ract

ion

with

org

anic

and

inor

gani

c soi

l com

pone

nts

Surfa

ce m

odifi

catio

n

DOC

Cl-

Mg2+

Ca2+

HCO3-

Na+

H+

“Pristine nanoparticle”

Figure 1. A conceptual overview of the environmental fate processes of metal-based ENPs in soil (aggregation/agglomeration, dissolution and sorption). The interaction of aggregated, free and dissolved ENPs in the pore water with other organic and inorganic components may lead to surface modifications, complexation and precipitation. All processes depend on ENP properties and soil properties.

16

Chapter 1

a surface coating, but limited data is available on the difference in fate and effects between coated and uncoated ZnO-NP in environmental media. So, it is needed to compare the fate and effects of coated and uncoated ZnO-NP to gain further insights in the influence of a NP coating on the dissolution and toxicity of ZnO-NP. In this thesis the coating material was also tested separately from the ZnO-NP for its toxicity in soil.

DissolutionSome types of metal-based ENPs, like ZnO-NP, are thermodynamically unstable and undergo chemical dissolution, which theoretically means that a metal ion detaches from the particle and migrates into the solution (Auffan et al., 2009; Misra et al., 2012; Borm et al., 2006). Such ionic metal species may be toxic themselves and a very crucial issue is to unravel if metal-based ENP toxicity is caused by soluble metal species derived from the ENPs or by the ENPs themselves. Thus, the extent of dissolution and the relative toxicities of perhaps the novel properties of the ENPs, or a combination of both needs to be investigated in order to better understand potential toxic effects.

It is generally assumed that smaller ENPs, with higher surface area, may have increased release of metal ions than larger particles (Borm et al., 2006). Different aquatic studies show that ENPs dissolve faster than larger sized materials of the same mass (see e.g. Wong et al., 2010; Reed et al., 2012), but this was not found for ZnO-NP in soil (Milani et al., 2010). Relatively fast dissolution of ZnO-NP has been observed in algal test media (Franklin et al., 2007; Aruoja et al., 2009), in natural water (Blinova et al., 2010; Poynton et al., 2011) and in kaolin suspensions (Scheckel et al., 2010). ZnO-NP dissolution most probably results in the formation of Zn2+, Zn(OH)+ and Zn(OH)3

- which are the dominant species in water at neutral pH (Bian et al., 2011). The number of studies on ENP dissolution in soil is very limited because of

the technically challenging nature of such work (Unrine et al., 2010). Data on the dissolution of ZnO-NP in soil will be useful to predict its bioavailability and toxicity to F. candida with time. Long-term studies on ZnO-NP dissolution are necessary, because it is unknown whether it is a slow or fast process. After continuous release of ZnO-NP into the environment, and subsequent transformation and dissolution of ZnO-NP, this may cause not only addition of Zn but also a mixture effect of different Zn forms, as illustrated in Figure 1.

Soil propertiesThe physico-chemical properties of the ENPs have a major influence on their environmental fate (Li et al., 2010, Li et al., 2013; Song et al., 2011b). But also the characteristics of the soil, such as pH and organic matter content are considered to play an important role in the environmental fate of ENPs. A lot of knowledge has been gained on the bioavailability of conventional metals in soils. Partition coefficients can describe the soil-solution interaction of zinc in soil (Brümmer et al., 1983), with smaller values corresponding to high metal concentrations in the pore water and higher values stronger sorption to the solid phase (McLaughlin, 2002). It is likely that soil pH and

17

1


natural organic matter content play an important role in determining ENPs speciation, solubility, movement, and eventual bioavailability (Kanel et al., 2011).

It is expected that a lower soil pH will increase ZnO-NP bioavailability due to lower sorption of ZnO-NP. At a lower pH the dissolution of ZnO-NP is expected to increase, which leads to more soluble Zn in solution and subsequently higher bioavailability. Natural organic matter (NOM) is also a major contributor to the ability of soils for retaining metals in an exchangeable form (Zeng et al., 2011). Negatively charged humic and fulvic acids may attract positively charged ZnO-NP and Zn2+ and may subsequently affect dissolution rates of ENPs in soil (Keller et al., 2010). In aquatic studies the bioavailability of ENPs is diminished in the presence of NOM due to the fact that NOM can coat the NP surface (Baalousha et al., 2008; Bian et al., 2011; Quik et al., 2010). For a better understanding of ENP behaviour as a function of soil properties, studies on multiple soils with different pH levels or organic matter content are necessary.

6 Ecotoxicity

Methodological approach for ENP toxicity testingThe study of ecotoxicity is the toxicology of a substance towards a range of organisms or towards ecosystems and involves knowledge of toxicology, chemistry and ecology (Truhaut, 1977). ENPs are a new class of synthetic pollutants and potential ecotoxicological effects need to be studied (Cattaneo et al., 2009). Testing ENPs in standardized laboratory tests may require adaptations to the existing test guidelines (Handy et al., 2012; Tiede et al., 2009). One of the adaptations would be the spiking procedure, which is used for the introduction of ENPs in environmental test media. Due to dissolution and aggregation test organisms may be exposed to other forms or sizes than primary ENPs. ENP properties are recommended to be evaluated in ecotoxicological studies, preferably at the beginning and at the end of the study, to take into consideration the possible transformations of ENPs in environmental media (Stone et al., 2010; Nel et al., 2006; Tiede et al., 2009). Aggregation may cause a non-homogeneous distribution of ENPs in the test media. It needs to be investigated whether aggregations and other transformations will influence the outcome of a toxicity test. Developing standardized spiking procedures for ENPs is needed to improve comparability between multiple soil toxicity tests. A range of spiking methods have been proposed for preparing test media including the use of surfactants and solvents, sonication, filtration and stirring for prolonged periods and pH manipulations (Tiede et al., 2009). To learn what type of application method is suitable for ENPs in soils, different spiking procedures need to be evaluated.

Another issue to be investigated for ENP testing is the expression of the dose and the quantification of ENPs in environmental media. If the nanoparticles themselves are causing an effect, it would be good to express the dose on the actual number

18

Chapter 1

of particles or on the total surface area of the particles. Nowadays, these two dose

metrics are used in human toxicology (Oberdörster et al., 2005), but they seem to

be impracticable for terrestrial ecotoxicology. It is extremely difficult to quantify the

total number of nanoparticles and their surface area in soil, due to interference of

soil particles which have a sorption surface area themselves. For metals in general,

verification of exposure concentrations in environmental media is required in

ecotoxicological studies. It needs to be investigated whether standard analytical

methods used for quantifying metal concentrations are applicable for metal-based

ENPs. When this is the case, exposure concentrations could be expressed on the basis

of the main component of the nanoparticles, e.g. mg ZnO-NP/kg dry soil or on the

bioavailable form of the nanoparticles, e.g. mg Zn/kg dry soil. Especially in case of

metal-based ENPs, the latter expression would be preferred as it also would enable

comparison with data obtained from studies with non-nano metal or with other salts

of the same metal. In order to enable relating potential ecotoxicity to the (nano) size

of ENPs, ecotoxicity data of ENPs needs be compared with a non-nano material with

similar composition. Testing a soluble salt of the metal of interest could unravel the

contribution of the soluble metal fraction to nanoparticle toxicity.

ZnO-NP ecotoxicity and mechanisms

Most of the ecotoxicity studies with ZnO-NP have been performed with aquatic species

and ecotoxicity studies with terrestrial species are limited (Kahru and Dubourguier,

2010). Also, the majority of the effect data that is available are from acute studies,

while chronic ecotoxicity data is lacking. Sub-lethal effects of ZnO-NP were found

for bacteria (Adams et al., 2006; Huang et al., 2008; Li et al., 2011a; Li et al., 2013;

Dimpka et al., 2011), crustaceans (Wiench et al., 2009; Heinlaan et al., 2008; Poynton

et al., 2011), algae (Franklin et al., 2007; Aruoja et al., 2009) and fish (Amutha and

Subramanian, 2009; Bai et al., 2010; Hao et al., 2012; Johnston et al., 2010). Some

of these studies suggest relatively high acute toxicity of ZnO-NP (in the low mg/l

levels), although the toxicity was highly dependent on the test species, the physico-

chemical properties of the ZnO-NP, test methods and test media properties. Particle

dissolution to ionic Zn and particle-induced generation of reactive oxygen species

(ROS) represent the primary modes of action for ZnO-NP toxicity in aquatic studies

(Ma et al., 2013). No major differences between the L(E)C50 for ZnO-NP and non-nano

ZnO were found in most of these organisms, suggesting that particle size did not

influence toxicity (Kahru and Dubourguier, 2010).

At this moment eight studies describe effects of ZnO-NP for different soil

invertebrates. Appendix I provides a summary of each study, including the test

organism, Zn compound(s) and size, test concentration, test media, exposure time,

outcome of the test and the main conclusion of the authors.

Tests were conducted with nematodes, earthworms, springtails and isopods, using

different test media, which makes it difficult to compare the outcome of the studies.

19

1


Several studies did not include a full dose-response curve, i.e. a range of ZnO-NP test

concentrations. With only a few test concentrations an EC50 value for the effect on

reproduction, that is the test concentration inducing 50% reduction on reproduction,

cannot be established. EC50 values are mostly used in risk and hazard assessment as

indicators for toxicity. In soil, ZnO-NP reduced earthworm reproduction by approx.

50% at 750 mg Zn/kg (Hooper et al., 2011) and reproduction ceased at 1000 mg/kg

(Canas et al., 2011). Manzo et al. (2011) determined the toxicity of ZnO-NP to the

springtail F. candida in OECD artificial soil at a single dose (230 mg Zn/kg) and did

not observe effects on reproduction at this level. Currently, it is unclear at what

concentration ZnO-NP are toxic for springtails. In-depth chronic toxicity studies of

ZnO-NP are necessary to resolve potential particle effects of ZnO-NP in soil and to

investigate whether toxicity is related to the soluble Zn fraction. Effects of the Zn salt

ZnCl2 to Folsomia candida have extensively been studied in soil. The 28-day EC50 in

Lufa 2.2 natural soil was estimated to be in between 348 mg Zn/kg dry soil (Smit et al.,

1996) and 476 mg Zn/kg dry soil (Nota et al., 2010).

Soil propertiesA negative correlation between metal bioavailability/toxicity and soil pH has been well

documented for the soil organism F. candida. Sandifer and Hopkin (1996) reported 28-d

EC50 values for zinc (added as Zn(NO3)2) in artificial soil of 590, 600 and 900 mg Zn/kg

dry soil at a pH of 4.5, 5 and 6, respectively. Crommentuijn et al. (1997) observed the

same relationship with soil pH for the 35-d EC50 of F. candida exposed to cadmium.

The influence of soil pH on the bioavailability and toxicity of ZnO nanoparticles to

F. candida has not been investigated for springtails. It is expected that a lower pH

level will increase ZnO-NP bioavailability and toxicity to soil organisms. No influence

of natural organic matter content in soil could be demonstrated for Zn toxicity to F.

candida (Vijver et al., 2001). Also, the organic matter content played no significant

role in grassland or heathland on sandy podzolic soil (range 2.2-8.6) on Zn uptake

by the earthworm Lumbricus rubellus (Ma et al., 1983). The influence of NOM on

ZnO-NP needs to be investigated for F. candida. It is expected that higher OM content

increases the aggregation and reaction of ZnO-NP with humic and fulvic acids, thereby

reducing Zn bioavailability and toxicity.

7 Test organism Folsomia candida

F. candida is considered as a highly valued model organism for soil. Springtails

(Collembola) have been used widely to assess the environmental impact of a range of

pollutants on soils due to their abundance and diversity (Fountain and Hopkin, 2001,

2005). Also, a standardized toxicity test has been developed by the Organisation

for Economic Co-operation and Development (OECD, 2009), i.e. “Test No. 232:

20

Chapter 1

Collembolan Reproduction Test in Soil” and by the International Organization for Standardization (ISO, 1999), i.e. Guideline 11267 “Soil quality - inhibition of reproduction of Collembola (Folsomia candida) by soil pollutants”.

Springtails are soil-living invertebrates frequently found in leaf litter and other decaying material, where they are primarily detritivores and microbivores. The springtail F. candida is a parthenogenetic species. The parthenogenesis and high rate of reproduction make F. candida suitable for studying the effects of pollutants. The adults of F. candida can produce 14 to 29 eggs with each oviposition. The eggs usually take an average of 7-10 days to hatch. F. candida has a mild preference for soils with a pH between 5 and 6, at which they achieve their highest reproduction (Fountain and Hopkin, 2005; Greenslade and Vaughan, 2003). An adult female may go through 45 molts in her lifetime. During molting the complete gut epithelium is refreshed and most likely, in this way also metals will be excreted from the body, as demonstrated for another springtail species, Orchesella cincta (Sterenborg et al., 2003). The body of F. candida consists of a head, three thoracic segments and six abdominal segments. From the first abdominal segment ventrally down projects the ventral tube. The ventral tube and the cuticle are the main exposure route for metals to enter the organism. The ventral tube is involved in fluid exchange with the external environment (Fountain and Hopkin, 2005). For partially-soluble nanoparticles such as ZnO-NP under investigation it is assumed that only the dissolved fraction is considered bioavailable for the springtail.

Zinc is an essential element in springtails and involved in the synthesis of nucleic acids and enzymes (Williams, 1984). Zinc is regulated to a more or less constant level in most animal species and in crustaceans the zinc level is approx. 70 µg/g dry weight (Depledge and Rainbow, 1990; Depledge, 1989). When internal zinc concentration exceeds a certain threshold, zinc toxicity usually involves inhibition of enzyme activity, disruption of cell membrane integrity and competition with other essential cations. Metal-binding proteins, called metallothioneins are involved in regulation and detoxification of metals. Nota et al., (2011) showed that the expression of the gene encoding a metallothionein-like protein was induced after springtail exposure to ZnCl2. The binding of zinc to metallotheionein in F. candida is not completely clear. There are indications that in invertebrates zinc may bind to a specific peptide called phytochelatin (Brulle et al., 2008; Sterenborg et al., 2003).

8 Aim and outline of the thesis

In summary, there is a need for research on the bioavailability and ecotoxicity of ZnO-NP to soil and water organisms. This research has been conducted within the framework of NanoFATE (Nanoparticle Fate Assessment and Toxicity in the Environment), a FP7-EU-funded integrated research project that aims to fill knowledge and methodological gaps of environmental risks posed by ENPs. Acute and chronic ecotoxicological

21

1


tests with ZnO-NP were carried out for a variety of water and soil species (algae,

nematodes, mussels, water fleas, isopods, earthworms and springtails). The three main

research components of NanoFATE are: 1) The particle chemistry and fate component

is centred on the development of production of ENPs and the use of advanced

analytical and image analysis techniques for tracking of these particles in different

environmental compartments; 2) The ecotoxicology and bioavailability component

focuses on assessing the ecotoxicity of selected ENPs including toxicokinetic and

toxicodynamic aspects; 3) The risk assessment and communication component

focuses on hazard characterisation, risk assessment, life cycle assessment and risk

communication of ENPs. The Animal Ecology group of the VU University Amsterdam

is leading and participating in the research program of the second component, which

comprises research within three work packages dealing with 1) ecotoxicology, 2)

bioavailability and relation between soil and water chemistry and particle properties

and 3) toxicokinetics and toxicodynamics.

The aim of this Ph.D. thesis was to perform a full fate and effect assessment of ZnO-NP

for the springtail F. candida in natural soils and to unravel the contribution of particulate

Zn and dissolved Zn to ZnO-NP toxicity. Long-term effects on the environmental fate

and bioavailability of ZnO-NP in soil were studied and ZnO-NP dissolution in soil was

determined under laboratory conditions. In order to link the ecotoxicity of ZnO-NP

to the environmental fate (dissolution and sorption), the influence of soil properties

was determined. Figure 2 provides a schematic overview of my research, including

different Zn forms, different soils regarding pH and organic matter content, pore water

phase and the model organism F. candida. To systemically test all ZnO nanoparticles

available on the commercial market was clearly not feasible, therefore three types of

ZnO-NP were selected, namely BASF products: Z-COTE® and Z-COTE® HP1 (coated

ZnO-NP with triethoxyoctylsilane) and ZnO Nanosun. In all experiments, I used the

springtail F. candida as a test organism to address the following research questions.

How does the method of soil spiking influence the ZnO-NP distribution in soil? Can the

presently used analytical techniques be applied to quantify Zn concentrations in soil?

Are ZnO-NP toxic to springtails? Are the effects size-related or related to dissolved

Zn? What is the effect of ageing on ZnO-NP dissolution and toxicity? Behave coated

ZnO-NP different from uncoated ZnO-NP with respect to dissolution and toxicity? And

what is the influence of soil properties on the dissolution and toxicity of ZnO-NP?

Research was carried out in different steps described below:

1. The applicability of two spiking procedures of ZnO-NP were investigated to obtain

an as homogenous ZnO-NP distribution in soil as possible (Chapter 2). In addition,

the spiking solution was characterized using Dynamic Light Scattering. Toxicity

of ZnO-NP was determined using standardised ecotoxicological testing methods.

2. The toxicity of ZnO-NP was determined in a 28-day toxicity test by studying the

survival and reproduction of F. candida. The EC50 values for ZnO-NP, non-nano

22

Chapter 1

ZnO and ZnCl2 were compared to determine the contribution of size-related effects and the influence of dissolved Zn to ZnO-NP toxicity in soil (Chapter 3).

3. Long-term fate and effects of coated and uncoated ZnO-NP were determined by spiking a large batch of soil with ZnO-NP, non-nano ZnO and ZnCl2 at different concentrations. Soils were equilibrated for one year in the laboratory and dissolution and toxicity were assessed at different time points (Chapter 4).

4. The effect of soil pH on ZnO-NP toxicity was determined in pH-amended field soil from the United Kingdom. Three soils with different pH levels (i.e. 4.5, 5.9 and 7.2) were tested in a 28-day toxicity test with F. candida (Chapter 5). The effect of soil properties on ZnO-NP dissolution and toxicity was further investigated using four natural soils, varying in natural organic matter content and pH (Chapter 6).

In the final chapter the results obtained in previous chapters are summarized and discussed.

Equilibrium times: 0, 3, 6, 12 months

Soils (mg Zn/kg)

pH amended: 4.5, 5.9, 7.2

Pore water (mg Zn/l)

Toxicity Folsomia candida

Centrifugation and 0.45 µm filtration

EC50 = f (pH, Zn2+, % OM)

coated ZnO‐NP

non‐nano ZnO

ZnCl2

uncoated ZnO‐NP

Dry Suspension

ventral tube

Ultrafiltration(3/100 kDa)

%OM: 2.4, 3.1, 10.6, 14.7

Lufa 2.2

Spiking?

Dissolution?

Soil properties?Size‐related orionic effect?

Effect ENP coating?

Figure 2. Schematic overview of the research performed for this thesis. Investigations revolved around bioavailability and toxicity of ZnO-NP, including different Zn forms, different soils regarding pH and OM content, equilibrium with the pore water and the link to the ecotoxicity towards Folsomia candida. Research questions are written in key words in balloons.

23

1


2Effect of different spiking procedures on the distribution and toxicity

of ZnO nanoparticles in soil

Abstract

Due to the difficulty in dispersing some engineered nanomaterials in exposure media, realizing homogeneous distributions of nanoparticles in soil may pose major challenges. The present study investigated the distribution of zinc oxide nanoparticles (30 nm) and non-nano zinc oxide (200 nm) in natural soil using two different spiking procedures, i.e. as dry powder and as suspension in soil extract. Both spiking procedures showed a good recovery (> 85%) of zinc and based on total zinc concentrations no difference was found between the two spiking methods. Both spiking procedures resulted in a fairly homogeneous distribution of the zinc oxide particles in soil, as evidenced by the low variation in total zinc concentration between replicate samples (< 12% in most cases). Survival of Folsomia candida in soil spiked at concentrations up to 6400 mg Zn/kg d.w. was not affected for both compounds. Reproduction was reduced in a concentration-dependent manner with EC50 values of 3159 and 2914 mg Zn/kg d.w. for 30 nm and 200 nm ZnO spiked as dry powder and 3593 and 5633 mg Zn/kg d.w. introduced as suspension. Toxicity of ZnO at 30 and 200 nm did not differ. We conclude that the ZnO particle toxicity is not size related and that the spiking of the soil with ZnO as dry powder or as a suspension in soil extract does not affect its toxicity to Folsomia candida.

Pauline L. Waalewijn-Kool, Maria Diez Ortiz and Cornelis A.M. van Gestel

Ecotoxicology 21, 2012, 1797-1804

26

Chapter 2

1 Introduction

As nanotechnology industry evolves and “nanotechnology enhanced” products enter the consumer world, environmental exposures are an inevitable consequence. Releases of nanoparticles (NP) may negatively affect the environment and toxic effects on organisms living in soils or water may occur. Ecotoxicological data are needed to establish sound risk assessment for this class of substances. The use of manufactured nanoparticles is a relatively new area of science and technology and the first papers on NP and ecotoxicity were published in 2006 (Kahru and Dubourguier, 2010).

The nano-size of these particles results in specific physicochemical characteristics that may differ from those of the bulk substance or larger sized particles. The high surface area in relation to the volume of the particle is the main cause of these differences. When particle size decreases the particle surface area increases exponentially. These characteristics may result in favourable properties for use in e.g. new cosmetic products, such as transparent sunscreens instead of the normal white ones. ZnO-NP is used in a wide range of cosmetic products such as foundations, hand creams and as a UV absorber in sunscreens. The physicochemical properties on the other hand also can lead to complex interactions of the (small) particles with the test media used in ecotoxicology. Test organisms therefore may not be exposed to the pristine NP powder or suspension as provided by a commercial supplier. Due to insolubility of nanoparticles in water and their aggregation in exposure media test organisms may be exposed to other forms or sizes of NP than were initially added to the test system (Heckmann et al., 2011). The behaviour of NP in soils is a complex process, due to their aggregation/agglomeration and abiotic factors affecting these processes such as pH, cation exchange capacity and natural organic matter content (Quik et al., 2010). In order to better understand how the presentation of NP in toxicity tests affects their uptake and toxicity, it is important to know the way nanoparticles interact with components of the test media.

The spiking of test media with NP for standardized ecotoxicity testing is an extremely important first step, because homogeneity of NP distribution in test media is difficult to establish and may influence the outcome of the toxicity test. Currently, NP are introduced in aquatic and terrestrial systems as dry powders or as suspensions, in stabilizing solvents, with or without sonication. Laban et al. (2010) compared stirring and sonication of Ag-NP solutions and found no difference in the percentage of dissolved Ag released in an aquatic test system. Some studies report the introduction of metal NP in soil as dry powder (Hu et al., 2010; Manzo et al., 2010) while others applied NP to the soil as a suspension in deionised water (Heckmann et al., 2011; Shoults-Wilson et al., 2011). Van der Ploeg et al. (2011) introduced C60 nanoparticles into soil, dissolved in an aqueous soil extract. The extract was obtained by stirring control soil in water. After filtering, C60 was added to the extract and the suspensions obtained were stirred to acquire an as stable and homogeneous NP suspension

27

2

Effect of different spiking procedures on the distribution and toxicity of ZnO nanoparticles in soil

as possible. This spiking method can influence the soil-water partitioning of the nanoparticles and in case of metal-based NP, substantial metal ions can be released from the NP. Further investigation of different spiking methods of NP is needed in order to develop a uniform approach for testing NP. It also remains uncertain to what extent the outcome of a toxicity test is influenced by the spiking method.

This study aimed at comparing different spiking methods, using ZnO NP. Characterization of NP in soil in general is rather difficult. Instead, total soil analyses and toxicity tests may provide insight into NP distribution in soil and potentially toxic effects that are related to NP concentration. ICP-MS and ICP-AES showed good recoveries (80-120%) for aluminium using soil spiked with dry Al2O3 NP powder (Coleman et al., 2010). And microwave digestion of soil spiked with Ag NP (suspension in deionised water) showed 100% recovery on average (Shoults-Wilson et al., 2011).

The springtail Folsomia candida (Collembola) has widely been used to assess the environmental impact of a range of pollutants on soils due to their abundance and diversity. F. candida is a parthenogenetic species and the rate of reproduction makes this species suitable for studying the effects on reproduction (Fountain and Hopkin, 2005). A widely-stated hypothesis is that nano-sized particles are more potent than larger particles of the same nominal substance because of their increased surface area per unit mass. This has been supported by animal studies with carbon black (Jia et al., 2005). Studies with metal-based nanoparticles found no clear relationship between NP size as such and the effect on soil organisms (Unrine et al., 2010).

The present study investigates the spiking of natural soil with differently sized ZnO particles (30 and 200 nm) using two different procedures, i.e. as dry powder and as a suspension in soil extract. Five sub-samples of spiked soil were analyzed to evaluate the distribution of zinc in the treated soil, and porewater samples were taken to evaluate possible differences in Zn dissolution. Differences in toxicity between the spiking methods and the particles were evaluated by exposing the springtail F. candida to the spiked soil. After four weeks exposure the effects on the survival and reproduction were determined and toxicity values (EC50/EC10) were estimated based on total zinc concentration in the soil.

2 Materials and methods

2.1 Nanoparticle characterisation and spiking proceduresTwo sizes of zinc oxide were applied, namely 30 nm (Nanosun Zinc Oxide P99/30) and 200 nm (Microsun Zinc Oxide W45/30). Figure 1 shows transmission electron micrographs (TEM) of the ZnO powders, which were dispersed in deionized water. Solutions were sonicated in a low power ultra sonication bath for 30 sec. and 1 drop (20 µl) was deposited on a carbon coated Cu TEM grid. Samples were dried at room temperature for several hours before examination in the TEM. Experiments were carried

28

Chapter 2

out on a JEOL 2010 analytical TEM, which has a LaB6 electron gun and can be operated between 80 and 200 kV. This instrument has a resolution of 0.19 nm, an electron probe size down to 0.5 nm and a maximum specimen tilt of ± 10 degrees along both axes. The instrument is equipped with an Oxford Instruments LZ5 windowless energy dispersive X-ray spectrometer (EDS) controlled by INCA. On the TEM images a small (approx. 20-25) number of particles were measured from about 4 or 5 TEM micrographs to get a rough particle size distribution for the primary particles. Each particle was measured individually from the TEM micrographs using Digital Micrograph program, which is a standard TEM instrument control and analysis program. The TEM images show that the ZnO particles were mainly equiaxial and rounded. Figure 1 shows the primary particle size distribution of the ZnO particles from the TEM images.

Seven concentrations (nominal range 100-6400 mg Zn/kg dry soil) and two controls without ZnO were tested. Loamy sand soil (LUFA-Speyer 2.2, Sp 2121, Germany, 2009) with a reported pHCaCl2 of 5.5, a total organic carbon content of 2.09%, a cation

Figure 1. Primary Particle Size Distribution (left) from Transmission Electron Micrographs (right) of 30 nm ZnO (top) with primary particle sizes of approx. 30-50 nm, and of 200 nm ZnO (bottom) with primary particle sizes of approx. 50-500 nm. Primary particles were dispersed in deionized water, sonicated and deposited on a carbon coated Cu TEM grid.

29

2


exchange capacity (CEC) of 10.0 cmolc/kg and a water-holding capacity (WHC) of

46.5% was used. For the dry spiking, powders were mixed in with 20 g dry soil in glass

jars. The treated soil was added to 180 g dry soil in glass containers. After mixing, water

was added to reach 50% of the WHC. Soil-water suspensions were prepared by mixing

air-dried soil with deionized water (Milli-Q) using a soil-water ratio of 2:5 (w/v) (van der

Ploeg et al., 2011). The suspensions were shaken at 180 rpm at ambient temperature for

one hour and filtered over a filter paper under vacuum. ZnO powders, corresponding

with the nominal concentrations, were weighed and added to 30 ml of the filtrates. Soil

solutions were shaken for two days at 180 rpm. Then, suspensions were mixed in with

200 g dry soil to obtain nominal test concentrations. Also with this procedure soil was

moistened to 50% of its WHC. The use of soil extract as spiking solution may prevent

the ZnO-NP to form aggregates and may diminish settlement or flocculation.

To visualize the particle size and the degree of aggregation of ZnO particles in the

spiking solution, images of the spiking solution (approx. 500 mg/l) were taken using

Transmission Electron Microscope (TEM) operating at 60 kV (JEOL 101, containing a 2 k

CCD camera). Samples were dropped on a 75 mesh copper Formvar grid and left to dry

before examination by TEM. The particle size distribution in the spiking solutions were

analysed by Nanoparticle Tracking Analysis (NTA) using NanoSight LM20 (size range

10-1000 nm). Videos were analyzed using NTA software (Version 1.5) and the minimum

required tracks of 100 were completed per video analysis. Therefore, spiking solutions were

100 times diluted in deionized water (Milli-Q) and followed in real-time for 90 seconds.

2.2 Zinc analysis in the soil and pore water

Five samples per treatment (± 100 mg dried soil) were randomly taken from the batches

of spiked soil and digested in a mixture of deionized water (Milli-Q), concentrated HCl

and concentrated HNO3 (1:1:4 by vol.) using an oven (CEM MDS 81-D). After digestion

for 7 hours at 140 °C, solutions were analysed for total zinc concentration by flame

Atomic Absorption Spectrometry (AAS) (Perkin Elmer AAnalyst 100). Certified reference

material (ISE sample 989 of River Clay from Wageningen, The Netherlands) was used

to ensure the accuracy of the analytical procedure. Measured zinc concentrations in

the reference material were within 10% of the certified concentrations. A two-sided

Student t-test was performed to compare spiking methods for each concentration.

Soil pore water was collected by centrifuging 50 g soil (Centrifuge Falcon 6/300 series,

CFC Free), after saturation with deionized water (Milli-Q) and one week equilibration. Soils

were centrifuged for 50 min. with a relative force of 2000 g over two round filters (S&S 597

Ø 47 mm, pore size 11 μm) and a 0.45 μm membrane filter (S&S Ø 47 mm), placed inside

the tubes (method described by Hobbelen et al., 2004). Approximately 10 ml pore water

per sample was collected for analysis by flame AAS (Perkin Elmer AAnalyst 100). Zinc

concentrations in the samples were also determined by flame AAS after ultrafiltration of

the soil pore water to obtain a particle-free extract. For this, soil solutions were centrifuged

in a 100 kDa ultrafiltration device (Amicon Ultra-15 Filters, Millipore) for 20 min. at 2000 g.

30

Chapter 2

2.3 Toxicity test with Folsomia candida

The springtail F. candida (Berlin strain; VU University Amsterdam) was cultured in pots

with a base of moist plaster of Paris mixed with charcoal at 20 ± 1 °C at a light/dark

regime of 12/12 h. The experiment was initiated with juveniles of the same age (10-12

days) that were obtained by synchronising the egg laying of the culture animals, fed

with dried baker’s yeast (Dr. Oetker). Toxicity tests were performed following ISO

guideline 11267 (ISO, 1999). The test was conducted in 100 ml glass jars containing

30 g moist soil. Five replicates for each zinc concentration and control were prepared.

At the start of the test, ten synchronised animals were transferred into each test jar.

The test jars were filled randomly and before introduction the animals were checked

under the microscope for a healthy appearance. The animals were fed a few grains of

dried baker’s yeast. The jars were incubated in a climate room at 20 ± 1 °C and with

a 12/12 hour light/dark cycle. Once a week, the moisture content of test soils was

checked by weighing the jars, and moisture was replenished with deionized water

(Milli-Q) when necessary. The jars were also aerated by this procedure.

After four weeks, the jars were sacrificed for determination of springtail survival

and reproduction. Each jar was emptied into a 200 ml beaker glass and 100 ml tap

water was added. The mixture was stirred carefully to let all the animals float to the

surface. After the adults were counted by eye, a picture of the water surface was taken

using a digital camera (Olympus, C-5060). The Cell^D imaging software was applied

to count the number of juveniles for determining the effect on reproduction. EC50

values, the actual concentration in the soil causing a 50% reduction in reproduction,

were estimated applying a logistic model according to Haanstra et al. (1985). In

addition, 10% effective concentration (EC10) values were obtained by modifying

the logistic model. A generalized likelihood ratio test (Sokal and Rohlf, 1995) was

applied to compare EC50 values obtained for both ZnO particles and for both spiking

procedures. All calculations were performed in SPSS Statistics 17.

3 Results

3.1 Zinc distribution in the spiking solution

Figure 2 shows transmission electron micrographs of Lufa 2.2 soil extract with 30 nm

and 200 nm ZnO, used for the spiking solution to reach 100 mg Zn/kg dry soil. Figure

2 clearly shows that the ZnO particles form aggregates in the soil extract. Free ZnO

particles, as shown in the TEM images (Figure 1), were not observed in the samples.

All aggregated ZnO particles were bound to natural organic matter (NOM) in the

soil extract. Aggregated ZnO particles seemed not to occupy all the empty spaces

available along the NOM (Figure 2). Attractive forces and binding of particles to each

other can not be avoided by the addition of NOM to the spiking solution. Nanoparticle

31

2


Tracking Analysis of the ZnO particles in the spiking solutions confirmed that ZnO particles were aggregated in the spiking solutions, but the histograms indicate a clear difference between the two ZnO particles (Figure 2). The NTA shows that the 30 nm ZnO has a strong peak for the primary particle size and aggregates/agglomerates in the 100-300 nm size range. The 200 nm ZnO, however, does not have a strong peak below approx. 100 nm, and the particle size range is much broader (200-600 nm).

3.2 Zinc concentrations in soil and pore waterBoth spiking procedures, i.e. dry powder and suspension spiking, showed a recovery above 85% of zinc in Lufa 2.2 soil (Table 1). Zinc concentrations in the soil were corrected for the zinc measured in the controls. The analyses of five subsamples, randomly taken from the treated soil, showed that spiking with dry powder or suspension did not influence

Figure 2. Particle Size Distributions determined by Nanoparticle Tracking Analysis (left) and Transmission Electron Micrographs (right) of the Lufa 2.2 soil extract spiked with 30 nm (top) and 200 nm ZnO (bottom). For the particle size distribution analysis, extracts spiked at 500 mg Zn/l were 100 times diluted in deionized water (Milli-Q) to obtain concentrations of approx. 5 mg Zn/l. The line represents cumulative particle sizes as a percentage of the total. The TEM images are from the non-diluted soil extracts (500 mg Zn/l), and show aggregates of ZnO particles and the binding of aggregates to dissolved natural organic matter.

32

Chapter 2

the distribution of zinc oxide particles in the soil. Based on these values no difference between the spiking procedures was found. Both spiking procedures resulted in a fairly homogeneous distribution of the ZnO particles in soil, as evidenced by the low variation between replicate samples (< 12% in most cases). The results of the two-sided Student t-test showed significant differences in total Zn concentrations in the soil between spiking methods in some cases, but differences were not consistent and not concentration related.

Zinc concentrations in the soil pore water were corrected for the zinc levels in the controls and ranged from 3.32 to 8.22 mg Zn/l for the lowest spiking concentration and from 18.3 to 25.8 mg Zn/l for the highest one (Table 2). No linear trend was observed between zinc concentrations in the soil pore water and the actual zinc concentrations in Lufa 2.2 soil. The maximum soluble zinc concentration was reached at intermediate spiking concentrations for both spiking procedures. Zinc concentrations in the pore water of soil spiked with ZnO as dry powder were slightly higher than the ones for suspension-spiked soils. This may suggest that ZnO-NP (or dissolved zinc) from suspensions are less available in the soil pore water due to binding of zinc to organic matter in the suspensions. However, the difference in soluble zinc is negligible based on measured total zinc concentrations in the soil (max. 1.24% for dry spiking and max. 0.841% for suspension spiking). Ultrafiltration did not reduce zinc concentrations in the porewater samples (Table 2), suggesting all zinc was available. The higher zinc concentrations after ultrafiltration in some samples are considered an experimental artefact.

3.3 Toxicity to Folsomia candidaSurvival of F. candida in soil spiked at concentrations up to 6400 mg Zn/kg was not affected for both ZnO powders. The average numbers of juveniles in the two control

Table 1. Average zinc concentrations (± SD, n=5) measured in Lufa 2.2 soil spiked with 30 and 200 nm ZnO as dry powder (D) and as suspension in soil extract (S). Zinc concentrations in the soil are corrected for the zinc levels measured in the controls. Recoveries (%) are presented in brackets.

Nominal conc.(mg Zn/kg d.w.) 30 nm, D 30 nm, S 200 nm, D 200 nm, S

Control 21.2 ± 2.1 17.9 ± 0.1 21.2 ± 2.1 17.9 ± 0.1

100 133 ± 15.5 (133) 122 ± 5.7 (122) 150 ± 15.3 (150) * 121 ± 11.0 (121) *

200 223 ± 7.97 (112) * 195 ± 13.3 (97.3) * 258 ± 15.7 (129) * 218 ± 13.5 (109) *

400 438 ± 17.4 (109) * 475 ± 11.0 (119) * 435 ± 18.9 (109) 458 ± 38.1 (115)

800 864 ± 41.2 (108) 904 ± 85.4 (113) 908 ± 65.4 (114) 993 ± 102 (124)

1600 1598 ± 112 (99.9) * 1899 ± 223 (119) * 1813 ± 318 (113) 1575 ± 62.2 (98.5)

3200 3280 ± 175 (103) 3322 ± 369 (104) 2913 ± 478 (91.0) 3075 ± 344 (96.1)

6400 5502 ± 274 (86.0) 5684 ± 967 (88.8) 6787 ± 205 (106) * 5608 ± 390 (87.6) *

* indicates a significant difference (p < 0.05) between the spiking methods for each concentration, using a two-sided Student t-test

33

2


soils were 206 (± 91.7, n=5) for dry spiking series and 81 (± 45.1, n=5) for suspension spiking. The effect of 30 and 200 nm ZnO particles on reproduction was concentration-dependent (Figure 3) with a steeper dose-response curve for dry spiking than for suspension spiking. The EC50 values for the effect on the reproduction of F. candida of 30 and 200 nm ZnO were 3159 (95% confidence interval: 126-5502) and 2914 (1813-6787) mg Zn/kg for the dry powder series. For the suspension-spiked soils, the EC50s were 3593 (122-5684) and 5633 (3711-5608) mg Zn/kg for 30 and 200 nm ZnO. Large 95% confidence intervals were estimated and no significant differences between the EC50 values were found when applying a generalized likelihood-ratio test (X2 < 3.84; n.s.). Corresponding EC10 values were 2559 (133-5502) mg Zn/kg for the 30 nm ZnO particles and 2730 (149-6787) for the 200 nm particles spiked in the soil as dry powder. For the 200 nm particles spiked as suspension EC10 was 3611 (121-5608) mg Zn/kg. No EC10 could be calculated for the 30 nm ZnO particles spiked as suspension. EC50 values based on porewater concentrations could not be estimated due to the small differences in zinc concentrations measured in the pore water.

4 Discussion

This study aimed at comparing two spiking procedures in order to find the most suitable method to introduce nanoparticle powders in soil ecotoxicological tests. Regarding the soil sample preparation for toxicity testing it may not be suitable to prepare the dosing using a dilution series of nanoparticles. Starting at the highest

Table 2. Zinc concentrations measured (n=1) in the pore water of Lufa 2.2 soil spiked with 30 and 200 nm ZnO as dry powder (D) and as suspension in soil extract (S) expressed as mg Zn/l. Soil pore water was collected one week after saturation of the soils with deionized water (Milli-Q). Zinc concentration in the control was deducted from the zinc concentrations measured in the soil pore water of the different treatments. Zinc concentrations measured in the soil pore water after ultrafiltration are presented in brackets

Nominal conc.(mg Zn/kg d.w.) 30 nm, D 200 nm, D 30 nm, S 200 nm, S

Control 0.60 (0.35) 0.60 (0.35) 0.63 (0.47) 0.63 (0.47)

100 6.78 (6.80) 8.22 (7.54) 4.55 (4.37) 3.32 (3.01)

200 12.0 (11.6) 11.6 (12.0) 6.98 (7.60) 5.99 (6.01)

400 20.9 (20.2) 23.8 (25.1) 15.9 (14.7) 14.1 (11.0)

800 21.8 (27.8) 21.8 (28.3) 19.5 (16.4) 15.7 (12.4)

1600 22.1 (26.9) 24.5 (31.7) 20.2 (25.6) 20.4 (19.3)

3200 23.8 (24.4) 25.8 (30.1) 18.3 (22.6) 18.6 (19.3)

6400 21.7 (24.9) 24.7 (24.6) 18.4 (21.6) 22.2 (24.1)

34

Chapter 2

Figure 3. Effect of 30 nm (Δ) and 200 nm (□) ZnO on the reproduction (average number of juveniles per test concentration is shown) of Folsomia candida after 28-d exposure in Lufa 2.2 soil. Actual exposure concentrations of zinc in the soil are provided on the x-axis applying suspension spiking (top) and dry spiking (bottom). Lines show fit obtained with a three parameter logistic dose-response model: ymax/(1+(concentration/EC50)b). The resulting equations are for 30 nm suspension spiking: 105/(1 + (conc./3593)0.973 (r2 = 0.140), for 200 nm suspension spiking: 93/(1 + (conc./5633)4.94 (r2 = 0.122), for 30 nm dry spiking: 192/(1 + (conc./3159)10.4 (r2 = 0.258) and for 200 nm dry spiking: 157/(1 + (conc./2914)43.2 (r2 = n.a.).

concentrations and diluting until the lowest test concentration may cause high deviations in the samples due to agglomeration of nanoparticles at each dilution step. The dosing is performed more precise when the amount of particles needed (weight in case of powders or volumes in case of suspensions) is added to the soil or other type of media for each test concentration separately.

0

50

100

150

200

250

300

350

400

1 10 100 1000 10000

num

ber o

f juv

enile

s

mg Zn/kg dry soil

30 nm

200 nm

0

20

40

60

80

100

120

140

160

1 10 100 1000 10000

num

ber o

f juv

enile

s

mg Zn/kg dry soil

30 nm

200 nm

35

2


Assessing NP or particle distribution in soil is difficult. In this study, we therefore focused on assessing the NP distribution in the soil extract used for suspension spiking. We showed that NTA analysis can be used for assessing particle size distributions in soil extracts. This technique, and many other characterization methods, provides the analysis at one time point and for one test concentration. In this study only one concentration (i.e. 5 mg Zn/l) has been analyzed at one time point. This concentration was 100 times lower than the Zn concentration in the spiking solution applied to reach the lowest concentration tested of 100 mg Zn/kg soil. Figure 2 shows one possible situation of the behaviour of ZnO particles in soil solution. Further in-depth characterization studies are necessary to assess NP behaviour (aggregation/agglomeration, particle size distributions) in soil and/or soil solution.

Instead of characterizing particle distribution in the soil, we analysed the soil for total Zn concentrations. The analyses of five subsamples, randomly taken from the treated soil, showed that spiking with dry powder or suspension did not influence the distribution of zinc oxide particles in the soil. The study showed good recoveries (> 85%) at all test concentrations with the ZnO particles. Both spiking procedures are currently applied in ecotoxicity testing and the most appropriate method has not been established yet among ecotoxicologists. A disadvantage of dry spiking is the static force of the particles that make them easily blown away. On the other hand dry particles are easily mixed in with dry soil particles. Introduction of nanoparticles in a soil extract may prevent agglomeration or flocculation as particles bind to natural organic matter or minerals. This procedure however, works better for dissolved nanoparticles than for insoluble particles.

The results obtained in this study showed that the survival of F. candida was not affected by ZnO particles up to 6400 mg Zn/kg d.w. The effect on reproduction was found to be concentration related and the EC50 values of the two compounds were not significantly different. Theoretically, it is expected that toxicity would be higher for 30 nm ZnO compared to 200 nm ZnO due to the larger surface area per volume, the NP therefore having more reactive sites to induce potential biological effects. Also, due to their small size ZnO-NP might have both greater mobility as well as potentially enhanced uptake across biological membranes. However, it seems that the size of ZnO does not contribute to a significant difference in the effect observed on springtail reproduction. This is in line with an aquatic toxicity study in which a 72-h IC50 of 60 mg Zn/l was found for the effect of both ZnO-NP (30 nm) and non-nano ZnO on the growth of the freshwater micro-algae Pseudokirchneriella subcapitata (Franklin et al., 2007). Wiench et al. (2009), testing different nano and non-nano ZnO particles, reported that the toxicity for Daphnia magna was independent of particle size, although the EC50 values for ZnO-NP and non-nano ZnO differed by a factor of 8. The characterization of particles in the spiking solution shows that the original ZnO particles were not present in their pristine form after their introduction into the soil extract (Figures 1 and 2). A wide range of particles sizes were present in the soil solution and therefore the relation between particle sizes and toxic effect could not be established.

36

Chapter 2

In this study porewater concentrations did not increase with increasing spiking concentrations, but showed a maximum of 20-25 mg Zn/l. Toxicity therefore cannot fully be explained by dissolved zinc measured in the pore water. The relationship between zinc concentrations in the soil and in the pore water suggests that a particle effect of ZnO can not be excluded yet. We show that ZnO-NP as such may have an effect on F. candida at concentrations between 1600 mg Zn/kg and 6400 mg Zn/kg. More studies are needed to gain insight into the processes occurring in the soil and pore water, in order to establish the ZnO-NP exposure to the animals.

This study shows that spiking with dry powder or suspension does not influence the distribution of zinc oxide nanoparticles in soil and their toxicity to F. candida. Both 30 nm and 200 nm ZnO particles were equally toxic to F. candida after 28 days exposure in natural soil.

37

2


3Chronic toxicity of ZnO nanoparticles, non-nano ZnO and ZnCl2

to Folsomia candida (Collembola) in relation to bioavailability in soil

Abstract

The chronic toxicity of zinc oxide nanoparticles (ZnO-NP) to Folsomia candida was determined in natural soil. To unravel the contribution of particle size and free zinc to NP toxicity, non-nano ZnO and ZnCl2 were also tested. Zinc concentrations in pore water increased with increasing soil concentrations, with Freundlich sorption constants Kf of 61.7, 106 and 96.4 l/kg (n=1.50, 1.34 and 0.42) for ZnO-NP, non-nano ZnO and ZnCl2 respectively. Survival of F. candida was not affected by ZnO-NP and non-nano ZnO at concentrations up to 6400 mg Zn/kg d.w. Reproduction was dose-dependently reduced with 28-d EC50s of 1964, 1591 and 298 mg Zn/kg d.w. for ZnO-NP, non-nano ZnO and ZnCl2, respectively. The difference in EC50s based on measured porewater concentrations was small (7.94-16.8 mg Zn/l). We conclude that zinc ions released from NP determine the observed toxic effects rather than ZnO particle size.

Pauline L. Kool, Maria Diez Ortiz and Cornelis A.M. van Gestel

Environmental Pollution 159, 2011, 2713-2719

40

Chapter 3

1 Introduction

In recent years, there is increasing interest in the environmental risks posed by engineered nanoparticles (NP), particles with at least one dimension of less than 100 nm and differing from non-nano material in physico-chemical properties, such as surface area and charge density (Handy et al., 2008). Zinc oxide, one of the most commonly used types of metal-based NP, is primarily used in electronics, personal care products and other applications requiring UV protection like sunscreens. Zinc oxide nanoparticles (ZnO-NP) can enter the environment via waste water at industrial sites or through domestic sewage from showering or swimming. NP can be transported to soil via sewage sludge, which is used for land application (e.g. fertilizer). Gottschalk et al. (2009) reported this to be the main route of soil exposure and estimated a predicted environmental concentration (PEC) of 0.093 µg/kg/y for European soils based on an annual production volume of 9845 tonnes; in case of sludge application on land, PEC is estimated at 3.25 µg/kg/y.

Consequently organisms living in the soil, such as earthworms, springtails, nematodes and isopods, may be harmed. Ecotoxicological studies with NP have mainly been performed with aquatic organisms (Heinlaan et al., 2008, Zhu et al., 2008; Wiench et al., 2009). A small but growing number of studies on the toxicity of NP to soil organisms has been published in the last two years (Peralta-Videa et al., 2011), assessing effects of short-term NP exposure to the isopod Porcellio scaber (Drobne et al., 2009; Jemec et al., 2008), the nematode Caenorhabditis elegans (Ma et el., 2009; Wang et al., 2009), and the earthworms Eisenia fetida (Hu et al., 2010; Unrine et al., 2010) and Lumbricus rubellus (Lapied et al., 2010; Lapied et al., 2011; van der Ploeg et al., 2011). In addition, more research is needed to provide insight into the ecotoxicological effects of chronic exposure to NP on organisms living in soil.

In case of metal-based NP like ZnO, TiO2, Ag and CeO, toxicity is at least partly due to the specific properties related to the small size and consequent high surface activity of NP, while effects may be further enhanced by the release of free metal ions (Auffan et al., 2009). If the free ions are more toxic than the original particles, this process of dissolution is likely to lead to an increase of the overall toxicity. Franklin et al. (2007) showed that the toxicity of ZnO-NP to the micro-algae Pseudokirchneriella subcapitata was attributable to dissolved zinc, while others could not explain the toxicity for the nematode C. elegans adequately by dissolution of the particles alone (Wang et al., 2009). Speciation of metal NP in soils is not yet understood. Given the well-known toxicity of the ionic forms of some metals, the solubility of metal-based NP may require particular attention. Dissolution of ZnO-NP has been observed in moderately hard reconstituted water to yield Zn concentrations of approx. 0.4 mg/l (Poynton et al., 2011). In soil, sorption is also an important factor that needs to be taken into account when performing toxicity or bioaccumulation tests. This depends strongly on soil characteristics such as organic matter content, cation exchange capacity and pH.

41

3

Chronic toxicity of ZnO nanoparticles, non-nano ZnO and ZnCl2 to Folsomia candida

Due to the complex behaviour of NP in soil, realising realistic exposure in ecotoxicity testing poses major challenges. Soil by definition is heterogeneous, which requires conscientious introduction of any test compound and frequently sub-samples are analyzed to ensure that spiked portions of soil are homogeneous. In general, manufactured NP are considered to be insoluble in water and tend to aggregate, making it a difficult task to obtain homogeneity of NP distribution in spiked soil. Depending on their physical and chemical properties as well as soil properties, NP tend to form aggregates and are likely to settle within a relatively short time (Klaine et al., 2008). Currently used spiking procedures may need to be adapted with respect to preparation of the soil to be spiked, introduction of NP to the prepared soil and mixing to evenly distribute NP throughout the soil.

Springtails (Collembola) are common and widespread arthropods that are abundant in soils throughout the world. They represent ecologically and relevant test species, because they are an integral part of soil ecosystems and are vulnerable to the effects of soil contamination. The cuticle and ventral tube (diameter approx. 5 µm) are important exposure routes for chemicals dissolved in soil pore water (Fountain and Hopkin, 2005). Folsomia candida, a parthenogenetic species, was chosen because of its relatively short generation time and the ease of culturing. The fact that there is only one limit-test with ZnO-NP and springtails yet published in the literature (Manzo et al., 2011) indicates the need for further research on these organisms.

The purpose of the present study was to determine the chronic toxicity of ZnO-NP to F. candida, by studying survival and reproduction of the springtail as effect parameters and to unravel the contribution of zinc oxide particle size and free zinc to NP toxicity. To help elucidate toxic effects of ZnO-NP in soil, two reference compounds, non-nano ZnO and ZnCl2 were studied for comparison. We hypothesized that ZnO-NP is more toxic than non-nano ZnO due to its larger specific surface area and that the release of free zinc from ZnO-NP, rather than the NP itself, is responsible for toxic effects.


2.1 Test compoundsZnO-NP powder, with a reported diameter size of < 200 nm, was purchased from BASF (Z-COTE®). Powders were coated with carbon and photos of the particles were taken using a (field emission) scanning electron microscope (JEOL JSM-6301F). Non-nano ZnO (Merck, pro analysi, > 99%) and ZnCl2 (Merck, zinc chloride pure) were used for comparison. Figure 1 shows that the diameter size of ZnO-NP powder is at the nano-scale (i.e. < 100 nm), although the fraction < 100 nm has not been established. The diameter size of non-nano ZnO powder is 200 nm in all cases, and therefore provides a good reference substance to study size-related effects of ZnO.

42

Chapter 3

2.2 Exposure of springtails to ZnO-NP, non-nano ZnO and ZnCl22.2.1 Test organismThe springtail F. candida (Berlin strain; VU University Amsterdam) was cultured in plastic containers with a moist bottom of plaster of Paris containing 10% charcoal, at 20 ± 1 °C at a light/dark regime of 12/12 h. The experiment was initiated with juveniles of the same age (10 - 12 days) that were obtained by synchronising the egg laying of the culture animals, fed with dried baker’s yeast (Dr. Oetker).

2.2.2 Method of exposureLoamy sand soil (LUFA-Speyer 2.2, Sp 2121, Germany, 2009) with a pHCaCl2 of 5.5, a total organic carbon content of 2.09%, a cation exchange capacity (CEC) of 10.0 meq/100g and a water-holding capacity (WHC) of 46.5% was used. The soil was oven-dried at 60 °C overnight prior to the experiment to eliminate undesired soil fauna. The test consisted of seven ZnO-NP concentrations (nominal range 100-6400 mg Zn/kg d.w.), five non-nano ZnO concentrations (400-6400 mg Zn/kg d.w.), five ZnCl2 concentrations (100-1600 mg Zn/kg d.w.) and a control without zinc. Slightly different exposure scales were chosen for ZnO-NP, non-nano ZnO and ZnCl2, because it was expected that soluble zinc compounds are more toxic to soil organisms than insoluble zinc compounds, at least shortly after the addition to the soil. To achieve an as homogeneous distribution as possible, the test compounds were introduced into the soil as aqueous solutions prepared in soil extracts according to van der Ploeg et al. (2011). For that purpose soil-water suspensions were prepared by mixing air-dried soil with deionized water (Milli-Q) using a soil-water ratio of 2:5 (w/v). The suspensions were shaken at 180 rpm at ambient temperature for one hour. Soil extracts were filtered under vacuum (Whatman filter paper, type 595) and ZnO-NP, non-nano ZnO and ZnCl2 were added to the filtrates. The solutions, which showed a milky-white colour, were shaken for two days at 180 rpm and carefully mixed with 200 g dry soil. Additional deionized water (Milli-Q) was added to achieve a soil moisture content of 23.3% (w/w) corresponding with 50% of the maximum WHC. Soils were equilibrated for one day before use in the toxicity test.

To visualize the size and degree of aggregation of the ZnO particles, photos of ZnO-NP and non-nano ZnO in the spiking solution (approx. 1000 mg Zn/l) were taken using transmission electron microscope (TEM) operating at 60 kV (JEOL 1010, containing a 2k CCD camera). Soil solution without ZnO particles was taken for comparison. Samples were dropped on a 75 mesh copper Formvar grid and left to dry before examination by TEM.

2.2.3 Toxicity testToxicity tests were performed following ISO guideline 11267 (ISO, 1999), using 100 ml glass test containers containing 30 g moist soil each. Five replicates for each zinc concentration and control were prepared. At the start of the test, ten synchronised animals were transferred into each test jar. The test jars were filled randomly and

43

3


before introduction, the animals were checked under the microscope for a healthy appearance. At the beginning of the experiment the animals were fed a few grains of dried baker’s yeast. The jars were incubated in a climate room at 20 ± 1 °C and at a light/dark regime of 12/12 h. Once a week, the moisture content of the test soils was checked by weighing the jars, and moisture was replenished with deionized water when necessary. The jars were also aerated by this procedure.

After four weeks, the jars were sacrificed for determination of springtail survival and reproduction. Each jar was emptied into a 200 ml beaker glass and 100 ml tap water was added. The mixture was stirred carefully to let all the animals float to the surface. The adults were counted by eye and a picture of the water surface was taken using a digital camera (Olympus, C-5060). The Cell^D imaging software was applied to count the number of juveniles for determining the effect on reproduction.

2.3 Zinc analyses

2.3.1 Zinc concentrations in the soilSoil samples were dried for 24 h at 60 °C. For total zinc analysis, ± 100 mg of dried soil (two replicates per treatment) was digested in a mixture of deionized water (Milli-Q), concentrated HCl and concentrated HNO3 (1:1:4 by vol.). Tightly closed bombs were placed in an oven (CEM MDS 81-D) at 140°C for 7 hours. After digestion, the solution was analysed by flame Atomic Absorption Spectrometry (AAS) (Perkin Elmer AAnalyst 100).

Certified reference material (ISE sample 989 of River Clay from Wageningen, The Netherlands) was used to ensure the accuracy of the analytical procedure. Measured zinc concentrations in the reference material were within 10% of the certified concentrations.

2.3.2 Zinc concentrations in the soil pore waterAt the end of the toxicity tests, soil pore water was collected by centrifuging 50 g soil (Centrifuge Falcon 6/300 series, CFC Free), after saturation with deionized water (Milli-Q) and two weeks equilibrium time. Soils were centrifuged for 50 min. with a relative force of 2000 g over two round filters (S&S 597 Ø 47 mm, pore size 11 μm) and a 0.45 μm membrane filter (S&S Ø 47 mm), placed inside the tubes (method described by Hobbelen et al., 2004). Approximately 10 ml soil pore water per sample was collected for analysis by flame AAS (Perkin Elmer AAnalyst 100).

Zinc concentrations in the samples were also determined by flame AAS after ultrafiltration of the soil pore water to obtain a particle-free extract. Soil solutions were centrifuged in a 100 kDa ultrafiltration device (Amicon Ultra-15 Filters, Millipore) for 20 min. at 2000 g.

The pH of the soil pore water was measured before and after ultrafiltration using a Consort P907 meter.

2.4 Data analysisUsing the measured soil porewater and total soil concentrations, sorption of zinc to the test soil was described by a Freundlich isotherm:

44

Chapter 3

Cs = Kf * Cw n

whereCs = concentration in soil (mg Zn/kg d.w.)Kf = Freundlich sorption constant (l/kg)Cw = concentration in the pore water (mg Zn/l) andn = shape parameter of the Freundlich isotherm

Estimates of Kf and n were obtained by linear regression in a double logaritmic plot of Cs versus Cw. In such a plot n is estimated from the slope. When n > 1, this may be indicative of agglomeration or aggregation due to saturation effects in the solution. In the case n < 1, binding sites in the soil may become saturated with zinc.

Taking into account that the number of binding sites is finite, sorption of zinc in the test soils treated with ZnCl2 was described by a Langmuir isotherm:

CmaxKLCw

Cs = ------------ 1 + KLCw

where Cs = concentration in soil (mg Zn/kg d.w.)Cmax = maximum sorption capacity (mg/kg)KL = Langmuir sorption constant (l/mg)Cw = concentration in the pore water (mg Zn/l)

The KL may be interpreted as the inverse of the porewater concentration at which 50% of the sorption sites is occupied. Estimates for Cmax and KL were obtained by nonlinear regression.

EC50, the actual concentration in the soil and soil pore water causing a 50% reduction in springtail reproduction, was estimated applying the logistic model of Haanstra et al. (1985). In addition, 10% effective concentration (EC10) values were obtained by modifying the logistic model. A generalized likelihood ratio test (Sokal and Rohlf, 1995) was applied to compare EC50 values obtained for different zinc forms (ZnO-NP, non-nano ZnO and ZnCl2). All calculations were performed in SPSS Statistics 17. The trimmed Spearman-Karber method was applied to estimate the LC50 value for the effect of ZnCl2 on springtail survival (Hamilton et al., 1977).

3 Results

3.1 Zinc distribution in the spiking solutionFigure 2 shows transmission electron micrographs of Lufa 2.2 soil solution without ZnO (A), with non-nano ZnO (B) and ZnO-NP (C, D). Figure 2C visualizes the binding of ZnO-NP to organic matter and shows a homogenous zinc distribution in the sample.

45

3


Free ZnO nanoparticles, as shown on the SEM pictures (Figure 1), were not observed. Non-nano ZnO was present as bigger particles and the binding to organic matter was less evident than for ZnO-NP (Figure 2B).

3.2 Zinc concentrations and pHThe results of the zinc measurements in the soil and in the soil pore water before and after ultrafiltration are presented in Tables SI-1 and SI-2 of the Supporting Information (SI). A good recovery (> 80%) was found for all soil samples except for soil spiked with the highest concentration of ZnO-NP (i.e. 6400 mg Zn/kg d.w.) providing an average recovery from seven subsamples of only 35%. Zinc concentrations in the soil pore water ranged from 1.85 to 12.6 mg Zn/l in soil spiked with ZnO-NP, which corresponds with a solubility of 0.286 to 0.083% of the measured total zinc concentration. Concentrations of 3.37 to 16.9 mg Zn/l (0.161 to 0.061%) and 1.81 to 612 mg Zn/l (0.315 to 8.55%) were measured for non-nano ZnO and ZnCl2, respectively. Ultrafiltration did not change zinc concentrations in the soil solutions. Results of pH measurements in the soil pore water are presented in Table SI-3 of the Supporting Information. After ultrafiltration pH values had increased in all soil solutions. For ZnCl2, the pH was slightly lower at high added zinc concentrations. This may be explained by the excess of zinc ions, added in high concentrations, causing a release of protons from sorption sites on the soils. Unfortunately, we are not able to explain the slight increase in pH values for ZnO-NP and non-nano ZnO soil solutions. All results are expressed on the basis of measured concentrations.

3.3 Zinc sorptionFor all three compounds, zinc concentrations in the pore water increased with increasing total soil concentrations. This is reflected by the Freundlich isotherm in test soils spiked with ZnO-NP, non-nano ZnO and ZnCl2 (Figure 3) yielding zinc sorption constants Kf of

Figure 1. Scanning electron micrographs of dry powders: non-nano ZnO (left) and ZnO-NP (right). Original magnification: ×70,000.

46

Chapter 3

61.7, 106 and 96.4 l/kg (with corresponding n values of 1.50, 1.34 and 0.42) respectively. The Langmuir isotherm gave a good description for the test soil spiked with ZnCl2 and its sorption constant KL was 0.0038 l Zn/mg with a Cmax of 2238 mg Zn/kg (see Figure SI-1).

3.3 Toxicity of ZnO-NP, non-nano ZnO and ZnCl2Survival of F. candida in soil spiked at concentrations up to 6400 mg Zn/kg d.w. with ZnO-NP and non-nano ZnO was not affected and comparable to that in the control (i.e. > 72%). Survival was affected by ZnCl2 with an LC50 of 1000 mg Zn/kg d.w (95% CI 861-1162). Table 1 summarizes the LC50, EC50 and EC10 values for the effect on F. candida exposed to soil spiked with ZnO-NP, non-nano ZnO and ZnCl2.

The average number of juveniles in the controls was 140 with a coefficient of variance of 40%. Reproduction was reduced in a dose-dependent manner (Figure 4) and EC50 values of 1964, 1591 and 298 mg Zn/kg d.w. were estimated for ZnO-NP, non-nano ZnO and ZnCl2, respectively. According to a likelihood-ratio test, the EC50 of ZnCl2 was significantly lower than the EC50s of ZnO-NP and non-nano ZnO, (χ2

(1) = 12.1 and 22.9, p < 0.001), while the EC50s of ZnO-NP and non-nano ZnO were not significantly different (χ2

(1) = 0.67). EC50 values of 10.1, 7.94 and 16.8 mg Zn/l were calculated for ZnO-NP, non-nano ZnO and ZnCl2, respectively based on measured concentrations in the soil pore water. The porewater EC50 for non-nano ZnO was significantly lower than

200nm

84nm

65nm

82nm

D

x 100,000

C

x 15,000500nm

B

x 15,000 500nm

A

x 15,000500nm

Figure 2. Transmission electron micrographs of Lufa 2.2 soil solution without ZnO (A), with non-nano ZnO (B) and ZnO-NP (C and D) (approx. 1000 mg Zn/l), showing the binding of NP to organic materials.

47

3


y = 1.50x + 1.79R2 = 0.972

y = 1.34x + 2.02R2 = 0.931

y = 0.424x + 1.98R2 = 0.992

1,5

2

2,5

3

3,5

4

0 0,5 1 1,5 2 2,5 3

log concentration soil pore water (mg Zn/l)

log

tota

l soi

l con

cent

ratio

n (m

g Zn

/kg)

1 10 100 1000concentration soil pore water (mg Zn/l)

bulk ZnOZnCl2ZnO-NPs

0

50

100

150

200

250

300

10 100 1000 10000

actual concentration (mg Zn/kg)

num

ber o

f juv

enile

s

A B

Actual concentration (mg Zn/kg dry soil) Porewater concentration (mg Zn/l)

Num

ber o

f juv

enile

s pe

r tes

t con

tain

er

non-nano ZnOZnCl2ZnO-NP

Figure 3. Measured zinc concentrations in soil as a function of zinc concentrations in soil pore water before ultrafiltration. Lines and equations show the fit of a Freundlich isotherm to the data for the sorption of Zn in Lufa 2.2 soil spiked with ZnO-NP (●), non-nano ZnO (□) and ZnCl2 (Δ).

Figure 4. Effect of ZnO-NP (●), non-nano ZnO (□) and ZnCl2 (Δ) on the reproduction (number of juveniles) of Folsomia candida after 28-d exposure in Lufa 2.2 soil. Actual exposure concentrations of zinc in the soil (A) and the soil pore water (B) are provided on the x-axis. Line shows fit obtained with a logistic model.

the one for of ZnO-NP (χ2(1) = 5.02, p < 0.05). Based on the concentration in the soil

pore water, the EC10 of ZnCl2 was 1.73 mg Zn/l, which was significantly lower value than the EC10 for ZnO-NP (χ2

(1) = 28.3, p < 0.05).

48

Chapter 3

4 Discussion

4.1 Characterization of NP in the spiking solutionTo date, very little work has been conducted on the identification of engineered NP in soil. The accurate detection of NP in soils requires their separation from natural soil solids (Lead and Wilkinson, 2006). In this study, transmission electron microscopy (TEM) was applied to visualize ZnO-NP in the soil solution (approx. 1000 mg Zn/l) used to spike the soils. Introduction of NP to soil in a filtered soil extract was chosen to prevent agglomeration or flocculation as particles bind to natural organic matter or minerals. Naturally occurring surfactants, such as humic and fulvic acids, may also help to prevent NP aggregation (Hyung et al., 2007). Figure 2 shows a homogenous distribution of the ZnO-NP in the sample and confirms the binding of NP to the organic materials as small aggregates or as crystals. The shape of the particles differs, which can be the start of dissolution and transformation of the particles in the soil solution. The spiking procedure used in this test showed a good recovery of zinc by chemical analyses except for the highest NP test concentration, which emphasizes the importance of expressing toxicity data on the basis of measured concentrations.

4.2 Toxicity to Folsomia candida

Influence of particle sizeThis study showed that the survival of F. candida was not affected by ZnO-NP and non-nano ZnO at concentrations up to 6400 mg Zn/kg d.w. Our hypothesis was that toxicity would be higher for ZnO-NP compared to non-nano ZnO due to the larger surface area per volume, the NP therefore having more reactive sites to induce potential biological effects. Also, due to their small size ZnO-NP might have both

Table 1. LC50, EC50s and EC10s for the effect of ZnO-NP, non-nano ZnO and ZnCl2 on the survival and reproduction of Folsomia candida after 28-d exposure in spiked Lufa 2.2 soil. EC50s and EC10s are presented as concentrations measured in the soil (mg Zn/kg d.w.) and in the soil pore water (mg Zn/l). 95% confidence intervals are presented in brackets.

Compound LC50(mg Zn/kg d.w.)

EC50(mg Zn/kg d.w.)

EC50(mg Zn/l)

EC10(mg Zn/kg d.w.)

EC10(mg Zn/l)

ZnO-NP > 3086 1964 a

(1635-2293)10.1 a

(7.83-12.4)1678 a

(1217-2139)9.47 a

(8.37-10.6)

non-nano ZnO > 6282 1591 a

- *7.94 b

- *1383 a

- *6.85 ab

- *

ZnCl2 1000(861-1162)

298 b

(181-415)16.8 ab

- *108 b

(7.45-209)1.73 b

- *

* Data did not allow calculating reliable 95% confidence intervalsa/b Different letters within a column indicate significant differences between EC50 or EC10 values according to a generalized likelihood ratio test (χ2

(1) > 3.84; p < 0.05)

49

3


greater mobility as well as uptake across biological membranes. Apoptotic activity was detected in the cuticle, intestinal epithelium and chloragogenous tissue of the earthworm Lumbricus terrestris exposed to silver NP at acute and sublethal concentrations (< 4 mg/kg) (Lapied et al., 2010). The effect on reproduction was found to be dose related and the dose-response curves for ZnO-NP and non-nano ZnO are fairly steep. Figure 4 shows that the numbers of juveniles produced were lower for ZnO-NP than for non-nano ZnO at concentrations below 1800 mg Zn/kg in the soil and 9.5 mg Zn/l in the soil pore water, while reproductive effects of non-nano ZnO were higher above these concentrations. However, the EC10 and EC50 values for ZnO-NP and non-nano ZnO were not significantly different. For this reason, it seems that the size of ZnO-NP did not contribute to a significant difference in the effect on springtail reproduction. This is in line with an aquatic toxicity study in which a 72-h IC50 of 60 µg Zn/l was found for the effect of both ZnO-NP (30 nm) and non-nano ZnO on the growth of the freshwater micro-algae Pseudokirchneriella subcapitata (Franklin et al., 2007). Also Wiench et al. (2009), testing different nano and non-nano TiO2 and ZnO particles, reported that the toxicity for Daphnia magna was independent of particle size, although the EC50 values for ZnO-NP (Z-COTE) and non-nano ZnO differed by a factor of 8, In our study, the difference of the EC50s is negligible (factor 1.2).

Also, in the previous study 30 nm and 200 nm ZnO were equally toxic to F. candida in soil (Chapter 2). The toxicity of 30 nm and 200 nm ZnO, based on total zinc concentrations in the soil, was lower than in this study with F. candida. ZnO-NP are commercially available as dry powders with particle sizes ranging from approximately 20 to 200 nm (Wang et al., 2004). The ZnO particles tested in this study were obtained from a different source (BASF Z-COTE® ZnO-NP and non-nano ZnO from Merck) compared to the ones used in our previous study (Nanosun ZnO-NP; Microsun non-nano ZnO). NP manufactured in different laboratories may have slightly different sizes and physicochemical properties, such as surface charge, size and shape. It is therefore likely that the use of different ZnO-NP has resulted in slightly different toxicities.

Influence of particle dissolutionZnCl2 was tested to compare the toxicity of ZnO-NP and that of free zinc. Our data on survival and reproduction (28-d LC50 1000 mg Zn/kg d.w.; 28-d EC50 298 mg Zn/kg d.w.) agreed with previous studies performed in similar soil types with ZnCl2 and F. candida (Lock and Janssen, 2001; Nota et al., 2010; Smit and van Gestel, 1996). Compared to ZnCl2, the EC50 for ZnO-NP in our study was almost 7-fold and the EC10 15-fold higher based on total zinc concentration in soil. This substantial difference in EC50 values based on soil concentrations indicates that zinc ions are much more toxic than NP. However, the difference in EC50 values based on soil porewater concentrations between the two compounds was small, and these values were in the range of EC50s (2.6-30 mg/l) found for water-extractable Zn in different soils (Smit and van Gestel, 1996; Smit et al., 1997). In toxicity tests with C. elegans and P. subcapitata ZnO-NP and ZnCl2 were also equally toxic regarding reproduction effects with EC50

50

Chapter 3

values of approximately 53 mg/l for the nematode and 60 µg Zn/l for the algae (Ma et al., 2009; Franklin et al., 2007). Ultrafiltration did not change the zinc concentration in the soil pore water, which indicates that the soluble NP fraction was completely dissolved and not present in the pore water as aggregates, agglomerates or intact particles. Free ZnO nanoparticles were not observed in the spiking solution (Figure 2) and it is unlikely that the NP would be released into the soil pore water. Soil porewater concentrations were only measured at one time point, so dynamics of the NP in the pore water was not investigated. The results suggest that toxicity can be related to the free Zn ions in the soil solution. The contribution of ZnO-NP to toxicity probably is very low, resulting in only slightly higher toxicity of the ZnO-NP when expressed on the basis of soil solution concentrations.

4.3 SorptionPartition coefficients for ZnO-NP in soil have not been published yet. Cornelis et al. (2010) reported Kr values for five different soils in the range of 77-2165 for Ag-NP and 1.1-2828 l/kg for CeO2. More than 7% of Ag-NP occurred as soluble Ag after 24 hours, which suggests higher solubility compared to ZnO-NP (0.286% for the lowest test concentration). We estimated the sorption constant Kf for zinc to be lower for ZnO-NP (61.7 l/kg) than for ZnCl2 (96.4 l/kg). In other studies, using soils with similar CEC values, higher zinc sorption coefficients were found for ZnCl2, for instance 214 l/kg in Panheel field soil and 463 l/kg in artificial soil with corresponding n values below one (0.47 in both cases) (Smit et al., 1997; Van Gestel and Hensbergen, 1997). The n value of 0.42 for ZnCl2 in Lufa 2.2 suggests saturation equilibrium in the soil, which is also confirmed by the more mechanistic Langmuir isotherm (Figure SI-1). The inverse of KL indicates that 50% of the sorption sites were occupied at a concentration of 264 mg Zn/l in soil pore water. The CEC of approximately 10 meq/100g, corresponding with 3250 mg Zn/kg, confirms that two third of the maximum sorption capacity of the Lufa 2.2 test soil is occupied with Zn2+ ions. ZnO-NP gave a higher n value than ZnCl2 (1.50 compared to 0.42), which suggests saturation effects in the soil pore water. This situation may indeed occur when agglomeration or aggregation occurs. The size of the agglomerates of ZnO-NP and ZnO-NP bound to organic matter was not measured, and agglomeration and binding to organic matter may decrease the bioavailability of ZnO-NP via the pore water. Non-nano ZnO showed a similar sorption constant as ZnO-NP, which indicates that sorption could be an important factor to explain the similar toxicity of the two compounds. A somewhat higher n value was estimated for ZnO-NP compared to non-nano ZnO and a slight increase of the zinc concentration was measured in the pore water. Further studies on the dissolution kinetics of ZnO-NP and non-nano ZnO are needed to unravel the mechanisms for NP toxicity for Folsomia candida.

51

3


0

200

400

600

800

1000

1200

1400

1600

1800

0 100 200 300 400 500 600 700

soil pore water concentration (mg Zn/l)

tota

l soi

l con

cent

ratio

n (m

g Zn

/kg)

5 Conclusion

In this study we characterized the toxicity and bioavailability of ZnO-NP, non-nano ZnO and ZnCl2 to the springtail Folsomia candida in a natural soil. We conclude that effects on springtail reproduction observed after 28 days is not related to ZnO particle size and most probably result from zinc dissolved from ZnO-NP, non-nano ZnO and ZnCl2.

Supporting Information

• Langmuir isotherm for the sorption of Zn in Lufa 2.2 soil spiked with ZnCl2 (Figure SI-1) • Average zinc concentrations and recoveries measured in the soil (Table SI-1) • Zinc concentrations measured in the pore water before and after ultrafiltration

(Table SI-2) • pH of the pore water before and after ultrafiltration (Table SI-3)

Figure SI-1. Langmuir isotherm for the sorption of Zn in Lufa 2.2 soil spiked with ZnCl2 (Δ) describing the relationship between measured zinc concentration in soil and corresponding zinc concentration in the soil pore water.

52

Chapter 3

Table SI-1. Average zinc concentration (± SD, n=2) measured in Lufa 2.2 soil spiked with ZnO-NP, non-nano ZnO and ZnCl2. Recoveries (%) are presented in brackets.

Nominal [Zn](mg Zn/kg d.w.)

Measured [Zn] in soil (mg Zn/kg d.w.)

ZnO-NP non-nano ZnO ZnCl2

0 14.9 ± 0.42

100 145 ± 12.4 (145) 129 ± 2.08 (129)

200 283 ± 13.8 (142) 259 ± 20.3 (130)

400 492 ± 69.0 (123) 470 ± 41.4 (118) 445 ± 7.03 (111)

800 965 ± 23.9 (121) 877 ± 10.1 (110) 913 ± 13.6 (114)

1600 1721 ± 4.67 (108) 1759 ± 81.7 (110) 1610 ± 20.3 (101)

3200 3086 ± 60.3 (96) 2628 ± 188 (82)

6400 2238 ± 290 (35)* 6282 ± 197 (98)

* n=7

Table SI-2. Zinc concentrations measured in the pore water of Lufa 2.2 soil spiked with ZnO-NP, non-nano ZnO and ZnCl2 expressed as mg Zn/l and as percentage soluble Zn (based on measured total Zn concentration in the soil). Soil pore water was collected two weeks after saturation of the soils with deionized water (Milli-Q). Zinc concentrations measured in the soil pore water after ultrafiltration are presented in brackets.


Measured [Zn] in soil pore water


mg Zn/l % mg Zn/l % mg Zn/l %

0 0.710 (1.67)

100 1.85 (2.00) 0.286 1.81 (1.86) 0.315

200 2.58 (2.73) 0.205 9.71 (11.2) 0.842

400 4.56 (4.28) 0.208 3.37 (3.30) 0.161 49.2 (57.2) 2.49

800 5.42 (5.32) 0.127 4.10 (4.94) 0.105 223 (248) 5.49

1600 9.57 (9.87) 0.125 8.82 (5.72) 0.113 612 (607) 8.55

3200 11.4 (11.6) 0.083 14.1 (15.2) 0.121

6400 12.6 (12.3) 0.127 16.9 (18.9) 0.061

53

3


Table SI-3. pH of the pore water of Lufa 2.2 soil spiked with ZnO-NP, non-nano ZnO and ZnCl2. pH values measured in the soil pore water after ultrafiltration are presented in brackets.


pH in soil pore water


0 6.53 (6.94)

100 6.73 (7.15) 7.46 (7.80)

200 6.84 (7.25) 6.97 (7.52)

400 6.83 (7.20) 7.02 (7.85) 6.40 (7.03)

800 7.08 (7.44) 7.30 (8.00) 5.86 (6.73)

1600 7.23 (7.35) 7.26 (7.84) 5.38 (6.31)

3200 7.38 (7.50) 7.50 (8.00)

6400 7.32 (7.57) 7.43 (7.73)

54

Chapter 3

4Sorption, dissolution and pH determine the long-term equilibration

and toxicity of coated and uncoated ZnO nanoparticles in soil

Abstract

To assess the effect of long-term dissolution on bioavailability and toxicity, triethoxyoctylsilane coated and uncoated zinc oxide nanoparticles (ZnO-NP), non-nano ZnO and ZnCl2 were equilibrated in natural soil for up to twelve months. Zn concentrations in pore water increased with time for all ZnO forms but peaked at intermediate concentrations of ZnO-NP and non-nano ZnO, while for coated ZnO-NP such a clear peak only was seen after 12 months. Dose-related increases in soil pH may explain decreased soluble Zn levels due to fixation of Zn released from ZnO at higher soil concentrations. At T=0 uncoated ZnO NP and non-nano ZnO were equally toxic to the springtail Folsomia candida, but not as toxic as coated ZnO-NP, and ZnCl2 being most toxic. After three months equilibration toxicity to F. candida was already reduced for all Zn forms, except for coated ZnO-NP which showed reduced toxicity only after 12 months equilibration.

Pauline L. Waalewijn-Kool, Maria Diez Ortiz, Nico M. van Straalen and Cornelis A.M. van Gestel

Environmental Pollution 178, 2013, 59-64

58

Chapter 4

1 Introduction

Nanotechnology has developed an increasing number of nano-based products that are currently applied in textiles, electronics, pharmaceuticals and cosmetics. This has raised scientific and public concerns about the potential impact of nanomaterials on the environment. Zinc oxide nanoparticles (ZnO-NP) are among the most commonly used nanoparticles having different uses such as environmental remediation and sunscreen application (Wang, 2004; Osmond and McCall, 2010). Due to their increased use and disposal ZnO-NP are likely to enter the environment, with soil being a potential sink, posing a hazard to soil organisms (Tourinho et al., 2012). Adverse effects on the reproduction of earthworms were shown for ZnO-NP (Canas et al., 2011; Hooper et al., 2011).

Standardized tests for regular chemicals are useful for determining the toxicity of nanoparticles (Kahru and Dubourguier, 2010). Short-term toxicity tests, however, lack the ability to study environmental fate processes, such as dissolution and sorption. Zn ion release from ZnO-NP was shown to be relatively fast in water (Poynton et al., 2011) and kaolin suspensions (Scheckel et al., 2010). Dissolution of metal-oxide nanoparticles depends on surface area, which is larger for smaller particles (Borm et al., 2006). Different aquatic studies show that nanoparticles dissolve faster than larger sized materials of the same mass (Wong et al., 2010; Reed et al., 2012). The dissolution of nanoparticles in soil may be different compared to liquid media. Nanoparticles tend to aggregate and may form coatings over mineral surfaces (Theng and Yuan, 2008). Soil organic matter has a high binding capacity for metal oxides and influences the dissolution of nanoparticles. Soil pH may play an important role in the dissolution of the amphoteric ZnO (Bian et al., 2011). The aqueous solubility of ZnO ranges from several thousand mg per litre at pH 6 to around 1 mg/l at pH 8 (Apte et al., 2009). Also, a coating of nanoparticles is likely to influence dissolution in soils, by preventing the release of metal ions. The majority of the nanoparticles are produced with a coating, but data is lacking on the difference in toxicity between coated and uncoated ZnO-NP.

Long-term exposures in the environment have shown to decrease zinc toxicity in soil over time (Lock and Janssen, 2002; Smit et al., 1997). Currently, it is not known whether long-term exposure of ZnO-NP also reduces their bioavailability and potential toxicity in soil.

Collembola are an integral part of soil ecosystems and are vulnerable to the effects of soil contamination. Folsomia candida has been used as a model organism for more than 40 years (Fountain and Hopkin, 2005). The 28-day EC50 for the toxicity of ZnCl2 in Lufa 2.2 soil has been reported to be between 348 and 476 mg Zn/kg d.w. (Smit et al., 1997; Nota et al., 2010).

This study aims to determine the toxicity of coated and uncoated ZnO-NP, non-nano ZnO and ZnCl2 to F. candida in equilibrated soil. Considering an expected slow Zn release, natural soil was equilibrated for one year in the laboratory. Pore water was collected from

59

4

Long-term equilibration and toxicity of coated and uncoated ZnO nanoparticles in soil

soils freshly spiked and after three, six and twelve month equilibration. At each sampling

time the toxicity of the four Zn forms to F. candida was compared to explain ZnO-NP toxicity

by total Zn in the soil or by soluble Zn concentrations in the pore water. We hypothesized

that 1) ZnO-NP toxicity is attributable to soluble Zn concentrations rather than to particle

size, 2) the coating of ZnO-NP reduces its dissolution and thereby its toxicity to springtails

and 3) that equilibration leads to increased toxicity of ZnO due to an increased dissolution.


2.1 Test compounds

Coated ZnO-NP (Z-COTE®HP1) and uncoated ZnO-NP (Z-COTE®) powder, both

with a reported diameter of < 200 nm, were obtained from BASF SE (Ludwigshafen,

Germany). The mass fraction (w/w) of the coated ZnO-NP was 96-99% zinc oxide and

1-4% triethoxyoctylsilane (coating). Triethoxyoctylsilane (CAS 2943-75-1, colourless

liquid) was purchased from Sigma-Aldrich Chemie BV (≥ 97.5%). Non-nano ZnO (Merck,

pro analysi, > 99%) and ZnCl2 (Merck, zinc chloride pure) were used for comparison

of particle size and with free zinc, respectively. ZnO-NP were characterized using

Transmission Electron Microscopy (TEM) and Dynamic Light Scattering (DLS); see

Supporting Information (SI) Figures SI-1-4.

2.2 Soil

Loamy sand soil (LUFA-Speyer 2.2, Sp 2121, Germany, 2009) with pHCaCl2 5.5, total

organic carbon content 2.09%, cation exchange capacity 10.0 meq/100 g and water-

holding capacity (WHC) 46.5% was used. The soil was oven-dried at 60 °C overnight

prior to the experiments to eliminate undesired soil fauna.

2.3 Spiking the soil

The first experiment consisted of seven concentrations of uncoated ZnO-NP (nominal

range 100-6400 mg Zn/kg d.w.), five non-nano ZnO concentrations (400-6400 mg

Zn/kg d.w.), six ZnCl2 concentrations (100-3200 mg Zn/kg d.w.) and two controls

without added zinc. The test compounds were introduced into the soil as aqueous

solutions prepared in soil extracts (van der Ploeg et al., 2011). Soil-water suspensions

were prepared by mixing air-dried soil with deionized water (Milli-Q) using a soil-water

ratio of 2:5 (w/v), shaken at 180 rpm at ambient temperature for one hour, and filtered

under vacuum (Whatman filter paper, type 595). Uncoated ZnO-NP, non-nano ZnO

and ZnCl2 were added to the filtrates (see Chapter 3 for TEM photos of ZnO particles

in soil solutions), shaken for two days at 180 rpm and carefully mixed with dry soil. Soil

moisture content was adjusted to 23.3% (w/w) with deionized water (Milli-Q).

In the second experiment seven concentrations of coated ZnO-NP (nominal range

100-6400 mg Zn/kg d.w.), triethoxyoctylsilane and two controls without added zinc

60

Chapter 4

were tested. The hydrophobic triethoxyoctylsilane coating makes the coated ZnO-NP

insoluble in water, therefore they were mixed with the soil as dry powders. Soil moisture

content was adjusted to 23.3% (w/w) with deionized water (Milli-Q).

For each Zn treatment, a total amount of 600 g soil was prepared.

To investigate the toxicity of the coating, five concentrations of triethoxyoctylsilane

were tested in a 28-day toxicity test with F. candida. For this purpose, 20 g dry soil

was spiked with triethoxyoctylsilane dissolved in acetone. After evaporation of the

acetone, 180 g dry soil was added and soils were mixed with a spoon to reach nominal

concentrations (75-1200 mg/kg). Deionized water (Milli-Q) was added to achieve a

soil moisture content of 23.3% (w/w). An acetone control and a water control were

included in the toxicity test.

2.4 Conditions for equilibration of the soil

Soils were equilibrated in glass jars with the individual Zn forms for one year, in a climate

room at 20 ± 1 °C. Twice a month, soil moisture content was checked by weighing the

jars, and moisture loss was replenished with deionized water. Soil moisture content

never decreased more than 2.5% upon replenishing the water loss. Once every 6-8

weeks and upon sampling, soils were mixed with a spoon. No growth of fungi was

visible over the course of the experiment. The lids of the jars were not tightly closed

so the jars were aerated continuously. After three, six and twelve months the soil was

sampled to collect pore water and to perform a toxicity test with F. candida.

2.5 Toxicity test

Toxicity experiments were carried out following ISO-guideline 11267 (ISO, 1999) using

survival and reproduction of F. candida as effect parameters. F. candida (Berlin strain;

VU University Amsterdam) was cultured in plastic containers with a moist bottom of

plaster of Paris containing 10% charcoal, at 20 ± 1 °C and 12/12h light/dark. Each

experiment was initiated with age-synchronized 10-12 day old juveniles.

At the start of each test (8 in total, four time points in the two experiments), ten

synchronized animals were transferred into 100 ml glass test containers containing

30 g soil each. Five replicates for each test concentration and control were prepared.

The test jars were filled randomly and before introduction the animals were checked

under the microscope for a healthy appearance. At the beginning of the experiments

the animals were fed a few grains of dried baker’s yeast. The jars were incubated at 20

± 1 °C and 12/12h light/dark. Once a week, the moisture content of the test soils was

checked by weighing the jars, and moisture loss was replenished with deionized water

when necessary. The jars were also aerated by this procedure. After four weeks, the

jars were emptied into a 200 ml beaker glass and 100 ml tap water was added. The

mixture was stirred carefully to let all the animals float to the surface. The number of

adults and juveniles were counted manually after taking a picture of the water surface

using a digital camera (Olympus, C-5060).

61

4


2.6 Zinc analysesAfter spiking and after one year, soil samples were oven-dried at 60 °C for total zinc

analysis. Approximately 100 mg dried soil (two replicates per test concentration and

time point) was digested in a mixture of deionized water (Milli-Q), concentrated HCl

and concentrated HNO3 (1:1:4 by vol.), in tightly closed teflon bombs in an oven

(CEM MDS 81-D) at 140 °C for 7 hours. The solution was analysed by flame Atomic

Absorption Spectrometry (AAS) (Perkin Elmer AAnalyst 100). Certified reference

material (ISE sample 989, River Clay from Wageningen, The Netherlands) was used to

ensure the accuracy of the analytical procedure. Measured zinc concentrations in the

reference material were within 10% of the certified concentrations.

At each sampling time, pore water from each test soil was collected by centrifuging

50 g soil (Centrifuge Falcon 6/300 series, CFC Free), after saturation with deionized

water and one week equilibration time (Tipping et al., 2003). Soils were centrifuged

for 50 min. at 2000 g over two round filters (S&S 597 Ø 47 mm, pore size 11 μm) and

a 0.45 μm membrane filter (S&S Ø 47 mm), placed inside the tubes (Hobbelen et al.,

2004). Pore water was analysed by flame AAS (Perkin Elmer AAnalyst 100). In case

of uncoated ZnO-NP, zinc concentrations in the pore water were also determined

by flame AAS after ultrafiltration to obtain a particle-free extract. Soil solutions were

centrifuged in a 100 kDa ultrafiltration device (Amicon Ultra-15 Filters, Millipore)

for 20 min. at 2000 g. This did not show differences in Zn porewater concentrations

before and after ultrafiltration for the uncoated ZnO-NP. Because of that and because

we expected the coated ZnO-NP to be less prone to aggregation, we tried the use of

3 kDa ultrafiltration devices for coated ZnO-NP.

The pH in the pore water was only measured at the beginning of the first

experiment. After six and twelve month equilibration, pHwater of the soil was measured.

In freshly spiked soil with coated ZnO-NP soil pHwater was also measured. Soils were

shaken with deionized water (5:1 liquid:soil) for 2 hours at 200 rpm. The pHCaCl2 of

soils spiked with triethoxyoctylsilane and of soils equilibrated with coated ZnO-NP for

twelve months were measured by shaking them with 0.01M CaCl2 (5:1 liquid:soil) for 2

hours at 200 rpm. pH was recorded using a Consort P907 meter.

2.7 Data analysisLC/EC50, the concentrations in soil and pore water causing 50% reduction in springtail

survival and reproduction, respectively, were estimated applying the logistic model

of Haanstra et al. (1985). A generalized likelihood ratio test (Sokal and Rohlf, 1995)

was performed to compare EC50 values obtained at different equilibration times. All

calculations were performed in SPSS 17.

Zn speciation was modelled for the lowest and highest Zn concentrations measured

in the pore water using Visual MINTEQ (http://www.lwr.kth.se/English/OurSoftware/

vminteq; Gustafsson, 2007). Speciation was also calculated for the concentrations

closest to the EC50 values.

62

Chapter 4

3 Results

This study is a follow-up of a previous experiment with ZnO-NP, non-nano ZnO and

ZnCl2. The results of the 28-days toxicity tests with F. candida in freshly spiked Lufa 2.2

soil are described in Chapter 3. These data represent time point zero (T=0) in this study.

3.1 Total soil analysis and pH measurements

Soil zinc concentrations measured at the start and end of the experiment are presented

in Table SI-1. Recovery was ≥ 91% for all soil samples at both sampling times. All

results are expressed on the basis of measured concentrations.

The pH slightly increased with increasing soil concentrations for coated and

uncoated ZnO-NP, and non-nano ZnO, and dose-related decreased for ZnCl2 at all

time points (Table SI-2). For all Zn forms, except for the lower ZnCl2 concentrations, pH

showed a steady decrease with time. The pHCaCl2 for soil spiked with triethoxyoctylsilane

was approx. 4.83 for all test concentrations.

3.2 Soil pore water analysis

3.2.1 ZnCl2Zn concentrations in pore water collected from soils equilibrated with ZnCl2 increased with

exposure concentration at all time points (Table SI-3). The maximum concentration was

612 mg Zn/l in freshly spiked soil at a nominal soil concentration of 1600 mg Zn/kg dry soil.

In equilibrated soil the maximum amounted to 2520, 2700 and 2196 mg Zn/l at a nominal

soil concentration of 3200 mg Zn/kg dry soil after 3, 6 and 12 months, respectively.

3.2.2 Uncoated ZnO-NP and non-nano ZnOZinc concentrations in pore water increased from 1.85 to 12.6 mg Zn/l in soil freshly

spiked with uncoated ZnO-NP with increasing soil concentrations (Figure 1, Table SI-3).

After three, six and twelve months, porewater concentrations for uncoated ZnO-NP

and non-nano ZnO increased with time, but peaked at intermediate concentrations.

The highest Zn concentration measured after one year was 67.1 mg Zn/l in soil at the

measured total concentration of 1027 mg Zn/kg for uncoated ZnO-NP. Ultrafiltration

did not affect the zinc concentrations, suggesting that no intact nanoparticles

were present in the soil solutions. For non-nano ZnO a maximum of 66.5 mg Zn/l

was measured in the pore water after one year from soil containing a measured

concentration of 941 mg Zn/kg.

3.2.3 Coated ZnO-NPFor the coated ZnO-NP, zinc concentrations in the pore water increased with increasing

soil concentrations during the first three months of equilibration (Figure 1; Table SI-3). At

the highest test concentration, 6.10 mg Zn/l was measured in the pore water in freshly

spiked soil, which increased to 22.5, 20.9 and 23.0 mg Zn/l after three, six and twelve

63

4


months, respectively. The maximum porewater concentration after six and twelve month equilibration was measured in soils spiked with 1600 mg Zn/kg, and was 26.6 and 36.6 mg Zn/l, respectively. These porewater Zn concentrations were lower than the ones measured for uncoated ZnO-NP at all time points and for coated ZnO-NP a substantial peak at intermediate concentrations was not observed until twelve months equilibration.

3.3 Ecotoxicity to Folsomia candidaEC50 values estimated for the four Zn forms are presented based on measured Zn concentrations in soil (Table 1) and on porewater concentrations (Table 2). Corresponding dose-response curves for coated and uncoated ZnO-NP, non-nano ZnO and ZnCl2 are shown in Figure SI-5.

In the first experiment (including uncoated ZnO-NP, non-nano ZnO and ZnCl2), control survival of the Collembola in the four toxicity tests with freshly spiked, three, six and twelve month equilibrated soil was 72, 88, 87 and 62%, respectively. The number of juveniles in the control with three month equilibrated soil was only 58 (± 18.3, n=10), but since survival was normal and consistent dose-response curves were seen the test was considered valid. In the controls with freshly spiked and with six and twelve month equilibrated soil the number of juveniles exceeded the minimum of 100 per test container set by ISO (1999), and was 140 (± 56, n=5), 494 (± 167, n=10) and 228 (± 122, n=10), respectively. Uncoated ZnO-NP and non-nano ZnO were toxic in freshly spiked soil with no significant difference between the two ZnO powders. No effect on collembolan survival was found in soil equilibrated for three, six or twelve months with uncoated ZnO-NP and non-nano ZnO. After three months equilibration the EC50 of uncoated ZnO-NP increased from 1964 to 2847 mg Zn/kg based on total soil concentrations and from 10.1 to 39.9 mg Zn/l based on porewater concentrations. After six months equilibration no EC50 could be estimated for ZnO-NP due to a flat dose-response curve and very high juvenile numbers in the control. However, almost

coated ZnO-NP uncoated ZnO-NP non-nano ZnO

measured concentration soil (mg Zn/kg d.w.)

pore

wat

erco

ncen

tratio

n(m

g Zn

/l)

0

10

20

30

40

50

60

70

0 2000 4000 6000 0 1000 2000 3000 4000 5000 6000 0 2000 4000 6000 8000

T=0T=3T=6T=12

Figure 1. Zinc concentrations measured in pore water (mg Zn/l) as a function of total zinc concentrations in Lufa 2.2 soil (mg Zn/kg d.w.) freshly spiked with coated and uncoated ZnO-NP, non-nano ZnO and ZnCl2 (T=0) and after three (T=3), six (T=6) and twelve months (T=12) equilibration.

64

Chapter 4

Table 1. EC50 values for the effect on the reproduction of Folsomia candida after 28-d exposure to Lufa 2.2 soil freshly spiked (T=0) with coated and uncoated ZnO-NP, non-nano ZnO and ZnCl2 and after three (T=3), six (T=6) and twelve months (T=12) equilibration. EC50 values are presented as total concentrations in the soil (mg Zn/kg d.w.). Corresponding 95% confidence intervals are presented in between brackets.

Time (months) coated ZnO-NP uncoated ZnO-NP non-nano ZnO ZnCl2

T=0 873a

(659-1087)1964

(1635-2293)1591

(-)299a

(181-415)

T=3 749a

(463-1035)2847

(-)3628<EC50<8359 912b

(-)

T=6 576a

(263-888)- - -

T=12 1817b

(1344-2291)>5855* >8359* 707b

(419-996)

* no 50% reduction in survival or reproduction was observed at the highest test concentration- Data did not allow estimating an EC50 value and/or 95% confidence intervals (see text)a,b indicate significant differences between LC/EC50 values at different time points according to a generalized likelihood-ratio test (χ2

(1) > 3.84; p < 0.05).

Table 2. EC50 values for the effect on the reproduction of Folsomia candida after 28-d exposure to Lufa 2.2 soil freshly spiked (T=0) with coated and uncoated ZnO-NP, non-nano ZnO and ZnCl2 and after three (T=3), six (T=6) and twelve (T=12) months equilibration. EC50 values are presented as soluble Zn levels in the pore water (mg Zn/l). Corresponding 95% confidence intervals are presented in between brackets.

Time (months) coated ZnO-NP uncoated ZnO-NP non-nano ZnO ZnCl2

T=0 4.08a

(3.86-4.30)10.1

(7.83-12.4)7.94(-)

16.8a

(-)

T=3 16.5b

(14.4-18.7)39.9(-)

- 180b

(-)

T=6 15.8b

(12.4-19.2)- 31.3

(11.6-51.0)-

T=12 18.0b

(1.85-34.1)>67.1* >66.5* 178b

(142-213)

* no 50% reduction in survival or reproduction was observed at the highest test concentration- Data did not allow estimating an EC50 value and/or 95% confidence intervals (see text)a,b indicate significant differences between EC50 values at different time points according to a generalized likelihood-ratio test (χ2

(1) > 3.84; p < 0.05).

50% reduction in reproduction was observed for the lowest test concentration (i.e. 116 mg Zn/kg), which was considered unrealistic. After twelve months equilibration, the reduction in reproduction was highest (46%) at 1027 mg Zn/kg, which contained the highest porewater concentration (i.e. 67.1 mg Zn/l).

65

4


For non-nano ZnO 37% reduction in reproduction was observed at 3628 mg Zn/kg and 65% at 8359 mg Zn/kg after three months equilibration. The data did not allow estimating an EC50, but 50% effect on reproduction is expected between these two exposure concentrations. After six months, no EC50 could be calculated but 62% effect was observed at the highest soil concentration. It was possible to calculate an EC50 of 31.3 mg Zn/l based on porewater concentrations. After 12 months, only 40% reduction in reproduction was found at the highest exposure concentration and therefore no EC50 was calculated. Overall, the toxicity tests with ZnO showed a decrease in springtail toxicity with time, which continued for one year.

The toxicity of ZnCl2 decreased with time shown by an increase in LC50 values from 955 (95% confidence interval 819-1114) to 1632 (1336-1927) and 1943 (1300-2586) mg Zn/kg in freshly spiked soil and after 3 and 6 months, respectively. After one year equilibration of ZnCl2-spiked soils springtail survival was no longer affected (LC50 > highest test concentration). The effect on reproduction also decreased with time shown by a three-fold increase in EC50 after three months equilibration. EC50 values were lowest for freshly spiked soil and increased after equilibration, also when expressed as porewater Zn concentrations. For the test with six month equilibrated soil no EC50 could be estimated due to the very high juvenile numbers in the control. Reproduction was already reduced by 65% at the lowest test concentration (123 mg Zn/kg), which was considered to be unrealistic and therefore no EC50 was calculated.

In the second experiment (including coated ZnO-NP), survival of the Collembola in the controls of the four toxicity tests performed with freshly spiked soil, three, six and twelve month equilibrated soil was 88, 85, 57 and 85%, respectively. The numbers of juveniles in the control with freshly spiked, three, six and twelve month equilibrated soil were 483 (± 96, n=10), 331 (± 142, n=10), 258 (± 114, n=10), 315 (± 96, n=10), respectively. No effect on collembolan survival was found in soil freshly spiked or equilibrated with coated ZnO-NP. The effect on reproduction in freshly spiked, three and six month equilibrated soil was similar, as shown by the EC50 values of 873, 749 and 576 mg Zn/kg, respectively. After one year equilibration the EC50 increased to 1817 mg Zn/kg. Based on soluble Zn concentrations the EC50 values increased from 4.08 to 15.8-18.0 mg Zn/l after three to twelve months equilibration. For the effect on collembolan survival and reproduction of triethoxyoctylsilane a LC50 of 653 mg/kg (95% CI = 514-793 mg/kg), an EC50 of 505 mg/kg (376-633 mg/kg) and an EC10 of 271 (106-437 mg/kg) were estimated (Figure SI-6).

4 Discussion

Several studies demonstrate that ZnO rapidly dissolves in soils. Priester et al. (2012) found high Zn accumulation in the leaves of soybeans after 50 days exposure to ZnO-NP in a soil at pH 6.78. As they measured similar Zn values for ZnCl2-exposed plants,

66

Chapter 4

this suggests most of the ZnO was dissolved. In soils with pH < 7 ZnO dissociates into free Zn2+ ions (Bodar et al., 2005). After dissociation, adsorption of Zn to various soil components such as organic matter, clay minerals and Fe and Mn (hydr)oxides occurs. Voegelin et al. (2005) showed a rapid dissolution of ZnO in a near neutral non-calcareous soil (pH 6.5) and rapid formation of Zn precipitates. They also found that dissolved Zn species in the soil do not diffuse from the solid phase where these Zn precipitates are formed rapidly. In soils at field capacity, slow diffusion at the surfaces of metal oxides and humics inhibits further dissolution of ZnO (Crout et al., 2006). This has also been observed in kaolin suspensions at pH 6, where ZnO-NP rapidly converted to Zn2+ that was bound in sorption complexes within one day and these sorption complexes were maintained through a 12-month ageing period (Scheckel et al., 2010). Donner et al. (2010) spiked different field soils with ZnSO4 and aged them for up to two years and found a decrease of available Zn with time. So, it can be assumed that ZnO-NP rapidly dissolved in soil and that the dissolution of ZnO was counteracted by fixation of Zn over time.

In this study, the porewater concentrations showed an interesting trend in Zn dissolution from both uncoated ZnO-NP and non-nano ZnO, with increasing Zn dissolution with time at low concentrations but not at high doses. Formation of agglomerates at high exposure concentrations could have hindered dissolution. However, it is more likely that the slightly increased pH levels with increasing exposure concentrations have reduced ZnO solubility. At higher soil concentrations there may be a higher degree of precipitation of released Zn ions, with sulphate, phosphate and other ions (McLaughlin, 2002) and metal fixation increases with pH (Crout et al., 2006). During the one-year equilibrium period the pH slightly decreased with time, explaining the increase of porewater concentrations with time.

Several studies have investigated the influence of primary particle size on NP solubility in water. David et al. (2012) observed similar dissolution rates of ZnO-NP with a primary particle diameter above 20 nm and non-nano ZnO, but a higher solubility for 6 nm ZnO-NP. Bian et al. (2011) showed that 15 nm ZnO-NP dissolve more readily in water than 240 nm ZnO. In our soils we found no clear differences in Zn dissolution between soils spiked with uncoated ZnO-NP and non-nano ZnO, despite the fact that the ZnO-NP were more than 50 times smaller than non-nano ZnO in the initial powders. So, the total surface area or primary particle size does not seem to play a major role in the dissolution of this type of ZnO-NP and this type of soil.

We did observe lower porewater Zn levels for coated ZnO-NP. In general, a surface coating is used to prevent aggregation of nanoparticles by steric or electrostatic mechanisms (Christian et al., 2008). Soluble Zn concentrations similar to the ones obtained from uncoated ZnO-NP after three months were reached only after twelve months equilibration. This suggests that a surface coating could delay the dissolution of ZnO-NP in soil. However, it is more likely that the higher pH levels for soils with coated ZnO-NP compared to the uncoated ZnO-NP have caused the lower Zn levels. The fact that the

67

4


peak in dissolved Zn appeared much later in case of coated ZnO-NP also suggests that the coating must have prohibited Zn release, and that it took at least several months to degrade the coating material. This seems in line with the half-life for triethoxyoctylsilane in soil of approx. 30 days estimated by a level III fugacity model in EPI-SuiteTM.

Both types of ZnO-NP (Z-COTE® and Z-COTE®HP1) have been tested for acute effects on Daphnia magna. Wiench et al. (2009) report 48-h EC50 values of 7.5 and 1.1 mg Zn/l for the uncoated and coated ZnO-NP, respectively. Also, in our toxicity tests EC50 values were higher for uncoated than for coated ZnO-NP, both when expressed on the basis of total soil concentrations and on porewater concentrations (1964 mg Zn/kg or 10.1 mg Zn/l versus 873 mg Zn/kg or 4.08 mg Zn/l; at T=0). The porewater-based EC50 values are in the same range as the values reported by Wiench et al. (2009) for daphnids. The hydrophobic character of the coated NP may have enhanced its toxicity. The coating triethoxyoctylsilane, which represents approx. 3% of the ZnO-NP, may have contributed to the overall toxicity. When applying a mixture toxicity approach, taking into account that on a mass basis the coating represented 3.85% of the Zn mass, it may be assumed that the EC50 for coated ZnO-NP of 873 mg Zn/kg dry soil corresponds with a triethoxyoctylsilane concentration of 33.6 mg/kg dry soil. Compared with the EC50 values for uncoated ZnO-NP and triethoxyoctylsilane these concentrations correspond with 0.44 and 0.067 Toxic Units (TU), respectively. This means that the toxic strength of the mixture equals 0.5 TU, so the coated ZnO-NP are much more toxic than expected from the EC50 values of uncoated ZnO-NP and triethoxyoctylsilane. The decrease of toxicity of the coated ZnO-NP at the end of the 12-month equilibration period may be explained from the loss of the coating.

In our ZnCl2-spiked soils toxicity decreased with time, while the dissolved Zn increased from 17.1% (of total Zn concentration at the highest spiking concentration) for freshly spiked soil to 34.9, 37.4 and 30.4% after three, six and twelve months equilibration, respectively. For the T=0 samples and the samples from the test with coated ZnO-NP, it seemed that toxicity could fairly well be explained from the Zn concentration in pore water, with EC50 values ranging between 8 and 20 mg Zn/l (Table 2). At later sampling times, toxicity of the uncoated ZnO-NP, non-nano ZnO and ZnCl2 could however, no longer be explained from porewater Zn concentrations. Since this might be due to complexation of Zn in the pore water, we modelled Zn speciation using Visual MINTEQ for measured soluble Zn concentrations of 1.18 and 2700 mg Zn/l (lowest and highest values measured in pore water). Speciation was also calculated for concentrations of 49.2 and 266 mg Zn/l which are closest to the EC50 values estimated at time zero and after six months, using a measured dissolved organic carbon content (Nica-Donnan DOC) of 315 mg/l for Lufa 2.2 soil (t = 20 °C). For these speciation calculations, we also needed porewater pH values. Unfortunately, we applied different methods for determining pH of the soil and pore water. Based on pHwater of pHCaCl2 values, porewater pH values were estimated using regressions derived by De Vries et al. (2005). These porewater pH values (Table SI-2),

68

Chapter 4

which showed a consistent decline with time, were used for speciation calculations. For ZnCl2 concentrations up to 49.2 mg Zn/l more than 95% of the dissolved Zn was estimated to be bound to DOC and not available in solution as Zn2+. At 266 mg Zn/l 65.3% was available as Zn2+, and at 2700 mg Zn/l 80.9% (Table SI-4). The increase in free zinc concentrations in soil solution coincides with a decrease in pH. According to the principles of the Biotic Ligand Model, competition of H+ ions and free zinc on the uptake sites of the animal therefore may explain the decreased toxicity of ZnCl2 over time (Thakali et al., 2006). For the ZnO-NP and non-nano ZnO in our study, all Zn measured in the pore water was bound to DOC. When plotting the estimated EC50 values based on porewater concentrations against pH, a significant negative correlation was found for all four Zn forms (Figure SI-7). For the coated ZnO-NP, EC50 with pH showed a less steep decline, which may be attributed to the degradation of the coating which apparently delayed release of Zn.

In this study we investigated the effect of time on ZnO-NP dissolution in soil and found that pH is a main factor. We conclude that the release of Zn in soils spiked with coated and uncoated ZnO-NP continued for one year and that the fixation of Zn contributed to a reduced bioavailability and springtail toxicity with time.


• ZnO nanoparticles characterization (Figures SI-1-4) • Dose-response curves for coated ZnO-NP (Figure SI-5a), uncoated ZnO-NP (Figure

SI-5b), non-nano ZnO (Figure SI-5c) and ZnCl2 (Figure SI-5d) • Toxicity of triethoxyoctylsilane (Figure SI-6) • EC50 values in relation to pH (Figure SI-7) • Measured zinc concentrations in soil (Table SI-1) • pH measurements of soil and pore water (Table SI-2) • Measured zinc concentrations in pore water (Table SI-3) • Modelled zinc speciation in pore water (Table SI-4)

69

4


ZnO nanoparticle characterizationCharacterization of the BASF (Z-COTE® and Z-COTE®HP1) ZnO nanoparticles was performed by Transmission Electron Microscopy (TEM) and Dynamic Light Scattering (DLS) in the laboratory of the University of Oxford.

For the uncoated ZnO, a concentration of approximately 1 mg ZnO/ml was dispersed in water and sonicated for 30 sec in low power US bath. A drop (20 µl) was deposited on a carbon-coated Cu TEM grid. Samples were dried at room temperature for several hours before examination in the TEM. The coated ZnO nanoparticles have a hydrophobic coating consisting of triethoxyoctylsilane which enables them to be easily incorporated into the oil phase of a formulation but they do not whet in water. So the ZnO nanoparticles were deposited on a TEM grid by dipping the grid through layer of particles on the surface of water. Experiments were carried out on a JEOL 2010 analytical TEM, which has a LaB6 electron gun and can be operated between 80 and 200 kV. This instrument has a resolution of 0.19 nm, an electron probe size down to 0.5 nm and a maximum specimen tilt of ± 10 degrees along both axes. The instrument is equipped with an Oxford Instruments LZ5 windowless Energy Dispersive X-ray Spectrometer (EDS) controlled by INCA. On the TEM images a small (approx. 20-25) number of particles were measured from about 4 or 5 TEM micrographs to get a rough particle size distribution for the primary particles. Each particle was measured individually from the TEM micrographs using Digital Micrograph program, which is a standard TEM instrument control and analysis program. The figures show particle size distribution after 10 minutes, measured by DLS.

0

10

20

30

0.1 1 10 100 1000 10000

Num

ber (

%)

Size (d.nm)

Size Distribution by Number

Record 1: Zcote 10min USbath 1 Record 2: Zcote 10min USbath 2Record 3: Zcote 10min USbath 3

Figure SI-1. Particle size distribution of uncoated ZnO-NP measured by Dynamic Light Scattering (DLS) after 10 minutes.

70

Chapter 4

0

10

20

30

40

0.1 1 10 100 1000 10000

Num

ber (

%)

Size (d.nm)


Record 4: Zcote HPI 10min USbath 1 Record 5: Zcote HPI 10min USbath 2Record 6: Zcote HPI 10min USbath 3

Figure SI-2. Particle size distribution of coated ZnO-NP measured by Dynamic Light Scattering (DLS) after 10 minutes.

0

10

20

30

40

0.1 1 10 100 1000 10000

Num

ber (

%)

Size (d.nm)



0

10

20

30

40

0.1 1 10 100 1000 10000

Num

ber (

%)

Size (d.nm)



Figure SI-3. Transmission Electron Microscopy (TEM) of uncoated ZnO-NP showed primary particles sizes of 30-300 nm, with a wide range of aspect ratios, some separated primary particles and some aggregates from a few particles to micron sized ZnO.

Figure SI-4. Transmission Electron Microscopy (TEM) of coated ZnO-NP showed primary particles sizes of 50-500 nm, almost all in aggregates.

71

4


0

100

200

300

400

500

600

0.01 0.1 1 10 100

0

100

200

300

400

500

600

0

100

200

300

400

500

6000

100

200

300

400

500

600

700

0

100

200

300

400

500

600

1 10 100 1000 10000

0

100

200

300

400

500

600

0

100

200

300

400

500

6000

100

200

300

400

500

600

700nu

mbe

rof j

uven

ilesp

er te

st ja

r

Zn concentration soil (mg Zn/kg d.w.)

T=3

T=6

T=12

T=0

porewater Zn concentration (mg Zn/l)

num

bero

f juv

enile

sper

test

jar

T=0

T=3

T=6

T=12

Figure SI-5a. Effect of coated ZnO-NP on the reproduction (number of juveniles) of Folsomia candida after 28-d exposure in Lufa 2.2 soil freshly spiked (T=0) and after three (T=3), six (T=6) and twelve (T=12) months equilibration. Measured concentrations of zinc in the soil (left) and the soil pore water (right) are provided on the x-axis. Line shows fit obtained with a logistic model.

72

Chapter 4

0

100

200

300

400

500

0.01 0.1 1 10 100

0

200

400

600

800

10000

20

40

60

80

100

1200

50

100

150

200

250

0

20

40

60

80

100

1200

50

100

150

200

250

0

100

200

300

400

500

1 10 100 1000 10000

0

200

400

600

800

1000

num

bero

f juv

enile

sper

test

jar

Zn concentration soil (mg Zn/kg d.w.) porewater Zn concentration (mg Zn/l)

num

bero

f juv

enile

sper

test

jar

T=0

T=3

T=6

T=12

T=0

T=3

T=6

T=12

Figure SI-5b. Effect of uncoated ZnO-NP on the reproduction (number of juveniles) of Folsomia candida after 28-d exposure in Lufa 2.2 soil freshly spiked (T=0) and after three (T=3), six (T=6) and twelve (T=12) months equilibration. Measured concentrations of zinc in the soil (left) and the soil pore water (right) are provided on the x-axis. Line shows fit obtained with a logistic model.

73

4


0

100

200

300

400

500

600

0.01 0.1 1 10 100

0

200

400

600

800

1000

0

20

40

60

80

100

0

50

100

150

200

250

300

0

100

200

300

400

500

600

1 10 100 1000 10000

0

200

400

600

800

1000

0

20

40

60

80

100

0

50

100

150

200

250

300nu

mbe

rof j

uven

ilesp

er te

st ja

r


num

bero

f juv

enile

sper

test

jar

T=0

T=3

T=6

T=12T=12

T=6

T=3

T=0

Figure SI-5c. Effect of non-nano ZnO on the reproduction (number of juveniles) of Folsomia candida after 28-d exposure in Lufa 2.2 soil freshly spiked (T=0) and after three (T=3), six (T=6) and twelve (T=12) months equilibration. Measured concentrations of zinc in the soil (left) and the soil pore water (right) are provided on the x-axis. Line shows fit obtained with a logistic model.

74

Chapter 4

0

100

200

300

400

500

0.01 0.1 1 10 100 1000 10000

0

200

400

600

800

1000

0

20

40

60

80

100

120

140

0

50

100

150

200

250

0

100

200

300

400

500

10 100 1000 10000

0

200

400

600

800

10000

20

40

60

80

100

120

140

0

50

100

150

200

250nu

mbe

rof j

uven

ilesp

er te

st ja

r


num

bero

f juv

enile

sper

test

jar

T=0

T=3

T=6

T=12

T=0

T=3

T=6

T=12

Figure SI-5d. Effect of ZnCl2 on the reproduction (number of juveniles) of Folsomia candida after 28-d exposure in Lufa 2.2 soil freshly spiked (T=0) and after three (T=3), six (T=6) and twelve (T=12) months equilibration. Measured concentrations of zinc in the soil (left) and the soil pore water (right) are provided on the x-axis. Line shows fit obtained with a logistic model.

75

4


0

2

4

6

8

10

12

0 500 1000 1500nominal concentration soil (mg/kg d.w.)

num

bero

f sur

vivor

s

050

100150200250300350400450500

0 500 1000 1500nominal concentration soil (mg/kg d.w.)

num

bero

f juv

enile

s

0

50

100

150

200

4.5 5.5 6.5 7.5

EC50

pore

wat

er(m

g Zn

/l)

pHporewater

coated ZnO-NPuncoated ZnO-NPnon-nano ZnOZnCl2

Figure SI-6. Effect of triethoxyoctylsilane on the survival (top) and reproduction (bottom) of Folsomia candida after 28-d exposure in Lufa 2.2 soil. Nominal exposure concentrations in the soil are provided on the x-axis. Lines show fit obtained with a three parameter logistic dose-response model.

Figure SI-7. Estimated EC50 values (mg Zn/l) for the effect of Zn on the reproduction of Folsomia candida based on porewater concentrations in relation to (estimated) pHporewater (see Table SI-2 for measured and estimated pH values). Data used from toxicity tests in Lufa 2.2 soil freshly spiked or equilibrated with coated ZnO-NP, uncoated ZnO-NP, non-nano ZnO and ZnCl2 (see Table 2).

76

Chapter 4

Tab

le S

I-1.

Ave

rag

e zi

nc c

onc

entr

atio

n (n

=4)

mea

sure

d i

n Lu

fa 2

.2 s

oil

spik

ed w

ith c

oat

ed Z

nO-N

P, u

nco

ated

ZnO

-NP,

no

n-na

no Z

nO a

nd Z

nCl 2

mea

sure

d a

t th

e st

art

(T=

0, n

=2)

and

aft

er t

wel

ve m

ont

hs e

qui

libra

tion

(T=

12, n

=2)

. Rec

ove

ries

(%) a

re p

rese

nted

in b

etw

een

bra

cket

s.

No

min

al [Z

n]

(mg

Zn/

kg d

.w.)

Mea

sure

d [Z

n] (m

g Z

n/kg

dry

so

il)

coat

ed Z

nO-N

Pun

coat

ed Z

nO-N

Pno

n-na

no Z

nOZ

nCl 2

T=0

(n

=2)

T=12

(n=

2)A

vera

ge

(n=

4)T=

0(n

=2)

T=12

(n=

2)A

vera

ge

(n=

4)T=

0(n

=2)

T=12

(n=

2)A

vera

ge

(n=

4)T=

0(n

=2)

T=12

(n=

2)A

vera

ge

(n=

4)

Co

ntro

l14

.018

.716

.332

.817

.925

.4

100

126

120

123

(123

)12

210

911

6 (1

16)

123

118

120

(120

)

200

240

257

249

(124

)26

926

326

6 (1

33)

237

221

229

(114

)

400

448

467

458

(114

)48

651

349

9 (1

25)

511

532

522

(130

)43

045

944

4 (1

11)

800

916

883

899

(112

)10

8097

410

27 (1

28)

901

981

941

(118

)99

183

191

1 (1

14)

1600

1695

1619

1657

(104

)18

2415

7617

00 (1

06)

1997

1798

1897

(119

)16

4216

9816

70 (1

04)

3200

2827

3409

3118

(97)

3393

3250

3321

(104

)38

5634

0036

28 (1

13)

3389

3103

3246

(101

)

6400

5764

7646

6705

(105

)52

3764

7358

55 (9

1)80

7586

4383

59 (1

31)

77

4


Tab

le S

I-2.

pH

of

Lufa

2.2

so

il (p

Hw

ater o

r p

HC

aCl2) o

r p

ore

wat

er (p

Hp

ore

wat

er) o

f so

il fr

eshl

y sp

iked

with

co

ated

ZnO

-NP,

unc

oat

ed Z

nO-N

P, n

on-

nano

ZnO

an

d Z

nCl 2

(T=

0) a

nd a

fter

six

(T=

6) a

nd t

wel

ve (T

=12

) mo

nths

eq

uilib

ratio

n. p

H v

alue

s m

easu

red

in t

he s

oil

are

the

aver

age

of t

wo

rep

licat

es. E

stim

ated

p

Hp

ore

wat

er v

alue

s, w

ritte

n in

bo

ld, w

ere

ob

tain

ed u

sing

a r

egre

ssio

n d

eriv

ed b

y D

e Vr

ies

et a

l. (2

005)

(pH

po

rew

ater

= 0

.958

* p

Hw

ater –

0.0

246

or

pH

po

rew

ater

=

0.62

2 *

pH

CaC

l2 +

2.3

3). p

H v

alue

s m

easu

red

in t

he p

ore

wat

er c

olle

cted

fro

m fr

eshl

y sp

iked

Luf

a 2.

2 so

il sp

iked

with

unc

oat

ed Z

nO-N

P, n

on-

nano

ZnO

an

d Z

nCl 2

afte

r ul

trafi

ltrat

ion

are

writ

ten

in b

etw

een

bra

cket

s.

No

m. [

Zn]

so

il(m

g Z

n/kg

d.w

.)

coat

ed Z

nO-N

Pun

coat

ed Z

nO-N

Pno

n-na

no Z

nOZ

nCl 2

pH

wat

er

T=0

T=6*

pH

CaC

l2

T=12

pH

po

rew

ater

T=0

pH

wat

er

T=6

pH

wat

er

T=12

pH

po

rew

ater

T=0

pH

wat

er

T=6

pH

wat

er

T=12

pH

po

rew

ater

T=0

pH

wat

er

T=6

pH

wat

er

T=12

Co

ntro

l6.

115.

835.

065.

476.

53(6

.94)

5.35

5.

105.

285.

03

100

6.11

5.83

4.98

5.42

6.73

(7.1

5)5.

53

5.27

5.18

4.94

7.46

(7.8

0)5.

28

5.03

5.50

5.

25

200

6.18

5.90

4.96

5.41

6.84

(7.2

5)5.

55

5.29

5.19

4.95

6.97

(7.5

2)5.

15

4.91

5.28

5.

03

400

6.36

6.07

5.02

5.45

6.83

(7.2

0)5.

63

5.37

5.25

5.01

7.02

(7.8

5)5.

97

5.70

5.70

5.44

6.40

(7.0

3)5.

02

4.79

5.18

4.

94

800

6.54

6.24

5.10

5.50

7.08

(7.4

4)5.

83

5.56

5.37

5.12

7.30

(8.0

0)5.

94

5.67

5.64

5.38

5.86

(6.7

3)5.

49

5.24

4.98

4.

75

1600

6.68

6.38

5.22

5.57

7.23

(7.3

5)6.

32

6.03

5.53

5.27

7.26

(7.8

4)6.

35

6.06

5.90

5.63

5.38

(6.3

1)5.

22

4.98

4.93

4.

70

3200

6.82

6.51

5.29

5.62

7.38

(7.5

0)6.

53

6.23

5.80

5.53

7.50

(8.0

0)6.

76

6.45

6.18

5.90

4.96

4.

734.

70

4.48

6400

6.97

6.65

5.44

5.71

7.32

(7.5

7)6.

86

6.55

6.14

5.86

7.43

(7.7

3)6.

86

6.55

6.24

5.95

* N

o p

H m

easu

rem

ent

per

form

ed fo

r th

is t

ime

po

int

78

Chapter 4

Tab

le S

I-3.

Zin

c co

ncen

trat

ions

mea

sure

d i

n th

e p

ore

wat

er (

mg

Zn/

l) o

f Lu

fa 2

.2 s

oil

spik

ed w

ith c

oat

ed Z

nO-N

P, u

nco

ated

ZnO

-NP,

no

n-na

no Z

nO

and

ZnC

l 2 d

irect

ly a

fter

sp

ikin

g (T

=0)

and

aft

er t

hree

(T=

3), s

ix (T

=6)

and

tw

elve

(T=

12) m

ont

hs e

qui

libra

tion.

Po

re w

ater

was

co

llect

ed o

ne w

eek

afte

r sa

tura

tion

of t

he s

oils

with

Mill

i-Q w

ater

. Zin

c co

ncen

trat

ions

mea

sure

d in

the

po

re w

ater

aft

er u

ltrafi

ltrat

ion

are

pre

sent

ed in

bet

wee

n b

rack

ets.

No

m. [

Zn]

so

il(m

g Z

n/kg

d.w

.)

Mea

sure

d [Z

n] p

ore

wat

er (m

g Z

n/l)

coat

ed Z

nO-N

Pun

coat

ed Z

nO-N

Pno

n-na

no Z

nOZ

nCl 2

T=0

T=3

T=6

T=12

T=0

T=3

T=6

T=12

T=0

T=3

T=6

T=12

T=0

T=3

T=6

T=12

Co

ntro

l0.

018

(0.0

15)

0.06

80.

084

0.11

10.

710

0.14

50.

585

0.23

100

0.79

8 (0

.605

)4.

344.

957.

101.

85 (2

.00)

1.11

(0.9

5)7.

46 (6

.34)

9.39

1.81

1.15

9.54

13.8

200

1.57

(1.3

0)3.

6610

.613

.12.

58 (2

.73)

2.01

(1.7

2)12

.60

(12.

8)20

.49.

719.

0731

.535

.6

400

2.74

(2.4

3)4.

3714

.121

.24.

56 (4

.28)

22.0

3 (1

9.3)

27.6

(27.

9)41

.93.

379.

4727

.131

.749

.275

.112

914

1

800

3.95

(3.3

6)16

.524

.526

.35.

42 (5

.32)

44.5

(37.

6)54

.6 (5

3.1)

67.1

4.10

42.5

46.3

66.5

223

179

266

200

1600

4.97

(4.1

3)20

.526

.636

.69.

57 (9

.87)

40.9

(33.

4)53

.1 (5

3.4)

63.4

8.82

33.7

38.5

45.0

612

830

*84

0

3200

6.02

(5.0

5)21

.523

.429

.011

.4 (1

1.6)

16.2

(12.

9)41

.1 (4

1.4)

51.4

14.1

13.9

17.3

21.8

2520

2700

2196

6400

6.10

(5.0

7)22

.520

.923

.012

.6 (1

2.3)

9.12

(6.4

6)27

.6 (3

2.1)

34.4

16.9

7.76

25.4

31.5

* E

xtre

mel

y hi

gh

Zn c

onc

entr

atio

n w

as m

easu

red

in t

his

sam

ple

and

the

refo

re c

ons

ider

ed in

valid

79

4


Tab

le S

I-4.

Vis

ual M

INTE

Q c

alcu

latio

ns o

f Zn

spec

iatio

n in

po

re w

ater

fro

m L

ufa

2.2

soil

spik

ed w

ith Z

nCl 2

or

ZnO

Ass

ump

tions

: DO

C =

315

mg

/l; t

=20

°C

; Nic

a-D

onn

an D

OC

Zn

form

Tim

e p

oin

tTo

tal s

oil

conc

. pH

Zn

in m

g/l

Zn

spec

iati

on

in m

ol/

lZ

n sp

ecia

tio

n in

%

Co

mp

one

ntD

isso

lved

in

org

anic

Bo

und

to

DO

CTo

tal

dis

solv

ed%

dis

solv

edb

oun

d t

o D

OC

free

or

com

ple

xed

free

as

Zn2+

ZnC

l 21.

18

T=0

Cl-

0.00

669

0.00

0.00

669

100%

0%0%

100

mg

/kg

H+

-1.7

2E-0

70.

0028

20.

0028

2

pH

=7.

46H

FA-

0.00

259

HFA

2-0.

0008

2

Zn2+

3.83

E-1

60.

0000

20.

0000

2

ZnC

l 249

.2C

l-0.

0267

40.

0000

0.02

6798

.2%

1.78

%1.

69%

T=0

H+

4.06

E-0

70.

0009

10.

0009

1

400

mg

/kg

HFA

-0.

0025

9

pH

=6.

40H

FA2-

0.00

082

Zn2+

3.14

E-0

60.

0001

70.

0001

8

ZnC

l 226

6C

l-0.

0535

0.00

0.05

3527

.0%

73.0

%65

.3%

T= 6

mo

nths

H+

2.98

E-0

6-0

.000

93-0

.000

93

800

mg

/kg

HFA

-0.

0025

9

pH

=5.

49H

FA2-

0.00

082

Zn2+

2.97

E-0

30.

0011

0.00

407

HFA

= H

umic

/Ful

vic

Aci

d; D

OC

= D

isso

lved

Org

anic

Car

bo

n

80

Chapter 4

ZnC

l 227

00C

l-0.

2140

00.

0000

110.

2140

02.

56%

97.5

%80

.9%

T=6

mo

nths

H+

0.00

0035

5-0

.000

739

-0.0

0073

3200

mg

/kg

HFA

-0.

0025

9

pH

5.3

5H

FA2-

0.00

082

Zn2+

0.04

0247

0.00

1056

0.04

13

ZnO

-NP

9.57

H+

-5.7

6E-0

80.

0014

10.

0014

110

0%0%

0%

T=0

HFA

-0.

0025

9

1600

mg

/kg

HFA

2-0.

0008

2

pH

=7.

23Zn

2+4.

40E

-10

0.00

015

0.00

015

ZnO

-NP

53.1

H+

4.65

E-0

7-0

.000

13-0

.000

1310

0%2.

49%

0%

T= 6

mo

nths

HFA

-0.

0025

9

1600

mg

/kg

HFA

2-0.

0008

2

pH

=6.

32Zn

2+2.

02E

-05

0.00

081

0.00

081

ZnO

-NP

67.1

H+

2.95

E-0

6-0

.001

38-0

.001

3899

.9%

0.00

9%0.

009%

T=12

mo

nths

HFA

-0.

0029

5

1600

mg

/kg

HFA

2-0.

0008

2

pH

=5.

53Zn

2+1.

33E

-07

0.00

150

0.00

150

HFA

= H

umic

/Ful

vic

Aci

d; D

OC

= D

isso

lved

Org

anic

Car

bo

n

81

4


5The effect of pH on the toxicity of ZnO nanoparticles to Folsomia candida

in amended field soil

Abstract

The effect of soil pH on the toxicity of 30 nm ZnO to Folsomia candida was assessed in Dorset field soils with pHCaCl2 adjusted to 4.31, 5.71 and 6.39. To unravel the contribution of particle size and dissolved Zn, 200 nm ZnO and ZnCl2 were tested. Zn sorption increased with increasing pH and Freundlich Kf values ranged from 98.9 to 333 l/kg for 30 nm ZnO and from 64.3 to 187 l/kg for ZnCl2. No effect of particle size was found on sorption and little difference was found in toxicity between 30 and 200 nm ZnO. The effect on reproduction decreased with increasing pH for all Zn forms, with 28-d EC50s of 553, 1481 and 3233 mg Zn/kg for 30 nm ZnO and 331, 732 and 1174 mg Zn/kg for ZnCl2, at pH 4.31, 5.71 and 6.39, respectively. EC50s based on porewater Zn concentrations increased with increasing pH for 30 nm ZnO from 4.77 to 18.5 mg Zn/l, while for ZnCl2 no consistent pH-related trend in EC50s was found (21.0-63.3 mg Zn/l). Porewater calcium levels in ZnCl2-spiked soils were ten times higher than in ZnO-spiked soils. Our results suggest that the decreased toxicity of ZnCl2 compared to 30 nm ZnO based on porewater concentrations was due to a protective effect of calcium and not a particle effect.

Pauline L. Waalewijn-Kool, Maria Diez Ortiz, Stephen Lofts, Cornelis A.M. van Gestel

Environmental Toxicology and Chemistry, 2013, accepted

84

Chapter 5

1 Introduction

Nowadays manufactured nanoparticles are applied in a variety of consumer products, such as cosmetics, electronics, fuels and sport clothes. With the increasing production of nanomaterials, the release into the environment has become unavoidable. The potential hazards for organisms living there need to be assessed for sound risk assessment of this new class of substances (Handy et al., 2008).

It is likely that nanoparticles will enter the soil environment through land application of sewage sludge derived from wastewater treatment (Tourinho et al., 2012). After release into soils, the fate and toxicity of nanoparticles will be modified by aggregation, sorption and dissolution. These processes are extremely complex in soil and depend on both soil and nanoparticle characteristics. Nanoparticles, having at least one dimension smaller than 100 nm, tend to aggregate and interact with environmental media due to their charged surface and surface potential (Baalousha et al., 2008). Negatively charged components of the soil, such as humic and fulvic acids, and clay particles, have the propensity to bind (aggregated) nanoparticles. In this respect, the metals in nanoparticles become bioavailable after dissolution, which is likely to occur for metal-based nanoparticles.

Zinc oxide nanoparticles (ZnO-NP) are one of the most commonly used types of nanoparticles in sunscreens (Osmond and McCall, 2010). They are also used for soil remediation processes, e.g. they were applied on semiconductor films to detect chlorinated phenols by quenching of visible emission (Kamat et al., 2002). Until recently little was known on the toxicity and risk of ZnO-NP to soil organisms. Toxicity tests have been published for earthworms (Hooper et al., 2011; Canas et al., 2011), isopods (Pipan-Tkalec et al., 2010) and springtails (Chapters 2, 3 and 4). Nevertheless, data is still inadequate to comprehensively assess the toxicity of ZnO-NP and to understand its toxic mechanism. Du et al. (2011) found that the toxicity for wheat plants was due to Zn dissolution from ZnO-NP. And Priester et al. (2012) found high Zn accumulation in the leaves of soybeans after 50 days exposure to ZnO-NP in a soil at pH 6.78.

Soil properties such as pH, organic matter (OM) content and cation exchange capacity (CEC) are important factors that determine the bioavailability of metals to soil organisms. Soil pH is the most important soil property affecting metal partitioning between the solid phase and the pore water. Increased dissolution with decreasing pH has been found in water suspensions with ZnO-NP, e.g. Dimpka et al. (2011) found less release of Zn ions at pH 7 than at pH 6 (4.66 and 0.98% at a nominal concentration of 500 mg/l). The influence of pH on the bioavailability and toxicity of ZnO-NP in soil is unknown. For metals, the effect on reproduction of Folsomia candida increases with decreasing soil pH. Sandifer and Hopkin (1996) assessed the influence of pH on zinc (Zn(NO3)2) toxicity to F. candida in artificial soil and found 28-d EC50 values of 590, 600 and 900 mg Zn/kg d.w. at a pH of 4.5, 5 and 6, respectively. Crommentuijn et al. (1997) observed the same relationship with soil pH for the 35-d EC50 of F. candida exposed to cadmium.

85

5

The effect of pH on the toxicity of ZnO nanoparticles to Folsomia candida in amended field soil

The present study aimed at determining the influence of soil pH on the bioavailability

and toxicity of ZnO particles and ZnCl2 to the springtail F. candida in amended field

soils. The pH levels tested fall within ranges likely to be encountered in natural soils.

It was expected that a lower pH would increase the dissolved Zn concentrations in the

pore water and that this would lead to an increased toxicity.


2.1 Soil propertiesSoil was sampled in an open heath land site in Wareham forest (Ordnance Survey

Grid Reference: SU108058, Dorset, United Kingdom). The vegetation on site was

dominated by heather (Erica sp.) with small trees. During sampling, large roots were

removed, and soil was collected from 0-30 cm depth. The soil was homogenised,

sieved through a 5 mm mesh and air dried (initial moisture content was ~14% w/w).

Three levels of calcium carbonate, i.e. 0.2%, 0.45%, and 1% w/w were added to the

soil to obtain three soils with nominal pHCaCl2 4.5 (soil 1), 5.9 (soil 2) and 7.2 (soil 3),

respectively. Loamy sand soil (LUFA-Speyer 2.2, Germany) was used as control soil for

springtail performance in the toxicity tests.

The pHCaCl2 of the treated soils was measured at the beginning of the toxicity

test. Soils (5 ± 0.1 g) were shaken with 25 ml 0.01M CaCl2 solution for 2 hours. After

settlement of the particles, the pHCaCl2 value of the soil solution was recorded using a

Consort P907 meter. The Water Holding Capacity (WHC) was determined following ISO

guideline 11267 (ISO, 1999). The organic matter content was determined as loss on

ignition at 500 °C in an ashing oven. The CEC was determined by the Silver Thiourea

Method (Dohrmann, 2006). Approximately 2 g dry soil was shaken with 25 ml 0.01M

silver thiourea complex cation (AgTU) solution for 3 hours to achieve a complete

exchange of all cations. Four blanks were included without soil. Ag was measured in

the supernatant solution by flame Atomic Absorption Spectrometry (AAS) (Perkin Elmer

AAnalyst 100). The decrease in Ag concentration is a measure for the CEC of the soil.

2.2 Test compoundsTwo types of zinc oxide with different sizes were applied in the toxicity tests: 30 nm

ZnO-NP (Nanosun Zinc Oxide P99/30) and 200 nm non-nano ZnO (Microsun Zinc

Oxide W45/30). Transmission Electron Micrographs and Particle Size Distribution of

the ZnO particles are shown in Chapter 2. Primary particle size of both powders was in

agreement with the size reported by the manufacturer. The effect of dissolved Zn was

investigated by running tests with the soluble salt ZnCl2 (Merck, zinc chloride pure).

Seven concentrations for ZnO (nominal range 100-6400 mg Zn/kg d.w.) and five

concentrations for ZnCl2 (nominal range 100-1600 mg Zn/kg d.w.) were tested for all

three soils. Test concentrations were based on toxicity data found in earlier studies

86

Chapter 5

with ZnO-NP and F. candida (Chapters 2 and 3). ZnO powders were mixed with 200 g dry soil in glass jars to reach nominal test concentrations. First the ZnO powders were mixed in with 10% of the dry soil using a kitchen spoon and this soil was then mixed with the remaining 180 g dry soil. After mixing, water was added to reach 45% of the WHC and the soils were equilibrated for 1-2 days before starting the toxicity tests. ZnCl2 was added to the soil as a solution in deionized water (Milli-Q).

2.3 Soil and porewater analysisTwo samples per test concentration (± 100 mg dried soil) were taken from the spiked soils and digested in a mixture of deionized water (Milli-Q), concentrated HCl and concentrated HNO3 (1:1:4 by vol.) using an oven (CEM MDS 81-D). After digestion for 7 hours at 140 °C, solutions were analysed for total zinc concentrations by flame AAS (Perkin Elmer AAnalyst 100). Certified reference material (ISE sample 989 of River Clay from Wageningen, The Netherlands) was used to ensure the accuracy of the analytical procedure. Measured zinc concentrations in the reference material were within 10% of the certified concentrations.

Pore water was collected only once at the beginning of each toxicity test by centrifuging 30 g soil, after saturation with deionized water (Milli-Q) and one week equilibration. Soils were centrifuged for 50 min. (Centrifuge Falcon 6/300 series, CFC Free) with a relative force of 2000 g over two round filters (S&S 597 Ø 47 mm, pore size 11 μm) and a 0.45 μm membrane filter (S&S Ø 47 mm), placed inside the tubes (method cf. Hobbelen et al., 2004). Approximately 7 ml pore water per sample was collected for Zn and Ca analysis by flame AAS (Perkin Elmer AAnalyst 100). For soil 1, zinc concentrations in the pore water were also determined by flame AAS after ultrafiltration to obtain a particle-free extract. Soil solutions were centrifuged in a 3 kDa ultrafiltration device (Amicon Ultra-15 Filters, Millipore) for 20 min. at 2000 g. For practical reasons (clogging of filters) we were not able to apply ultrafiltration for soils 2 and 3. In this paper, the Zn levels measured after 0.45 µm filtration are defined as the dissolved Zn fractions, which in theory could include intact ZnO nanoparticles, free Zn2+ ions, partly dissolved ZnO and Zn bound to dissolved organic carbon fractions passing the 0.45 µm filter.

Total organic carbon (TOC) was measured in the pore water using combustion (Rosemount Analytical, Dohrmann DC-190). A reference of sucrose was used at 1000 mg/l. A volume of 25 µl was injected for each measurement and two replicates per sample were analyzed.

The Zn2+ ion concentrations in the pore water from the three soils spiked with 30 and 200 nm ZnO and ZnCl2 were calculated with the speciation model WHAM7 using Ca, Zn and TOC concentrations (mg/l) of the pore water and soil pH.

2.4 Toxicity testsThe springtail F. candida (Berlin strain; VU University Amsterdam) was cultured in pots with a base of moist plaster of Paris mixed with charcoal at 20 ± 1 °C at a light/dark regime of 12/12 h. The experiments were initiated with juveniles of the same

87

5


age (10-12 days) that were obtained by synchronising the egg laying of the culture

animals, fed with dried baker’s yeast (Dr. Oetker).

Each soil was tested in a separate toxicity test, including the three Zn forms,

control soil without Zn and Lufa 2.2 soil as a control for springtail performance. The

ISO guideline 11267 for testing for chemical effects on the reproduction of springtails

was followed (ISO, 1999). Tests were conducted in 100 ml glass jars containing 30 g

moist soil and five replicates for each treatment were prepared. At the start of the

test, ten synchronised animals were transferred into each test jar. The jars were filled

randomly and before introduction, the animals were checked under the microscope

for a healthy appearance. The animals were fed a few grains of dried baker’s yeast (Dr.

Oetker). The jars were incubated in a climate room at 20 ± 1 °C at a light/dark regime

of 12/12 h. Once a week, the moisture content of test soils was checked by weighing

the jars, and moisture was replenished with deionized water (Milli-Q) when necessary.

The jars were also aerated by this procedure.

After four weeks, the jars were sacrificed for determination of springtail survival

and reproduction. Each jar was emptied into a 200 ml beaker glass and 100 ml tap

water was added. The mixture was stirred carefully to let all the animals float to the

surface. The number of adults and juveniles were counted manually after taking a

picture of the water surface using a digital camera (Olympus, C-5060).

2.5 Data analysisUsing the soil and porewater concentrations, sorption of zinc to the test soil was

described by a Freundlich isotherm:

Cs = Kf * Cwn , where

Cs = concentration in the soil (mg Zn/kg d.w.)

Kf = Freundlich sorption constant (l/kg)

Cw = concentration in the pore water (mg Zn/l) and

n = shape parameter of the Freundlich isotherm

When n > 1, this may be indicative of agglomeration and/or aggregation due to

saturation effects at higher concentrations in the pore water. In case n < 1, binding

sites in the soil may become saturated with increasing zinc concentrations in the pore

water. As a consequence, values of Kf always have to be interpreted together with the

corresponding value of n. See also Chapter 3 on ZnO-NP, non-nano ZnO and ZnCl2

sorption in Lufa 2.2 natural soil.

The trimmed Spearman-Karber method was applied to estimate LC50 values

for the effect of ZnCl2 on springtail survival (Hamilton et al., 1977). EC50 values for

reproduction were estimated applying the logistic model of Haanstra et al. (1985) and

were determined based on total Zn, dissolved Zn and free Zn2+ ion concentrations. A

generalized likelihood ratio test (Sokal and Rohlf, 1995) was applied to compare EC50

values obtained for each Zn form and each test soil and also to compare EC50 values

88

Chapter 5

based on dissolved Zn concentrations before and after ultrafiltration. Calculations were performed in SPSS Statistics 20.

3 Results

3.1 Soil propertiesTable 1 summarizes the soil properties that were determined in the Dorset test soils and Lufa 2.2 soil (pH, WHC, OM and CEC). The WHC of Dorset soil was 76.6 g/100 g and the OM content approximately 7.5%. The CEC slightly increased from 8.19 to 10.8 cmolc/kg soil with increasing soil pH.

At the beginning of the toxicity test the measured pHCaCl2 of the test soils was 4.31, 5.71 and 6.39 for soil 1, soil 2 and soil 3, respectively. After treatment with 30 nm ZnO the pH increased with increasing exposure concentrations up to 5.76 and 5.90 for soils 1 and 2, respectively. The pH of soil 3 treated with 30 nm ZnO slightly decreased to 6.26. The same pattern was observed in the three test soils treated with 200 nm ZnO. For soils spiked with ZnCl2, the pH slightly decreased with increasing exposure concentration in all three soils. At the highest test concentrations a pHCaCl2 of 3.96, 5.13 and 6.16 was measured for soils 1, 2 and 3, respectively. In the Supporting Information (SI), Table SI-1 shows all pH measurements in the three test soils treated with the three Zn forms.

3.2 Total soil analysisThe results of the zinc measurements in the three soils treated with 30 nm ZnO, 200 nm ZnO and ZnCl2 are presented in Table SI-2. Background (untreated) total Zn concentration was subtracted from those measured in each of the treatments. Recovery for all soil samples was above 80%. The coefficient of variance between replicates was always below 0.18. All results are expressed on the basis of measured Zn concentrations.

Table 1. Properties of Dorset test soils and Lufa 2.2 reference soil used to determine the influence of soil pH on the toxicity of ZnO nanoparticles to Folsomia candida

Soil pHCaCl2

(nominal)pHCaCl2

(measured)CaCO3

b

(% w/w)WHC

(g/100g)Organic

matter (%)CEC

(cmolc/kg soil)

1 4.5 4.31 0.2 76.6 7.39 ± 0.003 8.19 ± 0.74 (n=2)

2 5.9 5.71 0.45 76.6 7.63 ± 0.136 9.09 ± 0.05 (n=2)

3 7.2 6.39 1 76.6 7.65 ± 0.266 10.8 ± 0.73 (n=2)

Unadjusteda 3.0 3.0 - 76.6 8.41 ± 0.213 7.83 ± 0.16 (n=2)

Lufa 2.2 5.5 5.3 - 45.2 4.35 ± 0.090 8.24 ± 0.34 (n=2)

a Dorset soil not adjusted with CaCO3b added CaCO3

WHC = Water Holding CapacityCEC = Cation Exchange Capacity

89

5


3.3 SorptionThe results of the zinc measurements in the pore water from the three soils treated

with 30 nm ZnO, 200 nm ZnO and ZnCl2 are presented in Table 2. In soil 1, porewater

concentrations ranged from 1.08 to 24.1 and from 0.98 to 26.5 mg Zn/l for 30 nm and

200 nm ZnO, respectively. After ultrafiltration, porewater concentrations in soil 1 were

only lower at the lower and intermediate ZnO concentrations but not at the higher levels

or in the ZnCl2 spiked soil. The porewater concentrations in soil 2 and 3 spiked with

ZnO were slightly lower than in soil 1 at the lower and intermediate concentrations and

slightly higher at the higher treatment levels. Up to 41 mg Zn/l was measured for soil

2 spiked with 30 nm ZnO at 6400 mg Zn/kg d.w. At the highest test concentration for

200 nm ZnO 36.4 mg Zn/l was measured in soil 3. For ZnCl2-spiked soils the porewater

concentrations increased with exposure concentrations but decreased with increasing

pH. For example, dissolved Zn concentrations in soils spiked with 800 mg Zn/kg d.w.

were 192, 76.0 and 23.1 mg Zn/l in soils 1, 2 and 3, respectively.

For all Zn forms, zinc concentrations in the pore water increased with increasing total

soil concentrations. This is reflected by the Freundlich isotherm for which the R2 was above

0.87 for all test soils and Zn forms, indicating a good fit to the measured Zn concentrations.

Table 3 shows the Freundlich sorption constants Kf and shape parameter n for the three

Zn forms in the three test soils. Similar sorption constants were estimated for both ZnO

particles and sorption increased with increasing soil pH. The Kf values ranged from 98.9

to 333 l/kg and from 81.0 to 318 l/kg for 30 and 200 nm ZnO, respectively. Both ZnO

particles showed n parameters higher than 1 for soil 1 and values slightly below 1 for soils

2 and 3. This suggests saturation equilibrium in the soil at high pH and saturation effects

in the pore water at low pH. Sorption constants were higher for ZnO than for ZnCl2 in all

three test soils. The Kf values for ZnCl2 were 73.6, 64.3 and 187 l/kg (with corresponding n

values of 0.461, 0.581 and 0.475) for soils 1, 2 and 3, respectively.

3.4 Calcium and TOC measurements Ca concentrations in the pore water increased with increasing exposure concentrations

for ZnCl2-spiked soil and ranged from 48.9 to 180 mg/l in soil 1, from 51.7 to 484 mg/l

in soil 2 and from 33.8 to 434 mg/l in soil 3. For 30 nm ZnO, Ca concentrations

slightly decreased with increasing exposure concentrations from 16.4 to 11.4 mg/l

in soil 1, from 20.3 to 14.6 mg/l in soil 2 and from 48.1 to 10.9 mg/l in soil 3. Ca

concentrations in the pore water collected from soils treated with 200 nm ZnO showed

similar patterns (Table SI-3).

For ZnO, no clear pattern was observed between TOC level and total zinc

concentration but TOC levels did increase with increasing pH. For 30 nm ZnO, average

TOC concentrations of 452, 485 and 506 mg/l were measured in soils 1, 2 and 3,

respectively and for 200 nm ZnO these values were 431, 453 and 533 mg/l. TOC

concentrations for ZnCl2-spiked soil ranged from 259 to 319 mg/l in soil 1, from 282 to

357 mg/l in soil 2 and from 484 to 534 mg/l in soil 3 (Table SI-4).

90

Chapter 5

Tab

le 2

. Zi

nc c

onc

entr

atio

ns m

easu

red

in

the

po

re w

ater

(m

g Z

n/l)

of

the

thre

e te

st s

oils

sp

iked

with

30

nm Z

nO,

200

nm Z

nO a

nd Z

nCl 2.

Po

re

wat

er (

n=1)

was

co

llect

ed o

ne w

eek

afte

r sa

tura

tion

of

the

soils

with

dei

oni

zed

wat

er (

Mill

i-Q).

Zinc

co

ncen

trat

ions

mea

sure

d in

the

po

re w

ater

aft

er

ultr

afiltr

atio

n ar

e p

rese

nted

in b

etw

een

bra

cket

s

Soil

30 n

m Z

nO20

0 nm

ZnO

ZnC

l 2

(mg

Zn/

kg)

soil

1so

il 2

soil

3so

il 1

soil

2so

il 3

soil

1so

il 2

soil

3

Lufa

2.2

0.12

(0.1

0)0.

040.

06

Co

ntro

l0.

16 (0

.10)

0.02

0.02

100

1.08

(0.5

3)0.

390.

240.

98 (0

.58)

0.40

0.33

3.18

(3.4

0)2.

861.

36

200

1.78

(0.8

4)0.

840.

562.

02 (1

.02)

0.68

0.57

10.9

(10.

1)9.

091.

04

400

4.13

(1.9

6)1.

901.

174.

25 (2

.06)

1.76

1.68

43.0

(48.

9)41

.54.

58

800

6.18

(3.9

5)4.

314.

146.

58 (3

.31)

3.72

4.22

192

(*)

76.0

23.1

1600

9.98

(8.4

2)8.

7210

.711

.1 (8

.77)

10.3

14.2

231

136

3200

17.9

(19.

0)26

.322

.518

.0 (1

5.7)

20.8

28.2

6400

24.1

(30.

0)41

.028

.926

.5 (2

4.9)

34.3

36.4

* ul

trafi

ltrat

ion

faile

d fo

r th

is s

amp

le

91

5


3.5 Zn2+ ion concentrations Table SI-5 shows all calculated Zn2+ ion concentrations in the pore water from the three soils treated with 30 and 200 nm ZnO and ZnCl2 using WHAM7. Zn2+ ion concentrations for ZnCl2-spiked soils increased with increasing exposure concentrations and decreased with increasing soil pH. For example, Zn2+ ion concentrations in pore water from soils spiked with 800 mg Zn/kg d.w. were 2940, 803 and 77.1 µM in soil 1 (pHCaCl2=3.50), soil 2 (pH CaCl2=5.4) and soil 3 (pHCaCl2=6.20), respectively. The Zn2+ ion concentrations in the pore water from ZnO-spiked soils are roughly a factor 10-100 lower than in ZnCl2-spiked soils. For 30 nm ZnO, a maximum of 49.3 µM was calculated in soil 1 spiked with 800 mg Zn/kg d.w. (pHCaCl2=4.8). At higher exposure concentrations Zn2+ ion concentrations decreased to 6.5 µM with increasing pH up to 5.65. In the other two soils with higher pH levels, such a maximum in Zn2+ ion concentrations was not observed and the concentrations remained below 14.1 and 1.53 µM in soils 2 and 3, respectively. Soils spiked with 200 nm ZnO show a similar trend of Zn2+ ion concentrations with a maximum of 44.0 µM in pore water collected from soils spiked with 400 mg Zn/kg d.w. (pHCalCl2=4.65). At higher exposure concentrations the Zn2+ ion concentrations decreased when pH levels were above 5. In soils 2 and 3 the Zn2+ ion concentrations increased with increasing exposure concentrations and were 8.02 and 11.9 µM for the highest exposure concentration (6400 mg Zn/kg d.w.) in soils 2 and 3, respectively.

3.6 Toxicity dataControl performance of the collembolans was affected by soil pH. Control survival decreased with increasing pH, and was 88, 74 and 54% in soils 1, 2 and 3, respectively. The average number of juveniles in the controls also decreased with increasing pH and was 338, 142 and 37 for soils 1, 2 and 3, respectively, with coefficients of variance of 23.7, 16.8 and 52.0%.

Survival of F. candida in soils spiked at concentrations up to 6400 mg Zn/kg d.w. with 30 and 200 nm ZnO was not affected and comparable to that in the controls. Survival was affected by ZnCl2 and mortality decreased with increasing soil pH shown

Table 3. Freundlich sorption constants Kf ± standard error (s.e.) (l/kg) and shape parameter n ± s.e. for Zn partitioning in three soils spiked with 30 nm ZnO, 200 nm ZnO and ZnCl2

30 nm ZnO 200 nm ZnO ZnCl2

soil 1 Kf = 98.9 ± 1.13n = 1.23 ± 0.059

Kf = 81.0 ± 1.19n =1.27 ± 0.082

Kf = 73.6 ± 1.23n = 0.461 ± 0.060

soil 2 Kf = 269 ± 1.08n = 0.842 ± 0.038

Kf = 272 ± 1.11n = 0.886 ± 0.050

Kf = 64.3 ± 1.15n = 0.581 ± 0.038

soil 3 Kf = 333 ± 1.14n = 0.794 ± 0.060

Kf = 318 ± 1.13n = 0.757 ± 0.055

Kf = 187 ± 1.32n = 0.475 ± 0.104

92

Chapter 5

Tab

le 4

. EC

50 v

alue

s (w

ith 9

5% c

onfi

den

ce in

terv

als)

for t

he e

ffect

on

rep

rod

uctio

n o

f Fo

lso

mia

can

did

a af

ter 2

8-d

exp

osu

re to

30

and

200

nm

ZnO

and

Zn

Cl 2

in s

oils

with

thre

e d

iffer

ent p

H le

vels

. EC

50 v

alue

s ar

e p

rese

nted

as

tota

l Zn

conc

entr

atio

ns in

mg

Zn/

kg d

.w. (

left

), as

po

rew

ater

Zn

conc

entr

atio

ns

in m

g Z

n/l (

mid

dle

) and

as

Zn2+

ion

conc

entr

atio

ns (c

alcu

late

d u

sing

WH

AM

7) in

µM

(rig

ht)

30 n

m Z

nO20

0 nm

ZnO

ZnC

l 230

nm

ZnO

200

nm Z

nOZ

nCl 2

30 n

m Z

nO20

0 nm

ZnO

ZnC

l 2

Tota

l so

il (m

g Z

n/kg

d.w

.)P

ore

wat

er (m

g Z

n/l)

Free

Zn2+

(µM

)

soil

1(p

H 4

.5)

5531a

b

(209

-896

)88

91ab

(492

-128

6)33

11b

(264

-399

)4.

771a

(2.3

8-7.

16)

8.10

1a

(5.2

3-11

.0)

21.0

1b

(6.5

2-35

.5)

**

3201

(98.

8-10

00)

soil

2(p

H 5

.9)

1481

2a

(120

4-17

58)

1034

1a

(826

-120

6)73

212b

(543

-920

)8.

041a

(6.4

9-9.

59)

4.42

1b

(3.2

1-5.

62)

63.3

2c

(44.

9-81

.8)

3.72

1a

(-)

0.63

61b

(0.4

29-0

.844

)63

81c

(403

-872

)

soil

3(p

H 7

.2)

3233

12a

(92.

3-63

73)

1559

1a

(607

-251

0)11

742a

(546

-180

2)18

.51a

(0.6

70-3

6.3)

11.7

1a

(3.1

0-20

.3)

40.8

12a

(-)

1.51

2a

(0.2

24-2

.80)

4.79

2a

(1.1

1-8.

48)

1831b

(0-7

21)

1,2 in

dic

ate

sig

nific

ant

diff

eren

ces

bet

wee

n E

C50

val

ues

for

the

diff

eren

t so

ils a

cco

rdin

g t

o a

gen

eral

ized

like

liho

od

-rat

io t

est

(χ2 (1

) > 3

.84;

p <

0.0

5)a,

b in

dic

ate

sig

nific

ant

diff

eren

ces

bet

wee

n E

C50

val

ues

for

the

diff

eren

t Zn

form

s ac

cord

ing

to

a g

ener

aliz

ed li

kelih

oo

d-r

atio

tes

t (χ

2 (1) >

3.8

4; p

< 0

.05)

- D

ata

did

no

t al

low

cal

cula

ting

rel

iab

le 9

5% c

onfi

den

ce in

terv

als

* D

ata

did

no

t al

low

est

imat

ing

an

EC

50 v

alue

(see

tex

t)

93

5


by LC50 values of 643 and 815 mg Zn/kg d.w. in soils 1 and 2, respectively. Survival was no longer affected by ZnCl2 in soil 3 up to 1600 mg Zn/kg d.w.

Reproduction was reduced in a dose-dependent manner for the three Zn forms in all three test soils (See Figures SI-1-3 for all dose-response curves). The toxicity of 30 and 200 nm ZnO and ZnCl2 decreased with increasing soil pH. This was significant for the EC50 values obtained for 30 nm ZnO, which increased from 553 to 1481 mg Zn/kg d.w. in soils 1 and 2 (Table 4). For ZnCl2 the EC50 (331 mg Zn/kg d.w.) estimated in soil 1 was significantly lower than the EC50 (1174 mg Zn/kg d.w.) estimated in soil 3. Based on porewater concentrations the EC50 values also increased with increasing pH for 30 nm ZnO (from 4.77 to 18.5 mg Zn/l) (Table 4). Also for ZnCl2 the lowest EC50 was estimated in soil 1 (21.0 mg Zn/l) and higher EC50 values were estimated in soils 2 (63.3 mg Zn/l) and 3 (40.8 mg Zn/l). For 200 nm ZnO no significantly different EC50 values based on porewater concentrations were estimated for the three soils. EC50 values based on porewater Zn concentrations measured after ultrafiltration decreased from 4.77 to 2.44 mg Zn/l for 30 nm ZnO and from 8.10 to 4.52 mg Zn/l for 200 nm ZnO. According to a generalized likelihood-ratio test, these EC50 values were neither significantly lower for 30 nm ZnO (χ2

(1) = 1.71, p < 0.05), nor for 200 nm ZnO (χ2(1) = 2.54, p < 0.05).

Based on total Zn concentration EC50 values for ZnCl2 were lower than for ZnO in all soils, which was significant in soils 1 and 2 according to a generalized likelihood-ratio test. But EC50 values for ZnCl2 based on porewater concentrations were significantly higher than for ZnO in all soils. In most cases the EC50 values for 30 and 200 nm ZnO were similar in all soils, except for the one based on porewater concentrations in soil 2 which was lower for 200 nm ZnO.

The EC50 for 30 nm ZnO based on free Zn2+ ion concentrations was slightly higher in soil 2 (3.54 µM) than in soil 3 (1.52 µM). No EC50 could be estimated for 30 and 200 nm ZnO in soil 1, because the decrease in juvenile numbers was not related with increasing Zn2+ ion concentrations. For soils 2 and 3 the EC50 values were 0.635 and 4.79 µM for 200 nm ZnO. EC50 values for ZnCl2 were much higher than for ZnO and were 320, 638 and 183 µM for soils 1, 2 and 3, respectively (Table 4).

4 Discussion

The effect of soil properties must be investigated to understand speciation, bioavailability and toxicity of soluble metal-oxide nanoparticles, such as ZnO-NP. The environmental fate of nanoparticles in soils (aggregation, dissolution and sorption) is dependent on soil properties, such as soil pH and OM content. Natural soil from Dorset (UK) was adjusted for pH using CaCO3 to study the influence of soil pH on the toxicity of 30 nm ZnO to springtails. We used one test soil differing in pH, instead of natural soils differing in pH that may differ in other properties, such as granular composition, OM content and properties and ionic composition of the pore water.

94

Chapter 5

Unfortunately, the addition of the different Zn concentrations and forms changed the nominal pH levels aimed for. In addition, pH adjustment and the different Zn forms also affected Ca levels in the pore water. These factors increase the complexity of interpreting data from this study.

Metal sorption and dissolution are two processes that are strongly interconnected and both highly dependent on soil pH. Dissolved metal concentrations in the pore water were measured to compare the Zn release into the pore water from ZnO and ZnCl2-spiked soils. In general, sorption increases with increasing pH and dissolved metal concentrations increase with decreasing pH due to competition of H+ ions for metal binding on negatively charged soil particles. In the Dorset soils increased sorption with increasing pH was observed for all three Zn forms with similar Kf values for 30 and 200 nm ZnO. The Kf values obtained for both ZnO particles however were higher than for ZnCl2, indicating higher sorption of ZnO than the soluble metal salt ZnCl2. Different sorption constants were expected from the different physiochemical properties of ZnO and ZnCl2 (composition, solubility, size, surface charge). For ZnO much lower dissolved Zn levels were measured than for ZnCl2. The higher dissolved Zn concentrations in the ZnCl2 spiked soils were the result of higher water solubility of this metal salt. ZnO tends to aggregate/agglomerate and interact with the solid phase, and as a consequence releasing much lower Zn levels to the pore water. Interestingly, there was no difference in the Zn concentrations measured in the pore water from nano and non-nano ZnO spiked soils, which has been reported before in Chapters 2, 3 and 4 of this thesis. This suggests that the release of Zn from ZnO is not affected by particle size.

The present study found a clear linear correlation between dissolved Zn concentrations and pH for ZnCl2-spiked soils. Zn release from ZnO increased with decreasing pH for soils spiked with 100-1600 mg Zn/kg d.w., while for the two highest concentrations the highest dissolved Zn concentrations were measured in soil 2 (pHCaCl2 5.7). More acidic conditions would favor the dissolution for ZnO-NP (Miao et al., 2010). ZnO dissolution is expected below pH 7.5 in aqueous solutions with samples containing up to 1 g/l ZnO-NP (Bian et al., 2011). Our results from ZnO-spiked soils showed an increase in pH with increasing total Zn concentration, which could have hampered Zn release into the pore water from soils spiked with 3200 and 6400 mg Zn/kg d.w. At higher pH levels (between 6 and 9), solid Zn(OH)2 will precipitate from solution (Yamabi and Imai, 2002) and dissolved Zn2+ will form complexes with phosphate and carbonate (Ma et al., 2013). This may explain the difference in the Freundlich shape parameter n for 30 nm ZnO (n = 1.23) and ZnCl2 (n = 0.461) in soil 1. This difference suggests that at low pH the binding sites in soil became saturated at high Zn concentrations in the pore water of ZnCl2-spiked soils while for 30 nm ZnO the pore water reached saturation. In fact, we believe that other processes, such as agglomeration/aggregation and increased soil pH in ZnO spiked soil prevent Zn release into the pore water. The use of advanced microscopy and/or spectroscopy and size-fractionation methods would be needed to enable further unravelling sorption and aggregation behavior of ZnO-NP

95

5


in soil (Lead and Wilkinson, 2006). Zn concentrations in soil and pore water were measured on a short-time scale, namely directly after spiking. Ageing or equilibration of contaminated soil might provide more realistic insights into ZnO-NP dissolution rates and and increased sorption of ZnO-NP or of the released Zn.

An effect of soil pH on survival and reproduction of F. candida was found in the three control soils, which indicates that soil pH may act as a stress factor for F. candida. It is known that F. candida has a mild preference for soils with a pH between 5 and 6, at which they achieve their highest reproduction (Fountain and Hopkin, 2005; Greenslade and Vaughan, 2003). In preference tests with sand and pH ranging from 2 to 9, F. candida showed a weak and variable response to pH, suggesting F. candida is a weakly sub-neutral species (van Straalen and Verhoef, 1997). In our Dorset soils with three pH levels, reproduction was highest at pHCaCl2 4.31. Also, in soil at pHCaCl2 5.71 the number of juveniles was above 100 per replicate test jar, which is the minimum number needed to meet the validity criteria (ISO, 1999). Although the number of juveniles was below 100 per replicate test jar for soil at pHCaCl2 6.39, we still observed consistent dose-response relationships. Based on this and the normal reproduction at lower pH we conclude that our test animals were healthy and that the tests were valid also at pHCaCl2 6.39. A significant effect of soil pH on the reproduction of F. candida in the Dorset soils was found for ZnO and ZnCl2. Based on total soil concentrations the EC50 values increased with increasing pH for all Zn forms. This is in line with a study with earthworms (Heggelund et al., 2013) using the same soils, in which also higher toxicity of all Zn forms was found at lower soil pH. The finding that ZnCl2 was more toxic than ZnO is in agreement with earlier studies with earthworms (Hooper et al., 2011) and springtails (Chapter 3). No significant effect of ZnO particle size on springtail toxicity was found in all my experiments described so far in this thesis.

In order to make future predictions about the potential toxicity of ZnO-NP, it is important to understand the mechanisms of toxicity. Particle dissolution to ionic zinc, particle-induced generation of reactive oxygen species (ROS) and photo-induced toxicity represent primary modes of action for ZnO-NP toxicity (Ma et al., 2013). This study focused on the contribution of particle dissolution and potential particle effects in soil by studying the comparative bioavailability and toxicity of 30 nm ZnO and the metal salt ZnCl2. Based on porewater Zn concentrations the EC50 values for ZnO were lower than for ZnCl2. And based on free Zn2+ ion concentration the difference in EC50 values for ZnO and ZnCl2 was even larger. This may suggest that not only dissolved and free Zn2+ ion concentrations determined toxicity but that also a negative effect of the particles may have been present. To indicate if toxicity is enhanced by the nanoparticles, ultrafiltration of the pore water was performed which has been shown a suitable technique (Stone et al., 2010). For soil 1 the dissolved Zn concentrations in the pore water after ultrafiltration were somewhat lower for soils spiked with 30 and 200 nm ZnO at 100-1600 mg Zn/kg d.w. This suggests that the pore water may have contained nano-particulate Zn. The EC50 values based on total Zn concentrations are within this

96

Chapter 5

range (Tables 2 and 4). EC50 values based on dissolved Zn concentrations measured after ultrafiltration were slightly but not significantly lower. As a consequence, it remains uncertain to what extent nano-particulate Zn in the pore water contributed to toxicity.

The difference in EC50 values for ZnO and ZnCl2 based on porewater Zn concentrations can be explained by the Ca concentrations added to adjust the pH of the soils (0.2-1% CaCO3). Ca concentrations measured in the pore water increased with increasing Zn concentrations up to 484 mg/l for ZnCl2, but were much lower for soil spiked with ZnO. The Zn2+ concentrations were roughly a factor 10-100 lower in ZnO-spiked soils than in ZnCl2-spiked soils (Table SI-3). Considering the large difference in Ca concentrations in the pore water between these two Zn forms, and the fact that Ca may compete with Zn for binding and uptake sites on the organism (see e.g. Muyssen et al., 2006), it is likely that the high Ca level in the pore water has reduced Zn toxicity for ZnCl2-spiked soils. A protective effect of Ca in ZnCl2-spiked soils may explain the decreased bioavailability of Zn in the pore water. In addition, the decreased soil pH may have reduced Zn bioavailability in the pore water of ZnCl2-spiked soils due to competition with H+ ions. According to the terrestrial Biotic Ligand Model (Thakali et al., 2006) bioavailability and toxicity may increase with increasing pH due to H+ ions competing with Zn2+ ions for the same binding sites on the organism. As pH was higher for ZnO than for ZnCl2-spiked soils the reduced H+ activity may have resulted in higher Zn bioavailability for ZnO-spiked soils.

5 Conclusion

In the present study we showed that soil pH influences the toxicity of ZnO-NP to F. candida and that ZnO-NP and ZnCl2 were more toxic in acidic soil than in basic soil. In comparison with ZnCl2, ZnO-NP toxicity was lower based on total Zn concentrations and higher based on porewater Zn concentrations. Toxicity based on porewater Zn concentrations may however, also been affected by differences in Ca concentrations. These results indicate that toxicity of ZnO-NP can not be solely attributed to the dissolved Zn fraction. Our study suggests that ZnO-NP toxicity in soil is a Zn speciation problem rather than a particle problem, with soil pH having an important role in ZnO-NP dissolution, bioavailability and toxicity. Further research on the role of other soil properties, such as organic matter content, is also needed.

97

5



• All dose response curves based on total Zn concentrations (Figure SI-1), dissolved Zn concentrations (Figure SI-2) and free Zn2+ concentrations (Figure SI-3)

• pHCaCl2 measurements (Table SI-1) • Total zinc concentrations measured in soil (Table SI-2) • Calcium concentrations measured in the pore water (Table SI-3) • TOC concentrations measured in the pore water (Table SI-4) • Free Zn2+ concentrations in the pore water estimated with WHAM7 (Table SI-5)

98

Chapter 5

Fig

ure

SI-1

. Effe

ct o

f 30

nm Z

nO (t

op

), 20

0 nm

ZnO

(mid

dle

) and

ZnC

l 2 (b

elo

w) o

n th

e re

pro

duc

tion

(num

ber

of j

uven

iles)

of F

ols

om

ia c

and

ida

afte

r 28-

d

exp

osu

re in

Do

rset

so

il 1

(pH

CaC

l2 =

4.3

1), s

oil

2 (p

HC

aCl2

= 5

.71)

and

so

il 3

(pH

CaC

l2 =

6.3

9) b

ased

on

mea

sure

d c

onc

entr

atio

ns o

f zin

c in

the

so

il (m

g Z

n/kg

d

.w.).

Lin

e sh

ow

s fit

ob

tain

ed w

ith a

log

istic

mo

del

.

04080120

160

200

050

010

0015

0020

00

020406080100

120

020

0040

0060

0080

00

050100

150

200

020

0040

0060

0080

00

050100

150

200

050

010

0015

00

050100

150

200

020

0040

0060

0080

00

050100

150

200

250

020

0040

0060

0080

00

0

100

200

300

400

500

600

020

040

060

080

010

00

0

100

200

300

400

500

020

0040

0060

0080

00

0

100

200

300

400

500

020

0040

0060

0080

00numberof juvenilesper test jar

30 n

mZn

O

200

nmZn

O

ZnC

l 2

30 n

mZn

O30

nm

ZnO

200

nmZn

O20

0 nm

ZnO

ZnC

l 2Zn

Cl 2

[Zn]

soil(

mg

Zn/k

g d.

w.)

SOIL

1[Z

n] so

il(m

g Zn

/kg

d.w.

)SO

IL 2

[Zn]

soil(

mg

Zn/k

g d.

w.)

SOIL

3

99

5


Fig

ure

SI-2

. Effe

ct o

f 30

nm Z

nO (t

op

), 20

0 nm

ZnO

(mid

dle

) and

ZnC

l 2 (b

elo

w) o

n th

e re

pro

duc

tion

(num

ber

of j

uven

iles)

of F

ols

om

ia c

and

ida

afte

r 28-

d

exp

osu

re in

Do

rset

so

il 1

(pH

CaC

l2 =

4.3

1), s

oil

2 (p

HC

aCl2

= 5

.71)

and

so

il 3

(pH

CaC

l2 =

6.3

9) b

ased

on

solu

ble

Zn

conc

entr

atio

ns in

the

po

re w

ater

(mg

Zn/

l).

Line

sho

ws

fit o

bta

ined

with

a lo

gis

tic m

od

el.

04080120

160

200

050

100

150

020406080100

120

010

2030

40

050100

150

200

010

2030

40050100

150

200

250

010

2030

4050

04080120

160

200

010

2030

40

050100

150

200

050

100

150

200

250

0

100

200

300

400

500

600

050

100

150

200

250

0

100

200

300

400

500

010

2030

0

100

200

300

400

500

010

2030

numberof juvenilesper test jar

30 n

mZn

O

200

nmZn

O

ZnC

l 2

30 n

mZn

O30

nm

ZnO

200

nmZn

O20

0 nm

ZnO

ZnC

l 2Zn

Cl 2

[Zn]

por

ew

ater

(mg

Zn/l)

SOIL

1[Z

n] p

ore

wat

er (m

g Zn

/l)SO

IL 2

[Zn]

por

ew

ater

(mg

Zn/l)

SOIL

3

100

Chapter 5

Fig

ure

SI-3

. Effe

ct o

f 30

nm Z

nO (t

op

), 20

0 nm

ZnO

(mid

dle

) and

ZnC

l 2 (b

elo

w) o

n th

e re

pro

duc

tion

(num

ber

of j

uven

iles)

of F

ols

om

ia c

and

ida

afte

r 28-

d

exp

osu

re in

Do

rset

so

il 1

(pH

CaC

l2 =

4.3

1), s

oil

2 (p

HC

aCl2

= 5

.71)

and

so

il 3

(pH

CaC

l2 =

6.3

9) b

ased

on

free

Zn2+

co

ncen

trat

ions

in t

he p

ore

wat

er c

alcu

late

d

with

WH

AM

7 (µ

M).

Line

sho

ws

fit o

bta

ined

with

a lo

gis

tic m

od

el.

04080120

160

200

1.E-

091.

E-07

1.E-

051.

E-03

1.E-

01020406080100

120

1.E-

091.

E-07

1.E-

051.

E-03

050100

150

200

1.E-

091.

E-07

1.E-

051.

E-03

050100

150

200

1.E-

091.

E-07

1.E-

051.

E-03

1.E-

01050100

150

200

1.E-

091.

E-07

1.E-

051.

E-03

050100

150

200

250

1.E-

091.

E-07

1.E-

051.

E-03

0

100

200

300

400

500

600

1.E-

061.

E-04

1.E-

021.

E+00

0

100

200

300

400

500

1.E-

061.

E-05

1.E-

041.

E-03

0

100

200

300

400

500

1.E-

061.

E-05

1.E-

041.

E-03

numberof juvenilesper test jar

30 n

mZn

O

200

nmZn

O

ZnC

l 2

30 n

mZn

O30

nm

ZnO

200

nmZn

O20

0 nm

ZnO

ZnC

l 2Zn

Cl 2

free

[Zn2+

] (µM

)SO

IL 1

free

[Zn2+

] (µM

)SO

IL 2

free

[Zn2+

] (µM

)SO

IL 3

101

5


Table SI-1. pHCaCl2 of the three test soils spiked with 30 nm ZnO, 200 nm ZnO and ZnCl2.

Nominal conc.(mg Zn/kg d.w.)


soil 1 soil 2 soil 3 soil 1 soil 2 soil 3 soil 1 soil 2 soil 3

Control 4.31 5.71 6.39

100 4.24 5.42 6.66 4.06 5.48 6.34 4.31 5.58 6.81

200 4.28 5.45 6.79 4.27 5.54 6.30 4.01 5.51 6.83

400 4.43 5.44 6.70 4.37 5.63 6.20 4.03 5.41 6.61

800 4.63 5.49 6.55 4.44 5.75 6.06 3.96 5.28 6.41

1600 5.05 5.88 6.33 4.88 5.82 6.12 5.13 6.16

3200 5.48 5.90 6.27 5.53 5.88 6.13

6400 5.76 5.90 6.26 5.67 5.99 6.09

Table SI-2a. Average zinc concentrations (± SD, n=2) measured in soil 1 (pHCaCl2 = 4.31) spiked with 30 nm ZnO, 200 nm ZnO and ZnCl2. Zinc concentrations in the soil are corrected for the zinc levels measured in the controls. Recoveries (%) are presented in brackets.

Nominal conc.(mg Zn/kg d.w.) 30 nm ZnO 200 nm ZnO ZnCl2

Control 10.3 ± 1.5 10.3 ± 1.5 10.3 ± 1.5

100 119 ± 7.2 (119) 93.5 ± 12.0 (94) 104 ± 4.3 (104)

200 209 ± 6.6 (105) 242 ± 11.1 (121) 260 ± 1.1 (130)

400 450 ± 2.8 (113) 491 ± 42.7 (123) 434 ± 29.3 (108)

800 746 ± 27.2 (93) 662 ± 50.3 (83) 793 ± 63.5 (99)

1600 1760 ± 106 (110) 1280 ± 10.3 (80)

3200 3547 ± 507 (111) 3005 ± 376 (94)

6400 6084 ± 886 (95) 7419 ± 573 (116)

102

Chapter 5

Table SI-2b. Average zinc concentrations (± SD, n=2) measured in soil 2 (pHCaCl2 = 5.71) spiked with 30 nm ZnO, 200 nm ZnO and ZnCl2. Zinc concentrations in the soil are corrected for the zinc levels measured in the controls. Recoveries (%) are presented in brackets.


Control 5.1 ± 0.5 5.1 ± 0.5 5.1 ± 0.5

100 108 ± 2.9 (108) 106 ± 4.9 (106) 110 ± 9.9 (110)

200 286 ± 17.7 (143) 217 ± 4.0 (108) 247 ± 1.5 (123)

400 514 ± 44.8 (129) 511 ± 26.8 (128) 476 ± 1.0 (119)

800 825 ± 76.9 (103) 936 ± 97.7 (117) 891 ± 43.6 (111)

1600 1607 ± 9.7 (100) 1697 ± 186 (106) 1416 ± 86.2 (89)

3200 3960 ± 116 (124) 3610 ± 126 (113)

6400 7039 ± 375 (110) 7009 ± 382 (110)

Table SI-2c. Average zinc concentrations (± SD, n=2) measured in soil 3 (pHCaCl2 = 6.39) spiked with 30 nm ZnO, 200 nm ZnO and ZnCl2. Zinc concentrations in the soil are corrected for the zinc levels measured in the controls. Recoveries (%) are presented in brackets.


Control 5.8 ± 1.4 5.8 ± 1.4 5.8 ± 1.4

100 110 ± 20.2 (110) 113 ± 10.9 (113) 120 ± 13.3 (120)

200 252 ± 23.5 (126) 279 ± 20.5 (139) 219 ± 11.5 (109)

400 534 ± 22.6 (133) 512 ± 8.7 (128) 567 ± 35.8 (142)

800 770 ± 18.0 (96) 819 ± 21.6 (102) 957 ± 165 (120)

1600 1755 ± 77.1 (110) 1785 ± 65.5 (112) 1702 ± 254 (106)

3200 3381 ± 567 (106) 2577 ± 63.6 (112)

6400 6855 ± 549 (106) 6598 ± 392 (103)

103

5


Table SI-3. Calcium concentrations measured in the pore water (mg/l) of the three test soils spiked with 30 nm ZnO, 200 nm ZnO and ZnCl2. Pore water was collected one week after saturation of the soils with deionized water (Milli-Q).

Nominal conc.(mg Zn/kg)



Control 15.3 21.4 61.7

100 16.4 20.3 48.1 11.9 19.8 50.6 48.9 51.7 33.8

200 16.3 20.2 47.5 12.0 19.7 50.1 67.6 80.5 60.5

400 16.2 20.0 46.3 12.2 19.6 49.2 105 138 114

800 15.9 19.7 44.0 12.5 19.3 47.4 180 253 221

1600 15.2 19.0 39.3 13.2 18.6 43.8 484 434

3200 13.9 17.5 29.8 14.7 17.3 36.6

6400 11.4 14.6 10.9 17.6 14.8 22.2

Table SI-4. TOC concentrations (mg/l) in the pore water from the test soils treated with 30 nm ZnO, 200 nm ZnO and ZnCl2.

Nominal conc.(mg Zn/kg d.w.)



Control 443 449 517

100 507 467 533 479 460 526 319 318 534

200 472 459 510 433 421 543 282 282 484

400 436 447 481 423 447 540 259 296 477

800 434 518 480 394 459 533 294 357 473

1600 409 501 541 416 461 542 336 487

3200 442 507 555 428 463 544

6400 467 493 444 444 460 504

Average* 452 485 506 431 453 533 288 318 491

* average of all measurements (without control)

104

Chapter 5

Tab

le S

I-5.

Zn2+

ion

conc

entr

atio

ns (M

) in

the

po

re w

ater

cal

cula

ted

with

WH

AM

7 us

ing

cal

cium

and

zin

c le

vels

(mg

/l),

soil

pH

and

TO

C le

vel (

mg

/l) o

f th

e p

ore

wat

er.

No

min

al c

onc

.(m

g Z

n/kg

d.w

.)

30 n

m Z

nO20

0 nm

ZnO

ZnC

l 2

soil

1so

il 2

soil

3so

il 1

soil

2so

il 3

soil

1so

il 2

soil

3

Co

ntro

l2.

41E

-06

3.82

E-0

97.

10E

-09

100

1.45

E-0

51.

65E

-07

1.38

E-0

81.

39E

-05

1.33

E-0

76.

16E

-08

4.83

E-0

58.

41E

-06

4.04

E-0

9

200

2.24

E-0

53.

41E

-07

2.29

E-0

82.

76E

-05

2.45

E-0

78.

19E

-08

1.65

E-0

45.

12E

-05

1.70

E-0

7

400

4.03

E-0

59.

33E

-07

8.80

E-0

84.

40E

-05

3.69

E-0

74.

35E

-07

6.55

E-0

43.

74E

-04

4.40

E-0

6

800

4.93

E-0

51.

17E

-06

5.26

E-0

71.

38E

-05

4.67

E-0

71.

71E

-06

2.94

E-0

38.

03E

-04

7.71

E-0

5

1600

2.97

E-0

56.

16E

-07

1.53

E-0

61.

46E

-05

1.39

E-0

65.

99E

-06

3.03

E-0

31.

24E

-03

3200

1.80

E-0

54.

46E

-06

1.84

E-0

67.

07E

-06

3.96

E-0

61.

07E

-05

6400

6.50

E-0

61.

41E

-05

1.08

E-0

61.

34E

-05

8.02

E-0

61.

19E

-05

105

5


6Effect of soil properties on the toxicity of ZnO nanoparticles to Folsomia candida

in a comparison of four natural soils

Abstract

The last decade, a lot of attention has been given to the important role of natural organic matter (NOM) in stabilizing nanoparticle (NP) suspensions in which NOM generally decreases NP aggregation. Coating of NP with NOM may reduce NP bioavailability in water and soils and influences NP dissolution and bioavailability. This study investigated the influence of soil organic matter content and soil pH on the toxicity of ZnO-NP and ZnCl2 to Folsomia candida in four natural soils, ranging in organic matter content from 2.37 to 14.7% and pHCaCl2 levels from 5.0 to 6.8.

The porewater Zn concentrations of ZnO-NP spiked soils were much lower than the ones measured in ZnCl2 spiked soils at similar total soil concentrations, resulting in higher Freundlich sorption constants for ZnO-NP. For ZnCl2 the porewater Zn concentrations were significantly higher in less organic soils, while for ZnO-NP the highest soluble Zn level (23 mg Zn/l) was measured in the most organic soil, which had the lowest pH. Free Zn2+ concentrations were higher for ZnCl2 than for ZnO-NP with a strong effect of porewater pH (pHpw). The 28-d EC50 values for the effect of ZnCl2 on the reproduction of F. candida increased with increasing OM content from 356 to 1592 mg Zn/kg, but for ZnO-NP no correlation between EC50 and OM was found. ZnO-NP toxicity was more related to soil pH and EC50 values increased with increasing pHCaCl2 from 1695 to 4446 mg Zn/kg. EC50 values based on porewater concentrations increased with increasing dissolved organic carbon (DOC) and H+ levels for both Zn forms. This study shows that ZnO-NP dissolution and toxicity is dependent on soil pH and that DOC has a protective effect.

Pauline L. Waalewijn-Kool, Svenja Rupp, Stephen Lofts, Claus Svendsen and Cornelis A.M. van Gestel

108

Chapter 6

1 Introduction

Nanotechnology is a fast growing technology of global economic importance, with special regard to the invention of new manufacturing methods and materials on the nanoscale (i.e. <100 nm) (Royal Society and Royal Academy of Engineering, 2004). The properties of engineered nanoparticles that make them useful in a wide range of industrial applications, however, have led to concerns regarding their potential impact on environmental health (Scown et al., 2010). Due to their small size and high reactivity engineered nanoparticles are an emerging class of contaminants with potential of damaging the environment (Manzo et al., 2013). Zinc oxide nanoparticles (ZnO-NP) have been used in a variety of products and applications such as semi-conductors, catalysts, and paints, and increasingly in consumer products such as sunscreens because of the strong ultraviolet absorption properties of ZnO. Increased production and use of ZnO-NP suggest increased exposure for organisms living in the environment (Reed et al., 2012).

Release of ZnO-NP into the environment, such as through waste water treatment plant effluent, should increase environmental exposure, although this is difficult to quantify. Modelled data indicate that ZnO concentrations might be high enough to induce adverse effects on aquatic organisms (Gottschalk et al., 2009). Compared with other engineered nanoparticles, ZnO-NP has often been found to be among the most toxic one (Aruoja et al., 2009; Adams et al., 2006). The general belief is that these particles dissolve relatively quickly and that the Zn2+ ion is the main contributor to ZnO-NP toxicity. Knowledge on the environmental fate and effects of ZnO-NP is growing, but mainly focussed on the aquatic environment. Soils are a sink for most environmental contaminants after sewage sludge applications and need to be studied as well (Tourinho et al., 2012).

Environmental conditions may act on ZnO-NP to change their size, shape and surface chemistry. Changing these basic characteristics may result in speciation products that are significantly different from the initial ZnO-NP. The interaction of nanoparticles with natural organic matter (NOM) is now receiving considerable interest, in order to better understand how these interactions might affect the stability, aggregation and dissolution in aquatic media (Quik et al., 2010). In aqueous media it was found that NOM decreased NP aggregation and increased the colloidal stability of CeO2 nanoparticles (Quik et al., 2010). NOM originates from the breakdown of plant and animal tissue in the environment. The main constituents are humic acids, fulvic acids, and a hydrophilic fraction. Several metal oxide nanoparticles are stabilized in aqueous solutions by the adsorption of NOM, due to increased electrostatic repulsion (Domingos et al., 2009; Yang et al., 2009; Zhang et al., 2009). It has been demonstrated that NOM is able to coat individual nanoparticles (Lead and Wilkinson, 2006; Lowry et al., 2012). The thickness of the coating increases with increasing humic acid concentration (Baalousha et al., 2008). Such a natural coating causes charge neutralization and colloidal stability, as the surface of uncoated metal-based nanoparticles is mostly positively charged

109

6

Effect of soil properties on the toxicity of ZnO nanoparticles to Folsomia candida

at circumneutral pH while humic acid molecules are partially deprotonated and so partly neutralize the positive charge (Bian et al., 2011). This coating by humic acids could also imply that the release of the toxic metal ions is diminished by blocking the nanoparticle surface, but so far this has not been studied in much detail.

Understanding the interactions of manufactured nanoparticles in soils is difficult, because complex interactions occur with the solid phase and the pore water. It is known that a long time is needed to reach equilibrium for soil systems spiked with ZnO-NP (Scheckel et al., 2010; Chapter 4). Complex interactions of ZnO-NP with the soil matrix could diminish the exposure of soil organisms to the pristine ZnO-NP. Unfortunately, appropriate characterization techniques for nanoparticles in soil do not exist and most techniques start with pre-treatment of the soil by preparing a suspension of soil extracts (Lead and Wilkinson, 2006). Similar to aquatic solutions, NOM is likely to affect the transformations and speciation of ZnO-NP in soils. Our hypothesis claims that with increasing OM content, ZnO-NP interacts stronger with OM by forming OM-ZnO clusters, resulting in a decreased bioavailability, dissolution and toxicity. Other soil properties, such as pH and cation exchange capacitiy (CEC) may also affect the bioavailability and toxicity of ZnO-NP. The effect of soil pH has been studied in amended field soils and an increased toxicity of ZnO-NP with decreasing pH was demonstrated for earthworms (Heggelund et al., 2013) and springtails (Chapter 5).

This study investigated how soil characteristics, with a focus on organic matter content and pH, influence the bioavailability and toxicity of ZnO-NP and ZnCl2 to the springtail Folsomia candida. Springtails are abundant in most natural soils and F. candida represents the collembolans in ecotoxicological tests (Fountain and Hopkin, 2005). Four natural soils were tested that provide a range of organic matter contents (2.37-14.7%) and slightly different pH levels (5.0-6.8). This study does not differentiate between types of NOM (quality), but just considers the total content of organic matter in soils. For both Zn forms, the 28-d EC50 values were expressed on the basis of total Zn, porewater Zn and free Zn2+ ion concentrations.


2.1 Soil propertiesFour natural soils were collected from different countries within Europe; from Portugal (Coimbra), Germany (LUFA-Speyer 2.2, Sp 2121), The Netherlands (grassland, soccer field) and United Kingdom (North Wales), hereafter called soils 1, 2, 3 and 4, respectively. These soils were chosen because they represent a wide range in organic matter contents and slightly different pH levels. The soils were homogenised, sieved through a 5 mm mesh and air dried before spiking and toxicity testing.

The organic matter content was determined as loss on ignition at 500 °C in an ashing oven. The pHCaCl2 of the soils was measured in the middle of the toxicity test, after two

110

Chapter 6

weeks exposure, in two replicate soil samples of each treatment. Soils were shaken with 0.01 M CaCl2 solution (5:1 = solution:soil) for 2 hours at 200 rpm. After settlement of the particles, the pH of the soil solution was recorded using a Consort P907 meter. The Water Holding Capacity (WHC) was determined following ISO guideline 11267 (ISO, 1999). The CEC was determined by the Silver Thiourea Method (Dohrmann R. 2006). Approx. 2 g dry soil was shaken with 25 mL 0.01 M silver thiourea complex cation (AgTU) solution for 3 hours at 200 rpm to achieve a complete exchange of all cations. Four blanks without soil were included. Ag was measured in the supernatant solution by flame Atomic Absorption Spectrometry (AAS) (Perkin Elmer AAnalyst 100). The decrease in Ag concentration is a measure for the CEC of the soil.

2.2 Test compoundsZnO nanoparticle powders (Nanosun Zinc Oxide P99/30) with a reported diameter size of 30 nm were tested and characterized by Transmission Electron Micrographs and Particle Size Distribution. In Chapter 2 it was shown that the primary particle size of the ZnO nanoparticles was in agreement with the size reported by the manufacturer. The effect of dissolved Zn was investigated by running tests with the soluble salt ZnCl2

(Merck, zinc chloride pure).Seven concentrations for ZnO-NP (nominal range 100-6400 mg Zn/kg d.w.) and

five concentrations for ZnCl2 (nominal range 100-1600 mg Zn/kg d.w.) were tested. Test concentrations were based on toxicity data found in earlier studies with ZnO-NP and F. candida (Chapters 2, 3 and 4). ZnO-NP powder was mixed with 200 g dry soil in glass jars to reach nominal test concentrations. After mixing, water was added to reach 50% of the WHC. ZnCl2 was added to the soil as a solution in deionized water (Milli-Q). Spiked soils were equilibrated for five days before starting the toxicity tests.

2.3 Soil and porewater analysisTwo samples per test concentration (± 100 mg dried soil) were taken from the spiked soils and digested in a mixture of deionized water (Milli-Q), concentrated HCl and concentrated HNO3 (1:1:4 by vol.) using a microwave oven (CEM MDS 81-D). After digestion for 7 hours at 140 °C, solutions were analysed for total zinc concentrations by flame AAS (Perkin Elmer AAnalyst 100). Certified reference material (ISE sample 989 of River Clay from Wageningen, The Netherlands) was used to ensure the accuracy of the analytical procedure.

Pore water was collected at the beginning of each toxicity test by centrifuging 30 g soil, after saturation with deionized water (Milli-Q) and three days equilibration. Soils were centrifuged for 45 min. (Centrifuge Falcon 6/300 series, CFC Free) with a relative force of 2000 g over two round filters (S&S 597 Ø 47 mm, pore size 11 μm) and a 0.45 μm membrane filter (S&S Ø 47 mm), placed inside the tubes (method cf. Hobbelen et al., 2004). Approximately 7 mL pore water per sample was collected for Zn analysis by flame AAS. In porewater samples from control soils 2 and 3 also Ca concentrations were measured by flame AAS. The dissolved organic carbon (DOC) concentration in the pore water was measured by high temperature oxidation at 850°C - 900°C

111

6


and detection of the formed CO2 (liquiTOC). The pH of the pore water (pHpw) was

measured using a Consort P907 meter.

The Zn2+ ion concentrations in the pore water from the four soils spiked with

ZnO-NP and ZnCl2 were calculated with the speciation model WHAM7 using H+, Ca,

Zn and DOC concentrations (mg/l) of the pore water and soil pH as input variables.

2.4 Toxicity tests

The springtail F. candida (Berlin strain; VU University Amsterdam) was cultured in pots

with a base of moist plaster of Paris mixed with charcoal at 20 ± 1 °C at a light/

dark regime of 12/12 h. The experiments were initiated with juveniles of the same

age (10-12 days) that were obtained by synchronising the egg laying of the culture

animals, fed with dried baker’s yeast (Dr. Oetker).

Each soil was tested in a separate toxicity test, including the two Zn forms, control soil

without Zn and Lufa 2.2 soil as a control for springtail performance. The ISO guideline 11267

for testing of chemical effects on the reproduction of springtails was followed (ISO, 1999).

Tests were conducted in 100 mL glass jars containing 30 g moist soil and five replicates

for each treatment were prepared. At the start of the test, ten synchronised animals were

transferred into each test jar. The jars were filled randomly and before introduction, the

animals were checked under the microscope for a healthy appearance. The animals were

fed a few grains of dried baker’s yeast (Dr. Oetker). The jars were incubated in a climate

room at 20 ± 1 °C at a light/dark regime of 12/12 h. Once a week, the moisture content of

test soils was checked by weighing the jars, and moisture was replenished with deionized

water (Milli-Q) when necessary. The jars were also aerated by this procedure.

After four weeks, the jars were sacrificed for determination of springtail survival and

reproduction. Each jar was emptied into a 200 ml beaker glass and 100 ml tap water

was added. The mixture was stirred carefully to let all the animals float to the surface.

The number of surviving adults and juveniles produced were counted manually after

taking a picture of the water surface using a digital camera (Olympus, C-5060).

2.5 Data analysis

Using the soil and porewater concentrations, sorption of zinc to the test soil was

described by a Freundlich isotherm:

Cs = Kf * Cwn

where,

Cs = concentration in the soil (mg Zn/kg d.w.)

Kf = Freundlich sorption constant (l/kg)

Cw = concentration in the pore water (mg Zn/l) and

n = shape parameter of the Freundlich isotherm

EC50 values for the effect on reproduction were estimated applying the logistic model

of Haanstra et al. (1985) and were determined based on total Zn, dissolved Zn and free

112

Chapter 6

Zn2+ concentrations. A generalized likelihood ratio test (Sokal and Rohlf, 1995) was applied to compare EC50 values based on total Zn concentration obtained for both Zn forms and for each soil. Calculations were performed in SPSS Statistics 20.

3 Results

3.1 Soil propertiesTable 1 summarizes the properties of the four test soils used in this study. The organic matter content measured in the different soils increased from 2.37% in soil 1 up to 14.7% in soil 4. The DOC concentrations in the pore water increased with increasing OM content from approx. 64 mg/l in soils 1 and 2 up to 3605 mg/l in soil 4. Soils 1 and 2 had nearly the same pHCaCl2 (around 5.7-5.8), while the pHCaCl2 of soil 3 was slightly higher (6.8) and soil 4 was more acidic (pHCaCl2 5.0). This was probably due to liming of soil 3 in the field, because more than a 3-fold higher Ca level was measured in the pore water of this soil (221 mg/l) compared to Lufa 2.2 (64.3 mg/l). The CEC increased with increasing OM content for soils 1, 2 and 3. The CEC was slightly lower in the more acidic soil 4 (11.8 mval/100g) compared to soil 3 (20.0 mval/100g).

3.2 Total Zn analysis and soil pHTotal Zn analyses of the four test soils showed recoveries of Zn ranging from 66.9 to 127% (Supporting Information (SI), Tables SI-1-4). Zinc concentrations in the soil were corrected for the zinc measured in the controls, which may explain the variability in the Zn recovery. The total Zn concentration in the untreated soil was 75.6, 17.1, 39.4 and 71.5 mg Zn/kg d.w. for soils 1, 2, 3 and 4, respectively. As the measured zinc concentrations in the reference material were within 6% of the certified concentrations, the measured Zn concentrations were considered reliable. The measured Zn concentrations were used for all calculations.

The pHCaCl2 and pHpw increased with increasing ZnO-NP concentrations. The highest pHCaCl2 value at 6400 mg Zn/kg d.w. was 7.2, 7.3, 7.3 and 6.6 for soils 1, 2, 3 and 4,

Table 1. Properties of Coimbra (soil 1), Lufa 2.2 (soil 2), Dutch grassland (soil 3) and North Wales (soil 4) test soils used to determine the influence of soil properties on the toxicity of ZnO nanoparticles to Folsomia candida

Soil OM (%) DOC (mg/l) pHCaCl2 WHC (g/100g) CEC (mval/100g)

1 2.37 ± 0.06 64.9 ± 4.37 5.9 32 5.17 ± 2.47

2 3.09 ± 0.04 64.1 ± 4.93 5.7 45 6.34 ± 0.81

3 10.6 ± 0.31 265 ± 1.64 6.8 73 20.0 ± 0.8

4 14.7 ± 0.18 3605 ± 279 5.0 96 11.8 ± 0.2

OM = organic matter (± SD, n=2); WHC = water holding capacity; CEC = cation exchange capacity (± SD, n=2); DOC= dissolved organic carbon, measured in the pore water (± SD, n=2)

113

6


respectively. And the highest pHpw values at 6400 mg Zn/kg d.w. were 7.7, 7.9 and 7.9 for soils 1, 2 and 3, respectively. In soil 4, the highest pHpw (7.3) was measured at 3200 mg Zn/kg d.w. and dropped to 6.7 at 6400 mg Zn/kg d.w. The pH levels for ZnCl2

were lower than for ZnO-NP and decreased with increasing concentrations of ZnCl2. The lowest pHCaCl2 levels at 1600 mg Zn/kg d.w. were 5.6, 5.3, 6.7 and 4.9 for soils 1, 2, 3 and 4, respectively and the lowest pHpw 4.8, 5.0, 7.0 and 4.7, respectively. All pH values are given in Tables SI-5-8.

3.3 Porewater Zn concentration and sorptionThe Zn concentrations in the pore water of the four test soils are presented in Table SI-7 for ZnO-NP and in Table SI-8 for ZnCl2. In soils 2, 3 and 4 the Zn concentrations in the pore water increased with increasing ZnO-NP concentrations, up to a maximum of 2.83, 3.52 and 23.8 mg Zn/l, respectively. So, the highest porewater concentration was measured in soil 4 having the highest organic matter content and the lowest pHCaCl2. For soil 1 the maximum porewater concentration of 3.09 mg Zn/l was measured at an intermediate concentration of 800 mg Zn/kg d.w. The porewater Zn concentrations for ZnCl2 were much higher than for ZnO-NP spiked soils. For soils 1, 2, 3 and 4 the porewater Zn levels increased with increasing soil concentrations up to 1020, 720, 10.4 and 63.7 mg Zn/l, respectively. This difference in porewater concentration was clearly related with the OM content and pH of the four soils.

The porewater Zn levels were translated into Freundlich sorption constants for ZnO-NP and ZnCl2, indicating zinc availability in the four soils. Table 2 shows that the Kf value for ZnCl2 increased with OM content for soils 1, 2 and 3, and with CEC for all soils. The highest Kf value of 458 l/kg (n = 0.572) was calculated for soil 3 having the highest pH. For ZnO-NP the highest Kf value of 1146 l/kg (n = 1.08) was also calculated for soil 3. A relationship between Kf constants and OM content could not be discovered for ZnO-NP, as the lowest Kf value of 97.1 l/kg (n = 1.45) was estimated for the most organic soil (soil 4). For ZnO-NP, the Kf constants increased with increasing soil pH, suggesting higher availability in the more acidic soil 4. In all

Table 2. Freundlich sorption constants Kf (l/kg) and shape parameter n for Zn partitioning in four soils spiked with ZnO-NP and ZnCl2. See Table 1 for soil properties.

SoilZnO-NP ZnCl2

Kf n Kf n

1 - - 37.1 0.535

2 329 2.10 83.2 0.424

3 1146 1.08 458 0.572

4 97.1 1.45 146 0.608

- Kf and n values could not be estimated, because the log porewater concentrations did not increase linearly with log soil concentrations

114

Chapter 6

soils spiked with ZnO-NP the shape parameter n of the Freundlich isotherm was higher than one, suggesting saturation of the pore water or Zn precipitation. For all soils spiked with ZnCl2 the shape parameter n was below one, suggesting saturation of the solid phase at high porewater concentrations.

3.4 Free Zn2+ ion concentrationsThe free Zn2+ ion concentrations in the pore water calculated using WHAM7 are presented in Table SI-9 for ZnO-NP and in Table SI-10 for ZnCl2. The free Zn2+ ion concentrations for ZnCl2 were much higher than for ZnO-NP spiked soils. For ZnCl2, the free Zn2+ ion concentrations increased with increasing soil concentrations up to 5501, 4189, 15.0 and 176 µM for soils 1, 2, 3 and 4, respectively. The free Zn2+ ion concentrations for ZnCl2 increased also with decreasing pHpw; the lowest concentration of 15.0 µM was determined in soil 3, having the highest pHpw. The free Zn2+ ion concentrations for ZnO-NP increased with increasing soil concentration for soil 3 with a maximum of 2.70 µM. In the other three soils, the free Zn2+ ion concentrations showed a peak at intermediate soil concentrations of 11.2, 6.18 and 1.50 µM in soils 1, 2 and 4, respectively. For higher soil concentrations the free Zn2+ ion concentrations decreased to 4.38, 3.84 and 0.0001 µM for soils 1, 2 and 4, respectively, with increasing pHpw.

3.5 Toxicity Control performance of the collembolans was affected in the most organic soil with the lowest pHCaCl2. Control survival after 28 days exposure in control soils 1, 2, 3 and 4 was 86, 96, 90 and 68%, respectively. The average number of juveniles in the controls was 263, 475, 514 and 114 for soils 1, 2, 3 and 4, respectively, with coefficients of variance of 68, 60, 6.4 and 26%, respectively.

No effect on springtail survival was found in the four test soils spiked with ZnO-NP up to 6400 mg Zn/kg d.w. or with ZnCl2 up to 1600 mg Zn/kg d.w. Reproduction was reduced in a dose-dependent manner for the two Zn forms in all four soils (See Figures SI-1 and SI-2 for all dose-response curves). Table 3 shows the EC50 values for the effect of ZnO-NP and ZnCl2 on the reproduction of F. candida after 28 days exposure to the four soils. The EC50 values are presented as total Zn concentrations (mg Zn/kg d.w.), as porewater Zn concentrations (mg Zn/l) and as free Zn2+ ion concentrations (also in mg Zn/l in order to compare them with the EC50 values based on porewater Zn concentration).

The EC50 values for ZnCl2 increased with increasing OM content from 356 mg Zn/kg d.w. in soil 1 up to 1592 mg Zn/kg d.w. in soil 4. Linear regression of the relation between these EC50 values and soil OM content resulted in an R2 of 0.967 (Figure 1). The EC50 value for soil 3 was significantly higher than the ones for soils 1 and 2, according to a generalized likelihood-ratio test (χ2

(1) = 24.6 (soil 1 vs 3) and 21.7 (soil 2 vs 3), p < 0.05). No clear pattern could be seen for the relation between ZnO-NP toxicity and OM content of the test soils. The estimated EC50 values ranged between 1695 and 4446 mg Zn/kg d.w. The highest toxicity was observed in the most organic soil, but its EC50 value (1695 mg Zn/kg d.w.) was not significantly lower than the ones

115

6


for the other three soils with lower OM content. The lowest toxicity was observed in soil 3 with the highest pHCaCl2, but also this EC50 was not significantly higher than the ones for the other three soils with lower pH levels. Regression of EC50 values with soil pHCaCl2 resulted in an R2 of 0.787, showing a good correlation between toxicity and pH for ZnO-NP. For ZnCl2 the relation between EC50 values and soil pHCaCl2 based on total Zn concentrations was poor, and was a factor 2 weaker than for ZnO-NP (Figure 1). The EC50 values for ZnCl2 were significantly lower than for ZnO-NP in soils 1 and 2, according to a generalized likelihood-ratio test (χ2

(1) = 18.9 (soil 1) and 21.2 (soil 2), p < 0.05). In soil 4 the EC50 values for ZnO-NP and ZnCl2 were almost similar, namely 1695 and 1592 mg Zn/kg d.w., respectively.

Based on porewater Zn concentrations, the EC50 values for ZnO-NP ranged between 2.60 and 7.01 mg Zn/l and increased with increasing DOC concentrations of the pore water (R2 = 0.991, Figure 1). The highest EC50 value (7.01 mg Zn/l) was estimated for soil 4, but this was not significantly higher than the EC50 of 2.60 and 3.03 mg Zn/l estimated for soils 2 and 3, respectively. For all soils, higher EC50 values based on porewater concentrations were estimated for ZnCl2 than for ZnO-NP. The lowest EC50 of 10.2 mg Zn/l for ZnCl2 was estimated for soil 3, higher EC50 values for the other three soils ranged from 33.8 to 55.8 mg Zn/l. The EC50 for ZnCl2 related poorly to DOC concentrations of the pore water, but clearly decreased with increasing pHpw (R2 = 0.846, Figure 1).

For both Zn forms, the toxicity was higher on the basis of free Zn2+ concentrations than on the basis of porewater concentrations, as shown by the lower EC50 values

Table 3. EC50 values (with 95% confidence intervals) for the effect on reproduction of Folsomia candida after 28-d exposure to ZnO-NP and ZnCl2 in the four soils. See Table1 for soil properties. EC50 values are presented as total Zn concentrations in mg Zn/kg d.w. (left) as porewater Zn concentrations in mg Zn/l (middle) and as free Zn2+ ion concentrations in mg Zn/l (right)

Total soil (mg Zn/kg d.w.) Pore water (mg Zn/l) Free Zn (mg Zn/l)

ZnO-NP ZnCl2 ZnO-NP ZnCl2 ZnO-NP ZnCl2

soil 1 29621a (1389-4534)

3561b

(24-5378)> 3.09*

(-)54.9(-)

> 0.728*(-)

27.0(-)

soil 2 34931a

(358-6628)4391b

(316-561)2.60

(2.37-2.84)33.8(-)

0.368(0.259-0.476)

14.8(-)

soil 3 44461

(2830-6061)14332

(-)3.03

(2.53-3.53)10.2(-)

0.147(0.118-0.176)

0.969(-)

soil 4 16951a

(784-2605)159212a

(-)7.01

(5.22-8.81)55.8(-)

> 0.093* (-)

10.3(-)

1,2 indicate significant differences between EC50 values of ZnO-NP and ZnCl2 for the different soils according to a generalized likelihood-ratio test (χ2

(1) > 3.84; p < 0.05)a,b indicate significant differences between EC50 values based on total Zn concentration for the different Zn forms according to a generalized likelihood-ratio test (χ2

(1) > 3.84; p < 0.05)- Data did not allow calculating reliable 95% confidence intervals* EC50 is above highest porewater Zn/free Zn2+ ion concentration measured/calculated

116

Chapter 6

in Table 3. For ZnCl2 the EC50 based on free Zn2+ concentrations was 0.969 mg Zn/l for soil 3 and ranged from 10.3 to 27.0 mg Zn/l for the other three soils. For ZnO-NP, the estimated EC50 values based on free Zn2+ concentrations were 0.368 and 0.147 mg Zn/l for soils 2 and 3, respectively. It was not possible to estimate EC50 values on the basis of free Zn2+ concentrations for soils 1 and 4, due to the peak in free Zn2+ concentrations at intermediate soil concentrations making that reduction in reproduction was not consistently related with free Zn2+ concentration.

4 Discussion

In light of nanoparticle risk assessment, soil properties need to be taken into account to predict the bioavailability and toxicity of ZnO-NP. The most important soil properties determining equilibrium partitioning of metals in soils are the adsorption phases (clay, OM and hydroxides), the number of available sorption sites (CEC) and pH (Janssen et al., 1997). The aim of this study was to investigate the effect of soil properties (initially only OM content) on ZnO-NP dissolution and toxicity, but it is difficult to evaluate the effect of OM content separately from other soil properties such as pH. The presence of humic and fulvic acids and other organic substances that may adsorb to the surface of nanoparticles will likely change ZnO-NP dissolution and toxicity. In this study, the

ZnCl2ZnO-NPLinear (ZnCl2)Linear (ZnO-NP)

y = 779x - 3720R² = 0.380

y = 1555.4x - 7194.7R² = 0.7871

0500

100015002000250030003500400045005000

4 5 6 7 8

y = -21.1x + 163R² = 0.846

y = -2.50x + 22.4R² = 0.527

0

10

20

30

40

50

60

70

4 5 6 7 8 9

y = 107x + 135R² = 0.967

y = -65.4x + 3652R² = 0.116

0500

100015002000250030003500400045005000

0 5 10 15 20

y = 0.0061x + 32.6R² = 0.239

y = 0.0012x + 2.75R² = 0.991

0

10

20

30

40

50

60

70

0 1000 2000 3000 4000

pHCaCl2 pHpw

soil OM content (%) porewater DOC concentration (mg/l)

EC50

(mg

Zn/k

g d.

w.)

EC50

(mg

Zn/l)

Figure 1. Estimated EC50 values for the effect of Zn on the reproduction of Folsomia candida based on total Zn concentrations (mg Zn/kg d.w.) and porewater Zn concentrations (mg Zn/l) in relation to pHCaCl2, pHpw, soil organic matter content and DOC concentrations in the pore water (see Table SI-5 to SI-8 for measured pH values and Table 1 for soil properties). Data used from 28-d toxicity tests in the four test soils freshly spiked ZnO-NP and ZnCl2 (see Table 3)

117

6


dissolution and toxicity of ZnO-NP and ZnCl2 to F. candida have been determined in four natural soils, ranging in OM content from 2.37 to 14.7% and in pHCaCl2 from 5.0 to 6.8. The EC50 values for ZnCl2 based on total soil concentrations increased with increasing OM content, but no correlation between toxicity and OM content was found for ZnO-NP. Soil pH had a major influence on the EC50 values estimated on the basis of porewater Zn concentrations. It was already found that a decrease in soil pH increased ZnO-NP dissolution and induced higher toxicity to earthworms (Heggelund et al., 2013) and springtails (Chapter 5).

The pH of the soil increased after the spiking of ZnO-NP in contrary to a decrease of soil pH after spiking ZnCl2. The addition of ZnCl2 at high concentrations strongly decreases the pH through the displacement of protons from soil surfaces. The increase in soil pH for ZnO-NP can be explained by negatively charged OH- species that result from the reaction ZnO + H2O ↔ Zn2+ + 2OH-.

For ZnCl2, the porewater Zn concentrations were much higher in less organic soils (containing approx. 2.5% OM) than in soils containing approx. 10 to 14% OM. This is explained by the number of binding sites available to bind zinc, because it is known that humic and fulvic acids form chelates with Zn ions (Alloway, 1990). The porewater Zn levels in the more organic soils also showed that the CEC influenced the soluble Zn concentrations. The lowest porewater Zn concentrations were measured in the soil with the highest CEC. For both zinc forms, the lowest dissolved Zn concentrations and the highest sorption constants were estimated for the Dutch grassland soil with the highest pH. This can be explained by the absence of competition with H+ ions for Zn binding at the negatively charged phases in the soil when pH levels are increasing (Alloway, 1990).

The solubility of ZnO-NP, measured as dissolved Zn centrifuged over a 0.45 µm filter, is much lower than for the soluble metal salt ZnCl2 in all four soils. For ZnO-NP, the highest soluble Zn levels (max. 23 mg Zn/l) were measured in the most organic soil. It cannot be excluded that the dissolution of ZnO-NP in this soil was stimulated by the lower soil pH. And the larger amount of water needed to reach 50% WHC of this soil may have triggered a faster dissolution of ZnO-NP. Nevertheless, this is the first study that shows that ZnO-NP dissolution results in different available Zn concentrations in soils with different soil properties. The aggregation and dissolution of ZnO-NP in different soils have not been explored much. Aquatic studies demonstrated that OM can have a “masking” effect, either by direct coating of the nanoparticle surface or by minimizing dissolution (Lowry et al., 2012). Once coated and aggregated in either the solid phase or in the pore water, it can be more difficult for the ZnO nanoparticles to release Zn into the solution. In soils, the equilibrium processes of metals between pore water and solid phases are rather complex (Tipping et al., 2003). Although increasing solution pH will decrease metal solubility, it may increase dissolved metal concentrations because of formation of metal complexes with DOC in the porewater. DOC may coat reactive adsorption sites, inactivating them and therefore indirectly inhibiting metal adsorption. The shape parameter n of the Freundlich isotherm was higher than one for ZnO-NP

118

Chapter 6

spiked soils, supporting this idea of saturation of the pore water with Zn complexes. In our soils the OM could have enhanced the release of Zn from the ZnO-NP, by for example assisting in breaking up large aggregates (de-agglomeration). Several aquatic studies with ZnO-NP and humic acids or clay minerals support this hypothesis (Bian et al., 2011; Li et al., 2011a; Miao et al., 2010; Scheckel et al., 2010).

Our experimental data showed that both soil pH and DOC concentrations influenced ZnO-NP dissolution. The measured porewater Zn concentrations reflected to some extent the effect of pH on ZnO-NP solubility. The pH levels increased with increasing porewater Zn concentrations, and the latter peaked at intermediate soil concentrations for soil 1. The pH where the net surface charge of ZnO-NP is zero and highest stability of the particles is expected (point of zero charge, pzc), ranges between 7.5 and 9.8 in water (Kosmulski, 2004; Keller et al., 2010) and above 8 in soil (Collins et al., 2012). The pHCaCl2 values of our test soils were below this pzc, suggesting low stability of the particles. The pH values of the pore water of soils 1, 2, and 3, however, were closer to this pzc (in between 7 and 8). In theory, an increase in stability could lead to an increase in Zn release from the nanoparticles due to a larger surface area of the individual particles compared to aggregated nanoparticles. The peak in porewater and free Zn2+ concentrations for soils 1, 2 and 4 suggested a reduction of Zn release above a pHpw of 7. Such a peak has been observed before in Lufa 2.2 soil for ZnO-NP in aged soils (Chapter 4), but the present study showed that also in freshly spiked soils increasing pH can reduce ZnO-NP dissolution at high exposure concentrations and that its pore water can become saturated with Zn after 7 to 14 days.

For ZnO-NP no significant effect (increase or decrease) of soil OM content on springtail reproduction was detected, while the EC50 values for ZnCl2 increased with increasing OM content. The EC50 for ZnO-NP in Lufa 2.2 soil was 3493 mg Zn/kg d.w. which was comparable to the EC50 of 3159 mg Zn/kg d.w. estimated in previous experiment with this type of ZnO-NP (Chapter 2). The higher pH in the Dutch grassland soil most probably resulted in the lower toxicity observed for ZnO-NP and ZnCl2 based on total Zn concentrations. Zinc toxicity has been shown to be related to OM and clay content in freshly contaminated OECD artificial soil with 10% OM (Smit and van Gestel, 1998). The effect of OM on ZnO-NP toxicity is still poorly understood, with many studies producing contradictory results. Li et al. (2011a) found a reduced toxicity of ZnO-NP to bacteria with increasing DOC concentrations in aquatic media, while Blinova et al. (2010) reported that DOC could not decrease the acute toxicity of ZnO-NP to crustaceans in natural water. Li et al. (2011b) exposed Eisenia fetida for 96 hours to soil extracts and found that addition of humic acids enhanced ZnO-NP dissolution, but reduced toxicity.

Released Zn may form steady organic complexes, which is likely to occur when more DOC is present in the soil pore water (Fang et al., 2009). From our study, it remained unclear if the effect of DOC concentrations on the EC50 values based on porewater concentrations (for both Zn forms) was significant. A protective effect of DOC in the pore water may have occurred in our most organic soil, but it may also

119

6


have been an effect of pH. DOC concentrations of the pore water of soil 4 were approx. 50 and 10 times higher than of soils 2 and 3, respectively. The lower pHpw of soil 4 suggested that H+ ions may have competed with dissolved Zn species resulting in lower toxicity as well. The lowest EC50 based on porewater concentrations for both Zn forms was estimated for soil 3 and can be explained by a substantial higher pHpw compared to the pH levels in the other three soils. According to the theory of the terrestrial Biotic Ligand Model the reduction of competition with H+ ions may lead to higher toxicity (Thakali et al., 2006). This effect was also observed for the EC50 values based on free Zn2+ ion concentrations, in particular for ZnCl2, which suggests that the ionic Zn was a good predictor for toxicity.

In summary, the toxicity was related to soil OM content for ZnCl2 but not for ZnO-NP, suggesting that ZnO-NP dissolution is limiting Zn bioavailability. Soil pH and DOC play an important role in the dissolution of ZnO-NP (release into the pore water). Further research is necessary on interactions between ZnO-NP and soil components such as OM, affecting ZnO-NP toxicity, but it cannot be seen separately from the role of other soil properties, such as pH.


• Dose-response curves for ZnO-NP (Figure SI-1) and ZnCl2 (Figure SI-2) in the four test soils

• Total Zn analysis in the four test soils spiked with ZnO-NP and ZnCl2 (Table SI-1-4) • pHcacl2 of the four test soils spiked with ZnO-NP (Table SI-5) and ZnCl2 (Table SI-6) • Zn concentrations and pH in the pore water collected from soils spiked with

ZnO-NP (Table SI-7) and ZnCl2 (Table SI-8) • Calculated free Zn2+ ion concentrations in the pore water for collected from soils

spiked with ZnO-NP (Table SI-9) and ZnCl2 (Table SI-10) using WHAM7

120

Chapter 6

0

50

100

150

200

250

300

0 5 10 15 20 250

50

100

150

200

250

300

0 2000 4000 6000 8000 10000

0

200

400

600

800

1000

0.00 1.00 2.00 3.00 4.000

200

400

600

800

1000

0 2000 4000 6000 8000

0

200

400

600

800

1000

0.00 1.00 2.00 3.000

200

400

600

800

1000

0 2000 4000 6000 8000

0

200

400

600

800

1000

0 2000 4000 6000 80000

200

400

600

800

1000

0.00 1.00 2.00 3.00 4.00

num

bero

f juv

enile

sper

test

jar

Zn concentration soil (mg Zn/kg d.w.)

soil 1

porewater Zn concentration (mg Zn/l)

num

bero

f juv

enile

sper

test

jar

soil 1

soil 2 soil 2

soil 3 soil 3

soil 4 soil 4

Figure SI-1. Effect of ZnO-NP on the reproduction (number of juveniles) of Folsomia candida after 28 days exposure in four different test soils. See Table 1 for soil properties. Measured concentrations of zinc in the soil (left) and the pore water (right) are provided on the x-axis. Line shows fit obtained with a logistic model.

121

6


Figure SI-2. Effect of ZnCl2 on the reproduction (number of juveniles) of Folsomia candida after 28 days exposure in four different test soils. See Table 1 for soil properties. Measured concentrations of zinc in the soil (left) and the pore water (right) are provided on the x-axis. Line shows fit obtained with a logistic model.

0

200

400

600

800

0 500 1000 15000

200

400

600

800

0 500 1000 1500 2000

050

100150200250300350

0 20 40 60 800

50100150200250300350

0 500 1000 1500 2000

0

200

400

600

800

1000

0 5 10 150

200

400

600

800

1000

0 500 1000 1500 2000

0

200

400

600

800

1000

0 200 400 600 8000

200

400

600

800

1000

0 500 1000 1500

num

bero

f juv

enile

sper

test

jar


num

bero

f juv

enile

sper

test

jar

soil 1 soil 1

soil 2 soil 2

soil 3 soil 3

soil 4 soil 4

122

Chapter 6

Table SI-1. Average zinc concentrations (± SD, n=2) measured in soil 1 (Coimbra soil) spiked with 30 nm ZnO and ZnCl2. Zinc concentrations in the soil are corrected for the zinc levels measured in the controls. Recoveries (%) are presented in brackets.

Nominal conc. (mg Zn/kg) 30 nm ZnO ZnCl2

Control 75.6 ± 4.61

100 72.7 ± 42.3 (72.7) 79.9 ± 3.09 (79.9)

200 199 ± 7.21 (99.3) 134 ± 21.5 (66.9)

400 405 ± 100 (101) 328 ± 2.47 (82.0)

800 766 ± 49.7 (95.8) 785 ± 114 (98.1)

1600 1721 ± 50.2 (108) 1523 ± 20.5 (95.2)

3200 3361 ± 488 (105)

6400 6722 ± 164 (105)

Table SI-2. Average zinc concentrations (± SD, n=2) measured in soil 2 (Lufa 2.2) spiked with 30 nm ZnO and ZnCl2. Zinc concentrations in the soil are corrected for the zinc levels measured in the controls. Recoveries (%) are presented in brackets.


Control 17.1 ± 0.334

100 93.6 ± 2.99 (93.6) 107 ± 5.9 (107)

200 202 ± 1.18 (101) 192 ± 15.8 (95.8)

400 507 ± 93.4 (127) 424 ± 48.2 (106)

800 830 ± 66.6 (104) 788 ± 1.65 (98.5)

1600 1661 ± 73.5 (104) 1420 ± 193 (88.8)

3200 3164 ± 187 (98.9)

6400 6925 ± 272 (108)

123

6


Table SI-3. Average zinc concentrations (± SD, n=2) measured in soil 3 (Dutch grassland) spiked with 30 nm ZnO and ZnCl2. Zinc concentrations in the soil are corrected for the zinc levels measured in the controls. Recoveries (%) are presented in brackets.


Control 39.4 ± 2.96

100 86.0 ± 6.84 (86.0) 94.7 ± 7.16 (94.7)

200 189 ± 3.91 (94.5) 173 ± 0.93 (86.7)

400 392 ± 5.17 (98.1) 415 ± 7.83 (104)

800 820 ± 40.0 (102) 799 ± 52.9 (99.9)

1600 1644 ± 74.0 (103) 1441 ± 21.1 (90.1)

3200 3356 ± 158 (105)

6400 6174 ± 433 (96.5)

Table SI-4. Average zinc concentrations (± SD, n=2) measured in soil 4 (North Wales soil) spiked with 30 nm ZnO and ZnCl2. Zinc concentrations in the soil are corrected for the zinc levels measured in the controls. Recoveries (%) are presented in brackets.


Control 71.5 ± 1.35

100 113 ± 9.92 (113) 125 ± 0.19 (125)

200 257 ± 56.9 (128) 261 ± 23.8 (130)

400 455 ± 111 (114) 497 ± 27.1 (124)

800 948 ± 125 (118) 884 ± 130 (111)

1600 1365 ± 96.0 (85.3) 1705 ± 475 (107)

3200 3612 ± 158 (113)

6400 7997 ± 5.93 (125)

124

Chapter 6

Table SI-5. pHcacl2 of the four test soils spiked with ZnO-NP. Soil pH was measured approx. three weeks after spiking the soils, which is halfway the toxicity test with Folsomia candida

Nominal conc. (mg Zn/kg)

Coimbra (soil 1)

Lufa 2.2 (soil 2)

Dutch grassland (soil 3)

North Wales (soil 4)

Control 5.9 5.7 6.8 5.0

100 6.0 5.7 6.8 5.0

200 6.1 5.8 6.8 5.1

400 6.3 6.0 6.9 5.1

800 6.4 6.2 6.9 5.2

1600 6.7 6.5 7.0 5.7

3200 7.0 6.7 7.1 6.1

6400 7.2 7.3 7.3 6.6

Table SI-6. pHcacl2 of the four test soils spiked with ZnCl2. Soil pH was measured approx. three weeks after spiking the soils, which is halfway the toxicity test with Folsomia candida


Coimbra (soil 1)

Lufa 2.2 (soil 2)

Dutch grassland (soil 3)


Control 5.9 5.7 6.8 5.0

100 6.0 5.8 6.9 5.0

200 5.8 5.7 6.9 5.0

400 5.8 5.6 6.8 5.0

800 5.7 5.5 6.8 4.9

1600 5.6 5.3 6.7 4.9

125

6


Table SI-7. Zinc concentrations measured in the pore water (mg Zn/l) of the four test soils spiked with ZnO-NP. The pH of the pore water is also reported (pHpw). Pore water was collected three days after saturation of the soils with deionized water (Milli-Q) and approx. two weeks after spiking the soils


Coimbra (soil 1) Lufa 2.2 (soil 2) Dutch grassland (soil 3) North Wales (soil 4)

mg Zn/l pHpw mg Zn/l pHpw mg Zn/l pHpw mg Zn/l pHpw

Control 0.025 6.3 0.028 7.1 0.008 7.5 0.501 5.9

100 1.24 6.2 0.466 6.9 0.082 7.6 1.17 5.9

200 2.00 7.0 0.875 6.9 0.178 7.7 1.90 6.0

400 2.35 6.9 1.56 7.0 0.426 7.6 2.96 6.0

800 3.09 7.0 2.14 7.2 0.811 7.6 4.76 6.2

1600 2.46 7.6 2.61 7.4 1.78 7.6 6.74 6.7

3200 1.94 7.7 2.24 7.9 2.66 7.8 9.70 7.3

6400 1.88 7.7 2.83 7.9 3.52 7.9 23.8 7.0

Table SI-8. Zinc concentrations measured in the pore water (mg Zn/l) of the four test soils spiked with ZnCl2. The pH of the pore water is also reported (pHpw). Pore water was collected three days after saturation of the soils with Milli-Q water and approx. two weeks after spiking the soils


Coimbra (soil 1) Lufa 2.2 (soil 2) Dutch grassland (soil 3) North Wales (soil 4)

mg Zn/l pHpw mg Zn/l pHpw mg Zn/l pHpw mg Zn/l pHpw

Control 0.025 6.3 0.028 7.1 0.008 7.5 0.501 5.9

100 3.08 6.2 1.81 6.8 0.101 7.6 2.40 5.2

200 20.4 6.0 8.40 6.5 0.179 7.5 1.63 5.5

400 52.6 5.7 33.3 6.2 0.477 7.5 3.88 5.3

800 260 5.4 259 5.7 2.14 7.2 17.2 5.0

1600 1021 4.8 720 5.0 10.4 7.0 63.7 4.7

126

Chapter 6

Table SI-9. Predicted free zinc concentrations in the pore water (µmol) of the four test soils spiked with ZnO-NP using WHAM7. Calculations were based on total Zn concentrations (mg Zn/kg d.w.), pHCaCl2, porewater Zn concentrations (mg Zn/l), pHpw, cation exchange capacity (mval/100g), organic matter content (%), water holding capacity (ml/100g), dissolved organic carbon (mg/l), colloidal fulvic acid (mg/l) and calcium concentrations in the pore water (mg/l). Percentage free Zn2+ concentrations of the dissolved Zn fraction measured in the pore water are presented in between brackets.


Coimbra(soil 1)

Lufa 2.2(soil 2)

Dutch grassland(soil 3)


Control 0.076 (19.8) 0.032 (7.48) 0.003 (2.69) 0.047 (0.610)

100 5.76 (30.3) 0.822 (11.5) 0.032 (2.54) 0.108 (0.595)

200 6.37 (20.7) 1.78 (13.2) 0.068 (2.48) 0.133 (0.454)

400 8.12 (22.5) 3.58 (15.0) 0.210 (3.21) 0.150 (0.329)

800 11.2 (23.6) 4.96 (15.0) 0.453 (3.63) 0.120 (0.163)

1600 6.58 (17.4) 6.18 (15.4) 1.21 (4.39) 0.016 (0.016)

3200 4.53 (15.1) 3.84 (11.1) 1.93 (4.72) 0.0001 (0.0001)

6400 4.38 (15.2) 5.48 (12.6) 2.70 (4.99) 0.018 (0.005)

Table SI-10. Predicted free zinc concentrations in the pore water (µmol) of the four test soils spiked with ZnCl2 using WHAM7. Calculations were based on total Zn concentrations (mg Zn/kg d.w.), pHCaCl2, porewater Zn concentrations (mg Zn/l), pHpw, cation exchange capacity (mval/100g), organic matter content (%), water holding capacity (ml/100g), dissolved organic carbon (mg/l), colloidal fulvic acid (mg/l), chloride and calcium concentrations in the pore water (mg/l).


Coimbra(soil 1)

Lufa 2.2(soil 2)

Dutch grassland(soil 3)


Control 0.076 (19.8) 0.032 (7.48) 0.003 (2.69) 0.047 (0.610)

100 15.2 (32.2) 4.70 (16.9) 0.039 (2.52) 1.51 (4.10)

200 143 (45.5) 41.4 (32.0) 0.076 (2.77) 0.547 (2.18)

400 383 (47.3) 224 (43.7) 0.229 (3.12) 2.06 (3.46)

800 1735 (43.3) 1822 (45.6) 1.63 (4.96) 25.0 (9.45)

1600 5501 (35.0) 4189 (37.8) 15.0 (9.43) 176 (18.0)

127

6


7Summary and discussion

Although the size-dependent properties make engineered nanoparticles desirable for many applications, the increased use inevitable lead to its release into the environment, for instance via wastewater or sewage sludge. Contamination of the soil environment with engineered nanoparticles may be harmful for soil invertebrates living in there, such as earthworms, isopods and springtails. In the previous five experimental chapters the toxicity of ZnO-NP to the springtail Folsomia candida was described for different ZnO-NP (particle size and nanoparticle coating) and different soil properties (soil pH and organic matter content). Full dose-response curves for ZnO-NP were presented for springtail reproduction using a wide range of test concentrations (100-6400 mg Zn/kg d.w.). Also ZnO-NP bioavailability and dissolution were studied by equilibrating spiked soils for one year in the laboratory. A comprehensive ecotoxicological assessment of ZnO-NP was presented and the data provide new and essential information for future risk assessment of ZnO-NP in the soil environment. Here, I would like to discuss methodological issues regarding ecotoxicity testing of metal-based nanoparticles and to reflect on the different toxicity and behaviour of ZnO-NP and ZnCl2 in soils. Recommendations on ZnO-NP soil toxicity testing will be given and I will highlight some possibilities for further research. The chapter ends with concluding remarks on the main findings of this research.

1 Ecotoxicity testing of metal-based nanoparticles in soil

Due to their unique properties, existing analytical methods and test approaches for assessing environmental risk may not be appropriate for metal-based nanoparticles. Nano-ecotoxicology is currently in its infancy and environmental testing of nanoparticles may require developing new test guidelines (Kahru and Dubourguier, 2010; Handy et al., 2008, 2012). However, during investigations of the toxicity of metal-based nanoparticles it is important to take into account the existing knowledge on metal toxicity. At this early stage of nanoparticle research the approaches and tools that were developed for metal toxicity testing in the past can be useful. Additional nanoparticle characterization and adaptations to the test guidelines may be necessary, but only based on good scientific research and careful considerations. This is emphasized in a scientific letter in Appendix II of this thesis (van Gestel et al., 2010. Metal-based nanoparticles in soil: New research themes should not ignore old rules and theories. Comments on the paper by Hu et al., 2010). Methodological issues for soil toxicity testing of ZnO-NP and springtails are specifically addressed here.

Toxicity tests with F. candida usually are performed following ISO guideline 11267 (ISO, 1999) or the OECD guideline 232 (OECD, 2009). This method, which focuses on determining the effects on survival and reproduction of the springtails after 28 days of exposure, is suitable without amendments for determining the toxicity of ZnO-NP. The “only” aspects that need consideration are the method of spiking, analytical technique for total Zn analysis and soil type.

131

7

Summary and discussion

Considering the spiking method, two different methods have been tested and evaluated: from a suspension in a soil extract and as dry powder mixed in with dry soil (Chapter 2). The first method is based on dispersing dry powdered ZnO-NP in a soil extract (according to van der Ploeg et al., 2011). Due to association with the dissolved organic carbon and possibly other compounds in the soil extract, a suspension of ZnO-NP is obtained. At high concentrations, however, the suspension was not homogeneous as there still was precipitation. By shaking the suspension and bringing it “in one go” into the soil (in this way also reaching the desired soil moisture content), all ZnO-NP could be transferred into the test soil. This method produced a good recovery and homogeneous distribution of the Zn in the test soils (coefficient of variance (CV) for five subsamples per test concentration < 10%). Also the second method produced a good recovery and homogeneous distribution of Zn in the test soils (CV for five subsamples per test concentration < 10%). This experiment showed that both spiking methods resulted in homogeneous mixing of the ZnO-NP through the soil. The outcome of the toxicity tests with 30 nm ZnO was also reproducible in Lufa 2.2 soil using the dry spiking method with EC50 values of 3493 and 3159 mg Zn/kg (Chapters 2 and 6).

While working with the ZnO-NP powders and the soils, a personal preference for the dry spiking method was developed, because dry powder was mixed very easily with dry soil. Mixing the dry ZnO-NP directly into the soil allows the transformation and dissolution to occur in the soil, preventing that these processes already can occur in the suspensions. Using suspension spiking, one cannot be certain that the pristine ZnO-NP were tested. But this method is a more realistic scenario, because ZnO-NP will enter natural soils via the sludge resulting from waste water treatment, also in suspensions and not as dry powders. An advantage of suspension spiking is that the dissolution and particle size characterization of ZnO-NP can still be monitored in the soil extracts used for suspension spiking. The scientific debate on spiking nanoparticles to environmental media is ongoing. In the 2nd OECD Horizontal Workshop on Ecotoxicity and Environmental Fate (Berlin, 2013), it was suggested that spiking should at least be similar in ecotoxicity and fate tests for the sake of comparison. For soils, a suspension spiking is recommended, but dry spiking may also be justified.

The traditional metal analysis used in my experiments has shown to provide a reliable estimate of Zn concentration of a digested soil sample. Recoveries of Zn in spiked soil samples with ZnO-NP were most times higher than 80% using digestion with concentrated HCl:HNO3 (1:4) and detection by flame Atomic Absorption Spectroscopy (AAS). Also after one year ageing of ZnO-NP spiked soils the measured of total Zn concentrations were in good agreement with the nominal Zn concentrations. So, the Zn analyses showed that total Zn concentrations can be measured by AAS after digestion of the ZnO-NP spiked soils with aqua regia. It must be realized that the toxicity data is expressed on the basis of total Zn concentrations in soil (as mg Zn/kg) and not on the basis of g ZnO-NP/kg, for example. ZnO-NP powder was weighed on the basis of the amount of Zn needed to reach nominal test concentrations taking into account the molecular

132

Chapter 7

weight of ZnO and Zn. This enables comparison with data obtained from studies with non-nano ZnO and ZnCl2, but ignores the nanoparticle form of ZnO-NP in the soil during the 28 day exposure period of the test. The ZnO-NP could have been transformed or dissolved during the test period in the soil. Chemical transformation of test substances in test soils is not a new issue for ecotoxicity testing. Organic compounds may already degrade after 28 days and sparingly soluble metal compounds may partly dissolve. Transformation tests and ageing experiments are therefore proposed to understand the environmental fate in addition to ecotoxicity (McLaughlin et al., 2002).

It is very unlikely that the ZnO-NP remain intact nanoparticles in the soil, but only advanced characterization techniques, such as single particle ICP-MS and FFF could confirm this (Hassellöv et al., 2008; Gimbert et al., 2007). And still there is a great risk that sample preparation may change the properties of the nanoparticles or their aggregates/agglomerates. As a consequence, it will be very difficult to obtain proper insight in nanoparticle characteristics in soils. Another approach for nanoparticle characterization is to look at the Zn composition using electron microscopy techniques. Manzo et al. (2010) studied ZnO-NP at 230 mg/kg soil (nominal concentration) using scanning electron microscopy. After examination of this soil sample, it was not possible to identify any particle, agglomerate or mass portion that could be ascribed to the NP (Manzo et al., 2010). ZnO-NP were also not visible in Lufa 2.2 soil spiked at 6400 mg Zn/kg using SEM (analysis executed by Saskia Kars, Laboratory for Microanalyses at the VU University). Characterization of ZnO-NP in soil was not further investigated and also outside the scope of the ecotoxicological assessment in this thesis. To study ZnO-NP dissolution and bioavailability in soils, its pore water was collected and the dissolved Zn concentrations were measured by flame AAS. Filtration over a 0.45 µm membrane has already been shown to be a suitable method for determination of the dissolved metal fraction in the pore water (Harmsen, 2007; Hobbelen et al., 2004) and it was also applicable for nanoparticle spiked soils. One of the uncertainties with this dissolved Zn fraction is the Zn speciation and composition. Centrifugal ultrafiltration has been applied on the pore water in order to separate nanoparticulate Zn from the dissolved fraction. Unfortunately, no major differences in Zn concentrations were measured before and after ultrafiltration and therefore, this did not reveal potential toxicity of particulate Zn (Chapters 3 and 5).

Considering soil type, I recommend to use a natural soil instead of the OECD artificial soil described by the ISO guideline (ISO, 1999). OECD artificial soil can be easily prepared in the laboratory from silica sand, kaolinite clay and sphagnum peat, but it is not always as ‘standard’ as it intends to be. Substantial differences in soil properties were found for 25 artificial soils prepared according to OECD standardized procedures at various ecotoxicological laboratories. The total organic carbon content, for example, ranged from 1.4 to 6.1% which was responsible for variations in phenanthrene sorption and could also have led to differences in toxicity (Bielská et al., 2012). The use of natural soils provides more realistic exposure situations than artificial soils. And nowadays many soil properties are available for standardized natural soils,

133

7


including the Lufa soils (LUFA-Speyer, Germany, 2009). Therefore, standard soil Lufa 2.2 (pHCaCl2 = 5.5, ca. 2% OM) was used in my first three experiments (Chapter 2, 3 and 4). In the final two experiments (Chapter 5 and 6) four other natural soils from Europe, ranging in soil pH and organic matter content, were used. In order to evaluate toxicity of ZnO-NP and relate this to the various soil properties, pHCaCl2, OM content and CEC for these field soils needed to be determined in our laboratory.

2 Toxicity of ZnO nanoparticles in relation to particle size

In 28-day toxicity tests using different soils a low hazard of (uncoated) ZnO-NP was found for F. candida. EC50 values above 1000 mg Zn/kg were estimated for two types of uncoated ZnO-NP (BASF® Z-COTE, Nanosun ZnO), but also for non-nano ZnO. In order to truly test a particle size effect, the non-nano material (a conventional material with a size above 100 nm) must be exactly the same as the nanoparticle in every aspect, except for the size. This is often impossible to achieve (Handy et al., 2012). The inclusion of the non-nano material in ecotoxicity tests with nanoparticles may reveal insight in potential differences in toxicity between the two. In theory, higher toxicities for smaller sized particles than for larger particles are expected due to the fact that smaller sized particles may have more potential reactive sites for interaction with biotic ligands. In soils, an effect of particle size has not been found on ZnO-NP toxicity to F. candida, as the EC50 values for ZnO-NP and non-nano ZnO were not significantly different in any of my experiments (Chapters 2, 3, 4 and 5). In both Lufa 2.2 and Dorset soils the EC50 values based on total Zn and porewater concentrations were in the same range. ZnO-NP toxicity in soil can apparently not be attributed to the size of the ZnO nanoparticles. It is very likely that ZnO nanoparticles in soil undergo various transformations, such as aggregation, sorption and dissolution. There are indications that nanoparticles, due to their high surface reactivity, interact with dissolved organic carbon (Li et al., 2010; Li et al., 2013). This could lead to a decreased surface area and neutralization of a particle size effect. Due to sorption and dissolution behaviour of the particles, size is no longer the most important factor determining ZnO-NP bioavailability and toxicity in soil.

3 Toxicity of ZnO nanoparticles in relation to coating

A clear adverse effect on springtail reproduction was discovered for coated ZnO-NP (Chapter 4). In freshly spiked Lufa 2.2 soil the toxicity of triethoxyoctylsilane coated ZnO-NP was significantly higher (EC50 = 873 mg Zn/kg d.w.) than that of uncoated ZnO-NP (EC50 = 1964 mg Zn/kg d.w.). Only after one year equilibration the EC50 value for coated ZnO-NP (1817 mg Zn/kg d.w.) was similar to the one for uncoated ZnO-NP in freshly spiked soil. The properties of a nanoparticle coating are important,

134

Chapter 7

because they control the stability and interactions of the particles with components of

the soil. Thus, it is appropriate to investigate the influence of surface coatings on fate

and toxicity. Microbial mediated nanoparticle coating degradation has been observed

and this resulted in nanoparticle aggregation (Kirschling et al., 2011). The dissolved

Zn fraction may determine to a large extent the bioavailable fraction of ZnO-NP. In

my study, the dissolved Zn fraction was lower for coated ZnO-NP than for uncoated

ZnO-NP, suggesting that a coating could prevent the release of Zn. This implicitly

implies that toxicity was not related to particle size or to the dissolved Zn fraction, but to the physico-chemical properties of these coated ZnO-NP. From the colour (greyish instead of brown) and deviating structure of the spiked soils, it is understandable that the hydrophobic character of the coated ZnO-NP may have hindered the animals to thrive in the soil. The surface charge of the coated and uncoated ZnO-NP was studied in deionised water by Geert Cornelis (unpublished). He measured a zeta-potential of approx. -30 mV at pH 5 for the coated ZnO-NP (graph not shown in this thesis), which suggests “moderate” stability of these nanoparticles. The zeta-potential for uncoated ZnO-NP was approx. 0 mV at pH 5, so rapid coagulation and flocculation are expected for the uncoated ZnO-NP in solutions with a pH around 5. The higher stability of coated ZnO-NP may lead to higher bioavailability and toxicity than for uncoated ZnO-NP that loose their nanoparticle properties in environmental media. Zeta-potential measurements in soil extracts could be useful to investigate the influence of surface charge and to further examine the influence of a surface coating on ZnO-NP toxicity. General conclusions on the effect of nanoparticle coating cannot be drawn from my tests with only one single type of coating. More studies on different types of coating are needed to unravel the behaviour and toxicity of coated ZnO-NP.

4 Toxicity of ZnO nanoparticles in relation to dissolved Zn

ZnO-NP are soluble particles which means that an effect of the released metal ions can occur. In most aquatic and terrestrial studies, the toxic effects for ZnO-NP were

ascribed to the dissolved Zn fraction released from the ZnO-NP (Adams et al., 2006; Aruoja et al., 2009; Blinova et al., 2010; Franklin et al., 2007; Kim et al., 2011). The

contribution of the dissolved fraction to ZnO-NP toxicity in soil is unknown. The toxicity of the soluble metal salt ZnCl2 was assessed for comparison with the

toxicity of ZnO-NP. If the effects for ZnO-NP were not only nanoparticle effects additional effects of the Zn ions, in theory, could be demonstrated by the soluble metal salt ZnCl2. This was not that simple for soils, because the dissolution and sorption processes of ZnO-NP and ZnCl2 were completely different and resulted in another outcome of the toxicity tests than expected. Based on total Zn concentrations much higher toxicity was

found for ZnCl2 than for ZnO-NP. In Lufa 2.2 (pHCaCl2 = 5.5) for example the 28-day EC50 was 1964 mg Zn/kg for ZnO-NP and 298 mg Zn/kg for ZnCl2 (Chapter 2 and

135

7


3) and in the Dorset soil at pHCaCl2 = 5.9 this was 1481 mg Zn/kg for ZnO-NP and 732 mg Zn/kg for ZnCl2 (Chapter 5). According to the porewater hypothesis, toxicity could be better explained on the basis of porewater concentration, which is assumed to be the bioavailable fraction (van Gestel, 1997). Negative effects on springtail reproduction could be explained by the porewater concentrations, because clear dose-response curves were estimated on the basis of porewater concentrations. After twelve months ageing for example, when the porewater concentrations peaked at intermediate soil concentrations, the reduction in reproduction was highest (46%) at 1027 mg Zn/kg, which contained the highest porewater concentration (i.e. 67.1 mg Zn/l). However, based on porewater concentrations, the toxicity of ZnO-NP was much higher than that of ZnCl2. So, instead of a similar dose-response curve for ZnO-NP and ZnCl2, the EC50 values for ZnO-NP were lower than the ones for ZnCl2 based on the dissolved Zn fraction. Figure 1 illustrates the general pattern I found in all my dose response curves obtained for ZnCl2 and ZnO-NP based on total soil and porewater concentrations. These graphs and corresponding EC50 values suggest that the effects of ZnO-NP on juvenile production were related to the soluble Zn concentrations in the pore water, but that also other (particle-related?) effects may have occurred. However, direct particle effects of ZnO-NP on the springtail reproduction could not be proven from my experimental data. Ultrafiltration of the porewater samples did not result in a strong evidence for an effect of nanoparticulate Zn in the pore water, as discussed in Chapters 2 and 5. Mathematically, the higher EC50 values for ZnCl2 could have been a result of the much higher Zn concentrations measured in the pore water from ZnCl2 spiked soils than from ZnO-NP spiked soils. In Lufa 2.2 the porewater concentrations for the highest soil concentrations were 12.6 and 612 mg Zn/l, and in Dorset soils these were 41.0 and 231 mg Zn/l for ZnO-NP and ZnCl2, respectively. But it is more likely that the bioavailability of Zn released from ZnO-NP differed from the soluble salt ZnCl2 and that other factors in the pore water played a role. In Dorset soil higher calcium concentrations in the soil pore water were measured that may have induced a protective effect in ZnCl2 spiked soils, leading to higher EC50 values (Chapter 5). The presence of other cations that compete for binding sites with Zn in the soils is a common phenomenon and does explain the lower toxicity of ZnCl2 based on porewater concentrations in the Dorset soils according to the principles of the terrestrial Biotic Ligand Model (Thakali et al., 2006). Also, ZnO-NP and ZnCl2 addition differently affected soil pH, with ZnO-NP leading to increased pH levels while ZnCl2 caused a dose-related pH decrease. Lower EC50 values based on porewater concentrations for ZnO-NP may have resulted from higher pH levels than for ZnCl2. According to the free metal ion model, competition with H+ ions in the pore water can also reduce metal toxicity (Lofts et al., 2004; 2013).

To summarize, a particle size effect of ZnO-NP toxicity was not detected for F. candida, but the contribution of a nanoparticle coating and dissolved Zn was responsible for the mild effects of ZnO-NP. In my opinion, the contribution of dissolved Zn to ZnO-NP toxicity in soil cannot be fully unravelled by a simple comparison with the toxicity of ZnCl2.

136

Chapter 7

Figure 1. General pattern for the effect of ZnO-NP and ZnCl2 on the reproduction (number of juveniles) of Folsomia candida after 28 days exposure observed in this thesis. Total Zn concentrations in the soil (left) and dissolved Zn concentrations in the pore water (middle and right) are provided on the x-axis. See text for explanation.

020406080

100120140160180200

0 5 10 15 200

20406080

100120140160180200

0 5 10 15 20 25 300

20406080

100120140160180

0 1000 2000 3000 4000 5000 6000

num

bero

f juv

enile

s

conc. soil (mg Zn/kg) conc. pore water (mg Zn/l)

hypothesis

XZnO-NPZnCl2

conc. pore water (mg Zn/l)

observedobserved

5 Zinc uptake and regulation

It is likely that springtails can be affected following three options: 1) ZnO-NP dissolve in the soil, then the released Zn2+ moves onto the animal via the ventral tube, is internalized and triggers various negative responses; or 2) ZnO-NP first bind to surface membranes of the animal, then dissolves and zinc ions are taken up; or 3) ZnO-NP are taken up as particles and dissolve in the body, and thus leads to more Zn2+ inside the gut and/or cells. Zn uptake by F. candida exposed to ZnO-NP, non-nano ZnO and ZnCl2 for 28 days was investigated after finishing the toxicity test described in Chapter 3. Internal Zn concentrations from five animals selected randomly from each exposure concentration were measured using graphite furnace AAS. Unfortunately, the Zn absorption in the blanks was very high and therefore the Zn concentration in the animals was too high. In earlier analysis internal concentrations of 44.5 µg Zn/g d.w. were measured for F. candida kept in Lufa 2.2 soil and 126 µg Zn/g d.w. for animals from the breeding culture (Smit et al., 1996). Although the absolute internal Zn concentrations were too high, and for this reason not reported in this thesis, the analyses showed that internal Zn concentrations were regulated for the lower exposure concentrations and increased at higher soil concentrations. They dramatically increased at exposure concentrations above the EC50 values for all three Zn forms. This indicated that the effects on springtail reproduction are related to internal Zn concentrations. Van Gestel and Hensbergen (1997) reported effects on reproduction at internal concentrations of about 97 µg Zn/g d.w. after exposing F. candida to ZnCl2 in artificial soil for 28 days. It remains uncertain whether the effects were induced by particulate Zn, dissolved Zn or Zn2+ ions. Exposure of earthworms to ZnO-NP and ZnCl2 resulted in a similar internal distribution of Zn analysed using synchrotron x-ray microspectroscopy (Unrine et al., 2008). Characterization of gut or other tissue could reveal further insight into Zn speciation within the springtails and potential uptake of particulate Zn. A stable

137

7


isotope labelling approach to trace the uptake of ZnO-NP may be promising (Fabrega

et al., 2011; Larner et al., 2012). Buffet et al. (2012) showed that Zn from ZnO-NP could

be distinguished from natural Zn background in estuarine invertebrates using labelled

ZnO-NP. Gene expression profiles could unravel potential differences in detoxification

mechanisms of Zn taken up from ZnO-NP and ZnCl2. This might also shed light on the

possible contribution of particulate Zn to the toxicity of ZnO-NP (Poynton et al., 2011).

6 Fate processes of ZnO nanoparticles in soil

A synthetic view of the environmental fate processes of ZnO-NP in soil and the interactions

in pore water are illustrated in Figure 2. The dissolution (release of Zn into the pore water)

and sorption (binding of Zn to the solid phase) of ZnO-NP in soil is a complex and dynamic

process. The hydrolysis of ZnO-NP into Zn hydroxides is a fast process and when Zn2+ at

the particle surface reacts with OH- to Zn(OH)+ (aq), the resulting hydroxide layer may

prevent Zn2+ release (David et al., 2012; Mudunkotuwa et al., 2012). It is already known

that slow reactions convert the adsorbed Zn to non-labile Zn in the soil (fixation) (Crout

et al., 2006; Voegelin et al., 2002). Ageing processes such as going from bioavailable Zn

to non-labile Zn can take months or years (Degryse et al., 2009). Sorption and dissolution

depend strongly on pH and the available binding sites on soil components such as organic

matter (Chapter 6). Dissolved Zn may form complexes with dissolved organic carbon,

which is affected by competing ions such as protons and calcium.

Sorption

The behaviour of ZnO-NP and ZnCl2 in soil differs regarding its sorption (and

dissolution), and therefore results in different porewater exposures. Sorption constants

(Kf) for ZnO-NP were higher than for ZnCl2 with the shape parameter n higher than 1

for ZnO-NP and lower than 1 for ZnCl2. In the lower concentration range, curves can

be nearly linear (n 1), but at higher metal concentrations the sorption curve levels

off, as sorption sites become saturated (Degryse et al., 2009). Therefore, Kf values can

never be evaluated without taking the n value into account. The n values found in

this research suggested saturation of Zn in the solid phase for ZnCl2 spiked soils and

saturation of Zn in the pore water of ZnO-NP spiked soils. This suggests precipitation

or aggregation in the pore water for ZnO-NP, which is rarely observed for metal salts.

Further characterization of the NP forms and surface charge in the pore water are

needed to unravel the speciation and bioavailability of ZnO-NP.

Dissolution

Fairly low dissolved Zn concentrations were measured for ZnO-NP aged soils,

indicating that during ageing the release of Zn to the pore water is slow. For a nominal

concentration of 800 mg Zn/kg, the dissolved Zn fractions increased from 0.253% in

138

Chapter 7

freshly spiked soils to 1.95%, 2.39% and 2.94% after three, six and twelve months ageing. For ZnCl2 aged soils, the dissolved Zn fractions were much higher and a maximum of 30.4% was calculated at the highest nominal Zn concentration. In ZnCl2 spiked soils, the dissolved fractions remained the same for each test concentration during the ageing period, which can be explained by the use of closed systems. In experimental field plots with leaching, the water extractable Zn fraction at the highest nominal concentration (3200 mg Zn/kg spiked as ZnCl2) declined from 36.8 to 1.5% of the total Zn concentration in three successive years (Smit et al., 1997). Also in the pH-amended soils (no ageing) low amounts of Zn were present in the pore water, indicating low bioavailability of ZnO-NP in soil. The maximum percentage of dissolved Zn was 0.317%, calculated for the soil with the lowest pH.

The non-linear increase in porewater Zn concentrations for ZnO-NP with total soil concentrations can be explained by pH. The decrease in porewater concentrations at higher total soil concentrations is most likely a cause of higher pH levels. After six months ageing, the soil pH increased from 5.1 for the lowest test concentration to 5.6 for soils spiked with 800 mg Zn/kg (giving the highest dissolved Zn concentration) and to 6.55 for the highest soil concentration (6400 mg Zn/kg). This suggests that

Figure 2. Synthetic view of the environmental fate processes of ZnO-NP in soil (dissolution, sorption, and fixation) and the interactions in the pore water. OM = organic matter; DOC = dissolved organic carbon.

Zn(OH)2 (aq) Zn(OH)+ (aq)Zn2+ (aq)

Znlabile

Sorptionslow

Dissolutionfast

Znfixed

OM, pH

Ageingslow

Soil Pore water

Zndissolved Zn2+

Zn-DOC

CompetitionCa2+, H+

OM, pH

ZnO-NP

pH

pH

pH

?

139

7


the release of Zn is limited by a solid-phase reaction, reaching a pH-dependent limit around a pH of 5.6. Below this pH level H+ ions will force the soluble Zn into solution, while at higher pH levels the Zn will adsorb to the solid phases, as was also confirmed by the decrease in Freundlich sorption constants with decreasing pH (Chapter 5). The slight decrease of pH with time, probably due to microbial activity, may explain the continuous release of Zn during one year ageing. As the dissolution of ZnO-NP in soil was strongly influenced by pH it was not possible to establish a reliable dissolution rate for ZnO-NP. In addition, maximum dissolved Zn concentrations were measured for intermediate spiking concentrations so a single average dissolution rate would clearly overestimate the dissolution rate at the highest exposure concentrations. Modelling porewater concentrations from total ZnO-NP concentrations is still a future exercise when more experimental data on ZnO-NP dissolution is available. The voltammetric technique AGNES seems to be a promising tool for actual measurements of the free Zn concentrations in soil solutions (David et al., 2012).

7 Toxicity of ZnO nanoparticles in relation to soil properties

The influence of soil properties on ZnO-NP toxicity to F. candida was described in Chapter 5 and 6. No effect of ZnO particle size was found on toxicity in pH amended field soils, but the effect on reproduction significantly decreased with increasing soil pH. EC50s of 553, 1481 and 3233 mg Zn/kg were estimated for uncoated ZnO-NP, at pH 4.5, 5.9 and 7.2, respectively. No correlation between EC50s and OM content was found in four different natural soils. Also, in these soils ZnO-NP toxicity was more related to soil pH, as EC50s increased with increasing pH from 1695 to 4446 mg Zn/kg at pHCaCl2 5.0 to 6.8. These studies showed that ZnO-NP dissolution and toxicity is more dependent on soil pH than on OM content. Dissolved organic carbon in the pore water may protect springtails from adverse effects of dissolved nanoparticles, but further research is needed on this aspect.

8 Hazard assessment

Bioavailability and toxicity of metals and metal oxides may change over time because of changes in soil conditions affecting metal availability. Ageing decreased the toxicity of ZnO-NP and ZnCl2 already after three months, shown by significantly higher EC50 values for both Zn forms (Chapter 4). Longer equilibration times would provide information on the further development of the dissolved Zn fractions in the soil, but in all probability this would not change the hazard assessment considering the low toxicity for springtails in one-year aged soils. After one year ageing, negative effects on springtail reproduction were no longer found for uncoated ZnO-NP at a measured

140

Chapter 7

concentration of 5855 mg Zn/kg. The EC50 for ZnCl2 increased to 707 mg Zn/kg

after one year ageing. In this study, ZnO-NP, non-nano ZnO and ZnCl2 spiked soils

were aged in closed systems in the laboratory. Percolating the test soils with water

after spiking (leaching) would have increased the environmental realism of the toxicity

tests (Smit and van Gestel, 1998). Leaching has been demonstrated to be effective

for metal salts, but the applicability of leaching for nano-sized metals has not been

proven and was not performed in this study. The effect of leaching of ZnO-NP spiked

soils could be investigated as a next step in long-term soil studies. Toxicity data could be used to translate EC50 values into a Predicted No Effect

Concentration (PNEC) for hazard assessment in soils. The PNEC can be defined as the maximum concentration which is tolerated by an organism without producing any adverse effects. PNEC values are estimated by division of the lowest value for the toxicity with the relevant assessment factor (REACH guidance, 2008). The size of the assessment factor depends on the type of data that are available i.e. short-term or long-term toxicity tests, the number of trophic levels tested and the general uncertainties in predicting ecosystem effects from laboratory data (Crane et al., 2008). For deriving a PNEC for the terrestrial environment a leaching-ageing correction factor is usually applied to make toxicity field-relevant. The term leaching-ageing factor (L/A factor) refers to the combined effect of leaching (due to changing ionic strength) and ageing (due to long-term reactions) on Zn bioavailability and toxicity in soil (Smolders et al., 2009). This L/A factor can be quantified as the ratio between the EC50 values before and after equilibration (van Gestel et al., 2012). In this research in which only ageing was studied for one year without leaching, a factor of 2.36 for ZnCl2 and 2.98 for uncoated ZnO-NP can be derived from the EC50 values estimated at T=0 and T=12. In the literature, an L/A factor of 3 is proposed for Zn based on CEC, background Zn and pH of different soils tested (Smolders et al., 2009). So, multiplying EC50 values derived from short-term soil toxicity tests with a factor 3 for leaching/ageing effects seems to be safe enough for coated and uncoated ZnO-NP.

In conclusion

Scientific research and ecotoxicological testing of ENPs must be performed for environmental risk assessment and legislative purposes. Potential adverse effects on

our environment cannot be excluded at this stage of nanotechnology development. With the rapidly increasing nanotechnology we must not wait for uncontrolled high

exposures in the soil environment that may affect the organisms living there. The behaviour of ZnO-NP in soils suggests that the environmental risk assessment could

be similar to non-nano ZnO, as the ZnO-NP lose their pristine properties in natural soils. Nanoparticle coatings should get more attention in future environmental

research. ZnO-NP dissolution in soils is a slow process and ZnO-NP are less toxic than

141

7


the soluble metal salt ZnCl2. Based on total Zn concentrations in the soil, the hazard assessment of ZnCl2 using a L/A factor of 3 would still be safe for ZnO-NP. Based on porewater concentrations Zn toxicity may be reduced by a protective effect of DOC and protons. Soil pH and organic matter content are important soil properties to take into account, but soil pH is the most important soil property affecting ZnO-NP dissolution and toxicity. Although it is possible that ZnO-NP may create toxic effects by releasing the toxic Zn2+ ions, there are currently no conclusive data or scenarios that indicate that these effects will become a major problem in natural soils or that they cannot be evaluated using the current risk assessment of Zn.

142

Chapter 7

Bibliography

Adams, L.K., Lyon, D.Y. and Alvarez, P.J.J., 2006. Comparative eco-toxicity of nanoscale TiO2, SiO2 and ZnO water suspensions. Water Research 40, 3527-3532.

Aitken, R.J., Chaudhry, M.Q., Boxall, A.B.A. and Hull, M., 2006. Manufacture and use of nanomaterials: current status in the UK and global trends. Occupational Medicine 56, 300-306.

Alloway, B.J., 1990. Heavy metals in soils. Blackie and Son Ltd, UK. ISBN 0-216-92698-X.Amutha, C. and Subramanian, P., 2009. Tissues damaging effect of zinc oxide nanoparticle on

Oreochromis mossambicus. Biochemical and Cellular Archives 9, 235-239. Apte, S.C., Rogers, N.L. and Batley, G.E., 2009. Ecotoxicology of manufactured nanoparticles. In:

Lead, Smith (Eds.). Blackwell publishing Ltd, 267-301.Aruoja, V., Dubourguier, H.C., Kasemets, K. and Kahru, A., 2009. Toxicity of nanoparticles of CuO,

ZnO and TiO2 to microalgae Pseudokirchneriella subcapitata. Science of the Total Environment 407, 1461-1468.

Auffan, M., Rose, J., Wiesner, M.R. and Bottero, J.-Y., 2009. Chemical stability of metallic nanoparticles: a parameter controlling their potential cellular toxicity in vitro. Environmental Pollution 157, 1127-1133.

Auffan, M., Rose, J., Bottero, J.-Y., Lowry, G.V., Jolivet, J.-P. and Wiesner, M.R., 2009. Towards a definition of inorganic nanoparticles from an environmental, health and safety perspective. Nature Nanotechnology 4, 634-641.

Baalousha, M., Mancuilea, A., Cumberland, S., Kendall, K. and Lead, J.R., 2008. Aggregations and surface properties of ions oxide NPs: influence of pH and natural organic matter. Environmental Toxicology and Chemistry 27, 1875-1882.

Bai, W., Zhang, Z., Tian, W., He, X., Ma, Y., Zhao, Y. and Chai, Z., 2010. Toxicity of zinc oxide nanoparticles to zebrafish embryo: a physicochemical study of toxicity mechanism. Journal of Nanoparticle Research 12, 1645-1654.

Bian, S.W., Mudunkotuwa, I.A., Rupasinghe, T. and Grassian, V.H., 2011. Aggregation and dissolution of 4 nm ZnO nanoparticles in aqueous environments: influence of pH, ionic strength, size, and adsorption of humic acid. Langmuir 27, 6059-6068.

Bielská, L., Hovorková, I., Komprdová, K. and Hofman, J., 2012. Variability of standard artificial soils: Physico-chemical properties and phenanthrene desorption measured by means of supercritical fluid extraction. Environmental Pollution 163, 1-7.

Bleeker, E.A.J., Cassee, F.R., Geertsma, R.E., de Jong, W.H., Heugens, E.H.W., Koers-Jacquemijns, M., van de Meent, D., Oomen, A.G., Popma, J., Rietveld, A.G. and Wijnhoven, S.W.P., 2012. Interpretation and implications of the European Commission Recommendation on the definition of nanomaterial, RIVM Letter report 601358001/2012. National Institute for Public Health and the Environment, Bilthoven, The Netherlands.

Blinova, I., Ivask, A., Heinlaan, M., Mortimer, M. and Kahru, A., 2010. Ecotoxicity of nanoparticles of CuO and ZnO in natural water. Environmental Pollution 158, 41-47.

Bodar, C.W.M., Pronk, M.E.J. and Sijm, T.H.M., 2005. The European Union risk assessment on zinc and zinc compounds: the process and the facts. Integrated Environment Assessment Management 1, 301-319.

Borm, P., Klaessig, F.C., Landry, T.D., Moudgil, B., Pauluhn, J., Thomas, K., Trottier, R. and Wood, S., 2006. Research strategies for safety evaluation of nanomaterials, Part V: Role of dissolution in biological fate and effects of nanoscale particles. Toxicological Sciences 90, 23-32.

Bottero, J.-Y., Auffan, M., Rose, J., Mouneyrac, C., Botta, C., Labille, J., Masion, A., Thill, A. and Chaneac, C., 2011. Manufactured metal and metal-oxide nanoparticles: Properties and perturbing mechanisms of their biological activity in ecosystems. Comptes Rendus Geoscience 343, 168-176.

Brümmer, G., Tiller, K.G., Herms, U. and Clayton, P.M., 1983. Adsorption-desorption and/or precipitation-dissolution processes of zinc in soils. Geoderma 31, 337-354.

145

Bibliography

Brulle, F., Cocquerelle, C., Wamalah, A.N., Morgan, A.J., Kille, P., Leprêtre, A. and van den Bulcke, F., 2008. cDNA cloning and expression analysis of Eisenia fetida (Annelida: Oligochaeta) phytochelatin synthase under cadmium exposure. Ecotoxicology and Environmental Safety 71, 47-55.

Buffet, P.-E., Amiard-Triquet, C., Dybowska, A., Risso-de Faverney, C., Guibbolini, M., Valsami-Jones, E. and Mouneyrac, C., 2012. Fate of isotopically labeled zinc oxide nanoparticles in sediment and effects on two endobenthic species, the clam Scrobicularia plana and the ragworm Hediste diversicolor. Ecotoxicology and Environmental Safety 84, 191-198.

Canas, J.E., Qi, B., Li, S., Maul, J.D., Cox, S.B., Das, S. and Green, M.J., 2011. Acute and reproductive toxicity of nano-sized metal oxides (ZnO and TiO2) to earthworms (Eisenia fetida). Journal of Environmental Monitoring 13, 3351-3357.

Cattaneo, A.G., Gornati, R., Chiriva-Internati, M. and Bernardini, G., 2009. Ecotoxicology of nanomaterials: the role of invertebrate testing. Invertebrate Survival Journal 6, 78-97.

Cayrol, C., Sarraute, J., Tarroux, R., Redoules, D., Charveron, M. and Gall, Y., 1999. A mineral sunscreen affords genomic protection against ultraviolet (UV) B and UVA radiation: In vitro and in situ assays. British Journal of Dermatology 141, 250-258.

Christian, P., von der Kammer, F., Baalousha, M. and Hofmann, Th., 2008. Nanoparticles: structure, properties, preparation and behaviour in environmental media. Ecotoxiology 17, 326-343.

Coleman, J.G., Johnson, D.R., Stanley, J.K., Bednar, A.J., Weiss Jr, C.A., Boyd, R.E. and Stevens, J.A., 2010. Assessing the fate and effects of nano aluminium oxide in the terrestrial earthworm, Eisenia fetida. Environmental Toxicology and Chemistry 29, 1575-1580.

Collins, D., Luxton, T., Kumar, N., Shah, S., Walker, V.K. and Shah, V., 2012. Assessing the impact of copper and zinc oxide nanoparticles on soil: a field study. PLoS ONE 7, 42663-42663.

Cornelis, G., Kirby, J.K., Beak, D., Chittleborough, D. and McLaughlin, M.J., 2010. A method for determination of retention of silver and cerium oxide manufactured nanoparticles in soils. Environmental Chemistry 7, 298-308.

Crane, M., Handy, R.D., Garrod, J. and Owen, R., 2008. Ecotoxicity test methods and environmental hazard assessment for engineered nanoparticles. Ecotoxicology 17, 421-437.

Crommentuijn, T., Doornekamp, A. and van Gestel, C.A.M., 1997. Bioavailability and ecological effects of cadmium on Folsomia candida (Willem) in an artificial soil substrate as influenced by pH and organic matter. Applied Soil Ecology 5, 261-271.

Croteau, M., Dybowska, A.D., Luoma, S.N. and Valsami-Jones, E., 2011. A novel approach reveals that zinc oxide nanoparticles are bioavailable and toxic after dietary exposures. Nanotoxicology 5, 79-90.

Crout, N.M.J., Tye, A.M., Zhang, H., McGrath, S.P. and Young, S.D., 2006. Kinetics of metal fixation in soils: measurements and modelling by isotopic dilution. Environmental Toxicology and Chemistry 25, 659-663.

Darlington, T.K., Neigh, A.M., Spencer, M.T., Nguyen, O.T. and Oldenburg, S.J., 2009. Nanoparticle characteristics affecting environmental fate and transport through soil. Environmental Toxicology and Chemistry 28, 1191-1199.

David, C.A., Galceran, J., Rey-Castro, C., Puy, J., Companys, E., Salvador, J., Monne, J., Wallace, R. and Vakourov, A., 2012. Dissolution kinetics and solubility of ZnO nanoparticles followed by AGNES. Journal of Physical Chemistry 116, 11758-11767.

Degryse, F., Smolders, E. and Parker, D.R., 2009. Partitioning of metals (Cd, Co, Cu, Ni, Pb, Zn) in soils: concepts, methodologies, prediction and applications – a review. European Journal of Soil Science 60, 590-612.

Depledge, M.H., 1989. Re-evaluation of metabolic requirements for copper and zinc in Decapod crustaceans. Marine Environmental Research 27, 115-126.

Depledge, M.H. and Rainbow, P.S., 1990. Models of regulation and accumulation of trace metals in marine invertebrates. Comparative Biochemistry and Physiology 97C, 1-7.

De Vries, W., Schülze, Lofts, S., Tipping, E., Meili, M., Römkens, P.F.A.M. and Groenenberg, J.E., 2005. Calculation of Critical Loads for Cadmium, Lead and Mercury; Background Document

146

Bibliography

to a Mapping Manual on Critical Loads of Cadmium, Lead and Mercury. Alterra-report 1104. Alterra, Wageningen, The Netherlands.

Dimkpa, C.O., Calder, A., Britt, D.W., McLean, J.E. and Anderson, A.J., 2011. Responses of a soil bacterium, Pseudomonas chlororaphis O6 to commercial metal oxide nanoparticles compared with responses to metal ions. Environmental Pollution 159, 1749-1756.

Dohrmann, R., 2006. Cation exchange capacity methodology II: A modified silver–thiourea method. Applied Clay Science 34, 38-46.

Domingos, R.F., Tufenkji, N. and Wilkinson, K.J., 2009. Aggregation of titanium dioxide nanoparticles: role of a fulvic acid. Environmental Science and Technology 43, 1282-1286.

Donner, E., Broos, K., Heemsbergen, D., Warne, M.St.J., McLaughlin, M.J., Hodson, M.E. and Nortcliff, S., 2010. Biological and chemical assessment of zinc ageing in field soils. Environmental Pollution 158, 339-345.

Drobne, D., Jemec, A. and Pipan Tkalec, Z., 2009. In vivo screening to determine hazards of nanoparticles: Nanosized TiO2. Environmental Pollution 157, 1157-1164.

Du, W., Sun, Y., Ji, R., Zhu, J., Wu, J. and Guo, H., 2011. TiO2 and ZnO nanoparticles negatively affect wheat growth and soil enzyme activities in agricultural soil. Journal of Environmental Monitoring 13, 822-828.

European Union, 2011. Commission Recommendation of 18 October 2011 on the definition of nanomaterial (2011/696/EU). O.J.L. 275, 38-40.

Fabrega, J., Tantra, R., Amer, A., Stolpe, B., Tomkins, J., Fry, T., Lead, J.R., Tyler, C.R. and Galloway, T.S., 2011. Sequestration of zinc from zinc oxide nanoparticles and life cycle effects in the sediment dweller amphipod Corophium volutator. Environmental Science and Technology 46, 1128-1135.

Fang, J., Shan, X.Q., Wen, B., Lin, J.M. and Owens, G., 2009. Stability of titania nanoparticles in soil suspensions and transport in saturated homogeneous soil columns. Environmental Pollution 157, 1101-1109.

Fountain, M.T. and Hopkin, S.P., 2001. Continuous monitoring of Folsomia candida (Insecta: Collembola) in a metal exposure experiment. Ecotoxicology and Environmental Safety 48, 275-286.

Fountain, M.T. and Hopkin, S.P., 2005. Folsomia candida (Collembola): a “standard” soil arthropod. Annual Review of Entomology 50, 201-222.

Franklin, N., Rogers, N., Apte, S., Batley, G., Gadd, G. and Casey, P., 2007. Comparative toxicity of nanoparticulate ZnO, bulk ZnO, and ZnCl2 to a freshwater microalgae (Pseudokirchneriella subcapitata): the importance of particle solubility. Environmental Science and Technology 41, 8484-8490.

Gamer, A.O., Leibold, E. and van Ravenzwaay, B., 2006. The in vitro absorption of microfine zinc oxide and titanium dioxide through porcine skin. Toxicology in Vitro 20, 301-307.

Gimbert, L.J., Hamon, R.E., Casey, P.S. and Worsfold, P.J., 2007. Partitioning and stability of engineered ZnO nanoparticles in soil suspension using flow field-flow fractionation. Environmental Chemistry 4, 8-10.

Gottschalk, F., Sonderer, T., Scholz, R.W. and Nowack, B., 2009. Modeled environmental concentrations of engineered nanomaterials (TiO2, ZnO, Ag, CNT, fullerenes) for different regions. Environmental Science and Technology 43, 9216-9222.

Gottschalk, F. and Nowack, B., 2011. The release of engineered nanomaterials to the environment. Journal of Environmental Monitoring 13, 1145-1155.

Greenslade, P. and Vaughan, G.T., 2003. A comparison of Collembola species for toxicity testing of Australian soils. Pedobiologia 47, 171-179.

Gulson et al., 2010. Small amounts of zinc from zinc oxide particles in sunscreens applied outdoors are absorbed through human skin. Toxicological Sciences 118, 140-149.

Gustafsson, J.P., 2007. A Windows Version of MINTEQA2 Version 4.0, MINTEQA2 was Released by the US Environmental Protection Agency in 1999. Visual MINTEQ Version 2.53.

147

Bibliography

Haanstra, L., Doelman, P. and Oude Voshaar, J.H., 1985. The use of sigmoidal dose response curves in soil ecotoxicological research. Plant and Soil 84, 293-297.

Hamilton, H.A., Russo, R.C. and Thurston, R.V., 1977. Trimmed Spearman–Karber method for estimating median lethal concentrations in toxicity bioassays. Environmental Science and Technology 11, 714-719. Correction. Environmental Science and Technology 12, 217.

Handy, R.D., Kammer, F.V.D., Lead, J.R., Hassellov, M., Owen, R. and Crane, M., 2008. The ecotoxicity and chemistry of manufactured nanoparticles. Ecotoxicology 17, 287-314.

Handy, R.D., Owen, R. and Valsami-Jones, E., 2008. The ecotoxicity of nanoparticles and nanomaterials: current status, knowledge gaps, challenges, and future needs. Ecotoxicology 17, 315-325.

Handy, R.D., van den Brink, N., Chappell, M., Mühling, M., Behra, R., Dusinská, M., Simpson, P., Ahtiainen, J., Jha, A.N., Seiter, J., Bednar, A., Kennedy, A., Fernandes, T.F. and Riediker, M., 2012. Practical considerations for conducting ecotoxicity test methods with manufactured nanomaterials: what have we learnt so far? Ecotoxicology 21, 933-972.

Hao, L. and Chen, L., 2012. Oxidative stress responses in different organs of carp (Cyprinus carpio) with exposure to ZnO nanoparticles. Ecotoxicology and Environmental Safety 80, 103-110.

Harmssen, J., 2007. Measuring bioavailability: From a scientific approach to standard methods. Journal of Environmental Quality 36, 1420-1428.

Hassellöv, M., Readman, J.W., Ranville, J.F. and Tiede, K., 2008. Nanoparticle analysis and characterization methodologies in environmental risk assessment of engineered nanoparticles. Ecotoxicology 17, 344-361.

Heckmann, L.-H., Hovgaard, M.B., Sutherland, D.S., Autrup, H., Besenbacher, F. and Scott-Fordsmand, J.J., 2011. Limit-test toxicity screening of selected inorganic nanoparticles to the earthworm Eisenia fetida. Ecotoxicology 20, 226-233.

Heggelund, L.R., Diez-Ortiz, M., Lofts, S., Lahive, E., Jurkschat, K., Cedergreen, N., Spurgeon, D. and Svendsen, C., 2013. Soil pH effects on the comparative toxicity of dissolved zinc, non-nano and nano ZnO to the earthworm Eisenia fetida. Nanotoxicology, in press.

Heinlaan, M., Ivask, A., Blinova, I., Dubourguier, H. and Kahru, A., 2008. Toxicity of nanosized and bulk ZnO, CuO and TiO2 to bacteria Vibrio fischeri and crustaceans Daphnia magna and Thamnocephalus platyurus. Chemosphere 71, 1308-1316.

Hobbelen, P.H.F., Koolhaas, J.E. and van Gestel, C.A.M., 2004. Risk assessment of heavy metal pollution for detritivores in floodplain soils in the Biesbosch, the Netherlands, taking bioavailability into account. Environmental Pollution 129, 409-419.

Hooper, H.L., Jurkschat, K., Morgan, A.J., Bailey, J., Lawlor, A.J., Spurgeon, D.J. and Svendsen, C., 2011. Comparative chronic toxicity of nanoparticulate and ionic zinc to the earthworm Eisenia veneta in a soil matrix. Environment International 37, 1111-1117.

Hu, C.W., Li, M., Cui, Y.B., Li, D.S., Chen, J. and Yang, L.Y., 2010. Toxicological effects of TiO2 and ZnO nanoparticles in soil on earthworm Eisenia fetida. Soil Biology and Biochemistry 42, 586-591.

Huang, Z., Zheng, X., Yan, D., Yin, G., Liao, X., Kang, Y., Yao, Y., Huang, D. and Hao, B., 2008. Toxicological effects of ZnO nanoparticles based on bacteria. Langmuir 24, 4140-4144.

Hyung, H., Fortner, J.D., Hughes, J.B. and Kim, J.-H., 2007. Natural organic matter stabilizes carbon nanotubes in the aqueous phase. Environmental Science and Technology 41, 179-184.

ISO, 1999. Soil Quality Inhibition of Reproduction of Collembola (Folsomia candida) by Soil Pollutants. ISO11267. International Standardization Organization, Geneva.

ISO, 2008. Nanotechnologies – Terminology and definitions for nano-objects – Nanoparticle, nanofibre and nanoplate. ISO/TS 27687. International Organisation for Standardisation, Geneva, Switzerland.

Janssen, R.P.T., Peijnenburg, W.J.G.M., Posthuma, L. and van den Hoop, M.A.G.T., 1997. Equilibrium partitioning of heavy metals in Dutch field soils. I. Releationship between metal partition coefficients and soil characteristics. Environmental Toxicology and Chemistry 16, 2470-2478.

Jemec, A., Drobne, D., Remškar, M., Sepčić, K. and Tišler, T., 2008. Effects of ingested nano-sized titanium dioxide on terrestrial isopods (Porcellio scaber). Environmental Toxicology and Chemistry 27, 1904-1914.

148

Bibliography

Jia, G., Wang, H., Yan, L., Wang, X., Pei, R., Yan, T., Zhao, Y. and Guo, X., 2005. Cytotoxicity of carbon nanomaterials: single-wall nanotube, multi-wall nanotube, and fullerene. Environmental Science and Technology 39, 1378-1383.

Jiang, J., Oberdöster, G. and Biswas, P., 2009. Characterization of size, surface charge, and agglomeration state of nanoparticle dispersions for toxicological studies. Journal of Nanoparticle Research 11, 11-89.

Johnston, B.D., Scown, T.M., Moger, J., Cumberland, S.A., Baalousha, M., Linge, K., van Aerle, R., Jarvis, K., Lead, J.R. and Tyler, C.R., 2010. Bioavailability of nanoscale metal oxides TiO2, CeO2 and ZnO to fish. Environmental Science and Technology 44, 1144-1151.

Ju-Nam, Y. and Lead, J.R., 2008. Manufactured nanoparticles: an overview of their chemistry, interactions and potential environmental implications. Science of the Total Environment 400, 396-414.

Kahru, A. and Dubourguier, H.-C., 2010. From ecotoxicology to nanoecotoxicology. Toxicology 269, 105-119.

Kamat, P.V., Huehn, R. and Nicolaescu, R., 2002. A “sense and shoot” approach for photocatalytic degradation of organic contaminants in water. Journal of Physical Chemistry 106, 788-794.

Keller, A.A., Wang, H., Zhou, D., Lenihan, H.S., Cherr, G., Cardinale, B.J., Miller, R. and Ji, Z., 2010. Stability and aggregation of metal oxide nanoparticles in natural aqueous matrices. Environmental Science and Technology 44, 1962-1967.

Kim, S., Kim, J. and Lee, I., 2011. Effects of Zn and ZnO nanoparticles and Zn2+ on soil enzyme activity and bioaccumulation of Zn in Cucumis sativus. Chemistry and Ecology 27, 49-55.

Kirschling, T.L., Golas, P.L., Unrine, J.M., Matyjaszewski, K., Gregory, K.B., Lowry, G.V. and Tilton, R.D., 2011. Microbial bioavailability of covalently bound polymer coatings on model engineered nanomaterials. Environmental Science and Technology 45, 5253-5259.

Klaine, S.J., Alvarez, P.J., Batley, G.E., Fernandes, T.F., Handy, R.D., Lyon, D.Y., Mahendra, S., McLaughlin, M.J. and Lead, J.R., 2008. Nanomaterials in the environment: behavior, fate, bioavailability and effects. Environmental Toxicology and Chemistry 27, 1825-1851.

Kosmulski, M., 2004. pH-dependent surface charging and points of zero charge II. Journal of Colloid and Interface Science 275, 214-224.

Laban, G., Nies, L.F., Turco, R.F., Bickham, J.W. and Sepúlveda, M.S., 2010. The effects of silver nanoparticles on fathead minnow (Pimephales promelas) embryos. Ecotoxicology 19, 185-195.

Lapied, E., Nahmani, J.Y., Moudilou, E., Chaurand, P., Labille, J., Rose, J., Exbrayat, J-M., Oughton, D.H. and Joner, E.J., 2011 Ecotoxicological effects of an aged TiO2 nanocomposite measured as apoptosis in the anecic earthworm Lumbricus terrestris after exposure through water, food and soil. Environment International 37, 1105-1110.

Lapied, E., Moudilou, E., Exbrayat, J.-M., Oughton, D.E. and Joner, E.J., 2010. Silver nanoparticle exposure cause apoptotic response in the earthworm Lumbricus terrestris (Oligochaeta). Nanomedicine 5, 975-984.

Larner, F., Dogra, Y., Dybowska, A., Fabrega, J., Stolpe, B., Bridgestock, L.J., Goodhead, R., Weiss, D.J., Moger, J., Lead, J.R., Valsami-Jones, E., Tyler, C.R., Galloway, T.S. and Rehkämper, M., 2012. Tracing bioavailability of ZnO nanoparticles using stable isotope labelling. Environmental Science and Technology 46, 12137-12145.

Lead, J.R. and Wilkinson, K.J., 2006. Aquatic colloids and nanoparticles: current knowledge and future trends. Environmental Chemistry 3, 159-171.

Li, M., Lin, D. and Zhu, L., 2013. Effects of water chemistry on the dissolution of ZnO nanoparticles and their toxicity to Escherichia coli. Environmental Pollution 173, 97-102.

Li, M., Pokhrel, S., Jin, X., Madler, L., Damoiseaux, R. and Hoek, E.M.V., 2011a. Stability, bioavailability, and bacterial toxicity of ZnO and iron-doped ZnO nanoparticles in aquatic media. Environmental Science and Technology 45, 755-761.

Li, L.-Z., Zhou, D.-M., Peijnenburg, W.J.G.M., van Gestel, C.A.M., Jin, S.-Y., Wang, Y.-J. and Wang, P., 2011b. Toxicity of zinc oxide nanoparticles in the earthworm, Eisenia fetida and subcellular fractionation of Zn. Environmental International 37, 1098-1104.

149

Bibliography

Li, M., Zhu, L. and Lin, D., 2010. Toxicity of ZnO nanoparticles to Escherichia coli: mechanism and the influence of medium components. Environmental Science and Technology 45, 1977-1983.

Lock, K. and Janssen, C.R., 2001a. Modelling zinc toxicity for terrestrial invertebrates. Environmental Toxicology and Chemistry 20, 1901-1908.

Lock, K. and Janssen, C.R., 2001b. Zinc and cadmium body burdens in terrestrial oligochaetes: use and significance in environmental risk assessment. Environmental Toxicology and Chemistry 20, 2067-2072.

Lock, K. and Janssen, C.R., 2002. The effect of ageing on the toxicity of zinc for the potworm Enchytraeus albidus. Environmental Pollution 116, 289-292.

Lock, K. and Janssen, C.R., 2003a. Influence of ageing on zinc bioavailability in soils. Environmental Pollution 126, 371-374.

Lofts, S., Spurgeon, D.J., Svendsen, C. and Tipping, E., 2004. Deriving soil critical limits for Cu, Zn, Cd and Pb: A method based on free ion concentrations. Environmental Science and Technology 38, 3623-3631.

Lofts, S., Criel, P., Janssen, C.R., Lock, K., McGrath, S.P., Oorts, K., Rooney, C.P., Smolders, E., Spurgeon, D.J., Svendsen, C., van Eeckhout, H. and Zhao, F.-Z., 2013. Modelling the effects of copper on soil organisms and processes using the free ion approach: Towards a multi-species toxicity model. Environmental Pollution 178, 244-253.

Lowry, G.V., Gregory, K.B., Apte, S.C. and Lead, J.R., 2012. Transformations of nanomaterials in the environment. Environmental Science and Technology 46, 6893-6899.

Lukkari, T., Aatsinkia, M., Väisänen, A. and Haimi, J., 2005. Toxicity of copper and zinc assessed with three different earthworm tests. Applied Soil Ecology 30, 133-146.

Ma, H., Bertsch, P.M., Glenn, T.C., Kabengi, N.J. and Williams, P.L., 2009. Toxicity of manufactured zinc oxide nanoparticles in the nematode Caenorhabditis elegans. Environmental Toxicology and Chemistry 28, 1324-1330.

Ma, W., Edelman, Th., van Beersum, I. and Jans, Th., 1983. Uptake of cadmium, zinc, lead and copper by earthworms near a zinc-smelting complex: influence of soil pH and organic matter. Bulletin of Environmental Contamination and Toxicology 30, 424-427.

Ma, H., Williams, P.L. and Diamond, S.A., 2013. Ecotoxicity of manufactured ZnO nanoparticles – A review. Environmental Pollution 172, 76-85.

Manzo, S., Rocco, A., Carotenuto, R., De Luca Picione, F., Lucia Miglietta, M., Rametta, G. and Di Francia, G., 2011. Investigation of ZnO nanoparticles’ ecotoxicological effects towards different soil organisms. Environmental Science and Pollution Research 18, 756-763.

Manzo, S., Miglietta, M.L., Rametta, G., Buono, S. and Di Francia, G., 2013. Toxic effects of ZnO nanoparticles towards marine algae Dunaliella tertiolecta. Science of the Total Environment 445-446, 371-376.

Maynard, A.D., 2006. Nanotechnology: a research strategy for addressing risk, PEN 03. Washington DC: Woodrow Wilson International Center for Scholars, Project on Emerging Nanotechnologies.

McLaughlin, M.J., 2002. Bioavailability of metals to terrestrial plants. In: Allen, H. (Ed.), Bioavailability of Metals in Terrestrial Ecosystems: Importance of Partitioning for Bioavailability to Invertebrates, Microbes and Plants. Society of Environmental Toxicology and Chemistry (SETAC), USA.

Miao, A.-J., Zhang, X.-Y., Luo, Z., Chen, C.-S., Chin, W.-C., Schantschi, P.H. and Quigg, A., 2010. Zinc oxide-engineered nanoparticles: dissolution and toxicity to marine phytoplankton. Environmental Toxicology and Chemistry 29, 2814-2822.

Milani, N., McLaughlin, M.J., Hettiaratchichi, G.M., Beak, D.G., Kirby, J.K. and Stacey, S. 2010. Fate of nanoparticulate zinc oxide fertilisers in soil: solubility, diffusion and solid phase speciation. Proceedings, 19th World Congress of Soil Science, Soil Solutions for a Changing World. Brisbane, Australia, August 1-6, 172-174.

Misra, S.K., Dybowska, A., Berhanu, D., Luoma, S. and Valsami-Jones, E., 2012. The complexity of nanoparticle dissolution and its importance in nanotoxicological studies. Science of the Total Environment 438, 225-232.

150

Bibliography

Mudunkotuwa, I.A., Rupasinghe, T., Wu, C.-M. and Grassian, V.H., 2012. Dissolution of ZnO nanoparticles at circumneutral pH: a study of size effects in the presence and absence of citric acid. Langmuir 28, 396-403.

Muyssen, B.T.A., De Schamphelaere, K.A.C. and Janssen, C.R., 2006. Mechanisms of chronic waterborne Zn toxicity in Daphnia magna. Aquatic Toxicology 77, 393-401.

Navarro, E., Baun, A., Behra, R., Hartmann, N.B., Filser, J., Miao, A.-J., Quigg, A., Santschi, P. and Sigg, L., 2008. Environmental behaviour and ecotoxicity of engineered nanoparticles to algae, plants and fungi. Ecotoxicology 17, 372-386.

Nel, A., Xia, T., Mädler, L. and Li, N., 2006. Toxic potential of materials at the nanolevel. Science 311, 622-627.

Nota, B., Verweij, R.A., Molenaar, D., Ylstra, B., van Straalen, N.M. and Roelofs, D., 2010. Gene expression analysis reveals a gene set discriminatory to different metals in soil. Toxicological Sciences 115, 34-40.

Nota, B., Vooijs, R., van Straalen, N.M. and Roelofs, D., 2011. Expression of mtc in Folsomia candida indicative of metal pollution in soil. Environmental Pollution 159, 1343-1347.

Nowack, B. and Bucheli, T.D., 2007. Occurrence, behaviour and effects of nanoparticles in the environment. Environmental Pollution 150, 5-22.

Nowack, B., Ranville, J.F., Diamond, S., Gallego-Urrea, J.A., Metcalfe, C., Rose, J., Horne, N., Koelmans, A.A. and Klaine, S.J., 2012. Potential scenarios for nanomaterial release and subsequent alteration in the environment. Environmental Toxicology and Chemistry 31, 50-59.

Oberdörster, G., Oberdörster, E. and Oberdörster, J., 2005. Nanotoxicology: An emerging discipline evolving from studies of ultrafine particles. Environmental Health Perspectives 113, 823-839.

OECD, 2009. OECD Guidelines for testing chemicals No. 232: Collembolan Reproduction Test in Soil. Organization for Economic Co-operation and Development, Paris.

Osmond, M.J. and McCall, M.J., 2010. Zinc oxide nanoparticles in modern sunscreens: An analysis of potential exposure and hazard. Nanotoxicology 4, 15-41.

Peralta-Videa, J.R., Zhao, L., Lopez-Moreno, M.L., de la Rosa, G., Hong, J. and Gardea-Torresdey, J.L., 2011. Nanomaterials and the environment: A review for the biennium 2008-2010. Journal of Hazardous Materials 186, 1-15.

Piccinno, F., Gottschalk, F., Seeger, S. and Nowack, B., 2012. Industrial production quantities and uses of ten engineered nanomaterials in Europe and the world. Journal of Nanoparticle Research 14, 1-11.

Pinnell, S.R., Fairhurst, D., Gillies, R., Mitchnick, M.A. and Kollias, N., 2000. Microfine zinc oxide is a superior sunscreen ingredient to microfine titanium dioxide. Dermatologic Surgery 26, 309-314.

Pipan-Tkalec, Z., Drobne, D., Jemec, A., Romih, T., Zidar, P. and Bele, M., 2010. Zinc bioaccumulation in a terrestrial invertebrate fed a diet treated with particulate ZnO or ZnCl2 solution. Toxicology 269, 198-203.

Poynton, H.C., Lazorchak, J.M., Impellittero, C.A., Smith, M.E., Rogers, K., Patra, M., Hammer, K.A., Allen, H.J. and Vulpe, C.D., 2011. Differential gene expression in Daphnia magna suggests distinct modes of action and bioavailability for ZnO nanoparticles and Zn ions. Environmental Science and Technology 45, 762-768.

Priester, J.H., Ge, Y., Mielke, R.E., Horst, A.M., Moritz, S.C., Espinosa, K., Gelb, J., Walkerg, S.L., Nisbet, R.M., Ani, Y.-J., Schimel, J.P., Palmer, R.G., Hernandez-Viezcas, J.A., Zhao, L., Gardea-Torresdey, J.L. and Holden, P.A., 2012. Soybean susceptibility to manufactured nanomaterials with evidence for food quality and soil fertility interruption. Proceedings of the National Academy of Sciences, early edition, 1-6.

Quik, J.T.K., Lynch, I., van Hoecke, K., Miermans, C.J.H., De Schamphelaere, K.A.C., Janssen, C.R., Dawson, K.A., Stuart, M.A.C. and van de Meent, D., 2010. Effect of natural organic matter on cerium dioxide nanoparticles settling in model fresh water. Chemosphere 81, 711-715.

Reach Guidance R.4., 2008. Guidance on information requirements and chemical safety assessment Chapter R.4: Evaluation of available information. European Chemicals Agency.

151

Bibliography

Reed, R.B., Ladner, D.A., Higgins, P., Westerhoff, P. and Ranville, J.F., 2012. Solubility of nano-zinc oxide in environmentally and biologically important matrices. Environmental Toxicology and Chemistry 31, 93-99.

Royal Society and Royal Academy of Engineering, 2004. Nanoscience and nanotechnologies: opportunities and uncertainties. London, UK: The Royal Society.

Sandifer, R.D. and Hopkin, S.P., 1996. Effect of pH on the toxicity of cadmium, copper, lead and zinc to Folsomia candida Willem, 1902 (Collembola) in a standard laboratory test system. Chemosphere 33, 2475-2486.

Scheckel, K.G., Luxton, T.P., El Badawy, A.M., Impellitteri, C.A. and Tolaymat, T.M., 2010. Synchrotron speciation of silver and zinc oxide nanoparticles aged in a kaolin suspension. Environmental Science and Technology 44, 1307-1312.

Scown, T.M., van Aerle, R., and Tyler, C.R., 2010. Review: Do engineered nanoparticles pose a significant threat to the aquatic environment? Critical Reviews in Toxicology 40, 653-670.

Shoults-Wilson, W.A., Zhurbich, O.I., McNear, D.H., Tsyusko, O.V., Bertsch, P.M. and Unrine, J.M., 2011. Evidence for avoidance of Ag nanoparticles by earthworms (Eisenia fetida). Ecotoxicology 20, 385-396.

Smit, C.E. and van Gestel, C.A.M., 1996. Comparison of the toxicity of zinc for the springtail Folsomia candida in artificially contaminated and polluted field soils. Applied Soil Ecology 3, 127-136.

Smit, C.E., van Beelen, P. and van Gestel, C.A.M., 1997. Development of zinc bioavailability and toxicity for the springtail Folsomia candida in an experimentally contaminated field plot. Environnmental Pollution 98, 73-80.

Smit, C.E., and van Gestel, C.A.M., 1998. Effects of soil type, pre-percolation, and ageing on bioaccumulation and toxicity of zinc for the springtail Folsomia candida. Environmental Toxicology and Chemistry 17, 1132-1141.

Smolders, E., Oorts, K., van Sprang, P., Schoeters, I., Janssen, C.R., McGrath, S.P. and McLaughlin, M.J., 2009. Toxicity of trace metals in soil as affected by soil type and ageing after contamination: using calibrated bioavailability models to set ecological soil standards. Environmental Toxicology and Chemistry 28, 1633-1642.

Sokal, R.R. and Rohlf, F.J., 1995. Biometry. Third Edition. W.H. Freeman, San Francisco, CA, USA.

Song, Z., Kelf, T.A., Sanchez, W.H., Roberts, M.S., Rička, J., Frenz, M. and Zvyagin, A.V., 2011a. Characterization of optical properties of ZnO nanoparticles for quantitative imaging of transdermal transport. Biomedical Optics Express 2, 3321-3333.

Song, L., Vijver, M.G., Peijnenburg, W.J.G.M. and De Snoo, G.R., 2011b. Smart nanotoxicity testing for biodiversity conservation. Environmental Science and Technology 45, 6229-6230.

Spurgeon, D.J. and Hopkin, S.P., 1996. Effects of variations of the organic matter content and pH of soils on the availability and toxicity of zinc to the earthworm Eisenia fetida. Pedobiologia 40, 80-96.

Spurgeon, D.J., Svendsen, C., Rimmer, V.R., Hopkin, S.P. and Weeks, J.M., 2000. Relative sensitivity of life-cycle and biomarker responses in four earthworm species exposed to zinc. Environmental Toxicology and Chemistry 19, 1800-1808.

Sterenborg, I., Vork, N.A., Verkade, S.K., van Gestel, C.A.M. and van Straalen, N.M., 2003. Dietary zinc reduces uptake but not metallothionein binding and elimination of cadmium in the springtail, Orchesella cincta. Environmental Toxicology and Chemistry 22, 1167-1171.

Stone, V., Nowack, B., Baun, A., van den Brink, N., von der Kammer, F., Dusinska, M., Handy, R., Hankin, S., Hassellöv, M., Joner, E. and Fernandes, T.F., 2010. Nanomaterials for environmental studies: Classification, reference material issues, and strategies for physico-chemical characterisation. Science of the Total Environment 408, 1745-1754.

Tapiero, H. and Tew, K.D., 2003. Trace elements in human physiology and pathology: zinc and metallothioneins. Biomedicine and Pharmacotherapy 57, 399-411.

Thakali, S., Allen, H.E., Di Toro, D.M., Ponizovsky, A.A., Rooney, C.P., Zhao, F.J. and McGrath, S.P., 2006. A Terrestrial Biotic Ligand Model. 1. Development and application to Cu and Ni toxicities to barley root elongation in soils. Environmental Science and Technology 40, 7085-7093.

152

Bibliography

Theng, B.K.G. and Yuan, G., 2008. Nanoparticles in the soil environment. Elements 4, 395-399.

Tiede, K., Hassellöv, M., Breitbarth, E., Chaudhry, Q. and Boxall, A.B.A., 2009. Considerations for environmental fate and ecotoxicity testing to support environmental risk assessments for engineered nanoparticles. Journal of Chromatography A 1216, 503-509.

Tipping, E., Rieuwerts, J., Pan, G., Ashmore, M.R., Lofts, S., Hill, M.T.R., Farago, M.E. and Thornton, I., 2003. The solid-solution partitioning of heavy metals (Cu, Zn, Cd, Pb) in upland soils of England and Wales. Environmental Pollution 125, 213-225.

Tourinho, P.S., van Gestel, C.A.M., Lofts, S., Svendsen, C., Soares, A.M.V.M. and Loureiro, S., 2012. Metal-based nanoparticles in soil: fate, behavior and effects on soil invertebrates. Environmental Toxicology and Chemistry 31, 1679-1692.

Truhaut, R., 1977. Ecotoxicology: objectives, principles and perspectives. Ecotoxicology and Environmental Safety 1, 151-173.

Unrine, J,M., Bertsch, P.M., Hunyadi, S.E., Ma, H.B., Newman, L.A. and Williams, P.L., 2008. Spatial distribution and speciation of Au and Zn in terrestrial organisms exposed to Au and ZnO nanoparticles. In: Proceedings of the 235th ACS National Meeting. New Orleans, LA.

Unrine, J.M., Hunyadi, S.E., Tsyusko, O.V., Rao, W., Shoults-Wilson, W.A. and Bertsch, P.M., 2010. Evidence for bioavailability of Au nanoparticles from soil and biodistribution within earthworms (Eisenia fetida). Environmental Science and Techology 44, 8308-8313.

van der Ploeg, M.J.C., Baveco, J.M., van der Hout, A., Bakker, R., Rietjens, I.M.C.M. and van den Brink, N.W., 2011. Effects of C60 nanoparticle exposure on earthworms (Lumbricus rubellus) and implications for population dynamics. Environmental Pollution 159, 198-203.

van Gestel, C.A.M., 1997. Scientific basis for extrapolating results from soil ecotoxicity tests to field conditions and the use of bioassays. Ecological Risk Assessment of Contaminants in Soil, 25-50.

van Gestel, C.A.M., Dirven-Van Breemen, E.M. and Baerselman, R., 1993. Accumulation and elimination of cadmium, chromium and zinc and effects on growth and reproduction in Eisenia andrei (Oligochaeta, Annelida). Science of the Total Environment (Supplement), 585-597.

van Gestel, C.A.M. and Hensbergen, P.J., 1997. Interaction of Cd and Zn toxicity for Folsomia candida Willem (Collembola: Isotomidae) in relation to bioavailability in soil. Environmental Toxicology and Chemistry 16, 1177-1186.

van Gestel, C.A.M., McGrath, S.P., Smolders, E., Diez Ortiz, M., Borgman, E., Verweij, R.A., Buekers, J. and Oorts, K., 2012. Effect of long-term equilibration on the toxicity of molybdenum to soil organisms. Environmental Pollution 162, 1-7.

van Straalen, N. and Verhoef, H.A., 1997. The development of a bioindicator system for soil acidity based on arthropod pH preferences. Journal of Applied Ecology 34, 217-232.

Verwey, E.J.W. and Overbeek, J.Th.G., 1948, Theory of the stability of lyophobic colloids, Amsterdam: Elsevier .

Vijver, M., Jager, T., Posthuma, L. and Peijnenburg, W., 2001. Impact of metal pools and soil properties on metal accumulation in Folsomia candida (Collembola). Environmental Toxicology and Chemistry 20, 712-720.

Vijver, M.G., Vink, J.P.M., Jager, T., Wolterbeek, H.Th., van Straalen, N.M. and van Gestel, C.A.M., 2005. Biphasic elimination and uptake kinetics of Zn and Cd in the earthworm Lumbricus rubellus exposed to contaminated floodplain soil. Soil Biology and Biochemistry 37, 1843-1851.

Voegelin, A., Scheinost, A.C., Bühlmann, K., Barmettler, K. and Krestzschmar, R., 2002. Slow formation and dissolution of Zn precipitates in soil: a combined column-transport and XAFS study. Environmental Science and Technology 36, 3749-3754.

Voegelin, A., Pfister, S., Scheinost, A.C., Marcus, M.A. and Krestzschmar, R., 2005. Changes in zinc speciation in field soil after contamination with zinc oxide. Environmental Science and Technology 39, 6616-6623.

Wang, Z.L., 2004. Zinc oxide nanostructures: growth, properties and applications. Journal of Physics: Condensed Matter 16, 829-858.

153

Bibliography

Wang, H., Wick, R.L. and Xing, B., 2009. Toxicity of nanoparticulate and bulk ZnO, Al2O3 and TiO2 to the nematode Caenorhabditis elegans. Environnmental Pollution 157, 1171-1177.

Wiench, K., Wohlleben, W., Hisgen, V., Raske, K., Salinas, E., Zok, S. and Landsiedel, R., 2009. Acute and chronic effects of nano- and non-nano-scale TiO2 and ZnO particles on mobility and reproduction of the freshwater invertebrate Daphnia magna. Chemosphere 76, 1356-1365.

Williams, R.J.P., 1984. Zinc: what is its role in biology? Endeavour, New Series 8, 65-70.Wolf, R., Matz, H., Orion, E. and Lipozencic, J., 2003. Sunscreens – the ultimate cosmetic. Acta

Dermatovenerologica Croatica 11, 158-162.Wong, S.W.Y., Leung, P.T.Y., Djurišić A.B. and Leung, K.M.Y., 2010. Toxicities of nano zinc oxide

to five marine organisms: influences of aggregate size and ion solubility. Analytical and Bioanalytical Chemistry 396, 609-618.

Yamabi, S. and Imai, H., 2002. Growth conditions for wurtzite zinc oxide films in aqueous solutions. Journal of Materials Chemistry 12, 3773-3778.

Yang, K., Lin, D. and Xing, B., 2009. Interactions of humic acid with nanosized inorganic oxides. Langmuir 25, 3571-3576.

Zeng, F., Ali, S., Zhang, H., Ouyang, Y., Qiu, B., Wu, F. and Zhang, G., 2011. The influence of pH and organic matter content in paddy soil on heavy metal availability and their uptake by rice plants. Environmental Pollution 159, 84-91.

Zhang, Y., Chen, Y., Westerhoff, P. and Crittenden, J., 2009. Impact of natural organic matter and divalent cations on the stability of aqueous nanoparticles. Water Research 43, 4249-4257.

Zhu, X.S., Zhu, L., Duan, Z.H., Qi, R.Q., Li, Y. and Lang, Y.P., 2008. Comparative toxicity of several metal oxide nanoparticle aqueous suspensions to zebrafish (Danio rerio) early developmental stage. Journal of Environmental Science and Health Part A - Toxic/Hazardous Substances & Environmental Engineering 43, 278-284.

154

Bibliography

Appendix I Summary of ecotoxicity studies with ZnO nanoparticles (ZnO-NP) and terrestrial invertebrates

Test organism Zn compound(s) Concentration Test media Exposure time Outcome Conclusion Reference

Caenorhabditis elegans

ZnO-NP (1.5 nm),ZnCl2 (Sigma-Aldrich)

Mortality: 325-1625 mg Zn/lReproduction: 10-200 mg Zn/l

Buffered K-medium

Mortality:24 hoursReproduction:72 hours

LC50 = 789 mg Zn/l for ZnO-NP and 884 mg Zn/l for ZnCl2EC50 = 46 mg Zn/l for ZnO-NP and 59 mg Zn/l for ZnCl2

Similar effects of ZnO-NP and ZnCl2 on mortality and reproduction. Effects probably caused by dissolved Zn

Ma et al.,2009


ZnO-NP (20 nm), non-nano ZnO (532 nm), ZnCl2

ZnO: 0.4-8.1 mg/lZnCl2: 0.7-13.6 mg/l

Water Mortality: 24 hoursReproduction: 5 days

LC50 = 2.2, 2.3 and 2.9 mg/l and NOEC (offspring per worm) = 0.4, 0.8 and 0.7 mg/l for ZnO-NP, non-nano ZnO and ZnCl2, resp.

ZnO induced effects cannot be explained by soluble Zn only. Indications for nanoparticle effects

Wang et al.,2009

Eisenia fetida ZnO-NP (40-100 nm) 0.1, 10, 100, 1000 mg/kg

Artificial soil

28 days Reproduction decreased with increasing exposure concentration and ceased at 1000 mg/kg

There was a significant dose-response between ZnO-NP concentration and cocoon production

Canas et al.,2011

Eisenia fetida ZnO-NP (10-20 nm) 100-5000 mg/kg Artificial soil

7 days DNA damage and decreased response of antioxidant system at 1000 mg/kg. Internal Zn concentrations increased up to 50 µg/g weight.*

Indications of ZnO-NP uptake, although gut contents could not be cleared thoroughly. Indications of oxidative stress

Hu et al.,2010

Eisenia fetida ZnO-NP (30 nm) 50, 100, 200, 500, 1000 mg/kg

Agar 96 hours Mortality increased with increasing ZnO concentration; LC50 = 232 mg ZnO/kg agar. Slight decrease of SOD activity. Zn was distributed in organelles and cytosol for 76% at 100 mg ZnO/kg agar exposure

ZnO dissolution could not explain mortality (which was 1.4 mg/l and below LC50 of 9.4 mg/l determined in solution), mainly due to ingestion

Li et al.,2011

Eisenia veneta ZnO-NP (<100 nm)ZnCl2 (Sigma-Aldrich)

250, 750 mg Zn/kg Clay loam soil

21 days No effect on mortality for both Zn forms. At 750 mg Zn/kg, reproduction declined by 50% for ZnO-NP and was almost zero for ZnCl2. Internal Zn concentrations reached almost 500 µg/g d.w. for both Zn forms

Higher toxicity for ZnCl2 than for ZnO-NP. Zn from ZnO-NP was internalized in a different and less active form. Inclusion of ZnO particles within the body wall tissues was shown

Hooper et al.,2011

Eisenia veneta ZnO-NP (<100 nm) ZnCl2 (Sigma-Aldrich)

6-96 mg Zn/l deionised water

24 hours LC50 = 1.75 mg Zn/l for ZnO-NP and 6.50 mg Zn/l for ZnCl2

Water-based exposure showed similar LC50s for both Zn forms

Hooper et al.,2011

Folsomia candida

ZnO-NP (<50 nm), ZnCl2 (Sigma-Aldrich)

230 mg Zn/kg OECD artificial soil

28 days No effect on survival and reproduction Similar resultsfor both Zn compounds Manzo et al.,2011

Porcellio scaber

ZnO-NP (<100 nm), non-nano ZnO, ZnCl2

2000, 5000 mg/kg hazelnut tree leaves

28 days Bioaccumulation was higher at lower concentrations, but the bioaccumulation factor was below 1 for all Zn forms and test concentrations

Bioaccumulation was attributable to Zn dissolution rather than accumulation of ZnO particles

Pipan-Tkalecet al., 2010

* A concentration of approx. 5 µg/g weight was measured in control earthworms, which indicates that the test animals were probably Zn-deficient (see also Appendix II, van Gestel et al., 2010)

156

Appendix I

Appendix I Summary of ecotoxicity studies with ZnO nanoparticles (ZnO-NP) and terrestrial invertebrates

Test organism Zn compound(s) Concentration Test media Exposure time Outcome Conclusion Reference


ZnO-NP (1.5 nm),ZnCl2 (Sigma-Aldrich)

Mortality: 325-1625 mg Zn/lReproduction: 10-200 mg Zn/l

Buffered K-medium

Mortality:24 hoursReproduction:72 hours

LC50 = 789 mg Zn/l for ZnO-NP and 884 mg Zn/l for ZnCl2EC50 = 46 mg Zn/l for ZnO-NP and 59 mg Zn/l for ZnCl2

Similar effects of ZnO-NP and ZnCl2 on mortality and reproduction. Effects probably caused by dissolved Zn

Ma et al.,2009


ZnO-NP (20 nm), non-nano ZnO (532 nm), ZnCl2

ZnO: 0.4-8.1 mg/lZnCl2: 0.7-13.6 mg/l

Water Mortality: 24 hoursReproduction: 5 days

LC50 = 2.2, 2.3 and 2.9 mg/l and NOEC (offspring per worm) = 0.4, 0.8 and 0.7 mg/l for ZnO-NP, non-nano ZnO and ZnCl2, resp.

ZnO induced effects cannot be explained by soluble Zn only. Indications for nanoparticle effects

Wang et al.,2009

Eisenia fetida ZnO-NP (40-100 nm) 0.1, 10, 100, 1000 mg/kg

Artificial soil

28 days Reproduction decreased with increasing exposure concentration and ceased at 1000 mg/kg

There was a significant dose-response between ZnO-NP concentration and cocoon production

Canas et al.,2011

Eisenia fetida ZnO-NP (10-20 nm) 100-5000 mg/kg Artificial soil

7 days DNA damage and decreased response of antioxidant system at 1000 mg/kg. Internal Zn concentrations increased up to 50 µg/g weight.*

Indications of ZnO-NP uptake, although gut contents could not be cleared thoroughly. Indications of oxidative stress

Hu et al.,2010

Eisenia fetida ZnO-NP (30 nm) 50, 100, 200, 500, 1000 mg/kg

Agar 96 hours Mortality increased with increasing ZnO concentration; LC50 = 232 mg ZnO/kg agar. Slight decrease of SOD activity. Zn was distributed in organelles and cytosol for 76% at 100 mg ZnO/kg agar exposure

ZnO dissolution could not explain mortality (which was 1.4 mg/l and below LC50 of 9.4 mg/l determined in solution), mainly due to ingestion

Li et al.,2011

Eisenia veneta ZnO-NP (<100 nm)ZnCl2 (Sigma-Aldrich)

250, 750 mg Zn/kg Clay loam soil

21 days No effect on mortality for both Zn forms. At 750 mg Zn/kg, reproduction declined by 50% for ZnO-NP and was almost zero for ZnCl2. Internal Zn concentrations reached almost 500 µg/g d.w. for both Zn forms

Higher toxicity for ZnCl2 than for ZnO-NP. Zn from ZnO-NP was internalized in a different and less active form. Inclusion of ZnO particles within the body wall tissues was shown

Hooper et al.,2011

Eisenia veneta ZnO-NP (<100 nm) ZnCl2 (Sigma-Aldrich)

6-96 mg Zn/l deionised water

24 hours LC50 = 1.75 mg Zn/l for ZnO-NP and 6.50 mg Zn/l for ZnCl2

Water-based exposure showed similar LC50s for both Zn forms

Hooper et al.,2011

Folsomia candida

ZnO-NP (<50 nm), ZnCl2 (Sigma-Aldrich)

230 mg Zn/kg OECD artificial soil

28 days No effect on survival and reproduction Similar resultsfor both Zn compounds Manzo et al.,2011

Porcellio scaber

ZnO-NP (<100 nm), non-nano ZnO, ZnCl2

2000, 5000 mg/kg hazelnut tree leaves

28 days Bioaccumulation was higher at lower concentrations, but the bioaccumulation factor was below 1 for all Zn forms and test concentrations

Bioaccumulation was attributable to Zn dissolution rather than accumulation of ZnO particles

Pipan-Tkalecet al., 2010

* A concentration of approx. 5 µg/g weight was measured in control earthworms, which indicates that the test animals were probably Zn-deficient (see also Appendix II, van Gestel et al., 2010)

157

Summary of ecotoxicity studies with ZnO nanoparticles (ZnO-NP) and terrestrial invertebrates

Appendix II Letter to the Editor

Metal-based nanoparticles in soil: New research themes should not ignore old rules and theories. Comments on the paper by Hu et al. (2010) ‘Toxicological effects of TiO2 and ZnO nanoparticles in soil on earthworm Eisenia fetida.’ Soil Biology & Biochemistry 42, 586-591

Cornelis A.M. van Gestel, Pauline L. Kool and Maria Diez Ortiz

Soil Biology & Biochemistry 42, 2010, 1892-1893

158

Appendix II

During the last decade, the use and application of nanomaterials has shown an exponential growth, with nanoparticles increasingly being used in an increasing number of products (Aitken et al., 2006). This development also has raised concern on the emission to the environment and consequent potential ecotoxicological effects of these nanomaterials. Recent literature reviews have demonstrated the lack of knowledge in this field, with especially little data being available on effects on soil invertebrates (Handy et al., 2008; Klaine et al., 2008; Kahru and Dubourguier, 2010). These reviews also highlighted the specific properties of nanoparticles and the consequent requirements of proper tools for characterizing and quantifying their exposure concentrations in ecotoxicological tests. Emphasis is placed on determining properties like particle size distribution, surface area and charge density, and the possible need for alternatives for the traditional concentration-based (in mg kg-1) way of expressing exposure levels (Handy et al., 2008; Klaine et al., 2008).

In case of metal-based nanoparticles, like ZnO, TiO2, Ag and CeO, toxicity will at least partly be due to the release of free metal ions (Auffan et al., 2009), while effects may further be enhanced by the specific properties related to the small size and consequent high surface activity of the particles. Recently, some attempts have been made to determine toxicity and bioaccumulation of ZnO nanoparticles in soil invertebrates (Pipan-Tkalec et al., 2010; Hu et al., 2010), ZnO being one of the most commonly used types of metal-based nanoparticles. Hu et al. (2010) in this journal, presented data on the toxicity and bioaccumulation of Zn in the earthworm Eisenia fetida after 7 days exposure to ZnO nanoparticles in OECD artificial soil. Though interesting, this and other studies on metal-based nanoparticles (e.g. Jemec et al., 2008; Drobne et al., 2009; Pipan-Tkalec et al., 2010) trigger some concerns.

1. The attention for the specific properties related to the small particle size seems to dominate over the common practice of proper metal analysis and related quality control. Hu et al. (2010), for instance, do only briefly explain their method of Zn analysis in the earthworms and do not make any mentioning of quality control. It should be common practice to include the analysis of a certified reference material when determining metal concentrations in environmental samples and to report the outcome in the paper.

2. Exposure concentrations should be expressed in such a way that it is clear what their basis is. In case of the paper by Hu et al. (2010), for instance, nanoparticle (NP) concentrations are expressed on a g/kg soil basis. It remains unclear whether this really concerns g ZnO-NP/kg dry soil or g Zn/kg dry soil. Especially in case of metal-based nanoparticles, the latter expression would be preferred as it also would enable comparison with data obtained from studies with the bulk metal or with other salts of the same metal. This, by the way, does also require analytical verification of (metal) exposure concentrations in the soil.

3. Metal concentrations in organisms (or soil samples) expressed on a fresh or wet weight basis usually are difficult to interpret because fresh weight may be highly variable. In case of earthworms, fresh weight e.g. is dependent on soil moisture

159

Letter to the Editor. Metal-based nanoparticles in soil

content. Therefore, it would be much more consistent to express metal levels on a dry weight basis. In the paper of Hu et al. (2010), this problem may hamper proper interpretation of the earthworm Zn levels as shown in Fig. 1 of their paper. Nevertheless, it occurs that Zn levels in the earthworms are rather low in the control animals, with an estimated level of only 6.3 µg Zn/g fresh weight. Assuming a dry weight of the earthworms of 15%, a value not uncommon for E. fetida, this would correspond with a Zn level in the control earthworms of approx. 42 µg/g dry weight.

4. A proper interpretation of studies with metal-based nanoparticles may benefit from using existing knowledge on the metal physiology of the test organism. The level of 42 µg Zn/g dry weight derived from the study of Hu et al. (2010) is rather low and well below the level required for optimal functioning of the earthworms. Zn is an essential element, playing a role in numerous biological processes, especially due to its key role in stabilizing the structural domains of proteins (Tapiero and Tew, 2003). As a consequence, a minimum level of Zn is required. According to Depledge (1989), Zn level in crustaceans should be at least some 67.9 µg/g dry weight; considering the important and general role of zinc, this level also holds for other animals. The Zn concentrations in E. fetida reported by Hu et al. (2010) are well below this minimum level, suggesting their control worms were Zn deficient.

5. As an essential element, Zn concentrations in many organisms are strictly regulated to a more or less constant level, either in the whole body or at the level of specific tissues or organs, also at fluctuating environmental exposure levels (Depledge and Rainbow, 1990). This regulation seems to be enabled by an efficient Zn excretion mechanism (Depledge and Rainbow, 1990). Regulation of zinc levels in earthworms was best demonstrated for E. fetida and its conspecific E. andrei, with regulated levels in general ranging between 80 and 120 µg Zn/g dry weight (Van Gestel et al., 1993; Spurgeon et al., 2000; Lock and Janssen, 2001; Lukkari et al., 2005). Only at high exposure levels of Zn in the soil did the earthworms show increased body residue levels, often directly coinciding with toxic effects on survival and reproduction (Van Gestel et al., 1993; Spurgeon and Hopkin, 1996). In the study of Hu et al. (2010), the earthworms already at the lowest test concentration (100 mg Zn (or ZnO?)/kg soil) showed Zn body concentrations exceeding the levels reported to be toxic. At an estimated level of 32.2 µg Zn/g fresh weight, corresponding with approx. 215 µg Zn/g dry weight, the worms must already suffer from toxic effects. This seems to be confirmed by the reported stimulation or reduction of the activity of different enzymes in the earthworms. Unfortunately, Hu et al. (2010) also seem uncertain about their metal analysis, as they note that the increasing concentration in earthworms may be due to the fact that some soil particles may have been retained in the earthworm’ gut. Plotting their zinc body concentrations in the earthworms, read from Fig. 1 and expressed on a dry weight basis, against soil concentrations did not show a consistent trend, suggesting that the amount of soil remaining in the earthworm gut was not constant over the different exposure levels.

160

Appendix II

6. A final point of concern is test duration. Toxicity may develop with time, a phenomenon that may be explained from the kinetics of uptake and elimination of the test chemical. Contrary to cadmium, which only slowly accumulates, for Zn uptake in earthworms seems pretty fast with equilibrium reached within a few days (Vijver et al., 2005). Some studies, however, also show an initial overshoot of Zn uptake followed by a decline to the regulated level (see e.g. Lock and Janssen, 2001). So, for a proper interpretation of the effects of metals and related nanoparticles, a test duration of 7 days seems rather short. If also effects of the dissolution of metal-based nanoparticles have to be taken into account, a much longer test period will be required. Such longer test periods may also be necessary to enable proper interpretation of the biological meaning of biochemical responses. In the study of Hu et al. (2010), it remains e.g. uncertain how the responses of enzyme activities, which often don’t show a consistent trend with increasing exposure concentrations, do relate to ecologically more relevant endpoints like growth and reproduction.

To conclude, research on the potential effects metal-based nanoparticles is highly relevant but should not neglect rules and knowledge obtained on the ecotoxicity of different metal salts.

161

Letter to the Editor. Metal-based nanoparticles in soil

Samenvatting

Titel: Ecotoxicologische beoordeling van ZnO nanodeeltjes voor Folsomia candida

De ontwikkeling van nanotechnologieën en het gebruik van nanomaterialen

in onze maatschappij hebben de laatste jaren een sterke toename laten zien.

Voor consumenten, industrie en de medische sector zijn allerlei producten met nanomaterialen op de markt, waaronder elektronica en producten voor persoonlijke verzorging. Nanomaterialen (deeltjes met afmetingen in de orde van miljoenste millimeters) bezitten bijzondere chemische en fysische eigenschappen. Hiermee dragen nanomaterialen bij aan een scala aan productvernieuwingen. Nanomaterialen kunnen tijdens of na het gebruik van producten in het milieu terechtkomen, onder andere in het rioolwater. Dit kan ook leiden tot een ophoping van nanomaterialen in de bodem, wat negatieve effecten kan veroorzaken voor dieren die in de bodem leven, zoals regenwormen, pissebedden en springstaarten. Deze zouden kunnen sterven of in aantal kunnen afnemen na blootstelling aan deze contaminanten. Er zijn aanwijzingen dat nanodeeltjes schadelijk kunnen zijn voor water organismen zoals algen en watervlooien, maar over de risico’s van blootstelling en toxiciteit van nanomaterialen in de bodem is relatief weinig bekend.

In mijn proefschrift heb ik gekozen om de toxiciteit van zink oxide (ZnO) nanodeeltjes voor bodemdieren te onderzoeken, om drie redenen. Ten eerste worden ZnO nanodeeltjes tegenwoording veel gebruikt in zonnebrandcrèmes als UV beschermer en worden er veel verschillende ZnO nanodeeltjes geproduceerd. Ten tweede tonen recente studies aan dat ZnO nanodeeltjes deels oplosbaar zijn in water en daarom is er een verhoogd risico op schadelijke effecten van opgeloste deeltjes en vrije Zn ionen. De derde reden voor dit onderzoek is de beperkte hoeveelheid informatie over de ecotoxiciteit van ZnO nanodeeltjes voor bodemdieren. In al mijn experimenten heb ik de springstaart Folsomia candida gebruikt, omdat dit als een modelorganisme voor bodemecotoxiciteit wordt beschouwd. F. candida behoort tot de springstaarten (collembola), zes-potige, vleugelloze invertebraten, die hun naam danken aan een gevorkt apparaat aan het achterlijf waarmee ze in staat zijn grote sprongen te maken. F. candida is een witte springstaart van ongeveer 3 mm en heeft een ventrale tubus waarmee opname van metalen kan plaatsvinden vanuit het poriewater in de bodem.

Deze blootstellingsroute van metalen zou ook een rol kunnen spelen in de toxiciteit van opgeloste nanodeeltjes en daarom is een belangrijk doel mijn onderzoek om uit te vinden of mogelijke effecten op de voortplanting van de springstaarten veroorzaakt

worden door de ZnO nanodeeltjes zelf of door het opgeloste zink (Zn).

Vijf kernvragen werden gesteld (Hoofdstuk 1):1. In hoeverre kunnen de huidige protocollen voor toxiciteitstoetsen worden

toegepast voor het bepalen van de ecotoxiciteit van nanodeeltjes?

163

Samenvatting

2. Is de toxiciteit van ZnO nanodeeltjes gerelateerd aan de deeltjesgrootte of aan vrijgekomen Zn?

3. Hoe snel lossen ZnO nanodeeltjes op in de bodem en leidt dit tot afname van de toxiciteit van ZnO nanodeeltjes op lange termijn?

4. Wat is het effect van een coating om de ZnO nanodeeltjes op de toxiciteit voor F. candida?

5. Spelen bodemeigenschappen, zoals pH en organische stofgehalte, een rol in de oplosbaarheid en toxiciteit van ZnO nanodeeltjes?

Ik zal mijn bevindingen uit de verschillende experimenten kort uiteen zetten en op basis daarvan de mogelijke risico’s van ZnO nanodeeltjes voor springstaarten inschatten.

Toxiciteitstesten met F. candida worden meestal uitgevoerd volgens de ISO richtlijn 11267 (ISO, 1999). Deze methode richt zich op het bepalen van effecten op overleving en reproductie van de springstaarten na 28 dagen blootstelling en deze test is ook geschikt bevonden voor het bepalen van de toxiciteit van nanodeeltjes. De manier van doseren van de nanodeeltjes aan de grond staat nog wel ter discussie. Door hun reactieve oppervlak aggregeren nanodeeltjes in water wat het moeilijk maakt om een homogene suspensie te maken. Ik heb twee doseringsmethoden vergeleken en de homogeniteit van Zn in de grond gemeten door vijf willekeurige monsters te nemen uit grond behandeld met ZnO nanodeeltjes (Hoofdstuk 2). Aan de ene batch grond waren ZnO nanodeeltjes toegevoegd als droog poeder en aan de andere batch grond als een suspensie in een extract van dezelfde bodem. De monsters werden vergeleken op basis van het totaal Zn gehalte gemeten na destructie en analyse met atomaire absorptie-spectrometrie (AAS). Beide methoden hebben voor- en nadelen. Een nadeel van doseren met poeder is de electrostatische lading waardoor nanodeeltjes gemakkelijk worden weggeblazen tijdens het afwegen. Aan de andere kant is het mengen van droog poeder met droge grond gemakkelijk uitvoerbaar. Het doseren van de grond als een suspensie zou aggregatie van de nanodeeltjes kunnen voorkomen, omdat deze dan binden aan opgelost of gesuspendeerd organisch materiaal in het extract. Een nadeel is dat het oplossen van de nanodeeltjes reeds in de suspensie begint en dit de testresultaten kan beïnvloeden. Beide methodes lieten goede resultaten zien, met in alle monsters meer dan 85% van het toegevoegde Zn teruggemeten. En een variatie van minder dan 10% tussen de vijf replica’s wees op een redelijk homogene verdeling van de ZnO nanodeeltjes in de grond, voor beide doseringsmethoden.

Voor de tweede vraag is het effect op overleving en reproductie van F. candida onderzocht voor drie verschillende Zn vormen, namelijk ZnO nanodeeltjes (< 200 nm), ZnO deeltjes met afmetingen groter dan 100 nm (“niet-nano”) en het oplosbaar zout ZnCl2 (Hoofdstuk 3). In een 28 dagen studie met natuurlijke standaardgrond (Lufa 2.2) was het effect op reproductie dosis-gerelateerd. EC50 waarden (concentratie die 50% remming van de reproductie veroorzaakt) waren respectievelijk 1964, 1591 en 298 mg Zn/kg voor ZnO nanodeeltjes, niet-nano ZnO en ZnCl2. Er werden geen effecten op overleving van F. candida gevonden in grond behandeld met ZnO nanodeeltjes en

164

Samenvatting

niet-nano ZnO tot concentraties van 6400 mg Zn/kg. Mijn hypothese dat kleinere

deeltjes toxischer zouden zijn door hun grotere oppervlak per volume deeltje bleek

niet waar te zijn. De grootte van de ZnO deeltjes leverde geen significante verschillen

op in de EC50 waarden voor ZnO nanodeeltjes en niet-nano ZnO. EC50 waarden

gebaseerd op poriewaterconcentraties waren voor alle drie de Zn vormen in dezelfde

range (7,94-16,8 mg Zn/l) wat suggereert dat het vrijgekomen Zn verantwoordelijk is

voor de effecten op reproductie en niet de nanodeeltjes zelf.

In een langlopende studie is de oplosbaarheid en toxiciteit van dezelfde drie

Zn vormen en van met triethoxyoctylsilane gecoate ZnO nanodeeltjes onderzocht

(Hoofdstuk 4). Behandelde Lufa 2.2 gronden weren “verouderd” in glazen potten

in een klimaatkamer en na drie, zes en twaalf maanden werd een deel van de

grond gebruikt om de hoeveelheid beschikbaar Zn in het poriewater te meten en

een 28 dagen toxiciteitstest met F. candida uit te voeren. Zn concentraties in het

poriewater van grond behandeld met ZnO namen gestaag toe met de tijd en na drie

maanden en langer zagen we een piek in de poriewaterconcentraties bij middelhoge

bodemconcentraties. Na twaalf maanden was de hoogste poriewaterconcentratie bij

een bodemconcentratie van 800 mg Zn/kg 67,1 mg Zn/l. Dit is slechts 2,94% van de

totale hoeveelheid Zn in die bodem wat laat zien dat slechts een klein deel van de

totale hoeveelheid Zn in de bodem in het poriewater terechtkomt. Het is bekend dat

de oplosbaarheid van ZnO afhangt van de zuurgraad van de bodem en dat opgeloste

Zn vormen (“species”), zoals Zn(OH)2 (aq) en Zn(OH)+(aq), niet gemakkelijk desorberen

bij een wat hogere pH. Bij hogere pH vindt er Zn fixatie plaats in de vaste fase van de

bodem. De verhoging van de pH, die optrad bij toename van de totale Zn concentratie,

kan de daling in poriewater Zn concentraties bij hogere doseringen verklaren. Voor de

gecoate ZnO nanodeeltjes werd een dergelijke piek in poriewaterconcentraties ook

gevonden, maar pas na twaalf maanden veroudering. Een coating om de nanodeeltjes

kan dus het vrijkomen van Zn voor een bepaalde tijd tegenhouden. Na 28 dagen

blootstelling waren de gecoate ZnO nanodeeltjes significant toxischer voor F. candida

dan de niet-gecoate ZnO nanodeeltjes. De 28-dagen EC50 voor de toxiciteit van de

gecoate deeltjes was 873 mg Zn/kg en nam pas toe in twaalf maanden verouderde

grond, waarbij de waarde van de 28-dagen EC50 voor niet-gecoate ZnO nanodeeltjes

werd bereikt. Voor de andere drie Zn vormen nam de toxiciteit al drastisch af na drie

maanden veroudering. Na een jaar veroudering was de pH van de grond geleidelijk

afgenomen. De hierdoor opgetreden toename van de concentratie aan vrije H+ ionen

in het poriewater zou bijgedragen kunnen hebben aan een beschermend effect op de

toxiciteit van vrije Zn2+ ionen in het porie water. Volgens het “Biotic Ligand Model”

kan competitie tussen H+ ionen en andere cationen de toxiciteit van metaalionen

verminderen. In deze studie werd gezien dat niet alleen de tijd, maar ook de pH een

prominente rol speelt in de biobeschikbaarheid en toxiciteit van ZnO nanodeeltjes.

Het effect van bodemeigenschappen, zoals de zuurgraad en het organische

stofgehalte, op de toxiciteit van ZnO nanodeeltjes is onderzocht in verschillende

165

Samenvatting

natuurlijke gronden. Om het effect van de zuurgraad te bestuderen is de pHCaCl2 van een natuurlijke grond uit Dorset (Engeland) gesteld op drie niveaus, namelijk 4,5, 5,9 en 7,2 (Hoofdstuk 5). De oplosbaarheid van ZnO nanodeeltjes was het hoogst in de meest zure grond. De sorptie van Zn nam toe met toenemende pH wat resulteerde in toenemende Freundlich sorptieconstanten van 98,9 tot 333 l/kg (met corresponderende n waarden van 1,23 tot 0,794). De effecten op de reproductie van F. candida na 28 dagen blootstelling aan ZnO nanodeeltjes in deze gronden namen af met toenemende pH, met EC50 waarden van 553, 1481 and 3233 mg Zn/kg. EC50s gebaseerd op poriewater Zn concentraties namen ook toe met toenemende pH van 4,77 tot 18,5 mg Zn/l.

Organisch stof speelt een rol bij het stabiliseren van nanodeeltjes in suspensies waarbij organische stof in het algemeen aggregatie van nanodeeltjes vermindert. Wanneer organische stof reageert met ZnO kan dit voor een natuurlijke “coating” zorgen om de nanodeeltjes en dit zou de biobeschikbaarheid van nanodeeltjes kunnen verminderen. Vier natuurlijke gronden variërend in organische stofgehalte van 2,37 tot 14,7% en pH van 5,0 tot 6,8 werden behandeld met ZnO nanodeeltjes (Hoofdstuk 6). F. candida werd vervolgens blootgesteld gedurende 28 dagen in een toxiciteitstest. De hoogste oplosbaarheid werd gemeten in de meest organische grond (23 mg Zn/l), maar deze grond had ook de laagste pH. Er werd geen correlatie gevonden tussen de 28-d EC50s voor het effect van ZnO nanodeeltjes op reproductie van F. candida en het organische stofgehalte van de grond. De toxiciteit was meer gerelateerd aan de bodem pH en EC50s namen toe met toenemende pH van 1695 tot 4446 mg Zn/kg.

Concluderend kan worden gesteld dat de ZnO nanodeeltjes die ik heb getest minder toxisch waren dan vrij Zn (getest als ZnCl2), gebaseerd op totaal Zn concentraties in de bodem. In al mijn studies vond ik vergelijkbare EC50 waarden voor de giftigheid ZnO nanodeeltjes en niet-nano ZnO, wat aangeeft dat de deeltjesgrootte de toxiciteit voor F. candida niet beïnvloedt. Eenmaal in de grond verliezen ZnO nanodeeltjes hun karakteristieke eigenschappen. Door aggregatie, sorptie en oplossing is de nano-vorm van de deeltjes in de grond niet meer aanwezig zoals ze was in de poedervorm. Hoewel ZnO nanodeeltjes toxische effecten kunnen veroorzaken door het vrijkomen van Zn2+ ionen, zijn er geen bewijzen gevonden dat deze effecten een groot probleem worden in natuurlijke gronden. Zowel op korte als op lange termijn treden er minder negatieve effecten op voor (gecoate en niet-gecoate) ZnO nanodeeltjes dan voor vrij Zn, gebaseerd op totaal Zn in de bodem. Naar mijn mening zouden ZnO nanodeeltjes daarom geëvalueerd kunnen worden met de huidige risicobeoordeling van Zn. Wel zullen bodemeigenschappen mee genomen moeten worden, met name de pH. In zure grond zal het oplossen van ZnO nanodeeltjes gestimuleerd worden en kan de toxiciteit toenemen.

166

Samenvatting

Acknowledgements

After three and a half years my thesis is ready; NanoFATE continues a few months more. It has been a great experience working in an European project en collaborating with different universities and research partners. I have enjoyed working on this research topic and I would like to thank a few people that contributed to this thesis.

Nico, first of all I would like to thank you for being my promotor and for your interest in nanoparticle research. Thank you for your feedback and suggestions on experimental plans and scientific writings. During our meetings you were the “overseer” of my research and I want to thank you for that.

Kees, your supervision has been incredible! Thank you for the opportunity to learn about ecotoxicology and to learn together about nanoparticles. Thank you for your help with planning experiments and processing data. At TNO I already used EC50 values for risk assessment purposes, but now I know how much work is behind such one number. Working together on this project also included travelling abroad for training courses, congresses and NanoFATE meetings, which was a great opportunity to learn and to network. Thank you for being there for the big and small questions.

Rudo, thank you for your assistence in the laboratory, especially for your help with the AAS. Thank you also for being my paranymph.

Thanks Maria for being my paranymph and for your support. We started working with nanoparticles at the same time, and you had a lot of experience with soils. Thank you for guiding me the first weeks of my research at the VU. I have enjoyed working together with you in the laboratory, with springtails and also with bacteria. Thank you for your nice company at congresses and meetings. Also visiting and walking around with you in Oxford was great.

I would like to give thanks to four fantastic students, working with me during their master (Kim and Svenja), pre-master (Irene) and bachelor (Rebeca). Thank you girls for your input and contribution to this research. Thank you, Kim, for your patience and many trials with the Ag analyses. Svenja, your surgery did not withdraw you from mixing soils; thanks for your hard work. Irene, thank you for your work with CeO2 nanoparticles and very good literature review. Rebeca, your accuracy with the uptake and elimination study has been greatly appreciated.

Bi-weekly “ecotox” meetings were organized by Kees in which practical issues were discussed or scientific literature was read. I would like to thank the Ecotox Group for their input, suggestions and collaboration. Thanks Masoud and Chloris, for the discussions on BLM model, the metal speciation course in Wageningen and our trip to the YES-meeting in Krakow. Thanks for learning together.

167

Acknowledgements

I would also like to thank my roommates at the Animal Ecology Department. You made it a really nice workplace during these years of hard work, although we were sometimes walking in and out on different times. Thanks Yumi, Anna, Ben, Tjalf, Mohamed and Elaine!

I would like to express my thanks to Claus Svendsen and Dave Spurgeon (NERC - Centre for Ecology and Hydrology, UK) for their excellent coordination of the NanoFATE project. Thank you for bringing PhD students together. Paula, it has been a pleasure working with you and running similar experiments with the isopods and the springtails. Also having you around at the VU was a good time. Thank you NanoFATE members for your collaboration, in particular Steve Lofts (NERC, UK) for performing the speciation modelling and Kerstin Jurkschat (Oxford University, UK) for performing the transmission electron microscopy on the nanosun ZnO nanoparticles.

Thank you Saskia Kars (Laboratory for Microanalysis, VU) for the scanning electron microscopy of the BASF ZnO nanoparticles. Thanks for the time spend on characterizing these nanoparticles in soil. Thank you Rien Dekker (Functional Genomics, VU) for performing the transmission electron microscopy of the ZnO nanoparticles in the soil extracts and for attempting to prepare the springtails for internal NP characterization. That was not easy!

Thank you Joris Quik for assisting with the nanoparticle tracking analysis and for your kind invitation to discuss experimental approaches.

I am also thankful for the collaboration with dr. Merel van der Ploeg and Nico van den Brink of Wageningen University. Thank you for our discussions and sharing test protocols.

I would also like to add the members of the reading committee to this list of acknowledgements. Thank you Susana Loureiro (University of Aveiro, Portugal), Willie Peijnenburg (RIVM / University of Leiden), Bart Koelmans (University of Wageningen), Els Smit (RIVM) and Joris Koene (VU, Animal Ecology) for your participation in my graduation and your interest in my thesis.

Als laatste wil ik graag mijn familie en vriendinnen bedanken. Dank voor jullie interesse in mijn onderzoek en jullie betrokkenheid in mijn leven tijdens deze AIO-periode. Mam, jij bent één van de weinige die ook de inhoudelijke vragen stelde. Dat komt vast door je verleden als analist. Pap, bijna was je naar SETAC Milaan gekomen om mijn praatje te horen. Bedankt voor je meeleven de afgelopen jaren (van 1e jaars bioloogje tot nu). Marieke en Ingrid, dankbaar ben ik voor jullie zussen! Ondanks dat we kilometers bij elkaar vandaan wonen, voelt dat helemaal niet zo. Oma’s wat mooi dat jullie er allebei bij zijn. We denken ook aan opa, die dit graag nog had willen meemaken en zelfs mijn eerste artikel via “Google translate” in het Nederlands heeft geprobeerd te lezen. Wat fijn om ook zulke lieve en enthousiaste schoonfamilie er bij te hebben gekregen. Tirza, dankjewel voor alle gezellige koffie/thee pauzes. Het samen fietsen is er (nog) niet van gekomen.

Bart, nanodeeltjes en springstaartjes… er zijn grotere dingen in het leven. Dankjewel voor al je support. Nanno = Leuk !

168

Acknowledgements

Curriculum vitae

I was born in Leiderdorp, The Netherlands, on September 27, 1983. I finished Atheneum College Hageveld (high school) in Heemstede in 2001. After one year at University College Utrecht I became enthusiastic for Biology and started the Bachelor Biology in Utrecht in 2002. In 2005 I continued with the Master “Toxicology and Environmental Health”, also at University Utrecht. This included a 9-month internship at Kiwa Water Research (Nieuwegein) on the implementation of the Ames II test. This test is used to screen water samples for genotoxic compounds and is still running today. A 6-month internship followed at the National Research Centre for Environmental Toxicology in Brisbane, Australia. Also here, my research was on genotoxicity, but in smoke collected from (bush)fires. In 2007 I graduated the master cum laude and became a junior chemical risk assessor at TNO (location Quality of Life) in Zeist. I worked there for two years on different projects dealing with food safety and environmental toxicity within the framework of REACH, the European Community Regulation on chemicals and their safe use. In 2009 I applied for a PhD position at the VU University and started to explore the toxicity of metal-based nanoparticles in soils. The studies performed during this PhD project were part of the EU-funded project NanoFATE (2010-2014).

169

Curriculum vitae

171

172

Documents

Ecotoxicological assessment of ZnO nanoparticles to Folsomia … dissertation.pdf · Project CP-FP 247739 (2010-2014) under the 7th Framework Programme of the European Commission