Review Hazards and Risks of Engineered Nanoparticles for the Environment and Human Health

The objectives of this article are to: (1) investigate the current state of knowledge of the risks of engineered nanoparticles for the environment and human health, (2) estimate whether this knowledge is sufficient to facilitate their comprehensive and effective risk assessment and (3) provide recommendations on future research in the field of risk assessment of nanomaterials. In order to meet the objectives, the relevance of each of the four steps of the risk assessment methodology ( i.e. , hazard identification, dose-response assessment, exposure assessment and risk characterization) was evaluated in the context of the current state of knowledge of the risks of nanomaterials, limitations were identified and recommendations were given on how to overcome them.


Background
In contrast to the small size of the nanoparticles, the scale of their application is tremendous.Nanotechnology influences virtually all industrial and public sectors, including healthcare, agriculture, transport, energy, materials, information and communication technologies.

OPEN ACCESS
Both the potential benefits and the risks, associated with the application of engineered nanoparticles (ENPs) have been widely debated in recent years.In contrast to the dominating optimistic projections that nanotechnology will bring significant technological development and well-being to society, it is considered that exposure to certain ENPs may cause environmental problems and/or do harm to human health.Since the early discussions about the risks of ENPs, the chemical risk assessment (CRA) has been put forward as the most relevant approach to understand, evaluate and quantify these risks.Currently, a variety of methodologies are being internationally discussed and evaluated with great vengeance with the idea that, in the near future, it will be possible to perform complete and scientifically sound risk assessment of ENPs.

Objectives
The objectives of this article are to: 1. Investigate the current state of knowledge of the risks of ENPs for the environment and human health.2. Estimate whether this knowledge is sufficient to facilitate comprehensive and effective risk assessment of ENPs. 3. Provide recommendations on future research in the field of risk assessment of ENPs.

Methodology
This article is based on an extensive review of literature published in the period: January 1992-September 2009.The selected literature consisted mainly of scientific publications, but also books, information from conferences and patent data were used.

Nanotechnology and Nanoparticles
-Nanotechnology‖ is a field of applied science and technology, dealing with the organization and control of matter on the nano-scale (i.e., between 1 and 100 nm) and the manufacturing of products and devices with dimensions, lying within this size range.A nanometer (nm), from the Greek -nanos‖ for -dwarf‖, equals one billionth of a meter.
-Nanomaterials‖ are all materials with sizes on the nano-scale in at least one of their dimensions [1], while -nanoparticles‖ are materials, nano-sized in at least two dimensions [2].The nomenclature -nanoparticles‖ encompasses particles as well as fibrous materials and tubes, but it excludes materials, such as coatings, films and multilayers.
Two types of nanoparticles (NPs) can be distinguished: (1) naturally occurring NPs (e.g., produced naturally in volcanoes, forest fires or as combustion by-products) and (2) engineered nanoparticles (ENPs), deliberately developed to be used in application (e.g., carbon black, fumed silica, titanium dioxide (TiO 2 ), iron oxide (FO x ), quantum dots (QDs), fullerenes, carbon nanotubes (CNTs), dendrimers).Naturally occurring NPs do NOT fall in the scope of this article.The paper encompasses only ENPs.The main reasons why materials built of ENPs have different optical, electrical, magnetic, chemical and mechanical properties from their bulk counterparts are that in this size-range quantum effects start to predominate and the surface-area-to-volume ratio (sa/vol) becomes very large [1].The sa/vol of most materials increases gradually as their particles become smaller, which results in increased adsorption of the surrounding atoms and changes their properties and behavior.Once particles become small enough, they start to obey the quantum mechanical laws.Materials reduced to the nano-scale can suddenly show very different properties, compared to what they exhibit on the macro-scale, which enables unique applications.For example, opaque substances become transparent (copper); stable materials become combustible (aluminum); inert materials become catalysts (platinum); insulators become conductors (silicon); solids turn into liquids at room temperature (gold) [3].

Defining "Hazard" and "Risk"
The term -hazard‖ has many definitions.This paper uses the definition of the United States Environmental Protection Agency (EPA) which defines ‗hazard' as the -inherent toxicity of a compound‖ [5].According to this definition, if a chemical substance has the property of being toxic, it is therefore hazardous.Any exposure to a hazardous substance may lead to adverse health effects in individuals or even death.
EPA defines -risk‖ with respect to the above definition of -hazard‖ as -a measure of the probability that damage to life, health, property, and/or the environment will occur as a result of a given hazard‖ [5].According to this definition, if the probability of an exposure to a hazardous material is high and the consequences for the health or environment are significant, then the risk is considered to be high.It is important to consider both the frequency of the event and the degree of the hazard to estimate risk [2].
Usually two categories of risk are distinguished in literature: -known risks‖ and -potential risks‖.When the relation between a cause and an effect is established, we talk of -known‖ risks.The responsibility for such risks can generally be attributed.When the causal relationship is established, prevention is possible.When the relationship between a cause and damage is not well known, we talk of -potential‖ risks.In case of potential risks, it is unclear whether there is a danger, how significant the damage can be or what is the probability of its occurrence [2, after 6].This situation is characterized by a state of suspicion (not awareness) and it is generally admitted that a precautionary approach can be applied in order to prevent potential damage [2, after 6].The risks of ENPs for the environment and human health fail in the second category: potential risks.It is very important to assess the risks of hazardous agents.The likelihood that a hazardous substance will cause harm (the risk) is the determinant of how cautious one should be and what preventative or precautionary measures should be taken.

Risk Assessment of ENPs
Since the early debates about the potential hazards of ENPs, the risk assessment of chemicals (CRA) has been put forward as the most relevant approach to understand and quantify the related risks [7].CRA is a process, in which scientific and regulatory principles are applied in a systematic fashion in order to describe the hazards, associated with the environmental and/or human exposure to chemical substances.It is defined as -a process, intended to calculate or estimate the risk to a given target organism, system or (sub)population, including the identification of attendant uncertainties, following exposure to a particular agent, taking into account the inherent characteristics of the agent of concern, as well as the characteristics of the specific target system‖ [8].The CRA is a four-step process, consisting of: (1) hazard identification, (2) dose-response assessment, (3) exposure assessment and (4) risk characterization.Its main outcome is a statement of the probability that when humans or other environmental receptors (e.g., plants, animals) are exposed to a chemical agent, they will be harmed and to what degree.
The CRA methodology is internationally recognized and employed by major actors, such as the World Health Organization (WHO) and the Organization for Economic Co-operation and Development (OECD), as well as by several European and U.S. agencies [9].It is considered a valuable tool, very important for the regulation of chemicals.CRA is also a fundamental ingredient of the new European Union (EU) chemical regulation policy, known as Registration, Evaluation and Authorization of Chemicals (REACH).
In order to achieve the objectives of this study, the current state of knowledge of the risks of ENPs for the environment and human health were summarized and evaluated in relation to each of the four elements of the CRA framework, as more important scientific findings were highlighted and limitations were identified and discussed.

Hazard Identification
-Hazard identification‖ (HI) is defined as the -…identification of the adverse effects, which a substance has an inherent capacity to cause‖ [10, after 11].Until recently, much of the discussion about the environmental and health risks of ENPs was considered to be rather speculative than realistic.In the last few years, however, a number of experimental studies found that exposure to certain ENPs can lead to adverse health effects in living organisms.In 2007, Hansen et al. identified 428 studies reporting on toxicity of ENPs [12].In these studies, adverse health effects of 965 tested ENPs of various chemical compositions were observed [12].

Current state of knowledge
The following sections shortly describe some of the most important scientific findings, relevant for HI of ENPs.Their purpose is to summarize the current state of knowledge of the hazards of ENPs, based on experimental studies.For simplification, the studies are divided into two categories-in vivo and in vitro studies.

Carbon nanotubes (CNTs)
A study, performed by Lam et al. [13], demonstrated that single-walled carbon nanotubes (SWCNTs) are able to cause dose-dependent effects of interstitial inflammation and lesions in mice and rats (0-0.5 mg• kg -1 for 7 to 90 days).Warheit et al. [14] observed pulmonary grandulomas in rats after exposure to SWCNT soot (1 and 5 mg• kg -1 for 24 hours to 3 months).In contrast to Lam et al. [13], however, the effects, observed by Warheit et al. [14] were not dependent on dose.Smith et al. [15] tested the ecotoxicity of SWCNTs, dissolved in sodium dodecyl sulphate (SDS) and sonication on juvenile rainbow trout (0.1, 0.25 and 0.5 mg• L -1 for 24 hours to 10 days) and they observed a dose-dependent rise in ventilation rate, gill pathologies (oedema, altered mucocytes, hyperplasia), and mucus secretion with SWCNT precipitation on the gill mucus.They also observed a significant dose-dependent decrease in thiobarbituric acid reactive substances (TBARS), especially in the gill, brain and liver, which is an indication of oxidative stress.
Multi-walled carbon nanotubes (MWCNTs) were shown by Carrero-Sanchez et al. [16], to exhibit acute toxicity in rats with LD 90 of 5 mg• kg -1 .Long MWCNTs were shown by Poland et al. [17] to cause significant inflammation and tissue damage in mice, while shorter MWCNTs caused less inflammation, which suggests that CNT toxicity is influenced by the particle morphology.In addition, they concluded that water-soluble components of MWCNT do not produce strong inflammatory effects in mice.

C 60 fullerenes
Most studies on the toxicological effects of C 60 fullerenes suggest that these materials tend to induce oxidative stress in living organisms [18][19][20][21].Lai et al. [18] observed a significant increase in lipid peroxidation (LP) products (a sign of oxidative stress) after intravenous administration of 1 mg kg -1 C 60 (OH) 18 in male mongrel dogs.Oberdörster [19,20] studied the effects of C 60 fullerenes in the brain of juvenile largemouth bass and observed high LP levels (0.5 and 1 ppm for 48 h).Elevated LP was also observed by Zhu et al. [21] in the brain and gills of daphnia magna after exposure to hydroxylated C 60 fullerenes (C 60 (OH) 24 ) and tetrahydrofuran (THF)-dissolved C 60 , as it was shown that THF did not contribute to the effect.Sayes et al. [22] detected an increase in the numbers of bronchoalveolar lavage (BAL)-recovered neutrophils (i.e., white blood cells) after intratracheal instillation of C 60 and C 60 (OH) 24 in rats, 1 day after the exposure.They also observed a significant increase in LP values 1 week after the exposure.Accute effects of functionalized C 60 were also reported.Zhu et al. [21] estimated LC 100 in fathead minnow after exposure to 0.5 ppm of THF-dissolved C 60 for 6-18 hours.Chen et al. [23] observed a LD 50 of 600 mg• kg -1 polyalkylsulfonated C 60 in female rats after intraperitoneal administration (0-2,500 mg• kg -1 for up to two weeks).Oberdörster [24] tested uncoated, water soluble, colloidal C 60 fullerenes and estimated a Daphnid 48-hour LC 50 of 800 ppb.

Metal and metal oxide ENPs
Li et al. [25] found that metal ENPs induce more severe lung toxicity in mice than bulk particles from the same materials.Gordon et al. [26] tested the effects on humans of exposure to zinc (Zn) ENPs.After 2 hours of exposure to 5 mg• m -3 of Zn ENPs, the exposed individuals started feeling sore throat, chest tightness, headache, fever and chills.Beckett et al. [27] repeated that test in three trials, 2 hours each, but at lower concentration (i.e., 500 μg• m -3 ), and found no indication of adverse effects.The latter two studies suggest that Zn ENPs toxicity is concentration-dependent and the most probable uptake path is through the respiratory system.A study of Sayes et al. [22] concluded that environmental exposure to Zn ENPs causes pulmonary (lung) inflammatory response in mice.Wang et al. [28] found that Zn ENPs can cause severe symptoms of lethargy, anorexia, vomiting, diarrhea, loss of body weight and even death in mice when gastrointestinally administered, whereas they observed limited effect for micro-scale Zn at equal concentrations.Yang and Watts [29] tested the effect of Aluminium (Al) ENPs on the relative root growth (RRG) in Zea mays (corn), Glycine max (soybean), Brassica oleracea (cabbage), and Daucus carota (carrot).The study found that the ENPs significantly inhibited the growth of the plants after administration of 2 mg• mL -1 for 24 h.
Oberdörster [30] and Oberdörster et al. [31] observed that smaller TiO 2 ENPs tend to cause more severe pulmonary damage in mice than larger particles.In addition, Warheit et al. [32] found that smaller silicon dioxide (SiO 2 ) particles cause stronger lung inflammation in rats than larger ones.Wang et al., [33] noticed that the smaller the TiO 2 particle size is, the greater the concentration in the liver of mice is.Bourrinet et al. [34] reported hypoactivity, ataxia, emesis, exophthalmos, salivation, lacrimation, discolored and mucoid feces, injected sclera, and yellow eyes in dogs after single-dose intravenous bolus administration of 20 and 200 mg• kg -1 FeO ENPs and a significant increase in fetal skeletal malformations in rats and rabbits.

In vitro studies Carbon nanotubes (CNTs)
A number of cytotoxicity studies with SWCNTs were reported in the literature.Shvedova et al. [35] observed oxidative stress and cellular toxicity in human epidermal keratinocytes, after 2 to 18 hours exposure to unrefined (iron containing) SWCNTs in concentrations, ranging from 0.6 to 0.24 mg• mL -1 .Cui et al. [36] observed dose-and time-dependent inhibition of cell proliferation and a decrease in cell adhesive ability in human embryo kidney cells after exposure to SWCNTs in concentrations between 0.8 and 200 μg• mL -1 .Sayes et al. [37] found that the surface functionalization of SWCNTs plays an important role in their cytotoxicity towards human dermal fibroblasts.Bottini et al. [38] noticed that MWCNTs were more cytotoxic when oxized towards Jurkat T leukemia cells, whereas Monteriro-Riviere et al. [39] observed a decrease of the viability of human osteoblastic lines and human epidermal keratinocytes after exposures to 0.1, 0.2, and 0.4 mg• mL -1 of MWCNTs for 1 to 48 hours.Kang et al. [40] compared the cytotoxicity of commercially obtained MWCNTs in bacterial systems before and after physicochemical modification and they observed highest toxicity when the nanotubes were uncapped, debundled, short, and dispersed in solution.Kang et al. [40] concluded that there is need for careful documentation of the physical and chemical characteristics of CNTs, when reporting their toxicity.

C 60 fullerenes
Adelman et al. [41] observed a reduction of the viability of bovine alveolar macrophages after exposure to sonicated C 60 and increased levels of cytokine mediators of inflammation (i.e., IL-6, IL-8 and TNF), while Porter et al. [42] found that C 60 and raw soot were not toxic towards bovine-and human alveolar macrophages.The reason behind the discrepancy between the results of Adelman et al.Studies on the effects of ENPs on alveolar macrophages are very important because the alveolar macrophages are the first line of cellular defense against respiratory pathogens [11, after 43].Yamawaki and Iwai [44] observed dose-dependent cytotoxicity of C 60 (OH) 24 (1-100 µg• mL -1 for 24 hours), resulting in decreased cell density and lactate dehydrogenase (LDH) release in human umbilical vein endothelial cells cavity (a sign of increase in non-viable cell numbers).Rouse et al. [45] observed a dose-dependent decrease in the viability of human epidermeal keratinocytes after exposure to C 60 -phenylalanine, as no contribution to the effect was attributed to the phenylalanine groups.
Quantum dots (QDs) The toxicity of QDs was found to be influenced by several factors: (1) composition, (2) size, (3) surface charge and (4) coating of the QDs [7,[46][47][48].Jaiswal et al. [46] found that CdSe/ZnS QDs (i.e., CdSe QDs in a zinc sulfide (ZnS) matrix), coated with dihydrolipoic acid (DHLA) had no effect on mammalian cells, while Hoshino et al. [47] reported adverse effects on mouse lymphocytes after exposure to CdSe/ZnS QDs, coated with albumin.In addition, Lovrí c et al. [48] observed that smaller (2.2 ± 0.1 nm), positively charged QDs exhibit stronger cytotoxicity than larger (5.2 ± 0.1 nm), equally charged QDs under the same conditions.It was also found that the cytotoxicity of QDs is influenced by the exposure to light and by temperature [49,50].Green and Howman [49] observed 56% damaged DNA after exposure to CdSe/ZnS together with UV light versus only 29% after exposure to CdSe/Zn in the absence of UV light.Chang et al. [50] found that CdSe/CdS (i.e., CdSe QDs in a cadmium sulfide (CdS) matrix) were toxic to cancer cells at 37 º C, but at 4 º C they were not toxic at all.

Metal and metal oxide ENPs
Sayes et al. [51] found that anatase TiO 2 ENPs are able to kill human dermal fibroblast (HDF) cells at LC 50 of 3.6 μg• mL -1 , while Wang et al. [52] observed decrease in the viability of human lymphoblastoid cells due to exposure to TiO 2 ENPs (0-130 μg• mL -1 for 6-48 h).Chen & Mikecz [53] found that SiO 2 ENPs do significantly inhibit replication and transcription in human epithelial HEp-2 cells (25 μg• mL -1 for 24 h).Muller et al. [54] observed that Fe 3 O 4 ENPs, coated with dextran, decrease the viability of human monocyte macrophages.Alt et al. [55] found that nano-particulate silver (Ag) is an effective bactericide against S. epidermidis, while Baker et al. [56] noticed that it effectively kills E. coli bacteria too.Sayes et al. [57] observed an increase in the production of LDH levels (an indicator of inflammation) in immortalized rat lung epithelial cells after 1 hour exposure to Zn ENPs at 520 μg• cm -2 .

Limitations to hazard identification of ENPs
It is very important to note that the vast majority of the reviewed studies demonstrate some degree of hazardous effects on the tested organisms.Toxicity has been reported for many ENPs, as shown in the previous sections, but for most of them further investigation and confirmation are needed before hazard can be identified.A lot of studies, relevant for HI, have been carried out with different ENPs, but most of them were obviously not meant to facilitate risk assessment; they use non-standardized tests, differing greatly from each other in regard to endpoints, tested species, methods of administration, dose ranges and exposure periods [7].The lack of standardized testing results in nonreproducible results and makes the univocal HI of ENPs impossible.
Another significant drawback for the HI of ENPs is the serious lack of characterization data, which makes it difficult to identify which physical and/or chemical characteristics (or combinations of characteristics) determine the hazards, documented in the (eco)toxicological studies [12,58,59].

Dose-Response Assessment
-Dose-response assessment‖ (DRA) is defined as -…an estimation of the relationship between dose, or level of exposure to a substance, and the incidence and severity of an effect‖ [10, after 11].It is the process of characterizing the relationship between the dose of an agent, administered to or received by an individual, and the consequent adverse health effects.

The concept of -dose‖
In toxicological studies a -dose‖ is the quantity of anything that may be received by or administered to an organism.The -dose‖ is normally measured in mass units (i.e., μg, mg, g), as higher doses of the same compounds are expected to cause more severe adverse effects.
DRA studies with ENPs, however, suggest that the toxicity of some ENPs is not mass-dependent, but influenced by other physico-chemical characteristics (e.g., surface area, chemical composition, particle morphology) [7, after 60].Oberdörster et al. [61] and Stoeger et al. [62,63] found that the toxicity of low-soluble ENPs was better described by their surface area than by their total mass [7, after 61-63].Wittmaack [64,65] suggested the number of particles as the most appropriate dose metrics, while Warheit et al. [66,67] found that toxicity of some ENPs was associated with the number of their surface functional groups.Despite these findings, however, it is still largely unknown which properties influence the toxicity of most ENPs and this gap in knowledge is partly attributable to the fact that the tested ENPs are seldom well characterized.

Characterization of ENPs
Developing understanding about the physical and chemical properties of substances and materials is fundamental for their risk assessment [59].Studying the standard properties (e.g., composition, structure, molecular weight, melting point, boiling point, vapor pressure, octanol-water partitioning coefficient, water solubility, activity, stability) is sufficient for the characterization of most chemical compounds.For ENPs, however, more profound investigation is needed and other properties, such as particle size distribution, sa/vol ratio, shape, electronic properties, surface characteristics, state of dispersion/agglomeration and conductivity need to be studied [5].The high complexity and great diversity of ENPs, however, make their characterization very difficult [59].
As it can be inferred from the Table, most of the current research on the properties of ENPs is focused on the identification of metrics and associated methods for the measurement of ENPs and their properties.This type of research is fundamental in the sense that without reliable measurement methodology it would be impossible to develop good understanding of the physical and chemical properties of the ENPs.Only few comprehensive studies on the development of standard, well-characterised reference nanomaterials were published so far.To facilitate the appropriate interpretation of testing results, it is essential to select representative sets of ENPs, characterize them and share them among laboratories worldwide.Table 2.A summary of studies on the metrology, characterisation and standardization of ENPs in the period 2004-2009 and their total funding value (modified after [68]).EA is a very important element in risk assessment of ENPs, since if no exposure to ENPs occur, it would be impossible that they cause any harm and there would be no risk at all.EA can be divided into three sub-areas: (1) occupational exposure assessment (OEA), (2) environmental exposure assessment (EEA) (including indirect human exposure from the environment) and (3) consumer exposure assessment (CEA).

Environmental exposure assessment
The environment may be exposed to ENPs during all stages of their life-cycles: raw material production, transport and storage, industrial use (incl.processing and/or trade), consumer use, waste disposal (incl.waste treatment, landfill and recovery) [11] (Figure 1).
A very important element of the EEA of ENPs is the study of their environmental fate.The fate of ENPs, released in the environment is determined by their mobility in the different media (i.e., soil, water, air), as well as by their potential to biodegrade or undergo chemical transformation.

Environmental fate of ENPs
In order to determine the extent of environmental exposure to ENPs, it is necessary to understand their behavior in the environment.Until now, only a limited number of environmental fate studies with ENPs have been reported and the fundamental mechanisms behind their distribution are still not clearly understood (Table 3).Table 3.A summary of studies on the source identification and environmental fate of ENPs, done between 2004 and 2009, and their total funding value (modified after [68]).

Fate of ENPs in air
The fate of ENPs in the air is determined by three main factors: (1) the duration of time particles remain airborne, (2) their interaction with other particles or molecules in the atmosphere and (3) the distance they are able to travel in the air [68].The processes important to understand the dynamics of ENPs in the atmosphere are diffusion, agglomeration, wet and dry deposition and gravitational settling [68].These processes are relatively well understood from studying the air-suspended ultrafine particles and that knowledge can be applied to ENPs as well [69].In some cases, however, there can be considerable differences in behavior between ENPs and ultrafine particles, especially when the latter cannot agglomerate because they are coated [5].
With respect to the duration of time ENPs stay in the air, it is considered that they may follow the laws of gaseous diffusion [70].The rate of diffusion is inversely proportional to the particle diameter and the rate of gravitational settling is proportional to it [70].It is generally considered that particles in the nanoscale (d < 100 nm) have shorter residence time in the air, compared to medium-sized particles (100 nm < d < 2,000 nm), because they rapidly agglomerate into much larger particles and settle on the ground [71].Here again ENPs with anti-agglomerate coatings make an exception and their residence time cannot be predicted [71].It is considered that deposited ENPs are usually not likely to be re-suspended or re-aerosolized in the atmosphere [70,72].
Many nano-sized particles are photoactive [72], but it is still unknown whether they are susceptible to photodegradation in the atmosphere.ENPs also show high absorption coefficients [69], and many of them can act as catalysts.However, no information is currently available on the interactions between ENPs and the chemicals they absorb, and how this interaction might influence atmospheric chemistry.

Fate of ENPs in water
The fate of ENPs in water is determined by several factors: (1) aqueous solubility, (2) reactivity of the ENPs with the chemical environment and (3) their interaction with certain biological processes [5].Because of their lower mass, ENPs generally settle more slowly to the bottom than larger particles of the same material [5].However, due to their high surface-area-to-mass ratios, ENPs readily sorb to soil and sediment particles and consequently are more liable to removal from the water column [73].Some ENPs might be subject to biotic and abiotic degradation, which can remove them from the water column as well.Abiotic degradation processes that may occur include hydrolysis and photocatalysis [72].Near to the surface ENPs are exposed to sunlight.It is likely that light-induced photoreactions can account for the removal of certain ENPs and for changing the chemical properties of others [72].
In contrast to the removal processes mentioned above, some insoluble ENPs can be stabilized in aquatic environments.Hoon et al. [74] investigated the aqueous stability of MWCNTs in the presence of natural organic matter (NOM).MWCNTs were readily dispersed as an aqueous suspension and remained stable for over 1 month.Hoon et al. [74] found that NOM is more effective in stabilizing the MWCNTs in water than a solution of 1% sodium dodecyl sulfate (SDS), a commonly used surfactant to stabilize CNTs in the aqueous phase [74].The C 60 fullerenes were found to spontaneously form insoluble, dense aqueous colloids of nanocrystalline aggregates and remain in the aqueous phase for long periods [5].Another known interaction, which can delay nanoparticle removal from the water column, is the absorption of humic acid.Sea surface microlayers, consisting of lipid-carbohydrate-and protein-rich components along with naturally occurring colloids, made up of humic acid, may attach ENPs to their surfaces and transport them over long distances [75].

Fate of ENPs in soil
The behavior of ENPs in soil media can greatly vary, depending on the physical and chemical characteristics of the material.Some ENPs can strongly sorb to the soil particles and become completely inert and immobile [5].On the other hand, if ENPs do not sorb to the soil matrix, they might show even greater mobility than larger particles, because their small size might allow them to travel easily through the pore spaces between the soil particles.The possibility to sorb to soil and the respective sorption strength of ENPs is influenced by their size, chemical composition and surface characteristics [5].
Studies by Zhang [76], Lecoanet and Wiesner [77] and Lecoanet et al. [78] showed considerable differences in mobility of some insoluble ENPs in porous media.The properties of the soil, such as porosity and grain size, further influence the mobility of the particles.Just like the mineral colloids, the mobility of ENPs, agglomerated in colloid-like structures might be strongly affected by electrical charge differences in soils and sediments [76].Surface photoreactions might induce photochemical transformations on the soil surface [72].

Biodegradation and chemical transformation of ENPs
In some cases, the biological processes in the environment can lead to the complete degradation of ENPs and sometimes they can only change their physical and/or chemical properties [5].The mechanisms, which account for the biodegradation of ENPs are still not fully understood.The potential for biodegradation is strongly dependent on the material properties.Most of the ENPs in current use are composed of not easily biodegradable materials, such as ceramics, metals and metal oxides [5].Despite this, however, a study of Filley et al. [79] found that C 60 and C 70 fullerenes can be completely metabolized by certain fungi species in medium time periods (12 days), which suggests that fullerene carbon is subject to biodegradation.In contrast, it was shown by Fortner et al. [80] that C 60 fullerenes tend to form stable colloidal structures in water, which suggests a level of resistance to biodegradation.Some known biodegradable materials are certain polymer ENPs, used in drug delivery systems [81].
Certain ENPs undergo chemical transformation when released in the environment.An example are the zero valent iron (Fe 0 ) ENPs, used in environmental remediation [76], which are oxidized to FeO in the reaction path.Some other metal ENPs are converted to oxides in air or water (e.g., Zn, Cu, Si), which can be more toxic than their corresponding free metals [5].Some types of QDs were shown to degrade due to photolytic and oxidative reactions, as the degradation of their coatings can reveal toxic metalloid cores [82].

Occupational exposure assessment
Workers may be exposed to nano-scale materials while manufacturing these materials, formulating them into products, transporting them or handling them in the storage facilities.Because higher concentrations of nano-scale materials and higher frequency of exposure to them are more likely to happen in workplace settings, occupational exposures require special attention.

General considerations
The primary route of exposure for workers, involved in manufacturing ENPs, is considered to be through inhalation and/or dermal contact after the manufacturing process is complete [7].Exposure is less likely to occur during the manufacturing process itself, since most ENP manufacturing processes are performed in closed reaction chambers [7].Contamination and exposure of workers are more likely to happen while handling and bagging the materials and also during cleaning operations [83].
In the product formulation phase, occupational exposures are most likely to occur while unloading the materials from shipping containers and cleaning the process equipment and vessels.During product manufacturing, exposures to ENPs are highly process-specific.For example, workers who manually apply spray coatings are often exposed to very high particle concentrations [5].In contrast, particles, bound in nanocomposites are not likely to release and handling of composites would result in lower occupational exposure levels.High exposures are likely to occur during product machining (i.e., cutting, drilling and grinding), repair, destruction and recycling [84].

Experimental results
A review by Aitken et al. [70], aimed to identify potential exposure scenarios, related to the manufacture and use of ENPs, studied the production processes of fullerenes, CNTs, metals and metal oxides.The review identified four main groups of ENP production processes: vapour deposition, gas-phase, colloidal and attrition processes [70].According to the report, all production processes can potentially result in occupational exposure through inhalation, dermal or ingestion routes [70].
Maynard et al. [85] performed exposure measurements of airborne SWCNTs in production facilities to assess the propensity for aerosol particles to be released during agitation and to measure the size of particles released into the air while SWCNT material was removed from production vessels and handled prior to processing.Airborne concentrations of SWCNT were then estimated to be lower than 53 μg• m -3 , while hand glove deposits of SWCNT during handling were estimated to be in the range 0.2-6 mg per glove.The study concluded that occupational exposures of SWCNTs are most likely to happen during handling and bagging of the materials and there is high risk of dermal uptake [85].
In a recent study, Han et al. [86] measured occupational exposures in the production cycle of MWCNTs.Air samples were taken and the MWCNTs in the samples were counted using a transmission electron microscope (TEM).The results yielded that most of the MWCNT exposure levels (max.0.43 mg• m -3 ) were lower than the current threshold limit value (TLV) for carbon black (i.e., 3 mg• m -3 ).
Yeganeh et al. [87] measured the concentrations of airborne ENPs, released during manufacturing of carbonaceous nanomaterials, such as fullerenes and CNTs, in a commercial production facility.The mass concentrations (PM 2.5), the submicrometer size distributions and the photoionization potential (i.e., an indicator of carbonaceous content) of the particles were measured at three locations: inside the fume hood where nanomaterials were produced, just outside the fume hood, and in the background.Average mass concentrations and particle number concentrations were not significantly different inside the facility versus outdoors [87].However, large, short-term increases in PM 2.5 and particle number concentrations were associated with the physical handling of nanomaterials.In many cases, an increase in the number of sub-100 nm particles accounted for the majority of the increase in total number concentrations [87].Photoionization results indicated that the particles suspended during handling, inside the fume hood, were carbonaceous and therefore likely to include ENPs, whereas those suspended by other activities, taking place outside the fume hood, were not.Based on the results of the study, the engineering controls at the facility were effective at limiting exposure to ENPs [87].
Fujitani et al. [88] compared the particle size distributions and morphology of aggregated/ agglomerated fullerenes at the production facilities of Frontier Carbon Corporation in Japan, during work and non-work periods.After this they compared the results to the nearby outdoor air.They found that the concentration of particles with diameters, shorter than 50 nm was not larger during the removal of fullerenes from a storage tank for bagging and weighing than prior to the activity [88].It should be noted, however, that this size fraction is extremely difficult to quantify and it is really impossible to do it accurately, in spite of instrument manufacturer's reports.
Bello et al. [89] investigated the airborne exposures generated in a research lab during the dry and wet cutting of nanocomposites, consisting of advanced fibers and polymer matrces, containing CNTs.No significant difference in air concentrations during wet cutting, which is the usual procedure for such composites, was identified.Dry cutting, however, generated statistically significant quantities of nanoscale and fine particles, regardless of the composite type (e.g., CNT-carbon, CNT-alumina or their respective base composites) [89].
Biswas and Wu [90] concluded that there is linear dependence between the active operations in production and the concentrations of ENPs in the working settings, while several other authors suggested that the influences of background concentration (concentration of other particles than those of concern) as well as the potential spacial and temporal variations of exposure are very important and have to be taken into consideration [91][92][93].Whereas the fraction of the total ultrafine particle number oncentrations generally decreases, fine particle number concentrations increases with time and distance from the point of emission [90].
Major limitations to the OEA are that official data on the number of workers exposed to ENPs are not available, the concentations of ENPs in the working settings are seldom properly measured and the occupational exposure pathways are still not well studied.[7,94].

Consumer exposure assessment
Today, nanotechnology is available on the market for great variety of applications and it is expected that widespread consumer exposure via direct contact with ENP-containing products will take place.
Since the spectre of the nano-products is very diverse, it is expected that the nature of consumer exposure will be disparate too.
Hansen et al. [95] divided ENP-containing products into several categories (i.e., appliances, food and beverages, health and fitness, home and garden and goods for children).They found that the expected consumer exposure is highest for products in the categories -appliances‖, -health and fitness‖ and -home and garden‖.The following figure describes the distribution of ENPs versus product categories, based on the study of Hansen et al. [95].The next figure compares between the probability of exposure and the types of ENPs used in the manufacturing of the products.As it can be inferred from the study of Hansen et al. [95], the category of -unclassifiable‖ products, for which no information on the used materials is available, is the one, containing the highest number of products, which consumers are expected to be exposed to [95].The lack of information about the ENPs, used in these products, is alarming since some of these materials might be potentially hazardous for their users.
The assessment of the consumer exposure to ENPs is significantly restricted by the lack of access to information about which commercially available products contain ENPs, the exact nanomaterial content of these products and the consumer behavior towards them [7].For many products the number of users is also unknown [96].The industry-derived data is obscured from public knowledge and this is to the detriment of all stakeholders (i.e., the governments, the public and the private sector).

Risk Characterization
Risk characterization (RC) is the final step of the risk assessment procedure.RC is defined as -… estimation of the incidence and severity of the adverse effects likely to occur in a human population or environmental compartment due to actual or predicted exposure to a substance, and may include risk estimation‖ [10, after 11].In this phase, all information, gathered during the first three steps of risk assessment is taken together, weighted and the risk is quantified.

Completed risk characterization studies with ENPs
The quantitative RC compares the predicted environmental concentration (PEC) of a chemical agent with its predicted no-effect concentration (PNEC).The PNEC is the concentration, below which the exposure to the substance is not expected to cause adverse any effects, while the PEC is the prognosticated concentration of a chemical in the environment.The PEC/PNEC ratio is called risk quotient (RQ).If the RQ is lower than 1, it is considered that no further testing or risk reduction measures are needed [11].If it is greater than 1, further testing can be initiated to lower the PEC/PNEC ratio [9].If that is not possible, risk reduction measures should be implemented [11].
In 2008, Müller and Nowack [97] reported the first fully quantitative environmental risk assessment of ENPs.They used nano-particulate Ag at threshold concentrations of 20 mg• L -1 and 40 mg• L -1 and exposed B. subtilis and E. coli bacteria to it.The results showed that, at the above concentrations, Ag ENPs did not affect the integrity of the microorganisms (i.e., both concentrations were equivalent to NOEC).In addition, Müller and Nowack [97] calculated the PNEC values of nano-particulate Ag, TiO 2 and CNTs in water, which were 0.04 mg• L -1 , <0.001 mg• L -1 and <0.0001 mg• L -1 , respectively.Combining these PNEC-values with the predicted exposure, they calculated the environmental concentrations of the above ENPs in Switzerland, stemming from different industries (i.e., textiles, cosmetics, coatings, plastics, sports gear, electronics).Assuming worst-case exposure scenarios levels, Müller and Nowack [97] found that the RQs for Ag ENPs and CNTs were lower than 0.001, and concluded that there was little or no risk that these materials would do harm to aquatic organisms.Exposure to TiO 2 , however, might possibly pose risks, since its RQs were ranging from 0.7 to 16.
Park et al. [98] assessed the risk of cerium oxide (CeO 2 ) to cause lung inflammation.First, they estimated an internal dose of 3.8 -7 cm -2 cm -2 by converting the retained dose into surface area units and then dividing by the area of the proximal alveolar region of the lung.Then they compared this value to the highest No Observed Effect Level (NOEL), found in in vitro toxicity studies (i.e., 26.75 cm -2 cm -2 ).Assuming that in vitro exposure data was reliable, Park et al. [98] concluded that it was highly unlikely that exposure to CeO 2 at the monitored and modeled environmental levels would elicit pulmonary inflammation.

Limitations to risk characterization of ENPs
Each of the elements of risk assessment holds certain limitations and challenges.RC, being at the end of the line, sums all of these limitations [7].Toxicity has been reported on for many ENPs, but for most further investigation and confirmation are needed before hazard can be identified [7].DRA assumes that no-effect concentrations (NECs) are established and although a number of studies observed dose-response relationships, they do not explicitly state any NEC values.DRA is severely hindered by the fact that it is still unclear what the most suitable dose-descriptors for most ENPs are.EA is hampered by difficulties in monitoring nanomaterial exposure in the workplace and the environment, and by deep uncertaincies in regard to the environmental fate and the biological pathways of ENPs.

Recommendations on Future Research
We argue that future research strategies must have a strong focus on the characterization of ENPs to enable the identification of clear causality between their inherent properties and the adverse effects they cause.For all types of ENPs, the most suitable dose-descriptors need to be determined (e.g., surface area, mass, morphology, chemical composition).Prior to achieving this, it would be relevant to report doses with respect to several characteristics, instead of choosing an irrelevant one.Given the large number of diverse ENPs, we recommend the use of a tiered testing approach, in which in vitro screening tests are designed to uncover particular properties that would then trigger more extensive evaluation.In addition, to facilitate the appropriate interpretation of testing results, standard reference materials, testing methods, and reporting formats must be elaborated.The development of reproducible and validated test standards and protocols would help to ensure that toxicological studies with ENPs generate comparable, standardized outputs, which would greatly aid their univocal HI.
EA of ENPs must build on a realistic environmental, occupational and consumer exposure scenarios (ESs).The ESs must be based on known or anticipated ENP production, trade/transport, use and disposal figures and coupled with a range of possible loss routes (e.g., accidents, leaks).With respect to EEA of ENPs it is necessary to establish the degree of environmental mobility of ENPs and their potential to bioaccumulate in order to identify whether ENPs can be taken up by living organisms and cause harm to them.In order to facilitate effective OEA of ENPs it is essential to obtain detailed data on the sources, dispersion mechanisms and concentrations of ENPs in the working settings and well study their occupational exposure pathways.To aqurately estimate ENP concentrations, adequate measurement and sampling methods and tools need to be developed.To characterize the consumer exposures to ENPs, it is necessary to identify the ENP-containing products with a high priority for future exposure studies.Furthermore, it is essential to collect reliable data on how many consumers use these products, including information about which products they use, how often and for how long.The needs for further research with respect to environmental and health risk assessment of ENPs are summarized in the following table.Furthermore, it would be useful to compile results and establish open access databases, which can serve the international scientific society and reduce the duplication of research efforts.

Hazard identification and dose-response assessment-related studies
In January 2008, the US National Institute of Standards and Technology (NIST) issued its first reference standards for ENPs (i.e., 10, 30 and 60 nm gold spheres) [99].These new reference materials are three of the five ENP samples, currently being used in an American Society for Testing and Materials (ASTM) interlaboratory performance-benchmarking study, intended to improve the consistency and quality of testing results across organisations.The comparative study began in January 2008, as NIST works in partnership with other national measurement institutes in several countries, as well as with the European Commission's Joint Research Centre Institute for Reference Materials and Measurements (IRMM) [100].
A team of materials scientists and toxicologists announced in 2008 the formation of the International Alliance for Nano Environmental, Health and Safety (EHS) Harmonization (IANH).IANH is a new international research network to establish protocols for reproducible in vivo and in vitro testing of ENPs [101].The goals of the alliance are to: 1. Create standard ENP toxicological testing protocols to use in a round-robin study and obtain identical toxicological results for ENPs.2. Implement a round-robin laboratory set of tests, based on the standard protocols.3. Facilitate further development of protocols, which take into consideration the properties of ENPs and their relationships to toixicity.
The alliance activities will potentially validate existing toxicity assessment strategies and thereby improve reproducibility and overall confidence in the reported results.
In 2009, the Institute of Occupational Medicine (IOM) launched the ENPRA, a project under the European 7th Framework Program.One of the main objectives of the project is to identify the critical ENP characteristics, responsible for the toxicity, observed in in vivo and in vitro studies [102].The € 3.7 million worth ENPRA project would combine the knowledge and capabilities of 15 European and six U.S. partners, including three U.S. Federal Agencies: EPA, the National Institute for Occupational Safety and Health (NIOSH) and the National Institute of Environmental Health Sciences (NIEHS) [102].
Two large-scale projects, which plan to develop and employ tiered approaches, using in vitro screening tests, are currently in progress (Table 5).The objective of the SUNANO project is to establish and validate toxicological test systems in order to perform an integrated hazard characterization of free nanoparticles of Ag, zirconium dioxide (ZrO 2 ), SiO 2 and TiO 2 , using a new tiered approach [103].The toxicity of the particles will be assessed as a function of size, shape and dose (i.e., mass and surface area).The NIRT project studies the mechanisms of interaction between carbon-based ENPs and cells, identifying potential adverse effects.The study uses in vitro and in vivo (rat) models and a tiered testing approach.6).The outputs of these studies may deliver important knowledge of the environmental pathways of ENPs, their biological and chemical transformation potentials as well as of their tendency to bioaccumulate.OEA of ENPs requires detailed information on the sources ENPs in occupational settings, their concentrations, the number of workers, exposed to the particles, and the pathways of exposure.A number of studies, currently in progress, are expected to deliver important knowledge of these aspects and thus contribute to the OEA of ENPs (Table 7).To identify consumer products with a high priority for future exposure studies, it is essential to know how many consumers use products, containing ENPs, which products they use, how often and for how long.Furthermore it is important to study the behavior of ENPs in consumer products and identidy potantial exposure pathways.The following table lists several ongoing studies, expected to fill some of the current data gaps.

Managing Uncertainty
As it was identified in the previous sections, a decent number of limitations and severe gaps in knowledge exist in regard to each of the four elements of CRA, when it is applied to ENPs.Considering these flaws, one can conclude that the risk assessment of ENPs is hampered solely by the lack of scientific knowledge.There is another viewpoint, however, it might be that conventional CRA methodology is simply inadequate to apply to ENPs and it should be substituted for an alternative methodology.
It would be naive to state that the conventional CRA should be entirely substituted for an alternative methodology due to the multiple and profound limitations.It is, however, also difficult to accept that it will be able to overcome these limitations in near future and effectively serve its purpose.It has been suggested by several authors that the CRA framework can be applied effectively to nanoparticles if it is aided by certain tools to reduce the deep uncertainties, which currently pervade every step of the procedure [7, after 104,105].Some approaches, which are considered relevant in this respect are the: (1) Multi Criteria Decision Analysis, (2) Weight-of-evidence and (3) Expert Elicitation.
The common purpose of Multi Criteria Decision Analysis (MCDA) methods is to evaluate and choose among different decision alternatives, based on multiple criteria, using systematic and structured analysis [7].MCDA methods have evolved as a response to the inability to effectively analyze multiple streams of conflicting information [106].There are different MCDA methods, based on different theoretical foundations, such as optimization, goal aspiration and outranking [106].MCDA tools could help in deciding what criteria to use to judge ENPs, to determine the relative importance of each of the criteria, score it, and finally compare the scores to identify the best alternative [106].MCDA can be used for toxicity and risk assessment of selected ENPs as well as for the development of regulatory criteria for them [107].
Another tool, widely used in risk assessment applications and considered applicable to the risk assessment of ENPs, is the Weight-of-evidence approach [105,107].Using the Weight-of-evidence methodology, assessors weigh various lines of evidence and apply professional judgment and/or calculations to decide where the weight of evidence lies-or, whether the various lines of evidence point to a potential risk or not [106].Weight-of-evidence evaluations may be either qualitative or quantitative, as the quantitative approach is often preferred because its results are considered more consistent and less subjective [108].
Expert Elicitation (EE) is the synthesis of opinions of experts on a subject where high uncertainty is present due to insufficient or conflicting data.EE is essentially a methodology, based on scientific consensus, used to quantify uncertainty [109,110].The subjective judgment of experts is usually represented as a -subjective‖ probability density function (PDF), as effort is made to minimize subjective judgment and the errors related to it in the elicited outcomes [109,110].

Conclusions and Recommendations
This chapter presents the main conclusions drawn from this study.In addition, suggestions and recommendations on further research are given and focal points are discussed.

Objective 1: Investigate the Current state of Knowledge of the Risks of ENPs for the Environment and Human Health
It is possible that ENPs can cause novel environmental problems; impose risks to human health or do both.It is impossible, however, at this point of time and stage of knowledge to make any collective judgment about the potential risks of exposure to nanomaterials.ENPs are expected to affect living organisms in different ways than their bulk alternatives and considering their significant diversity, it is anticipated that ENPs would also differ a lot from each other in terms of toxicity.
Most of the reviewed toxicity studies with ENPs demonstrate some degree of hazardous effects on the tested organisms.Some in vivo toxicological studies with CNTs suggest that these materials tend to cause interstitial inflammation and lesions in mammals [13,14].Shorter MWCNTs were shown to cause less severe inflammation than longer MWCNTs and dissolved MWCNTs caused almost no adverse response, which suggests that MWCNT toxicity is influenced rather by particle morphology than by chemical composition [17].This notion was supported by Kang et al. [40], who compared the cytotoxicity of commercially obtained MWCNTs in bacterial systems before and after physicochemical modification and concluded that uncapped, debundled, short, and dispersed in solution MWCNTs exhibit highest cytotoxicity.Most in vivo studies with C 60 fullerenes suggest that these materials tend to induce oxidative stress in living organisms [18][19][20][21].The toxicity of QDs was found to be influenced by their composition, size, surface charge and coating, as well as by the exposure to light and on temperature [46][47][48][49].Smaller QDs were shown to be more toxic than larger QDs [51], while the exposure to UV light and higher temperatures also tend to increase the tocicity of QDs.Exposure to Zn ENPs causes pulmonary (lung) inflammatory response in mammals [22,26], as the smaller the particle size is, the greater the adverse effect is [25].Exposure to FeO and TO 2 ENPs decreases the viability of human monocyte macrophages and human lymphoblastic cells, respectively [52,54].It was also shown that nano-Ag acts as an effective bactericide [55,56].Despite that most (eco)toxicity studies with ENPs observed some degree of toxicity, it is still unclear which physical and/or chemical characteristics of ENPs are responsible for it.The uncertainty in this respect is mainly due to the fact that most ENPs, used in toxicity tests, are seldom well characterized.
Since a very limited number of studies are made in the field of environmental fate of ENPs, their behavior in the environment is still largely unexplored.When addressing the environmental fate of ENPs, most of the literature uses imprecise general considerations and comparison with data, obtained for larger particles.It is very important to study the environmental fate of ENPs in order to understand their pathways of environmental and human exposure.Occupational exposures to ENPs are most likely to occur via inhalation and/or dermal contact [7,88].In most cases exposures are more likely to occur after the manufacturing process is complete (i.e., while handling and bagging the materials and also during cleaning operations) [85,88].Hansen et al. [96] estimated that the expected consumer exposure is highest for appliances, health, fitness, home and garden products.They noted that for most products, no information on the used materials is available, which is alarming since some of these materials might be potentially hazardous for their users.

Objective 2: Estimate Whether Current Knowledge is Sufficient to Facilitate Comprehensive and Effective Risk Assessment of ENPs
The above analysis identified a number of severe limitations and flaws in relation to each of the four steps of the risk assessment procedure, when it is applied to ENPs.Toxicity has been reported for many ENPs, but for most of them further investigation and confirmation are needed before hazard can be identified [12].Currently, most laboratories use non-standardized tests, generating non-reproducible results, which make the univocal hazard identification of ENPs very difficult.The DRA of ENPs is restricted by the enormous deficit of characterization data [59], which makes it impossible to determine which properties account for the inherent hazards of ENPs and identify appropriate dose metrics.Furthermore, despite that dose-response estimation assumes that no effect thresholds can be established, most studies, reporting dose-responce relationships, do not state any NECs [7].EEA is hindered by the fact that the biological and environmental pathways of ENPs are still largely unexplored.OEA of ENPs is hindered by the lack of information about: (1) the number of workers, exposed to ENPs, (2) the type of ENPs workers are exposed to, (3) the occupational exposure pathways, and (4) the concentrations of ENPs in the working settings [85].The deficiency of data in this regard is partly explained by the lack of adequate occupational exposure measurement methods and tools.CEA of ENPs requires complex modeling, addressing aspects such as ENP production volumes, number of nano-products, their market distribution, ENP releases from the products throughout their life-cycle etc. [96].There is insufficient knowledge with respect to these parameters, which is partly attributable to the lack of studies and partly to a lack of access to industry-derived data.
As it was shown, in most cases, the available information about the risks of ENPs for the environment and human health is insufficient to facilitate comprehensive and effective risk assessment.Each step of the procedure is hindered by serious data flaws, as RC, being at the end of the line, sums all the limitations.In order to facilitate sound risk assessment of ENPs, more research is needed.

Objective 3: Provide Recommendations on Future Research in the Field of Risk Assessment of ENPs
Based on the information presented in the previous chapters, it becomes obvious that there are many data deficits in regard to the EHS of ENPs, which require ungent research activities.In order to facilitate accurate risk assessment of ENPs, it is essential to start with adequate property characterization to understand the causality between the properties of ENPs and their inherent toxicity and identify suitable dose metrics for them.Furthermore, given the great diversity of ENP types, it would be useful to use tiered testing approaches, in which in vitro screening tests are used to detect any potential hazards and indicate whether more extensive in-vivo evaluation is necessary.The credibility and the broad acceptance of the test results will depend greatly on the establishment and utilization of standard reference materials, testing methods, and reporting formats.Standard testing would generate reproducible results and greatly aid the univocal HI of ENPs.
The most urgent research need in regard to the environmental exposure of ENPs is to establish the degree of their environmental mobility and bioavailability.Understanding the environmental fate of ENPs would greatly help us assess their exposure of ecosystems and of humans via the environment.Furthermore, for accurate exposure estimation it is necessary to obtain reliable environmental concentration data and therefore adequate sampling and monitoring technologies need to be developed.To facilitate effective OEA of ENPs it is essential to identify the exposure sources and pathways of ENPs in the working settings as well as to study the mechanisms behind their dispersion and measure their concentrations.For the latter, the development of adequate measurement techniques is a must.In order to facilitate proper CEA of ENPs future research should address multiple aspects (e.g., ENP global production volumes, number of products entailing ENPs, current and future market penetrations of these products, ENP releases throughout their life-cycles).Furthermore, a comprehensive inventory of consumer exposure data needs to be elaborated and made easily accessible to scientists and risk managers.
The deep uncertaincies, which currently pervade every step of the risk assessment of ENPs, make the procedure uncapable of properly serving its purpose.Data gaps with respect to EHS of ENPs are gradually filled by new research, but this process advances at low speed, while risk assessment results are urgently needed to triger adequate regulatory response.We recommend that the current risk assessment approach is adapted to reflect the challenges, discussed above, as it is aided by appropriate tools to manage the present uncertainties.Some approaches, which are considered useful in this respect are the Multi Criteria Decision Analysis, the Weight-of-evidence and the Expert Elicitation [106][107][108][109][110]. Implementing some of these non-conventional tools in the risk assessment framework holds promise to reduce uncertainties and deliver accurate risk characterization results very soon.This would enable current regulation to adequately reflect the risks of ENPs and protect the environment and the community.
and Porter et al. can be attributed to the fact that they used very different methods.Porter et al. used transmission electron microscopy (TEM) to image the distributions of the fullerenes within the macrophages, while Adelman et al. used a viability assay, based on metabolic activity as primary parameter.
Possible pathways of occupational, environmental and human exposure to ENPs.

Table 4 .
Needs for further research with respect to the risk assessment of ENPs.

Table 5 .
[103]es, aimed to develop and employ tiered approaches for HI of ENPs (modified after[103]).Environmental exposure assessment-related studies A number of studies, focused on the environmental fate of ENPs are currently in progress (Table

Table 7 .
[68]ing studies on the sources, dispersion mechanisms and concentrations of ENPs in occupational settings (modified after[68]).