Engineered nanomaterials: exposures, hazards, and risk prevention

A. The objectives of this review

Although there has been considerable work to advance nanotechnology and its applications, understanding the occupational, health and safety aspects of engineered nanomaterials (ENMs) is still in its formative stage. The goals of this review are to describe some general features of ENMs, how a worker might be exposed to ENMs, some potential health effects, and approaches to minimize exposure and toxicity. The target audience includes industrial hygienists, investigators working with these materials, institutes and universities conducting research, and start-up companies that may not have the necessary occupational health and safety expertise, knowledge, and/or staff.

A comprehensive review described the field of nanotoxicology six years ago, including some mechanisms of toxicity, portals of ENM entry, their translocation, and the state of their risk assessment at the time [1]. More recent reviews have focused on the major challenges, key questions, and research needs to assess ENM toxicity and risk [2–7]. This review addresses issues not extensively covered in prior reviews, including recent exposure-assessment studies, and engineering and personal protective equipment (PPE) options and their efficacy to minimize ENM exposure. This review also includes accepted but not yet published reports, recently completed studies not yet published, and ongoing work. Our goal was to provide up-to-date information on ENM exposures, their health hazards, and ways to minimize risk.

B. Engineered nanomaterials

Nano is a prefix derived from the Greek word for dwarf. The parts of the U. S. National Nanotechnology Initiative (NNI) definition that are relevant for this review define nanoscale materials as having at least one dimension in the range of 1 to 100 nanometers (nm), with properties that are often unique due to their dimensions, and that are intentionally manufactured [8]. There are many definitions of nanoscale materials, which generally encompass the same bounds on ENM size [9, 10]. This is in contrast to naturally occurring and unintentionally-produced materials on the same scale, which are referred to as ultrafine particles. The term ultrafine has been used by the aerosol research and occupational and environmental health communities to describe airborne particles smaller than 100 nm in diameter [11]. Ultrafine particles are not intentionally produced. They are the products of combustion and vaporization processes such as welding, smelting, fuel combustion, fires, and volcanoes [1, 12, 13]. In this review, intentionally-manufactured nanoscale materials will be referred to as ENMs. They are usually produced by bottom-up processes, such as physical and chemical vapor deposition, liquid phase synthesis, and self-assembly [5, 14].

The health and environmental effects of ENMs are not well understood, leading some to caution development of this technology [15–19]. Some understanding of ENM effects can be derived, however, by analogy from ultrafine particles, which have been shown to produce inflammation, exacerbation of asthma, genotoxicity, and carcinogenesis following inhalation. The following sections describe ENMs, and some of their uses and uncertainties, providing the context of this review.

C. Common ENM size, composition, and quality

Figure 1 relates ENM size to other chemical and biological materials. There are a staggering number of ENM compositions and shapes. Over 5000 patents have been issued for carbon nanotubes (CNTs) and > 50,000 varieties of CNTs have been produced [20]. The sheer number of ENMs contributes to the lack of our adequate understanding of ENM health and safety. They are primarily composed of carbon or metal/metal oxide, as illustrated by the representative manufactured nanomaterials selected for testing by the Organisation for Economic Co-operation and Development (OECD) [21]. Carbon-based ENMs include single-walled and multi-walled carbon nanotubes (SWCNTs and MWCNTs), graphene (a single sheet of carbon atoms in a hexagonal structure), spherical fullerenes (closed cage structures composed of 20 to 80 carbon atoms consisting entirely of three-coordinate carbon atoms, e.g., C₆₀ [Buckyballs, buckminsterfullerene]), and dendrimers, which are symmetrical and branched. SWCNTs and MWCNTs are ~1 to 2 and 2 to 50 nm wide, respectively, and can be > 1 μm long. The C₆₀ diameter is ~1 nm. Metal and metal oxide ENMs most commonly studied are cadmium in various complexes, gallium arsenide, gold, nickel, platinum, silver, aluminum oxide (alumina), cerium dioxide (ceria), silicon dioxide (silica), titanium dioxide (TiO₂, titania), and zinc oxide. The size of ENMs is in the same range as major cellular machines and their components, such as enzymes, making it likely that they will easily interact with biochemical functions [22].

Some ENMs contain contaminants, such as residual metal catalysts used in the synthesis of CNTs. ENM toxicity has been attributed to these residual metals, as discussed in II, B, 1. ENM exposure effects in the lung. The physico-chemical properties of ENMs, when tested prior to their use, are often different from those stated by the supplier [23, 24]. A major cause of changes in the physico-chemical properties of ENMs over time and in various media is agglomeration, discussed in II, A, 2. The physico-chemical properties of ENMs that impact their uptake. When ENMs are not sufficiently characterized to identify their composition or properties it makes the prediction of toxicity, when added to the insufficient understanding of their biological effects, even more difficult [25].

D. Some uses of ENMs and the projected market and workforce

There is considerable interest in developing ENMs because their properties differ in fundamental and valuable ways from those of individual atoms, molecules, and bulk matter. Nanoscale products and materials are increasingly being used in optoelectronic, electronic (e.g., computer hard drives), magnetic, medical imaging, drug delivery, cosmetic and sunscreen, catalytic, stain resistant fabric, dental bonding, corrosion-resistance, and coating applications [26]. Major future applications are expected to be in motor vehicles, electronics, personal care products and cosmetics, and household and home improvement. These applications capitalize on their electromagnetic, catalytic, pharmacokinetic, and physico-chemical properties, including strength, stiffness, weight reduction, stability, anti-fogging, and scratch resistance. Current products contain various ENMs including nanotubes, metal oxides, and quantum dots (semiconductors developed as bright, photostable fluorescent dyes and imaging agents). Nanowerk identified ~2500 commercial nanomaterials, including ~27% metal oxides, 24% CNTs, 18% elements, 7% quantum dots, and 5% fullerenes [http://www.nanowerk.com/phpscripts/n_dbsearch.php]. There are > 1000 consumer products available that contain ENMs. They are primarily composed of silver, carbon, zinc, silica, titania and gold. The main application is in health and fitness products [27, 28]. Three to four new nanotechnology-containing consumer products are introduced weekly into the market, according to The Project on Emerging Nanotechnologies [http://www.nanotechproject.org/inventories/consumer/].

The anticipated benefits of ENM applications resulted in expenditure of $18 billion worldwide on nanotechnology research and development in 2008. In 2004 Lux Research predicted that nanotechnology applications will become commonplace in manufactured goods starting in 2010 and become incorporated into 15% of global manufacturing output in 2014 [https://portal.luxresearchinc.com/research/document_excerpt/2650]. The ENM workforce is estimated to grow ~15% annually [29]. An epidemiological feasibility study of CNT workers initiated in 2008 revealed most manufacturers were small companies that had no environmental/occupational health and safety person and little knowledge about this topic [30]. By 2015, the global market for nanotechnology-related products is predicted to employ 2 million workers (at least 800,000 in the U.S.) to support nanotechnology manufacturing, and $1 trillion in sales of nanotechnology-related products [31].

E. Uncertainties regarding the adverse effects of ENMs

There have been concerns about the safety and public acceptance of this burgeoning technology, particularly in the past 5 years, due to the lack of much information about potential adverse effects [32]. This resulted in an increase from 2.9 to 6.6% of the NNI budget for environmental health and safety from 2005 to 2011. Prior to 2005 it does not seem funds were specifically allocated for this purpose nor was the U.S. National Institute for Occupational Safety and Health (NIOSH) a contributor to NNI funding [33, 34]. The United Nations Educational, Scientific and Cultural Organization (UNESCO) compared the concerns of the public over new products with their perception of genetically modified foods/organisms to nanotechnology. They noted that the lack of knowledge can result in restrictions, outright bans, and international conflicts over production, sale, and transport of such materials [35]. Public acceptance can influence the success of an emergent technology, as public opinion is considerably influenced by information prior to the adoption of the technology. However, individuals form opinions often when they do not possess much information, based on factors other than factual information, including values, trust in science, and arguments that typically lack factual content [36]. This creates a challenge to earn public acceptance of nanotechnology.

There is a notable lack of documented cases and research of human toxicity from ENM exposure. It is widely recognized that little is known about ENM safety. An uncertainty analysis revealed knowledge gaps pervade nearly all aspects of ENM environmental health and safety [4]. Owing to their small size and large surface area, ENMs may have chemical, physical, and biological properties distinctly different from, and produce effects distinct from or of a different magnitude than, fine particles of similar chemical composition. This is discussed in II, A, 2. The physico-chemical properties of ENMs that impact their uptake. ENM properties often differ from individual atoms, molecules, and from bulk matter. These differences include a high rate of pulmonary deposition, the ability to travel from the lung to systemic sites, and a high inflammatory potential [1]. Further contributing to our lack of understanding of the potential health effects of ENMs is that most production is still small scale. As such, potential adverse effects from the anticipated increase in large scale production and marketing of ENM-containing products and use are generally unknown. Furthermore, the number of novel ENMs being created continues to grow at a high rate, illustrated by the accelerating rate of nanotechnology-related patent applications [37, 38].

A. Hazard identification

In the occupational context, hazard identification can be re-stated as "What effects do ENMs have on workers' health?" to which NIOSH has stated: "No conclusive data on engineered nanoparticles exist for answering that question, yet. Workers within nanotechnology-related industries have the potential to be exposed to uniquely engineered materials with novel sizes, shapes, and chemical properties, at levels far exceeding ambient concentrations...much research is still needed." [http://www.cdc.gov/niosh/topics/nanotech/about.html].

Information about ENMs might be obtained from well-documented retrospective analyses of unintended exposures. The most extensive exposures to ENMs likely occur in the workplace, particularly research laboratories; start-up companies; pilot production facilities; and operations where ENMs are processed, used, disposed, or recycled [51]. Occupational hygienists can contribute to the knowledge and understanding of ENM safety and health effects by thorough documentation of exposures and effects. In the U.S., NIOSH is responsible for conducting research and making recommendations for the prevention of work-related illnesses and injuries, including ENMs. The U.S. Occupational Safety and Health Administration (OSHA) is responsible for making and enforcing the regulations.

1. The key routes of ENM exposure

Figure 3 illustrates the four routes that are most likely to result in ENM exposure of the five organ systems which are the major portals of ENM entry: skin, gastrointestinal tract, lung, nasal cavity, and eyes [22]. It also illustrates the most likely paths of translocation (re-distribution or migration), enabling ENMs to reach organs distal to the site of uptake.

The inhalation route has been of greatest concern and the most studied, because it is the most common route of exposure to airborne particles in the workplace. The skin has also been investigated. Most studies have shown little to no transdermal ENM absorption. Oral (gastrointestinal) exposure can occur from intentional ingestion, unintentional hand-to-mouth transfer, from inhaled particles > 5 μm that are cleared via the mucociliary escalator, and of drainage from the eye socket via the nasal cavity following ocular exposure. Direct uptake of nanoscale materials from the nasal cavity into the brain via the olfactory and trigeminal nerves has been shown. Each of these routes is discussed in more detail below.

Routes that avoid first-pass clearance and metabolism in the gastrointestinal tract and liver include uptake (absorption) from the nasal cavity (either into systemic circulation or directly into the brain), orotransmucosal (e.g., buccal [from the cheek] and sub-lingual), and transdermal. These routes may present a greater risk of ENM-induced adverse effects because more ENM is likely to reach the target organ(s) of toxicity.

B. The effects of ENM exposure on target organs and those distal to the site of uptake

Public concerns about ENMs and health may arise with reports of some effect(s) in a laboratory study or their presence in human tissue (or another organism). Any report must be interpreted carefully before concluding ENMs are risky for one's health. To start with, risk is defined as a joint function of a chemical's ability to produce an adverse effect and the likelihood (or level) of exposure to that chemical. In a sense, this is simply a restatement of the principle of dose-response; for all chemicals there must be a sufficient dose for a response to occur. Additionally, advances in analytical chemistry have led to highly sensitive techniques that can detect chemicals at remarkably low levels (e.g., in parts per billion or parts per trillion). The detectable level may be far lower than any dose shown to produce an adverse effect. Further, a single finding in the literature may garner public attention, and it may be statistically significant, but its scientific importance remains uncertain until it is replicated, preferably in another laboratory. In this regard, a follow-up study may be warranted to characterize the relevant parameters of dose, duration, and route of exposure, as outlined in the Risk Assessment/Risk Management framework.

The above discussion reflects many of the issues that have gained prominence in the fields of risk perception and risk communication (see for example [109, 110]), neither of which were dealt with by the NRC in their landmark publication.

The knowledge of ultrafine-particle health effects has been applied to ENMs. However, the toxicity from ultrafine materials and ENMs is not always the same [111]. Similarly, the effects produced by ENM components do not reliably predict ENM effects. For example, toxicity was greater from cadmium-containing quantum dots than the free cadmium ion [112]. Some metal and metal oxide ENMs are quite soluble (e.g., ZnO), releasing metal ions that have been shown to produce many of the effects seen from ENM exposure [113, 114]. Therefore, one cannot always predict ENM toxicity from the known effects of the bulk or solution ENM components.

C. Dose-response assessment

Exposure in experimental studies is typically expressed as dose, usually on a mass/subject body weight basis, or as concentration. Dose or concentration may not be the best metric to predict ENM effects [42, 53, 137]. Neutrophil influx following instillation of dusts of various nanosized particles to rats suggested it may be more relevant to describe the dose in terms of surface area than mass [138]. The pro-inflammatory effects of in vitro and in vivo nanoscale titania and carbon black best correlated when dose was normalized to surface area [122]. Secretion of inflammatory proteins and induction of toxicity in macrophages correlated best with the surface area of silica ENM [139]. Analysis of in vitro reactive oxygen species generation in response to different sized titania ENMs could be described by a single S-shaped concentration-response curve when the results were normalized to total surface area, further suggesting this may be a better dose metric than concentration [53]. Similarly, using surface area as the metric, good correlations were seen between in vivo (PMN number after intratracheal ENM instillation) and in vitro cell-free assays [42].

Nonetheless, most studies of ENMs have expressed exposure based on dose or concentration. The relatively small amount of literature has generally shown dose- or concentration-response relationships, as is usually the case for toxicity endpoints. Ceria ENM uptake into human lung fibroblasts was concentration-dependent for several sizes, consistent with diffusion-mediated uptake [58]. Positive, dose-dependent correlations were seen in blood, brain, liver, and spleen following iv ceria infusion in rats, measured by elemental analysis as cerium [23], as well as brain titanium after ip titania injection [140], and lung cobalt after inhalation of cobalt-containing MWCNTs [141]. Concentration-dependent inhibition of RAW 264.7 (murine) macrophage cell proliferation was seen following in vitro SWCNT exposure, as was lipopolysaccharide-induced COX-2 expression, up to 20 μg/ml [142]. Intratracheal instillation of MWCNTs (average length ~6 μm) or ground MWCNTs (average length ~0.7 μm) produced dose-dependent increases in LDH activity and total protein, but no dose-dependent effect on the number of neutrophils or eosinophils, or TNF-α, in rat lung bronchoalveolar lavage fluid [143]. Activated Kupffer cell count increased with iv ceria dose; the increase in hippocampal 4-hydroxy-2-trans-nonenal and decrease in cerebellar protein carbonyls (indicators of oxidative stress) were dose-dependent up to a maximum that did not increase further at the highest dose [23].

Some studies demonstrating adverse effects of CNT introduction to the lung have been criticized for using doses or concentrations that far exceeded anticipated human exposure [144]. Most studies assessing potential adverse effects of ENMs have utilized a single exposure. Both of these features make extrapolation of results to prolonged or episodic (periodic) human exposure difficult. However, the study of acute high doses/concentrations to probe potential effects is a standard approach in toxicology and experimental pathology for initially surveying adverse effects (i.e., hazard identification). When adverse effects are seen following some reasonable (e.g., sublethal) dose, subsequent studies must define exposures that do, and do not, result in adverse effects.

D. The clearance of ENMs, their translocation to distal sites, and persistence

As with the above studies that inform about uptake, the clearance and translocation of ENMs from the initial site of exposure to distal sites is best understood from whole-animal studies.

The solutes of dissolved particles in the lung can transfer to blood and lymphatic circulation. Some ENMs in the airway wall that slowly dissolve or are insoluble will be cleared within a few days from the lung by cough or the mucociliary escalator. Slowly dissolving and insoluble ENMs that reach the alveoli may be taken up by macrophages. Macrophage-mediated phagocytosis is the main mechanism for clearing foreign material from the deep lungs (alveoli) and from other organs. Macrophages are ~20 μm in diameter and able to phagocytose materials up to 15 μm in length. They engulf the particle in a vacuole (phagosome) containing enzymes and oxidizing moieties that catabolize it. Particles resistant to catabolism may remain inside the macrophage. After the death of the macrophage the material may be engulfed by another cell. Therefore, it may take a long time for insoluble material to be cleared from the body. The elimination half-live of insoluble inert particles from the lung can be years [145]. This raises the question of the ultimate fate of "poorly digestible" ENMs that are engulfed by macrophages in the lung, liver (Kupffer cells), brain (microglia), and other organs.

Some ENMs, e.g., those that have a high aspect ratio, are not effectively cleared by macrophages. Alveolar macrophages that cannot digest high-aspect-ratio CNTs (termed "frustrated phagocytosis") can produce a prolonged release of inflammatory mediators, cytokines, chemokines, and ROS [146]. This can result in sustained inflammation and eventually fibrotic changes. Studies have demonstrated MWCNT-induced pulmonary inflammation and fibrosis, similar to that produced by chrysotile asbestos and to a greater extent than that produced by ultrafine carbon black or SWCNTs [117]. Greater toxicity from a high-aspect-ratio metal oxide (titania) ENM has also been shown in cells in culture and in vivo[147]. Studies such as these have raised questions (and concern) about the long-term adverse effects of ENM exposure.

Translocation of ENMs from the lung has been shown. After MWCNT inhalation or aspiration they were observed in subpleural tissue, the site of mesotheliomas, where they caused fibrosis [148, 149]. Once ENMs enter the circulatory system across the 0.5-μm thick membrane separating the alveoli from blood, the sites of reticuloendothelial system function (including the lymph nodes, spleen, Kupffer cells, and microglia) clear most ENMs. Thirty to 40 nm insoluble ¹³C particles translocated, primarily to the liver, following inhalation exposure [150]. Similarly 15 and 80 nm ¹⁹²iridium particles translocated from lung to liver, spleen, heart, and brain. The extent of translocation was < 0.2%, and greater with the smaller ENMs [151].

ENMs have also been shown to translocate following injection. Indirect evidence was shown of fullerene distribution into, and adverse effects in, the fetus 18 h after its injection into the peritoneal cavity of pregnant mice on day 10 of gestation [152]. Following subcutaneous injection of commercial 25 to 70 nm titania particles into pregnant mice 3, 7, 10, and 14 days post coitum, aggregates of 100 to 200 nm titania were seen in the testes of offspring at 4 days and 6 weeks post-partum and in brain at 6 weeks post-partum. Abnormal testicular morphology and evidence of apoptosis in the brain indicated fetal titania exposure had adverse effects on development. The authors attribute these effects to ENM translocation across the placenta [153]. ENM excretion into milk and oral absorption post-partum might contribute to ENM presence in the offspring, but we are unaware of any studies assessing ENM translocation into milk. Non-protein bound substances generally enter milk by diffusion, and reach an equilibrium between milk and blood based on their pKa and the pH difference between blood and milk, described by the Henderson-Hasselbalch equation. Given the size of most ENMs, it is unlikely they would diffuse across the mammary epithelium. Within 40 weeks after a single intrascrotal injection of MWCNTs most rats died or were moribund with intraperitoneal disseminated mesothelioma, which were invasive to adjacent tissue, including the pleura. Fibrous MWCNT particles were seen in the liver and mesenteric lymph nodes, suggesting peritoneal effects might have been due to MWCNT translocation [154].

The distribution of carbon-, metal- and metal oxide-based ENMs after translocation from the lung, skin or intestine is similar to that seen after their iv administration. They generally appear as agglomerates in the liver and spleen [23, 93, 132, 151, 155–158]. The ENMs are usually in the cytoplasm, with little indication that they enter the nucleus [132, 134, 158–160].

Due to their small size ENMs may gain access to regions of the body that are normally protected from xenobiotics (sanctuaries), such as the brain. This feature has suggested their potential application for drug delivery to the brain, which is being extensively pursued [161–164], but at the same time it raises concern about central nervous system distribution of ENMs when exposure is not intended. Studies have generally found << 1% of the administered dose of ceria and iridium ENMs translocate to the brain after inhalation exposure or iv injection [23, 132, 151]. Anionic polymer ENMs entered the brain more readily than neutral or cationic ones. Both anionic and cationic ENMs altered blood-brain barrier integrity [165].

The persistence of ENMs may be a major factor contributing to their effects. Many ENMs are designed to be mechanically strong and resist degradation [22]. Referring to nanoscale fiber-like structures, it has been stated: "The slower [they] are cleared (high bio-persistence) the higher is the probability of an adverse response" [166]. The analogy of high-aspect-ratio ENMs to asbestos is one of the contributors to this concern. The prolonged physical presence of ENMs, that are not metabolized or cleared by macrophages or other defense mechanisms, appears to elicit ongoing cell responses. The majority of CNTs are assumed to be biopersistent. For example, two months after the intratracheal instillation of 0.5, 2 or 5 mg of ~0.7 μm and ~6 μm MWCNTs, 40 and 80% of the lowest dose remained in the lungs of rats, suggesting adequate persistence to cause adverse effects that are summarized in II, B, 1 ENM exposure effects in the lung[143]. Following oral administration, 50-nm non-ionic polystyrene ENMs were seen in mesenteric lymphatic tissues, liver, and spleen 10 days later [167]. Following iv administration, carboxylated-MWCNTs were cleared from circulation and translocated to lung and liver; by day 28 they were cleared from the liver, but not from the lung [168]. No significant decrease of the amount (mass) of cerium was seen in the liver or spleen of rats up to 30 days after iv administration of 5 or 30 nm ceria. Hepatic granuloma and giant cells containing agglomerates in the cytoplasm of the red pulp and thickened arterioles in white pulp were seen in the spleen (unpublished data, R. Yokel) [159, 169].

In summary, the persistence of ENMs in tissue raises justifiable concerns about their potential to cause long-term or delayed toxicity.

E. The physico-chemical properties of ENMs that impact their hazard - The role of surface coating in ENM effects

Many surface coatings have been investigated in order to develop ENMs as carriers for drug delivery. Surface modifications can prolong ENM circulation in blood, enhance uptake at a target site, affect translocation, and alter excretion. When ENMs enter a biological milieu they rapidly become surface coated with substances such as fulvic and humic acids and proteins, all of which can alter their effects [142, 170, 171]. When 3.5, 20, and 40 nm gold and DeGussa P-25 titania ENMs were incubated with human plasma, proteins appeared to form a monolayer on the ENMs. The abundance of plasma proteins on gold approximated their abundance in plasma, whereas some proteins were highly enriched on titania [172]. Metal oxide and carbon-based ENMs rapidly adsorb proteins [66], resulting in changes in their zeta potential (electrical potential at the ENM surface) and toxicity [142, 171]. For circulating ENMs, the surface coating is extremely important, because this is what contacts cells [173].

Although it is understood that ENMs will be surface coated with proteins, lipids or other materials, which may or may not persist on the ENM surface when they enter cells, little is known about the surface associated molecules on ENMs within cells. It is likely, however, that surface coatings profoundly influence ENM effects within cells. Although surface functional groups are known to modify ENM physico-chemical and biological effects, there is little information on the influence of functional groups on health effects. This further complicates the prediction of ENM toxicity in humans from in vitro, and perhaps in vivo, studies.

F. The effects of ENMs at distal sites

Reported systemic effects of pulmonary-originating CNTs include acute mitochondrial DNA damage, atherosclerosis, distressed aortic mitochondrial homeostasis, accelerated atherogenesis, increased serum inflammatory proteins, blood coagulation, hepatotoxicity, eosinophil activation (suggesting an allergic response), release of IL-6 (the main inducer of the acute phase inflammatory response), and an increase of plasminogen activator inhibitor-1 (a pro-coagulant acute phase protein) [118]. Elevation of the serum analyte ALT was reported up to 3 months after intratracheal MWCNT instillation, suggesting ENM-induced hepatotoxicity [174]. The translocation of ENMs and their release of cytokines and other factors could potentially affect all organ systems, including the brain. For example, daily ip injection of titania for 14 days resulted in a dose-dependent increase of titanium and oxidative stress and a decrease of anti-oxidative enzymes in the brain of rats [140].

A. Engineering controls

ENM exposure can be reduced through the use of engineering controls, such as process changes, material containment, and enclosures operating at negative pressure compared to the worker's breathing zone; worker isolation; separated rooms; the use of robots; and local exhaust ventilation (LEV). Given the lack of occupational exposure standards to provide guidance, the most prudent approach is to minimize exposure. A survey found that engineering controls in Switzerland were more commonly used in the production of powder than liquid ENMs. For the former, the use of PPE (masks, gloves, safety glasses, or full protective suits) were the norm, although used by only ~16, 19, 19, and 8% of the workers, respectively [195]. This low use of PPE is thought to reflect the early stage of development of the ENM industry. It is anticipated that as this industry matures and knowledge is gained, control will more commonly include superior methods in the hierarchy of exposure control [196]. An international survey of ENM industry managers conducted in 2009-2010 by the University of California Center for Environmental Implications of Nanotechnology that focused on industry controls of ENM exposure, use of PPE, environmental risks, and perceptions revealed that 46% of the companies had a nano-specific environmental health and safety program, compared to 58% in a 2006 survey [197]. Some companies (a minority) were using inappropriate occupational environmental clean-up methods, such as sweeping and compressed air [198]. These results suggest more widespread adoption of nano-specific environmental health and safety programs and the use of PPE in the absence of superior controls are appropriate.

B. Administrative controls

When engineering controls are not feasible for reducing exposure, administrative controls should be implemented. These are policies and procedures aimed at limiting worker exposure to a hazard [209]. These could include a nanoscale material hygiene plan; preparation, training in, and monitoring use of standard operating procedures; reduction of exposure time; modification of work practices; and good workplace and housekeeping practices. For example, one laboratory was thoroughly cleaned after high air concentrations of nanoscale materials were measured in a facility engaged in the commercial compounding of nanocomposites [183]. A large decrease of airborne 30 to 100 nm particles resulted. Subsequent routine maintenance kept the particles below those originally observed, leading the authors to conclude that this administrative control was beneficial in reducing potential exposure. Biological monitoring and medical examination, a component of secondary prevention (Figure 4), is another administrative control [209]. It is discussed below.

C. Personal protective equipment

The last line of defense in the hierarchy of exposure control is the use of PPE, such as respirators, protective clothing, and gloves.

1. Respirators

Major types of respiratory protection include dust masks, filtering facepiece respirators, chemical cartridge/gas mask respirators, and powered air-purifying respirators. Examples can be seen at OSHA's Respiratory Protection Standard site [http://web.utk.edu/~ehss/pdf/rpp.pdf].

NIOSH classifies filter efficiency levels as Type 95, 99, and 100, which are 95, 99, and 99.7% efficient, respectively. The filter's resistance to oil is designated as N, R, and P; N (not resistant to oil), R (resistant to oil), and P (oil proof). Some industrial oils can remove electrostatic charges from filter media, reducing filter efficiency. Efficiency of N filters is determined using 300-nm median aerodynamic, charge neutralized, NaCl particles at a flow rate of 85 l/min; R and P filters are tested with dioctyl phthalate oil. The European Standards (EN 143 and EN 149) rank filtering facepiece (FFP) respirators as FFP1, FFP2, and FFP3, which are 80, 94, and 99% efficient, respectively, indicated by CE (for Conformité Européene) on complying products. They are tested with non-neutralized NaCl at 95 l/min.

Particles > 100 nm are collected on filter media by two mechanisms: 1) inertial impaction in which air flow deviates around the fiber but large denser-than-air particles do not and impact the fiber due to their inertia, as shown in Figure 5; and 2) interception where the particle trajectory takes it within a particle radius of the fiber, which captures the particle. Airborne nanoparticles behave much like gas particles. Particles < 100 nm are collected by diffusion. Charged particles are trapped by electrostatic attraction, which involves an electrically charged particle and an electrically charged (electret) fiber. Electret filters are constructed from charged fibers. This appears to be a significant mechanism for respirator trapping of ENMs [210]. Neutral particles that pass through a charged fiber can be polarized by the electric field, thereby inducing charge to the particle. In dry conditions, ENM penetration decreases with time. With continued use, however, ENM penetration through an electrostatic filter increases; this was suggested to be due to the humidity of exhalation [211]. Soaking fiber filters in isopropanol removes electrostatic charge. Studies treating filtering facepiece respirators with isopropanol, and then drying them, showed increased penetration of particles > 30 nm [210], indicating electrostatic charge is a significant mechanism of fiber entrapment of ENMs above this size.

Figure 6 shows results of some studies that assessed the efficacy of different types of dust masks and filtering facepiece respirators to retain ENMs. Most of the studies were conducted with different sizes of NaCl, but a few used silver, graphite or titania. The results show that dust masks purchased at hardware or home improvement stores would not be expected to protect the wearer. The results also show that the NIOSH and CE respirators generally limit penetration of ENMs to concentrations below their ranked efficiency level, which is based on 300 nm particles, except for the N99 filter, which did not retain more than 99% of nanoscale NaCl (Panel D). Most of the results shown were obtained with a flow rate of 85 l/min (modeling heavy work load). Flow rate affects particle penetration; an example is shown in Panel C where 30 l/min (modeling low/moderate intensity work) and 85 l/min flow rates were compared. Increasing flow rate increased penetration. This was further shown in a comparison of 30, 85, and 150 l/min flow rates with N95 and N99 filtering facepiece respirators [212], which is not shown in Figure 6. This highest flow rate was intended to model an instantaneous peak inspiratory flow during moderate to strenuous work. A similar result of ENM penetration positively correlating with air flow rate is shown in Panel F, where 5.3 and 9.6 cm/sec face velocity rates were compared.

The results shown in Figure 6 also indicate ENM penetration is influenced by particle composition. Panel G shows greater penetration of titania than graphite through FFP3 respirators under the same experimental conditions. Panel H shows greater penetration of 20 to 30 nm NaCl than silver ENMs of the same size through P100 filters. These results suggest further work is warranted to understand the influence of the physico-chemical properties of ENMs, particularly size, charge, and shape, on their penetration through filtering facepiece respirators. An issue that significantly impacts filtering facepiece respirator effectiveness is its seal around the face. "The biggest source of leakage is around the respirator seal because of poor fit" (Ronald E. Shaffer quoted in [213]). It has been estimated that 20% or more leakage occurs in respirators that are not properly fitted [214]. This underscores the importance of a proper fit for face mask respirators.

There is a particle size that maximally penetrates each filter material; the most penetrating particle size (MPPS). The results shown in Figure 6 indicate that the MPPS is ~40 to 50 nm for ENMs. This is approximately the same size of spherical ENMs that appear to contribute to their greatest differences in biological systems from solution and bulk forms of the same materials, as discussed in II, A, 2. The physico-chemical properties of ENMs that impact their uptake. This feature raises concern because the size of ENMs that may have the greatest effects in people are those that are best able to penetrate filtering facepiece respirators.

Until results are obtained from clinical-laboratory or work-place studies, traditional respirator selection guidelines should be used. These are based on OSHA Assigned Protection Factor values (the workplace level of respiratory protection that a respirator is expected to provide), the British Standards Institution Guide to implementing an effective respiratory protective device programme (BS 4275), and BS EN 529:2005 (Respiratory protective devices. Recommendations for selection, use, care and maintenance. Guidance document).

D. Biological monitoring and medical examination

Secondary prevention in the continuum of the prevention and heirarchy of exposure control (Figure 4) includes biological monitoring and medical examination, the early detection of asymptomatic disease, and prompt intervention when the disease is preventable or more easily treatable [221]. Occupational health surveillance is the process by which information obtained from any activity in the continuum of prevention and heirarchy of exposure control is collected and used to support or alter what is done at a step higher in the heirarchy, as shown in the right upward pointing arrow in Figure 4 and discussed in [194]. Occupational health surveillance is the ongoing systematic collection, analysis, and dissemination of exposure and health data on groups of workers for the purpose of early detection and injury. It includes hazard surveillence, the periodic identification of potentially hazardous practices or exposures in the workplace, assessing the extent to which they can be linked to workers, the effectiveness of controls, and the reliability of exposure measures. A goal is to help define effective elements of the risk management program for exposed workers. Occupational health surveillance also includes medical surveillance, which examines health status to determine whether an employee is able to perform essential job functions [222]. It is required when there is exposure to a specific workplace hazard (OSHA, 29 CFR 1910.1450). This is different than medical screening or monitoring, a form of medical surveillance designed to detect early signs of work-related illness by administering tests to apparently healthy persons to detect those with early stages of disease or those at risk of disease. NIOSH concluded: "Currently there is insufficient scientific and medical evidence to recommend the specifc medical screening of workers potentially exposed to engineered nanoparticles" [222].

E. Diagnosis, therapy, and rehabilitation

The third level in the continuum of prevention and heirarchy of exposure control, tertiary prevention, includes diagnosis, therapy, and rehabilitation. Owing to the lack of documented episodes of ENM exposure in humans that have resulted in adverse outcome, there is little experience with treatments of ENM exposure. One example that illustrates clever application of the knowledge of ENM properties was the use of UV light to visualize and treat the accidental dermal exposure of a human to quantum dots suspended in solution [223].