This article was originally written by the National Nanotechnology Initiative (NNI) as a guide for reporters and journalists writing about nanotechnology health and safety risks. It offers good background knowledge in regards to understanding the promises and dangers in nanotechnology.
Why Small Particles Are A Big Story
For decades, scientists have anticipated from theory that if they could manipulate individual molecules, they could engineer materials with electronic, optical, and other properties not observed in bulk - and open new frontiers in electronics, medicine, and consumer products. Rather as cells use a few amino acids to assemble proteins with a wide range of characteristics and functions, nanotechnology may make it possible to design and engineer materials at the molecular level to have specific properties. “There is plenty of room at the bottom” is an often-quoted prophetic quip of the late Caltech physicist Richard A. Feynman in 1959.
Half a century later, the promise of nanotechnology is becoming reality - not only in the lab but already in some commercial consumer products ranging from sunscreens to self-cleaning windows. More exciting are possibilities of targeted cancer therapies, where a tumor may be eradicated without making the rest of the body sick. Environmental researchers are investigating the use of engineered nanoscale materials (engineered nanomaterials for short) to purify or desalinate water, to improve energy efficiency, or to clean up hazardous wastes. Indeed, people are starting to talk about engineered nanomaterials as a completely new class of materials, and nanotechnology as being a new industrial revolution—as significant to the twenty-first century as the first industrial revolution was to the nineteenth century and the information-technology revolution was to the twentieth.
But with such a revolutionary new technology come questions about occupational, consumer, and environmental safety and health. If engineered nanomaterials have physical properties different from their bulk counterparts, might they also pose new risks to human health in their manufacture, use, and disposal?
As yet, no one knows. Current data basically suggest “it depends.” But researchers both in government and private industry are keen to find out.
First, toxicity itself can be useful. Indeed, it is highly sought for certain applications, such as cancer therapies. (Also, keep in mind that often toxicity depends on dose and administration: even table salt is toxic in high doses.)
Second, if toxicity is known, handling and packaging procedures can be devised to mitigate risks of undesired exposure in manufacturing processes, as is routinely done in industries using hazardous materials. Safe-handling procedures for engineered nanomaterials may need to differ from those now used for larger micrometer-sized particulates—especially important for nanomanufacturing workers. Questions have also been raised about the safety of engineered nanomaterials in consumer products or in implantable medical devices, or to plants and animals in the environment after disposal.
Third, nanotechnology developers are heeding a lesson in perceived risk from an unrelated high-tech field: consumer resistance that arose at the introduction of crops and products using genetically modified organisms (GMOs). In part, that resistance arose because biotech companies introduced GMO products without much open discussion of legitimate questions and concerns in the general public, with the result that the public felt it had to accept risks to health and environment while benefits were limited to increased profits for large agribusiness. The result was widespread public mistrust and suspicion. Wanting to avoid a similar fate (especially given that concern and calls for regulation already have been expressed in some quarters), nanotech developers are pursuing what they call “responsible development.” That specifically includes encouraging early, forthright press coverage of work in assessing risks as well as benefits of engineered nanomaterials, as well as straightforward regulations devised through transparent processes.
This article has three purposes: to sketch essential basics of the physics and biology of engineered nanomaterials (and, for that matter, also natural and incidental nanoparticles), to highlight key issues and resources, and - most importantly - to warn about contradictory findings and pitfalls of logic and to suggest insightful questions for sources.
Disagreement On Classification
According to the National Academies, a distinction is made between three types of nano-scale particles (often abbreviated in the literature as “NSPs”): natural, incidental, and engineered. Natural nanoparticles occur in the environment (volcanic dust, lunar dust, magnetotactic bacteria, mineral composites, etc.). Incidental nanoparticles, sometimes also called waste or anthropogenic particles, occur as the result of manmade industrial processes (diesel exhaust, coal combustion, welding fumes, etc.). Both natural and incidental nanoparticles may have irregular or regular shapes. Engineered nanoparticles most often have regular shapes, such as tubes, spheres, rings, etc.
Engineered nanomaterials can be produced either by milling or lithographic etching of a large sample to obtained nanosized particles (an approach often called “top-down”), or by assembling smaller subunits through crystal growth or chemical synthesis to grow nanoparticles of the desired size and configuration (an approach often called “bottom-up”). Since the specific production technique might influence human health risk, ask sources to specify.
Recent questions about toxicity are directed at engineered nanomaterials. Nonetheless, the literature about natural and incidental nanoparticles is helpful, because more is known about them (in part, because of research on smog, welding fumes, coal dust, and ultrafine aerosols), and because information about their behavior can be helpful for understanding the behavior of engineered nanoparticles.
Also according to the National Academies, nanoscale materials—whether engineered or natural—so far seem to fall into four basic categories. The group currently with the largest number of commercial nanomaterials is the metal oxides, such as zinc or titanium oxides, which are used in ceramics, chemical polishing agents, scratch-resistant coatings, cosmetics, and sunscreens. A second significant group is nanoclays, naturally occurring plate-like clay particles that strengthen or harden materials or make them flame-retardant. A third group is nanotubes, which are used in coatings to dissipate or minimize static electricity (e.g., in fuel lines, in hard disk handling trays, or in automobile bodies to be painted electrostatically). The last group is quantum dots, used in exploratory medicine or in the self-assembly of nanoelectronic structures. But be aware: not every official source finds the same categorization useful. For example, the U.S. Environmental Protection Agency divides engineered nanoparticles into carbon-based materials (nanotubes, fullerenes), metal-based materials (including both metal oxides and quantum dots), dendrimers (nano-sized polymers built from branched units of unspecified chemistry), and composites (including nanoclays).
Disagreement On Definition
Most U.S. and British nanotech experts define NSPs as particles smaller than 100 nanometers (nm)—that is, 0.1 micrometer or micron (μm)—in any one dimension. Thus, a fiber thinner than 100 nm would be considered an NSP, even if it were several micrometers long. This definition, however, is not universal. In Japan, particles between 50 and 100 nm are classed as “ultrafines” and only those below 50 nm in one dimension are considered genuine NSPs. That being said, even some U.S. agencies also use the term “ultrafines” to describe particles under 100 nm (although usually in the context of only natural or incidental nanoparticles—seldom referring to engineered nanoparticles).
To resolve such confusion, ISO, IEC, ANSI, ASTM, and other national and international standards bodies are now discussing the standardization of terminology, metrology, characterization, and approaches to safety and health. Until all that is finalized, ask sources to clarify definitions and assumptions underlying their specific work. The distinctions might be crucial to the physics and biology being reported.
Just how small is 100 nm? It’s about one hundred-thousandth the diameter of a human hair (which is 50 to 100 μm). More usefully, 1 μm (1,000 nm) is about the size of a bacterium, about the limit of what is visible through most light microscopes. In contrast, 100 nm is about the size of a virus, a tenth the size of a bacterium. NSPs, like viruses, are invisible even through the best light microscope, because they are smaller than wavelengths of light (which range from about 700 nm in the red to 400 nm in the violet); they can be imaged only with some higher-resolution instrument such as a scanning electron microscope. 1 nm is about the size of a single sugar molecule.
Four Anticipated Generations
Already, scientists are talking in terms of generations of engineered nanomaterials. First-generation is passive nanostructures, such as individual particles, coatings, etc. - types of engineered nanomaterials already incorporated into some consumer products. Second-generation is nanostructures that perform an active function, such as transistors or sensors, or that react in an adaptive way; many are under development. Third-generation engineered nanomaterials might be three-dimensional systems that could self-assemble or be used to target drug delivery to specific parts of the body, anticipated to be developed about 2010. Fourth generation is anticipated to be molecular structures by design.
A simple thought experiment shows why nanoparticles have such phenomenal surface area per unit volume. A solid cube of a material 1 cm on a side—about the size of a sugar cube—has 6 square centimeters of surface area, about equal to one side of half a stick of gum. But if that volume of 1 cubic centimeter were filled with cubes 1 mm on a side, that would be 1,000 millimeter-sized cubes (10 x 10 x 10), each one of which has a surface area of 6 square millimeters. The total surface area of the 1,000 cubes adds up to 60 square centimeters—about the same as one side of two-thirds of a 3 x 5 notecard—because one must count the surface areas of all the millimeter cubes even in the interior of the original volume. But when that single cubic centimeter of volume is filled with cubes 1 nanometer on a side—yes, 1021 of them, each with an area of 6 square nanometers—their total surface area comes to 60 million square centimeters or 6,000 square meters. In other words, a single cubic centimeter of cubic nanoparticles has a total surface area a third again larger than a football field!
Figure 1. Surface area diagram
[Source: Trudy E. Bell; graphics courtesy of Nicolle Rager Fuller]
The Surprising Physics Of Engineered Nanomaterials
At the nanoscale, fundamental mechanical, electronic, optical, chemical, biological, and other properties may differ significantly from properties of micrometer-sized particles or bulk materials.
One reason is surface area. Surface area counts because most chemical reactions involving solids happen at the surfaces, where chemical bonds are incomplete. The surface area of a cubic centimeter of a solid material is 6 square centimeters—about the same as one side of half a stick of gum. But the surface area of a cubic centimeter of 1-nm particles in an ultrafine powder is 6,000 square meters—literally a third larger than a football field. (See Figure 1, above.)
Thus, collections of NSPs with their enormous surface areas can be exceptionally reactive (unless a coating is applied), because more than a third of their chemical bonds are at their surfaces. For example, nanoparticles of silver have been found to be an effective bactericide—inspiring several companies to design reusable water-purification filters using nanoscale silver fibers.
At what size do a material’s properties start changing? Is it a gradual transformation as one proceeds from large to small, or is there a threshold below which the properties abruptly change? Both may be true, actually. Quantum-size effects begin to significantly alter material properties (such as transparency, color of fluorescence, electrical conductivity, magnetic permeability, and other characteristics) whenever they dominate thermal effects, which for many materials is around 100 nm. For electronic properties, quantum-size effects increase inversely with decreasing particle size. Yet, for some materials, other distinct properties become pronounced at particular sizes - for example, gold nanoparticles have greatly increased catalytic properties at 3 nm. Characterizing material effects at different sizes is a hot area of basic research.
Figure 2. Structures of Diamond, Graphite and Buckminsterfullerene
Carbon and some other elements (including sulphur, tin, and oxygen) are found in multiple structural forms, called allotropes, which have significantly different properties. For example, in crystalline form, pure carbon is found as graphite (very soft), diamond (very hard), and various sizes of Buckminsterfullerenes (depending on the number of carbon atoms).
Engineered nanomaterials with the identical chemical composition can have a variety of shapes (including spheres, tubes, fibers, rings, and planes). Moreover, every one of these shapes may have different physical properties, because the pattern of molecular bonds differ even though they are composed of the same atoms.
For example, until 1985, it was believed that pure carbon came in only two crystalline forms: graphite (whose hexagonal crystal lattice lies in a two-dimensional plane) or diamond (whose cubic crystal lattice extends in all three dimensions). That year, hollow cages of 60 carbon atoms in a soccerball shape were first made in the laboratory (and also independently discovered in distant stars and in combustion byproducts) - a new crystalline form of carbon so significant it was recognized by the Nobel Prize in Chemistry in 1996. The new form, quite stable, was named a buckyball or fullerene after the architect Richard Buckminster Fuller, inventor of the geodesic dome of the same shape. Since then, stable fullerenes of 70, 74, and 82 carbon atoms have also been synthesized. (See Figure 2, above)
Similarly, titanium dioxide (TiO2) has been synthesized in NSPs of at least two different shapes and crystalline structures, each of which may have different toxicities. Although titanium dioxide is normally opaque white - indeed, is used to make white paints - as engineered nanoparticles, its optical qualities change, allowing it to become transparent. Yet it still effectively blocks ultraviolet light, a combination of properties attractive to makers of cosmetics and sunscreens.
Other properties matter. Other material properties that may be more important than just size include charge, crystal structure, surface coatings, residual contamination depending on method of synthesis, and tendency of individual nanoparticles to aggregate into larger clumps.
Hazard, Risk and Other Terms Of Art
If the physical properties of NSPs are so different from bulk materials, what might be the implications for toxicology and the risk of human exposure? First, some essential definitions:
Several everyday words have specific meanings in the fields of risk analysis, toxicology, or occupational safety and health.
“Hazard” is the potential to cause harm; it is an intrinsic property of a material. Sulfuric acid, for example, is a hazardous material by virtue of its chemistry. Nothing can change that, short of altering its chemistry to become something else.
“Risk” is the likelihood of harm occurring; it is a combination of a hazard with the probability of exposure and the magnitude and frequency of doses. Risks, unlike hazards, can be managed and minimized: a hazardous material poses low risk if the chances of exposure and the magnitude and frequency of the dose that might be received through that exposure are low. Leaving an unlabeled paper cup of concentrated sulfuric acid on a kitchen counter poses high risk because the chance of exposure and the potential dose are high; but the same acid, if properly labeled and locked in a chemistry lab to which only trained personnel have access, poses minimal risk.
“Exposure” is a combination of the concentration of a substance in a medium multiplied by the duration of contact. For example, dilute sulfuric acid that splashes and is quickly washed off is a low-exposure dose that may only redden the skin; concentrated sulfuric acid allowed to sit on skin is a high-exposure dose that likely will cause serious burns.
“Dose” is the amount of a substance that enters a biological system and can be measured as systemic dose, the total amount taken up by the biological system, or as the amount in a specific organ (skin, lung, liver, etc.). And herein lie more unanswered questions.
Questions About Dosimetry
Up to now, exposure to dust and toxic doses have been measured in terms of mass per unit volume, commonly milligrams per cubic meter. However, even very low concentrations of NSPs - whether natural, incidental, or engineered - in the air represent a phenomenal number of particles, as is well known from measurements of ultrafine pollutants. Exposing lab rats to 100-nm titanium dioxide particles has evoked the same amount of pulmonary inflammation as 10 times greater mass of larger (1–2.5-μm) particles. In fact, in at least some cases, the amount of inflammation seems to be better correlated to particle surface area of administered NSPs than to their mass. Thus, some toxicologists are now wondering whether surface area would be a better measure of dose for NSPs than mass. Until researchers know which counts most, many investigators are starting to specify both in their papers.
The Surprising Toxicology Of Nanoparticles
Size may have another crucial biological consequence: where nanoparticles end up in the body.
A complex of physical factors such as aerodynamics, gravity, and mass causes the largest inhalable dust particles to deposit primarily in the nose and throat. Any toxic effects occur at that site (for example, nasal cancers due to wood dust). Smaller particles are deposited in upper airways and are expelled by the “mucosociliary escalator;” the fingerlike cilia and the mucous lining of the trachea and bronchial tubes, which together move particles up into the throat and nose, where they are coughed, sneezed, blown out, or swallowed. Any toxic effects usually result from absorption through the gut (lead poisoning for example).
The next smallest particles penetrate deeper into the alveolar region (where oxygen and carbon dioxide are exchanged in and out of the blood) and are usually cleared when alveolar macrophages (special monocytic scavenger cells in the lungs) engulf the particles and carry them away. But if a high concentration of NSPs is inhaled, the sheer number of particles - especially if they do not agglomerate - can overwhelm those clearance mechanisms, and they can penetrate to different parts of the respiratory tract. Toxic effects are usually due to killing of the macrophages, which causes chronic inflammation that damages lung tissue (asbestosis and silicosis are examples).
At sizes less than 100 nanometers, inhaled particles begin to behave more like gas molecules and can be deposited anywhere in the respiratory tract by diffusion. Like gases, NSPs—whether natural, incidental, or engineered—simply because of their “nanoscopic” size, can pass through the lungs into the bloodstream and to be taken up by cells, within hours reaching potentially sensitive sites such as bone marrow, liver, kidneys, spleen, and heart.
As particles become small compared to the size of a cell, they can begin to interact with the molecular machinery of the cell. The central nervous system’s olfactory bulb (where aromatic molecules are detected) seems to be able to absorb NSPs smaller than 10 nm from the nasal cavity - which then can travel along axons and dendrites to cross the blood-brain barrier.
Inhalation is not the only route into the body. When ingested, NSPs can end up in the liver, the spleen, and the kidneys. When touched, NSPs in the range of 50 nm and smaller tend to penetrate the skin more easily than larger particles (although other aspects such as charge and surface coatings of the particles are also important), sometimes, being taken up by the lymphatic system and localizing in the lymph nodes. (See Figure 3, below.)
By the same token, the mucosociliary escalator is also not the only way out of the body. There is evidence suggesting that nanoparticles could be excreted through urine. However, excretion routes for nanoparticles (urine, feces, sweat) are likely to vary depending on exposure route, size, charge, surface coating, chemical composition, and many other factors.
For incidental exposure, all this uptake of NSPs into internal organs could be of concern. But for therapeutic exposure, it is exciting, as it suggests that engineered nanomaterials can be used to target therapies to specific organs, even ones normally quite difficult to reach (such as the brain).
So far, results from different investigators are more suggestive than definitive. More research needs to be done on methods of administration, means of uptake, and on the body’s clearance mechanisms. Also, when nanometer-sized particles are generated in combustion processes, most collide with other particles, are held together by the strong surface tension, and agglomerate into larger particles. The distribution of particles sizes will depend on the density of nanometer particles at the point of generation. One of the early priorities for nanotechnology health research is to gain a better understanding of the particle sizes that are likely to be associated with the production of engineered nanoparticles.
Still, size is not the only thing that matters for potential toxicity.
Figure 3. Biokinetics of nanoscale particles
Nanoscale particles can end up in different parts of the body depending on size and other characteristics as well as routes of entry. Although many uptake and translocation routes have been demonstrated, others still are hypothetical and need to be investigated. Translocation rates are largely unknown, as are accumulation and retention in critical target sites and their underlying mechanisms. These, as well as potential adverse effects, largely depend on physicochemical characteristics of the surface and core of NSPs. Both qualitative and quantitative changes in NSP biokinetics in a diseased or compromised organism also need to be considered.
Although the shapes of NSPs also give them unique properties, under the Toxic Substances Control Act (TCSA) engineered nanoparticles may not be viewed as new compounds unless they have a unique composition. For example, TiO2 nanoparticles are handled the same way with respect to regulation as bulk TiO2, even though the two forms have different properties.
Some studies show that the materials having the same composition but of different shapes as well as sizes have different toxicities - moreover, not with a linear relationship as one might expect. For example, one study showed that nanoparticles 50 to 130 nm across of quartz-crystalline silica (a substance known to be toxic) were less toxic than 1.6-μm particles - but that 10-nm particles were actually more toxic. But route of entry into the body as well as dose also affect toxicity.
Bulk carbon in macroscopic components is medically useful because it is not poisonous to or rejected by the body. Yet, some researchers have observed from experiments that carbon nanotubes (especially single-walled or multi-walled carbon nanotubes) seem to be more toxic than other forms of carbon. Others have debated that claim because the nanotubes used had trace impurities of iron or solvents. Indeed, some studies suggest that other forms of nanoscale carbon such as C60 fullerenes might prevent toxicity by being antioxidants.
Possibly at stake here, or in similar debates over other engineered nanomaterials, may be the purity of the engineered nanomaterials. At this stage, people don’t have absolutely repeatable control on manufacturing processes; nanotech production is now roughly where the production of indium gallium arsenide phosphide (InGaAsP) semiconductor lasers were in the early to mid 1980s - relatively low yield of reliable production. Thus, buckyball products from one supplier are not necessarily identical to those from another, so toxicity may differ. Ask sources careful questions about the size of particles, their manufacture, experimental methods, whether they characterized the materials themselves at the time when they performed the experiment or simply believed the statements made by the supplier, and the comparison of their results with other studies.
With more research under way, there are more and new publications reporting on nanotoxicology. Until more is certain, the National Institute for Occupational Safety and Health (NIOSH) has announced research needs and interim guidelines for protecting workers in nanotech industries in its report Approaches to Safe Nanotechnology.