A PAIR OF SCIENTISTS, sporting white clean-suits complete with helmets and face masks, approach a prefab agricultural greenhouse in a clearing at Duke University’s Research Forest. Inside are two long rows of wooden boxes the size of large horse troughs, which hold samples of the natural world that surrounds them—the pine groves and rhododendron thickets of North Carolina’s piedmont, which at this moment are alive with bird song.

Looking a lot like the government bad guys in E.T., the two men cautiously hover over a row of boxes containing native sedges, water grasses, and Zebra fish to spray a fine mist of silver nanoparticles over them. Their goal: to investigate how the world inside the boxes is altered by these essentially invisible and notoriously unpredictable particles.

The researchers are part of a multidisciplinary coalition of scientists from Duke, Stanford, Carnegie Mellon, Howard, Virginia Tech, and the University of Kentucky, headquartered at Duke’s Center for the Environmental Implications of NanoTechnology (CEINT), that represents one of the most comprehensive efforts yet to measure how nanoparticles affect ecosystems and biological systems.

So far the questions about whether nanoparticles are an environmental risk outnumber the answers, which is why the Duke scientists take the precaution of wearing clean-suits while dosing the boxes—no one’s sure what exposure to a high concentration of nanoparticles might do. Among the few things we do know about them are that they sail past the blood-brain barrier and can harm the nervous systems of some animals.

The regulation of nanoparticles has been recommended for more than a decade, but there’s no agreement on exactly how to do it. Meanwhile, the lid has already been lifted on nanotechnology. The use of man-made nanoparticles has spread into almost every area of our lives: food, clothing, medicine, shampoo, toothpaste, sunscreen, and thousands of other products.

Regulatory structures, both here and abroad, are completely unprepared for this onslaught of nanoproducts, because nanoparticles don’t fit into traditional regulatory categories. Additionally, companies often shield details about them by labeling them “proprietary”; they’re difficult to detect; we don’t have protocols for judging their effects; and we haven’t even developed the right tools for tracking them. If nanotechnology and its uses represent a frontier of sorts, it’s not simply the Wild West—it’s the Chaotic, Undiscovered, Uncontrollable West.

And yet, when I visit the boxes on a warm spring day filled with the buzzing of dragonflies and the plaintive call of mourning doves, they look perfectly benign and could easily be mistaken for a container garden. But there are hints that more is going on: each “mesocosm” (a middle ground between microcosm and macrocosm) is studded with probes and sensors that continually transmit data to CEINT’s central computer.

As I instinctively squint my eyes to try and locate evidence of the silver nanoparticles inside each box, I realize I might as well be staring down at these research gardens from another arm of the galaxy. The scale of these two worlds is so disparate that my senses are destined to fail me.

AS WITH MANY THINGS that are invisible and difficult to understand—think subatomic particles such as the Higgs boson, muons, gluons, or quarks—any discussion of nanoparticles quickly shifts into the realm of metaphor and analogy. People working in nanoscience seem to try to outdo each other with folksy explanations: Looking for a nanoparticle is like looking for a needle in the Grand Canyon when the canyon is filled with straw. If a nanoparticle were the size of a football, an actual football would be the size of New Zealand. A million nanoparticles could squeeze onto the period at the end of this sentence.

But what is a nanoparticle? The very simplest explanation is that a nanoparticle is a very small object. It can consist of any bit of matter—carbon, silver, gold, titanium dioxide, pretty much anything you can imagine—that exists on the scale of nanometers. One nanometer equals one-billionth of a meter. A nanoparticle may range in size from one nanometer to one hundred nanometers, although the upper boundary remains a matter of debate among scientists.

Nanoparticles exist in nature, but they can also be manufactured. One way is top-down: grinding up things that are big until they are really, really small, an approach used in nanolithography for electronics. Or you can make them from the bottom up, following instructions that read like a chemistry textbook: mixing one chemical with another by pyrolysis (heating a material in a partial vacuum), or with electrolysis (running a current through a liquid), or by other means.

But what do they look like? Raju Badireddy, a postdoctoral researcher, is happy to satisfy my curiosity. He greets me with a smile at the door to one of CEINT’s basement labs and guides me around his little domain. For much of his work, Badireddy uses a “dark field” microscope that excludes certain wavelengths of light, reducing the “noise” in the image to provide unparalleled clarity. Sensing my anticipation, he doses a slide with silver nanoparticles similar to those in the mesocosm boxes in the forest, and slips it under the lens.

As I look into the scope, it fairly takes my breath away. There are so many dots of light that I’m reminded of staring up at the Milky Way on a trip across the Tibetan Plateau years ago. Yet the silver dots throb and undulate as if alive. Here and there, giant spheres of dust, as large as Goodyear blimps, porpoise through the nanoparticles. I pull back from the oculars, feeling as if I’ve intruded upon something private. This world is so close—it’s even inside me—yet it looks so other, so mysterious.

Scientists don’t really have a full theoretical foundation to explain reality at this scale. But all agree that one of the most important aspects of nanoparticles is that they are all surface. Consider a conventional chemical process: When one element is reacting with another, it’s really just the surface molecules that are involved in the lock-and-key dance of classical chemistry. The vast majority of the molecules remain interior, and stable. But there are many fewer molecules in a nanoparticle, so most of the molecules are on the outside, thus rendering nanoparticles more reactive.

Myriad surface imperfections cause randomness to dominate the nano world. If you hit a billiard ball with a clean shot at the macro level, you can have a good idea where it will go. But at the nano level, a billiard ball might shoot straight up, or even reverse direction. These bits of matter are hot to trot: ready to react, to bond, and to do so in unpredictable ways.

This makes life at the nano scale more chaotic. For instance, aluminum is used everywhere to make soda cans. But in nanopowder form, aluminum explodes violently when it comes in contact with air. At the macro level, gold is famously nonreactive. At the nano level, gold goes the opposite way, becoming extremely reactive. Bulk carbon is soft. But at the nano level, if you superheat it, the molecules bend into a tube that is very strong and semiconductive. In the nano world, gravity fades to the background, becoming less pronounced, the melting temperature of materials changes, and colors shift. At 25 nanometers, spherical gold nanoparticles are red; at 50 nanometers they are green; and at 100 nanometers they’re orange. Similarly, silver is blue at 40 nanometers and yellow at 100 nanometers.

So chemistry and physics work differently if you’re a nanoparticle. You’re not as small as an atom or a molecule, but you’re also not even as big as a cell, so you’re definitely not of the macro world either. You exist in an undiscovered country somewhere between the molecular and the macroscopic. Here, the laws of the very small (quantum mechanics) merge quirkily with the laws of the very large (classical physics). Some say nanomaterials bring a third dimension to chemistry’s periodic table, because at the nano scale, long-established rules and groupings don’t necessarily hold up.

These peculiarities are the reason that nanoparticles have seeped into so many commercial products. Researchers can take advantage of these different rules, adding nanoparticles to manufactured goods to give them desired qualities.

Scientists first realized that nanomaterials exhibit novel properties in 1985, when researchers at Rice University in Houston fabricated a Buckminsterfullerene, so named because the arrangement of sixty carbon atoms resembles the geodesic domes popularized by architect Richard Buckminster Fuller. These “Buckyballs” resist heat and act as superconductors. Then, in 1991, a researcher at the Japanese technology company NEC discovered the carbon nanotube, which confers great strength without adding weight. Novel nano materials have been reported at a feverish pace ever since.

With these engineered nanoparticles—not even getting into the more complex nanomachines on the horizon—we can deliver drugs to specific cells, “cloak” objects to make them less visible, make solar cells more efficient, and manufacture flexible electronics like e-paper.

In the household realm, nanosilica makes house paints and clothing stain resistant; nanozinc and nano–titanium dioxide make sunscreen, acne lotions, and cleansers transparent and more readily absorbed; and nanosilicon makes computer components and cell phones ever smaller and more powerful. Various proprietary nanoparticles have been mixed into volumizing shampoos, whitening toothpastes, scratch-resistant car paint, fabric softeners, and bricks that resist moss and fungus.

A recent report from an American Chemical Society journal claims that nano–titanium dioxide (a thickener and whitener in larger amounts) is now found in eighty-nine popular food products. These include: M&Ms and Mentos, Dentyne and Trident chewing gums, Nestlé coffee creamers, various flavors of Pop-Tarts, Kool-Aid, and Jell-O pudding, and Betty Crocker cake frostings. According to a market report, in 2010 the world produced 50,000 tons of nano–titanium dioxide; by 2015, it’s expected to grow to more than 200,000 tons.

AT FIRST some in the scientific community didn’t think that the unknown environmental effects of nanotechnology merited CEINT’s research. “The common view was that it was premature,” says CEINT’s director, Mark Wiesner. “My point was that that’s the whole point. But looking at risk is never as sexy as looking at the applications, so it took some time to convince my colleagues.”

Wiesner’s team at CEINT chose to study silver nanoparticles first because they are already commonly added to many consumer products for their germ-killing properties. You can find nanosilver in socks, wound dressings, doorknobs, sheets, cutting boards, baby mugs, plush toys—even condoms. How common is the application of nanoparticles? It varies, but when it comes to socks, for example, hospitals now have to be cautious that the nanosilver in a patient’s footwear doesn’t upset their MRI (magnetic resonance imaging) machines.

Wiesner and his colleagues spent several months designing the experiments that will help them outline some general ecological principles of the unique nanoverse. He knew they wanted to test the particles in a system, but a full-scale ecosystem would be too big, too unmanageable, so they had to find a way to container-ize nature. They considered all sorts of receptacles: kiddie pools (too flimsy), simple holes in the ground (too dirty, too difficult to harvest for analysis), concrete boxes (crack in winter). Finally, they settled upon wooden boxes lined with nonreactive, industrial rubber: cheap to build, easy to reuse, and convenient to harvest.

They built thirty boxes and a greenhouse to hold them. The large number would make it easier to replicate experiments, and to answer the spectrum of questions being posed by CEINT’s interdisciplinary team. The ecologists were interested in community diversity and how the biomass shifts over time. The biologists wanted to know whether the nanoparticles become concentrated as they move up the food chain. The toxicologists wanted to track where the particles went and how fast they got there. The chemists wanted to know about reactivity.

Whatever the goal of the experiment it houses, each mesocosm features a slanted board upon which a terrestrial ecosystem slowly gives way to an aquatic one. It’s a lot more complicated than a test tube in a lab, but it remains an approximation. The team had hoped to run streams through the mesocosms, but the computing power and monitoring vigilance necessary to track nanoparticles in the streams proved prohibitive.

In 2011, the team dosed the boxes with two kinds of nanosilver made on campus: one coated in PVP, a binder used in many medicines, and the other coated in gum arabic, a binder used in numerous products, including gummi candies and cosmetics. Both coatings help to stabilize the nanosilver. In some boxes, the researchers let the silver leach slowly into the box. In other boxes, they delivered the silver in one big pulse. In some, they introduced the silver into the terrestrial part of the box; in others, they put the silver into the water.

Then the researchers watched and waited.

READING THROUGH DESCRIPTIONS of nanoparticle applications can make a person almost giddy. It all sounds mostly great. And the toxicology maxim “Dose makes the poison” leads many biologists to be skeptical of the dangers nanoparticles might pose. After all, nanoparticles are pretty darn small.

Yet size seems to be a double-edged sword in the nanoverse. Because nanoparticles are so small, they can slip past the body’s various barriers: skin, the blood-brain barrier, the lining of the gut and airways. Once inside, these tiny particles can bind to many things. They seem to build up over time, especially in the brain. Some cause inflammation and cell damage. Preliminary research shows this can harm the organs of lab animals, though the results of some of these studies are a matter of debate.

Some published research has shown that inhaled nanoparticles actually become more toxic as they get smaller. Nano–titanium dioxide, one of the most commonly used nanoparticles (Pop-Tarts, sunblock), has been shown to damage DNA in animals and prematurely corrode metals. Carbon nanotubes seem to penetrate lungs even more deeply than asbestos.

What little we know about the environmental effects of nanoparticles—and it isn’t very much—also raises some red flags. Nanoparticles from consumer products have been found in sewage wastewater, where they can inhibit bacteria that help break down the waste. They’ve been found to accumulate in plants and stunt their growth. Another study has shown that gold nanoparticles become more concentrated as they move up the food chain from plants to herbivores.

“My suspicion, based on the limited amount of work that’s been done, is that nanoparticles are way less toxic than DDT,” says Richard Di Giulio, an environmental toxicologist on the CEINT team. “But what’s scary about nanoparticles is that we’re producing products with new nanomaterials far ahead of our ability to assess them.”

As a society, we’ve been here before—releasing a “miracle technology” before its potential health and environmental ramifications are understood, let alone investigated. Remember how DDT was going to stamp out malaria and typhus and revolutionize agriculture? How asbestos was going to make buildings fireproof? How bisphenol A (BPA) would make plastics clear and nearly shatterproof? How methyl tertiary-butyl ether (MTBE) would make gasoline burn cleanly? How polychlorinated biphenyls (PCBs) were going to make electrical networks safer? How genetically modified organisms (GMOs) were going to end hunger?

The CEINT scientists are trying to develop a library that catalogues all the different kinds of engineered nanoparticles. They’re designing methods for assessing potential hazards, devising ways to evaluate the impact nanoparticles have on both terrestrial and aquatic ecosystems, and creating protocols that will help shape environmental policy decisions about nanoparticles.

Wiesner says the boxes in the forest provide “ground truth” for experiments in the lab. Sometimes, he says, environmental research leads to generalizations that become so abstracted that they have no relationship to reality. The example he likes to give is Freon: if you were to study the toxicology of Freon in the traditional way, you’d never get to the ozone hole. “Nature changes things,” Wiesner says. “So we need to be able to understand those transformation processes, and we need to understand them in complex systems.”

The first large set of CEINT experiments ended about a year ago, and the team spent most of last year figuring out where the nanoparticles went, what they did, and how they added up. They superimposed a grid on each box, then harvested the plants and animals section by section. They clipped the grasses, sorted them by type, and ground them up. They took bore samples of the soil, the water, and the rocks. They anesthetized and flash froze the vertebrates. Then they started measuring the nanoparticle concentrations in the plants, the animals, and core-sample slices.

But consider the magnitude of the scientific problems that face the scientists at CEINT, or anyone else trying to answer a multitude of questions as nanotech applications gallop into the market and man-made nanoparticles begin to litter our world. Just try tracking something a billion times smaller than a meter in even a modestly sized ecosystem, say, a small wetland or a lake. Do carbon nanotubes degrade? And if not, then what? And how do you tell the nanotubes from all the other carbon in your average ecosystem? Even if we did regulate nanoparticles, how would we detect them? There’s no “nanoprobe” that could find them today, and given the challenges of developing such a thing, the team at CEINT considers it unlikely that there will be one any time soon. Thus, gathering evidence of nanoparticles’ effects—whether positive or negative—turns out to be a titanic task. Simply finding them in the experiment samples seems about as complicated as finding that needle in a haystack the size of the Grand Canyon.

LEE FERGUSON, a chemistry professor who directs the nanoparticle analysis, meets me in the basement of the CEINT building and leads me on a tour of all the hulking, pricey instruments the researchers use. Despite the cutting-edge aura of this machinery, none of it is fully up to the task of locating and analyzing the proverbial nanoneedle.

“With nanoparticles, we’re playing catch-up as a scientific community—not only to ask the right questions, but to have the right tools to investigate them,” Ferguson says as he pushes through a door into the first lab. “We were well prepared to answer questions about PCBs—we’d spent half a century refining the chemistry and the instruments that were used to analyze the molecules in those chemicals. But simply measuring nanoparticles is a challenge. It’s one thing if they’re concentrated, but if you’re looking for nanoparticles in soil, for instance, you just can’t find them.”

He spends the next hour showing me how the CEINT team has back-engineered methods to detect and characterize nanoparticles. The fluorometer aims three lasers at carbon nanotubes. Another instrument uses ultrasonic waves to flush out its tiny quarry. Across campus, huge electron microscopes train electron beams on the nanoparticle samples, projecting their images onto a charge-coupled device camera, like the ones used on the Hubble Telescope, and atomic force microscopes form images of them by running a probe over samples like a hypersensitive, high-tech record player.

As the team’s methods continue to advance, their experiments have resulted in some surprising data. “After we dosed the water, we took some of it to the lab and exposed fish to it,” says Wiesner’s research assistant, Benjamin Espinasse. “Some of the particles turned out to be more toxic in the lab. And the reverse also happened: some things didn’t appear to be toxic in the lab, but they were more toxic in the boxes. It seems that the organic matter in the mesocosms changed the coatings of the particles, making them more toxic or less toxic,” Espinasse continues. “We could never have imagined that.”

While CEINT has only published the results of the preliminary mesocosm experiments, the team has been able to make a few conclusions: When the nanoparticles come in a burst, they tend to stay in the soil. But if they bleed into the system slowly, they filter into the water column. Regardless, nanoparticles seem to have a tendency to stick around—that was also the case with DDT.

Meanwhile, CEINT has begun a new set of experiments in the boxes: testing nanoparticles that have been combined with various other substances.

“The materials we most see now are nanomaterials incorporated into other products: textiles, foams, mattresses, nanotubes in display screens,” Wiesner explains. “How it will get out into the environment will be very different than just the pristine particle.”

And then there are the nanobots to plan for. “As we get closer to even simple nanobots, we will need to understand how to do research on them, too,” Wiesner says. Although they remain a marvel of the future, scientists are working toward nanomachines that may someday be able to replicate red blood cells, clean up toxic spills, repair spinal cord injuries, and create weapon swarms to overwhelm an enemy. Researchers are already working on simple versions of nanobots using the chemical principles of attraction and repulsion to help nanostructures arrange and build themselves in a process akin to the way DNA works: a strand of DNA can only split and rebuild in one particular way, and the desired structure is preserved, no matter how many times the DNA replicates.

As if trying to figure out the effects of simple nanoparticles weren’t enough of a futuristic challenge, concerns surrounding nanobots that replicate like DNA are so theoretical they’re spoken about in narratives resembling science fiction. Sun Microsystems founder Bill Joy famously warned that, if released into the environment, self-assembling and self-replicating nanomachines could spread like pollen or bacteria, and be too tough and too small to stop before invading every part of the biosphere, chewing it up and reducing all life on earth to “gray goo.” In nanotech circles, this is called the “gray goo problem,” but no one really knows if this vision is prophetic or simply hysterical.

DOWN THE BASEMENT HALLWAY, postdoc Badireddy motions to me to join him at a computer monitor next to the dark field microscope in his lab. He clicks on a movie he’s made from images he’s captured. It shows silver nanoparticles interacting with bacteria.

At first, the nanoparticles don’t seem to be doing much. Then, all of a sudden, they start to clump to the outside of a bacterium. The nanoparticles build up and build up until the bacterium’s cell membrane bursts. Then the nanoparticle clumps dissolve into small units before clumping back up again and attacking more bacteria. “The whole cycle happens in about thirty minutes,” Badireddy says. “It’s so fast. If you leave the nanoparticles overnight, when you come back in the morning, all the bacteria are ground mush.”

If you’re looking for stink-free athletic socks, maybe this is a good thing. But could that same process someday turn out to have some sort of nasty biological effect? We just don’t know yet.

“The fact that they re-cycle suggests they might persist for a long time,” Badireddy says as we watch the movie a second time. “They might enter the food chain. And then, who knows what will happen?”

Is nanotechnology a panacea or Pandora’s box? Listen to a conversation with Heather Millar.