Nanodynamite

One of the greatest challenges in all of technology right now is improving energy storage.

It's an enormous challenge on many fronts and on many scales, with examples ranging from utility-scale batteries as big as a trailer down to the button-size cells that keep our quartz wristwatches ticking. And while we know how to make batteries bigger—add more cells—we are up against fundamental limits as we try to scale down rechargeable energy sources into the micro realm and beyond.

It's probably not a challenge you've thought much about, unless you're designing the next generation of nanoscale memory-storage devices, circuits, magnets, or biosensors, which all need extremely small and rechargeable power sources. Most of the many researchers working on rechargeable power sources these days are trying to wring a few more kilometers out of an electric-car battery pack or a few more minutes out of a cellphone charge. But the challenges of the micro realm are just as fascinating, even if they're not of commercial consequence yet.

But they may soon be. Researchers already envision tiny systems whose realization is blocked by the lack of sufficiently small power sources. These future systems include cardioverter-defibrillators the size of apple seeds, implanted relatively unobtrusively in a heart patient to automatically control heart arrhythmias. Also, microsize systems could one day attack cancerous cells with localized pulses of energy inside the body. Another possible application for these puny power­houses is in "smart dust"— sensors that float in the air, collecting information about temperature, airborne pollutants, and other characteristics.

We think that a solution to this long-standing problem is finally within reach. Working at MIT's Strano Laboratory, along with colleagues from the Royal Melbourne Institute of Technology, in Australia, we have been pursuing a new approach to energy storage and power generation. Our experimental system, based on one of the new materials that have come from nanotechnology—­carbon nanotubes—generates power in a way that has no macroscopic analogy. By coating a nanotube in fuel and igniting one end, we set off a combustion wave along it and learned that a nanotube is an excellent conductor of the heat from the burning fuel. Even better, the combustion wave creates a strong electric current.

Illustration: Emily Cooper

How to Catch a Thermopower Wave: A single carbon nanotube becomes a thermopower wave generator when it's coated in fuel and ignited on one end, which sets off a chemical reaction. The reaction starts moving along the nanotube's walls even faster than the fuel itself is consumed. This results in a high-speed thermopower wave, in which heat from the reaction feeds back to the fuel. As this wave travels, it pushes electrical carriers to the end of the nanotube, thus creating the electric current. Click to enlarge.

Before we explain how our nanotube-­based power source works, it is worth considering the alternatives. For instance, why not just keep scaling down chemical batteries? The simple reason is that battery performance begins to degrade when engineers reduce a battery's size to several tens of micrometers. The flow of ions is disrupted, sharply reducing the battery's power density—the amount of power it produces per unit of mass.

What about other energy-storage technologies, such as rocket engines, which convert chemical energy to mechanical work? They don't typically scale to the sizes needed for the next generation of technologies, either. To date, scaled-down rocket engines can sustain power densities of only around 0.1 watt per kilogram—a minute fraction of the 200 W/kg that average lithium­-ion batteries achieve.

A few newer technologies have the potential to power future micro- and nanoscale systems. Very small fuel cells, about the size of a peppercorn, might yet make promising power sources. But despite many years of research and innovation, fuel cells of this size have yet to produce sufficiently high levels of power per unit of mass. Fuel cells designed to work with microscale systems produce only 100 W/kg.

To understand how our tiny power sources work, first consider the carbon nanotube, a hollow, submicroscopic  tube made of a chicken-wire-like lattice of carbon atoms. A single nanotube has an average diameter of 5 or 6 nanometers, but some have diameters as small as 1 nm. Their lengths vary: A nanotube can be as short as tens of nanometers, but researchers have grown some that are many millions of times as long—up to tens of centimeters. Nanotubes are extremely strong and have a high density of electrical carriers—electrons and electron deficiencies, or holes—which enable the nanotubes to conduct both electricity and heat very well.

Researchers have been studying carbon nanotubes for decades. But new advances over the past few years have made them easier to produce and use experimentally. These factors prompted us to look for ways to use carbon nanotubes to harness thermal energy in a previously undiscovered way. Basically, we wanted to take advantage of their shape and strength to propagate an explosive chemical reaction along the outside of the tube. If we could use the tube to harness the energy of such strong reactions, we reasoned, we would have a small system with exceptionally high power density.

In 2009, we began coating carbon nanotubes with a highly energetic fuel. When we ignited one end using either a laser beam or a hot wire, we found that the chemical reaction set off a wave of heat that propagated along the nanotube's walls like a flame along the length of a lit fuse. The resulting combustion wave traveled down the nanotube at speeds 10 000 times as fast as the fuel would burn in open air—from 0.01 up to 2 meters per second. That's because the core of the system—the nanotube—conducts heat so well. The heat entering the carbon nanotube propagates much faster than within the fuel itself, allowing the heat to ignite more fuel as it travels. The nanotube guides the decomposition of the fuel on the surface while keeping the reaction moving in one direction, thus serving as a guide for the wave and allowing a large current to flow unimpeded.

We call this combustion wave a thermopower wave because as it transmits energy from one place to another, it couples with the nanotube's electrical carriers, setting them in motion along the conducting tube. The moving heat source sweeps the coupled electrical carriers along, an effect we call electron entrainment. The carriers all move together in a single direction, creating an electrical current that is extremely large relative to the mass of the system. In some of our latest experiments, the power generated exceeded 7 kilowatts per kilogram, or about three to four times what is possible with the best lithium-ion batteries currently available.

As far as we know, we are the first to harness these thermopower waves to convert chemical energy to electric power, and we're still trying to understand some of the fine points of how these waves propagate heat and electrical carriers to produce electricity. What we do know is that the propagation of the charge carriers depends on the thermoelectric effect—the voltage that results from a steady temperature difference across a wire. Depending on the nanomaterial, either electrons or holes flow from the hot side to the cold side; their density determines the current.

The main challenge with any power generator is, of course, to maxi­mize its output. One way to accomplish this is by choosing materials with good thermoelectric properties. To understand what makes a good thermoelectric material, recall that power equals voltage times current. For a thermoelectric power generator, this translates to the voltage differential across the wire multiplied by its induced current. That means there are only two ways to increase power output—increase the voltage by increasing the temperature differential between the two ends of the wire or increase the current by decreasing the wire's resistance. So in choosing the material for the wire, researchers strive to maximize both the temperature differential—you want the greatest possible difference between the hot end and the cold end—and minimize the resistance.

It's no small feat. That's because materials that are highly conductive of electricity—metals, for example—are usually also good conductors of heat. So in almost all materials with low electrical resistance, it is impossible to sustain a large temperature differential between two points on the material—the high thermal conductivity precludes such a differential. That's exactly what French physicist Jean-Charles Peltier and Estonian-German physicist Thomas Johann Seebeck quickly learned when they independently discovered the thermo­electric effect in the first half of the 19th century—many metallic materials are weak when it comes to maximizing thermoelectric properties.

The first real breakthrough in designing materials for thermoelectric power generation came in 1958, when Tasmanian physicist H. Julian Goldsmid, now emeritus professor at the University of New South Wales, noticed that certain semiconducting materials—bismuth telluride and antimony telluride—have something unusual in common: high electric conductance and low thermal ­conductivity. This meant it might be possible to maintain a large temperature gradient in these materials. Early experimenters even dreamed of producing unlimited amounts of power; they imagined very long wires and rods, with one end in a hot environment, such as a geyser or volcanic crater, and the other end in a cool environment, perhaps several meters underground. Free and everlasting electric power!

Well, not quite. Although semiconducting materials like bismuth telluride and antimony telluride have lower thermal conductivity than any metal except mercury, even a small amount of conductivity is enough to prevent a material from sustaining a temperature difference between its two sides forever. Eventually, heat moves from one side to another, and a temperature balance is reached.

Researchers have tried for years to exploit the thermoelectric effect by manipulating materials with maximal electrical conductivity and minimal thermal conductivity. But despite extreme efforts, they have been stymied by the difficulty of that process. Ten years ago, scientists at the Research Triangle Institute, in North Carolina, were able to combine bismuth telluride and antimony telluride to make a material with almost no thermal conductivity at all. However, the fabrication process faced practical difficulties, and it's still far from being cost-effective for power genera­tion. Researchers at Boston College; Wayne State University, in Michigan; Nanjing University, in China; and the Korea Institute of Science and Technology are also studying similar synthesized structures, but none are viable for everyday use yet.

That's where carbon nanotubes come in. Carbon nanotubes have extremely high thermal conductivity because of their crystal-like molecular structure. And like bismuth telluride, they also have high electrical conductivity. Our discovery that a thermopower wave works best across these tubes because of their dual conductivity turns conventional thermo­electricity on its head: It's the first nanoscale approach to power generation that exploits the thermo­electric effect but sidesteps the feasibility issues associated with minimizing thermal conductivity.

Here's why: The thermoelectric effect says that the more efficiently a material can sustain a temperature differential across it, the more electrical potential it has. So materials with very little thermal conductivity but high electrical conductivity can sustain high electrical potential, which determines a material's voltage. The ­thermoelectric effect suggests that the faster a material conducts heat, the more quickly it loses its electrical potential.

But if, as with our system, voltage is related to how fast electrical carriers are moving—and the ­carriers are moving because of a high-­temperature reaction—then thermal conductivity isn't a drawback at all. In fact, it's essential. It turns out that sustaining a temperature difference across a material isn't really important: It's all about how fast you can move electrical potential down a wire. Our trick of using an explosive reaction exploits a carbon nanotube's thermal conductivity and its high concentration of electrons to speed up the chemical reaction. That reaction—rather than the ends of the tube—is what provides the needed temperature difference.

We're not the only ones trying to better understand how carbon nanotubes can generate power by means of the ­thermoelectric effect. Researchers at institutions such as Sungkyunkwan University, in South Korea; Texas A&M; Wake Forest University and NanoTechLabs, in North Carolina; and Victoria University, in New Zealand, are now looking at why carbon nanotubes can produce thermoelectric power and how to harness it better. What we know so far is that with thermal waves, the movement of a temperature gradient from one end of a material to the other is what creates voltage. And the faster it moves, the better: Higher thermal conductivity and higher temperatures mean a stronger electrical current.

A thermopower-wave generator produces up to 0.2 to 0.3 volts and 0.1 to 0.2 amperes of electrical current. The current is generated as a pulse, typically several milliseconds long. We can also increase the ­capacity and current by arranging nanotubes in parallel or increase the voltage by putting a number of nanotubes in series.

The current appears to scale up or down with wave velocity. We initially assumed that the true mechanism here wasn't wave velocity but rather wave-front temperature, itself the effect of a stronger thermal reaction that swept more electrical carriers along the tube. But after testing several fuels that reacted at different speeds on a nanotube, we confirmed that wave velocity is in fact the more important factor. We are still trying to explain why this is the case in order to understand more about how thermal waves couple with electrons and electron holes. One of our future goals is to find equations that help us calculate the relationship between the temperature, the voltage, and the wave‑front behavior. So far, we've seen that the fuel reaction needs to generate localized temperatures around a thousand degrees Celsius in order to start a reaction that's fast enough to kick off the thermal wave.

While that might seem like an absurdly high temperature for use in any practical application, the high heat is contained within an area smaller than a cell. It's so highly localized and insulated by the nanotubes that we think the high temperature would be safe in almost any device, even ones inside the human body.

Early in our research, we began to see ways that we could modify and control carbon nanotubes, hoping to demonstrate their usefulness in future systems. Many electrical applications require only the sorts of power pulses that thermopower wave generators provide.

Because the length of the nanotube determines the duration of a reaction and how many charge carriers are entrained by the system, it therefore determines the energy and duration of the power pulse. By changing the length and choosing a fuel that supplies the right energy, we can effectively set the system's propagation velocity.

During our experiments, we made thermopower wave generators that had many different dimensions and properties and tested them in different conditions. Interestingly, the smallest generators produced the largest power densities.

Thermopower wave generators are also remarkable in other ways. For instance, before the nanotubes are ignited, the chemical energy can be stored in the fuel coating indefinitely. Batteries can leak or erode over time, but carbon nanotubes stay completely intact until lit, as well as after a reaction. They also have simple designs, which can take advantage of standard industrial micro- and nanofabrication technologies, and they have potential as nanogenerators. And they're rechargeable—reapplying the fuel is all it takes to launch another thermopower wave. Carbon nanotubes are sturdy enough to remain intact after the reaction, even though reaction temperatures exceed 1000 °C around the nanotube and the devices operate in open air.

Encouraging as some of these results have been, there is much we need to learn before we can turn carbon nanotube generators into a commercially viable power source. For instance, the 200 to 300 millivolts our systems have been able to put out so far isn't enough for most applications. We hope to find out whether different materials or mixtures of materials could produce more voltage. We'd also like to try liquid and gaseous fuels, which could work together with microfluidic systems.

On another front, we recently discovered that changing the conductivity of carbon nanotubes, by doping or other means, alters the propagation velocity of the thermal waves along the tube. And by changing certain properties of the fuel it should, in theory, be possible to produce both alternating and direct current with a ­single reaction; we've already seen that the rate of a reaction can decrease and increase as it travels.

Right now, we're trying to better understand the physics of thermopower waves. Already there are multiple angles to explore when it comes to taming these exotic waves and, ultimately, finding out if they're the wave of the future.