For the first time, researchers have developed a way to accurately measure how heat flows within carbon nanotubes–tiny molecular wires that could someday be used to make circuits that are much faster and more energy efficient than today’s. The results show that nanotube heating is more complicated than previously thought–a fact that could be crucial in enabling engineers to build carbon-nanotube electronics.

Traditional semiconductors, such as silicon, undergo heating, says Phaedon Avouris, leader of IBM’s nanoscale science and technology group in New York, where the work was carried out. “It’s one of the limitations in improving speed,” says Avouris. But the study published by his team today in Nature Nanotechnology “goes beyond the simple observation of heat” in carbon nanotubes, he says. “It goes to the atomic level of how heat is generated and dissipated.”

In particular, Avouris’s team found that when an electrical current is applied to a nanotube transistor, some atomic vibrations can produce heat of up to 1,000 °C, while other vibrations produce a relatively cool temperature of 400 °C. This is contrary to the behavior of most materials, which maintain a relatively uniform heat.

Moreover, the researchers found that the electrical properties of a nanotube, and the manner in which heat is transferred to a substrate made from the silicon dioxide, are both affected by the vibrations of atoms on the surface of the substrate. This means that the substrate used with nanotube transistors will play an important role in determining the electrical properties of the transistor, and the manner in which heat can be removed.

Since about 1998, when the first carbon-nanotube transistor was demonstrated, researchers have dreamed of next-generation electronics made from such components. Nanotubes have novel properties that allow electrons to zip through them quickly, at low power, and researchers believe that they could act as the active component in transistors, outpacing those made of silicon in terms of speed, energy efficiency, and compactness. But understanding how nanotubes heat up when an electric current is passed through them has been a roadblock to building reliable nanotube circuits.

Mathais Steiner, a researcher at IBM’s nanoscale science and technology group, who also worked on the project, says that in the past couple of years, scientists have turned their attention to the way that nanotubes heat, but this property has been difficult to measure. “The problem is that it’s difficult to probe properties of the active channel [the region of nanotube used as the electrical switch in a transistor] because we’re talking about one molecule,” he says. “People were wondering how to get data and perform experiments. This is the first one to get results.”

To measure the heat of a nanotube, the researchers relied on standard optical spectroscopy techniques. In other words, they shined a tightly focused laser beam onto a single nanotube transistor while passing an electrical current through the target. When the light hits the nanotube, it is absorbed and scattered in ways that reveal the manner in which the nanotube’s atoms are vibrating. These atomic vibrations, known as phonons, occur at different frequencies, or modes. And these modes, explains Avouris, indicate temperatures inside the nanotube.

The reason why the IBM team was able to achieve results that have eluded others, explains Avouris, is its degree of sensitivity. The researchers built their experiment to specifically probe a single nanotube, a process that requires a tight focus of light, and to physically isolate one nanotube from all the others. But simply looking at the results from a single spectroscopy study wouldn’t be enough, says Avouris. So his team looked at a combination of different measurements, or “thermometers,” including the characteristics of electron transitions within the nanotube. Combining these measurements produced enough data to construct a complete heat-flow model.

It has been a challenge to get an accurate view of the thermal properties of nanotubes, especially in relation to a substrate and the surrounding environment, says Eric Pop, a professor of electrical engineering at the University of Illinois, Urbana-Champaign. “The work confirms ideas that had been floating around for a while, but that people weren’t too sure about.” This includes the idea that a nanotube’s environment plays a strong role in determining how it dissipates heat. “Maybe a year ago, it would have been fair to say that nanotubes are excellent heat conductors, but now we know they’re [also] actually quite sensitive to their environment,” Pop says.

For engineers, the IBM findings are significant because they point the way to important heat-management techniques of circuits. Since the substrate has shown to be an important factor in the heat flow and electrical properties of nanotubes, engineers could start to think about modifying them, wrapping materials around nanotubes, or exploring different materials to bond nanotube transistors to a substrate.