Temperatures taken in the realm of the tiny

To take the temperature of very small things, scrap the thermometer. Aluminum, silicon and other materials can serve as their own thermometers, researchers report in the Feb. 6 Science, enabling temperature readings of objects nanometers in size.

The researchers probed the temperature at various points on a tiny aluminum wire by measuring electrons that were fired through the wire. The electrons’ energies exposed subtle changes in the aluminum’s density, which corresponds to the temperature. If the technique passes further scrutiny — and not everyone believes it will — it could be used to measure the temperature of individual transistors.

Temperature, which is related to the energy distribution of particles in a substance, is typically deduced indirectly by measuring another property. In the macroscopic world, that’s pretty easy to do: A bulb thermometer, for example, infers temperature from the expansion and contraction of a liquid (traditionally mercury) as it is heated or cooled. But even specialized thermometers present complications at small scales because they transfer heat to or from the object being measured, skewing the sensitive measurement.

Physicist Matthew Mecklenburg of the University of Southern California in Los Angeles and colleagues wondered whether certain materials could expose their own temperatures. Thinking about bulb thermometers, the scientists knew that heat-driven expansion of a material should also cause the material’s electrons to spread out from each other. So the researchers set out to quantify the change in density and electron spacing and to use it to determine temperature.

They fired a beam of electrons at an 80-nanometer-thick aluminum wire placed atop a roughly 10-nanometer-thick silicon nitride plate. As the electrons passed through the sample at about half the speed of light, Mecklenburg says, they disturbed the aluminum’s resident electrons just as a boat zooming through a still lake disturbs the water. By measuring the energies of the electrons exiting the material, the researchers determined the frequency of those waves of disturbance. That frequency is determined by the electron spacing, which in turn depends on the temperature.

The self-thermometer method enabled the researchers to determine the spectrum of temperatures along the wire as a current heated it up (see diagram). In another experiment, the scientists heated one side of the wire and charted the progression from hot to cold. And finally, when the researchers compared their temperature reading of a small aluminum wafer with that from a commonly used but less spatially sensitive thermometer, they found the readings agreed to within about 10 percent.

David Cahill, a materials physicist at the University of Illinois at Urbana-Champaign, says the technique has crucial fundamental flaws. “I think this is absolutely wrong,” he says. He argues that there are factors other than electron density that determine the temperature of a particular spot on a material. He is also concerned that the aluminum can’t expand freely (and thus become less dense) because it is sitting atop the silicon nitride plate. Mecklenburg responds by saying there is no evidence that any other factors influence the temperature reading.

Despite his concerns, Cahill says that an improved version of the technique could benefit materials scientists, who would be able to probe a sample’s temperature as they manipulate it under an electron microscope. Mecklenburg says he and his colleagues have bigger plans. New experiments reveal that silicon, the material central to electronics, works nearly as well as aluminum as a self-thermometer. Over the years, computer chip manufacturers have shrunk silicon-based transistors to fit more on each chip. The trade-off is that electric current tends to spill out of the transistors and escape as heat. By probing the temperature of individual transistors, engineers could potentially determine where and how the heat escapes.