YORKTOWN HEIGHTS, N.Y. Carbon nanotubes have exceptional electron mobility, but their extremely small size makes it exceedingly difficult to integrate them into transistors. One barrier to commercialization of nanotube-based circuitry is the lack of quick and easy ways to accurately measure their performance.
IBM's T. J. Watson Research Center has demonstrated a characterization technology for carbon-nanotube transistors that is quick and accurate. By focusing a laser on the nanotube that is being tested, IBM measures the shift in Raman phonon frequency to determine whether the nanotube is metallic or semiconducting and to measure its electron- or charge-density. This type of probing method, absent until now, is essential to characterizing nanotube materials capable of commercialization.
"We have found a correlation between a shift in a nanotube's Raman phonon frequency and a change in its charge density," said Phaedon Avouris, an IBM fellow and the manager of Nanoscale Science at the Research Center. "Now we have a quantitative way of finding out what each part of a nanotube circuit is doing, something we didn't have before."
IBM's metrology technique, which is likely to be widely adopted by other nanotube researchers, depends on the close coupling between electronics and lattice vibrations in a semiconductor, known as "phonons." When the laser beam probe hits a nanotube, some of its energy is transferred to vibrations in the lattice, or phonons, thereby shifting its resonant frequency. The electron density in the nanotube also affects the frequency of this coupling because some of the phonons' energy goes into slowing down the electrons, resulting in a higher order object called a polaron: the electron plus its polarization field. The polaron affects the lattice distortions with its polarization field, leaving a cloud of phonons in its wake as it moves through a material.
"Our technique uses the Raman effect, focusing a laser beam on a spot where some of its energy causes the atoms to start vibrating and the rest scatters," said Avouris. "By measuring the slightly different color of the scattered light, we learn about the vibrations: how the phonons are creating polarons."
Usually, the effects of Raman scattering are used to study the vibrations or phonons; but IBM instead used the change in Raman phonon frequency to sense the electron density in the carbon nanotube material.
"We turn it around from phonons affecting electrons to the reverse, where the electrons have an effect on the phonon," said Avouris. "What we found was that the Raman phonon frequency is changed by coupling to electron transitions: the higher the electron density, the higher the Raman phonon frequency."
In its experimental setup, IBM calibrated the relationship between electron density and a shift in the phonon frequency by using a gate to vary the electrostatic doping of a nanotube and measure the resulting change in the Raman frequency. By running all parameters through their full ranges and recording the results, IBM has now created a calibrated tool to measure electron density in any part of a circuit by merely shining a laser on it.
"Now we can measure the variations among different nanotubes as well as the variations along the length of a single nanotube," Avouris said.
As a bonus, the team also discovered that they can tell the difference between a metallic and a semiconducting nanotube by the way the latter distorts the bandwidth of returned scattered signals.
Next, IBM plans to characterize combinations of various materials in search of the optimal combination to harness the high-electron mobility of nanotubes while shielding them from unintentional coupling to their environment. Because nanotubes are so sensitive, extremely careful construction methods must be used to shield them from environmental effects. With its new methrology technique, IBM hopes to characterize environmental effects.
"One of the main differences with nanoscale structures is [that] their larger proportion of surface area (especially with nanotubes, which are just single layer of atoms rolled together) interacts very easily with anything it touches: gases, solids, the substrate you put it [in]; anything in its environment," Avouris said. "Electrically, these environmental interactions have the same process as doping a semiconductor: increasing or decreasing electron distribution. Now we have a way of measuring these environmental effects."
In addition, IBM hopes this noninvasive technique for probing materials' electron densities will also be able to be adapted for use with other semiconductors.