Carbon nanotubes-cylindrical carbon atoms with incredible strength and conductivity-hold great promise for creating new micron-sized low-power electronic devices. However, finding a way to build a reliable computing platform based on carbon materials has always been a major challenge facing researchers.

Now, a team of mechanical and materials engineers at Georgia Institute of Technology has devised a method for identifying changes in the performance of transistors made from a network of carbon nanotubes. This new method can help researchers create more reliable devices and ultimately use the technology to deal with various applications such as wearable electronic devices, sensors and antennas.

"The use of carbon nanotubes to manufacture thin film transistors with good performance and repeatability has always been a challenge because of random defects in the manufacturing process," said Satish Kumar, associate professor at the George W. Woodruff School of Mechanical Engineering. "These random defects cause changes in the properties of the nanotubes-differences in length, diameter, and chirality. All of these affect the conductivity of the nanotubes, leading to changes in these properties.

"What we are doing now is to create a systematic method to estimate these changes that can improve the reliability of carbon nanotube network equipment," he said. The results of this study are sponsored by the National Science Foundation.

Although early research has studied how to improve the production method of carbon nanotubes to achieve higher uniformity, Kumar's team focused on analyzing performance changes in a statistical way, so that the performance characteristics are more estimable.

"This kind of analysis is essential for exploring the reliability and stability of circuits based on carbon nanotube networks and designing techniques that help reduce circuit performance variations in various electronic applications," said Kumar in a thesis by Jialu Chen, a graduate student in Georgia Write. technology.

Although some carbon nanotubes conduct electricity in the same way as semiconductors such as silicon, some carbon nanotubes have conductivity more similar to metals. The latter type is called metallic carbon nanotube. The ubiquity of such metallic carbon nanotubes in the network is related to performance problems.

The study found that metallic carbon nanotubes cause more performance changes in thin-film transistors with short channels than in thin-film transistors with long channels, which means that device designers can use long-channel Thin network to achieve higher performance. Thin film transistors.

The researchers also found that as long as the nanotube network is dense, changes in the length of the carbon nanotubes seem to have less impact on performance. "Our research results show that the performance variability of thin film transistors can be reconstructed by the distribution functions of related parameters, which will help us create more reliable circuits to realize the next generation of low-cost flexible microelectronics," Kumar said.