Nanomemory – Devices Based on Nanotubes and Nanowires

by | Nov 22, 2010

French scientists have made light-sensitive memory devices by combining carbon nanotubes and silicon nanowires.

French scientists have made light-sensitive memory devices by combining carbon nanotubes and silicon nanowires.

As computer chips become ever smaller and more complex, scientists are interested in carbon-based components for devices such as memory chips. Nanoscale memory devices are key elements of future nanoelectronics circuits. For improved performance, the down-scaling of both the size and the programming speed are equally important.

Carbon nanotubes have already been used to make field-effect transistors (CNTFETs). These transistors are very sensitive to their environment and are thus also of interest as sensors. In a recent paper, French scientists led by Vincent Derycke at Saclay have shown that they can make very efficient CNTFETs that work as light sensors and memory elements together. “To be useful, devices based on nano-objects need to bring more than their small size; they need to show extra functionality when compared with conventional transistors” says Derycke.

Their devices combine thin silicon wires as gate electrodes, individual carbon nanotubes as a semiconducting channel, and a polymer layer as a light-sensitive material. Electron-hole pairs are generated in the polymer in response to light and, once separated and trapped, cause the device to switch on. The gate electrode is then used to control the amount of trapped charge and thus to adjust the memory level.

The programming speed of such devices increases dramatically when the thickness of the gate oxide is decreased. The team has used their system to study scaling rules and the nanoscale charge-trapping mechanism for gate oxides. The group believes that the information gained from learning about their current system will enable programmable circuits with high tolerance to defects and variability to be built. “Conventional circuit architectures are not adapted for future nanoscale circuits. We believe that nanoscale devices with memory capabilities would be better exploited in an adaptive circuit with learning capabilities well suited to tolerate imperfections” says Derycke.

Related posts: