IBM brings carbon nanotubes a step closer to usurping silicon
Silicon’s reign as the standard material for microchip semiconductors may be coming to an end. Using standard semiconductor processes, scientists from IBM Research have succeeded in precisely placing over 10,000 working transistors made from carbon nanotubes onto a wafer surface – and yes, the resulting chip was tested, and it worked. According to IBM, “These carbon devices are poised to replace and outperform silicon technology allowing further miniaturization of computing components and leading the way for future microelectronics.”
Silicon transistors have been getting smaller and smaller over the years, while simultaneously improving in performance. This has fueled the development of microelectronics, allowing devices to be made smaller, use less power, and run faster. Due to silicon’s physical properties, however, there’s a limit to just how small of an object can be made from it. The IBM researchers believe that we are nearing that limit, and that another material must be used if we don’t want the evolution of our computers, smartphones and other electronics to stall out.
Carbon nanotubes tick many of the boxes for being that new material. First of all, they’re tiny – an individual transistor made from them measures just tens of atoms across. They’re also more electrically conductive than silicon, allowing electrons (and thus data) to travel through them more quickly – computer models have suggested that electronic circuits incorporating the nanotubes should perform about five to ten times better than silicon-based circuits. Additionally, their tube shape makes them “ideally shaped for transistors at the atomic scale,” said IBM.
The company also stated that ultimately, over a billion transistors per chip will be required for commercial electronics. Previously, scientists had only managed to place no more than a few hundred nanotubes on a wafer. Challenges have included finding a way to precisely place them, and ensuring that they are all semi-conductive – ordinarily, in any batch of carbon nanotubes, some end up being semi-conductive while some are metallic.
IBM has apparently found a way around these challenges.
The process begins with carbon nanotubes being mixed with a surfactant (aka a detergent). A specially-prepared substrate is then immersed in that solution. The substrate has a silicon oxide base, with trenches made of chemically-modified hafnium oxide running evenly through it. The semi-conductive nanotubes chemically bond to those trenches, while the rest of the base remains clear. Orderly rows of purely conductive carbon nanotubes are the result, packed onto the base at a density that should allow for approximately one billion per square centimeter.
The researchers claim that the technique could be easily implemented using common chemicals and existing semiconductor fabrication facilities.
Source: IBM Research