Engineers at Stanford have created the first working carbon nanotube computer - validating the technology as a potential replacement for silicon semiconductors in electronics.
Max Shulaker, doctoral student in electrical engineering at Stanford, holds a wafer filled with carbon nanotube processors. To his left, a basic CNT computer using this technology is sandwiched beneath a probe card. Image credit: Norbert van der Groeben/Stanford University
Researchers have been working for many years to develop new technologies that will enable the march of progress dictated by Moore's Law to continue beyond the practical limits of silicon-based electronics.
Carbon nanotubes have been one of the favourites in the race, amongst other advanced materials like graphene, quantum dots and ferromagnetics.
Whilst many of these technologies have been used to make individual transistors, very few fully functional systems have been demonstrated. Manufacturing the materials on a large enough scale, and controlling the layout and operation of the chips themselves, have been the major stumbling blocks.
Now, a Stanford team have announced a method to create a processor from carbon nanotubes, and have been able to run a basic operating system on their proof-of-concept device, swapping between counting and number sorting calculations.
This is a strong indicator that carbon nanotubes may take over from silicon as the semiconductor industry's material of choice in the future - once the shortcomings of silicon's poor nanoscale heat dissipation and problems with lithography techniques it impossible to go any smaller.
Semiconductor giant Intel's latest statements hint at a roadmap down to a 7nm transistor size. It is not clear whether silicon can be taken any further than this - carbon nanotubes, on the other hand, could allow feature sizes a full order of magnitude smaller, along with huge benefits in power usage and heat management. Image credit: Norbert van der Groeben/Stanford University
Individual carbon nanotube transistors were first fabricated about 15 years ago. The challenge since then has been to manufacture CNTs on a large scale whilst maintaining the rigorous quality standards needed for semiconductor applications.
A single chip could potentially contain billions of carbon nanotubes, which must be laid down in straight lines - and any defects in their structure or alignment could render the entire device unusable.
The Stanford team devised an ingenious quality control method, to ensure that every CNT on the chip is perfect, using an approach they call "imperfection-immune design" - simply working around inherent errors in the manufacturing process.
The two main fabrication defects to be dealt with are misaligned nanotubes, and metallic nanotubes, which conduct electricity much better than their more desirable semiconducting counterparts, but also cannot be switched off - making them useless as transistors.
This very property was used to remove them. The researchers switched off all the semiconducting nanotubes, which had no effect on the defective metallic tubes. The circuit was then subjected to a very high current, which caused the conducting tubes to burn up from the sheer amount of electricity passing through them - neatly removing them from the circuit.
The misaligned CNTs were not so easily dealt with - however, the engineers were able to develop a sophisticated circuit design algorithm, which is able to work around the defective nanotubes without even needing to know where they are.
Computers based on advanced nanomaterials like CNTs often seem like science fiction. With the furious pace of progress dictated by Moore's Law, however, these technologies will begin to appear in the commercial realm much sooner than we might think.
This incredible achievement by the Stanford team has demonstrated the viability of a CNT-based computer. It is now up to the semiconductor industry to take up the baton and get the technology ready for the real world in time for silicon's inevitable expiration date, whenever that may be.
Sources & more information