I have been following carbon nanotubes for over 20 years. They looked very promising from the very beginning as a substitute of silicon for chips. Carbon nanotubes can create much smaller transistors, that in turns would result in faster switching, lower power consumption and higher density >>> hence faster processing.
Over these twenty years researchers have been able to demonstrated the feasibility of creating a transistor using carbon nanotube with the first single bit processor based on carbon nanotubes going back to 2013.
However, one thing is to demonstrate that you can create a transistor using nanotubes, quite a different story is to show that you can create a chip containing thousands of transistors (current silicon chips contain billions of them), and quite another different story is to find an industrial process (that can deliver volumes at an affordable price).
Now researchers at the MIT have announced the construction of a microprocessor chip completely based on carbon nanotubes.
It contains over 14,000 carbon nanotube transistors, that is huge although it is nothing if you compare to silicon chip today. It is actually au pair with silicon chips that were produced in 1980.
To be able to create a functional chip the researchers had to overcome a number of hurdles that have blocked the progress so far. One is the tendency of nanotubes to cluster and form lumps, losing the desired architecture. To avoid this the researchers subjected the wafer to vibrations that untangled the nanotube. Another stumbling block so far has been that along with carbon nanotube you also get a few metallic nanotubes and these disrupt the flow of electrons. The workaround was to design specific architectures that are not affected by the presence of a few metallic nanotube.
All in all, we can see that they have managed to create a carbon nanotube chip. The next step, and not an easy one, is to find an industrial process to manufacture them. This is crucial and it is the starting step for a continuous increase in performance, what happened with silicon (the Moore’s law). It will likely take several more years but I am pretty sure that the way will be found, also because we have reached the end of the silicon path.