[Image Credit – Phys.org]
For decades now, all of our computers have run on silicon processors and memory chips, the capacities of which have increased exponentially every year according to Moore’s Law. However, the physical limitations of silicon are about to be reached, limiting the potential progress of silicon as a conductive material around the year 2020. This fast-approaching deadline has sent scientists and engineers scurrying to find a new material to replace silicon as their semiconductor of choice. And that new material looks to be everyone’s favourite trope of science fiction meets science fact: carbon nanotubes.
What are Nanotubes?
Carbon nanotubes, first discovered in 1991 by researcher Sumio Iijima at NEC (Japan), are one atom thick layers of graphene rolled up into a tube that measures 1.2 nanometers in diameter. They are the fourth form of carbon after graphite, diamonds, and fullerenes. They were immediately hailed as a possible replacement for the channel in silicon transistors, conducting electrons over 70 times faster than standard silicon chips, making them one of the most conductive materials ever discovered. However, there are many logistical barriers that must first be overcome before carbon nanotubes can be used in computers on a commercial level.
When nanotubes are made, because of their chirality, a portion of them comes out metallic instead of semiconducting, which can cause transistors to short out. As a result, finding a reliable way of separating out the metallic nanotubes is essential. A lot of progress has been made on this front, but the challenge remains. Another issue is placement: the development of a reliable, non-lithographic way to place billions of nanotubes on a chip exactly where they are needed. The final challenge is the performance of nanotubes. Until now, silicon transistors have outperformed their carbon cousins at every turn – but a recent finding shows that that is about to change.
A New Hope
For the first time ever, scientists have built a carbon nanotube transistor that performs better than a silicon one. This is big news for several reasons. First, as mentioned above, there is a hard physical limit on the number of silicon transistors that can be fit on a chip, so nanotubes can make electronic devices even smaller than they are now. Second, the increased speed of conduction means that information can be transferred more quickly inside the chip, giving future nanotube chips even faster clock speeds. And third, because carbon is so much more conductive than silicon, devices will use much less power to produce the same result. Great for cool, battery sipping phone CPU chips. It’s little wonder that producing working carbon nanotube transistors is the holy grail of the electronics industry.
In theory, computing speeds could be up to five times faster than they are at the moment, or use five times less energy. This means longer lasting phone batteries, or much faster wireless communications and processing speeds. A fully working nanotube computer has yet to be built, however, so until one is perfected it’s hard to say for sure. But so far, the signs look promising.
Assuming that the problems with producing nanotubes can be sorted out, the next step is increasing manufacturing efficiency. So far, transistor-holding wafers up to 2.5 cm by 2.5 cm (1 inch by 1 inch) have been produced, and scaling this process is essential. Another issue is electromagnetic interference: it is currently unknown how much the extra conductive nanotubes will need to be shielded to prepare them for commercial rollout. However, once the wafers have been scaled up, and working prototypes have been built, this should be straightforward to determine.
In the future, even carbon nanotubes will become obsolete, so it’s worth paying attention to other materials too. 2-D black phosphorus is an alternative to nanotubes and graphene, and has the advantage of having a bandgap. It works well alongside silicon photonics devices, which are also skyrocketing in popularity. 3-Molybdenum disulphide is also a contender for the next generation of semiconductors, because of its novel stacking properties: a single layer of it is a normal semiconductor, but add a second layer and it becomes piezoelectric. There are definitely exciting times ahead in the world of semiconductor technology, and it’s well worth keeping abreast of the latest news.