IBM and TSMC are already manufacturing the first 2nm chips. These two companies have recently confirmed that they already have this photolithography in place, and while at the moment they are producing these semiconductors solely in a test environment, large-scale manufacturing is not too far off.
In fact, TSMC has confirmed that in 2022 it will start mass production of chips at 3 nm, and in 2023 it will do the same with its 2nm integration technology. This rapid development reminds us that each step we take brings us a little closer to the physical limit of silicon technology, but in reality, this challenge is not new.
The main semiconductor manufacturers and numerous research groups linked to some of the most important universities on the planet have been working for many years in a solution to this challenge. Decades, even.
Currently there are several lines of research open, and possibly the solution will require betting on one of them, but it is even more likely that the way forward will invite us to allow several to coexist of the proposals that researchers are currently working on.
However, we do not need to exaggerate either: silicon technology as we know it is left sane for a while. Although everything goes as planned by TSMC and the mass manufacturing of semiconductors with 2 nm photolithography begins in 2023, both this process and those that will come after it can be further refined, so it is reasonable to expect that they will accompany us at least throughout this decade.
In any case, in this article we suggest you take a look at two of the research lines promising ones that researchers are currently working on. They are not the only ones that are giving us attractive results, and if you are interested in this report we will prepare another that allows us to explore more options, but they are two solutions that invite us to face the future of semiconductors with optimism.
Gallium arsenide semiconductors are promising. A lot of
The gallium arsenide it is a peculiar semiconductor. Even, in a way, daring. And it is that although it is not part of the lineage of elemental semiconductors, among which is, of course, silicon, it has properties that make it very attractive and have placed it in the spotlight of the electronics industry .
For a long time, manufacturers of photoelectric cells and telecommunications equipment, among others, have been forced to share it with consumer electronics brands, so that soon we users will be aware of the impact that you already have, and will have, in our lives.
The elemental semiconductors They are characterized by being made up of a single chemical element, but gallium arsenide (GaAs), as we can guess even though we don’t know much chemistry, is made up of gallium (Ga) and arsenic (As).
In elements with electrical conduction capacity some of the electrons of their atoms, known as free electrons, can pass from one atom to another when we apply a potential difference at the ends of the conductor.
Precisely, this capacity for the displacement of electrons is what we know as electric current, and we all intuitively know that metals are good conductors of electricity. Curiously, they are so because they have many free electrons that can move from one atom to another and, thus, they manage to transport the electrical charge.
Gallium arsenide is a semiconductor, and this implies that under certain circumstances it is capable of carrying electrical charge. When the right conditions exist the mobility of its electrons it is much higher than in semiconductors like silicon or germanium. And this means that its ability to carry electrical charge is also superior.
Another very interesting property of this compound is its high saturation rate. This parameter reflects the maximum speed at which electrons can travel through their crystal structure. This maximum speed is limited by the scattering suffered by the electrons during their movement.
The most interesting and easy-to-understand conclusion of all we’ve seen so far is to accept that, when the right conditions are met, electrons they move more and faster in gallium arsenide than in silicon. And this property has very important repercussions.
One of them is that gallium arsenide transistors can work at frequencies above 250 GHz, which is quite an impressive figure. In addition, they are relatively immune to overheating and produce less noise in electronic circuits than silicon devices, especially when it is necessary to work at high frequencies.
So far we have only investigated the most attractive properties of this semiconductor, which are precisely those in which it has an advantage over silicon. But this last element also has its strengths, and they are important, so the most reasonable thing is to consider gallium arsenide as a complement to silicon, or an alternative to this in certain applications where it is necessary to work at high frequencies.
Carbon nanotubes are (almost) ready for rescue
The first carbon nanotube transistors were produced by IBM more than two decades ago, reminding us that this is not really a new technology. As we can guess, this material is made up of very thin sheets of carbon atoms that adopt a structure with a peculiar tubular geometry.
What makes it so attractive is precisely that it is an excellent semiconductor, which postulates it as an ideal candidate to manufacture high-performance chips that, in addition, on paper should have a very high energy efficiency. The path that this technology has traveled over the last two decades has been arduous, but it has undergone a remarkable development that invites us to view carbon nanotubes with optimism.
One of the most relevant milestones took place at the MIT (Massachusetts Institute of Technology) in 2019. And, as the prestigious scientific journal Nature reported at that time, a group of researchers from this university managed to fine-tune a 16-bit microprocessor made up entirely of 14,000 carbon nanotube transistors.
There is no doubt that this chip is very simple if we compare it with the microprocessors that we can currently find inside our computers and mobile phones, but, even so, it represents a great advance if we keep in mind that the carbon nanotube chip is more complex that had been manufactured only a few years earlier, in 2013, had only 178 transistors.
In addition, the MIT researchers achieved something that further adds to their achievement: to make their carbon nanotube microprocessor they used exactly the same manufacturing technology that is used in the production of current processors. And this means that its technology should be able to be refined and scaled to enable the mass production of more complex chips with relative ease and in the same facilities where silicon semiconductors are produced today.
One of the main obstacles that the manufacture of chips with carbon nanotubes must overcome, and what has prevented them from being an alternative to silicon in practice, is that it is difficult to achieve that this material have the necessary purity. Currently, carbon nanotubes are produced with a purity of 99.99%, and the researchers assure that their manufacture must be refined until it allows us to reach a purity of 99.999999%, thus, with no less than six decimal places.
Outside of science it may seem that the difference between these two figures is miniscule, but it is not. In fact, this difference in purity causes the carbon nanotubes stop behaving like a semiconductor, and become a metal. And, logically, it is a problem if we want to use them to produce chips.
In any case, if we look back for a moment and contemplate where we came from, we can be reasonably optimistic. Silicon promises to give us a few more years of service, and perhaps during this time the researchers will be able to solve the challenges which still raises the development of complex chips using carbon nanotube transistors. Fingers crossed that our expectations are fulfilled.