About the race to make the smallest chip in the world Essay

There are already processors smaller than a cell. Can they become smaller than an atom? We tell you the lofty quest for the most extreme miniaturization

If things are made of atoms ... Can you make things smaller than an atom? It is not about philosophy; it is a question that highlights one of the limits that currently faces the accelerated technological development. According to Moore's Law, enunciated by Gordon Moore, co-founder of Intel, in 1965, technological power doubles every two years, and at lower cost. This means an exponential growth in the number of transistors that can be fed into a processor.

Although not a law itself, but rather an empirical trend, Moore's Law has been met with some accuracy. However, it now encounters a physical limit: the size of the transistors is reaching atomic orders: you can’t make transistors smaller than an atom; In addition, the smallness presents / displays other problems related to the quantum nature of the matter or the dissipation of heat. In short: the process of miniaturization and increase of power that has caused us to carry prodigious machines in our pockets is faced with a wall.

To give us an idea of the progress, the physicist Ramón Aguado, a researcher at the Center for Scientific Research (CSIC) proposes a nice metaphor: "In a microprocessor, in the 1970s, there were two thousand transistors: the number of spectator’s theater. Today in that theater we have managed to put one billion people, that is to say, there are a billion transistors in a processor ". ENIAC, built in 1943 to calculate ballistic trajectories and considered the first computer, occupied a whole room and weighed 27 tons. He could make 5",000 sums or 3",000 multiplications per second. Today any smartphone has a power more than a thousand times higher and fits in the palm of the hand. How did this prodigy work?

How small things were made

It should be understood from the beginning. In an abstract plane, the information is reduced to the digits 0 and 1 (hence the digital). In practice those 0 and 1 are translated into electric current (1) and absence of current (0). To build computers it is necessary to be able to handle thousands of small switches that let or not electricity. In ENIAC these switches were vacuum valves, a kind of bulbs that took up a lot of space and melted frequently. Progress had a turning point with the invention, in 1947 and thanks to the physics of the semiconductors (particularly the silicon, from there Silicon Valley), of a new type of switch: the transistor.

With the advent of the transistor begins the process of miniaturization and Moore's Law. At the end of the 1950s, the integrated circuit made it possible to manufacture processors in an industrial way; Today they are manufactured in wafers by the process of photolithography, that is, they are printed with light, similarly to an analog photograph. Each time you can put more transistors in each processor, they are becoming smaller. And more transistors (more zeros and switches), more operations per second, more computing power, more and more advanced technological applications.

Physical limits

In addition to the size of transistors, increasingly close to the size of atoms, in the world of increasingly smaller problems arise. One is the appearance of quantum effects: small-scale matter is governed by other laws, below it is quite likely that a particle crosses a potential barrier. It would be the equivalent of a ball crossing a wall in our world. It is called tunnel effect and is one of the strange qualities of the quantum. "With the tunneling effect the current starts to get less accurate and the operating thresholds start to become worse"," explains Aguado.

Another of the physical limits is heat dissipation. "The smaller the transistor the more powerful the processors, but also generate more heat"," says Aguado, "one of the technological challenges is to optimize speed with respect to heat." Because if there is too much heat the appliance can simply be burned. Hence, computers usually carry a system of fans that produce that characteristic sound and smell.

What solutions are proposed? There are several: the creation of integrated circuits with two-dimensional layers of transistors, the investigation of new materials ("eg semiconductor alloys such as silicon with germanium, or materials such as graphene or phosphorene", notes Viñals), transistors of A single electron (which consume less power) or the parallel use of several processors, as already used. Since the time will come when you will not be able to miniaturize more and thus make the processors more powerful, the technology offers these secondary ways to move forward.

Quantum computing

But the path that seems to imply a qualitative change is quantum computation. In this discipline the possible states are not only 0 and 1, but a superposition of both states, as in the famous case of Schrödinger's cat that is dead and alive at the same time. Thus, from the traditional bit, we move to the qubit. The computing capacity would grow in such a way that it is still difficult to imagine its consequences. At the moment one of the most indicated effects is that a quantum computer could fly through the air all the systems of encryption, for example the one of the credit cards. "To encrypt a credit card, you basically have to break down two factors into a very large number. It is an operation that a normal computer could take thousands of years, but a quantum computer could perform in a reasonable time, "explains Joaquín Fernández Rossier, researcher in nanoelectronics at the University of Alicante and the International Iberian Nanotechnology Laboratory (INL). "But it will not be faster to face normal problems"," he adds, "but they will be able to efficiently solve what normal computers can’t solve."

All major technology companies have already positioned themselves on quantum computing and are working on their development. The IBM Q division of IBM introduced in May 2016 its first prototype processor, consisting of five qubits. Google and Nasa are also working on them (with the participation of researchers from the University of the Basque Country). Canadian company D-Wave Systems already sells quantum computers: its D-Wave 2000 Q model, available since June and valued at 15 million dollars, has two thousand qubits. For its part, the European Commission announced last year the launch of a program of 1 billion euros dedicated to the development of this computer, the Quantum Flagship. They call it the Second Quantum Revolution. "They are long-term projects, we will not have a quantum computer at home in five years"," says Fernandez Rossier, "quantum computers are still large and very sophisticated: they need superconductors, temperatures near absolute zero, etc."

But it's not enough?

The question that can be assaulted in the face of the need for technological progress to continue is the following: is not the technological level that humanity has reached enough for the moment? Is it necessary to maintain exponential growth as prodigious as that which has been experienced over the past 50 years?

How to cite this essay: