Almost all technologies, although they tend to evolve, eventually become obsolete. This pattern has not been spared, and silicon electronics. It is easy to see that recently its progress has slowed down significantly and has generally changed the direction of its development.
The number of transistors in microcircuits no longer doubles every two years, as it used to be. And today, the performance of computers is not increasing by increasing their operating frequency, but by increasing the number of cores in the processor, that is, by expanding the possibilities for parallel operations.
It's no secret that any modern computer is built from billions of small transistors , which are semiconductor devices that conduct electricity when a control signal is applied.
But the smaller the transistor, the more pronounced the parasitic effects and leaks that interfere with its normal operation, and represent an obstacle to creating even more compact and faster devices.
These factors determine the fundamental limit on the path of miniaturization of the size of the transistor; therefore, a silicon transistor, in principle, cannot have a thickness of more than five nanometers.
The physical reason lies in the fact that electrons moving through a semiconductor waste their energy simply because these charged particles have mass. And the higher the frequency of the device is made, the greater the energy losses in it become.
With a decrease in the size of the element, although it is possible to reduce the loss of energy in the form of heat, it is not possible to prevent the influence of the atomic structure. In practice, the atomic structure itself begins to become a hindrance, since the element size of 10 nanometers achieved to date is comparable in order of magnitude to only a hundred silicon atoms.
But what if you try to use light instead of current? After all, photons, unlike electrons, have neither charge nor rest mass, while they are the fastest particles. In addition, their fluxes at different wavelengths will not interfere with each other during synchronous operation.
Thus, with the transition to optical technologies in the field of information management, there could be many advantages over semiconductors (with heavy charged particles moving through them).
Information sent by means of a light beam could be processed directly in the process of its transfer, and the energy consumption would not be as significant as when transferring by means of a moving electric charge. And to carry out parallel computations, the applied waves of different lengths would allow, and no electromagnetic interference would be fundamentally safe for the optical system.
The clear advantages of the optical concept over the electric one have long attracted the attention of scientists. But today, computational optics remains largely hybrid, that is, combining electronic and optical approaches.
By the way, the first prototype of an optoelectronic computer was created back in 1990 by Bell Labs, and in 2003 Lenslet announced the first commercial optical processor EnLight256, capable of performing up to 8000000000000 operations on 8-bit integers per second (8 teraop). But despite the steps already taken in this direction, questions still remain in the field of optical electronics.
One of these questions was as follows. Logic circuits imply the answer "1" or "0" depending on whether two events occurred - B and A. But the photons do not notice each other, and the circuit's response must depend on two light beams.
Transistor logic, which operates with currents, does this easily. And there are a lot of similar questions. Therefore, there have been no commercially attractive optical devices based on optical logic until now, although there were some developments. So, in 2015, scientists from the laboratory of nanophotonics and metamaterials at ITMO University demonstrated in an experiment the possibility of manufacturing an ultrafast optical transistor consisting of only one silicon nanoparticle.
To this day, engineers and scientists of many institutions are working on the problem of replacing silicon with alternatives: they are trying graphene , molybdenum disulfide, thinking about using particle spins and, of course, about light as a fundamentally new way of transmitting and storing information.
The light analogue of a transistor is a fundamental concept that requires a device capable of selectively transmitting or not transmitting photons. In addition, a splitter is desirable that can split the beam into parts and remove certain light components from it.
There are already prototypes, but they have a problem - their size is gigantic, they are more like transistors from the middle of the last century, when the computer age was just beginning. Reducing the size of such transistors and splitters is not an easy task.
Meanwhile, scientists from the Skoltech Hybrid Photonics Laboratory, together with colleagues from IBM, at the beginning of 2019 managed to build the first optical transistor capable of operating at a frequency of 2 THz and at the same time not requiring cooling to absolute zero.
The result was obtained using the most complex optical system, which was created by long painstaking work of the team. And now we can say that photonic processors performing operations at the speed of light are in principle real, as real as fiber optic communication.
The first step has been taken! A miniature optical transistor that does not require cooling and is capable of operating a thousand times faster than its electronic semiconductor ancestor has been created.
As noted above, one of the fundamental problems on the way of creating elements for "light" computers was that photons do not interact with each other, and the movement of light particles is extremely difficult to control. However, scientists have found that the problem can be dealt with by resorting to so-called polaritons.
Polariton is one of the recently created virtual particles, similar to the photon, and capable of exhibiting the properties of waves and particles. A polariton includes three components: an optical resonator, consisting of a pair of reflector mirrors, between which a light wave is sharpened, and a quantum well. A quantum well is represented by an atom with an electron revolving around it, capable of emitting or absorbing a quantum of light.
The polariton quasiparticle already in the first experiments showed itself in all its glory, showing that it can be used to create transistors and other logical elements of light computers, but there was one serious drawback - work was possible only at ultra-low temperatures near absolute zero.
But scientists have solved this problem. They learned to create polaritons not in semiconductors, but in organic analogs of semiconductors, which retained all the necessary properties even at room temperature.
Polyparaphenylene , a recently discovered polymer, similar to those used in the production of Kevlar and various dyes, has approached the role of such a substance .
Three such transistors have already allowed researchers to assemble the first logic lighting devices that reproduce the operations "AND" and "OR". The result of the experiment gives reason to believe that the road to the creation of light computers - economical, fast and compact - is finally open.
This post was resteemed by @steemvote and received a 5.97% Upvote. Send 0.5 SBD or STEEM to @steemvote
Downvoting a post can decrease pending rewards and make it less visible. Common reasons:
Submit