Imprimir Republish

Physics

Valuable impurity

Behavior of intruding atoms might mean new applications for semiconducting materials

IF / UFBASuddenly, after calculations that took eight months of work to come to maturity, what seemed to be simply some impurity inside groups of atoms has turned out to be an optical phenomenon that may be able to generate chips (integrated circuits) for smaller and faster computers than today’s, capable of processing information in the form of an electric current or of light. The discoveries made by the team led by physicist Antonio Ferreira da Silva, a researcher from the Physics Institute of the Federal University of Bahia (UFBA), also represent a possibility for the survival of silicon, the material currently used in the chips and the target of intensive research, because its capacity for being compacted is now close to the physical limit.

Through mathematical models, Silva has shown that the behavior that seemed to be just random noise – those impurities – in experimental chips made with gallium nitride (GaN) and gallium arsenide (GaAs) was actually a level of energy – or, as physicists would say, a characteristic pattern of wave frequency – given off by a cluster of three atoms of silicon. It is these atoms, on their own or together, that form the impurities, which for decades have been added to semiconductors with the purpose of increasing the capacity of these materials for transmitting light or electric current. Published in October in Applied Physics Letters, the results indicate that this behavior of the silicon clusters – much more organized than used to be imagined – could be used to process information at quite specific levels of energy, expanding the properties of the semiconductors.

Silva and two students, Jailton Souza de Almeida and Adriano Jesus da Silva, worked with the teams of Clas Persson, Rajeev Ahuja, from Uppsala University, and Patrick Norman, of Linkopings University, both in Sweden, analyzing the experiments carried out at the Naval Research Laboratory, in the United States. It was there, last year, that researchers detected energies of around 19 milli-electron volts in crystals of gallium nitride – and straight away regarded it as just noise, something undesirable and of no importance. At the end of last year, when he reassessed the results, Silva had the impression that it was not simply a question of noise.

Later experiments, carried out at the Naval Research Laboratory itself, proved it: what they had observed – and originally disregarded – was a level of energy compatible with what appeared in a cluster of three atoms of silicon. A cluster, therefore, active in a range of the electromagnetic spectrum corresponding to infrared, which could be manipulated to read, store and transmit information by means of light and an electric current.

Promising material
Gallium nitride and gallium arsenide, which make up the chips in which these phenomena were investigated, are materials that are synthesized in the laboratory, unlike silicon, a mineral found in nature. Researched above all in the United States and in Japan – in Brazil, there are also groups at the University of São Paulo (USP) and at the Campinas State University (Unicamp) -, these materials, in particular gallium arsenide, were already put into chips with special application, made by companies like America’s Lucent Technologies. They do not yet have such an accessible price as silicon, but they are equally semiconductors: like wild cards of the physical-chemical world, they show a mixed behavior when stimulated by an electric current or by light, being able both to conduct an electrical current and emit luminosity.

This property makes the semiconductors the ideal material for the creation of optic-electronic devices, like LEDs (light emitting diodes), detectors and lasers, which use both electrons, which conduct an electric current, and photons, luminous particles, to process information. Furthermore, the semiconductors can be operated at room temperature and emit visible light – the region of the electromagnetic spectrum between red and purple, captured naturally by human eyes. “For this reason”, explains Silva, “the semiconductors can be used to create practically any kind of optic-electronic device”.

But semiconductors also have their deficiencies. The main one of them is that to transmit the information, they need relatively high energy for activation (around 3 electron volts), three times greater than crystalline (pure) silicon takes. This is the point where the silicon atoms used as impurities come in, to reduce the energy needed to start the conduction of light or of an electric current. The structure of the crystals of gallium nitride or arsenide, created artificially in the laboratory, is extremely regular – layer on layer of hexagons or cubes, with about 10²² (number 1 followed by 22 zeroes) atoms per square centimeter.

The quantity of silicon atoms that go in as impurities is far less than in the crystal gallium nitride or gallium arsenide strictly speaking, but even so is something astronomical: about 1015 to 1018 per cubic centimeters. “The impurities help to process information rapidly and using little energy, in the region of milli-electron volts”, explains. As crystalline silicon is a poor emitter of light, the intention now is to combine the best of both worlds: a swift response, when necessary, and the optical properties of the compounds of gallium nitrides and gallium arsenides.

The problem is that silicon may not appear only as isolated atoms in the middle of the crystal chain, but also asclusters. According to Silva, it is difficult to know how these atoms together – which even behave as a molecule – have an influence on the optical and electricity conducting properties of the whole crystal network. Silva’s work with the Swedes is making clearer the distinction between the pure material and the impure, by establishing the difference in the absorption or emission of energy of each kind of material, and, at the same time, showing how the atoms join up inclusters.

Knowing how the clusters work increases the control over the energy emitting process: depending on the purpose devised for the chips, the manufacturer can modulate the properties of the device to be created with the semiconducting material. “Silicon chips could be produced, like those that are in use today, but with areas of gallium nitride, gallium arsenide, or even another kind of silicon – the porous form, which hasa rugose surface – in which luminous information can also be processed”, the physicist suggests.

According to him, the mixed system – an optical and electronic device – would in principle be extremely practical for mitigating the problem of miniaturization with silicon chips: the information processing capacity would grow, no longer by means of reducing the components of the chips, as is the current tendency, but by the possibility of semiconducting materials processing simultaneously light and an electrical current.

Silva is not researching just gallium compounds. His work with Iuri Pepe, also from UFBA, resulted in articles recently accepted by the Physical Review B and the Journal of Applied Physics dealing with the possibilities of controlling the properties of other technologically important, such as silicon carbides or bismuth iodine and antimony, used, for example, in X-ray or infrared detectors.

Republish