guia do novo coronavirus
Imprimir Republish

Computing

Changes in sight

Hafnium oxide is a candidate for replacing silicon oxide in the production of chips

The quest for a new raw material capable of replacing silicon oxide (SiO2) in the production of the chips embedded into computers and in all kinds of electronic equipment is at the center of studies by Brazilian researchers. A group from the Physics Institute (IF) of the Federal University of Rio Grande do Sul (UFRGS), coordinated by Professor Israel Baumvol, has tested alternatives to SiO2 for international giants from the information technology sector, like IBM, Texas Instruments, Nortel and Motorola. The biggest bet has been placed on hafnium oxide, based on the metal of the same name.

The concern exists because the era of silicon is threatened to reach its end, after over four decades and having produced fundamental technological advances in people’s daily lives, generating wealth and economies of scale in production. According to the famous “law” known in information technology circles as Moore’s Law (Gordon Moore, the founder of the microprocessor company Intel), processors double their capacity every two years, integrating a far greater number of circuits in just one chip. This situation is reducing so much the thickness of the insulating layer of silicon oxide that very soon there will be no way of preventing leakage of current in the integrated circuits. At that point, the reliability of electronic equipment will be jeopardized and the information technology sector may face a period of stagnation.

Today, a chip’s insulating layer measures from 2 to 2.5 nanometers, and it urgently needs a replacement for SiO2 that can hold the current and avoid overheating, in concentrations that are not so thick. It will also need to become economically viable and to perform this role for at least ten years. o find this replacement, the industry launched itself into a veritable race, some four years ago, mobilizing the main research centers in the semiconductor area in the United States, Europe and Japan, to reach this objective.

Baumvol’s study with several candidates to a replacement for SiO2 indicates that Hafnium oxide (HfO2) may win this position. In the laboratory, it is a material that has proved to stand up to the task of keeping the integrity of the silicon, which does the work of storing information, while the SiO2 acts as an insulator.

Stages of production
The replacement of the material, though, implies significant changes in a good part of the process of producing chips, which consists of 20 to 30 stages. “Altering one of the first stages of the process can affect all the subsequent ones, calling for investments in equipment”, he explains. Besides this, none of the possible replacements for SiO2 is to be found in profusion in nature, and none of them is cheap. “Hafnium oxide, for example, is never to be seen in a pure state, as it has at least 3% of zirconium in its composition, and this calls for new technologies for purification”.

Aware that he is in a “game worth billions” – the company that patents the first effective new technology will make a lot of money by licensing production -, Baumvol does not however see any prospects for Brazil benefiting, in the short term, from the advances at the IF. “Making broader use of the knowledge built up here, that has tripled our productivity in research in the last three years and transformed us into a international point of reference in the area of semiconducting materials, depends on the formulation of national policies with the objective of producing chips in the laboratory”, he says.

Republish