Imprimir Republish


Race for the chip

Brazil experiments new ways for reducing its dependence on abroad in the semiconductor area

eduardo cesarSemiconductor components produced in Brazil by Aegis: 30% goes abroadeduardo cesar

Present in every kind of electronic device, from the television set to the microwave oven, from the cell phone to the fuel injection system in automobiles, besides, of course, in computers, semiconductors have a characteristic that is always implicit in its description: evolution. Not that other products from the world of electronics or biotechnology, for example, are out of the evolutionary stage, but with semiconductors evolution happens as minutes pass.

The demands for miniaturization, for speed in data processing, from the advance of software, and for memory capacity make these little silicon pads one of the supports of present-day civilization. Such importance is also measured by the volume of researches in this area. Semiconductors are constant targets for researchers from universities and from the whole electrical-electronic and information technology industry world-wide, which is always alert to the demands from market in the near and distant future.

In Brazil, although we do not have a large-sized semiconductor industry with cutting edge products, which led the country to spend US$ 2 billion in 2003, according to figures from the Brazilian Electrical and Electronic Industry Association (Abinee), research often surprises. Here, it is carried out fundamentally at the universities and is already gathering good fruits.

Among the most recent practical results, we find computer memories 250 times more powerful that are being analyzed by two multinationals, a microsensor of blood pressure for surgical operations, besides several alternatives to the compounds of silicon (Si), the main material used in the manufacture of semiconductor components (materials with an intermediate level of electrical conductivity, between conductors, like copper and other metals, and insulators like rubber and ceramics, which make better use of electrons in processing information).

Within Brazilian academic research, besides São Paulo, important semiconductor study fronts were also opened up in the 1990’s, in Rio Grande do Sul, in Santa Catarina, and in Pernambuco. The University of São Paulo (USP), for example, concentrates most of the studies in this area on the Polytechnic School (Poli) and on the Physics Institute (IF). At the Poli, the researches of the Integrated Microsystems Division (DMI), coordinated by Professor Nilton Itiro Morimoto, have already brought several results. “We have developed a disposable blood pressure sensor to monitor patients under surgery that resulted in a small start up company”, Morimoto says.

The system, which uses the technology for MEMs (Micro-Electro-Mechanical Systems), semiconducting devices known as micromachines, is assembled on a ceramic substrate and connected to the patient and to the electronic monitoring apparatuses. Located in São Paulo, the company is Torr Microssistemas, which was funded by FAPESP’s Small Business Innovation Research Program (PIPE) and should, in three months, start distributing the product in a market whose similar sensors are imported.

The production of Professor Morimoto’s group can be measured by the number of papers published annually. “We are responsible for about 60% of the papers presented in the area of microelectronics processes at the Symposium on Microelectronics Technology and Devices, an international meeting organized annually by the Brazilian Microelectronics Society (SBMicro) and by the Brazilian Computing Society (SBC). For Morimoto, the incentive from the new industrial policy for creating a semiconductor industry is a praiseworthy initiative.

“A semiconductor factory in Brazil means that the country will have the capacity for adding high value to its products, just by putting a minimum of electronic intelligence in them It means that the country will be able to export more and import fewer electronic products and components, and at the same time, creating thousands of jobs of an extremely high level, both in terms of salary and intellectually”, he says.

A similar opinion is held by his colleague Inés Pereyra, the head of the Electronic Systems Engineering Department of the Polytechnic School. For Inés, mastering this technology is a strategic issue, because a great deal of the technological development of the last few decades is based on it. “We could say, without exaggerating, that the economic independence of any country involves mastering semiconductor technology, and this, evidently, calls for investments in research and the existence of factories in the country”, she says. The researcher is coordinating a thematic project that has as its focus research into new semiconducting and insulating materials and the development of micro and optoelectronic devices based on these materials.

Inés’s researches involve producing and studying thin films of materials like silicon carbide (SiC) and silicon oxynitride (SiOxNy), produced with processes involving low temperatures. Silicon carbide is an alternative semiconductor to silicon in the manufacture of devices that work in environments with high temperatures, that are chemically aggressive, or have high radiation, such as sensors for the aerospace and steel industries. Silicon oxynitride is an insulator than can replace silicon oxide in many applications. “We tried to improve the properties of these two materials, to develop micro and optoelectronic devices, such as transistors, light emitting and detecting diodes and optic guides (used in telecommunications via optical fiber)”, says Inés.

Besides applied research, the professors from the Poli also do theoretical research. This is the case of physicist João Francisco Justo Filho, who is using computer simulations to investigate the electronic, structural, and optical properties of semiconducting alloys. According to the researcher, these simulations are low cost tools that can assist in the development and manipulation of materials, and, in turn, in the creation of new devices and production processes. “Our systems for study are made up of sets of atoms, fundamental elements in the context of materials science. It is important to understand their microscopic properties, such as the type of interactions between them and their neighbors, because it is they that determine a major part of the macroscopic properties of materials.”

One of the lines of research of Justo Filho, who has a project under FAPESP’s Young Researcher Program, is aimed at the development of software for simulating nanostructured materials, particularly silicon nanostructures. The researcher’s first step towards creating it was to develop a computer code that made it possible to investigate the thermal properties of materials. The next stage was to create software for visualizing the simulations, where films can be constructed, showing the temporal evolution of the atoms. “One of the simulations that we carried out successfully was to observe the behavior of a silicon nanowire being submitted to a constant tension under certain temperature conditions”, the researcher states. “In my view, silicon nanowires are the next step in nanotechnology. They will be able to be used in optoelectronic applications (lasers, mainly).”

Growth of nanofilms
Equally focused on semiconducting nanostructures is physicist José Roberto Leite, the coordinator of the New Semiconducting Materials Laboratory (LNMS), of the Physics Institute of USP in São Paulo. “We are working on the growth of nanofilms and characterizing and applying them in nanoelectronic devices like LEDs (light emitting diodes), LDs (laser diodes), sensors, and detectors. These devices are of great importance in optic recording (CDs and DVDs) and in telecommunications, amongst other areas. “We developed new LDs and LEDs of great interest, because they will be able to replace, in the future, household incandescent bulbs, with more luminosity and lower spending of energy.”

Another novelty, guaranteed by a patent and intended to improve computer memories, has been transformed into an industrial negotiation, as Pesquisa FAPESP showed in its issue No. 97. Up until now, two large multinational companies, which prefer not to reveal their names, have contacted Professor Elson Longo, from the Federal University of São Carlos (UFSCar) and the coordinator of the Ceramic Materials Development Multidisciplinary Center (CMDMC) and asked for more information about the production process and the new formulation of a chip potentially capable of increase computer memories 250 times, based on a barium and lead titanate compound. “They contacted us and took reports to their head offices, and now they are analyzing them”, says Longo.

At Unicamp’s Physics Institute, the quest for new semiconducting materials is also in the focus of the researchers. The Device Research Laboratory (LPD) is working on synthesizing new materials and on processing and characterizing optical devices and systems. “We are researching into semiconductors made from elements like gallium, phosphorus, nitrogen, and antimony, amongst others, which are more suitable for manufacturing optoelectronic devices”, explains physicist Mauro Monteiro Garcia de Carvalho. These devices are used mainly in the manufacture of lasers, light amplifiers, solar cells and light emitters and detectors, much used in the telecommunications sector, in the storage of data, as in CD and DVD devices, in electronic displays, and in laser equipment for medical and dental applications. “We are in the forefront of research, both in synthesizing and in developing new semiconductor lasers.”

Created in 1974 as the Electronics and Devices Laboratory (LED) and restructured in 1993, when it gained its current name, Unicamp’s Semiconductor Components Center is one of the few Brazilian microelectronics laboratories that are carrying out complete processes for making integrated circuits. “We are doing research and development in devices with CMOS (Complementary-Metal-Oxide-Semiconductor) technology, which accounts for over 85% of the chips made at the world level”, says Jacobus Swart, a coordinator of the center and a professor of Unicamp’s School of Electrical Engineering. Several projects have already been carried out.

“We are also researching into micromanufactured sensor technology, into MEMs, which are called micromachines and are used as sensors of pressure, gases, and radiation.” MEMs are semiconductor devices that typically measure less than 100 micrometers or micra (1 micrometer is equal to a thousandth of a millimeter) and can be used in various industries. “They are sensors made from microelectronic processes. By reducing their size, we also reduce their cost, we improve their performance, and we increase their reliability”, he explains.

For Swart, overcoming Brazil’s technological backwardness in this area will not be an easy task. He believes that the government will only succeed in creating a competitive semiconductor industry in the country if economic conditions are created, enough to attract investments from abroad, and incentives are granted to smaller companies that may be able to grow in specific market niches. Another aspect, according to him, are the human resources. “Without specialized personnel, this industry cannot work. In Brazil, there are only about 500 researchers with direct experience in silicon semiconductors. It is a low figure for our country. It would be ideal if we had at least double that.”

The government’s objective is to stimulate native entrepreneurs to invest in this area and to attract direct investments to the country, with the installation of multinationals from the sector, like the major manufacturers of integrated circuits, also called chips, fundamental for the workings of all electronic equipment. “The semiconductor industry is crucial for innovation, for all the sectors that are growing most in the world: information technology, telecommunications, and entertainment”, summarizes Sérgio Bampi, the president of the Brazilian Microelectronics Society and the coordinator of the multidisciplinary program in Microelectronics at the Federal University of Rio Grande do Sul (UFRGS).

Going into this competitive sector, dominated by such countries as the United States, Germany, Japan, Ireland, Korea, and Taiwan is easily explained. The world semiconductor industry shows some gigantic figures and is the industry that is growing most in the world. According to the Semiconductor Industry Association (SIA) of the United States, the sector’s global sales reached US$ 166.4 billion last year – in 2000, at the height of the bubble of the information technology and Internet companies, they came to US$ 204 billion. In the last 20 years, the segment has had an average annual growth of 16%, against the 3 to 4% of the world economy in general. In Brazil, according to Bampi, the semiconductor market grosses around US$ 3.5 billion a year, adding together the imported components in isolation, those added to ready-made imported products, plus the smuggling market.

Regardless of the interest, or otherwise, of the multinational manufacturers in producing chips in Brazil, in Porto Alegre, in Rio Grande do Sul, the Center of Excellence in Advanced Electronic Technology (Ceitec), is emerging, funded by the Financier of Studies and Projects of the Ministry of Science and Technology, state and municipal governments, besides companies. Still in the project stage, the institution will be a nucleus specialized in carrying out projects and the manufacture of prototypes of integrated circuits, particularly those endowed with CMOS technology.

To do so, the prototyping center will have a class 1000 clean room (with a concentration of particles in suspension in the air of less than 1000 particles per cubic foot) of 800 square meters, with internal class 100 environments (less than 100 particles per square foot) and class 10. A support building will also be built, which will include the center’s administrative sector, the greater part of which will be set aside for carrying out projects in integrated circuits and the training of human resources.

While Ceitec has not begun to operate, the Federal University of Rio Grande do Sul has several research projects for the development of semiconductors and chip engineering. The institution’s postgraduate program in microelectronics, for example, is working on an integrated circuit project, on the development of devices with an MOS (Metal Oxide Semiconductor) structure, of software for integrating chips, and of the so-called hardware and software embedded system-on-chip, regarded as one of the future paths of the chip, with memory, processors and activators of data input and output in the same device.

One of the most important studies, coordinated by Professor Israel Baumvol, has tested alternatives to silicon oxide (SiO2) in the production of chips, in particular with an oxide, silicates and aluminates of hafnium, a metallic chemical element (see Pesquisa FAPESP No. 82). “My research activity is closely related to the production of semiconductor components for the next ten to 20 years”, Baumvol says. “If Brazil has a semiconductor industry, our work will be of vital importance. If we do not have one, the only beneficiaries of our researches will be the major international manufacturers”, says he.

Discrete components
At the moment, there are only four factories working in the productive chain of semiconductors installed in Brazil: Aegis, Semikron, Heliodinâmica and Itaucom. The first two produce the so-called discrete power components – like diodes and thyristors –, which are simpler than integrated circuits. “Diodes and thyristors are devices that work like a key, letting electric current pass or blocking it”, explains engineer Wanderley Marzano, the director-president of Aegis, located in São Paulo. “Our devices are made of silicon and are intended for the capital goods industry.

They are used in sources of direct current that are make up automation and rectification equipment (transforming alternating current into direct current), amongst others.” Aegis’s monthly production is about 6,000 wafers, but it could be far larger. “I do not use even 25% of my installed capacity”, says the businessman, who complains about the instability of the Brazilian market and of the difficulties in selling to customers abroad. Even so, 30% of his production is exported to a dozen countries, including the United States, China, Taiwan, Germany, Italy, and France.

Heliodinâmica, in turn, produces mainly solar cells (which are made of silicon), and Itaucom works in the assembly of integrated circuits. The components come from abroad in the form of a processed silicon wafer, and the company carries out their assembly and testing for putting them on the market. Although it is very limited today, the Brazilian semiconductor manufacturing complex has been bigger in the past. At the beginning of the 1970, the country had a world class laboratory, the Microelectronics Laboratory of the University of São Paulo (USP), which was close to the state of the art in integrated circuit research. And in the 1980s, there were 23 companies installed, the majority belonging to major international groups.

Changes to the course
“The opening up of the market carried out by the Fernando Collor administration made the domestic companies close their doors, and the foreign ones leave the country”, explains José Elis Ripper Filho, the director-president of Asga, located in the city of Paulínia, in the state of São Paulo, which in those days made semiconductors and today produces advanced equipment for telecommunications via optical fiber. With the end of the trade barriers, the makers of the end goods started to import kits, ready for assembly in Brazil.

“Accordingly, the components purchase started to be done abroad”, Ripper says. “Importing a complete kit came to be more advantageous for the end assembler, since it makes possible a reduction in the cost of its own engineering and simplifies the supply chain”, points out the document of the National Microelectronics Program – A Contribution for Formulating a Structured Action Plan, produced by the Ministry of Science and Technology, in December 2002. A way will have to be negotiated, if cutting edge semiconductors start being produced in Brazil again.

Inside the productive chain

The factories that make up the productive chain of integrated circuits can be divided, in a simple manner, into three categories: the design houses, responsible for designing the circuits, the silicon foundries, which handle the manufacture properly speaking, which comprises the physicochemical processing of the circuits that make up the so-called front end stage, and the companies in charge of assembling, encapsulating, and testing the product, which make up the back end stage. Of them all, the foundries, precisely the ones that do not exist in the country, are the ones that add most value to the product.

The amount of the investment for setting up these factories is very varied. According to a study carried out by the Ministry of Science and Technology, at the end of 2002, the cost of setting up a design house would vary from US$ 1 million to US$ 5 million, with the funds concentrated basically in software, training and workstations. Foundries, in turn, have a far more diversified cost, depending on the area of work. Factories specialized in the prototyping of small series, such as the production of simple CMOS devices on a small scale, call for relatively small investments, from US$ 10 million to US$ 100 million.

Then there are the factories that are suppliers for specialized segments, like automobile components, sensors, transceptors (that send and receive radio signals) and microelectromechanical systems (MEMs), costing from US$ 300 million to US$ 600 million. “These foundries comprise the majority of the wafer factories in the world and allow for a lot of innovation in electronic products”, says Sérgio Bampi, the president of the Brazilian Microelectronics Society. The third category of these companies is made up of megafactories, like the American ones Intel and AMD, which produce mainly microprocessors and memories with cutting edge technology. The cost of implanting them is extremely high, ranging from US$ 1 billion to US$ 3 billion.

The projects
1. Production, Characterization and Applications of Semiconductor and Insulating Alloys (nº 00/10027-3); Modality Thematic Project; Coordinator Inés Pereyra – USP; Investment R$ 287,049.00 and US$ 399,205.00
2. Experimental and Theoretical Study of Epitaxial Semiconductor Nanostructures Derived from Compounds III-V (nº 98/12779-0); Modality Thematic Project; Coordinator José Roberto Leite – USP; Investment R$ 416,700.00 and US$ 502,219.00
3. Theoretical Modeling of Electronic and Structural Properties of Semiconductor Alloys (nº 00/11438-7); Modality Young Researcher; Coordinator João Francisco Justo Filho – USP; Investment R$ 71,547.00 and US$ 45,449.00
4. Complete Encapsulation of Disposable Transducers of Blood Pressure (nº 01/08711-6); Modality Small Business Innovation Research Program (PIPE); Coordinator Edgar Charry Rodriguez – Torr; Investment R$ 155,100.00
5. Optoelectronic Integration Platform Based on Selective and Non-Selective Epitaxial Growth by Epitaxy of Chemical Beams (nº 98/14560-6); Modality Thematic Project; Coordinator Mauro Monteiro Garcia de Carvalho – Unicamp; Investment R$ 539,193.00