DRESDEN, Germany Researchers at a nanotechnology symposium here said they are attempting to use the technology to plug gaps in the roadmap for future semiconductor production.
Meanwhile, policy makers are urging researchers to focus nanotechnology research on practical applications as they seek to faster return on government investments.
Infineon Technologies AG (Munich, Germany) is developing nanotechnology applications for communication technology. Karl Joachim Ebeling, an Infineon researcher, described an SiGe transistor that could be used to control semiconductors with a bandwidth of 60 GBit/s. But he said designers still don't yet know what the components full capacity. "That's hard to measure in areas like this," Ebeling said. What is certain, however, is that the SiGe transistor offer higher voltages for switching that are "at least as good as with a CMOS," he added.
While Infineon is using up to 45-nm node planar MOSFETs on ultrathin silicon-on-insulator (SOI) wafers, the company said it is looking at nonplanar FinFET structures on SOI for sub-30-nm nodes. Ebeling said Infineon is working on FinFETs with high on-off current ratios. The company intends to use the devices in its FinFET memories the first component of this type in the world, as Ebeling claimed.
Chip designers also face a familiar problem in trying to apply nanotechnology: Mask costs increase exponentially as structures become smaller. "We are anticipating prices of 4.5 million ($5.3 million) for a mask set," said Ebeling. Then there's the exploding cost of exposure systems. Steppers alone are expected to cost about of 70 million ($82 million) each.
To escape this price explosion, nanotechnology experts are seeking alternatives in the form of a "bottom-up" approach. This involves examining organizational principles in nature and trying to duplicate them. A prime example is the carbon nanotube. Scientists hope one day to be able to manufacture transistors based on the structure that are many times smaller than the tiniest structures that can be fabricated today.
The current state of semiconductor research means the practical implementation of the nanotechnology remains years way, experts here said. Nevertheless, researchers recently discovered that carbon structures could be placed on silicon wafers. That opens up the prospects of carbon circuits that are compatible with conventional silicon electronics. Infineon researchers are trying to integrate nanotechnology components into "wearable" applications.
Chip makers are also using nanotechnology to confront challenges like improving the performance of existing designs. According to Ehrenfried Zschech of AMD's Dresden lab, chips with 100 billion transistors will be possible by 2010. The semiconductor industry plans to reach this goal through the use of SOI, strained silicon and metal gates.
To continue along this path into the future, the industry is searching for new materials and processes to further miniaturize structures. Copper materials remain the state-of-the-art for interconnects, but experts are searching for better insulating materials providing the lowest possible dielectric constant.
Currently, the use of spin-on polymer materials (SOP) and plasma enhanced inorganic hybrid materials is being tested in the lab. These materials are expected to yield a dielectric constant in the range between 2.8 and 3.0. But some researchers are already searching for materials with an "ultra-low" dielectric constants.
AMD is also considering the use of nanotubes in future chip designs They could replace copper interconnects, for example, according to the paper by Zschech. On the other hand, optical interconnects are regarded as useful for transferring signals over greater distances from chip to chip or from board to board. Experts remain divided over whether optical interconnects offer an advantage within the chip. According to Zschech, innovations in packaging technology are needed to solve long-term connection problems.
As in the United States and elsewhere, nanotechnology research is enjoying political clout here. Edelgard Bulmahn, Germany's federal research minister, who rushed to the nanotechnology symposium from the opening of AMD's new chip facility here, said German government spending on nanotechnology research totals about 100 million ($117 million) annually.
However, Bulmahn said research remains too far from solving real-world problems. "We need to bring this knowledge together with product development and marketing," he stressed. "The goal of the research policy must be to generate marketable and competitive products."
-Christoph Hammerschmidt is editor-in-chief of EETimes.de.