Increasing the running speed of microelectronics can open up new application areas: high-speed wireless communications, car collision avoidance or high-definition non-invasive imaging for security scanners. But microcircuits that can operate at over 100GHz, in order to implement these new products, demand performance of three times this speed at transistor level.
The EU therefore supported the three-year DotFive project, with the goal of designing Hetero-junction Bipolar Transistors (HBTs) that could reach 500GHz (or 0.5 THz). The challenge was to double the frequency compared to the state of the art at the submission of the project.
“And we did reach those numbers!” says the project coordinator Gilles Thomas of STMicroelectronics, France.
The project consortiumDOTFIVE Partners included four technology providers: two companies, Infineon and STMicroelectronics, and two research institutes. All partners made substantial progress, the two research institutes have even produced transistors running at the targeted speed, with Germany’s IHP Microelectronics GmbH getting the best results so far.
The DotFive team tried more than one approach to the problem: one of the project’s workpackages tried to evolve existing architectures, while an other workpackages tried ‘breakthrough’ architectures. As Mr. Thomas explains, “the architecture that did best is the one that minimises most the ‘parasitic effects’ in the transistor (such as capacitances, resistances and access resistance) and shows the best ‘self-alignment’ of base, emitter and collector.”
Working together to build a new market
One of the key successes of the project was therefore getting all the partners aligned on using the same methodology, electrical characterization technique,and modelling techniques so that the results were comparable among the project participants.
“In order to operate at these speeds we had to understand factors never encountered before,” says Mr Thomas, “as well as the physics of their effects. This could not be done without collaboration.”
“The kind of technology development carried out by DotFive is pre-competitive,” he explains further. The teams shared their computer aided design (CAD) platforms, measurement techniques, model parameters and some data processing.
Just as when the GSM standard for mobile phones was agreed, even if companies involved in a new technology are in competition with each other at the product level, they still need to collaborate in order to develop and agree the basic technology and standards.
“Europe is the leading producer of this type of technology,” according to Mr. Thomas, “we established the roadmap for radio frequency (RF) technology, so we’d better cooperate to maintain our lead.” This is one reason why the EU contributed funding of EUR 9.7 million towards an overall project budget of EUR 14.74 million.
“In order to produce the products that will enlarge the market five years down the road, we need to talk to each other today,” he says.
Results of first year already in commercial production
“In terms of commercialisation,” continues Mr Thomas, “we wanted to complete three cycles of learning – and incremental improvement of the design, process and tools – over the three years of the project.”
The results from the first year’s cycle are already embedded in the circuit designs in preparation, with an increase in circuit speeds from 77GHz to 120GHz.
“We are now in the qualifying stage of the results of the third cycle,” says Mr Thomas, with radar demos running at 140GHz.
“Car radars have moved into a new generation thanks to this project,” says Mr Thomas. “In addition to the 77GHz band allocated by international standards, we expect a new 120GHz band to be opened up for longer-range radar.”
“We would also like to develop imaging systems using millimetre waves,” he says. These lie above the 100 GHz range, between microwave and the infra-red radiations. Such imaging systems could contribute to public safety by improving security scanners.
“Currently, such systems exist but they are expensive, bulky and use a lot of electricity,” he explains, as they are built from discrete components not microcircuits. And because they don’t use integrated components, they cannot be assembled into large arrays, meaning their resolution remains poor.
“For scanners, if we can succeed with the miniaturisation and integration of our new high-speed components on silicon,” says Mr Thomas, “it will be like moving from 1950s computers, which filled an air-conditioned room, to PCs.”
The different project partners are now taking different routes to market with their products. Having established the basic technologies, the work is moving from pure research towards commercial development.
“We have now begun a new project, funded by the EUREKA programme’s CATRENE cluster on microelectronics, to develop a BiCMOS technology based on 500GHz HBTs plus digital CMOS for industrial production,” says Mr Thomas.
These would be able to integrate the RF components with the digital image processing on the same chip. Such revolutionary microcircuits may well contribute to continuing European success in these markets, as well as changing our lives through new, as yet unheard of, applications.
DotFive was funded by the EU’s Seventh Framework Programme (FP7), under the ICT sub-programme and budget line for ‘Next-generation nanoelectronics components and electronics integration’.
Source: Gilles Thomas, STMicroelectronics, France (on Cordis web site).