Shape of things to comeGoing beyond Moore’s Law

Moore’s law, formulated by Intel’s Gordon Moore in 1965, states that the number of transistors in a microchip doubles every two years, but this is just not fast enough for current innovation. Now computer scientists are reaching beyond that limit by packing more functions onto a single chip, and European researchers are setting the standards. Thanks to the work of the Leuven, Belgium-based European Network for the Co-ordination of Advanced Semiconductor Technologies (ENCASIT), Moore’ law can now be exceeded. “The main driver for ENCASIT is mobile computing, handheld devices and medical technologies,” says Ken Ball, technical manager for the project. When a medical system is deployed in the body, for example, it requires more than one chip in the package to accomplish a specific task, and these chips need to be in one, very small, package: Just think of advanced micro-medical devices that need to reach a specific location, or even specific cancer cells within an organ. It is a complex task, and demands a complex device. Until now, Moore’s law made sure microchip design stayed abreast of current innovation. “Moore’s Law describes the doubling of [microchip] speed, function, density, almost anything, during a period of two years or one generation of a product,” says Ball. This is fast, but not fast enough to keep pace with current innovation.

Chip designers and manufacturers thus started stacking, or packing, several functional chips into a single, extremely small, package. The process starts with a semiconductor die, a small piece of semiconducting material onto which a functional circuit is fabricated, using a wide array of techniques from etching to deposition. Several die are combined together to make a system in package (SIP). Often, though, there is a problem, because each die comes from a different manufacturer before final assembly, and up to now there was no standard way to define semiconductor dies. “It’s a problem that’s ten years old,” says Ball. “A company would source components for the package from different suppliers. But the information was in different formats, arriving on paper, and transcribed by hand. It most often meant the package needed a redesign.” This brings us back to Europe, where, ten years ago the European Commission funded the Known Good Die project (KGD) to promote sourcing of good die suppliers. The EU continued to support the effort through ENCAST and now ENCASIT. “Many of the KGD project members are still working in ENCASIT, so there has been a continuity right through the work,” says Ball. The effort has produced important results. A standard way to describe the physical and electrical properties of a semiconductor die now exists. “It has led to co-design, where the clients work with all suppliers at the same time to design a chip,” says Ball. Even more importantly, ENCASIT communicates these new standards through a large number of seminars, conferences, newsletters, and even a dedicated Die Products Users Club that regroups suppliers and their clients. “The Die Product Users Club is particularly important for SMEs, companies that need small quantities of die products but that often find it difficult to source reliable suppliers,” says Ball.

Some problems are yet to be overcome. “Often the chip information comes from the marketing side of the suppliers’ business,” says Ball. “But co-design needs information from the manufacturing side.” Even worse, much of it is still by hand. It is an irony that semiconductor packaging, a key technology of the information age, still uses paper. “We now hope to develop an electronic data sheet using standard descriptions of the outline and terminals on the one hand, and the electric circuit design on the other.” Visual standards are another area that needs work. This will be after the lifetime of this project, which ends in early 2008, and Ball says the consortium hopes to launch another. “A lot of the partners are eager to continue this work.” In a famous lecture in 1959, physicist Richard Feynman announced “There is plenty of room at the bottom,” inviting the audience to enter the new field of physics that was to become nanotechnology. The work done by ENCASIT means that, as Feynman predicted, engineers will continue to exploit vast areas of microscopic space.