IC technology R&D for the next century

Summary form only given. As we have gone through the era of integrated circuits, more strictly silicon-based CMOS integrated circuits, where integration density has increased from hundreds of transistors on a chip to almost one quarter of a billion transistors on a chip, a way to develop technology itself has come to the point where we must examine any possible new model for research and development. Technology trends which have been well discussed in forums such as consortia in the USA, Japan, Europe, South-east Asian countries are now converging reasonably, with some differences due to the differences of so-called "technology drivers". Common parameters are the minimum geometries, though they have increasingly become more conceptual parameters as opposed to what can be found on a real chip. Obviously, the density of active elements per unit area and the number of interconnect layers are more important parameters for high-density memories and high performance processors, respectively. The purpose of this talk is not to discuss the technology trend itself, but to examine how we have developed technology and how we can possibly continue in the future.