Archived Viewpoints

About Advanced Silicon Microelectronics/ULSI

July 2008

Integrated circuits have, since their commercial introduction in 1961, become the enabling technology for virtually all modern data-transfer and storage operations, from electrically controlled domestic appliances to complex telecommunications networks. The dramatic progress from silicon chips containing only a few semiconductor junctions to today's state-of-the-art circuits containing millions of transistors has resulted from the evolutionary refinements in microfabrication processes and materials and the development of a system and software infrastructure to take full advantage of these devices.

Ultra-large-scale integration generally refers to today's leading-edge chip densities, although even higher levels of integration are already in development. The spread of IC-based systems into every facet of modern life is limited at the one extreme by the economic feasibility of providing electronically controlled functions and at the other by the technological limits on the quantity of data that researchers can manipulate. At the same time, higher integration levels are reducing the cost of processing power (cost per operation or instruction).

Observers have historically predicted the development of future generations of integrated circuits with fair accuracy, based on projecting constant, incremental improvements in both technology and economies of scale. These development cycles—which Moore's law (for Gordon Moore of Intel) embodies—have held fast for the 30 or so years that the semiconductor industry has been in existence. But more recent observations indicate that these cycles are gradually shortening, challenging the assumption that current technology will continue to yield to research and development in a predictable way. Pessimistic commentators suggest the approach of a technology "wall" defined by the smallest features that can be controllably etched to form individual transistors. However, so far, the ingenuity of semiconductor scientists appears far from exhausted. Novel technologies, new materials, and creative chip architectures will likely emerge to continue the advancement of integrated circuits.