From circuit miniaturization to the scalable IC

Abstract

For more than three decades, the main driving force behind the integrated circuit (IC) revolution has been the steady scaling downwards of their lithographic design rules. It may therefore come as a surprise to those not familiar with their earlier history that the original IC did not, and actually could not, embrace the scaling concept. Instead, the early work was intended to achieve a revolution in circuit miniaturization by bringing all circuit components down to the same tiny size as the transistor. For these circuits to become truly scalable ICs, new transistor structures had to be invented and adapted to IC needs. Until this happened the cost-effectiveness of the IC was highly controversial. Five years after the IC was invented, plans were being put in place by IBM and AT&T to utilize flipchip transistor technology rather than ICs. But these, like the original ICs, were also unscalable, and consequently lost out to a new generation of ICs that were scalable. Two revolutions were driven by the subsequent evolution of design rules; first TTL/minicomputers, and then MOS/microcomputers. K. Marx proved disastrously wrong in the political arena, but his aphorism "evolution carried far enough becomes revolution" fits IC history like a glove.