Design for manufacturability: challenges and opportunities

Abstract

Design for manufacturability (DFM) is one of the foremost concerns in the semiconductor industry today. This work describes a few of the key issues facing circuit designers and tool developers in the DFM space and points to future areas of research interest. In particular, we show that comprehending and exploiting systematic pattern dependencies can be a powerful knob to improve performance under variability. Furthermore, we describe a cost-driven approach to mask-level correction that points to new areas of research that would allow design rules to vary depending on context and/or instance criticality. This and other opportunities in DFM research are summarized