Layout based Monte-Carlo simulation (LBMCS) for complex back end of line (BEOL) design rule study

Abstract

While BEOL process/integration stride to explore new technologies to support scaling needs, BEOL design rules are hindering design scaling due to the increasing process complexity. A design rule calculation tool that can consider all relevant process variation and calculate certain failure rate under the combination of these variation is in great need. In this paper, we introduce a Layout Based Monte Carlo Simulation (LBMCS) tool for this purpose and demonstrate three use cases. The first case is the 'bird beak' fail where sharp angle may form when via is placed near metal above inner vertex due to Self-Aligned Via process. LBMCS can help to predict the fail rate considering all variations while analytical prediction is almost impossible. The second case is via contact size failure that involves 5 process steps. While the complex calculation can be easily handled, we show that the LBMCS outputs can help to analyze and determine the most sensitive Process Assumption (PA), which can help to improve the yield and Design for Manufacturing (DfM) optimization. The third case is to help resolving competing rules in a design arc. We propose a new design check-off flow based on LBMCS that goes beyond design rules.