Optical Proximity Correction Using a New Hyper Error Estimation Method

Abstract

In recent years, semiconductor manufacturing process has made great progress. To avoid lithography hotspots and enhance the yield of integrated circuits, we can use Model-based Optical Proximity Correction (MBOPC) to improve image fidelity and printability. However, the optical lithography simulation of MBOPC is a time-consuming calculation. In this paper, we propose an effective MBOPC to obtain a modified mask with high-resolution. The proposed OPC flow is divided into three steps: (1) Pre-simulation generates a set of modified value for each pattern, (2) Gridbased Partition can speed up MBOPC process and overcome alignment problem, and (3) a set of hotspot detection formulas detects the variation sub-area. The first and second steps will improve performance of lithography, and the third step will improve image fidelity. The experimental results of our procedure show that the average of edge placement error (EPE) within the marked area can be decreased from 259.76um to 7.24um, and bitmap error (BME) within the marked area can be decreased from 20.01% to 3.15%.