Optical proximity correction using diffused aerial image model

The aerial threshold model has been popularly used due to its fast computat ional speed and convenience. But it does not guarantee accurate prediction of CD (Critical Dimension), especially to correct optical proximity effect because it does not take into account the resist process in the model. Ther efore, there has been a strong demand for an improved simulation model whic h can more accurately predict the CD's for various patterns in shapes and s izes within an acceptable computational time budget and with an easiness to calibrate parameters from experimental data. To meet such demands, we have developed a new approximation model called DAIM (Diffused Aerial Image Mod el) where a resist process is represented by diffusion and amplification fu nction. In this paper, we compared the simulation results using the DAIM wi th experimental data for three types of patterns (line/space, isolated line and isolated space) with respect to their linearity. Compared with classic al threshold models, the DAIM resulted in a dramatic improvement in accurac y without any negative effect on computational speed and also offers simpli city in analysis. We also found that the mask error effect becomes very sig nificant as the pattern size becomes smaller, thus making OPC (Optical Prox imity Effect Correction) less effective. This means that the mask CD contro l is very crucial together with proper OPC using an accurate simulation mod el in order to achieve sub-0.25 mu m patterning with high-fidelity.