Process Model Guided Photoresist Formulation Optimization

Ever since the debut of Chemically Amplified Resist (CAR), the optical resolution in lithography has been continually improving since 0.25 micron all the way to 5 nm with 248 nm, 193 nm, 193 nm immersion, and Extreme Ultra-Violet (EUV) lithography. The key behind the success of CAR platform in resolution accomplishment through many generations is chemical amplification in deprotection reaction, which not only drastically improves sensitivity, but also makes the cross section profile in the photoresist much more vertical. Throughout a nearly 30-year span of integrated circuit generations, the amplified deprotection reaction is done through photo-generated acid diffusion. Not until year 2004, the measurement of the diffusion length is not accurate enough to enable effective modeling [1]. Now it becomes a consensus that the improvement of CAR resolution must be accompanied by the reduction of the diffusion length. Due to the fact that the reduction of the photoacid diffusion length may reduce chemical amplification, an increase of PhotoAcid Generator (PAG) loading and/or a reduction of deprotection activation energy may be needed. The deprotection activation energy depends on both the PAG and the polymer designs, which can also affect linewidth bias between dark field and bright field conditions. This will generated difference in Optical Proximity Effect (OPE). We have used a self-developed physical aerial image simulation model to fit to the wafer exposure data of Exposure Latitude (EL) in Focus Exposure Matrix (FEM) and linewidth through pitch (an example linewidth through pitch data are shown in Fig. 1) to extract physical parameters such as, the effective photoacid diffusion length, the isolated to dense linewidth bias and compare the extracted data to the commonly accepted industry standard [2] or the target requirement from the customer processes. The difference can serve as a guide for formulation improvement and we will demonstrate that this is a very efficient formulation optimization methodology. Our model parameters can also cover and be applied to situations where Photo Decomposable Base (PDB), Negative Toned Developing (NTD), and EUV exposure are used for advanced photoresist platform development in the future.