Physical model-based ArF photoresist formulation development

Since the logic 28 nm technology and beyond, ArF immersion lithography has been widely used in manufacturing. To fully utilize the potential of the lithographic resolution, process simulation has been used since the lithography process setup step. The accuracy of the model prediction can be a key factor in the process stability, defectivity, and yield of the final product. To get better model prediction, it is very important to develop a process model with as many parameters as possible with physical meanings. Since the analysis of exposure data with a physical model can provide insights into the process and the photoresist material, it may greatly accelerate the photoresist formulation development or improvement process. From another aspect, as the function of the photoresist is to record the aerial image information on wafer, theoretically, the photoresist image should not deviate much from the aerial image or it will inevitably reduce the information content of the aerial image, i.e., through various kinds of averaging, mixing, etc. Since the emergence of Optical Proximity Correction (OPC), the time that takes to finalize a photolithography process has been significantly lengthened and made more complicated. To make the OPC model more reliable and less dependent on patches, a more physical OPC model is necessary. Therefore, it is clear that if we can develop a photoresist formulation to realize a more physical photoresist image, or a photoresist image closer to the aerial image, we can greatly improve the photolithography process performance and the time consumed for the process setup including OPC. In this paper, we focus on several physical model parameters for photoresist, especially the effective photoacid diffusion length and photo decomposable base parameters. We will show that these two parameters can provide a good description of photoresist imaging behavior, and the resist formulation can be improved to match the physical model prediction with more accuracy.