Substrate-Specific OPC Models for Gate Photomask Dimension Control

The patent segments the gate photomask patterns into two distinct categories: those on Active Area substrates and those on Shallow-trench isolation substrates. Separate OPC models are created for each substrate type, allowing each model to be optimized for its specific substrate characteristics. This segmentation resolves the contradiction by maintaining process simplicity through automated classification while achieving high precision through substrate-specific correction models.

The patent applies local quality by creating specialized OPC models tailored to each substrate type's unique optical characteristics. The Active Area model and Shallow-trench isolation model have different parameters and correction rules optimized for their respective substrate properties. This allows each local region (substrate type) to receive appropriately optimized correction, resolving the contradiction between unified simple processing and location-specific precision requirements.

The patent manages complexity by segmenting the problem into two distinct, well-defined substrate types with dedicated models. Rather than creating a single complex model attempting to handle all variations, the segmentation approach creates two simpler, specialized models. The automated classification system routes patterns to the appropriate model, preventing the need for complex decision-making within each model itself.

The patent implements a universal OPC processing framework that can handle multiple substrate types through a single integrated system. The classification mechanism and processing pipeline are designed to work with both Active Area and Shallow-trench isolation models uniformly. This universality reduces overall system complexity by providing a consistent interface and workflow, despite the presence of multiple specialized models.