Edge Fragment Correlation for OPC Convergence

The patent segments the OPC process into two distinct phases: a rough OPC pass that handles individual edge fragments independently, and a fine OPC pass that processes edge fragments in correlated groups based on calculated correlation values. This segmentation allows the system to handle the increased complexity of advanced technology nodes by dividing the problem into manageable stages, improving OPC convergence without requiring complete redesign of the entire OPC process.

The patent performs preliminary calculation of edge fragment correlation values and identification of correlated edge fragment groups before executing the fine OPC pass. By pre-processing and organizing edge fragments into correlated groups based on their mutual influence, the system prepares the data structure needed for efficient convergence in the fine OPC stage, addressing the convergence failure issue before the main correction process begins.

The patent implements a partial action approach by selectively applying the computationally intensive fine OPC pass only to identified correlated edge fragment groups, rather than processing all edge fragments with full correlation analysis. The rough OPC pass handles the majority of edge fragments with simpler processing, while the fine OPC pass focuses computational resources on specific correlated groups, thus improving overall convergence without proportionally increasing total runtime.

The patent performs preliminary identification and grouping of correlated edge fragments before the fine OPC pass, pre-processing the data to organize it in a way that optimizes subsequent computational efficiency. By calculating correlation values and identifying correlated groups in advance, the system reduces the computational burden during the actual fine OPC execution, balancing convergence improvement with runtime considerations.

The patent introduces image intensity slope and amplitude sensitivity information as intermediary parameters that mediate between the physical optical effects and the OPC correction process. These intermediaries provide a quantitative framework for calculating edge fragment correlations, enabling accurate prediction of how mask errors propagate to wafer features. By using these intermediary measurements, the system achieves high image reproduction accuracy without requiring direct complex physical measurements during OPC execution.

The patent utilizes parameter changes in image intensity slope and amplitude sensitivity as indicators of edge fragment correlation strength. By monitoring and calculating these parameters, the system dynamically identifies which edge fragments are strongly correlated and require joint processing. This parameter-based approach transforms the complex physical interaction problem into a manageable computational task, improving image reproduction accuracy through quantitative parameter analysis rather than complex physical modeling.