Semiconductor Mask Pattern Data Correction via Marginal Error Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The optical proximity effect in semiconductor manufacturing leads to inaccuracies in transferring mask patterns onto wafers, and existing methods struggle to evaluate and correct for fatal errors caused by infinite variations in peripheral patterns, making realistic evaluation impossible within practical time frames.
Innovation Solution
A pattern data creating method that identifies marginal error patterns, generates peripheral environment patterns to deteriorate the transfer fidelity of these patterns, and repeatedly applies OPC and lithography simulation to correct the evaluation target cell patterns until fatal errors are removed, ensuring accurate pattern transfer.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If process simulation is applied to evaluate all possible peripheral pattern variations, then evaluation completeness is improved, but evaluation time becomes infinitely long
Solution Approach 1:
The patent extracts only the critical marginal error patterns from the complete set of peripheral pattern variations. By identifying patterns with insufficient margins that are most likely to cause fatal errors, the method evaluates only these extracted patterns rather than all possible variations, thereby achieving reliable evaluation within practical time limits.
Solution Approach 2:
The patent segments the evaluation process into distinct stages: first identifying marginal error patterns through initial simulation, then creating deteriorated peripheral environment patterns specifically targeted at these marginal cases, and finally performing focused simulation on these segmented critical cases. This segmentation transforms an intractable complete evaluation into manageable focused evaluations.
2Manufacturing precision
If OPC processing is repeatedly applied to correct marginal error patterns, then pattern transfer accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent implements a feedback loop where simulation results are used to identify marginal error patterns, which then trigger OPC processing, followed by re-simulation to verify correction. This closed-loop feedback system systematically improves pattern transfer accuracy by continuously evaluating and correcting errors until no marginal patterns remain.
Solution Approach 2:
The patent performs preliminary identification of marginal error patterns before final pattern creation. By detecting patterns with insufficient margins in advance and applying OPC processing to these specific cases before manufacturing, the method prevents fatal errors from occurring rather than correcting them after the fact, thereby improving accuracy efficiently.
Data Source
AI summary
A pattern data creating method according to an embodiment of the present invention comprises: extracting marginal error patterns using a first result obtained by applying process simulation to mask pattern data based on an evaluation target cell pattern, applying the process simulation to mask pattern data based on an evaluation target cell pattern with peripheral environment pattern created by arranging a peripheral environment pattern in the marginal error patterns such that a second result obtained by creating mask pattern data and applying the process simulation to the mask pattern data is more deteriorated than the first result, and correcting the evaluation target cell pattern or the mask pattern data based on the evaluation target cell pattern when there is a fatal error.