Multi-Beam Writing Data Compression for Curve Approximation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing complexity of semiconductor circuit designs, particularly with finer linewidths, leads to a significant increase in the amount of writing data required for multi-beam writing apparatuses, especially when approximating figures with curves and straight lines, resulting in inefficiencies and longer processing times.
Innovation Solution
A method for generating writing data that calculates control points and vertices for curves and straight lines, expressing positions as displacements from adjacent points, and organizing data into sections to reduce the overall data volume, allowing for more efficient processing and handling of complex patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If high-precision approximation is performed for figures with curves, then manufacturing precision is improved, but the amount of writing data increases
Solution Approach 1:
The patent segments the figure data into two distinct sections: curve figures and straight-line figures. This segmentation allows each type to be processed and stored using its own optimized data structure, reducing the overall data volume while maintaining high approximation precision for curves through dedicated curve representation methods
Solution Approach 2:
The patent changes the data representation parameters by using displacement values relative to placement origins rather than absolute coordinates. This parameter transformation reduces data redundancy and enables more efficient storage of curve approximation data while preserving manufacturing precision
2Manufacturing precision
If the number of vertices is increased for high-precision approximation, then manufacturing precision is improved, but device complexity increases
Solution Approach 1:
By dividing figure data into separate curve and straight-line sections, the patent reduces the complexity of processing each individual section. Curve sections can use specialized approximation algorithms while straight-line sections use simpler representations, avoiding the need to process all vertices uniformly
Solution Approach 2:
The patent performs preliminary organization of figure data into categorized sections before processing. This preliminary action groups curve vertices and straight-line vertices separately, enabling more efficient subsequent processing and reducing the computational complexity during the actual writing operation
Data Source
AI summary
In one embodiment, a writing data generating method is for generating writing data used in a multi charged particle beam writing apparatus. The method includes calculating, for a figure containing a curve and a straight line included in design data, a plurality of control points representing the curve and a plurality of vertices of the curve and straight line, and expressing a position of each of the control points and vertices as a displacement from an adjacent control point or vertex to generate the writing data.