Lithography Pixel Dose Bias Correction for Uniform Pattern Edges
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional bias correction methods in lithography fail to achieve consistent edge biasing for patterns with small minimum shape dimensions, especially in electron beam lithography, due to variations in shot dosage and overlapping shots, leading to non-uniformity in biasing results.
Innovation Solution
A method and system that utilize a multi-beam energy source to calculate and adjust pixel dosages in real-time, relocating pattern edges by a target bias using a graphics processing unit (GPU) for inline processing, enhancing dose margin and achieving consistent edge bias across patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional bias correction methods are used, then processing is simpler, but manufacturing precision deteriorates for patterns with small minimum shape dimensions
Solution Approach 1:
The patent changes the parameter being corrected from geometric dimensions to exposure dosage. By adjusting the dosage of pixels adjacent to pattern edges based on their distance from the edge and the pattern's minimum dimension, the method achieves consistent edge bias for small features without complex geometric processing. The dosage adjustment formula incorporates the pattern's minimum dimension to normalize the bias correction across different feature sizes.
Solution Approach 2:
The patent replaces conventional geometric bias correction methods with a dosage-based correction approach. Instead of modifying pattern geometry through complex calculations and iterations, the system substitutes this with a straightforward dosage adjustment process that uses the pattern's minimum dimension as a reference parameter to achieve uniform edge bias across all patterns.
2Manufacturing precision
If multiple iterations of geometric processing are performed to achieve uniform bias, then manufacturing precision improves, but productivity deteriorates
Solution Approach 1:
The patent performs preliminary calculation of the pattern's minimum dimension and uses this value to pre-determine the dosage adjustment parameters. By having this reference value ready before the exposure process, the system eliminates the need for multiple iterative geometric processing steps, achieving uniform edge bias in a single pass and significantly improving throughput.
Solution Approach 2:
The patent transforms the bias correction problem from a geometric iteration problem to a single-step dosage parameter adjustment. By changing the correction parameter from geometric offset to exposure dosage based on the pre-calculated minimum dimension, the system achieves edge uniformity without requiring multiple processing iterations, thereby maintaining high productivity.
3Measurement precision
If shot dosage is increased to improve signal for small patterns, then measurement precision improves, but object-generated harmful factors worsen due to overlapping shots
Solution Approach 1:
The patent applies local quality by differentiating dosage adjustment based on the local context of each pixel. Pixels adjacent to edges of small patterns receive higher dosage adjustments than those near large patterns, with the adjustment量 determined by the pattern's minimum dimension. This localized approach improves edge detection accuracy for small features while avoiding excessive dosage that would cause overlapping shot problems in other areas.
Solution Approach 2:
The patent changes the dosage parameter dynamically based on the pattern's minimum dimension and the pixel's distance from the pattern edge. This parameter adjustment ensures that small patterns receive sufficient signal for accurate edge detection while the dosage increase is localized and controlled, preventing the harmful overlapping shot effects that would occur with uniform dosage increase across the entire substrate.
Data Source
Figure 1
Figure 2
Figure 3A~3F
AI summary
Methods include inputting an array of pixels, where each pixel in the array of pixels has a pixel dose. The array of pixels represents dosage on a surface to be exposed with a plurality of patterns, each pattern of the plurality of patterns having an edge. A target bias is input. An edge of a pattern in the plurality of patterns is identified. For each pixel which is in a neighborhood of the identified edge, a calculated pixel dose is calculated such that the identified edge is relocated by the target bias. The array of pixels with the calculated pixel doses is output. Systems for performing the methods are also disclosed.