Two-Level Pattern Generation for Lithography via Error Diffusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current charged particle beam lithography systems face challenges in achieving smaller critical dimensions, such as 22 nm, while maintaining sufficient wafer throughput and addressing dose variation and feature placement accuracy issues due to beamlet size and proximity effects.
Innovation Solution
A method for generating a two-level pattern using a multi-beamlet system that assigns multi-level values to pixel cells based on relative coverage and dose levels, employing error diffusion algorithms to form a multi-level pattern which is then converted to a two-level pattern for exposure, allowing for improved dose control and feature placement accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional raster scan method with beamlet modulation is used, then lithographic processing can be performed, but feature placement accuracy deteriorates due to beamlet size and proximity effects
Solution Approach 1:
The patent segments the continuous beamlet exposure process into discrete pixel cell exposures with multi-level dose values. Each pixel cell is assigned a specific dose level based on error diffusion calculations, allowing independent control of each cell's exposure amount. This segmentation enables precise feature placement by controlling which pixel cells receive exposure and at what dose level, overcoming the proximity effect limitations of conventional beamlet scanning.
Solution Approach 2:
The patent changes the dose parameter from a binary on/off state to multiple discrete dose levels (multi-level pattern). By assigning different dose levels to different pixel cells based on error diffusion algorithms, the system can precisely control the exposure amount for each cell, improving feature placement accuracy and reducing patterning errors caused by beamlet size and proximity effects.
2Manufacturing precision
If beam size is reduced to achieve smaller critical dimensions, then resolution improves, but dose variation between adjacent cells increases
Solution Approach 1:
The patent implements feedback through error diffusion algorithms that calculate the dose error for each pixel cell based on the difference between the desired multi-level dose and the actual assigned dose. The system uses this error information to adjust the exposure of subsequent pixel cells, compensating for dose variations caused by beamlet size and proximity effects. This feedback mechanism ensures uniform dose distribution across the patterned area while maintaining precise critical dimension control.
3Manufacturing precision
If error diffusion algorithm is applied to generate multi-level pattern, then feature placement accuracy improves, but computational complexity increases
Solution Approach 1:
The patent segments the error diffusion calculation into independent pixel cell processing units. Each pixel cell's dose level is calculated based on local error diffusion from neighboring cells, allowing the complex computation to be divided into manageable discrete steps. This segmentation reduces computational complexity compared to calculating the entire pattern simultaneously, while still achieving high pattern accuracy through the cumulative error diffusion effect.
4Manufacturing precision
If multi-level pattern is converted to two-level pattern for exposure, then lithographic processing becomes feasible, then patterning errors are reduced, but information loss occurs during conversion
Solution Approach 1:
The patent converts the harmful effect of information loss during multi-level to two-level pattern conversion into a beneficial process. By using error diffusion algorithms during the conversion, the system calculates and distributes the dose error across neighboring pixel cells, transforming the information loss into a controlled error distribution that actually improves pattern accuracy. The error diffusion process ensures that the cumulative effect of the two-level pattern closely matches the intended multi-level pattern, reducing patterning errors.
Data Source
Figure 1~2
Figure 3~4B
Figure 5A~6B
AI summary
The invention relates to a method of generating a two-level pattern for lithographic processing by multiple beamlets. In the method, first a pattern in vector format is provided. The vector format pattern is then converted into a pattern in pixmap format. Finally, a two- level pattern is formed by application of error diffusion on the pixmap format pattern.