Multi Beam Writing Data Transfer Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Multi-beam writing techniques for semiconductor devices face challenges with high data processing demands and long data transfer times due to the need for parallel processing and large-capacity memory in multi-pass writing methods, which complicates the control of irradiation time periods for multiple beams.
Innovation Solution
A multi-charged particle beam writing apparatus with a movable stage, an aperture plate, and data processing circuitry that generates and corrects bitmap data for irradiation time periods, using dose correction units and data transfer processing circuitry to efficiently transfer dose data to blanking plates through signal line groups, allowing for reduced data processing and memory requirements by changing signal line groups for each writing pass.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multi-pass writing is performed with parallel processing on multiple computers, then the throughput is improved, but the data transfer time and memory capacity requirements increase
Solution Approach 1:
The patent divides the writing area into multiple regions and assigns each region to a specific computer for processing. Instead of having all computers process all data, each computer handles only its assigned region's data throughout the multi-pass writing process. This segmentation of processing responsibilities significantly reduces the amount of data each computer needs to store and transfer, while maintaining parallel processing throughput benefits
Solution Approach 2:
The patent introduces a spatial dimension to data distribution by mapping specific writing regions to specific computers. Rather than distributing data across computers in a traditional manner, the system assigns regions in the writing area to computers, creating a many-to-one mapping between beams and computers that reduces overall data transfer requirements
2Productivity
If multi-pass writing is performed with parallel processing on multiple computers, then the throughput is improved, but the memory capacity requirements increase
Solution Approach 1:
The writing area is segmented into multiple regions, with each region assigned to a specific computer. This segmentation allows each computer to store only the bitmap data for its assigned region rather than the entire writing area, significantly reducing memory capacity requirements while enabling parallel processing for improved throughput
Solution Approach 2:
Each computer is configured with memory capacity matched to its specific processing requirements for its assigned region. This local optimization ensures that memory resources are efficiently utilized without over-provisioning, as each computer's memory capacity is tailored to its local processing needs rather than requiring uniform high capacity across all computers
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This configuration reduces the amount of data transferred and memory capacity needed, enhancing the efficiency of multi-pass writing by optimizing data processing and transfer in multi-beam writing systems.
Implementation Method 1
an aperture plate including a plurality of openings through which the charged particle beam passes to form multiple beams
Implementation Method 2
a blanking plate including a plurality of blankers each performing blanking deflection on a corresponding one of the multiple beams
Data Source
AI summary
In one embodiment, a multi charged particle beam writing apparatus includes a blanking plate including a plurality of blankers, bitmap generation processing circuitry generating bitmap data for each writing pass of multi-pass writing, the bitmap data specifying irradiation time periods for a plurality of irradiation positions, a plurality of dose correction units configured to receive bitmap subdata items obtained by dividing the bitmap data from the bitmap generation processing circuitry, and correct the irradiation time periods to generate a plurality of dose data items corresponding to respective processing ranges, and data transfer processing circuitry transferring the plurality of dose data items to the blanking plate through a plurality of signal line groups. Each of the signal line groups corresponds to the blankers located in a predetermined region of the blanking plate. The data transfer processing circuitry changes the signal line groups, used to transfer the plurality of dose data items generated by the respective dose correction units, for each writing pass.