Dose Correction Method for Semiconductor Process Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current semiconductor manufacturing processes face challenges in efficiently transitioning between different equipment or resist types due to the need for extensive recalibration and computation-heavy correction methods, which are time-consuming and costly, especially when maintaining identical behavior across process changes.
Innovation Solution
A method that determines dose corrections to replace or complement geometry corrections, allowing for process matching using a single differential model, which reduces calibration and correction efforts and maintains an adequate process window, enabling flexible matching and interpolation/extrapolation between measurement points.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If geometry correction is applied to match the new process to the reference process, then the manufacturing precision is improved, but the device complexity increases and the computation cost increases
Solution Approach 1:
The patent extracts the essential matching parameter (edge displacement) from the complex geometry correction process and isolates it into a separate, manageable component. By focusing only on the critical difference between processes rather than correcting all geometric parameters, the method reduces complexity while maintaining matching accuracy.
Solution Approach 2:
The correction process is segmented into distinct steps: first calculating edge displacement using a differential model, then applying this displacement to the target design. This segmentation separates the complex geometry correction into a simple displacement application, reducing overall system complexity.
2Manufacturing precision
If extensive recalibration and computation-heavy correction methods are used to maintain identical behavior across process changes, then the manufacturing precision is improved, but the productivity decreases
Solution Approach 1:
The patent performs preliminary calibration between the reference process and new process before actual production. The differential model is pre-established and stored, so that when process changes occur, the pre-calculated edge displacement corrections can be immediately applied without extensive recalibration, thus maintaining productivity.
Solution Approach 2:
The method creates a simplified copy of the reference process behavior through the differential model, which captures the essential characteristics needed for matching. This copy allows rapid application of corrections without replicating the entire complex calibration process, thereby improving productivity.
3Productivity
If a single differential model is used for process matching, then the productivity is improved and the computation burden is reduced, but the measurement precision may be insufficient when measurement points are not well scattered
Solution Approach 1:
The patent implements a feedback mechanism where the differential model is calibrated using metrology results from both the reference process and new process. The model continuously adjusts based on measured edge displacements, ensuring that even with limited measurement points, the calibration accuracy is maintained through iterative refinement.
Solution Approach 2:
The method changes the approach from using multiple complex calibration parameters to using a single edge displacement parameter derived from the differential model. This parameter change simplifies the calibration process while maintaining accuracy by focusing on the most critical difference between processes.
4Adaptability or versatility
If sizing correction is applied beyond certain limits to achieve process matching, then the adaptability is improved, but the reliability decreases due to process window reduction
Solution Approach 1:
The patent applies local quality by allowing edge displacement corrections to be applied selectively to different regions of the design based on local requirements. Rather than applying uniform sizing corrections that could push the process beyond reliable limits, the method adjusts only the necessary edges locally, maintaining process window adequacy while achieving adaptability.
Data Source
Figure 1
Figure 2
Figure 3a~3b
AI summary
The invention discloses a method to easily determine the parameters of a second process for manufacturing from the parameters of a first process. Metrics representative of the differences between the two processes are computed from a number of values of the parameters, which can be measured for the two processes on a calibration layout, or which can be determined from pre-existing values for layouts or reference data for the two processes by an interpolation/extrapolation procedure. The number of metrics is selected so that their combination gives a precise representation of the differences between the two processes in all areas of a design. Advantageously, the metrics are calculated as a product of convolution of the target design and a compound of a kernel function and a deformation function. A reference physical model of the reference process is determined. A sizing correction to be applied to the edges of the design produced by the reference process is calculated. It is then converted, totally or partially, into a dose correction.