Machine-Learned Dose Mapping for Wafer Deformation Mitigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern semiconductor manufacturing faces challenges in mitigating complex wafer deformations caused by high aspect ratio structures, leading to misalignment and degradation of device quality, particularly in correcting anisotropic and non-uniform stresses that traditional stress-compensation layers and ion implantation methods struggle to address effectively.
Innovation Solution
A machine learning model is trained to determine dose maps for a stress-modification beam applied to a stress-compensation layer on the substrate, using a combination of measured wafer deformation data and physics models to predict optimal beam parameters, such as ion species, energy, and dose distribution, to correct both isotropic and anisotropic stresses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional stress-compensation layers and ion implantation methods are used, then some stress compensation is achieved, but they struggle to effectively address complex anisotropic and non-uniform stresses
Solution Approach 1:
The patent uses machine learning to determine optimized beam parameters (ion species, energy, dose distribution) based on measured wafer deformation. The system adjusts these parameters dynamically to match the specific stress pattern, enabling effective correction of complex anisotropic stresses that traditional fixed-method approaches cannot handle.
Solution Approach 2:
The patent replaces traditional mechanical/physical stress compensation methods (fixed ion implantation patterns, uniform stress-compensation layers) with a machine learning-based system that uses computational models to predict and optimize beam parameters, allowing adaptive response to complex stress distributions.
2Productivity
If machine learning models are used to compute dose maps, then fast and accurate computation is achieved, but the system complexity increases
Solution Approach 1:
The patent uses machine learning models that have been trained on simulation data to copy the results of complex physics simulations. Instead of performing computationally intensive physics calculations in real-time, the system uses pre-trained neural networks that can quickly predict dose maps by pattern recognition, achieving both speed and accuracy.
Solution Approach 2:
The machine learning models are trained in advance using simulation data and historical wafer deformation measurements. This preliminary training allows the system to make accurate predictions without requiring complex real-time calculations, thereby reducing computational burden during actual wafer processing.
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
The approach enables fast and accurate computation of dose maps for efficient mitigation of wafer deformations, ensuring precise alignment of features and improving the quality of manufactured semiconductor devices by effectively flattening the substrate.
Implementation Method 1
a stress-modification beam (SMB) that, being applied to a stress-compensation layer (SCL) formed on the substrate, causes modification of the deformation of the substrate
Data Source
AI summary
Disclosed systems and techniques that deploy machine learning models (MLMs) for mitigation of stresses and deformations of substrates. The techniques include obtaining a training input that includes a map of deformation of a substrate, processing the training input using the MLM to generate an MLM output, the MLM output predicting a dose map for a stress-modification beam (SMB) that, being applied to a stress-compensation layer (SCL) formed on the substrate, causes modification of the deformation of the substrate. The techniques further include training the MLM using the predicted dose map and deploying the trained MLM for processing of one or more additional substrates.