Semiconductor Yield Prediction From Process Geometry Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current semiconductor manufacturing processes lack a reliable method to predict the yield of functional integrated circuits (ICs) until the final stage of production, leading to inefficiencies and low yield ratios due to the inability to assess the quality of ICs until all process steps are completed.
Innovation Solution
A method is developed to predict yield by obtaining a trained model that translates geometrical and process parameters into yield parameters, using geometric models and free parameters to construct yield probability models for device elements, allowing for early prediction of yield based on process data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If final stage electrical testing is performed to determine yield, then accurate yield measurement is achieved, but production time is extended and efficiency is reduced
Solution Approach 1:
The patent applies preliminary action by performing yield prediction during the manufacturing process using process parameter data and trained models, rather than waiting until final testing. This allows yield to be estimated early based on intermediate measurements and process data, significantly reducing the time loss associated with waiting for final stage electrical testing while maintaining reasonable accuracy through the predictive modeling approach.
2Manufacturing precision
If process parameters are controlled to improve yield, then manufacturing precision is improved, but process complexity increases
Solution Approach 1:
The patent introduces an intermediary predictive model that translates process parameter data into yield predictions. This intermediary system simplifies the control complexity by providing a clear causal relationship between process parameters and yield outcomes, allowing manufacturers to adjust process parameters based on model predictions without needing to understand the complex underlying manufacturing processes. The model acts as a mediator that converts complex process data into actionable yield insights.
Data Source
AI summary
A method for predicting yield relating to a process of manufacturing semiconductor devices on a substrate, the method including: obtaining a trained first model which translates modeled parameters into a yield parameter, the modeled parameters including: a) a geometrical parameter associated with one or more selected from: a geometric characteristic, dimension or position of a device element manufactured by the process and b) a trained free parameter; obtaining process parameter data including data regarding a process parameter characterizing the process; converting the process parameter data into values of the geometrical parameter; and predicting the yield parameter using the trained first model and the values of the geometrical parameter.