Semiconductor Binning Feed-Forward Adjustment for Metallization Layers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Semiconductor manufacturing process variations often result in chips performing below designed standards, leading to inconsistent performance levels and reduced profitability due to inefficient binning practices, where chips are sorted into lower performance bins and sold at lower prices.
Innovation Solution
A machine learning model is developed to determine correlations between metallization layer process parameters and measurement/test data, allowing for adjustments to be made to subsequent metallization layers to improve chip performance, thereby upgrading chips to higher performance bins and increasing profitability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional binning sorting is performed based on final chip performance measurements, then chips are sorted into performance bins, but many chips end up in lower performance bins and are sold at lower prices, reducing manufacturer profitability
Solution Approach 1:
The system performs preliminary analysis of process parameters during metallization layer fabrication to predict final chip performance before completion. This allows identification of chips likely to underperform early in the manufacturing process, enabling targeted adjustments to be made during subsequent processing steps to improve their final performance and bin assignment.
Solution Approach 2:
The system dynamically adjusts process parameters of subsequent metallization layers based on predicted performance deficiencies. By changing parameters such as deposition conditions, patterning parameters, or etch conditions in response to early performance indicators, the system compensates for variations that would otherwise lead to lower performance bins.
2Manufacturing precision
If process parameters are adjusted dynamically during fabrication based on real-time data, then chip performance can be improved, but the complexity of the manufacturing process increases
Solution Approach 1:
The system implements a feedback loop where measurement and test data from intermediate process steps are continuously analyzed, and process parameters are adjusted accordingly. Machine learning models predict final performance based on current process data, and this information feeds back to control systems that modify subsequent processing parameters to compensate for detected variations and achieve target performance levels.
3Measurement precision
If more measurement and testing operations are performed during fabrication, then performance prediction accuracy improves, but manufacturing time and cost increase
Solution Approach 1:
The system performs preliminary measurements and testing at intermediate stages of fabrication rather than waiting for final chip completion. These early measurements provide sufficient data for machine learning models to predict final performance with high accuracy, allowing the system to identify performance issues before all manufacturing steps are complete, thereby reducing the need for extensive final testing.
Data Source
AI summary
One or more processors determine a predicted sorting bin of a semiconductor device, based on measurement and test data performed on the semiconductor device subsequent to a current metallization layer. A current predicted sorting bin and a target sorting bin are determined by a machine learning model for the semiconductor device; the target bin include higher performance semiconductor devices than the predicted sorting bin. The model determines a performance level improvement attainable by adjustments made to process parameters of subsequent metallization layers of the semiconductor device. Adjustments to process parameters are generated, based on measurement and test data of the current metallization layer of semiconductor device, and the adjustment outputs for the process parameters of the subsequent metallization layers of the semiconductor device are made available to the one or more subsequent metallization layer processes by a feed-forward mechanism.