Reconstructing Semiconductor Component Positions on Wafers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In semiconductor component packaging, the traceability of components to their original wafer positions is lost after cutting, making it difficult to assign semiconductor components to their original wafers, especially for PowerMOS components without memory, leading to a combinatorial problem with factorial complexity.
Innovation Solution
A method using a machine learning system to determine an assignment rule between wafer-level test results and final test results without requiring unique identifiers, by creating a dataset, training a regressor, calculating a cost matrix, and optimizing the assignment rule to minimize costs, allowing for a one-to-one assignment and reconstruction of component positions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If wafer cutting and packaging are performed, then semiconductor components can be separated and packaged, but traceability of components to original wafer positions is lost
Solution Approach 1:
The patent applies preliminary action by performing wafer-level tests before cutting and packaging, capturing test results and position data while the components are still on the wafer. This preliminary data collection enables later reconstruction of component positions without requiring physical markings or unique identifiers on the components themselves.
Solution Approach 2:
The patent creates a digital copy of the wafer layout and test results, storing position information and test data as digital records. This digital copying allows the system to reconstruct original positions from test data alone, eliminating the need for physical traceability markers on the semiconductor components.
2Loss of information
If unique identifiers are added to enable traceability, then component assignment to wafers becomes possible, but the complexity and cost of the system increases
Solution Approach 1:
The patent extracts the traceability function from the physical components by using external digital records. Instead of embedding unique identifiers in the components, the system extracts position and test data from the wafer-level testing process and stores it digitally, separating the traceability function from the semiconductor components themselves.
Solution Approach 2:
The patent introduces test results and position data as intermediary information between the wafer and the packaged components. These intermediary records serve as the linkage that enables traceability without requiring direct modification of the semiconductor components or addition of unique identifiers.
3Loss of information
If rough matching between loose components and wafer components is performed, then some traceability is maintained, but several thousand components cannot be assigned to multiple wafers
Solution Approach 1:
The patent segments the traceability problem into individual component-level records by maintaining separate position and test result data for each component on the wafer. This segmentation enables precise one-to-one matching between components and their original positions, allowing accurate reconstruction even when components are separated during packaging.
Solution Approach 2:
The patent uses feedback from wafer-level test results to reconstruct component positions and assign them to packaged components. The test results provide feedback information that enables the system to accurately match components to their original locations, improving assignment accuracy beyond rough matching.
Data Source
AI summary
A method determines an assignment rule in order to combine test results from different tests of the same semiconductor device. The method includes fitting a model, such as a linear regression model, using the model to predict the test data, calculating a cost matrix based on the predictions, and applying the Hungarian method to the cost matrix to obtain a new assignment rule and repeating these steps multiple times.