Virtual Metrology for Memory Die Profile Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The manufacturing of memory devices faces challenges due to the high cost and time-consuming nature of extensive testing, which limits the number of test samples and types of tests that can be performed, leading to inefficient quality control and process adjustments.
Innovation Solution
The implementation of machine learning techniques to interpolate virtual test results for devices that have not been tested, based on correlations with other sets of tests, allowing for more rapid and accurate feedback on processing parameters and improving manufacturing efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If extensive testing is performed to improve quality control, then measurement precision and reliability improve, but manufacturing cost and time increase
Solution Approach 1:
The patent creates virtual copies of physical test results through machine learning models. Virtual metrology data and virtual early failure rate data are generated as digital replicas that mimic the outcomes of extensive physical testing, allowing quality control without the time and resource costs of actual extensive testing
Solution Approach 2:
The patent replaces the mechanical testing system with an information-processing system. Instead of physically testing devices to obtain quality data, machine learning models process existing manufacturing data to generate virtual quality metrics, substituting physical measurement with computational prediction
2Measurement precision
If more test samples are tested to improve quality control, then measurement precision improves, but manufacturing cost increases
Solution Approach 1:
Virtual copies of test results are generated through machine learning, eliminating the need to physically test additional samples. The virtual metrology and virtual early failure rate data provide the same quality control information as extensive physical testing would, without the associated costs
Solution Approach 2:
The manufacturing process itself generates the data needed for quality control. Existing manufacturing parameters and process data are reused to train models and generate virtual quality metrics, eliminating the need for separate expensive testing operations
3Reliability
If extensive testing is performed to improve quality control, then reliability improves, but productivity decreases
Solution Approach 1:
Virtual copies of quality assurance data are generated instantaneously through machine learning models. The virtual early failure rate and virtual metrology data provide reliability assessment without requiring time-consuming physical testing, maintaining productivity while ensuring quality
Solution Approach 2:
The machine learning models are trained in advance on historical manufacturing data. Once trained, they can rapidly generate virtual quality metrics for new devices without requiring extensive new testing, enabling real-time quality control that maintains manufacturing throughput
Data Source
AI summary
To provide more test data during the manufacture of non-volatile memories and other integrated circuits, machine learning is used to generate virtual test values. Virtual test results are interpolated for one set of tests for devices on which the test is not performed based on correlations with other sets of tests. In one example, machine learning determines a correlation study between bad block values determined at die sort and photo-limited yield (PLY) values determined inline during processing. The correlation can be applied to interpolate virtual inline PLY data for all of the memory dies, allowing for more rapid feedback on the processing parameters for manufacturing the memory dies and making the manufacturing process more efficient and accurate. In another set of embodiments, the machine learning is used to extrapolate limited metrology (e.g., critical dimension) test data to all of the memory die through interpolated virtual metrology data values.