Semiconductor Measurement Deviation Root-Cause Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Identifying the root cause of measurement deviations in semiconductor specimens with multi-layer structures is challenging due to the numerous unknown structural properties of each layer, making it difficult to correlate these properties with resulting signals, and destructive tools like TEM are often required for analysis.
Innovation Solution
A computerized system and method that uses a simulation model to analyze actual and additional measurements across varying scanning parameters, mapping these measurements in a multi-dimensional property space to identify the root cause of deviations without destructive tools.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If destructive tools like TEM are used to analyze measurement deviations, then accurate identification of structural properties is achieved, but the specimen is damaged and cannot be used for further examination
Solution Approach 1:
The patent creates a simulation model that replicates the physical measurement process and generates simulated measurements corresponding to different structural properties. This virtual copy allows analysis of measurement deviations without physically damaging the specimen, resolving the contradiction between accurate identification and specimen integrity
Solution Approach 2:
The patent introduces a simulation model as an intermediary between the actual measurements and the structural properties. This intermediary translates measured signals into inferred structural properties through virtual experimentation, enabling non-destructive analysis while maintaining measurement accuracy
2Measurement precision
If multiple examination steps are performed to identify root cause of deviations, then measurement precision is improved, but examination time and complexity increase
Solution Approach 1:
The patent pre-generates a simulation model with simulated measurements for various structural properties before actual examination. This preliminary preparation enables rapid comparison with actual measurements during examination, achieving accurate root cause identification without time-consuming iterative physical experiments
Solution Approach 2:
The simulation model serves multiple functions simultaneously: it generates expected measurements, identifies deviations, infers structural properties, and determines root causes. This multi-functionality consolidates multiple examination steps into a single integrated process, reducing overall examination time while maintaining precision
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
Enables accurate identification of the root cause of measurement deviations in semiconductor specimens, improving process control and yield by correlating structural properties with measurement signals, thereby enhancing fabrication precision and reliability.
Implementation Method 1
The system further comprises an electron beam tool operatively connected to the PMC and configured to scan the semiconductor specimen using the set of scanning parameters with predefined values and generate image data representative thereof by collecting backscattered electrons emitted from the specimen
Data Source
AI summary
There is provided a system and method of examining a specimen. The method comprises obtaining an actual measurement in response to scanning the specimen using a set of scanning parameters with predefined values; obtaining a simulated measurement based on design data; comparing the actual measurement with the simulated measurement to identify a deviation with respect to a predefined tolerance; identifying at least one structural property as root cause of the deviation by: obtaining one or more additional actual measurements in response to scanning the specimen using one or more varying values of at least one scanning parameter; and providing the actual measurement and the additional actual measurements to a simulation model representative of simulated measurement distribution in a multi-dimensional property space characterized by the structural properties of the plurality of layers and the at least one scanning parameter, thereby identifying the at least one structural property.