Cross-Wafer Model Constraining Semiconductor Metrology Parameters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current optical metrology systems face challenges in accurately measuring structural parameters of semiconductor devices due to increasing complexity and miniaturization, leading to high parameter correlation, increased measurement time, and reduced illumination intensity, especially when dealing with high-aspect ratio three-dimensional structures and opaque materials.
Innovation Solution
The development of an optimized measurement model that constrains parameter variations across a semiconductor wafer using a cross-wafer model, reducing parameter correlation and enabling more accurate measurements with fewer technologies and reduced wavelength ranges, by characterizing specimen parameters as functions of location and incorporating process-induced spatial patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If measurements are performed over a large range of machine parameters (wavelength, azimuth, angle of incidence), then measurement coverage and characterization capability are improved, but measurement time and computation time increase significantly
Solution Approach 1:
The patent applies partial action by selecting only the most informative subset of machine parameters for measurement. Instead of measuring across the full range of available parameters, the system identifies and measures only at specific wavelengths and angles that provide maximum information for the given target structure, thereby reducing measurement time while maintaining characterization capability
Solution Approach 2:
The patent uses preliminary action by pre-calculating and storing measurement models for different target configurations before actual measurement. The system prepares lookup tables and reference data in advance, allowing rapid comparison with actual measurements without performing full computations during the measurement process itself
2Adaptability or versatility
If measurements are performed over a large range of machine parameters, then characterization capability is improved, but illumination intensity at any particular wavelength decreases
Solution Approach 1:
The patent applies local quality by concentrating measurement resources at specific wavelengths and angles where they provide maximum information. Instead of distributing illumination intensity uniformly across a broad spectrum, the system intensifies illumination at selected wavelengths that are most sensitive to the target structure's characteristics, improving signal quality without increasing total energy consumption
3Adaptability or versatility
If the number of parameters required to characterize complex structures increases, then measurement comprehensiveness is improved, but parameter correlation increases and measurement reliability decreases
Solution Approach 1:
The patent applies segmentation by dividing the complex measurement problem into separate, independent sub-problems. The measurement process is split into multiple stages: first measuring certain parameters at specific conditions, then using those results to inform subsequent measurements of other parameters. This segmentation reduces parameter correlation by ensuring that each measurement is optimized for its specific parameter rather than trying to measure all parameters simultaneously
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
This approach results in more accurate and efficient measurements with reduced computation time, enabling the use of high-intensity light sources within limited wavelength ranges and improving the reliability of structural and material characteristic assessments in semiconductor fabrication processes.
Implementation Method 1
optical radiation to penetrate to the bottom layers
Implementation Method 2
scatterometry and reflectometry implementations
Implementation Method 3
scatterometry and reflectometry implementations
Data Source
AI summary
An optimized measurement model is determined based a model of parameter variations across a semiconductor wafer. A global, cross-wafer model characterizes a structural parameter as a function of location on the wafer. A measurement model is optimized by constraining the measurement model with the cross-wafer model of process variations. In some examples, the cross-wafer model is itself a parameterized model. However, the cross-wafer model characterizes the values of a structural parameter at any location on the wafer with far fewer parameters than a measurement model that treats the structural parameter as unknown at every location. In some examples, the cross-wafer model gives rise to constraints among unknown structural parameter values based on location on the wafer. In one example, the cross-wafer model relates the values of structural parameters associated with groups of measurement sites based on their location on the wafer.