Hybrid Metrology for Semiconductor Layer Characterization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current metrology techniques for nanosheet and nanowire technology are inadequate for accurately measuring the thickness of each layer, Germanium concentration in Silicon Germanium layers, strain in Silicon and Silicon Germanium layers, and dielectric properties of Nitride layers, often requiring multiple techniques and sacrificing accuracy or throughput.
Innovation Solution
The equation-based hybridization technique uses the raw signal from a first tool, such as an X-ray tool, as a constraint for a second tool, like an optical scatterometer, to improve measurement accuracy without reducing throughput, by employing raw signal data as a boundary condition for calculations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single metrology technique is used, then throughput is maintained, but measurement accuracy and precision are insufficient for multiple parameters
Solution Approach 1:
The patent combines data from multiple metrology tools (scatterometry tool and spectroscopic ellipsometry tool) into a unified hybridization model. The scatterometry raw signal serves as a constraint for the ellipsometry calculation, merging the strengths of both techniques to achieve accurate multi-parameter measurement without requiring sequential measurement of each parameter separately, thereby maintaining high throughput while improving measurement precision.
2Measurement precision
If multiple metrology techniques are used sequentially, then measurement accuracy improves, but throughput decreases due to extended measurement time
Solution Approach 1:
The patent performs preliminary action by obtaining the scatterometry raw signal first and using it as a pre-determined constraint for the subsequent ellipsometry calculation. This preliminary use of the first measurement tool's data eliminates the need for iterative adjustments and reduces the time required for the second measurement tool, thereby reducing total measurement time while maintaining the accuracy benefits of using multiple techniques.
3Measurement precision
If traditional hybridization methods are used, then some measurement parameters are obtained, but the goodness of fit and overall accuracy are insufficient
Solution Approach 1:
The patent implements feedback by using the scatterometry raw signal as a constraint that feeds into the ellipsometry calculation model. This feedback mechanism ensures that the hybridization model consistently converges to accurate solutions with high goodness of fit (0.99 or better). The feedback loop validates the model parameters against the actual measured signal, improving accuracy while the automated computational process manages the complexity of the hybridization approach.
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 enhances measurement accuracy and throughput, providing precise results for critical parameters in next-generation device architectures, with a goodness of fit of 0.99, effectively addressing the limitations of traditional hybridization methods.
Implementation Method 1
receiving, using a processor, a raw signal from a first tool representing a measured parameter of a semiconductor device
Implementation Method 2
receiving, using a processor, a raw signal from a first tool representing a measured parameter of a semiconductor device
Implementation Method 3
receiving, using the processor, data on the measured parameter from a second tool
Data Source
AI summary
A computer-implemented method for measuring a parameter of a semiconductor. A non-limiting example of the computer-implemented method includes receiving, using a processor, a raw signal from a first tool representing a measured parameter of a semiconductor device. The method also receives, using the processor, data on the measured parameter from a second tool, and calculates, using the processor, the measured parameter based on the data received from the second tool and on a constraint based on the raw signal from the first tool.