Optimizing Hybrid Metrology: Rigorous Implementation of Bayesian and Combined Regression

Area of Science:

• Metrology
• Semiconductor Manufacturing
• Data Analysis

Background:

• Hybrid metrology combines diverse measurement techniques for enhanced critical dimension determination.
• The semiconductor industry increasingly relies on hybrid metrology for accurate 3-D structure characterization.
• Accurate uncertainty estimation is crucial for reliable metrology results.

Purpose of the Study:

• To explore the benefits and challenges of hybrid metrology in the semiconductor industry.
• To present a hybrid metrology approach combining optical critical dimension (OCD) and scanning electron microscope (SEM) measurements.
• To address error analysis challenges, particularly the impact of correlated errors on measurement uncertainty.

Main Methods:

• Integration of optical critical dimension (OCD) and scanning electron microscope (SEM) measurement data.
• Development of methodologies for error analysis in hybrid metrology.
• Utilizing hypothetical examples and real measurement data to validate the approach.

Main Results:

• Demonstration of hybrid metrology's potential for feasible 3-D attribute measurements.
• Identification of challenges in error analysis, especially concerning correlated errors and the chi-squared function.
• Illustration of solutions for comparing results from different instrument models.

Conclusions:

• Hybrid metrology offers improved quantitative characterization and uncertainty estimation for 3-D structures.
• Addressing systematic and correlated errors is essential for robust hybrid metrology.
• The presented methods provide solutions for reliable data integration and analysis in hybrid metrology.