Diffraction amplification enables imaging- and lens-free defect detection.

Area of Science:

• Optics and Photonics
• Materials Science
• Computational Imaging

Background:

• Conventional optical defect detection often relies on complex imaging systems.
• Subtle defects can be challenging to detect due to low contrast and noise.
• Existing methods may require specialized equipment, phase retrieval, or annotated data.

Purpose of the Study:

• To develop a lens-free optical detection method leveraging natural signal amplification.
• To establish a direct pathway from wavefront modulation to defect detection.
• To create a simple, low-cost, and highly reliable defect inspection framework.

Main Methods:

• Utilizing Fresnel diffraction patterns as the primary detection domain.
• Employing a single inverse diffraction step for computational localization of defects.
• Analyzing diffraction phenomena directly, bypassing conventional image formation.

Main Results:

• Demonstrated reliable detection of wavelength-scale defects on semiconductor wafers and display panels.
• Achieved high sensitivity under strong noise (σ²=0.05) and extremely low contrast (C=0.005).
• Showcased potential for spectral scalability across visible, near-infrared, and extreme ultraviolet ranges.

Conclusions:

• Diffraction saliency offers a scalable framework for optical defect detection.
• The proposed method enables simple hardware implementation for inline inspection in complex environments.
• This lens-free, computation-driven approach bypasses traditional imaging limitations for enhanced reliability and speed.