Measurement for high aspect ratio trenches based on inversion of pupil wavefront
Year: 2026
Type: article
Abstract: High-aspect-ratio (HAR) trench structures are critical components in advanced semiconductor and integrated circuit manufacturing. However, achieving rapid, non-destructive, and high-precision measurement of their morphological parameters (depth and linewidth) remains a significant challenge. This study proposes a non-destructive metrology method based on the inversion of pupil wavefront (IPW). To overcome the computational complexity of 3D physical modeling and enhance inversion efficiency, the physical interaction process is rigorously simplified into a 2D waveguide model by leveraging the translational symmetry of the trench structure. This approach introduces a unique orthogonal decoupling strategy: the unmodulated wavefront in the YOZ plane (parallel to the trench) is utilized for focus detection algorithm to achieve precise focusing at the bottom of trench; subsequently, the structurally modulated wavefront in the XOZ plane (vertical to the trench) is then extracted to retrieve the linewidth and depth by matching a pre-computed database. By implementing phase-shifting interferometry within a Linnik interference system, the pupil wavefront is accurately reconstructed. Experimental validations are conducted on four typical HAR trenches with nominal morphological parameters ranging from 5 to 10 μm and 100 to 300 μm (linewidth and depth). The results demonstrate that the values acquired by IPW are in good consistency with scanning electron microscopy (SEM) benchmarks, with absolute deviations for both linewidth and depth strictly controlled within 0.02 ∼ 0.08 μm. By avoiding extensive mechanical scanning, the proposed IPW method compresses the measurement period to a few seconds, significantly improving the detection efficiency for complex HAR microstructures.
Source: Optics & Laser Technology
Institution: Nanjing University of Science and Technology
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $3,230
OpenAlex ID: W7221950999
