TL;DR
This paper introduces a novel, non-invasive method to precisely characterize the aberrations of extreme ultraviolet (EUV) microscopes using readily available photomask blanks. This technique allows for in-situ evaluation of imaging system performance, achieving high accuracy and paving the way for better quality control in advanced lithography.
What Problem Does This Paper Address?
Characterizing the performance of imaging systems, especially in specialized regimes like extreme ultraviolet (EUV) lithography, is crucial for ensuring high-quality imaging and manufacturing. Traditional methods for measuring aberrations often involve complex, expensive hardware (like interferometers or adaptive optics) or require custom-fabricated, high-fidelity test objects. These approaches are often impractical or too costly for in-situ measurements, particularly in demanding environments like ultra-high vacuum systems used in semiconductor manufacturing. The challenge is to find a simple, cost-effective, and non-invasive way to accurately assess the aberrations across the entire field of view of an EUV microscope.
Methods
The core of this work is a computational procedure that leverages the inherent, atomic-scale surface roughness of standard EUV photomask blanks. Instead of using precisely engineered test patterns, the authors treat the photomask's random surface as a "weak phase object" that generates speckle patterns when illuminated coherently. By acquiring multiple images of this speckle pattern under varying, steerable illumination angles, the system's aberrations can be computationally reconstructed.
The key steps involve:
- Speckle Generation: Coherently illuminating a blank EUV photomask with plane waves at different angles.
- Image Acquisition: Capturing multiple in-focus images of the resulting speckle pattern, along with one out-of-focus image for initial calibration.
- Fourier Transform Analysis: Computing the Fourier transform of the acquired images. The interference between the object's Fourier spectrum and the system's pupil function creates characteristic patterns.
- Aberration Reconstruction: Using a mathematical model that relates the illumination angles, the system's pupil function (including aberrations), and the measured Fourier spectra. A gradient descent optimization algorithm is employed to iteratively solve for the wavefront error function (WEF), which describes the system's aberrations.
- Illumination Wavefront Recovery (Extension): The method can be extended to also reconstruct the illumination wavefront, which is important for understanding the light source itself.
This approach is novel because it avoids the need for specialized test objects and complex optical setups, relying instead on the intrinsic properties of a standard component and sophisticated computational analysis.
Datasets
The primary "dataset" used in this study consists of ten coherently illuminated images of a blank EUV photomask acquired using the SHARP EUV microscope. One image was taken with central illumination and defocus for initial calibration, while the other nine were acquired with varying illumination angles. These images were then processed into numerous sub-regions for detailed analysis. No new benchmark datasets were introduced; rather, the study utilized a readily available industrial component as the test object.
Key Insights and Ideas
- Photomask Roughness as a Universal Probe: The inherent, random surface roughness of standard photomask blanks is sufficient to act as a highly effective, self-calibrating probe for characterizing imaging system aberrations. This eliminates the need for custom test objects.
- In-Situ, Non-Invasive Characterization: The method is entirely computational and requires only standard imaging data, making it ideal for in-situ measurements in demanding environments like ultra-high vacuum, where invasive techniques are difficult.
- Field-Dependent Aberration Mapping: The technique allows for the mapping of aberrations across the entire field of view, revealing how image quality degrades away from the center. This is critical for understanding the practical usable area of a microscope.
- Illumination Wavefront Recovery Potential: The framework can be extended to characterize not only the microscope's aberrations but also the wavefront of the illumination source, providing a more complete picture of the imaging system's performance.
Key Results
- High Accuracy Aberration Measurement: The method achieved a measurement accuracy of better than $\lambda/182$ for wavefront errors.
- Minimal Aberrations in the Sweet Spot: The SHARP microscope demonstrated minimal aberrations, with a wavefront error magnitude of less than $\lambda/21$ rms averaged over the central 5 µm x 5 µm region of the 30 µm field of view.
- Consistency with Theory: The observed field variation of aberrations was consistent with theoretical predictions for a single-lens system, primarily attributed to Petzval curvature.
- Practical Usability: The results suggest that the method is robust enough to characterize the performance of next-generation industrial EUV microscopes.
Limitations and Future Work
- Telecentricity Assumption: The current method is directly applicable to telecentric imaging systems. For non-telecentric systems like SHARP, the local angle estimates are influenced by variations in the chief ray angle across the field of view. Accounting for this non-telecentricity is identified as a subject for future work.
- Illumination Angle Estimation Ambiguity: While the sign ambiguity of illumination angles does not affect aberration recovery, it needs to be resolved for illumination wavefront reconstruction. The paper proposes a method for disambiguation using pairs of images with a known relative change in illumination angle.
- Computational Intensity: While non-invasive, the computational reconstruction process can be intensive, requiring optimization algorithms.
Future work will focus on extending the method to fully account for non-telecentric systems and further refining the illumination wavefront reconstruction capabilities.
Summary updated