Off-axis aberration estimation in an EUV microscope using natural speckle

Aamod Shanker, Antoine Wojdyla, Gautam Gunjala, Jonathan Dong, Markus P. Benk, A. R. Neureuther, Kenneth A. Goldberg, Laura Waller2016 · DOI

TL;DR

This paper presents a novel method to estimate aberrations in EUV microscopes by analyzing the natural speckle patterns generated by surface roughness on a mask blank. By leveraging the phase-to-intensity transfer function, they can directly recover the microscope's pupil function, offering a powerful tool for system calibration and characterization.

What Problem Does This Paper Address?

Extreme ultraviolet (EUV) lithography is crucial for manufacturing advanced semiconductor chips, but it operates at very short wavelengths (around 13.5 nm). At these wavelengths, even seemingly smooth surfaces exhibit roughness at the scale of the wavelength, leading to a phenomenon called "speckle" in the resulting images. This speckle, while often considered a nuisance, can actually contain valuable information about the imaging system itself. Specifically, understanding and quantifying aberrations (imperfections in the optical system that distort the image) in EUV microscopes is critical for achieving high-resolution imaging and accurate lithography. Traditional methods for aberration estimation can be complex and time-consuming. This paper tackles the challenge of efficiently and directly estimating these aberrations by harnessing the information encoded within the natural speckle patterns.

Methods

The core of this paper lies in the application of the "phase-to-intensity transfer function" (PITF) theory. For weak phase objects (like EUV lithography masks where phase variations are small), the intensity of the speckle image in the frequency domain is directly related to the system's pupil function and the object's phase spectrum. The authors derive this relationship, showing that the speckle spectrum essentially "images" the PITF.

The key steps involve:

  1. Modeling the object: A thin object with both absorption and phase variations is considered, with the phase component being the source of speckle.
  2. Image formation: The process of image formation in a microscope, including the effect of the pupil function and off-axis illumination, is described. Off-axis illumination causes the object spectrum to shift relative to the pupil.
  3. Deriving the PITF: Through Fourier transforms and approximations for weak phase objects, the authors derive the PITF, which links the object's phase spectrum to the image intensity spectrum.
  4. Speckle spectrum analysis: For broadband speckle, the object's phase spectrum is assumed to be uniformly distributed. This simplifies the relationship, allowing the speckle intensity spectrum to directly reveal the PITF.
  5. Pupil recovery: By analyzing the morphology of the speckle spectrum, the authors demonstrate how to recover the microscope's pupil function, which is directly related to the system's aberrations.

The method is applied to the SHARP EUV microscope, a reflection geometry microscope. They show that by analyzing the speckle patterns from a mask blank, they can recover the pupil function, which in this case revealed a defocus aberration.

Datasets

The paper uses a "mask blank" as the object to generate natural speckle. This is not a pre-existing dataset in the traditional sense but rather a physical object used within the experimental setup. The "dataset" is the speckle intensity images captured by the SHARP EUV microscope. The paper also includes simulations to validate their findings, using synthetic pupil functions and object phase spectra.

Key Insights and Ideas

  • Speckle as a diagnostic tool: The paper highlights that speckle, often seen as an artifact, can be a powerful source of information for characterizing optical systems.
  • Direct aberration estimation: The PITF provides a direct link between the speckle intensity spectrum and the system's aberrations, bypassing the need for more complex phase retrieval algorithms in certain cases.
  • Off-axis illumination advantage: Off-axis illumination, while shifting the object spectrum, is crucial for separating the pupil and its conjugate in the speckle spectrum, enabling the recovery of the transfer function.
  • Reflection microscopy specificities: The paper notes that the reflection geometry of the SHARP microscope leads to specific aberration characteristics, such as a defocus normal to the incident beam.

Key Results

The authors successfully demonstrated the estimation of aberrations in the SHARP EUV microscope using natural speckle. They showed that the speckle spectrum directly images the phase-to-intensity transfer function. Specifically, they were able to:

  • Recover the pupil function of the microscope from the speckle spectrum.
  • Identify a defocus aberration in the system.
  • Show that the morphology of the speckle spectrum changes with illumination angle, reflecting the corresponding shift in the pupil function.
  • The recovered pupil can be used for tool calibration and potentially for object phase recovery.

Limitations and Future Work

  • Weak phase object approximation: The method relies on the weak phase object approximation, which may limit its applicability to objects with significant phase variations.
  • Broadband speckle assumption: The direct imaging of the PITF assumes broadband speckle, where the object phase spectrum is uniformly distributed.
  • Higher-order aberrations: While the paper suggests that more rigorous fitting can estimate higher-order aberrations, this was not fully explored in the presented work.
  • Future work: The authors propose extending this method to estimate aberrations over the full 0 to 2π range of azimuthal illumination angles, similar to diffraction tableaux in electron microscopy. This could lead to comprehensive characterization of illumination-dependent aberrations and improved system alignment. The method is also seen as valuable for techniques like Fourier Ptychography, which require accurate pupil estimates at various illumination angles.

Summary updated

Abstract

Surface roughness on a flat object causes natural speckle when imaged by an extreme ultraviolet (EUV) microscope under sufficient coherence. Using a phase-to-intensity transfer function theory, direct estimation of aberrations from the spectrum of the speckle intensity is demonstrated for various illumination angles.
BibTeX
@misc{shanker2016off,
  author = {Aamod Shanker and Antoine Wojdyla and Gautam Gunjala and Jonathan Dong and Markus P. Benk and A. R. Neureuther and Kenneth A. Goldberg and Laura Waller},
  title = {{Off-axis aberration estimation in an EUV microscope using natural speckle}},
  year = {2016},
  doi = {10.1364/isa.2016.ith1f.2},
}

Loading related content...

Figure